Академический Документы
Профессиональный Документы
Культура Документы
51
October 2010
10.3310/hta14510
Additionally the HTA website allows you to either print out your order or download a blank order form.
Payment methods
Paying by cheque
If you pay by cheque, the cheque must be in pounds sterling, made payable to University of Southampton
and drawn on a bank with a UK address.
Paying by credit card
You can order using your credit card by phone, fax or post.
Subscriptions
NHS libraries can subscribe free of charge. Public libraries can subscribe at a reduced cost of £100 for
each volume (normally comprising 40–50 titles). The commercial subscription rate is £400 per volume
(addresses within the UK) and £600 per volume (addresses outside the UK). Please see our website for
details. Subscriptions can be purchased only for the current or forthcoming volume.
The website also provides information about the HTA programme and lists the membership of the v arious
committees.
Systematic review of the links between
human resource management
practices and performance
*Corresponding author
Patterson M, Rick J, Wood S, Carroll C, Balain S, Booth A. Systematic review of the links
between human resource management practices and performance. Health Technol Assess
2010;14(51).
Abstract
Systematic review of the links between human
resource management practices and performance
M Patterson,1* J Rick,2 S Wood,3 C Carroll,4 S Balain5 and A Booth4
1
Institute of Work Psychology, University of Sheffield, Sheffield, UK
2
Health Services Research Group – Primary Care, University of Manchester, Manchester, UK
3
School of Management, University of Leicester, Leicester, UK
4
School of Health and Related Research (ScHARR), University of Sheffield, Sheffield, UK
5
Management School, Lancaster University, Lancaster, UK
*Corresponding author
Background: In recent years human resource were derived from the NHS Performance Indicators
management (HRM) has been seen as an important and the NHS Improvement Plan. The quality criteria
factor in the successful realisation of organisational used to select papers incorporated a longitudinal
change programmes. The UK NHS is undergoing study design filter to provide evidence of the causal
substantial organisational change and there is a need to direction of relationships between HRM and relevant
establish which human resource (HR) initiatives may be outcomes. Single HRM practices were considered.
most effective. Within the health-specific literature, focus was on the
Objectives: To assess the results from a wide-ranging impact of HRM on patient outcomes. Information is
series of systematic reviews of the evidence on HRM presented on the reliability of measures in each of the
and performance. The first part assesses evidence intermediate outcome areas.
on use of HRM in the UK and fidelity of practice Results: Work design practices that enhance
implemented. The second part considers evidence employee autonomy and control influenced a number
for the impact of HRM practices on intermediate of outcomes and there was consistent evidence
outcomes, which can impact on final outcomes, such for the positive impact of increased job control
as organisational performance or patient care. on employee outcomes, such as job satisfaction,
Data sources: The following databases were absence and health. For employee participation, the
searched: Applied Social Sciences Index and Abstracts small number of studies reviewed supported the
(ASSIA), British Nursing Index (BNI), Business Source involvement of employees in design/implementation
Premier, Campbell Collaboration, Cochrane Central of changes that affect their work. In health literature
Register of Controlled Trials (CENTRAL), Cochrane in particular, employee involvement through quality
Database of Systematic Reviews (CDSR), Cumulative improvement teams resulted in improved patient
Index to Nursing and Allied Health Literature outcomes. Findings were positive for the impact of
(CINAHL), Database of Abstracts of Reviews of training on the intended outcomes of the initiatives.
Effectiveness (DARE), DH-Data, EMBASE, Health Support for the impact of performance management
Management Information Consortium (HMIC), practices was apparent, in particular feedback on
International Bibliography of the Social Sciences (IBSS), performance outcomes and the use of participative
King’s Fund database, MEDLINE, NHS Economic goal setting. Strong associations were found among all
Evaluation Database (NHS EED), National Research intermediate outcomes, and the relationship between
Register (NRR), PREMEDLINE, PsycINFO, ReFeR, most intermediate behaviours and outcomes were
Social Sciences Citation Index (SSCI) and Science significant.
Citation Index (SCI). The searches were conducted in Limitations: Limited evidence was available on the
May/June 2006. use of HRM and on the implementation of policy.
Review methods: Broad categories of HRM Also, the specific practices studied within each HRM
interventions and intermediate outcomes were category differ so there was little evidence to show
generated: 10 HRM categories and 12 intermediate whether similar practices have the same effects in
outcome categories. Seven patient final outcomes health and non-health settings.
iii
Conclusions: Some potentially effective practices is highlighted. However, future multilevel studies that
for both health and non-health areas were identified, embrace the individual, team and organisational level
and HRM methods could be used to support change are needed. Studies should look into interventions
processes within the NHS; the findings relating to aimed at improving HR outcomes and performance,
work organisation are particularly promising with and allow for pre- and post-intervention measurement
regard to changes in methods of service delivery. of practices and outcomes.
Using training to support the implementation of change
iv
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Contents
List of abbreviations ................................. vii Definitions of selected intermediate
outcomes ................................................. 69
Executive summary .................................. ix Distinguishing final outcomes and
intermediate outcomes ........................... 74
1 Introduction .............................................. 1
Context for the review: overview of the 6 The impact of HRM practices on
literature on HRM and organisational intermediate outcomes among health
performance ........................................... 1 professionals .............................................. 75
Review aim, scope and objectives ............... 4 Introduction ............................................... 75
Report ......................................................... 4 Review methodology .................................. 75
Results ......................................................... 76
2 Methodology ............................................. 7 Conclusions ................................................. 82
Introduction ............................................... 7
Applying systematic review methodology 7 The impact of HRM practices on final
within the HRM literature ..................... 7 outcomes in the health sector ................ 85
Overview of systematic review Introduction ............................................... 85
methodology .......................................... 8 Patient outcomes ......................................... 85
Applying the systematic review Review methodology .................................. 87
methodology .......................................... 8 Results ......................................................... 88
Summary of specific research questions Conclusions ................................................. 124
and methodological approaches ............ 14
8 Impact of HRM practices in non-health
3 HRM practices and their reported use settings ....................................................... 129
in the UK ................................................... 17 Introduction ............................................... 129
Introduction ............................................... 17 Review methodology .................................. 129
What HRM practices are described in the Conclusions ................................................. 207
literature? ............................................... 17
How widespread is the use of these HRM 9 Intermediate and final outcomes:
practices in the UK? ............................... 19 correlations and effects ........................... 209
Summary .................................................... 21 Introduction ............................................... 209
Conclusion .................................................. 49 Review methodology .................................. 209
Findings for Objective 3.1 .......................... 211
4 Implementation fidelity in HRM ............. 51 Findings for Objective 4.2 .......................... 217
Introduction ............................................... 51 Findings for Objective 4.1 .......................... 218
Policy–practice link and implementation Conclusions ................................................. 222
fidelity ..................................................... 51
Dimensions of fidelity ................................. 52 10 Measuring the intermediate outcomes
Overview of findings from fidelity studies of HRM: the reliability of measures ........ 223
in social policy ........................................ 52 Introduction ............................................... 223
Fidelity of HR policy .................................. 59 What is measurement reliability? ............... 223
A framework for implementation fidelity ... 64 Which intermediate outcomes is it valid
Conclusion .................................................. 65 to measure in this context? .................... 224
Reliability of intermediate outcome
5 Identifying linkages between HRM measures ................................................. 225
practices and performance outcomes ... 67 Conclusions ................................................. 234
Introduction ............................................... 67
Intermediate outcomes linking HRM and 11 Conclusions and recommendations ....... 237
performance ........................................... 67 Introduction ............................................... 237
Selecting intermediate outcomes ............... 68 Systematic review methodology .................. 237
v
vi
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
List of abbreviations
All abbreviations that have been used in this report are listed here unless the abbreviation is well
known (e.g. NHS), or it has been used only once, or it is a non-standard abbreviation used only in
figures/tables/appendices, in which case the abbreviation is defined in the figure legend or in the
notes at the end of the table.
vii
Executive summary
Context for this review longitudinal study design filter, as this may
provide evidence about the causal direction
The UK NHS is the largest employer in Europe, of relationships between HRM and relevant
involving complex and diverse services and work outcomes in a way that cross-sectional data
roles. The NHS continues to undergo substantial cannot.
organisational change, as indicated by increased • The review considers single HRM practices
emphasis on performance targets representing and is not confined to collectivities of them,
drivers at local, national and clinical level. or ‘bundles’ as they are known in the high-
performance management literature.
Human resource management (HRM) is being • The review covers issues around the
seen as a vital element in the successful realisation implementation of HRM in practice and
of these change programmes and is being given the measurement of relevant intermediate
a greater prominence than it has traditionally. It outcomes in the HRM performance chain.
is therefore timely to assess the evidence we have • Within the health-specific literature, the review
on what human resource (HR) initiatives are most is focused, in particular, on the impact of HRM
effective. on patient outcomes.
No single HRM practices or bundle of practices conclude that each of the intermediate outcomes
were found to be a panacea. However, our review identified in this review may contribute uniquely
does enable us to identify some potentially effective to efforts to understand and manage employee
practices for both health and non-health areas. behaviours.
• In the area of work design, practices that The review also explored the correlations between
enhance employee autonomy and control are intermediate outcomes and productivity-enhancing
influential in relation to a number of outcomes behaviours (e.g. individual job performance,
and there is consistent evidence for the positive employee turnover). The relationships between
impact of increased job control (in various most intermediate outcomes and behaviours
forms) on employee outcomes such as job were significant and of small to moderate
satisfaction, absence and health. strength. The premise here is that intermediate
• In the parallel field of employee participation, outcomes are determinants of salient employee
the small number of studies reviewed here behaviours, which, in turn, enhance organisational
supports the widely advocated principle performance. These data do not prove a causal link
of involving employees in the design and but do demonstrate associations.
implementation of changes (e.g. job redesign)
that affect their work. Specifically in the health Impact of intermediate
literature, employee involvement through outcomes on final outcomes
quality improvement teams was found to
be effective in terms of improved patient This review was unable to identify any longitudinal
outcomes. evidence to assess whether intermediate outcomes,
• In the area of training, findings are consistently such as job satisfaction or burnout, impact
positive for the impact of training on the on patient-care outcomes. In the non-health
specific intended outcomes of the training field, a small number of longitudinal studies
initiatives. were identified that examined the impact of
• Support for the impact of performance intermediate outcomes (mostly average employee
management practices is found and job satisfaction) on organisational performance.
particularly the importance of feedback While the studies in this review show associations,
on performance outcomes and the use of the evidence on the casual direction of this
participative goal setting. relationship is mixed. This relationship is a
crucial link for the premise that HRM influences
Such evidence points to the HRM methods that final outcomes partially through its impact on
can be used to support and enhance change employee outcomes such as job satisfaction, and
processes within the NHS. The findings in the we clearly need more substantial data sets for surer
work organisation area are particularly promising interpretation.
in the light of considerable changes in methods
of service delivery that are ongoing in the NHS. Measuring intermediate
Opportunities for job and service redesign within outcomes in the NHS
the NHS offer great scope for future exploration.
The use of training to support the implementation The report presents information on the reliability
of change is also highlighted in the good practice of measures in each of the intermediate outcome
around implementation fidelity identified by this areas identified for review. Where possible, the
review, and therefore is important evidence on the specific measures used in the included studies were
process of HRM policy development and practice. reported on. Where an intermediate outcome area
was not covered by the studies included in this
Relationships between review, the subject experts on the research team
intermediate outcomes identified an appropriate measure for inclusion.
Details of the measures, their items and reliabilities
The relationships amongst intermediate outcomes are presented in Chapter 10.
were also examined. Moderate to high correlations
were found between all of the intermediate These measures represent a basic toolkit that could
outcomes for which data were available. The be used or adapted for future NHS-based research
associations, although strong, do not suggest of the HRM performance link.
construct redundancy and it is reasonable to
xi
xii
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Chapter 1
Introduction
of high-quality performance, thereby sharpening approach and performance, i.e. that the impact
the competitive edge of Western organisations of human capital will vary between contexts. In
and improving public sector delivery. They the past few years, more studies have either been
contend that there is a unique set of practices or conducted singly in the service sector or included
approaches to management that will, regardless that sector within the scope of the study,15,25–27 and
of the context, outperform all others. Within the the results have been mixed. Studies specifically
strategy literature, the resource-based theory of the on health care are emerging, for example West et
firm4–6 highlights how human resources, and the al.28 have investigated the link between HRM and
knowledge embodied in them, may be decisive for patient mortality.
long-term competitive advantage. In the operations
management literature7–10 the emphasis is on how Assessing the evidence
HRM practices and the knowledge and skills of
workers are decisive in exploiting lean production11 Notwithstanding the above, an increasing number
and other modern management methods (e.g. see of writers question whether the portrayal of the
Dean and Bowen12 on TQM). studies as providing conclusive evidence for a
universal link between HRM and performance is
Research evidence premature, and that there is a clear need to delve
more deeply – both into the studies and into the
The last decade has witnessed a stream of research reality of high-involvement management and its
assessing high-involvement management and links to other aspects and methods of management.
its associated concepts. The basic hypothesis There are a number of reasons for extending the
underlying this work is that high-involvement debate.
HRM systems will have positive effects on
organisational performance. As academics First, the results of the various studies are neither
have sought to move beyond advocacy of high- as clear-cut nor as uniform as some have concluded
involvement systems, they have focused on testing (see Wood,29 particularly table 1 for a summary of
this hypothesis. A spate of research studies in the the main results; and Wall and Wood30,31). Within
1990s were motivated by this, with many surmising studies, there is unevenness in the findings between
that high-involvement HRM systems do perform performance measures; while some results point
best.13–20 Several overviews21,22 concluded that, on to universal effects, others do not. Moreover, in
the basis of the first few studies, the universalistic many studies only the universal hypothesis has
hypothesis is supported. The increasing use of the been tested so one cannot rule out the contingency
term ‘high-performance’ model to describe the set argument that the effects of high-involvement
of practices may imply that the matter is indeed management on performance are contingent
settled, much as a drug might be named by the on a third factor, either a dimension of the
disease it is known to cure. context or the strategy of the organisation, even
when a positive link between high-performance
The initial overviews of these studies22,23 tended management and performance has been found.
to present them in a homogeneous way, taking
for granted that they are studying the same Second, it is not always clear whether high
phenomenon and that the results are broadly involvement or performance management is being
the same, concluding they had successfully defined simply as the combined use of best practice
demonstrated the link between the human capital in each domain of human resources, as a synergistic
approach and key organisational outcomes. These set of practices, or by a more fundamental
reviews have led some to conclude that the main managerial orientation.32 In fact, the discussion so
issue now is to explain the link between high- far has begged the question of how we define and
involvement management and performance, i.e. identify a system. Most studies have defined an
to assess the mechanisms between them or get HRM system a priori, often with little theoretical
inside the ‘black box’ between the HRM system justification, as the emphasis has been on testing
and outcomes. A second issue might be the extent the high involvement–performance link, rather
to which the link will be found in all contexts, and than first investigating the relationship between
particularly in service industries such as health practices or the nature of any systems. Researchers
care. The majority of the early studies were in have measured HRM on the basis of differentiating
manufacturing, and it is significant that one of the organisations by their usage of a set of practices,
few early studies in services24 found evidence of a the precise means varying between studies. Then
contingent relationship between the human capital they correlate these differences with performance
2
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
measures. The problem of simply aggregating systems, with some studies treating them as part
practices to provide an overall measure of HRM is of high-involvement management or at least
that each practice is treated implicitly as if it were performance management, whereas others do not.
equally important. With only a few exceptions,33,34
studies have provided insufficient information Fourth, there has been insufficient empirical
about which individual or subset of practices has exploration of the mechanisms or intervening
the strongest effects and which may be marginal variables that might explain any link between
or even irrelevant. In one case13 where this was HRM and performance. Of course it must be
done, and it was shown that only some of the acknowledged that there is little point in investing
practices affect performance, and, moreover, that resources in exploring the reasons for a link
their strength varied between industrial context, until it has been established. Yet, labels such
though the authors still framed their conclusions as ‘high commitment’ and ‘involvement’ imply
in terms of an overall (high-performance work) certain mechanisms that are concerned with the
system yielding superior performance. Aggregating extent of commitment or engagement on the
practices does not allow one to test whether the part of the work force, which is mirrored in the
effects of one practice are enhanced by the use importance given to morale in statements about
of another, i.e. whether they are synergistically the modernisation of health care, and these could
related, and begs the question of whether they tend be explored more than they have been. Much
to be used together.32,35 of the theoretical justification that precedes the
empirical studies also implies that skills, knowledge
Third, there is considerable variability across and learning might have an independent effect on
studies in the practices included either in the performance as important as employees’ attitudes.
model of high-involvement management or the
set of practices used to test its link to performance. Some studies have included intervening variables,
Some of the differences are terminological, which have either centred on intermediate HR
perhaps reflecting disciplinary biases or a quest outcomes (e.g. labour turnover in Huselid18)
to differentiate one’s wares. But a core difference or on commitment and satisfaction.36,37 Several
within the literature can be identified. On the authors hint at, but do not explore in depth, other
one hand are those authors who see the core mechanisms; for example, Guest and Conway38
of high-involvement management as changes imply that procedural justice may be important.
in work organisation, job design and employee In addition to morale and staff development,
involvement methods (particularly idea capturing), an important emphasis in statements on the
with the other practices acting as supports to modernisation of health care is what might be
help motivate and equip people to work in a termed the strategic integration of individuals into
new, more flexible and proactive way. On the the organisation, so that they both understand
other hand are those who do not prioritise the and share the goals that achieving patient
task system. They may then concentrate more on care entails and, consequently, orientate their
skill acquisition, following an emphasis on the behaviour towards common visions. Studies
resource-based theory of the firm, according to outside the narrow confines of the HRM studies,
which genuine competitive advantage reflects an particularly building on the job redesign and lean
organisation having a unique capability or set of production literature, point to the importance
resources. Therefore, primacy is placed on the of the kinds of attitudes that advocates of
competencies of individuals, rather than whether total quality management and other modern
they are empowered. Alternatively, they may treat management methods highlight, including
high-involvement management as entailing the quality consciousness,39 continuous improvement
use of the whole gamut of sophisticated personnel orientation40 or flexible work orientation.41
management methods, and conceive them as
primarily operating through people’s commitment Finally, the majority of studies have concentrated
or sense of whether they are being treated fairly on a narrow range of HRM practices in isolation,
and consistent with their psychological contract. the exception being those that have included
The different approaches need not affect TQM in their analysis,20,31,34 but there is a general
greatly the practices researchers include in their failure to assess the effectiveness of HRM and
studies, but the inclusion of job design and work related practices relative to both key elements of
organisation practices is one major source of management, such as leadership, and practices
diversity between the studies, as Wood and Wall31 further afield, such as R&D expenditure.
show. The other major difference is incentive
3
In addition, despite the differences in focus and assess the reliability of measures of some outcomes
measures, most of the studies share a basic research of HRM.
design, which has a number of weaknesses. These
include: The brief for the review was that its scope should
be broad, covering, where feasible, evidence from
• The use of cross-sectional designs, which limits non-health sectors, although evidence from the
the ability to make causal inferences (e.g. health sector should be emphasised.
to decide whether HRM practices promote
performance, or rather better performance The objectives of the review (source: NHS/SDO
encourages greater investment in HRM). Invitation to Tender 2004) are to:
• A reliance on a single data source (e.g. a chief
executive or HRM director) for information 1. describe HRM practices and measure the
on HRM practices and performance, which fidelity with which they may be implemented
may result in measures of unknown reliability 2. review the reliability of methods of measuring
and common-method bias (see, for example, particular intermediate outcomes (intended
the debate between Huselid and Becker42 and outcomes of HRM that may affect patient care,
Gerhart et al.43 on the reliability of measures e.g. morale)
based on single respondents). 3. review the literature on the correlation of
• Many have small samples, low response rates intermediate outcomes to each other
and concentrate on a limited range of sectors. 4. investigate the correlation of intermediate
outcomes with final outcomes
However, even the most searching reviews that 5. investigate the sensitivity of intermediate
have highlighted problems in the studies have outcomes to individual HRM practices.
concluded that there is nonetheless sufficient
promise in the body of evidence to consider major A pictorial summary of the project and its
investments in the area.30 It is thus timely to invest objectives (adapted from that provided in the
in detailed assessment of what we know. original brief) is shown in Figure 1 (for details of the
full proposal see Appendix 1).
Objective 1: Objective 3:
Describe HRM Links between
methods intermediate outcomes
Intermediate
Intermediate
outcomes:
HRM outcomes: Final
Policy Staff
practice Staff behaviour outcomes
psychological
outcomes
outcomes
4 FIGURE 1 Human resource management policy, performance model and project objectives.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Chapter 2
Methodology
Introduction is examined in this review is multidisciplinary and
heterogeneous. Finally, the review is composed
Chapter 2 provides a broad description of the of five (subsequently six) distinct and yet related
methodological approaches used at different objectives – not a single well-focused research
stages in the review. First, the research context question but rather a series of separate and equally
is considered and the challenges of applying elusive questions.
systematic review methodology to the HRM
literature are discussed. Next, the systematic Systematic review was the method of choice
review approaches used are explained and the because the methodology offers a number of
specific questions for the review established. The models to be applied or modified depending
combinations of search terms and databases used on the questions being asked. Systematic review
are detailed and a summary of the methodological is an evolving methodology, primarily designed
approaches used in addressing each of the specific to answer questions regarding the effectiveness
research questions is presented. of interventions by synthesising findings from
large numbers of studies. As such systematic
reviews are ‘a scientific tool which can be used to
Applying systematic review summarise, appraise and communicate the results
methodology within the and implications of otherwise unmanageable
quantities of research. They differ from traditional
HRM literature reviews and commentaries produced by “content
Systematic review methodology has a very experts” in that they use a replicable, scientific and
limited pedigree within HRM, which has only transparent approach that seeks to minimise bias.’47
comparatively recently begun to adopt some of the Systematic reviews follow a process of constructing
tools of evidence-based practice.44 Nevertheless, a clearly defined research question with inclusion
some key studies do exist. For example, Levy and exclusion criteria, comprehensive searching to
and Williams45 conducted a systematic review identify all relevant literature, quality assessment
of 300 studies examining the social context of of selected studies, extraction of relevant data from
performance appraisal and Taris46 examined 16 these studies, and synthesis of the study results. In
studies dealing with the burnout–performance cases where there is sufficient similarity in the focus
relationship. The lack of a significant body of of studies, in terms of interventions and outcomes
exemplars within HRM has limited the capacity of researched, a meta-analysis may be performed: a
this particular field to develop tailored methods of quantitative synthesis. The final result of the review
systematic review and to stimulate methodological is thus a conclusion based on a critical synthesis
innovation in the identification, assessment and of the comprehensive, unbiased selection and
synthesis of this type of data. As a consequence, appraisal of the best available research examining
there are very few instances of methodological the topic under consideration, which typically, in
papers on conducting systematic reviews within the health area, is an intervention.
HRM. The review team, therefore, faced a
number of challenges in conducting this series of This methodology was originally developed for
systematic reviews and in leading the adaptation analysing the effectiveness of pharmaceutical
and development of methods to tackle practical interventions in medicine, based on quantitative
problems generated by these review questions. data from randomised controlled trials (RCTs).
The challenges were compounded by a corpus of Systematic reviews in the medical field now have
literature that conceives the concepts of both HRM an accepted, standardised methodology.47,48 In
and performance (and indeed their intersection) recent years, the advantages of systematic review
in different ways, asking comparable but not over more conventional forms of literature review
identical questions, and using different types of have been recognised and examples have appeared
research design. Furthermore, the literature that in other fields.49,50 This has led to the piloting and
8
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Once this had been done, each research challenging. Chapter 3 details the process of
question was then typically broken down into identifying HRM practice terms, and the specific
the components of ‘population’, ‘intervention’, intervention and outcome terms used by the
‘comparators’ and ‘outcome’; this ‘PICO model’ respective reviews are given in Chapters 6–8.
provides a structure for devising research questions,
constructing search strategies and defining the Developing the protocol
inclusion and exclusion criteria for the review.47 It
is not necessary for all of the elements of the PICO The protocol provides an explicit outline and
model to be present in every question. Where a plan to be followed by the review. Conventional
PICO model has been used in these reviews, it systematic review methods require that a protocol
is detailed in the specific methodology for that describe clearly defined populations, interventions
chapter. and outcomes for the review, and provide details
of the literature-searching procedure, quality
Scoping exercise assessment, and data extraction and synthesis to
be conducted.47 As the outcome of the scoping
The scoping exercise defines key issues for a review process, the protocol specifies the exact content and
as clearly as possible in terms of the population, processes of the review, from which there should be
interventions and outcomes to be studied, and no deviations. It facilitates the performance of the
the inclusion and exclusion criteria to be applied. review by itemising exactly what the review does
Together with the definition of the research and does not include, and specifying the methods
question, this constitutes the ‘problem formulation of quality assessment and synthesis. It should also
stage, which involves the development of key state any language or geographical restrictions that
research questions, construction of the definitions might lead to bias in the review.61,62
of the key concepts … and the establishment of
inclusion and exclusion criteria for the review’.60 It is increasingly recognised that protocols for
reviews in disciplines where the literature is a less
For each review, a series of scoping exercises was known quantity, i.e. potentially heterogeneous,
undertaken to identify the best search terms with qualitative, and lacking the tradition of a
regard to the balance of specificity (relevance) recognised hierarchy of study design and quality
and sensitivity (inclusivity). The nature of the assessment criteria,54,55,60,63 may be required to be
HRM literature made the identification of specific broad, flexible and open to change. In short, as
intervention and outcome terms particularly Eakin and Mykhalovskiy64 suggest, the question is
10
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
In conventional review methodology, literature from three databases using standard and more
searching is refined by the scoping process and sensitive versions of the pay and training filters.
then carried out once.47 However, where concepts The number of potential HRM categories and
or interventions are evolving or cannot be easily practices was very large. A complete and formal
defined, supplementary, purposive literature process of reciprocal translation to produce a
searches may be performed. This may reflect more clearly defined list would have been an
an increasing understanding of the topic or the extremely resource-intensive task and would have
subsequent identification of specific questions produced an unmanageable number of results
raised by the findings of the review.52,63 Searching from any searches performed. On account of the
becomes an iterative rather than one-off process, multifaceted nature of our study, with its multiple
the so-called ‘spiral approach’ as described by questions, it was decided that the approach of
Grant et al.71 This was often how we had to proceed leading the review by specific HRM interventions
in this review, where pre-defining outcomes was not viable. The original idea of structuring this
could have limited the findings. However, such a review by HR intervention was therefore rejected.
process is still systematic and consistent with the
principles of systematic review methods. All search Problems encountered using particular HRM
strategies and sources for all included papers are terms required an alternative approach to the
documented to make the process transparent, intervention terminology. The team tested the
auditable and reproducible. viability of constructing an intervention search
filter using only generic HRM terms, such as
Developing the search terms human resource, HRM, HR and high-performance
work practices or management, and database
Scoping searches were performed using two of keywords, such as ‘personnel management’. This
the less ambiguous categories from the list of removed the problem posed by constructing
HRM practices generated for the review: pay and strategies using potentially non-HRM-related
training. Even relatively non-sensitive scoping terms, such as ‘training’, ‘communication’ or ‘pay’,
searches of the principal health and psychology which retrieved large numbers of studies and
databases (MEDLINE and CINAHL) produced were, arguably, as generic and non-specific as this
large numbers of research studies, which, although overarching terminology. Scoping searches using
they used these terms, were not the focus of this approach produced manageable numbers of
the review and would provide no data on HR potentially more relevant studies and, therefore,
performance linkages. Additionally, the scoping this method was adopted. The resulting HRM
searches indicated that even larger numbers of search strategy is discussed more fully in Chapters
citations would be retrieved by the more sensitive 5–7, and example searches are provided in
searches planned for the full review. Table 2 Appendices 6 and 7.
provides an illustration of the number of hits
TABLE 2 Results of scoping searches for two HR practices with patient outcomes
– Comments,
HR + Patient editorials, letters,
practice Database Hits outcomes reviews English only + Longitudinal
Pay systems a
MEDLINE 1350 417 379 367 55
Pay MEDLINE 58,126 7443 6387 5572 1385
EMBASE 40,179 5882 4873 4213 914
CINAHL 13,909 1344 1283 1247 209
Traininga MEDLINE 11,607 939 823 793 163
Training MEDLINE 178,806 10,028 8804 8269 1341
EMBASE 209,317 28,908 21,084 19,283 4001
CINAHL 9968 1274 1219 1198 160
11
12
DOI: 10.3310/hta14510
HRM
IOs
POs
Longitudinal
Meta-analysis/review
IO measures
MEDLINE, CINAHL
PsycINFO
Business Source
Premier
SSCI
ASSIA, SCI,
EMBASE, Cochrane
library
HMIC, King’s Funda
IBSS,a BNI
DH-Data, NRR,
ReFeR
PREMEDLINE
Report chapter
FO, final outcome; IO, intermediate outcome; PO, patient outcome; ReFer, Research Findings Register.
a Of the two databases listed, only this database was searched.
Health Technology Assessment 2010; Vol. 14: No. 51
13
Methodology
Synthesis aims to ‘draw together, contextualise • Question 1.1 What HRM practices are described
and interpret the findings from the separate in the literature?
[studies]’.53 Methods of synthesis are determined
by whether data are qualitative or quantitative, Method Content analysis was performed on
and by the questions being asked of the data.75 six, recent major reviews of HRM practices,
Meta-analysis is ‘the quantitative synthesis of supplemented by additional literature identified
data, in which evidence is pooled using statistical for Objective 6. Practices were then categorised
techniques’.75 A narrative review seeks to ‘capture under broad headings of HRM, using accepted
and describe, rather than “average out” the definitions of these categories, for example work
heterogeneity between studies’, and its choice as design, performance management.
a method of synthesis may be ‘predicated on the
diversity and complexity of the field’.53 Several • Question 1.2 How widespread is the use of
methods of synthesising qualitative data are being HRM practices in UK organisations?
developed, such as thematic analysis, grounded
theory, meta-ethnography and realist synthesis.75 Method Evidence was drawn from the best available
The method applied to each review is determined surveys of HRM practices in the UK. These
by the heterogeneity of the included studies and surveys were identified from the expert knowledge
their data. In this review, synthesis is primarily of members of the project team, personal
descriptive and quantitative in recording the communications from colleagues, and systematic
studies that meet specific characteristics, and is searching of electronic databases and relevant
accompanied by a narrative synthesis.76 There websites. Surveys were included if they contained
were very few studies in this review that addressed UK data. They were then weighted according to
the same basic research question (i.e. examining specific criteria. Results were tabulated.
the same relationship or set of relationships)
precluding the use of meta-analysis for most • Question 1.3 How can the fidelity with which an
intervention groups. Narrative synthesis was HRM practice is implemented be measured?
retained to ensure consistent synthesis across all
interventions and enabling comparisons to be Method A critical review was undertaken of research
made. on implementation fidelity and relevant studies
from the field of HRM. Studies were identified
by systematic searching of electronic databases
Summary of specific and informal methods, such as serendipity. A
research questions and model, a measure and a guidance document were
developed, based on this literature review.
methodological approaches
The review team drew on the most recent Objective 2
developments in systematic review methodology
in identifying and then resolving the challenges of • Question 2.1 What are the most reliable
this evidence base. measures of the intermediate outcomes
identified for this review?
A number of different research methods were
used to identify evidence in relation to each of the Method The intermediate outcomes to be assessed
specific research questions. This section provides a by this review were identified by content analysis
summary of the methodological approaches used of the relevant literature and a Delphi process
in relation to each of the objectives and specific involving HRM experts. The existing review
research questions. and theory literature on the measures of these
14
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
15
Chapter 3
HRM practices and their
reported use in the UK
Introduction too broad to be influenced by HR policies/
practices alone, but were of potential interest
This chapter of the report presents findings in to the review for their impact on organisational
relation to Objective 1. Specifically, it addresses or individual performance. Examples include
the first two research questions identified under bullying/harassment, socialisation, etc.
Objective 1 (Question 1.3, on implementation • A sizeable number of HR policies/practices,
fidelity, is dealt with in Chapter 11): 29 in total, were considered to be part of HR
maintenance functions (i.e. administration)
• Objective 1: 1.1 What HRM practices are rather than proactive HR practices aimed at
described in the literature? enhancing performance. Examples include
• Objective 1: 1.2 How widespread is the use of equal opportunity, dispute resolution policies,
HRM practices in UK organisations? and health and safety policies.
The chapter first describes the process by which The classification of HRM practices into broad
HRM practices were identified, and then goes on categories was reviewed by the expert advisers. As a
to report findings on the extent of use of these result, the terms were refined according to the most
practices in the UK. appropriate HR category as defined and used in
the literature and additional HRM practice terms
were suggested, which had not been identified in
What HRM practices are the review of reviews:
described in the literature?
• The higher level concepts category was
This phase of the research sought to audit the main renamed ‘bundles’ of practices. Over the last
HRM practices described in the literature. The aim 10–15 years there has been a growing body
was to identify a complete list of HRM practices of research on ‘bundles’ of practices. This
to be developed into a search strategy. Data from
three sources were used to develop the list.
TABLE 4 Initial categorisation of HRM
Step 1
1 Work design
The research team and expert advisers generated 2 Training and development
a list of 13 headings. Eleven of these were broad
3 Communication
HRM categories under which it was anticipated
that the majority of individual HRM practices 4 Single status/harmonisation/status differentials
could be grouped, which could then be used as 5 Appraisal/performance management
a framework to shape the reviews. Initial HRM 6 Higher-level concepts (HPWPs, HI, HC, etc.)
practice terms are presented in Table 4. 7 Administration
8 Recruitment and selection (including placements)
It was recognised from the outset that HRM
9 Compensation and reward systems
practices were unlikely to all fall into neat
categories. In addition to the 11 broad categories 10 Family friendly
of HRM practice identified, two further categories 11 Employee representation/involvement/participation
were used to group individual practices: 12 Total quality management
13 Miscellaneous
• The ‘Miscellaneous’ category was used to
group practices that were not subject to direct HC, high commitment; HI, high involvement; HPWP,
manipulation by the HR department, or were high-performance work practice.
17
category was therefore intended to capture particular the inconsistency of terminology and the
research that focused on combinations of HRM blurred boundaries of some of the concepts, meant
practices that were intentionally designed to be that the broad categories identified through this
used in concert as well as for multiple practices process would not be a suitable as the sole basis
opportunistically used together. for the subsequent review of evidence. Ultimately,
• Articles in the ‘Miscellaneous’ category were a combination of broader search terms and study
reappraised and either recategorised under design criteria were used (which are described in
one of the main headings or excluded from Chapters 2, 5 and 6).
this assessment of the main HRM practices
described in the literature. Step 3
Step 2 The systematic reviews undertaken to meet
Objectives 3–5 generated over 450 papers (see
In a parallel process, a list of HRM practices was Chapter 5 for a description of these searches). The
compiled from six major recent reviews of the individual HR practices, policies or interventions
HRM literature.30,35,77–80 This process identified 245 that had been the focus of these studies were
practices. mapped on to the list of HR practices generated by
Steps 1 and 2. This was done in order to identify
A process of reciprocal translation81 was applied any gaps in the list of HR practice terms and
to the 245 practices identified by the review of further refine the categories. The exercise revealed
reviews, with the intention of building a consensus that:
about the main HRM practices covered in the
literature. Reciprocal translation involves the • A large body of literature in the health
comparison of themes across papers and an sector focused on specific HR practices not
attempt to establish whether there is equivalence commanding much attention in the non-health
in the themes of one paper with those from sector – specifically ‘working hours’, ‘staff
another. The aim is to ensure that a key theme levels’ and ‘skill mix’. These practices were
captures similar policies or practices from different included under the ‘staffing’ practice category
papers. In this exercise, the reciprocal translation along with selection and recruitment.
was approached by two of the researchers • All TQM practices identified (a separate
independently attempting to compare HRM category on the original list) were, in fact,
practice terms across review papers and group training interventions, so were included under
them under one of the broad HRM categories the broad ‘training’ heading.
listed in Table 4.
The final list of 10 broad HRM categories is as
Initially, a substantial number of practices did follows:
not appear in any of the categories or could be
allocated to more than one. For example, appraisal- 1. work design
based pay can be seen as a part of the broad HRM 2. staffing
category of appraisal/performance management 3. training and development
and of compensation and reward systems. Likewise, 4. compensation and rewards
introducing autonomous work teams can be seen as 5. communication
both an employee involvement strategy and a work 6. family friendly
design approach. 7. single status/status harmonisation
8. employee representation, involvement and
The process of reciprocal translation made it participation
apparent that the same term could be used to 9. appraisal/performance management
describe different practices in different papers (e.g. 10. bundles.
‘goal setting’ to mean performance management
in one context and employee involvement in Appendix 2 contains a list of these broad HRM
another). categories with definitions. Appendix 3 presents
the specific HRM practices identified in the
At this stage in the research, it became apparent literature grouped under these 10 broad categories.
that the diffuse nature of the HRM literature, in
18
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
CRANET, Cranfield Network on International Human Resource Management; IPD, Institute of Personnel and
Development; NESS, National Employers Skills Survey.
a Some of the CIPD surveys contain these data, but others do not.
Seven surveys sampled the workplace, all of which information about staff experiences of working
reported UK data (although three surveys were within the NHS, the annual staff survey differs
Europe wide). In the remaining surveys, CIPD considerably from the HR surveys reviewed here in
members were being surveyed rather than actual that:
workplaces, limiting the potential value of these
sources. Only one specific health sector survey of • It is targeted at the individual employee, rather
HRM practices was identified by this review.102 The than the HR function.
study reported on a non-representative sample of • It asks questions about individual staff
general practices in the UK. Some HRM practices, perceptions or experiences rather than the
which duplicate those covered by other surveys, are absence or presence of specific HR practices.
included because this was the only specific health • The nature of the data means that findings are
service survey. reported at the individual level (i.e. percentage
of all staff responses to a particular item)
This section examines evidence on the extent of for the NHS as a whole, or by type of Trust
use of HRM practices among UK employers. As [e.g. all Primary Care Trusts; all Ambulance
a result, the focus is on the most recent data for Trusts, etc.], or by Strategic Health Authority,
any practice rather than on all available data. or by individual Trust, rather than at the
For example, only the most recent WERS82 data organisational level (i.e. percentage of Trusts/
were recorded here. Data from an earlier WERS83 organisations reporting the presence of an HR
survey were recorded only if the information was practice).
not collected for the subsequent survey. Also, it
was outwith the remit of this review to examine Comparing such data with HR practice measures
developments in coverage of practices over time. is extremely problematic. For example, the
proportion of employees reporting that they have
Health sector-specific surveys had an appraisal in the last 12 months may not
bear a strong relation to the proportion of NHS
The most comprehensive surveys within the health organisations with appraisal policies in place, or to
sector are those conducted by the Healthcare the existence of an appraisal policy in any specific
Commission (HCC). Whilst providing extensive organisation. Accepting these limitations, it is
20
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
possible to interpret some of the NHS Staff Survey at least some of the workforce. However, the
findings as a proxy indicator for staff experience of figures for companies where the entire workforce
particular HR practices in the NHS. is covered by the practice would be considerably
lower. For example, in the Wood et al.99 study of
Analysis of data from the NHS Staff Survey focuses UK manufacturing firms, 70% of responding
on 26 key areas (key scores). These are a mixture organisations had team working but only 38% used
of attitudinal and experiential measures, the it for the majority of the workforce.
latter providing an indication of the percentage of
NHS staff who have experienced a particular HR The information on staffing is mainly confined
practice (i.e. a proxy indicator). to whether selection tests are used and whether
specific recruitment procedures are used to
Where a proxy indicator for staff experience of attract minority or discriminated groups, for
an HR policy is available from the National NHS example disabled or elder workers or whether exit
Survey 2007,103 this has been indicated in the right- interviews are used. None of these practices shows
hand column of Table 6. Findings from the survey particularly widespread coverage, with none of
are summarised later (see Table 17) and discussed the practices being used in more than 50% of the
below (see Data from the NHS Staff Survey 2007, units. Competence tests are used in 46% of the
below). workplaces.
These surveys provide information on the extent Family-friendly practices are covered well in WERS
to which different HRM practices are used. The and their use varies considerably between practices.
results for work organisation practices (team In the cases of practices where there is a statutory
working, job rotation, role empowerment and the right to provide leave, for example maternity and
formality of job descriptions) show that for all of paternity leave, the study focused on the provisions
these practices over 70% of the units surveyed, beyond the statutory minimum, for example paid
the workplace or company had the practice for leave. Aside from these practices, the ability to
21
NHS103 Staff
Survey
WERS ESWT CVTS EPOC CIPD IPD NESS CIPD CIPD CIPD CIPD CIPD CIPD CIPD 2007 proxy
200682 200686 200187 199688 200689 199990 200591 200292 200393 200594 200595 200696 200697 200698 indicator
Work design
Staffing:
recruitment and
selection
HRM practices and their reported use in the UK
Training and
development
Compensation
and rewards
Communication
Family friendly/
work–life balance
Single status/
harmonisation
Employee
involvement
Performance
management
Bundles of
practices/HPWPs
DOI: 10.3310/hta14510
Percentage of workplaces
Practice: interview or survey with practice [overall Data collection Sample size;
Category question/report definition, where available (public/private/health)] method workplace size Source
Job characteristics Formal job description 72 (na/na/na) Manager telephone 27,172 managers/ NESS (2005:
Percentage of workplaces answering ‘yes’ to the survey establishments 84)91
question, ‘What percentage of your staff has a formal
written job description?’
Job rotation Job rotation (used frequently, occasionally or rarely) 77 (na/na/na) Postal survey 635 CIPD members in UK CIPD (2006: 6)97
organisations; no limit
23
24
TABLE 8 Reported use of HRM practices: recruitment and selection
Percentage of workplaces
Practice: interview or survey question/report with practice [overall (public/ Data collection Sample size;
Category definition, where available private/health)] method workplaces size Source
Selection Personality tests used routinely for some occupations 19 (18/20/17) Manager interviews 2024 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, more employees (2006: 77)82
‘When filling vacancies at this workplace do you ever
conduct any type of personality or attitude test?’
Performance tests used routinely for some occupations 46 (63/42/50) Manager interviews 2024 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, more employees (2006: 77)82
‘When filling vacancies at this workplace do you ever
conduct any type of performance or competency test?’
Preference for internal applicants 22 (12/25/na) Manager interviews 2024 managers; 10 or WERS
Percentage of workplaces specifying the answer to the more employees (2004) p. 984
question, ‘Which of these statements best describes your
approach to filling vacancies as internal applicants are
HRM practices and their reported use in the UK
25
26
TABLE 9 Reported use of HRM practices: training and development
Percentage of workplaces answering ‘some’, ‘a lot’ or ‘all’ of time more employees (2006: 92)82
to the question, ‘What proportion of [the largest occupational
group of staff] is formally trained to be able to do jobs other
than their own?’
Some core employees trained in team working, communication 48 (65/45/na) Manager interviews 2012 managers; 10 or WERS
or problem-solving more employees (2004: 11)84
Percentage of workplaces answering ‘yes’ to this choice in the
question, ‘Does the training cover any of the matters listed
[here]?’
Job-specific training 81 (percentage of the 64% of Manager telephone 27,172 managers/ NESS
Percentage of workplaces answering ‘yes’ to this choice in the employers who offer training) survey establishments (2005: 73)91
question, ‘Over the past 12 months which, if any, of the following (na/na/na)
types of training and development has this establishment funded
or arranged for staff employed at this location?’
Health and safety training 80 (percentage of the 64% of Manager telephone 27,172 managers/ NESS
Percentage of workplaces answering ‘yes’ to this choice in the employers who offer training) survey establishments (2005: 73)91
question, ‘Over the past 12 months which, if any, of the following (na/na/na)
types of training and development has this establishment funded
or arranged for staff employed at this location?’
Training in new technology 59 (percentage of the 64% of Manager telephone 27,172 managers/ NESS
Percentage of workplaces answering ‘yes’ to this choice in the employers who offer training) survey establishments (2005: 73)91
question, ‘Over the past 12 months which, if any, of the following (na/na/na)
types of training and development has this establishment funded
or arranged for staff employed at this location?’
Percentage workplaces with
Practice: interview or survey question/report definition, practice [overall (public/ Data collection Sample size;
Category where available private/health)] method workplaces size Source
Supervisory 42 (percentage of the 64% of Manager telephone 27,172 managers/ NESS
Percentage of workplaces answering ‘yes’ to this choice in the employers who offer training) survey establishments (2005: 73)91
question, ‘Over the past 12 months which, if any, of the following (na/na/na)
types of training and development has this establishment funded
DOI: 10.3310/hta14510
In-house development programmes (used frequently, 96 (na/na/na) Postal survey 635 CIPD members CIPD
occasionally or rarely) in UK organisations; (2006: 6)97
no limit
continued
27
28
TABLE 9 Reported use of HRM practices: training and development (continued)
Graduate development programmes 55 (na/na/na) Postal survey 635 CIPD members CIPD
Percentage of workplaces reporting that they use this practice in UK organisations; (2006: 6)97
frequently, occasionally or rarely no limit
Assessment centres 59 (na/na/na) Postal survey 635 CIPD members CIPD
Percentage of workplaces reporting that they use this practice in UK organisations; (2006: 6)97
frequently, occasionally or rarely no limit
Development centres 56 (na/na/na) Postal survey 635 CIPD members CIPD
Percentage of workplaces reporting that they use this practice in UK organisations; (2006: 6)97
frequently, occasionally or rarely no limit
MBAs 86 (na/na/na) Postal survey 635 CIPD members CIPD
Percentage of workplaces reporting that they use this practice in UK organisations; (2006: 6)97
frequently, occasionally or rarely no limit
Action learning sets 56 (na/na/na) Postal survey 635 CIPD members CIPD
Percentage of workplaces reporting that they use this practice in UK organisations; (2006: 6)97
frequently, occasionally or rarely no limit
DOI: 10.3310/hta14510
29
30
TABLE 10 Reported use of HRM practices: compensation and reward
Percentage of workplaces answering ‘yes’ to the question, interviews more employees (2006: 191)82
‘Does this company operate any of the employee share
schemes listed on this card for any employees at this
workplace?’
Cash bonus or incentive plans 66 (32/88/na) Postal survey 535 organisations; no CIPD
limit (2006: 22)89
Team performance-related pay for managers 8 (4/11/na) Postal survey 1158 organisations; IPD
na (1999: 3)90
Team performance-related pay for non-managers 8 (2/12/na) Postal survey 1158 organisations; IPD
na (1999: 3)90
Skill- or competency-related pay for managers 6 (4/7/na) Postal survey 1158 organisations; IPD
na (1999: 3)90
Skill- or competency-related pay for non-managers 11 (9/14/na) Postal survey 1158 organisations; IPD
na (1999: 3)90
Job evaluation scheme 20 (42/16/28) Manager 2059 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, interviews more employees (2006: 244)82
‘Are there any formal job evaluation schemes here? These
are schemes for comparing systematically the relative
value of different jobs in order to settle their relative rates
of pay’
Percentage workplaces with
Practice: interview or survey question/report practice [overall (public/private/ Data collection Sample size;
Category definition, where available health)] method workplaces size Source
Employee Employer provision of child-care assistance 8 (18/5/na) Manager 2054 managers; 10 or WERS
benefits Percentage of workplaces answering ‘yes’ to the question interviews more employees (2006: 255)82
DOI: 10.3310/hta14510
continued
Health Technology Assessment 2010; Vol. 14: No. 51
31
32
TABLE 10 Reported use of HRM practices: compensation and reward (continued)
Entitlement to more than 4 weeks’ annual leave for non- 59 (na/na/na) Manager 1994 managers; 10 or WERS
managerial employees interviews more employees (2004: 21)84
Percentage of workplaces answering ‘yes’ to this choice
in the question, ‘Is the largest occupational group at this
workplace entitled to any of these non-pay terms and
conditions?’
Entitlement to sick pay in excess of statutory 62 (na/na/na) Manager 1994 managers; 10 or WERS
requirements for managerial employees interviews more employees (2004: 21)84
Percentage of workplaces answering ‘yes’ to this choice
in the question, ‘What about managers at this workplace,
are they entitled to any of these non-pay terms and
conditions?’
Entitlement to sick pay in excess of statutory 54 (na/na/na) Manager 1994 managers; 10 or WERS
requirements for non-managerial employees interviews more employees (2004: 21)84
Percentage of workplaces answering ‘yes’ to this choice
in the question, ‘Is the largest occupational group at this
workplace entitled to any of these non-pay terms and
conditions?’
Occupational sick pay 83 (95/83/na) Postal survey 535 organisations; no CIPD
limit (2006: 29)89
Face to face Meetings between senior management and the entire 79 (89/77/na) Manager 2057 managers; 10 or WERS
workforce interviews more employees (2006: 135)82
Team briefings or meetings with entire workforce 71 (81/68/na) Manager 2057 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, ‘Do interviews more employees (2006: 135)82
you have meetings between line managers or supervisors and
all the people for whom they are responsible? … these are
sometimes known as “briefing groups” or “team briefings” ’
Regular meetings with feedback 63 (81/59/na) Manager 2017 managers WERS
Percentage of workplaces answering ‘some’ or ‘a lot’ of time interviews (overall) – 554 (2006: 139)82
to the question, ‘What proportion of the time at meetings managers (public),
is usually available for questions from employees or for 1463 managers
employees to offer their views?’ (private); 10 or more
employees
continued
Health Technology Assessment 2010; Vol. 14: No. 51
33
34
TABLE 11 Reported use of HRM practices: communication (continued)
35
36
TABLE 12 Reported use of HRM practices (family friendly/work–life balance)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
37
38
TABLE 12 Reported use of HRM practices (family friendly/work–life balance) (continued)
Fully paid paternity or fully paid discretionary leave for fathers 55 (84/49/na) Manager 1863 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, interviews more employees (2006: 258)82
‘Would any fathers taking (paternity or fully paid discretionary
leave) from this workplace receive their normal, full rate of
pay?’
Extrastatutory leave for carers (e.g. paid parental leave, 82 (98/77/na) Manager 1879 managers; 10 or WERS
emergency leave, etc.) interviews more employees (2006: 269)82
Paid parental leave or special paid leave for parents 25 (47/21/na) Manager 1975 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, interviews more employees (2006: 258)82
‘With the exception of maternity leave, paternity leave and
time off for emergencies, how do mothers and fathers usually
take time off to look after their children? Paid parental leave …
take special paid leave’
Special paid leave for family emergencies 49 (80/43/na) Manager 1988 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to the question, interviews more employees (2006: 258)82
‘With the exception of maternity leave, paternity leave and
time off for emergencies, how do mothers and fathers usually
take time off to look after their children? Take special paid
leave?’
Percentage workplaces with
Practice: interview or survey question/report practice [overall (public/ Data collection Sample size;
Category definition, where available private/health)] method workplace size Source
Leave for carers of older adults 6 (16/4/na) Manager 1985 managers; 10 or WERS
DOI: 10.3310/hta14510
Percentage of workplaces answering ‘yes’ to the question, interviews more employees (2006: 258)82
‘Are any employees entitled to any of the following? A specific
period of leave for carers of older adults?’
Career breaks/sabbatical: total availability 42 (na/na/na) Postal survey 585 HR professionals; CIPD
no limit (2005: 6)94
Career break schemes 20 (na/na/na) Postal survey 453 CIPD members CIPD
whose organisations (2005: 6)94
offer working from
home; no limit
Retirement Sabbaticals 17 (na/na/na) Postal survey 453 CIPD members CIPD
whose organisations (2005: 6)94
offer working from
home; no limit
39
40
TABLE 13 Reported use of HRM practices: single status/status harmonisation
continued
Health Technology Assessment 2010; Vol. 14: No. 51
41
42
TABLE 13 Reported use of HRM practices: single status/status harmonisation (continued)
Age monitoring recruitment and selection 20 (na/na/na) Manager interviews 2030 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to this choice in more employees (2006: 248)82
the question, ‘Do you monitor recruitment and selection
by any of these characteristics?’
Review procedures for recruitment and selection 16 (na/na/na) Manager interviews 2030 managers; 10 or WERS
regarding age more employees (2006: 248)82
Percentage of workplaces answering ‘yes’ to this choice in
the question, ‘Do you review recruitment and selection
procedures to identify indirect discrimination by any of
these characteristics?’
Age monitoring on promotion 7 (na/na/na) Manager interviews 2030 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to this choice in more employees (2006: 248)82
the question, ‘Do you monitor promotions by any of these
characteristics?’
Review procedures for promotion regarding age 9 (na/na/na) Manager interviews 2030 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to this choice more employees (2006: 248)82
in the question, ‘Do you review promotion procedures
to identify indirect discrimination by any of these
characteristics?’
Review of relative pay rates regarding age 6 (na/na/na) Manager interviews 2030 managers; 10 or WERS
Percentage of workplaces answering ‘yes’ to this choice in more employees (2006: 248)82
the question, ‘Do you review relative pay rates by any of
these characteristics?’
Percentage workplaces with
Practice: interview or survey question/report practice [overall (public/ Data collection Sample size;
Category definition, where available private/health)] method workplaces size Source
Training interviewers in diversity 53 (72/49/na) Postal survey 659 HR professionals CIPD
DOI: 10.3310/hta14510
43
44
TABLE 14 Reported Use of HRM Practices (Employee Involvement)
45
46
TABLE 16 Reported use of operational practices associated with HRM
Practice: interview or survey question/ Percentage workplaces with practice Data collection Sample size;
Category report definition, where available [overall (public/private/health)] method workplaces size Source
Just-in-time (JIT) JIT system of inventory control 29 (na/na/na) Manager 1926 managers; 25 WERS (1998:
interviews employees or more 10)83
JIT na (78/na/45) Manager 1115 UK manufacturing Wood et al.
telephone survey or health organisations; (2004: 426)99
more than 150;
employees
Total quality TQM na (88/na/73) Manager 1115 UK manufacturing Wood et al.
management telephone survey or health organisations; (2004: 426)99
(TQM) more than 150
employees
47
HRM practices and their reported use in the UK
reduce hours, for example to switch from full- to Data from the NHS Staff Survey
part-time work, is the most frequently provided 2007
practice, with 70% of workplaces having it, while Data from the NHS Staff Survey103 is presented
the provision of child-care facilities is the least in Table 17. Caution is needed when considering
used. these data in relation to the data on HR practices
in general. The NHS data shows the proportion
The Workplace Employment Relations Survey of NHS staff who report work experiences that
offers the only information that there is on single are consistent with the HR practices considered
status, revealing that 48% of workplaces provide here. It is not, however, indicative of the number
the same level of benefits for all staff (i.e. they of Trusts in which such HR policies/practices exist,
provide or do not provide a particular benefit nor does it help address other important questions
for managers and non-managerial staff alike). raised in this section, such as the percentage of the
Five surveys are questions on this specifically. workforce covered by a particular HR practice in a
The nearest proxy is equal opportunities, which given organisation.
is covered extensively in surveys, although only a
minority of organisations go beyond a formal equal Accepting the limitations of the NHS Staff Survey103
opportunities policy (73%) or training interviewers data in relation to the focus of this chapter, it can
in diversity (53%). be considered as providing proxy indicators of staff
experiences in areas. This is of particular interest
Information on employee involvement methods where organisational level HR practice data are not
covers union representation, consultation, available.
suggestion schemes and employee surveys. The
usage of such practices varies considerably but in Specifically the NHS Staff Survey103 data indicate
no case are they used in a majority of workplaces. that 39% of NHS employees work in a well-
structured team environment. This contrasts
Information on performance management with data from the survey of HR practices, which
focuses exclusively on appraisal. The majority of indicate that 72% of all organisations surveyed
workplaces have some form of appraisal, but only have some employees working in teams and 79% of
a minority (27% in WERS) do it more than once a health sector organisations make some use of team
year. working.
Separate figures for the health sector are available With regard to training, 54% of employers
for only a small number of the practices (Tables surveyed on HR practices had funded or arranged
7–16, column 3), and indeed for the majority of some training over the previous 12 months.
HRM categories (e.g. family-friendly practices) we No breakdown is available for health sector
have no such data. For those for which we have organisations, but 94% of NHS employees report
data, in the majority of cases the frequency of taking part in some form of training over the
use in the health sector is not that different from previous 12 months. Eighty-one per cent of
the overall use, and in most cases is even more employers from the HR practice surveys provided
similar to the use in the public sector. The two job specific training and 80% health and safety
exceptions are related to trade unions. The first is training. From the NHS Staff Survey, 77% of staff
the determination of pay by collective bargaining, reported receiving job relevant training, 71%
as this exists in 82% of public workplaces but only health and safety training.
in 36% of health sector workplaces and 14% of the
private sector. The second is union recognition, There are no health sector-specific breakdowns
which is granted in 90% of public workplaces, available for the wide range of communication
41% of health sector workplaces, and only 16% of practices covered by the HR practice surveys.
private sector workplaces. The NHS staff survey indicates that 23% of
NHS employees feel that senior management
Although evidence is reported on some HRM communicate effectively with staff.
practices within each of the broad categories
included in this review, the survey coverage is by no No breakdown is available for the percentage of
means comprehensive. It is clear from these results health sector employers offering family friendly/
that crucial elements of the HRM system are not work–life balance HR practice; however, the NHS
covered by a lot of the surveys. Staff Survey103 indicates that between 2% and 32%
48
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
of NHS employees make use of one of a range appraisal in the previous 12 months. Interestingly,
of specific practices in this area, with 73% of all this drops to 24% when factors indicating a well-
employees reporting that they use flexible working structured performance appraisal were taken into
options. account. Fifty-two per cent of staff (as measured by
the NHS Staff Survey103) reported that they agreed
Eighty-nine per cent of health sector organisations a personal development plan (PDP) as part of their
report having an equal opportunities policy in appraisal.
place, but no further breakdown of specific HR
practices in this area is available. Thirty-nine per
cent of NHS employees report receiving some Conclusion
form of diversity training. Levels of training in
age, disability, gender, race, sexual orientation or Human resource practice surveys and the NHS
religion-related equality/diversity range from 14% Staff Survey cover some common themes. NHS
to 18% of employees surveyed. Staff Survey103 data help to supplement the picture
provided by organisational level surveys of HR
With regard to employee involvement, there is practices by demonstrating the degree to which
some evidence from HR practice surveys in relation NHS employees in general may experience the
to health sector organisations. This mainly relates results of HR practices in place in organisations.
to union membership and joint consultative Some of the NHS Staff Survey data provide better
committees (Table 14). No breakdown is available proxy indicators than others (e.g. evidence of a
for the percentage of health sector organisations PDP) and receiving training identified in a PDP
using other employee involvement HR practices are arguably stronger indicators of the existence
(such as suggestion schemes or problem-solving of a specific HR practice in relation to appraisal
groups). Findings from the NHS Staff Survey103 and development than, say, the extent to which
indicate that 23% of NHS employees feel that perceptions of the quality of communication
senior managers try to involve staff in important between senior managers and staff indicate the
decisions and 31% feel that senior managers existence of an HR communication practice.
encourage staff to make suggestions to improve
services. Responses to the NHS Staff Survey103 at an
individual Trust level can help that Trust to assess
With regard to direct participation, 48% of NHS the effectiveness of a range of HR practices within
staff surveyed felt that they were involved in the organisation. Further breakdown of WERS
decision-making, and/or consulted about decisions data to reveal the extent of HR practices in the
that affected them. health sector would enable more meaningful
comparisons to be made between data collected at
Finally, with regard to performance management, the organisational level and individual data such as
89% of health sector employers reported using the NHS Staff Survey. Equally, systematic collection
performance appraisals (compared with 71% for of HR practice data at Trust level would enable
all organisations). From the NHS Staff Survey, more detailed analysis to identify the presence and
61% of respondents said that they had received an effectiveness of such practices.
49
Chapter 4
Implementation fidelity in HRM
Introduction This section reviews the literature on
‘implementation fidelity’, develops a framework for
This report seeks to provide as requested in the understanding and explaining processes at work in
invitation to tender (ITT) ‘an account of the evaluating and achieving implementation fidelity,
different human relations policies and methods and places this evidence and framework within the
to determine the fidelity with which they were context of HRM and policy. Subsequent guidance
implemented’. Implementation fidelity relates to outlines questions to be asked when implementing
the fidelity with which both policies and research- a policy. A checklist is presented for evaluating
based interventions are implemented by those and explaining the degree of fidelity achieved in
responsible for delivering them. In this case, the implementation of any such policy.
concern was for the fidelity with which real-world
HRM policies are implemented as practices. The
aim was to understand better how knowledge about Policy–practice link and
the effectiveness of HRM practices, evaluated implementation fidelity
and reported in previous sections, may best be
transferred into practice. There is no shortage of research evaluating the
impact of HRM practices on performance amid
Certain practices have been found by this report growing awareness of the need for evidence-based
to be effective, for example in reducing mortality practice.57 However, if a practice of demonstrated
rates or patient waiting times, but these practices effectiveness is to have a positive impact it needs
may only have the same potential effect when to be implemented faithfully. This idea is captured
reproduced by policy-makers and practitioners if by the concept of implementation fidelity: ‘the
they are implemented properly, as they should be, degree to which … programs are implemented
i.e. with high fidelity to the original practices or … as intended by the program developers’.104
interventions examined in the research. Unless an evaluation assesses whether a policy
has been implemented properly, it is not possible
This section reviews existing literature on this to know whether lack of impact is due to poor
concept and measurement of fidelity, with the aim implementation, (a so-called ‘Type III error’),105
of understanding the process of implementation and covered by the ‘thesis of comprehensiveness’,106
fidelity and thus informing and improving or inadequacies of the policy itself. Also, it would
the success with which HRM policies may be be unclear whether positive outcomes produced
implemented. With this in mind, the report by a policy could be improved still further, if the
was also required to provide guidance on how policy had in fact not been implemented fully.
fidelity could be measured, and to produce
a ‘checklist of actions, which an auditor for a Monitoring of fidelity is of particular concern
resource management method, such as Investors in HRM because of a long-standing concern
in People, would use’, i.e. a checklist to be followed about the gap between organisational HR policy
by those seeking to evaluate fidelity (ITT). The prescriptions and actual practice.107 Several scholars
guidance was intended to inform implementation ascribe to poor implementation the divide between
by outlining questions that need to be asked at espoused people-management policies and the
every stage in the process if high fidelity is to be reality experienced by employees subject to the
achieved. A tool to evaluate the fidelity with which policies.107–110
policies have been implemented is provided as
a checklist. This evaluation could be undertaken This section critically reviews the literature on
by researchers, policy-makers or practitioners, or implementation fidelity and considers the concept
anyone seeking to assess whether a policy has been in relation to the implementation of policies in
implemented properly in practice, i.e. as intended HRM. It begins by examining existing descriptions
by those designing and seeking to deliver the of a model of implementation fidelity. This is
policy. followed by a review of research that specifically
51
uses the term ‘implementation fidelity’ and uses an HRM policy may not be a simple one-stage
this concept to evaluate the implementation and process; it may take time to be implemented
effect of interventions. This literature focuses on fully and, once it has been implemented, this
interventions to help and support people with level of implementation needs to be sustained.
mental health problems, criminal behaviour or Implementation therefore requires ongoing
addiction issues. This research provides examples monitoring and support.54 Few studies examine
of how to measure implementation, describes how to sustain implementation over time. The
how to achieve high levels of fidelity and, in some second issue concerns the appropriateness
cases, evaluates associations between level of of adaptations to fit ‘local’ conditions. Some
implementation fidelity and the actual impact of an specifications allow for local adaptation or, even
intervention. The concept of fidelity is then applied if they do not explicitly do this, local adaptations
to HRM policy, both as a model and with reference may aim to improve a practice so that it fits more
to existing HRM ‘rhetoric versus reality’ literature, with the local context. Indeed, Blakely et al.115 refer
which deals with some elements of implementation to a proadaptation perspective that implies that
fidelity, although it is less sophisticated and uses successful interventions are those that adapt to
different terms. Finally, a new framework for local needs. Yet, others argue that the case for local
understanding the process of evaluating fidelity is adaptation may well have been exaggerated, where
proposed. This framework is then used to develop the evidence does not necessarily support it.116
guidance on evaluating fidelity, and a checklist for The intermediate position is that implementation
assessing the degree of fidelity with which a policy can be flexible as long as there is fidelity to the
has been implemented. essential elements, the absence of which would
adversely affect the capacity of the practice to
achieve its goals, and without which the policy
Dimensions of fidelity cannot be said to have been implemented in any
meaningful fashion.104,111,112 However, the question
In the literature that specifically names and remains about how to identify such essential
discusses the concept, implementation fidelity elements. Possible options are either canvassing
is described in terms of ‘five dimensions’ or opinions of the practice designers themselves, or
‘components’.104,111–113 These are adherence to a conducting a sensitivity analysis or evaluation to
policy or intervention, dose, quality of delivery, determine which components are most vital to its
participant responsiveness, and programme effectiveness. To date only the more unreliable
differentiation. former method has been used to identify such
elements, and limitations of this approach are
Within this framework, adherence is defined as acknowledged.116–118 What is more, there is a
whether a policy, practice, service or intervention growing appreciation that implementation data
is being delivered as it was designed or written’.111 should accompany research on the effectiveness
Dosage refers to the amount of a practice received of interventions to facilitate their proper
by participants, in other words, whether the implementation.119,120 This involves specification
frequency and duration of the practice is as full as of the essential components of an intervention or
intended.104,111 It may be, for example, that not all HRM practice because this defines exactly what
of the aspects of a HRM practice are delivered or needs to be implemented.
are delivered but not as often as required. Quality
of delivery has been defined as ‘the manner
in which a teacher, volunteer or staff member Overview of findings from
delivers a program’.111 Participant responsiveness fidelity studies in social
is ‘the extent to which participants are engaged
by and involved in the activities and content of
policy
a program’.104 Programme differentiation, the There is a growing body of literature on
fifth dimension, is defined as ‘identifying unique implementation fidelity. Studies summarised
features of different components or programs’ and in Table 18 are a sample of those published
‘which elements of … programs are essential’.104 since the reviews conducted by Dusenbury et
There are certain overlaps here with the concept of al.104 and Mihalic.111,112 This literature focuses on
process evaluation.114 interventions to help and support people with
physical and mental health problems, criminal or
Two other issues need to be addressed. The first antisocial behaviour, or addiction issues.
is implementation over time. Implementation of
52
TABLE 18 Studies addressing implementation fidelity: 2004 to present
p. 11)
O’Brien (p. 244) ‘the extent to Prevention Nurse–Family Partnership Site development, guidelines, Unclear Unknown
(2005)121 which the program is programme: training, an information
replicated with fidelity to the Nurse home visits to support system to record data for
intervention tested’ young mothers on low measures, ongoing quality
incomes improvement assessment and
feedback
Resnick et al. (p. 140) ‘Treatment fidelity Health and Exercise plus vs exercise-only Training manuals, standardised Programme All elements
(2005)122 impacts the internal and medicine intervention programme: training program, ongoing description (p. 140)
external validity of a study, Exercise intervention for direct observation of
as well as the effect size of women, post hip fracture. interventions with feedback,
a tested intervention … If Individualised programme and ongoing training
the treatment being tested involving weekly or monthly
does not adhere to the stated exercise, education,
Elliott and None given Prevention Blueprint VP programme and Capacity building/site Initiatives described in Core only
Mihalic LST programme development, training, published reports ‘… judged by program
(2004)116 ongoing technical assistance developers as essential for
program effectiveness … not
established by experimental
manipulation or other formal
tests’ (p. 49)
Rinaldi et al. (p. 282) ‘Fidelity refers to the Health and ETHOS team vs ETHOS None given IPS model (Bond et al. Core only
(2004)118 degree of implementation of social care plus part-time vocational 1997)129 ‘… critical ingredients of
an evidence-based practice’ specialist: evidence-based supported
To support education and employment for people with
employment of young people severe mental illness’ (p. 282)
(aged 17–30) with mental Items identified by
illness ‘brainstorming sessions
1 year among the authors’ (Bond et
al. 1997: 271)129
continued
Health Technology Assessment 2010; Vol. 14: No. 51
53
54
TABLE 18 Studies addressing implementation fidelity: 2004 to present (continued)
ACT, Assertive Community Treatment; ATP, Adolescent Treatment Programme; ETHOS, Early Treatment and Home-based Outreach Service; IPCC, Intensive Psychiatric
Community Care; IPS, Individual Placement and Support; LST, Life Skills Training; MAPS, Marriage and Parenting for Stepfamilies; na, data not available; PMTO, Oregon Model of
Parent Management Training; PSRM, Personal and Social Responsibility Model; SECT, Supported Employment Consultation and Training; SIL, social interaction learning; VP, violence
prevention.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Elliott and Mihalic116 report that ‘the available and clear recommendations were almost twice
research demonstrates that fidelity is related as likely to be followed as vague and complex
to effectiveness’, a finding echoed by other recommendations.137 Generally, research indicates
reviews111,112,114 and the primary research.117,127,127,136 that simple but specific policies or interventions
Much of this research, summarised in Table 19, are more likely to be implemented with high
seeks only to measure implementation fidelity, fidelity than complex or vague ones. As such,
but some studies do aim to analyse findings about the comprehensiveness and nature of an HRM
fidelity in relation to the interventions’ outcomes policy’s description may moderate how far the
too. Forgatch et al.117 have demonstrated how the implemented practice actually adheres to the
level of implementation fidelity to a parent training policy.
programme, designed to improve interaction
between parents and stepchildren, predicted the Two further points are raised by this research.
level of improvement in parenting practices. Firstly, the level of implementation fidelity is not
determined purely by the simplicity or specificity
Resnick et al.125 found that employment outcomes of a policy or intervention, strategies are usually
among people with mental health issues ‘were employed to facilitate the achievement of the
weakest for those in poorly implemented Intensive highest possible fidelity. Seven of the studies
Psychiatric Community Care (IPCC) programs’ reviewed here report the employment of strategies
(p. 142). The same correlation was found in a to support adherence to the policy or intervention
similar study by McGrew and Griss,127 which also as it was designed, and to ensure high quality of
looked at supported employment programmes. delivery. In some cases, these strategies include
Resnick et al.122 also found that participant ongoing monitoring and feedback to ensure
outcomes were good when fidelity to an exercise continuing adherence and quality of delivery
programme was high, but no cases of low fidelity over time.116,121,122 Monitoring such as this exists
were present in this study, so a positive correlation not only to optimise, but also to standardise
could not be confirmed. The number of studies implementation fidelity, i.e. to ensure that everyone
evaluating this association is small, but this research is receiving not only the most effective training or
does indicate that the degree of fidelity with which support to aid implementation, but also exactly
a policy or intervention is implemented will be a the same training.138 These studies aim not only
factor in its achievement of successful outcomes. to evaluate fidelity, but also to achieve the highest
fidelity possible. However, the potential role of
Research has also found that detailed or specific these strategies in implementation fidelity has
policies are more likely to be implemented not been considered by research. It is possible
with high fidelity than those that are vague. that these strategies, like the nature of a policy
For example, well-planned violence protection or intervention’s description, may potentially
interventions, where all the key components moderate the degree of adherence: the more that is
are identified in advance, have been found to done to help implementation, through guidelines,
produce higher levels of adherence than less well- training, support and monitoring, the higher
structured interventions.112 Specificity enhances the level of fidelity achieved. This is especially
delivery. Indeed, by far the majority of policies and the case for complex interventions, which can be
interventions implemented in our sample are very multifaceted and more vulnerable to variation
detailed and are based on published models (Table in their implementation.139 The role of such
18: Source of data 1). In one case where the source strategies in optimising fidelity and standardising
is unclear,121 strategies to support implementation what is being implemented arguably is even more
fidelity, for example guidelines, suggest that a important in this case, and even more likely to
highly specific model is behind the intervention. influence the level of fidelity achieved. However,
High specificity may be prerequisite to meaningful there is currently no research to confirm the
measurement. potentially important role of facilitation strategies
as moderators of the relationship between policies
Specificity does not mean complexity. A policy and actual implemented practices, but they should
or intervention can be specific but simple, and be considered as such.
there is evidence that it is easier to achieve high
fidelity of simple guidelines or policies with Secondly, there is no generic measure of fidelity.
fewer ‘response barriers’ than complex ones.54,104 All of the studies used different measures. In each
For example, a study of guidelines intended for case the content of measures is determined by the
general practitioners (GPs) found that detailed content or description of the HRM practice or
55
O’Brien Outcomes Data collected by local Unclear Only given for 1 of the Good implementation fidelity
(2005)121 Outcomes are perceived programme staff Dose is acknowledged (p. measures Indicators ‘in expected direction’ but
to be indicative of level of 251) based on frequency and ‘weaker’ than RCTs (p. 251)
implementation fidelity (p. duration of nurse home visits, Reduction in maternal smoking and
246) but no formal measure is given drug use, but below target levels
Measures relate to maternal Positive impact on immunisation rates
and child functioning. Results
compared to national Positive impact on language
benchmark standards development
Positive impact on maternal education
Resnick et al. Checklists, logbooks, Checklists completed Delivery (adherence) No information given Good implementation fidelity
(2005)122 calendars by direct observation of Receipt ‘that the treatment has 90%+ score for delivery, receipt and
random sessions been received’ (p. 142) (dose) enactment
Logbooks of staff Enactment ‘that the treatment
delivering the has been … understood …
intervention (self-report) that the individual performs
Calendars of patients treatment-related … skills and
receiving the strategies’ (p. 142) (participant
intervention (self-report) responsiveness)
Elliott and Fidelity score (%) VP: Not clear how these Adherence No information given Good implementation fidelity
Mihalic No. of core components data are collected Dose: lessons taught ‘at Adherence levels high
(2004)116 ‘correctly and fully LST: Checklist rated by recommended levels … dosage Dosage levels for LST low
implemented’ (p. 49) divided classroom observers elements in the prescribed
by total number of core amounts’ (p. 50)
components to give fidelity
score
Rinaldi et al. Supported EFS ‘Independent Adherence Bond et al. (1997) (cited) Good implementation fidelity
(2004)118 15-item scale; items use 1–5 assessment’ (p. 282) provide validity and reliability 71 out of possible 75 on fidelity score
Likert scale information on the Fidelity
Scale
DOI: 10.3310/hta14510
continued
Health Technology Assessment 2010; Vol. 14: No. 51
57
58
TABLE 19 Studies measuring implementation fidelity (continued)
ATP, Adolescent Treatment Programme; DACTS, Dartmouth Assertive Community Treatment Scale; EFS, Employment Fidelity Scale; FIMPY, Fidelity of Implementation Rating
System; IPS, Individual Placement and Support; LST, Life Skills Training; QSEIS, Quality of Supported Employment Implementation Scale; SE, supported employment; SECT,
Supported Employment Consultation and Training; VP, violence prevention.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
intervention as it appears in the policy document embedded in policy. Typically, such policies and
or as espoused by the designers of the intervention. programmes involve installing new practices or
Nine of the studies collected data on fidelity maintaining central activities or approaches within
using measures that were developed specifically core areas of HRM, for example recruitment and
for the intervention being studied. For example, selection, training, employee involvement, and
the Individual Placement and Support Scale (IPS equal opportunities. Assessing their fidelity is about
Fidelity Scale) was developed to measure fidelity gauging whether policies are implemented fully as
to interventions that were designed to help people practices.
with mental health issues into employment,127 and
the service delivery tracking log, completed by Adherence is then about assessing whether
social workers, was used to monitor implementation practices are implemented in accordance with
of an intervention to protect children potentially policy. ‘Dosage’ refers to ensuring that all aspects
at risk.124 Two other studies evaluated fidelity using of the policy are covered, and all employees and
outcomes: they assumed that positive outcomes are sites to be covered by a particular element are
a proxy for high fidelity of implementation,121,123 actually involved in it. They must also be exposed
which, in effect, negates the primary reason for to it as frequently, and for as long, as prescribed
assessing fidelity, i.e. to see if it moderates the link by policy. ‘Quality of delivery’ refers to how well
between policy and outcomes. the practice is delivered or performed by those
responsible for doing so. For example, a policy may
Data collection on fidelity involves independent prescribe that training is delivered by specialist
observation or reporting in four of the studies, and trainers and involve feedback on quality from
the self-report of staff delivering or participants participants. In this sense, quality of delivery is
receiving the intervention in the remaining seven part of implementation adherence. Alternatively,
studies. The measurement of fidelity in two studies a policy may simply state that training in a
combines data from independent observers with particular skill should be given to all new members
data from staff and participants. Independent of staff but make no specific prescriptions to
observation of delivery offers the best source of guarantee the quality of the training. Quality of
data about both adherence and quality of delivery. delivery is therefore no longer a dimension of
Six of the eleven studies give validity and reliability implementation fidelity, an element of adherence,
information on the scales used. Three studies aim but rather a moderator of the relationship
to measure whether the ‘essential’ elements of the between a policy and successful implementation.
intervention or practice have been implemented, Participant responsiveness concerns the reaction
with the opinions of experts being used to identify of employees to the practices. It may involve
these elements.116–118 assessing whether employees are engaged in each
element of the practice to the extent that they
The implementation fidelity measures applied by should be, and as such may also moderate the
this research literature are intervention specific and implementation–outcome relationship. Programme
heterogeneous in their data collection methods, differentiation is relevant if certain components
but they all seek to measure implementation are core. For example, if employee involvement is
fidelity by measuring adherence, and have as their key to an HRM policy we would naturally expect
starting point highly specific descriptions of their practices associated with the high-involvement
interventions. They also aim to achieve high levels model to be used. Other practices may include
of fidelity by employing similar support strategies, supporting training programmes. A ‘programme
and seek to provide accurate measurement differentiation’ analysis, to use the earlier
of implementation fidelity. However, none terminology, may reveal that supporting practices
examines whether any of the following may have are just as important as core practices, or that
a moderating effect; the complexity or simplicity their delivery is a vital adjunct if the core practices
of the interventions’ descriptions; strategies to are to have an impact. Indeed, within employee
facilitate implementation, quality of delivery; or the involvement it is widely recognised that the not-
responsiveness of participants. so-essential practices support the essential ones,
and hence are necessary if the performance effects
of the policy are to be realised. In effect they are
Fidelity of HR policy practices that ensure the quality of its delivery.
Organisations establish HRM policies, and As with all evaluations, an assessment of the fidelity
occasional change programmes, which may become of policies is an important complement to the
59
evaluation of its impact. However, there are several This implies that the effect of fidelity on outcomes
other reasons why it is important to assess the of policies will be moderated by the level of
fidelity of HRM policies and practices. specificity of a policy. Alternatively, it may be the
case that low specificity leads to non-adherence,
First, the issue of policy implementation may because if policies are vague they can be easily
contain interesting questions in its own right neglected, even allowing for the implied ease at
precisely because adherence is not such a which it might be met. In the extreme, policies may
straightforward matter. Policy-makers themselves be so vague as to contain no clear intentions in
would no doubt value understanding the which case assessing the policy against intentions
organisational processes through which policies are is not possible. In less extreme cases, where some
implemented and the factors (organisational, group intentions can be inferred, little of interest may
and individual) that may facilitate and impede emerge from an assessment of the implementation
implementation fidelity. For example, Cunningham fidelity of a vague policy.
et al.107 in an analysis of four case studies, identified
a range of factors that militated against the In HRM, policies are often not clearly specified if
implementation of a policy on job security for they exist at all. They are often not self-contained,
ill workers. These included contradictory policy have evolved incrementally and statements only
requirements, weakness in training, and various refer to the most salient, perhaps most novel,
work and budgetary pressures. Implementation features. Other elements are taken for granted
fidelity is therefore inherently important in or left unspecified. Modern managerialism may
understanding and evaluating how HRM policies encourage policy-makers and managers to develop
are implemented. policy statements that simply convey homilies and
exhortations to go in a particular direction.141
Second, non-adherence to a policy may adversely
affect employees. For example, if an organisation Certainly, the presence of a policy should not be
has a policy of paying high wages but does not give inferred from HR practices, and, in the same way,
out wage increases, employees’ pay satisfaction and the presence of practices should not be inferred
perception of trust in their employer may decline from the existence of a policy. Indeed, this is a
and affect their performance or commitment to the recognised issue in HRM research.142–144 As a result
organisation. It is often claimed that management research increasingly looks to measure practices by
espouses policies that they do not operationalise collecting data from multiple sources rather than
fully, or even at all, precisely because the policies relying on a single managerial respondent, whose
are designed to make them appear employee- response about uptake may be determined by their
centred but their implementation falls considerably knowledge of policy rather than practice.
short of the implied aspiration. This claim often
falls under the rhetoric-versus-reality label107,140 As a consequence, most recent HRM research
and is applied especially to employee involvement concentrates on practices and not policies or
and equal opportunities, but also sometimes to programmes. More specifically, attempts to link
TQM. For some employees, perhaps especially HRM to performance, i.e. gauge its impact, focus
management, this policy infidelity may affect their on practices, perhaps because of the problems of
morale and commitment. Indeed, the disparity vagueness surrounding policies or because small
between an organisation’s practice and its professed and medium firms may well not have formal
values may be one possible cause of burnout or policies, or simply because they want to focus
disengagement amongst employees. on what organisations do rather than what they
profess to do. Some studies measure HR practices
Finally, policies are important because they drive by coverage, i.e. the proportion of staff covered
and shape practices, and should be accompanied by a practice, but if we do not have information
by strategies to achieve adherence. As a result, about the level of coverage stated in any policy
policies may also ensure that a practice is then we cannot use it as a measure of dosage. For
implemented well, i.e. the quality of delivery is example, a practice may cover 60% of staff. This is
high. One problem that has to be faced, however, the measure of the coverage of a practice, typically
is that the lower the level of specificity the greater assessed by surveys, such as the figures reported
the possibility that the policy will be adhered to, above. However, if a practice is found to cover
on the grounds that it is easier for practice to fall 60% of staff but should in fact cover 100%, i.e. all
within parameters that are broad, or at least to staff then it becomes a measure of implementation
appear to be within them when they are opaque. fidelity because this suggests that the organisation
60
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
has an acknowledged benchmark for the practice, the HR studies sought to evaluate whether policy
perhaps provided by written policy. is being implemented rather than to do so while
also trying to achieve high levels of fidelity or
It may even be possible to employ theoretical evaluate its relationship with performance, the
or model benchmarks against which to measure need for support strategies is recognised: three
adherence. For example, if an organisation studies report how such strategies had been used by
claims to be using TQM then seeing whether organisations.107,140,147 Like much implementation
reported practices adhere to the accepted model fidelity research, they focused entirely on
of TQM could offer a measure of implementation adherence (including dose). The principal
fidelity.34,144 However, the danger is that the difference between the two literatures concerns
benchmark is a pure ‘textbook’ model that is data collection: the HR literature used only self-
unlikely ever to be implemented in this form. report of participants rather than independent
Nonetheless, Patterson et al.34 successfully audited observation – employee experience rather than the
companies against pre-determined definitions report of quantitative data.
of ‘good practice’ for advanced manufacturing
technologies, TQM, JIT, job enrichment and skill In many cases, studies compare details about
enhancement, and, in so doing, circumnavigated policy given by HR managers or documents, with
the implementation problem by assuming that the report of employees. Disparities between
the measures reflected both the existence and the the two are highlighted. In most cases, greater
effective use of a practice. weight is given to the report of the employees,
even though self-report of employees can be
Finally, the question that naturally follows an unreliable measure of a practice and its
an assessment of how fully a policy has been implementation.110,143,148,149 No validity or reliability
implemented concerns the effectiveness of the information was given about any measures.
resulting practice. Only two studies have sought Measures used were basic by comparison with
to measure effective implementation in the HRM the fidelity scales, logs and checklists used in the
field.145,146 Khilji and Wang145 argued that high fidelity literature (Table 21). HR studies reported
levels of implementation of practices correlate little more than employees’ opinions about whether
with high levels of performance, with satisfaction or not a policy has been implemented; fidelity
as a mediator. However, this study only compared studies often reported percentages or exact fidelity
manager and employee perceptions rather than scores. Also, none of the HR studies sought to
actual implementation fidelity and assumed relate fidelity to outcomes.
that satisfaction was a recognised mediator of
the relationship between implemented practice Research needs to test whether specificity of HRM
and performance, the practices being the prior policy aids or hinders implementation fidelity
variable. The study was also unable to determine (adherence and quality), and to ascertain if this
the direction of causality in these relationships. effect varies according to context. This is a gap
Another study aimed to measure how well HRM in the HR literature and implementation fidelity
practices were implemented by measuring research. Research also fails to address adequately
outcomes: successful outcomes implied successful the issue of local adaptation, which in HR is
implementation or practices.146 This relationship particularly pronounced in multinationals and
cannot be assumed, however. On the whole, large public bodies where adaptations to local
studies have ignored whether how well practices demands and differing cultures is acknowledged as
are implemented affects outcomes, and simply legitimate, and may be incorporated in the policy
concentrated instead on studying the effects of the or may be implicit in management’s approach
adoption of practices, regardless of how well they to it. The effect of specificity may in fact be
are implemented. moderated by the extent to which local adaptation
is allowed or more importantly vital for successful
In the field of HRM, few studies have collected performance.
and evaluated data on the implementation
of policy (Table 20), the equivalent of the The sophisticated knowledge of the concept and
interventions described in the implementation measurement of fidelity apparent in the specific
fidelity literature (Table 18). Despite a lack of implementation fidelity research and associated
concern about implementation fidelity within the literature is clearly lacking in the field of policy
HRM literature, there are similarities between the implementation research in HRM.
literature described in Tables 18 and 20. Although
61
‘Core’
Definition of Strategies to support elements only?
Study fidelity Domain Intervention implementation fidelity Source of data 1, programme How identified?
Cunningham None HRM Management of long-term Training: pp. 283–4 Policy documents Unknown
et al. (2004)107 absence: p. 278
Currie and None HRM Teamworking: pp. 588–9 Selection of managers, Semi-structured interviews with all Unknown
Procter Six offices in one Allocation of work to groups training of managers, grades of staff (including HR staff)
(2003)147 UK organisation rather than individuals. Change training of employees, team
in management style development activities (e.g.
Implementation fidelity in HRM
63
Implementation fidelity in HRM
Potential moderators:
1. Comprehensiveness of policy
description
2. Strategies to facilitate
implementation
3. Quality of delivery
4. Participant responsiveness
Adherence in practice:
Details of content
Policy Coverage Performance outcome
Frequency
Duration
Evaluation of
implementation Evaluation
fidelity
place as often as it should?; Did it last as long evaluate the fidelity with which policies have been
as it should?; and Did it contain what it should implemented. This evaluation could be undertaken
have done? However, implementation fidelity by researchers, policy-makers or practitioners,
might also require that the participants actually indeed anyone seeking to assess whether a policy
acquire knowledge of a new skill as a result of the had been implemented properly in practice, i.e. as
training. Whether or not they did so might again intended by those designing and seeking to deliver
be influenced by the complexity of the training the policy.
programme, the quality with which the training
was delivered (quality of delivery), the strategies
that were in place to support the delivery and Conclusion
implementation of the training (such as a manual
or guidelines), and whether the participants Introduction of new practices, or the revision of
themselves were actually receptive to the training existing practices, requires formulation of policy or
(participant responsiveness). programmes. There is a growing interest in data
for assessing the quality of implementation.119,120,151
Whether the new knowledge acquired affects However, even if the requisite data are provided
outcomes or performance is a separate question, within examples of effective HRM interventions,
beyond fidelity. This is the remit of measures it is still necessary to measure fidelity to
of the practice–performance relationship. gauge whether evidence-based policies are
Implementation fidelity is therefore a potential being implemented properly, in accordance
moderator of this relationship, as represented with a specific policy. Achievement of high
by the broken line in Figure 2. A sensitivity implementation fidelity is one way of replicating
analysis based on performance outcomes could the success achieved by original research.
then determine which, if any, components or Obviously, the successful development and
combination of components are essential if the implementation of policy is governed by more than
practice is to have its desired effect. The scope for the principles of evidence-based practice,152 but the
adaptability would then be known, and the issue of above framework helps to understand why and how
external validity addressed more fully. policy implementation should be evaluated.
Finally, the fidelity with which apparently less This review concerns the entire HRM performance
complex HRM practices are implemented may chain, from HR policy through to performance
still be affected by the moderators described here. outcomes. This chain has several stages, from
For example, appraisals: Have all those who are policy to practice, from practices to intermediate
eligible been covered? Are they appraised as often outcomes, and from intermediate outcomes to
as they should be? Does the appraisal last the set final outcomes. The link between HRM practices
amount of time? Are the required elements of the and intermediate outcomes (e.g. job satisfaction,
process covered? The description of the process, burnout), and the link between such intermediate
the support provided for its implementation, the outcomes and final outcomes (e.g. productivity/
quality of any training given to the managers patient care outcomes), are often measured by the
delivering the appraisals, their own delivery of the research and are analysed in depth elsewhere in
new practice, and the enthusiasm of both managers this review. The HR policy to practice link, called
and employees for the practice may all affect how here implementation fidelity, is measured much
far the policy is adhered to in practice, i.e. the level less if at all. This section therefore considered
of implementation fidelity achieved. the first stage in the HR policy to performance
chain, critically examined current literature on the
The framework in Figure 2 outlines and clarifies concept, outlined its importance, described how it
processes to be taken into account in any evaluation can be measured and the problems presented by
of implementation fidelity. This framework measuring it. The relevance of this lies in how it
informs the guidance presented at Appendix 4 and highlights the processes involved in implementing
checklist in Appendix 5. The guidance outlines practices of demonstrable effectiveness and how
the questions that need to be asked at every stage the implementation process may be evaluated.
in the process if high levels of implementation As such, it is of value to policy-makers and those
fidelity are to be achieved. The checklist is a tool to seeking to develop evidence-based practice.
65
Chapter 5
Identifying linkages between HRM
practices and performance outcomes
Introduction be the mediating path between HRM and final
organisational performance outcomes, although
This chapter provides context and theoretical we would expect HRM practices to also directly
background for Objectives 2–6. Specifically, it influence performance outcomes by creating
discusses current thinking about intermediate operational and structural efficiencies. It is a short
outcomes of HRM. First, the links between HRM step to specify the HRM practices that promote
and performance are considered. The chapter high knowledge, skills and abilities,22 for example
then goes on to describe the method used for through careful attention to recruitment, selection,
identifying the intermediate outcomes considered appraisals and investment in training. Less
in the review and provides conceptual definitions straightforward is the series of mental, emotional
for each intermediate outcome. and attitudinal processes156 that are frequently
proposed to link HRM and performance,
and which are implied by labels such as high-
Intermediate outcomes commitment and high-involvement HRM.
linking HRM and
performance Intermediate outcomes
Two of the principal aims of this review are to Although there are a multitude of variables
evaluate the research evidence linking HRM that are implicated in this process, some of the
practices and organisational performance and more important proposed mental, emotional
also the evidence for mechanisms or intermediate and attitudinal states are employee well-being,
outcomes of HRM that may explain any link an effective psychological contract, perceived
between HRM and performance; specifically, organisational support, fairness and justice,
how and why HRM practices may impact on motivation, job satisfaction, commitment job
performance. While researchers have devoted a involvement and organisational climate.156 An
great deal of attention to examining the linkage organisation’s HR architecture will be closely
between HRM practices and firm performance, tied to these states. For example, satisfaction of
there has been less empirical work attempting to employees, perceptions of fairness and trusting
better understand possible causal mechanisms. management–employee relationships may be
Without examining intervening variables or influenced by HR practices that provide employees
intermediate outcomes of HRM, it is difficult to with opportunities for participation and autonomy.
rule out alternative explanations for an association Intrinsic motivation could well be influenced by
between HRM and performance.23 work design interventions, such as job enrichment
and empowerment. Firms offering job security
There are a number of rationales and and organisation pay programmes, such as profit-
models explaining a link between HRM and related pay, can engender employee perceptions’
performance,13,22,153,154 and while there are of organisational support, and their identification
differences there are also constant themes that with, and commitment to, the organisation.
are shared across the approaches. At a general
level, most models are based on the assumption However, a lack of clarity remains about the degree
that superior HRM practices are believed to result of overlap between the constructs describing
in more skilled, motivated, satisfied, committed, employee states and their causal ordering (e.g.
and hence productive, employees, who, in turn, Does satisfaction precede commitment, and
promote a more effective firm.155 That is, employee under what circumstances?; Is job satisfaction
skills, psychological states (such as attitudes and a consequence of job involvement or is the
motivation) and specific behaviours (such as job relationship reciprocal?). Also the relationships
performance and absenteeism) are purported to between any one individual’s feelings,
67
attitudes, motives and performance will not be most popular conceptualisation of OCB is a five-
straightforward.157 As Ostroff and Bowen155 point dimension model proposed by Organ,159 consisting
out, ‘Lower performance is only one possible of altruism, courtesy, conscientiousness, civic virtue
response to dissatisfaction. A dissatisfied employee and sportsmanship.
could file a grievance, engage in sabotage, ask for a
transfer, try to improve performance or fail to pass Although much of the work on OCB takes the
on important information.158 In contrast, a satisfied position that there is a clear boundary between in-
employee could work harder, engage in more role and extra-role behaviour, the distinction is not
citizenship behaviours, improve skills or make as clear as much of the literature suggests.161
suggestions for improvements in work procedures’
(p. 228). But although the relationship between Research has suggested that employees who are
satisfaction and performance at the individual level satisfied with their jobs and committed to their
may be not be large, the collective effects for the organisation, and feel that they are treated fairly,
organisation as a whole may be enhanced due to are more likely to engage in OCBs.162 Consequently,
the overall cumulative impact of the responses of OCBs have recently emerged as a potential
satisfied or dissatisfied employees.155 important linking mechanism between employee
attitudes, such as commitment and satisfaction, and
Productivity-related employee organisational outcomes. Any single incidence of
behaviours citizenship may have a minor or moderate effect –
‘however, across multiple acts of citizenship from a
How do employee mental, emotional and single employee, and across multiple employees,
attitudinal states get converted into organisational the aggregate impact should be more substantial’
performance? The assumption is through their (Ostroff and Bowen:155 227). However, so far,
influence on employee behaviours salient to arguments for an association between employee
effective organisational performance, i.e. when OCB and organisational performance are
these mental, emotional and attitudinal states are conceptual rather than empirical.
favourably directed then employees are more likely
to display performance-orientated behaviours.156 In summary, despite the paucity of evidence on
Three pertinent sorts of employee behaviour are these relationships (evidence that will be reviewed
task performance, withdrawal behaviours and in later chapters), it has been consistently theorised
organisational citizenship behaviour (OCB).154 that satisfaction, motivation and other mental,
emotional and attitudinal states of employees
Task performance behaviours include job-related tasks are key factors in determining appropriate
and activities comprising the employee’s formal productivity-related employee behaviour,
role. which ultimately contribute to organisational
performance.
Withdrawal behaviours refer to absenteeism and
turnover. A large number of variables may serve as
intermediate outcomes linking HRM practices and
Organisational citizenship behaviours include final performance outcomes, from job satisfaction
components of so-called prosocial behaviours, to organisational commitment, and from burnout
and refer to constructive or cooperative gestures to organisational support. Given the vast range of
that are not mandatory, but which contribute potential intermediate outcomes available from
to organisational effectiveness.159 OCB refers to the research, and the limitations of project time
work-related behaviour that is outside traditional and resources, it was necessary to select a limited
task performance, in-role job performance and number of intermediate outcomes.
the formal organisational reward system, but
is beneficial to the organisation.160 Typically,
employees who engage in OCBs are those who Selecting intermediate
go the extra mile for their organisations and outcomes
therefore contribute to its effective functioning.
This discretionary work performance has been It is important to state at this stage that while we
conceptualised in a number of ways (e.g. contextual support the notion that HR practices have a direct
performance, prosocial organisational behaviour, impact on the knowledge and skills of employees
extra role performance), but the research literature that may translate to improved organisational
has focused most prominently on OCB. The performance, here, and in line with the research
68
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
brief, we concentrate on the role of employee a newer concept in the literature but is increasingly
mental, emotional and attitudinal states as seen as important employee orientation for
intermediate outcomes of HRM practices. performance, and organisational climate and
organisational support play a prominent role in
The intention was to select only those variables for some models of HRM and performance.
which high-quality data were available, and which
were likely to prove most instructive about the Selected intermediate outcomes
relationships being examined by this report. • Motivation.
• Job satisfaction.
For this reason, a list of potential intermediate • Organisational commitment.
outcomes was compiled from the following sources: • Occupational/professional commitment.
• Engagement.
• The longitudinal research examining the link • Burnout.
between HRM practices and their outcomes • Job involvement.
(Chapters 6–8). • Turnover intentions.
• A list of 40 systematic reviews identified • Psychological contract.
by a search of electronic databases (ASSIA, • Organisational justice (distributive,
MEDLINE, EMBASE, CINAHL, PsycINFO, procedural).
Cochrane Library, Emerald, King’s Fund, • Organisational support.
DH-Data, HMIC, SSCI, SCI) using the search • Organisational climate.
terms ‘human resource’ or ‘HRM’ or ‘high
performance’ or ‘high involvement’ or ‘high Excluded intermediate outcomes
commitment’ and ‘systematic review’ or ‘meta- The following potential intermediate outcomes
analysis’. were in the initial list identified from the sources
• Six recent major HRM reviews (Hyde et al.,77 above, but were excluded because they were only
Boselie et al.,78 Wall and Wood,30 Michie and cited by a single source:
West,79 National Institute of Clinical Studies,80
Wood35). • work-related irritation
• Literature modelling the mediators and • job anxiety/tension
moderators of the HRM practice–performance • job exhaustion
relationship).153,154,163–167 • health complaints
• psychosomatic complaints
This produced a comprehensive list of the • psychophysiological stress reactions
intermediate outcomes cited by both secondary • somatic complaints
research and primary longitudinal research into • physical symptoms
HRM. If two or more of the above sources cited an • psychological well-being
intermediate outcome then it was included. • depression
• strain (depression /anxiety)
In addition, intermediate outcomes that would be • occupational stress
otherwise excluded were included, nevertheless, if • frustration
they were present in a meta-analysis that reported • life distress
correlations between this intermediate outcome • life satisfaction
and three or more selected intermediate outcomes • self-esteem
identified specifically to address Objectives 2 and 3 • conscientiousness
(see Chapter 9). For example, job involvement was • level of participation
originally excluded as an intermediate outcome • transfer of learning
because it was only mentioned by one of the • overflow.
four sources described above, but meta-analyses
reported correlations between organisational
commitment, professional commitment, burnout Definitions of selected
and OCB and this variable. It was then included intermediate outcomes
as an additional intermediate outcome of interest.
Finally, a decision was made in consultation Motivation
with the expert panel to include three ‘wildcard’ Work motivation can be generally defined as
variables – employee engagement, organisational the direction, intensity and persistence of work
support and organisational climate. Engagement is behaviour;168 i.e. what drives a person’s choice
69
of what to do, how hard they try and how long itself. In addition, the concept is sometimes
they keep trying. Motivation is clearly critical to separated into intrinsic satisfaction, which involves
performance. Indeed, work performance is often satisfaction with supervision, co-workers and the
used as a measure of a person’s motivation,169 work itself, and extrinsic satisfaction that targets
although there are many other factors that features extrinsic to task activities, such as pay and
influence performance other than motivation (e.g. promotion.
ability, resources, employee collaboration and
cooperation). There are many motivation theories Organisational commitment
that have been advanced to help understand
behaviour at work but it is often difficult to know Employee commitment to an organisation has been
which theory is helpful in any individual case.169 defined in a variety of ways but is most popularly
Some theories are more relevant to the notion that operationalised as the level of attachment and
HRM practices effect employee motivation, such loyalty to an organisation among its employees.172
as goal-setting theory, equity theory and job design The concept has been particularly popular in the
theory. HRM literature as many scholars advocate that
the traditional relationship between employer and
Job satisfaction employee, based on control, must be replaced by a
high-commitment approach.1
Job satisfaction is the most widely researched
concept in organisational psychology and Mowday et al.173 who were influential in the early
organisational behaviour. Locke170 estimated work on commitment, characterised it as a strong
that over 3300 studies on job satisfaction had belief in and acceptance of the organisation’s goals
been conducted up to 1973. Judge et al.171 then and values, a willingness to exert considerable
identified a further 7856 studies on job satisfaction effort on behalf of the organisation, and a strong
since 1973 using the PsycINFO database. Studies desire to maintain membership in an organisation.
and papers have been written about its definition, They have developed a widely used scale, the
meaning, specific dimensions of job satisfaction, its Organisational Commitment Questionnaire (OCQ),
antecedents and its consequences. Its popularity to measure these elements. But the definition and
stems from the key role it plays in many theories the measure have been criticised for conflating
and models of organisational practices and commitment with outcomes such as effort and
processes, individual attitudes and behaviours. propensity to stay.
Job satisfaction is important as an end in itself
for employees and also because of proposed Meyer and Allen174 have suggested definitions and
relationships with individual and organisational measures based on three broad themes. Affective
effectiveness (e.g. turnover and performance). It commitment describes an employee’s liking for an
is a widely held belief that ‘a happy worker is a organisation and emphasised identification and
productive worker’ (although there is considerable involvement in the organisation. Employees with a
debate about the validity of this statement), and job strong affective commitment continue employment
satisfaction is one way of measuring happiness at with the organisation because they want to do so.175
work. Continuance commitment refers to an awareness of
the costs associated with leaving the organisation.
Locke170 defined job satisfaction as ‘a pleasurable Employees whose primary link to the organisation
or positive emotional state resulting from the is based on continuance commitment remain with
appraisal of one’s job or job experiences’. Thus the employer because they need to do so. Finally,
job satisfaction involves affective (emotional) states normative commitment reflects a feeling of obligation
or how one feels about one’s job, and cognition, to continue employment. Employees with a high
deriving from appraisal of the situation. Although level of normative commitment feel that they ought
there is some debate about the emphasis placed to remain with the organisation.
on affect and cognition, it is probably most helpful
to think of job satisfaction as the interplay of The HRM literature has, not surprisingly,
cognition and affect, or thoughts and feelings.171 concentrated on the affective aspect of
commitment because of the assumed associated
Job satisfaction can be thought of as an overall positive organisational gains. The general theory
assessment of one’s job but which comprises various of organisational commitment predicts that high
job facets, such as satisfaction with pay, recognition, commitment should result in greater motivation
promotion, co-workers, supervision and the work and performance, lower absenteeism and lower
70
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
labour turnover,172 with consequent organisational Although the concept is relatively new, work
benefits. engagement does overlap with established concepts
also included in this review as intermediate
Occupational/professional outcomes. Vigour is conceptually similar to
commitment motivation, and the dimension of dedication shares
conceptual space with those of job involvement
Much of the theoretical and empirical work on and commitment.180 Engagement has been
commitment has focused on the organisation; defined as the opposite of burnout,181 as, contrary
however, it is clear that workers have multiple to employees suffering from burnout, engaged
commitments at work (such as work itself) and the employees are energetically and effectively
organisational unit (such as the team, department connected to their work, and able to cope well with
or branch). An important commitment domain the demands of the job.182
outside the employing organisation is commitment
to the profession or occupation as well as the Engagement is influenced by the availability
family. of job resources that reduce job demands, help
achieve work goals and stimulate learning and
As with organisational commitment it is the state development.183 Examples of possible resources are
of psychological attachment that is important training and coaching, performance feedback and
in the occupational commitment literature.176 participative management.
Occupational commitment is the individual’s
psychological attachment to, and identification Burnout
with, his/her occupation, based on an affective
reaction to that occupation.177 Burnout is defined as ‘a persistent, negative, work-
related state in normal individuals that is primarily
The strength of employees’ occupational characterised by exhaustion, which is accompanied
commitment may have important consequences by distress, a sense of reduced effectiveness,
for organisations because of its potential links to decreased motivation and the development of
retention and work performance. Also, the salience dysfunctional attitudes and behaviours at work’.184
of occupational commitment may be on the rise as It is a psychological strain resulting from chronic
a means of coping with the uncertainty of extensive work stress, such as job demands.
organisational change.178 Workers may increase
their identification with their occupation on the The concept of burnout was initially associated
basis that they retain more control over that aspect with human services work such as health care
of their working life. and social work, where employees are engaged in
‘people’ work, for example, providing patient care.
Although there has been debate over the It is usually characterised as a three-dimensional
possible inherent conflict between occupational syndrome. Le Blanc et al.185 describe the
commitment and organisational commitment, dimensions in relation to burnout in nursing staff
more recent work argues that the two forms of in three stages:
commitment are not incompatible, but rather they
are positively associated.177 • First, high emotional demands in interpersonal
relationships with patients lead workers to feel
Work engagement emotionally exhausted.
• Second, as a coping strategy they may try
Work engagement refers to ‘a positive, fulfilling, to protect themselves by detaching from
work-related state of mind characterized by their patients by treating their patients in
vigour, dedication and absorption’.179 Vigour is an indifferent and cynical way.186,187 This
characterised by high levels of energy and mental detached attitude towards patients is called
resilience while working, a desire to invest effort in depersonalisation.
one’s work and persistence in the face of adversity. • Finally, as a result of this dysfunctional attitude,
Dedication refers to a strong psychological workers are unable to perform adequately, and
involvement in one’s work, together with feelings the quality of their care is impaired. In turn,
of pride, significance, enthusiasm and inspiration. this may lead to a decline in their feelings
Finally, absorption refers to total concentration on, of personal accomplishment or professional
and pleasant immersion in, work. efficacy.
71
More recent research has argued that burnout by describing the mutual obligations or exchange
exists outside the realm of human services. relationship perceived by the employee to bind him
Consequently, the concept has been adapted or her with the organisation.197 The individual’s
to apply to all employees, by replacing the psychological contract reflects his of her
depersonalisation dimension with cynicism, which understanding of each party’s obligations, derived
reflects indifference towards work in general, not in part from HRM practices, and also verbal
necessarily towards other people.188 statements and actions each party has made.
Burnout may have negative consequences not just Psychological contracts have important implications
for the individual (such as health problems), but for employee attitudes and behaviours, especially
also in relation to organisational outcomes such as when employees perceive that obligations
absenteeism and performance. or promises explicitly or implicitly made by
an employer have been unfulfilled. Contract
Job involvement perceptions are associated with attitudes such as
mistrust, employee behaviour (including effort and
The concept of job involvement has been the contribution), frequency of absence, and intention
subject of a large volume of research for over to stay or leave.198
40 years. Although it is subject to some definitional
confusion, Brown189 notes that most research Organisational justice
has followed Lawler and Hall’s190 definition of
job involvement as ‘psychological identification Organisational justice is one of the most popular
with one’s work’ and ‘the degree to which the job research topics in HRM, organisational psychology
situation is central to the person and his [or her] and organisational behaviour.199 The literature
identity’. The more recent concept of employee has ‘grown around attempts to describe and
engagement treads on similar territory. explain the role of fairness as a consideration in
the workplace’,200 and much of the research in
Multiple outcomes are often attributed to high this area starts from the premise that employees’
job involvement. Job involvement has been judgements of fairness are important in
considered important for promoting employee determining organisational outcomes such as job
motivation,191 goal-directed behaviour, personal satisfaction, organisational commitment, intention
growth and satisfaction, and consequently to quit and OCB.201 What is just or unjust is based
performance.190,192,193 For example, Janssen194 on individuals’ perceptions of fairness.
describes it as ‘the key that unlocks motivation’
to exert extra effort. Increasing job involvement Organisational justice research has primarily been
has been advocated as a means to ‘enhance concerned with distributive justice and procedural
organisational effectiveness and productivity by justice. Distributive justice deals with perceived
engaging employees more completely in their fairness of the distribution of outcomes, such as
work and making work a more meaningful and pay and promotion decisions, whereas procedural
fulfilling experience’.189 Developing enriched jobs justice focuses on the fairness of the process by
is regarded as an important means for stimulating which outcomes are determined.202 The process for
job involvement.192,195 making organisational decisions may be equally or
more important to employees as their perception
Turnover intentions of outcome fairness.
Morrell et al.196 describe turnover as ‘voluntary Interactional justice,203 as more recently advanced
cessation of membership of an organisation by an in the justice literature, relates to the quality of
employee of that organisation’. Turnover intention the interpersonal treatment people receive when
or intention to quit is the psychological state that procedures are implemented, and can be seen as
precedes this cessation. taking two forms.204,205 First, interpersonal justice,
which pertains to the degree to individuals feel
Psychological contract they are treated with respect, politeness and
honesty in the process. The second, informational
The relationship between employees and justice, reflects employees’ perceptions about
employers is central to HRM. The concept of the whether they feel that they are provided with
psychological contract explores this relationship adequate explanations for the use of procedures or
for outcome distribution.
72
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Employees who perceive themselves to be An important distinction has been made between
supported and valued are more likely to respond psychological climate and organisational
in a favourable manner. Organisational support climate.213 Psychological climate is based on
theory proposes three psychological mechanisms individuals’ perceptions of the organisation, but
underpinning the consequences of support.208 when individuals in the same unit (e.g. team,
First, on the basis of reciprocity, organisational department) or organisation agree or share their
support should produce an obligation in employees perceptions, organisational climate emerges,
to care about the organisation’s welfare and goal represented by the mean score across all employees
attainment. Second, the support and respect within a particular unit or organisation. The
provided by the organisation should satisfy rationale behind aggregating individual data to
socioemotional needs, leading to employees a unit level is the assumption that organisational
psychologically identifying with the organisation. collectives have their own distinct climate.
Third, perceived organisational support should
increase employees’ beliefs that the organisation Culture and climate are similar concepts since
recognises increased performance. These processes both describe employees’ experiences of their
should result in positive attitudes (e.g. commitment organisations. Organisational climate, according
and satisfaction) and behaviours (e.g. reduced to Schneider,214 represents the descriptions
turnover, increased performance). of the things that happen to employees in an
organisation, they tell us what is happening,
Perceived organisational support has been found whereas culture comes to light when employees
to be related to, but distinct from, organisational are asked why these patterns exist. The question
commitment, procedural justice and job is answered in relation to shared values, common
satisfaction.208 assumptions, and patterns of beliefs held by
organisational members and it is these which
Organisational climate define organisational culture.
Central to most, if not all, models of organisational Schneider214–216 eschews the use of general
behaviour are employee perceptions of the work multidimensional measures of climate and argues
environment, referred to generally as organisational for a facet-specific climate approach where
climate.209 Primarily understood as an intervening climate has a focus and is tied to something of
variable between the context of an organisation interest. Schneider suggests that the dimensions
and the behaviour of its members, and attempting of organisational climate will differ depending on
to understand how employees experience their the purpose of the investigation and that general
organisations, the concept has inspired many measures of organisational climate will contain
descriptions and operationalisations. The dimensions that are not relevant for each specific
dominant approach conceptualises climate as study. This line of argument has encouraged the
employees’ perceptions of what their organisation development of measures with a particular strategic
is like in terms of its routines, practices, procedures focus such as service217 and safety,218 which are
and rewards.210
73
74
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Chapter 6
The impact of HRM practices on intermediate
outcomes among health professionals
Introduction comparison was specified, but a longitudinal
filter was developed and used in the search. No
One objective of this review is to evaluate whether specific outcomes were stipulated.
HRM practices have an impact on intermediate • The second, a more specific search for HRM
outcomes, such as job satisfaction. The findings and patient outcomes (reported fully in
presented in Chapter 6 focus exclusively on Chapter 7) used a more focused search strategy
evidence from the health sector and examine the in which the population was working adults,
extent to which HRM practices have an impact on the intervention was fixed as one of the broad
intermediate outcomes. Specifically, it addresses: categories of HRM practice (see Chapter 3 for
details) and the outcomes identified as a range
• Objective 5: 5.1 What is the evidence for the of patient outcomes (again, described fully in
impact of HRM practices on the intermediate Chapter 7).
outcomes identified for this review?
In both searches there was no exclusion of studies
This chapter provides a narrative summary for based on formal quality assessment. However,
relevant longitudinal research relating specifically longitudinal study design was a criterion for
to the health sector and health professionals. This inclusion in the review and served as a form of
synthesis also enables a comparison of this research quality assessment. This ensured that the search
and its findings with that of the non-health sector. was focused on causal relationships.
This chapter first summarises the specific method The search strategies and methodology used for
for this part of the review then gives an overview of identifying longitudinal studies in Chapters 7 and
the intermediate outcomes for which evidence was 8 therefore apply here also; only the inclusion
found. Finally, narrative syntheses and data tables criteria are different. To be included in this part
are presented for each intermediate outcome. of the review, studies had to satisfy the following
criteria.
75
76
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
(1982)228 concurrent controls Medical vs surgical ward performance; affective Goldberg (1978)239 satisfaction than medical wards, and
UK 164 student nurses duty states; somatic symptoms; male wards were higher than female
social dysfunctions; wards
13 weeks Male vs female ward depression; anxiety;
duty Other outcomes mixed
sickness/absence
Jung et al. Prospective cohort with Workload No Job satisfaction; nurses Job satisfaction: Price and No effect: Work delegated increased
(1994)232 historical controls redistribution ability to supervise nursing Mueller (1986)241 with no significant effect on
USA 265 nurses programme: decrease assistants, coordination outcomes
nurse workload, of care; quality of nursing
21 months increase use and care; workload; patient
productivity of nursing satisfaction
assistants, educate
nurses to direct and
supervise ancillary staff
Nielsen et Prospective cohort with Psychological demands No Job satisfaction; self-rated Marmot et al. (1991)242 No effect: Job satisfaction
al. (2002)229 concurrent controls Control (decision health; perceived stress; Positive effect: Decision authority
Denmark 2068 employees authority and skill absence from work; labour had positive impact on self-rated
discretion) turnover health, perceived stress and
5 years
Social support absence. Skill discretion had positive
impact on health and absence but
Meaning of work no impact on perceived stress
Predictability Negative effect: Psychological
demands had a negative impact on
self-rated health
Study, Design, sample size, Fidelity
country duration Intervention(s) measure Outcomes Outcome measure(s) Results +ve/–ve
de Jonge et Prospective cohort with Job demand No Job satisfaction Scale for job satisfaction Positive effect: Workplace social
al. (2001)233 historical controls Job autonomy Emotional exhaustion, work was constructed by the support was positively related to
Netherlands 261 employees motivation authors job satisfaction
Workplace social
12 months support Negative effect: Job demand was
DOI: 10.3310/hta14510
JDI, job descriptive index; JDS, Job Diagnostic Survey; MPS, motivating potential score; QICC, quality indicator for client care.
Health Technology Assessment 2010; Vol. 14: No. 51
79
80
TABLE 24 Burnout and morale
Australia Unknown strategic planning committee care performance; Average waiting time (against outcomes
consisting of physicians patient satisfaction benchmark)
3 years and nursing staff reviewing
working practice. Areas in Time to craniotomy for patients
ED defined by function, staff with acute head injuries
work in functional teams, Patient and relative satisfaction
staff redeployed to functional judged by number of complaints
area (acute, emergency,
observation) of greatest need
Lovgren et al. Prospective cohort with Introduction of a new No – but Burnout; job Pines et al. 1981 (p. 204), 250 Negative effect: Increase in
(2002)237 historical controls care policy, which made strategies such satisfaction; creative Sandman 1992;238 Karasek and burnout
Sweden 225 health staff statements about greater as training were and innovative climate Theorell 1990;246 Rasmussen Significant reduction in
work satisfaction and an open put in place to aid 1996;247 Ekvall et al. 1983 (p. satisfaction, and decrease in
1995–8 climate implementation 204)251 climate on 2–3 of the wards
studied
JDI, Job Descriptive Index; JDS, Job Diagnostic Survey; MPS, motivating potential score.
Outcomes Conclusions
The study evaluated creative and innovative climate
in the setting using Ekvall et al.’s251 instrument Thirteen different studies from the longitudinal
for creative organisational climate. The study also literature identified for Chapters 7 and 8 examined
measured the outcomes of job satisfaction and the relationship between an HRM practice or
burnout. job characteristic and one of the intermediate
outcomes selected by this review within the health
Results sector. All studies examined some facet of work
The introduction of this particular new care policy design, such as teamworking,248,249 reallocation or
had a significant negative effect on all the staff reorganisation of work,228,230–232,237 job demands and
outcomes studied, including creative organisational control,229,233 role conflict,234,235 and variation252 or
climate. task complexity.236 A subset assessed the impact of
additional practices on intermediate outcomes.
Omissions These practices were pay234,236 and employee
Although widely studied in other settings, involvement.235 Consequently, most HRM practices
longitudinal research on climate in the health we have considered in this review, i.e. recruitment
sector is rare. and selection, staffing, training and development,
communication, family friendly, harmonisation and
Summary performance management have not been subject
The included study found that the introduction of to longitudinal research in the health sector that
a new care policy had a significant negative effect has examined their relationship with the relatively
in terms of a decrease in climate scores across most common intermediate outcomes described above.
of the wards studied within this setting of a Swedish
hospital, contrary to expectations. Very few intermediate outcomes of potential
relevance were studied: job satisfaction, burnout,
82
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
83
Chapter 7
The impact of HRM practices on final
outcomes in the health sector
Introduction The tender specified ‘patient’ outcome variously as
patient care and patient outcomes, which include
This chapter presents findings from a systematic patient satisfaction, mortality rates, hospital acute
review of the evidence on the impact of HRM infection rates, clinical quality standards, and
practices on final outcomes in the health sector. patient safety and well-being. However, a more
The findings reported here relate to objective six of specific definition of patient care outcomes is not
the overall review, specifically: given. An earlier report commissioned by the
SDO on the impact of HRM in the health sector77
• Objective 6: 6.2 Which HRM practices have an failed to examine the relationship between HRM
impact on patient care outcomes? practices and clearly defined patient care outcomes.
The type and nature of the patient care outcomes
Chapter 7 describes aspects of the methodology therefore lacked definition. Scoping exercises were
specific to the health sector systematic review. It undertaken to examine a broad spectrum of patient
then presents the narrative syntheses and data outcomes that appear in the medical and health
tables summarising the evidence for the impact of literature, including clinical outcomes, which are
HRM outcomes on patient care outcomes. treatment- or condition-specific. This literature was
vast and unmanageable, and outcomes were diverse
The narrative syntheses and data tables drawing and heterogeneous. It was decided to prioritise a
together evidence for the impact of HRM on set of patient outcomes based on Department of
intermediate outcomes in the health sector are Health and NHS performance literature.
presented in Chapter 6.
Policy-makers and practitioners want information
Chapter 2 of this report outlines the systematic that is locally applicable, timely and relevant.52,263
review methodology approaches used to guide all Given that the invitation to tender for this
the reviews. The search strategies for this review research described the target audience as ‘NHS
were based on the broad HRM categories described managers, HR practitioners and researchers, …
in Chapter 3 and on patient outcomes. commissioners, … and the NHS HR community
as a whole’, it was decided to focus on patient
In this chapter, the classification of patient care outcomes of relevance to these groups.
outcomes is described first then the specific The Healthcare Commission NHS Performance
methodology is described. The chapter then goes Indicators264 and the NHS Improvement Plan265
on to describe the results. therefore provided a potentially good rationale
for examining a limited, but specific, list of patient
care outcomes.
Patient outcomes
This review focused on seven patient outcomes
One of two principal questions to be addressed is derived from the Healthcare Commission
whether there is any demonstrable link between NHS Performance Indicators264 and the NHS
HRM practices and final performance outcomes in Improvement Plan265 (Table 28). There are no
health, and between intermediate outcomes (such standard definitions for any of these outcomes.
as morale and motivation) and final performance Definitions used in this review are either those
outcomes in health. A clear definition of this provided by the Healthcare Commission264
particular outcome offers the only viable approach and NHS Improvement Plan265 or the terms
to dealing with this question and reducing the as they appear in the health services research
number of outcomes to be investigated. literature. For example, patient satisfaction has
85
no standardised measure, but is a self-explanatory help maintain the profile of infection control.’
term widely used in the literature. Definitions This outcome does not include adverse events or
appearing in this review’s protocols relating to medication errors.
patient care outcomes, and the related literature
searching, are perhaps broader and more vague Patient centred-care refers to ‘health care that is
than would be required by standard systematic provided in partnership with patients, their carers
review methodology, but this reflects the non- and relatives, respecting their diverse needs,
standardised nature of the outcomes being preferences and choices’ (NHS 4). This involves
measured. Further research may add to this list, giving ‘patients and carers information when
but it provides a sound starting point for looking they want or need it, seeking their feedback and
at specific patient care outcomes of demonstrable consulting them regarding their care, and treating
relevance to the target audience. their choices regarding their care or food with
respect’ (NHS 4). This also includes taking into
Patient care outcomes considered by the review are: account their views and choices regarding access to
services and treatments (NHS 5). It is measured in
Patient safety refers to ‘activities that prevent or two indicators under the Patient Focus standards
reduce the risk of harm to patients’ (NHS 1). This for Acute Trusts. These are the ‘Outpatient and
includes activities to reduce ‘the risk of hospital A&E patient surveys: better information, more
acquired infection’ and MRSA, and to achieve choice’ and ‘Outpatient and A&E patient surveys:
high standards of hygiene and cleanliness (NHS 1 building closer relationships’ indicators. The
and 6). MRSA is also a standards indicator under former measures ‘patients’ experience of these
the Clinical Focus element of the Healthcare aspects of care, such as information about care and
Commission Performance Indicators, which treatment, and involvement in care’, and the latter,
states, ‘Infection control (IC) is an integral part ‘feedback from patients … taking account of their
of healthcare and the NHS is working to improve views and priorities … and placing the patient at
hospital cleanliness and reduce infection rates. the centre of health services’.
Rates of infection with MRSA are an increasing
public concern and reducing them is a key priority Patient waiting times refers to ‘delay at … [the]
for the Government. Including this indicator will stage of service delivery’ (NHS 5). It is an
86
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
integral element of the Healthcare Commission have their limitations.278 For example, Mitchell
Performance Indicators. Waiting times makes and Shortell279 point out that mortality may be
up five of the eight Key Targets and 10 of the a good ‘bottom-line outcome’, but it is arguably
16 Patient Focus Targets for Acute Trusts. These more dependent on patient characteristics than
targets range from standards regarding waiting organisational factors, such as HRM practices.
times in A&E and waiting times from GP referral to
consultation, to the number of cancelled operations
and delays in transfers. They also constitute three Review methodology
of the eight Key Targets and 3 of the 13 Access to
Protocol development
Quality Services indicators for Primary Care Trusts.
Protocols were produced for the review question
Patient quality of life is a common measure in ‘Which HRM practices have an impact on patient
the health services research literature.268,269 It care outcomes?’ but much was left open because
is measured in the Healthcare Commission of the lack of clear definitions and because the
Performance Indicators using ‘feedback from literature itself was an unknown quantity. The
patients … taking account of their views and literature was so diverse and heterogeneous that
priorities’. Patient surveys and complaints nothing was fixed beyond the broad categories
constitute 5 of the 16 Patient Focus Targets for of HRM practice (Chapter 3) and the patient
Acute Trusts, and 5 of the 13 Access to Quality outcomes described above, and the plan for the
Services indicators for Primary Care Trusts. systematic search of the literature by various
methods. There was to be no exclusion of studies
Patient satisfaction is a common measure in based on formal quality assessment. However,
health services research270–273 and is covered by a longitudinal study design was a criterion for
substantial number of the Healthcare Commission inclusion in the review and served as a form
Performance Indicators. These indicators measure of quality assessment. The details of the data
‘feedback from patients … taking account of their extraction and synthesis remained open and
views and priorities’. Satisfaction, as with patient flexible. Such flexibility is permitted by evolving
quality of life, makes up 5 of the 16 Patient Focus systematic review methods, especially when dealing
Targets for Acute Trusts, and 5 of the 13 Access to with complex literatures.52,56,58,60,63
Quality Services indicators for Primary Care Trusts.
Search strategy
Patient mortality rates is a common measure in
the health services research literature274,275 and is Literature search strategies were developed and
covered by 2 of the 10 Clinical Focus standards performed on the following electronic databases:
of the Healthcare Commission Performance ASSIA, BNI, Business Source Premier, Campbell
Indicators for Acute Trusts, and two of the nine Collaboration, CENTRAL, CDSR, CINAHL,
Improving Health indicators for Primary Care DARE, DH-Data, EMBASE, HMIC, IBSS, King’s
Trusts. Fund database, MEDLINE, NHS EED, NRR,
PREMEDLINE, PsycINFO, ReFeR, SSCI and SCI.
Patients’ length of stay and re-admission rates Search strings were modified for the different
are common measures in the health services databases to take account of the different keywords
research literature.274,276,277 Re-admissions that are and thesauri they use. Examples of the search
‘potentially avoidable’ are also covered by 2 of filters and strategies can be found in Appendices 6
the 10 Clinical Focus standards of the Healthcare and 7. The searching of electronic databases using
Commission Performance Indicators for Acute pre-designed search filters was supplemented by
Trusts. reference tracking of selected studies.
87
which were rejected when the full paper was the health sector and is not used to describe any
retrieved because they did not meet the inclusion of the interventions in this sample of longitudinal
criteria. Therefore, in total, reference tracking research. However, the interventions fall under the
yielded 21 additional references. A member of heading of ‘work or job design’ because they all
the project team also encountered a relevant involve the redesign or extension of job content.
paper while working on another project. Overall, The absence of any existing systematic review or
therefore, 99 studies were identified (Table 29). meta-analysis of health sector studies described
The QUOROM flow chart280 in Figure 3 explains as work design may be explained by difficulties
how many studies were rejected and at which stage, in creating a clear, structured definition of work
and the means by which included studies were design or operationalising the concept as a search
identified. strategy. It may also have been prevented by the
potentially limited number of relevant studies
Data extraction was performed on the resulting and the heterogeneity of their interventions and
99 studies, which were then categorised by outcomes, as demonstrated by this sample. In
intervention following the framework of broad fact, there is only one meta-analysis covering job
HRM practices (developed earlier in the review and enrichment, an element of work design, in the non-
described in Chapter 3). The tables below describe health sector.281
the key data for the studies and the accompanying
narrative offers a basic synthesis of these data. Details of studies
Some studies appear under more than one Fourteen studies met the inclusion criteria (Table
category, either because the interventions being 30). All studies were published in English, with
examined have a number of different components the earliest being published in 1992. Eight studies
or because the nature of the intervention means it were conducted in the USA, four in the UK, one in
can be justifiably categorised under more than one Hong Kong and one in Australia. One study used
heading, for example education and feedback. an experimental design, but most of the studies
used a prospective longitudinal design, and there
Work design were also three retrospective studies. The studies
varied in duration, from 6 weeks to 3 years. The
All of these studies consider interventions or sample sizes ranged from as few as 10 staff or 38
practices changing or affecting the content of a job patients to 879 staff or 5037 patients, depending
or the way it is to be done. The term ‘work design’ on the outcomes being studied. In three studies
is rarely used in the literature relating to HRM in the relevant patient sample size was not given.
Systematic reviews:
HRM practice Subcategories No. of studies health sector
Work design 14 0
Staffing Staffing levels and working 19 2
hours
Innovation 13 5
Substitution 23 5
Transfer 8 5
Training and development 16 5
Compensation and rewards 3 7
Communication 0 0
Family friendly 0 0
Single status 0 0
Employee involvement 5 0
Performance management, 7 0
appraisal
Bundles 0 0
89
References satisfying
inclusion criteria from
Serendipity (n = 1)
reference tracking of
included papers (n = 21)
Final total of studies
satisfying inclusion
criteria (n = 99)
In all cases the unit of analysis consisted of units staff members to facilitate the revision or extension
or specific services or departments in hospitals, of roles.232,248
including emergency departments.
The remaining studies all involved elements of
HRM practice job redesign, including within the interventions
All of the interventions considered here involved components such as team working231,286 and
the introduction of new systems or programmes redesigned task allocation.231,287,288 One study
affecting either who performed what tasks within a measured the fidelity with which the new systems
team or how work was allocated to team members. were implemented by recording when protocols
However, analysis of the literature reveals that some required by the new system were activated.284
of the interventions studied may be placed into two However, none of the other studies reported any
groups, each containing three studies. measures of implementation fidelity for the systems
or programmes being examined.
In the first group, three studies, two by the
same authors, examined the implementation Outcomes
of an outcomes management model of working The three outcomes management studies all
developed from quality circles: jobs were researched the impact of this model on patient
redesigned with the intention of achieving outcome length of stay, but had no other outcomes in
benchmarks.282–284 common. Two of the three studies evaluating skill
enhancement for NPs measured both patient
In the second group of three studies, the satisfaction and staff satisfaction. One study used
intervention focused on skill enhancement of nurse validated scales for both of these outcomes,232 but
practitioners (NPs), i.e. nurses taking on expanded the other study used its own instruments.248 The
roles. This involved an increase in workload and remaining seven outcomes measured by these
the assumption of more duties by these staff, studies were all different. The latter study also
or a redistribution of workload combined with assessed other staff psychological and behavioural
increased supervision of more junior members outcomes, including the ability to supervise staff
of staff.232,248,285 In two of these studies, this job and workload.
90 redesign was accompanied by education of relevant
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
The remaining studies measured a wide variety and retrospective designs all produced the same
of patient care outcomes. Patient satisfaction was inconsistent results.
measured by five studies and waiting times by four.
Overall, unlike the majority of other interventions Omissions
considered by the longitudinal research included The unit of analysis used by these studies did not
in this review, several work design studies did include the hospital or primary or community
examine staff psychological outcomes. Six studies care. Several staff outcomes were measured
measured job satisfaction, employee satisfaction or by these studies, but no study controlled for
morale. None of these outcomes was analysed as a any staff variables and none examined these
moderator or mediator but only as direct outcomes, staff psychological or behavioural variables as
and the job and employee satisfaction measures moderators or mediators of the relationships
used by these studies were all different. being evaluated. Unlike many other interventions
considered by this review, only two patient safety
Results indicators were examined as an outcome by any
Two of the three studies evaluating the outcomes of these studies: errors and urinary tract infection
management model found that this intervention (UTI) rates; and only one study considered
had a positive effect on all outcomes, which mortality rates.
included length of stay and patient mortality
rates.282,283 However, the third study found that the Summary
impact of this intervention on outcomes such as The job design interventions considered here had
length of stay was positive, but not significant.284 All certain basic elements in common, but they also
three studies used a prospective cohort design with differed markedly in parts of their content. Neither
historical controls. outcomes management nor skill enhancement
interventions demonstrated consistent results,
In the group of skills enhancement studies, one but other job and work design interventions all
intervention (workload redistribution) had no had positive effects on staff satisfaction, patient
significant positive or adverse effect on staff or satisfaction and patient waiting times.
patient satisfaction.232 A second study, using a
superior experimental design, also examining the Staffing level and working hours
relationship between increased role responsibility
and these outcomes, found that the intervention These studies consider research into two elements
had a positive effect on morale, but no significant of staffing practice: staffing level and staff working
impact on patient satisfaction.248 The weakest hours. The staffing level literature considers
study in this group in terms of study design only whether increased ratios of staff–patients improve
measured the impact of the intervention on length patient care outcomes or whether decreased
of stay, and found a significant positive effect.285 ratios have an adverse effect on these outcomes.
This ratio may relate either to numbers of staff
In the remaining studies, the work redesign or to grades of staff, such as registered nurses
intervention significantly improved patient (RNs). There is one systematic review of studies
satisfaction in almost every study that measured of staffing levels in the health sector.295 Two of the
this outcome,230,249,287,289 although one study longitudinal studies described below have been
found that the provision of continuity of care included in this previous systematic review.296,297
by physicians, as opposed to their rotation, However, the majority have not been included in
made no difference to patient satisfaction or existing systematic reviews and therefore add to
patients’ perception of the atmosphere in a the review literature on this topic. There are no
ward.288 However, all three studies examining the systematic reviews or meta-analyses on staffing
impact of interventions on staff psychological levels in the non-health sector HRM literature.
outcomes found a significant positive impact
on these outcomes,230,231,249 and all four studies The working hours literature considers both
examining waiting times found significant positive the variable of total numbers of hours worked
outcomes.249,286,287,290 In the four studies examining and interventions that affect the number of
financial performance or cost as an outcome, the hours worked by health professionals, especially
results were positive for three of the interventions, reductions or rescheduling. There are two meta-
and only had no effect in the case of re-engineering analyses that include studies of health sector
work design for patient-focused care.230 Prospective working hours, reflecting the sizeable number of
91
Design, sample
Study, country size, duration Intervention(s) FM Outcomes Outcome measure(s) Results +ve/–ve
Burns et al. Prospective cohort Outcomes No Ventilator duration Ventilator duration: Data collection by OMs Positive effect: All
(2003)282 with historical management: developed ICU length of stay ICU length of stay: Data collection by OMs outcomes
USA controls from quality circles:
pathway, protocols, Hospital length of stay; Hospital length of stay: Data collection by
1105 patients mortality; costs OMs
four nurses as OMs
30 months Mortality: Data collection by OMs
Costs: Routine collection
Russell et al. Prospective cohort Outcomes No details Length of stay; rates of Length of stay: Routine data collection Positive effect:
(2002)283 with historical management: UTI and skin breakdown; Rates of UTI and skin breakdown: Routine Significant
USA controls (generated from quality discontinuation of catheter data collection improvements on all
524 patients circle): and mobilisation; cost savings outcomes
Discontinuation of catheter and
OM (NPs), pathway and
The impact of HRM practices in the health sector
USA controls programme: decrease satisfaction; coordination Job satisfaction: Price and Mueller (1986)241 with no significant
265 nurses nurse workload, of care; quality of nursing effect on outcomes
increase use and care; workload; patient Coordination of care: Price and Mueller
21 months productivity of nursing satisfaction; final outcomes: (1986)241
assistants, educate job satisfaction Quality of nursing care: Atwood and
nurses to direct and Hinshaw (1988)291
supervise ancillary staff Workload: Atwood and Hinshaw (1988)291
Patient satisfaction: Press–Ganey Survey
Schweer et al. Retrospective Role of trauma No Length of stay Positive effect:
(2004)285 Unknown paediatric NPs Significant decrease
USA (trauma PNP) was re- in LOS
3 years engineered to a joint
practice model with
trauma PNP assuming
continued
Health Technology Assessment 2010; Vol. 14: No. 51
93
94
TABLE 30 Work design (continued)
Design, sample
Study, country size, duration Intervention(s) FM Outcomes Outcome measure(s) Results +ve/–ve
Browne et al. Prospective cohort Seamless model of ED No Patient waiting times Average waiting time (against benchmark) Positive effect:
(2000)249 with historical management: designed Critical care performance Time to craniotomy for patients with acute Significant positive
Australia controls by a strategic planning head injuries effects on all outcomes
committee consisting of Patient satisfaction
Unknown Patient and relative satisfaction judged by
physicians and nursing Staff morale; final outcome:
3 years staff reviewing working staff morale number of complaints
practice. Areas in ED Turnover rate and sick leave of 2 days or less
defined by function,
staff work in functional
teams, staff redeployed
to functional area
The impact of HRM practices in the health sector
(acute, emergency,
observation) of greatest
need
Jones et al. Prospective cohort Re-engineered work No Patient satisfaction; employee Press–Ganey Survey Positive effect:
(1997)230 with historical design for patient satisfaction; collaboration; Attitude survey programme for Health Care Employee satisfaction,
USA controls focused care; use cost of care; length of stay; (1985)240 patient satisfaction,
119 staff of multidisciplinary final outcome: employee length of stay
collaboration satisfaction Routine data collection
Patient numbers No effect: Costs
unknown Negative effect:
15 months Collaboration
decreased
O’Connor Prospective cohort Daily patient allocation No Staff perception of changes; Staff perception of changes: job satisfaction: Positive effect: All
(1993)231 with historical (old system) vs team activity analysis; QICC; final 20-item scale (own scale)
UK controls nursing (continuity of outcomes: job satisfaction; Activity analysis: observation exercise
38 patients care) (new system) staff activities
QICC: 85-item questionnaire based on
10 staff: activity Goldstone et al. (1984)292
24 staff: satisfaction
6 months
DOI: 10.3310/hta14510
Design, sample
Study, country size, duration Intervention(s) FM Outcomes Outcome measure(s) Results +ve/–ve
Lau and Leung Prospective cohort Small team consultation No Waiting times Interval between registration and being seen Positive effect:
(1997)286 with historical vs common patient pool by doctor Significant reduction in
Hong Kong controls waiting times
2583 patients
8 months
Sparr et al. Retrospective House officer rotation No Patient satisfaction; Patient satisfaction questionnaire (PSS) No effect
(1994)288 120 patients vs no rotation patient perception of ward (Speedling et al. 1983);293 revised version of
USA atmosphere Ward Atmosphere Scale (WAS) (1974)294
6 weeks
CCC, clinical care coordinators; ED, emergency department; ICU, intensive care unit; LOS, length of stay; MIU, Minor Injuries Unit; OM, outcomes manager; QICC, quality
indicator for client care; trauma PNP, trauma paediatric NP.
Health Technology Assessment 2010; Vol. 14: No. 51
95
The impact of HRM practices in the health sector
studies in this field.298,299 However, none of the measured the patient safety variable ‘unplanned
studies described below has been included in these extubations’. The remaining study from this cluster
existing reviews. This sample therefore adds to considered the staff outcome: medically certified
the review literature on this topic. There is only sick leave.
one meta-analysis in the non-health sector HRM
literature reviewing studies on a similar subject.300 The studies comparing different lengths of working
hours can be separated further into several smaller
Details of studies groups examining three or more of the same
Nineteen studies met the inclusion criteria (Table or similar outcomes. Four studies investigated
31). All were published in English, with the earliest the relationship between this intervention and
being published in 1988, and all but three of the mortality rates. In only one study was this the only
studies were conducted in North America. Most outcome; the remainder all evaluated multiple
of the studies used a retrospective or prospective outcomes. Three examined the relationship with
longitudinal design, but there was one RCT.301 length of stay, medication errors and adverse
The studies varied in duration, from 3 months events, although these last-named outcomes were
to 7 years. The sample sizes ranged from 347 to only nominally the same. In total, an additional 15
more than 300,000 patients, although in five cases different outcomes were also examined by these
the study’s sample size was not given. In terms of studies. Five of the seven studies assessed multiple
units of analysis, seven studies used the hospital as outcomes rather than a single outcome.
the unit and 12 used units or specific services or
departments in hospitals. Results
Two of the studies examining the impact of total
HRM practice number of hours worked found that this variable
Analysis of the literature reveals that these had no impact on patient satisfaction,297,305 and
longitudinal studies may be examined in two others found no significant relationship with
three separate groups: six studies examine the mortality or prescribing errors.296,306 Only one study
relationship between patient care outcomes and found lower numbers of hours to be associated with
the total number of hours worked by staff; six lower satisfaction among patients, and one study
studies look at how nurse or medical staff–patient found that higher nurse intensity, that is, more
ratios impact on patient care outcomes; and seven hours, led to decreased length of stay.307 All of these
studies compare the effect of shifts of different studies used either retrospective or prospective
lengths being worked by physicians, for example study designs.
long hours compared with an overall reduced
total numbers of hours or shifts of shorter hours. The results for the four nurse–patient ratio studies
In the case of this last grouping of studies, the examining the outcome of mortality rates were
intervention was often introduced as a consequence inconsistent. The two prospective studies found
of new working hour limits. The fidelity with which that there was no significant association with
these interventions were implemented was clearly mortality rates regardless of the ratio,308,309 while
measured in only one study,302 but there was also the two studies using a less powerful retrospective
some monitoring of compliance in another.303 design did find that higher ratios of staff–patient
produced lower mortality rates.310,311 The results
Outcomes of these latter two studies may also have been
In the first two groups, all but one of the studies determined by their patient population of newborn
examined the relationship between staffing or children, which was very different from the adult
working hours and a single patient care outcome. populations of the two studies that reported no
Three of the studies that examined total number of effect. A lower nurse–patient ratio was found to
hours worked all looked at the outcome of patient have the negative effect of increased unplanned
satisfaction. These three studies used two different extubations and increased sick leave absences by
measures and the two studies which used the same head nurses.312,313
patient satisfaction measure reported different
findings.297,304 The three other studies examining Two of the seven studies examining reduced
total number of hours worked looked at three very working hours recorded some negative outcomes,
different outcomes; length of stay, mortality and such as an increase in certain complications,
prescribing errors. Four of the nurse–patient ratio negative patient safety indicators and delays in
studies looked at the same outcome: in-hospital the provision of some diagnostic tests. However,
mortality. One of the nurse–patient ratio studies reduction in working hours consistently appears
96
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
not to have any adverse effect on the patient liaison (using specialists in one sector to educate
outcomes of mortality rates, length of stay, transfers and support staff in another)’.318 The following
to intensive care, discharge disposition, prescribing synthesis focuses on studies researching innovation:
or medication errors, complications, re-admissions, the introduction of a new type of worker to existing
reinterventions, cost per patient, adverse events or teams or services. There are five recent systematic
perception of care.302,303,314–316 In some cases, this reviews or meta-analyses on staffing in the health
intervention was associated with positive outcomes sector workforce, reflecting both the sizeable
regarding length of stay, medication errors and number of studies and the growing interest in this
maternal and neonatal health outcomes.260,301,314 field.295,319–322 However, only one of the studies
This was the same across study designs: RCTs described below has been included in any of these
generated the same outcomes as simple prospective systematic reviews.323 There is no meta-analysis in
or retrospective designs. In fact, some of the the non-health sector HRM literature reviewing
positive outcomes were found by the more powerful studies on a similar subject.
RCT.301
Details of studies
Omissions Thirteen studies met the inclusion criteria (Table
None of the studies reported here took emergency 32). All were published in English, with the earliest
departments or primary or community care being published in 1988. Eight studies were
as its unit of analysis. Also, staff psychological conducted in the USA, two from Australia, and
or behavioural outcome (sick leave) was only one each in the UK, Switzerland and Turkey. Of
measured by one of these studies, but no study the 13 studies, one was experimental,324 nine were
controlled for any staff variables and none prospective in design and three were retrospective.
examined these variables as moderators or The studies varied in duration, from 1 month to
mediators of the relationships being evaluated. 5 years, and the sample sizes ranged from 203 to
6909 patients. However, no sample size was given
Summary by five studies. Almost all of these studies were
According to this sample of longitudinal research, conducted on units, wards or specific hospital
neither lower numbers of hours worked nor the services or departments. There was one study of a
reduction of working hours appears to have an community nursing home.
adverse effect on mortality rates, length of stay,
transfers to intensive care, discharge disposition, HRM practice
prescribing or medication errors, complications, Analysis of the literature reveals that the
re-admissions, reinterventions, cost per patient or interventions studied may be examined in
adverse events. These variables may potentially three groups. Five studies look at whether the
have a negative effect on patient satisfaction, introduction of physicians with specific training
although they may equally have some positive or skills has any impact on patient care outcomes.
effect on some patient outcomes. The evidence on Five studies compare the impact on patient care of
staff–patient ratios is inconsistent for mortality as teams made up of physicians alone versus teams
a patient outcome, although those studies with a composed of physicians and newly introduced
superior design suggest there is no adverse effect. NPs. Finally, three studies examine the impact of
However, the research also indicates that lower the creation of new NP or senior nurse services.
nurse–patient ratios may lead to adverse events The units of analysis were all either emergency
and increased absences among senior nursing staff. departments or intensive care units (ICUs). None
of the studies made any attempt to measure the
Introduction of a new type of fidelity with which the new staffing practice was
worker to existing teams or being implemented.
services
Outcomes
The staffing category includes skill mix at work, The five studies which looked at the introduction
which involves ‘practices affecting enhancement of new types of physicians assessed a small number
(increasing depth of job by extending roles of of consistent outcomes. Four of the five studies
skills), substitution (working across professional assessed one outcome only, either length of stay,
divides or exchanging one type of worker for mortality or discharge rates. Length of stay was
another), delegation, innovation (creating new evaluated by three of the five studies and mortality
jobs by introducing a new type of worker), transfer by two. There were only two additional patient
of services, relocation of staff and services, and outcomes.
97
3 months
Beglinger Retrospective No. of hours worked by No Length of stay Routine data collection Positive effect: Decreased
(2006)307 Unknown nurses length of stay
USA 13 years
Tourangeau et Retrospective Predictors: No 30-day mortality 30-day risk-adjusted No effect: All other predictors
al. (2002)296 46,941 patients Higher dose of nurse staffing rates mortality rates: Routine had no significant effect
Canada (greater number of nursing data collection
12 months
hours)
Sovie and Jawad Retrospective No. of hours worked per No Patient satisfaction Picker Institute Inpatient No effect
(2001)297 12 hospitals (1997) patient-day Questionnaire
USA 16 hospitals (1998) Press–Ganey Survey
24 months
Amaravadi et al. Prospective cohort with Different night time No ‘Clinical and In-hospital mortality No effect
(2000)309 concurrent controls nurse–patient ratios in ICU economic outcomes’
USA 366 patients (postoperative care)
5 years (1994–8)
Dimick et al. Prospective cohort with Nurse–patient ratio (night- No In-hospital mortality Routine data collection No effect: No association with
(2001)308 historical controls time in ICU) mortality regardless of ratio
USA 569 patients
5 years (1994–8)
Study Design, sample size,
country duration Intervention(s) FM Outcomes Outcome measure(s) Results +ve/–ve
Marcin et al. Retrospective Nurse–patient ratio No Unplanned Routine data collection Negative effect: Lower nurse–
(2005)312 1004 patients extubations patient ratio significantly
USA associated with increased
4 years unplanned extubations
DOI: 10.3310/hta14510
Callaghan et al. Retrospective Nurse–patient ratio No Mortality (survival to Mortality (survival to Positive effect: Highest infant–
(2003)310 692 patients hospital discharge) hospital discharge): staff ratio (I.71+) produced
Australia routine data collection significantly lower mortality
4 years
Bourbonnais et Retrospective Nurse–patient ratio No Medically certified Routine data collection Negative effect: Lower nurse–
al. (1992)313 1165 nurses sick leave patient ratio significantly
Canada associated with increased
41 months taking of sick leave by head
nurses (not all nurses)
Stilwell et al. Retrospective Level of staffing/all staff– No Mortality Mortality per 1000 births Positive effect: Lower staff
(1988)311 Unknown patient ratio (medical, ratio, higher mortality
UK nursing, midwives)
7 years (1977–83)
Landrigan et al. RCT Traditional schedule No Medical errors; Routine data collection Positive effect: Intervention
continued
Health Technology Assessment 2010; Vol. 14: No. 51
99
100
TABLE 31 Staffing level and working hours (continued)
(1993)316 historical controls residents’ working hours transfers to ICUs; its reliability or validity transfer to intensive care,
USA 526 patients cardiopulmonary discharge disposition and
resuscitation length of stay
13 months attempts; discharge Negative effect: More patients
disposition; length of suffered complications or had
stay; diagnostic test delays
In-hospital medical
complications; delays
in ordering and
performing diagnostic
tests
Poulose et al. Retrospective Resident work hours limit Yes: surveillance PSIs Rates per 1000 discharges No effect: For three of five
(2005)302 85,000+ patients performed to per quarter of: APL, patient safety indicators
USA evaluate compliance PEDVT, FB, PTX, Negative effect: For two of five
7 years (1995–2001) with limit (p. 848) postoperative WD patient safety indicators
Howard et al. Retrospective Residents’ work hours limits Some monitoring In-hospital mortality In-hospital mortality for No effect: Mortality rates
(2004)303 300,000+ patients of compliance, the following diagnoses: declined but no different
USA but authors CHF, AMI and iatrogenic between experiment and
4 years (1988–91) recognise this was PNU control
not universal, and Source: New York
therefore cite this SPARCS database (p. 2)
as a limitation of
the study
AMI, acute myocardial infarction; APL, accidental puncture or laceration; CHF, congestive heart failure; FB, foreign body left during procedure; ICU, intensive care unit; PEDVT,
postoperative pulmonary embolus or deep vein thrombosis; PNU, pneumonia; PSI, patient safety indicators; PTX, pneumothorax; SPARCS, Statewide Planning and Research Co-
operative System; WD, wound dehiscence.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
101
Australia 6909 patients nurses trained to perform Arrival at triage to time first seen times for most frequent
ED clinicians’ role (non- by doctor categories and numbers
12 weeks medical technicians) who leave without being
seen
Christmas et al. Prospective cohort with Surgical residents with No No. of patients No details of measure No effect (positive): on
(2005)329 historical controls no work hour limitations admitted to trauma mortality and cost per
USA Unknown vs surgical residents with service patient
limited working hours plus Hospital length of Positive effect: Reduction
3 years NPs (p. 918) stay in all lengths of stay
Intensive care stay
Floor length of stay
Mortality
Direct cost per
continued
Health Technology Assessment 2010; Vol. 14: No. 51
103
104
TABLE 32 Innovation (continued)
Australia nurse)
6 years
Spisso et al. Prospective cohort with Introduction of NPs No Cost–benefit Cost–benefit: FISCAL and routine Positive effect: Length
(1990)323 historical controls Quality of care: collected patient data of stay, waiting times
USA Unknown waiting times/length Quality of care: Waiting times/ and patient complaints
of stay; patient length of stay (routine data decrease; time saved for
24 months team increased; costs
satisfaction collection); patient satisfaction:
number of patient complaints offset
Impact of healthcare
team: time spent Impact of health-care team: team
doing certain tasks spent doing certain tasks: 30-day
review
Barr et al. Prospective cohort with NP services to cope with No Patient satisfaction Patient satisfaction Positive effect: High
(2000)332 historical controls increased demand patient satisfaction, but
UK (NI) 241 patients not a comparison with a
control
1 month
CCM, Critical Care Medicine; ED, emergency department; LOS, length of stay; PNP, paediatric nurse practitioner.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
unlicensed assistive personnel, unregulated staff and adverse events. Unlike the studies focusing
and nursing assistants. Therefore, despite the almost exclusively on nurses, there was little
apparent homogeneity of the intervention, the consideration of patient satisfaction or patient
substitutes were not always the same. However, 13 safety.
of these 14 studies focused exclusively on nursing
staff. Mortality was the only outcome to be examined by
more than one study in the group examining the
In the same way, the second largest group (a group replacement of physicians by NPs or senior nurses.
of five studies) examined the substitution of one The remaining eight outcomes were all different.
type of physician for another, such as part-time for Patient satisfaction was measured by one study;
full-time physicians. However, in each case both once again the researchers used their own scale.
the physician being replaced and his/her substitute
were different. Results
Results for interventions examining the
The final group of four studies examined the substitution of one type of nurse for another, or
replacement of physicians by NPs or senior for a more junior member of health staff, were
nurses. Once again, however, the details of each inconsistent. Two prospective studies found that
intervention were different. None of the studies the higher the ratio of RNs, the higher the patient
made any attempt to measure the fidelity with satisfaction,305,337 but another study, using a higher-
which the intervention was being implemented. quality experimental design, found that using
RNs alone was less effective at improving patient
Outcomes satisfaction than replacing some RNs with NAs.338
In the group of 14 studies examining the The use of more RNs or regular nurses compared
substitution of one type of nurse for another, or to other grades was also positively associated
for a more junior member of health staff, only with lower mortality rates by all five studies that
four patient outcome measures were examined examined this outcome, although all of these
by more than one study: patient satisfaction, studies were retrospective, which is the weakest
mortality, quality of nursing care and length of form of longitudinal study design. All-RN staffing
stay. Six studies considered the outcome of patient versus staffing by temporary nurses or other grades
satisfaction and all used different measures or was found to have a positive impact on length of
tools, only three of which were named scales (Press– stay, and, in a study of a burns unit, infection rates
Ganey Survey; Burroughs Patient Satisfaction among patients. Employee nurses were also found
Measure; Patient Judgement of Hospital Quality to have a better impact on quality of nursing care
Questionnaire). Quality of nursing care was and patient satisfaction than contract nurses.339
also only measured by a recognised instrument However, the remaining three studies found that
(Hinshaw and Atwood336 scale) in one of the two the substitution of RNs with others generally had
studies examining this outcome. Four studies no effect on outcomes.232,340,341
evaluated mortality rates as an outcome and three
length of stay. Cost was the only other outcome to Results from those studies examining the
be measured by more than one study. These twelve substitution of one physician for another were
studies also measured an additional 13 different again inconsistent. In two cases there was no effect
outcomes, several of which related to patient safety, on mortality rates,342,343 but one did have a positive
such as patient falls. Jung et al.232 also measured effect.344 However, the latter was a retrospective
a number of staff psychological and behavioural study344 and one of the studies that found no effect
outcomes, including job satisfaction, ability to was a RCT,343 the results of which may be more
supervise staff and workload. This was the only one robust. Cross-coverage by physicians was found
of the 23 substitution studies to measure any staff to have a negative effect on adverse events when
outcomes. compared to residency345 but otherwise there was
no significant effect, either positive or negative, on
The group of five studies examining the almost any of the remaining outcomes.342,343,346 Only
substitution of one physician for another made a the study of full- versus part-time primary care
more limited but consistent choice of outcomes. physicians controlled for any staff variables, but
Mortality was measured as an outcome by three these were purely demographic.346
studies, and was the only outcome examined by two
of these studies. Other outcomes measured were The studies examining the replacement of
length of stay, transfer time, patient satisfaction physicians by NPs or senior nurses generally found
105
that such a substitution had no adverse effect patients. All eight studies were conducted on units,
on the outcomes measured, regardless of study wards or specific services.
design.290,347–349 Two studies in this group also
indicated that this intervention had a positive effect HRM practice
on patient satisfaction and some waiting times, Analysis of the literature reveals that the
although these studies used less reliable prospective interventions studied may be examined in two
designs.290,348 groups.335,355–357 Four studies all examine the same
intervention: the specific transfer of services to a
Omissions particular team of physicians, to the exclusion of
All of the principal patient care outcomes were physicians who were previously involved in the care
covered by this sample of studies. However, only under a different model. These studies examined
one study measured any staff psychological or the impact of a move from open ICUs, in which the
behavioural outcome, and no study controlled physician responsible for admission retained some
for any of these variables or examined them as input into a patient’s care, to closed units in which
moderators or mediators of the relationships being care was controlled exclusively by the intensive unit
evaluated. critical care team. The remaining four studies may
also be grouped together because they examine
Summary the effect on patient care outcomes of the presence
This sample describes longitudinal research or absence of certain physicians, although in each
relating to substitution of staff members in the case the unit of analysis, and, therefore, the type
health sector literature. This research suggests that of physician, was different. None of these studies
staffing by RNs only, rather than their replacement made any attempt to measure the fidelity with
with other grades of staff, is better for mortality which the intervention was being implemented.
rates, length of stay and some infection rates
but has limited impact on other outcomes. The Outcomes
replacement of one type of physician with another The group of studies evaluating open versus closed
was also shown to have neither an adverse nor ICUs all considered mortality as an outcome. In
positive impact on patient outcomes. The most one study it was the only outcome measured, but
consistent results were found for the replacement in the remainder it was one of multiple outcomes:
of physicians with senior nursing staff: there were length of stay and duration of mechanical
no adverse effects as a result of this substitution. ventilation were examined as outcomes by two of
these studies, but eight further different outcomes
Transfer or relocation of staff were also studied.
and services
In the same way, in the remaining four studies, two
The following synthesis focuses on studies outcomes were examined by two studies: mortality
researching transfer or relocation of services. There rates and length of stay.253,358 Mortality was the only
are five recent systematic reviews or meta-analyses outcome measured by one study.359 The remaining
on staffing in the health sector workforce, reflecting studies measured an additional nine different
the sizeable number of studies in this field and outcomes. Only one study focused on patients’
the growing interest in this field.295,319–322 However, perceptions of quality of care and used the Picker
only one of the studies described below has been Commonwealth Institute Questionnaire.253
included in any of these systematic reviews.335
There is no meta-analysis in the non-health sector Results
HRM literature reviewing studies on a similar The studies examining the impact of the presence
subject. or absence of a physician found inconsistent results
regarding mortality rates and waiting times,335,355,357
Details of studies and found that the practice had no effect at all
Eight studies met the inclusion criteria (Table 34). on length of stay.356,357 These studies all used a
All were published in English, with the earliest retrospective study design.
being published in 1984. Six of the studies were
conducted in the USA, one in the UK and one in In the group of four studies examining open
Australia. The studies used either a prospective versus closed ICUs, three found that the
or a retrospective design. The studies varied practice of closing the unit significantly reduced
in duration, from 6 months to 3 years, and the mortality.253,359,360 However, one of the two studies
sample sizes ranged from 274 to more than 22,000
106
TABLE 33 Substitution
100 patients
Taiwan both nurses and NAs to carry skill mix model
12 months out nursing activities’ (p. 408)
An NA is ‘anyone who was
qualified to assist in nursing tasks
through obtaining a nursing care
training certification’ (p. 408)
Fontaine Experimental Two different models of after- No Cost; quality of nursing care Cost: No. of hospital days Positive effect: Employee
and Rositani 373 patients hours nursing care: nurse Patient/family satisfaction Quality of nursing care: nurses produce
(2000)339 employees vs contract nurses Own scale significantly better
8 months outcomes than contract
USA Patient/family satisfaction: nurses
Own questionnaire
Hall et al. Prospective ‘Nurse staff mix’ (p. 1099): No Patient functional health FIM; Medical Outcome Positive effect: Better
(2003)337 with concurrent All RN outcomes; pain; patient Study SF-36 (acute form) FIM, social function
continued
Health Technology Assessment 2010; Vol. 14: No. 51
107
108
TABLE 33 Substitution (continued)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
109
110
TABLE 33 Substitution (continued)
(2001)346 132 primary care gender; staff years in practice; part-time primary care
USA physicians staff location physicians
Unclear
Rudy et al. Experimental Residents vs acute care NPs and No Length of stay; mortality; adverse Length of stay: Routine data No effect
(1998)349 389 patients physician assistants drug reaction; completeness of collection
USA admission notes; re-admission Mortality: Routine data
14 months
collection
Adverse drug reaction:
Routine data collection
Completeness of admission
notes: Examination by
researchers
Re-admission: Routine data
collection
Hoffman et al. Prospective cohort ACNP and attending physician No Mortality rates No. of deaths of patients No effect
(2005)347 with concurrent vs critical care/pulmonary fellow who did or did not have
USA controls and attending physician in an ICU treatment limitations
526 patients (skill mix)
31 months
ACNP, acute care nurse practitioner; BSI, bloodstream infection; FTE, full-time equivalent; PCA, patient care associate; PCP, primary care physician; RPN, registered practical
nurses.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
that used a superior prospective design found no impact on physician performance and health-care
such positive effect.358 The impact of this practice outcomes.363–365 The two other reviews examine
on all other outcomes was also inconsistent. For systems of continuing professional development to
example, two studies examined length of stay and enhance the recruitment and retention of specific
found differing results.253,358 The different results health professionals.299,366 There are also three
were not related to quality of study design as both meta-analyses of training practices from the non-
studies employed a form of prospective design. health sector. These focus on elements of training
that are very different from those examined for
These eight studies covered all of the key patient health, such as altering organisational behaviour,367
outcomes, although the patient safety variables mentoring368 and evaluating the effectiveness of
were limited. organisational training.369 None of the studies
reviewed below appear in any of these systematic
Omissions reviews. A narrative synthesis has been undertaken
None of the studies reported here took the hospital due to the general heterogeneity of the studies in
or primary or community care practices as their this sample.
unit of analysis, and none controlled for any staff
variables or examined them as moderators or Details of studies
mediators of the relationships being evaluated. Sixteen studies met the inclusion criteria (Table
35). All studies were published in English, with the
Summary earliest being published in 1988. Eight studies were
This sample describes longitudinal research conducted in the USA, two in Argentina, and one
relating to the transfer or relocation of staff and each in the UK, the Netherlands, Sweden, Pakistan,
services. Despite the apparent homogeneity of Taiwan and Macedonia. Most of the studies
such practices or interventions, there is still a great employed a prospective longitudinal design, but
deal of heterogeneity between these studies in there were two RCTs256,370 and one experimental
terms of outcomes and units of analysis. Also, only study.371
one study232 measured any staff psychological or
behavioural outcome. The studies varied in duration, from 6 months
to more than 6 years. The length of the study
The results derived from similar interventions period was not given by one paper. The sample
(such as the move to closed from open ICUs) either sizes ranged from as few as 35 patients or 76 staff
had a positive effect, or no effect at all, in equal to 1411 patients or 649 staff, depending on the
measure. These findings, in combination with outcomes being studied. The relevant patient
the small number of studies, means that no firm sample size was not given in two studies. The unit
conclusions can be reached about the relationship of analysis was a hospital unit or ward in 13 of
between patient outcomes and the location of the studies. In one study the unit of analysis was
physicians on- or off-site, or the transfer of services a primary care practice and in two others it was a
to exclude certain types of physician. hospital.
113
The impact of HRM practices in the health sector
similarities, however, the type of training used was one study,258 but in another study, with a superior
different for each study and ranged from leaflets to design, the combination of improved knowledge
formal lectures and seminars. The majority of these and improved patient satisfaction was not found.256
studies concerned both nurses and physicians. The latter was a RCT.
Other than the group of training and feedback
studies, no study clearly measured the fidelity with Five studies found that training led to a significant
which the training programmes were implemented, reduction in mortality rates. This was the case not
although two studies did acknowledge barriers to only for studies using the prospective cohort and
full implementation. retrospective case–control research designs,254,259,379
but also for high-quality longitudinal designs, such
Outcomes as RCTs370 and quasi-experimental designs.371 Two
Almost one-half of the studies focused on a single studies examined length of stay as an outcome,
specific outcome: infection rates. Three of these with differing results. The more powerful quasi-
studies looked at ventilator-associated pneumonia, experimental study found that the training
two at bloodstream infections (BSIs), one at UTIs intervention had a positive effect on length of
and one at nosocomial infections generally. In each stay,371 whereas the more limited prospective
case, the measure was the number of infections per cohort design of Lawrence et al.379 found no such
1000 days of the relevant device. positive outcome. However, the interventions and
populations of the two studies were different.
Unlike many of the other interventions considered
by HRM studies in the health sector, several Omissions
training studies did examine staff outcomes, Several staff outcomes were measured by
usually knowledge or beliefs. In each case, this these studies, but no study controlled for any
outcome was measured in conjunction with at least staff variables and none examined these staff
one other outcome, usually patient satisfaction. psychological or behavioural variables as
In three studies, increased staff knowledge was moderators or mediators of the relationships
a specific and primary outcome of the training being evaluated. Unlike many other studies of
intervention. However, in each case the knowledge HRM practices in the health sector, there are
being assessed was different, so the scales used to no longitudinal studies here looking at whether
measure this knowledge were different. In the same training or education can have an impact on
way, three studies measured patient satisfaction waiting times or patient safety indicators other
or perceptions of care, and all three studies used than infection rates.
different measures.
Summary
Results The training interventions considered here have
With a single exception – Cohran et al.’s study372 certain basic elements in common, but they also
– the training interventions designed to address differ in elements of their content and delivery.
the specific outcome of infection rates all had The training ranged from leaflets to formal lectures
a significant positive effect on infection rates. and seminars. The primary outcomes examined by
Four of the five studies that combined training most of these studies were also different.
with some form of performance monitoring,
such as evaluation and feedback, had positive A group of seven studies all examined the impact
outcomes,373–376 and both studies examining of some form of infection control training on
education alone had positive outcomes.377,378 All of infection rates, and six of these studies found
the studies used a prospective cohort design with that training significantly reduced infection rates.
historical controls. However, none of these studies used a high-quality,
powerful form of longitudinal study design.
For the remaining studies, the results of those
training interventions that aimed to increase staff Several studies also found that training had a
knowledge were consistently positive.256,257,259 All positive effect on staff knowledge and mortality
four studies used a full range of study designs, rates. These studies not only used simple
from a RCT to a retrospective case–control study. prospective and retrospective designs, but also
Where staff knowledge was measured, significant superior longitudinal study designs, such as RCTs
improvements were found in two instances. and quasi-experimental design. The effect of
Improved knowledge was also reported in training from those studies is, overall, positive.
conjunction with greater patient satisfaction by
114
TABLE 35 Training
continued
Health Technology Assessment 2010; Vol. 14: No. 51
115
116
TABLE 35 Training (continued)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
117
118
TABLE 35 Training (continued)
(p. 250)
Harrison and Prospective cohort with CEP for nurses on knowledge No Nurses’ attitude and Miller–Dodder revision of Positive effect: Significant
Novak (1988)258 historical controls of and attitudes towards the knowledge; patient Palmore’s Facts on Aging improvement in nurses’
USA 76 nurses elderly satisfaction and Quiz knowledge and attitudes
perceptions of nursing Kogan’s Attitude Toward Old No effect: Patient
63 patients care; final outcome only: People Scale satisfaction
Unclear nurses’ knowledge and
attitudes LaMonica–Oberst Patient
Satisfaction Scale
Birnbaum et al. Retrospective Training in advanced cardiac life No Specific medical knowledge Range of specific behaviours/ Positive effect: Significant
(1994)259 869 patients (ischaemic support gains knowledge of staff reduction in mortality
USA heart disease) Mortality rates Mortality rates from rates
24 months Quality of care ischemic heart disease
Enhancement of Explicit measures of quality
knowledge and skills of care (e.g. number with
intravenous line in first hour
post training vs pre-training)
CEP, Continuing Education Program; CDCP, Centers for Disease Control and Prevention; CVC, central venous catheter; DDD, door-to-drug delivery; DSM IV, Diagnostic and Statistical
Manual of Mental Disorders-Fourth Edition; IPV, intimate partner’s violence; IVD, intravascular device; IVSP, Intravascular Surveillance and Education Programme; ORM, out-of-range
measurement; PMEE, pain management experience evaluation; PMIG, patient’s pain management interview guide; PN, practice nurse; PSSS, Patient Safety and Satisfaction Survey; SAS,
Staff Attitudinal Survey;
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
4 years
Armour et al. Retrospective Influence of bonuses on No Increased chance of Increased chance of Positive effect: Bonus eligibility
(2004)395 6749 patients Colorectal Cancer Screening referring patient for referring patient for significantly associated with
USA colorectal cancer colorectal cancer physicians asking for more
24 months screening screening: Routine data colorectal screening tests for
collection patients
Lee et al. Retrospective Primary care physician No Number and cost of Number and cost of No effect
(1999)394 582 patients payment: Salary vs FFS specialist diagnostic specialist diagnostic
Canada services and medicines services and medicines:
3 years Routine data collection
No. of days in hospital
Causes of death No. of days in hospital:
Routine data collection
Causes of death: Routine
data collection
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
123
The impact of HRM practices in the health sector
instruction, handouts or monthly presentations. outcomes have not been considered by the research
Each of these studies viewed this combination of in relation to this type of HRM intervention or
interventions as additive rather than synergistic, practice, such as waiting times, mortality and
and therefore not a bundle as understood by length of stay, as well as patient safety indicators
current HRM research.20,404,405 In only one study other than nosocomial infection rates. No staff
did the intervention have additional components, outcome was measured by any of these studies,
which included the posting of visual reminders and no study controlled for any staff variables, and
the use of penalties and incentives.375 However, none examined staff psychological or behavioural
the interventions were not all aimed at the same variables as moderators or mediators of the
population of health staff: physicians and nurses relationships being evaluated.
were the target of the intervention in three studies,
and either nurses or physicians alone were affected Summary
by the intervention in the two remaining studies. The interventions and outcomes considered by
Of the two studies that examined the practice of these studies were quite homogeneous and, even
feedback alone, one examined the outcome of though the number of outcomes researched in
infection rates, just like the education and feedback relation to this intervention was extremely limited,
studies, but the other examined the impact on this sample does suggest that the practice of
patient satisfaction of feedback to physicians. This feedback, both alone and in conjunction with
is the only study of this type. education or instruction, can have a significant
positive effect on infection rates, as well as patient
These studies are unusual because only one satisfaction.
does not measure compliance with the practices
described by the educational component of the
intervention. The majority, therefore, seek to Conclusions
measure the fidelity with which at least part of the
intervention or practice is being implemented. Longitudinal studies identified by the broad
This is because this monitoring or surveillance HRM and patient care outcomes literature search
is also part of the intervention: the feedback conducted for this review focus on a small number
component is informed by the measurement of of distinct HRM practices.
compliance. This is not a requirement for most
HRM interventions. These are:
continued
Health Technology Assessment 2010; Vol. 14: No. 51
125
126
TABLE 38 Performance management (continued)
patient to recommend
the physician to others
Kelleghan et al. Prospective cohort Feedback to physicians Yes – Reduction in VAP Reduction in VAP: Routine Positive effect
(1993)373 with historical Education programme monitoring data collection of cases per
USA controls and 1000 ventilator-days
Unknown surveillance of
compliance
4 years
Nettleman et Prospective cohort Feedback to physicians Yes – MRSA rates MRSA rates per 1000 Positive effect
al. (1991)407 with historical Handouts monitoring of patient-days
USA controls compliance
Monitoring
Unknown
Monthly presentations
27 months
CVC, central venous catheter; IVD, intravascular device; VAP, ventilator-associated pneumonia.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
(for example by replacing one type of worker with interest among health sector researchers in a small
another who costs less).321 Alternatively, this focus number of practices relating to staff skill mix
may be because the increasing numbers of patients and working hours. In some cases, these small
using UK health services have an impact on staff groupings also measure the same outcomes (albeit
workload, which may, in turn, affect the quality of rarely using the same or equivalent measures).
patient care.320 However, with the exception of practices relating
to skill enhancement, or the skill mix variables
This differs markedly from the non-health sector involving the introduction or substitution of staff,
HRM review and meta-analysis literature, which the number of studies contained in these groupings
has almost no meta-analyses at all on the skill mix is often small, rarely larger than four or five. There
of teams, staffing ratios or working hours. Instead, are also often differences between these studies in
it focuses principally on practices that are covered terms of location, unit of analysis and study length.
much less by literature on the health sector, such as Such heterogeneity precludes the pooling of data
employee involvement391,401,409 and recruitment and for quantitative synthesis and means we must add
selection.281,410,411 This review found only a small a strong caveat to the drawing of conclusions from
number of longitudinal studies looking at quality any narrative synthesis of these studies.
circles (employee involvement) in relation to
patient outcomes. HRM practices or interventions Interestingly, almost all of the studies identified
relating to communication, family friendly working in this review report neither negative nor adverse
and harmonisation are other major categories effects for any of the practices evaluated, which
for which no longitudinal studies from the health may suggest a degree of publication bias affecting
sector were identified by this review. the results. The interventions that appear to report
the most consistent positive impact on patient
There is a great deal of heterogeneity among and staff outcomes are training, performance
the studies reviewed in all of the categories. For management (feedback) and employee
example, the studies placed within the categories involvement (quality circles). Beneficial effects were
of pay, employee involvement and training are found on infection rates, mortality, length of stay,
all very different in terms of both practices and patient satisfaction and staff knowledge. The many
outcomes, and this prevents anything but the work design variables, such as the substitution or
most basic narrative synthesis. There is a greater introduction of staff, skill enhancement, reduced
degree of homogeneity among the studies under working hours and the relocation of staff or
the categories of work design, performance services, produced more inconsistent results. In
management and staffing. Within these categories, some studies these interventions had no effect (or
some of the studies examine the same or very no adverse effect), and only in a few cases were
similar practices and may be grouped together. the effects positive, such as for the addition of new
Once again, this probably reflects the greater nurses or physicians.
127
Chapter 8
Impact of HRM practices in non-health settings
Introduction Search strategy
Literature search strategies were developed
This chapter reports on findings from a systematic and performed on Business Source Premier
review of the impact of HRM practices on and PsycINFO; these two electronic databases
intermediate and final outcomes in non-health have the most comprehensive coverage of peer-
sector settings. The findings address research reviewed journals in the field of psychology and
questions under Objectives 5 and 6 of the review, management. Search strings were modified for the
specifically: different databases to take account of the different
keywords and thesauri they use. Examples of
• Objective 5: 5.1 What is the evidence for the the search filters and strategies can be found in
impact of HRM practices on the intermediate Appendix 8.
outcomes identified for this review?
• Objective 6: 6.1 Which HRM practices have an The searching of electronic databases using pre-
impact on performance outcomes? designed search filters was supplemented by
reference tracking of selected studies. The titles,
This chapter gives details of the aspects of the abstracts and full papers of studies identified by
review methodology specific to this review. both electronic database searching and reference
Narrative syntheses for each of the HRM categories tracking were sifted using the following inclusion
and data tables are then presented. criteria.
Study selection
Review methodology
Sift criteria were developed by the team and
The broad approach to systematic review piloted on a sample of titles and abstracts.
methodology adopted to identify the evidence Adjustments were made to improve specificity, and
presented in Chapters 6–8 is described in Chapter all titles and abstracts identified by the electronic
2. This section describes aspects of the review searches and reference tracking were sifted by one
methodology specific to this search. team member, with 50% being sifted by a second
reviewer. The inter-rater reliability, using k-statistic,
Protocol development for the double-sifted articles was 0.92.
Simple protocols were developed to guide the Unclear articles went to another team member for
search, although some parameters were left open a second opinion. If still unclear, the full paper was
due to the lack of consistent definition/use of retrieved for sifting.
HRM terms in the literature. The population for
the searches was identified as working adults and The criteria identified papers as relevant to the
the intervention was determined as any HRM or review on the basis that they:
personnel management practice. No comparison
was specified, but a longitudinal filter was • were longitudinal in design
developed and used in the search. This ensured • included an intervention which could be
that the search was focused on causal relationships. described as some form of HRM practice, and
No specific outcomes were stipulated. • evaluated the intervention on some form of
outcome measure.
The data extraction and synthesis was left open
at this stage to allow development of a synthesis Studies were excluded if they were unpublished
framework most fitting to the data retrieved. This theses.
approach is consistent with the latest developments
in applying systematic review methodology to This not only acknowledged potential difficulties
complex literatures.52,56,58,60,63 in obtaining these materials, but also, more
129
130
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
The remainder of the chapter presents the is conceptually similar to the job characteristics
narrative synthesis of findings for each of the main approach to job design proposed by Hackman and
HRM categories (or subclusters). Oldham.412 The difference in the two approaches
is the level at which the concepts of variety,
Work design identity, significance, autonomy and feedback are
applied. The job characteristic model proposes the
A total of 49 studies were identified under the application of these principles at the individual
broad HRM category of work design. The studies work level; the autonomous teamwork model tries
fell into distinct, heterogeneous clusters and to implement the same variables at a group or team
syntheses under work design are presented as level in a bid to influence work-related outcomes.
follows:
Details of studies
• teamwork Eight studies met the inclusion criteria (Table 40).
• job complexity All the studies were published in English, with
• job demands and control three conducted in UK, two each in Australia and
• job rotation, enlargement and enrichment Canada, and one in USA. There were five studies
• role conflict/clarity. with quasi-experimental methods, and three
studies with a prospective longitudinal research
Teamwork design. The shortest time duration for a study
Studies on teamwork largely look at how was just 2 months, whereas the longest study was
autonomous or semi-autonomous teams can carried out over a period of 8 years. The sample
have a significant impact on certain work-related size of these eight studies varied from 34 to 545
outcomes. Autonomous teamwork approaches employees.
propose that increases in work factors (e.g. variety,
identity, significance, autonomy and feedback) are HRM practice/intervention
causally linked to work-related outcomes, such as Four research papers investigated the impact of
motivation, job satisfaction and productivity. It semi-autonomous or autonomous work groups.
131
2 months Employee turnover; Productivity, accidents and absenteeism: practices on all the outcomes
None Company data
absenteeism
Banker et al. Quasi- Practices: (as a result of relocation) No Quality; labour Quality: Percentage of total units Positive effect:
(1996)417 experimental High-performance work teams productivity produced that were defective labour Both, quality and labour
USA 41 employees (characterised by decision- Productivity: Ratio of the number of units productivity improved
21 months making authority, mandated team produced to total production hours
membership, information rich and
expanded problem-solving domain)
Moderators/mediators:
Managerial policies:
Workforce policies (overtime,
headcount additions, headcount
deletions)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
133
134
TABLE 40 Work design (teamwork) (continued)
GHQ, General Health Questionnaire; OCQ, Organisational Commitment Questionnaire; JDS, Job Diagnostic Survey.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Other practices investigated in this section were the in turn, led to poor organisational commitment
impact of work group fit, high-performance work and high job turnover. A similar finding was
teams, work group stability, and team working, as reported for the link between organisational
opposed to individual work arrangements. commitment and work teams, with employees
showing high commitment if they worked in teams
Outcomes with similar demographics.419
A range of outcome variables was measured in
these studies. The most frequently studied type of Mixed results were found for the relationship
outcome variable in relation to teamwork was some between team working and absenteeism. One
form of performance measure (e.g. individual/team study found work teams to exert a positive impact
performance, quality or productivity), with five of on absenteeism,416 but the study by Cordery et
the eight studies investigating at least one of these. al.418 found a negative impact of teamwork on
absenteeism.
All other outcome variables in the studies included
here were employee centric, such as motivation, job Omissions
satisfaction, organisational commitment, employee The longitudinal studies on work teams largely
turnover, absenteeism, morale, change in employee concentrated on performance outcomes or
attitude and mental health. employee-centric outcomes. There were virtually
no longitudinal studies that investigated how, or
Job satisfaction was the most studied employee- if, team working can lead to learning outcomes for
centric outcome, with five of the eight studies employees. There are also no longitudinal studies
measuring how it changed with the implementation included here that explore the impact of working
of one form of teamwork or another. Job turnover in teams on the financial bottom line of a firm or
was the next most commonly studied outcome, with business unit.
four studies investigating it as an outcome of group
work. Other variables were studied by one or two Summary
papers. There were eight studies on teamwork and its
impact on various work-related outcomes, with
Results one-half of the papers concentrating on semi-
There was some support for improved team and autonomous or autonomous teams. Most of these
individual performance under autonomous or studies have explored the impact of autonomous
semi-autonomous teamwork, with two out of three team working on performance, quality or
studies arriving at this conclusion.413,414 One study productivity issues. Almost all the studies found
failed to find any such significant relationship.415 these two variables to be positively linked.
Both quality and labour productivity were also
found to have improved as a result of introducing The relationships of work teams with job turnover
high-performance work teams, characterised and absenteeism were not so straightforward. Some
by decision-making authority, mandated team studies found teamwork to have a positive impact
membership, and an expanded, information rich, on both factors, whereas others found them to
problem-solving domain.416,417 have a negative impact. There was some evidence
of this relationship being moderated by group and
There was very strong support for the positive demographic fit.
impact of work teams on job satisfaction, with all
five studies coming to this conclusion.413–415,418 The studies in this area mostly focused on
performance and employee-centric outcomes and
Five studies investigated the impact of work teams there were no longitudinal studies investigating the
on job turnover and the results were equivocal, impact of work groups on learning and growth of
with three studies finding support for the positive employees.
impact of team working on job turnover,413,416,419
whereas two other studies found team working to Job/task complexity
have a negative impact on employee turnover.415,418 Studies of job/task complexity focused on two
Kirchmeyer419 found that a better work group aspects of complexity:
fit had a positive impact on organisational
commitment and led to reduced turnover, whereas • Psychological complexity refers to the
dissimilarity in terms of age, education and lifestyle psychological dimensions of tasks, for example
led to poor integration with the work group, which, variety and perceived job scope. It is a more
135
subjective measure of job complexity and routine or simple tasks to complex tasks – on job
has been measured using tools such as the satisfaction. Other than task complexity, this study
JDS developed by Hackman and Oldham.195 also included wages and performance appraisal
Such an approach to job/task complexity has satisfaction as predictors for the given outcome.
often been equated with the concept of job
enrichment.427 Outcomes
• Task–person job complexity refers to tasks All three studies used different employee-centric
that can be objectively classified as simple or outcomes, for example employee burnout,
complex, based on the mental demands that a employee cardiovascular health, and employee job
given job places on the person and the extent satisfaction, as dependent variables.
of education, skill or training required on the
part of the person to do that job.428 Results
None of the three studies investigated the impact
Details of studies of job/task complexity directly. The studies all
Three studies met the inclusion criteria (Table 41). examined the interaction of job/task complexity
All of the studies were published in English, with with other variables, or study job/task complexity as
two being conducted in the USA and the one in a moderator in the analysis.
Finland. The first study was quasi-experimental,
with retrospective analysis of the impact of The studies by Kalimo et al.429 and Schaubroeck et
certain HRM practices on the given outcome. al.428 concluded that the impact of job complexity
The remaining two studies used a prospective was moderated by, or interacted with, the personal
longitudinal design. The quasi-experimental study resources and personality type of an individual to
was conducted over 10 years, whereas the other determine how it would impact on an individual’s
two studies were conducted over 7 and 4 years, health. The study by Kalimo et al.429 found that
respectively. The sample sizes of the three studies higher job complexity was associated with lower
were 174, 110 and 672 employees, respectively. burnout. The other variables that helped prevent
burnout were appreciation of work, feedback
HRM practice/intervention and role clarity. They concluded that sense of
The three studies in this synthesis had distinct coherence (SOC) was the most important personal
concepts of job/task complexity. However, the determinant of burnout and individuals with high
studies can be broadly classified as using either SOC were significantly more resistant to burnout
psychological or task–person complexity as than those with low SOC. Schaubroeck et al.428
a variable in their investigations. One study found that job complexity (both types) interacted
examined only task–person complexity, one study with employee personality type (types A and B) to
used only psychological complexity, and one study impact on cardiovascular health. They concluded
looked at both task–person and psychological that high job complexity for type B individuals
complexity. leads to low cardiovascular disorders, whereas high
job complexity for type A individuals leads to high
In the first study, by Kalimo et al.429 psychological cardiovascular disorders.
task complexity was studied as one of the many
variables, measured using the Occupational The study by Blau236 used task responsibility as a
Stress Questionnaire (OSQ).430 This study used construct and measured it as a continuum ranging
personal resources (sense of coherence, self-esteem from routine to complex tasks. He concluded that
and sense of competence) as moderators of the task complexity was significantly positively related
impact of task complexity and other predictors on to job satisfaction. Besides task complexity, he
employee burnout. found both higher wages and satisfaction with
supervisory performance appraisal to be significant
The second study, by Schaubroeck et al.,428 determinants of employee job satisfaction.
investigated the impact of both psychological task
and task–person complexity, as moderated by the Omissions
impact of an individual’s personality (type A or All studies have focused on employee-centric
type B) on his/her cardiovascular health. The third health-related outcomes, for example burnout
study in this analysis, Blau,236 investigated the and cardiovascular health, or on psychological
impact of task complexity – varying on a scale from outcomes, such as job satisfaction.
136
TABLE 41 Work design (job complexity)
Design,
Study, sample size, Practice(s)/intervention(s);
country duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Kalimo et al. Quasi- The following work characteristics were No Employee MBI–GS; Schaufeli et al. Positive effect:
(2003)429 experimental studied using the OSQa) burnout 1996)188
DOI: 10.3310/hta14510
DOT, Dictionary of Occupational Titles; MBI, Maslach Burnout Inventory – General Scale.
a Elo A-L, Leppanen A, Lindstrom K, Ropponen T. OSQ. Occupational Stress Questionnaire: User’s Instructions. Helsinki: Institute of Occupational Health; 1992.
Health Technology Assessment 2010; Vol. 14: No. 51
137
Impact of HRM practices in non-health settings
Summary Outcomes
These studies used the concept of job/task All 13 studies have used employee-centric
complexity in two ways: psychological complexity outcomes (e.g. job satisfaction, absenteeism,
of jobs, i.e. jobs higher or lower on the five turnover and work motivation) as dependent
dimensions of job scope, as defined by Hackman variables. The majority of studies also measured
and Oldham,431 or jobs higher or lower on employee health-related outcomes, including
task–person complexity, as determined by some mental health, occupational stress, psychological
objective criteria of a job being simple or complex. well-being, self-rated health, perceived strain,
Overall, the studies seem to suggest that either: emotional exhaustion, anxiety, depression, immune
functions and cardiovascular health.
• Task complexity interacts with other variables,
for example, individual personality factors, to Job satisfaction was the most frequently studied
have impact on given outcomes; or, outcome, with five studies investigating how
• The effects of task complexity are moderated demands and control have an impact on employee
by variables, such as an individual’s personal job or work satisfaction.
resources including SOC, self-esteem and sense
of competence. Two studies investigated the impact of demands
and control on employee learning outcomes, such
Job demands and control as skill utilisation, self-efficacy and motivation to
The studies in this section investigated how learn.
demands and control at work can have important
consequences for various work-related outcomes. Results
Most of these studies have tested Karasek’s The studies on job demands and control have used
demand–control model433 in one form or another. Karasek’s433 demand–control model as a starting
Studies examined various aspects of job demands: point to investigate how various combinations of
physical, psychological, workload, monitoring and demand and control impact on given outcomes.
complexity. Control has been studied as decision
control, skill discretion, degree of autonomy and Of the five studies that investigated the impacts
predictability. of demands and control on job satisfaction, two
concluded that high demands and low control lead
Details of studies to poor job satisfaction,233,434 two studies found that
Thirteen studies met the inclusion criteria (Table high control led to higher job satisfaction,435,436 and
42). All of the studies were published in English. one study failed to find any impact for high control
Five were conducted in the Netherlands, four in on job satisfaction.229
the UK, two in Denmark and one each in Sweden
and Hong Kong. Five of the 13 studies had a Absence was another outcome that was found to
quasi-experimental design and the remaining respond well to high job control. Three studies
eight studies were of prospective longitudinal concluded that high job control leads to lower
design. The majority of studies (8 out of 13) were absence.229,437,438 The study by Smulders and
conducted over a 12-month period or less. The Nijhuis437 concluded that high job demands, when
shortest time duration for a study was just 1 month, coupled with high control, can lead to low absence.
whereas the longest was carried out over 5 years.
Sample size varied from 144 to 11,799 employees. Job turnover was found to be negatively associated
with high demands, but high control was not found
HRM practice/intervention to have any significant relationship with employee
All of the studies in this section investigated the turnover.229,439
impact of work demands and individual levels
of control. The majority of studies (8 out of 13) All of the health and psychophysiological outcomes
measured some form of social support at work were found to be negatively related to high job
as a potential moderator of the demand–control demands with low control, and to respond well to
relationship. Locus of control (LOC)/personality enhanced job control.
was used as moderator by two studies, and one
study each used work values (traditional versus Two studies that explored the relationship of
modern) and extent of predictability (being demands and control on employee learning and
informed of future events at work) as moderating self-efficacy concluded that increased control leads
variables.
138
TABLE 42 Work design (job demands and control)
Johnson et al. Quasi-experimental Practices: No Cardiovascular Mortality data were Positive effect:
(1996)442 11,799 employees Work control (a 12-item scale disease mortality or obtained from National Higher psychological demands were found
Sweden measuring decision authority and risk Death Registry to be associated with lower cardiovascular
4 years
skill discretion on the job) disease morality or risk of it
Psychological demands (time Negative effect:
pressure and extent to which the Men exposed to lower control jobs had a
job was mentally strenuous) substantially elevated risk for cardiovascular
Physical job demands; job hazards; disease mortality relative to men with a work
work social support history of high-control jobs
Moderators/mediators: None
Moyle and Parkes Quasi-experimental Practices: (as a result of relocation) No Strain indicators: Job satisfaction: 10-item Positive effect:
(1999)436 175 employees Demand and control (an 1. Job satisfaction measure Perceived control and high managerial
UK 7 months adaptation from Karasek (1979),433 2. Psychological Psychological distress: 12- support buffered the negative impact of work
continued
Health Technology Assessment 2010; Vol. 14: No. 51
139
140
TABLE 42 Work design (job demands and control) (continued)
2 years
authority; predictability of work; events at work) predicted lower absence rate
skills; support from colleagues;
supervisory support; meaning of
work)
Moderators/mediators:
None
Daniels and Guppy Prospective Practices: No Occupational stress; Occupational stress was Positive effect:
(1994)447 longitudinal Job control (job autonomy and psychological well- measured using a 18-item High job control and high social support had
UK 244 employees participative decision-making) being questionnaire, designed a positive impact on psychological well-being
for the study by the
1 month Social support authors Interaction effect:
Moderators/mediators: Psychological well-being LOC moderated the impact of job control
LOC (Warr 1990;445 and and social support on psychological well-
GHQ12 by Goldberg and being in such a way that persons with internal
Williams 1988 448) LOC benefited more from high job control
than persons with external LOC
Moyle (1998)435 Prospective Practices: No Mental health and job Mental health: GHQ Positive effect:
UK longitudinal Demand and control satisfaction (used as (Goldberg 1978)239 High control had a direct positive impact on
148 employees indices of strain) Job satisfaction: 10- job satisfaction
Managerial support
12 months item scale adapted from Managerial support has a positive impact on
Role ambiguity Parkes (1993)449 job satisfaction and mental health
Moderators/mediators:
Negative effect:
None
High demand and low control had a negative
impact on both concurrent and future mental
health
Design, sample Practice(s)/intervention(s);
Study, country size, duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Smulders and Prospective Practices: No Absence Two measures of absence Positive effect:
Nijhuis (1999)437 longitudinal Job demands were used: Job control was significantly associated with
Netherlands 1755 employees Job control 1. Absence rate: Days low absence rate
11 months absent/100 calendar Job demands was also associated with low
Physical working conditions days
DOI: 10.3310/hta14510
absence rate
Moderators/mediators: 2. Absence frequency: No effect:
None No. of spells of
absence/per year per Neither job control nor job demand was
person (irrespective of associated with absence frequency
their duration)
Jia et al. (2000)450 Prospective Practices: No Immune function; Blood test Interaction effect:
Hong Kong longitudinal Worker control: Includes job symptoms of The relationship between decision control
492 employees control and decision control; respiratory illness and poor health was stronger for more
12 months Measures: Smith et al. (1997)451 Job traditional employees than for less traditional
complexity: House (1980);432 Job employees. Also, job control interacted with
demands: Includes job complexity, job demands and job self-efficacy to predict
responsibility for others, workload health outcomes
and monitory pressure
continued
141
142
TABLE 42 Work design (job demands and control) (continued)
de Croon et al. Prospective Practices: No Voluntary turnover Voluntary turnover data Negative effect:
(2004)439 longitudinal taken from company
Impact of HRM practices in non-health settings
Job demands and control (with Jobs high on demands and low on controls
Netherlands 564 employees jobs high on demands and low on records (defined in the study as stressful work) had a
2 years controls being defined as ‘stressful negative impact on voluntary turnover
work’) Mediation analysis revealed that this
Moderators/mediators: relationship between stressful work and
Strain (defined in this study as turnover was mediated through psychological
‘need for recovery after work’ and fatigue
‘fatigue’)
de Langeff et al. Prospective Practices: No Mental health: Mental health Negative effect:
(2004)434 longitudinal Job demands and Job control Defined in the Depression:11-item High level of demand with low job control
Netherlands 668 employees current study as version of the CES-D and poor social support led to poor mental
Social support from supervisors comprising three
4 years scale health (higher depression, low job satisfaction
Moderators/mediators: variables: and high emotional exhaustion)
Job satisfaction: New
None 1. Depression single-item measure; Also, there was a reciprocal relationship
2. Job satisfaction Burnout: Seven-item between job demands, social support and
3. Emotional dichotomous subscale of emotional exhaustion
exhaustion the MBI
CES-D, Center for Epidemiologic Studies Depression Scale; LOC, locus of control; MBI, Maslach Burnout Inventory.
a Schaufeli WB, Van Dierendonck D. The construct validity of two burnout measures. Journal of Organizational Behavior 1993;14:631–47.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
to better skill utilisation and also has a positive engaged in.454 It refers to horizontal expansion
impact on active learning.440,441 of the jobs, increasing the role responsibilities
and range of activities being carried out by an
Later adaptations of Karasek’s433 model of employee.453
demands and control include support as a third • Job enrichment leads to more holistic changes
strand. This is reflected in the use of managerial in job roles and responsibilities. In contrast
support or social support measures in eight of the with the horizontal expansion involved in
13 studies. There was clear evidence in this set of job enlargement, job enrichment involves
studies for the positive impact of social support on vertical expansion.412 An enriched work design
all given outcomes when used in conditions of high means that employees have discretion to take
job control. on broader and more proactive tasks as and
when required. It can be done by increasing
Omissions employee responsibility, autonomy and regular
These 13 studies have covered a wide variety of feedback.
health and employee-centric outcomes. However,
there were no studies exploring how job demands Details of studies
and job control, in their various combinations, Twenty-one studies met the inclusion criteria
influence job performance, productivity or (Table 43). All the studies were published in English,
customer/client outcomes. with eleven being conducted in USA, three in the
UK, three in Canada, two in the Netherlands, and
Summary one each in Sweden and Israel. There were four
The results of the 13 studies meeting the criteria studies with RCT designs, six studies used quasi-
for inclusion in this review provided consistent experimental methods, and eleven studies had a
evidence for the positive impact of increased job prospective longitudinal research design. There
control and the negative impact of high demands was huge variation in the duration of the studies in
on the given outcomes. Job satisfaction, absence this section, with the shortest being just 2 months
behaviour and health outcomes in particular were and the longest being 4.5 years. The sample size
positively influenced by higher control. A large of these 21 studies varied from 31 employees to
proportion of these studies (5 out of 13) were 80 manufacturing firms with between 60 and 1150
quasi-experimental in design. Over one-half of the employees.
studies investigated the influence of managerial or
social support as a variable alongside demands and HRM practice/intervention
control. Authors used the terms job enlargement and
job enrichment in their own distinct ways.
The studies in this area mostly focused on health Campion and McClelland455 used the terms ‘task
and employee-centric outcomes and there were enlargement’ and ‘knowledge enlargement’ as
no longitudinal studies investigating the impact two types of job enlargement, with knowledge
of demand and control on performance or enlargement being conceptually similar to
productivity. job enrichment. Therefore, the Campion and
McClelland455 study has been included as a study
Job rotation, enlargement and of both job enlargement and enrichment. Parker454
enrichment used job enlargement as horizontal enlargement
The concepts of job rotation, enlargement and and job enrichment for jobs that were high on
enrichment are all suggested answers to the autonomy and control. Other variations of the
problems posed by job simplification. use of the term job enrichment included high job
stimulation, high motivating potential score (MPS),
• Job rotation involves doing more than a single high authority and control, high job quality, high
type of repetitive task. The inherent logic of job scope, and high intrinsic motivating properties
job rotation is to help break the monotony of the job.
arising out of doing the same task repeatedly,
also making the employee more versatile as There were two studies on job rotation:
he/she learns various other tasks. It does not, one analysed the impact of job rotation on
however, change the content of the job.453 psychophysiological stress reactions and
• Job enlargement is intended to increase musculoskeletal symptoms456 and one explored the
the breadth of activities that employees are impact on performance.457
143
McClelland 515 employees enlargement of performance); job had positive impact on satisfaction, mental
(1993)455 efficiency workload, errors and led to better customer
2 years Moderators/mediators:
USA Satisfaction; mental service
None
load Negative effect:
Task enlargement led to more mental overload,
greater chance of making errors, and lower job
efficiency
Parker Prospective Practices: No RBSE RBSE scale was designed by Positive effect:
(1998)454 longitudinal Job enlargement (measured as the the author Job enrichment was positively associated with
UK 459 employees horizontal range of the jobs) RBSE
18 months Job enrichment (measured as jobs
high on autonomy and control)
Communication; training (on
quality management)
Moderators/mediators:
None
Axtell and Prospective Practices: No RBSE RBSE scale was designed by Positive effect:
Parker longitudinal Job enlargement (measured as the the author Job enrichment was positively associated with
(2003)458 94 employees horizontal range of the jobs) RBSE
UK 18 months Job enrichment (measured as jobs Negative effect:
high on autonomy and control) Job enlargement had a lagged negative impact on
Communication; training (on RBSE
quality management)
Moderators/mediators:
None
Study, Design, sample Practice(s)/intervention(s);
country size, duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Locke et al. Experimental Practices: No Productivity; Outcomes for example Positive effect:
(1976)466 design Job enrichment absenteeism; absenteeism, turnover, Absenteeism, turnover and productivity were
USA 95 employees turnover; complaints complains and disciplinary positively influenced by job enrichment
Moderators/mediators: and disciplinary actions and productivity
14 months None actions; attitudes were taken from company No effect:
records. Attitudes were No changes were observed in attitudes of the
DOI: 10.3310/hta14510
continued
145
146
TABLE 43 Work design (job rotation, enlargement and enrichment) (continued)
Job enrichment (through Task perceptions Time 3 and 4 measures of performance were
USA 564 employees enhancement of responsibility, (MPS score); Minnesota Satisfaction significantly (positive) related with the time 1 and
4 years authority and accountability; job satisfaction; Questionnaire (Weiss et al. 2 MPS of the job
increased feedback) organisational 1967)473 There was a significant increase in MPS,
Moderators/mediators: commitment; Organisational commitment satisfaction, commitment, and performance over
absenteeism; (Porter et al. 1974)244 time
None propensity to quit
Propensity to quit (Seashore No effect:
et al. 1982)474 There was no significant change in absenteeism
Performance was assessed and propensity to quit as a result of increased
through supervisor scores job enrichment
Absenteeism was taken from
company records
Morgeson Quasi- Practices: No Overall job Positive effect:
and experimental satisfaction; training
continued
Health Technology Assessment 2010; Vol. 14: No. 51
147
148
TABLE 43 Work design (job rotation, enlargement and enrichment) (continued)
Moderators/mediators:
None
Meyer et al. Prospective Practices: No Organisational Organisational commitment: Positive effect:
(1991)468 longitudinal Job quality: Comprising a commitment Meyer and Allen (1984)477 Job quality had a significant effect on affective
Canada 104 employees composite score from the (affective and commitment
following variables: Job challenge, continuance
11 months commitment) No effect:
participation, self-expression,
personal importance, role Job quality had no effect on continuance
clarity, feedback, confirmation of commitment
expectations
Moderators/mediators:
None
Barnett et al. Prospective Practices: No Psychological distress Used items from 2 scales Negative effect:
(1995)470 longitudinal Job role quality on psychological distress: A decrease in job quality over time was related
USA 210 employees Derogatis, (1975),478 Barnett with an increase in psychological distress but this
Moderators/mediators: et al. (1993)479
2 years relationship was not affected by the gender
Gender
Krausz et al. Prospective Practices: No Intention to leave Self-made scale Positive effect:
(1995)252 longitudinal Job scope [measured using a -tem High job scope led to low intention to leave
Israel 146 employees scale based on Hackman and
12 months Oldham’s (1975)195 JDS scale]
Moderators/mediators:
None
Study, Design, sample Practice(s)/intervention(s);
country size, duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Taris et al. Prospective Practices: No Depression Translation of Depression Negative effect:
(1998)471 longitudinal Intrinsic job characteristics Adjective Checklist (Rooijen Intrinsic job characteristics and depression had
DOI: 10.3310/hta14510
Netherlands 593 employees (comprising jobs high on variety, 1979480 and Lubin 1965)481 a significant reciprocal negative effect on each
4 years autonomy, opportunity to engage other
in meaningful tasks, and feedback) Job change did moderate the effect between
Moderators/mediators: time 1 depression and time 2 job characteristics
Job change
Blocher Prospective Practices: No Performance Reviewer’s overall evaluation Negative effect:
(1979)457 longitudinal Job rotation of workers’ performance, A lapse in performance in consecutive
USA 58 employees rated on an ordinal scale assignments was observed for employees within
Moderators/mediators: from ‘superior’ to ‘needs
5 years the same industry
None improvement’
No effect:
No change in performance was observed for
employees who rotated between assignments in
different industries
DBP, diastolic blood pressure; GNS, growth need strength; MBI, Maslach Burnout Inventory; SBP, systolic blood pressure.
Health Technology Assessment 2010; Vol. 14: No. 51
149
Impact of HRM practices in non-health settings
Three studies investigated the combined effect level if the seniors were assigned consecutively
of job enlargement and job enrichment.454,455,458 within a particular industry, whereas no lapse
Sixteen studies investigated the sole impact of job in performance was observed for those seniors
enrichment on various outcomes. who rotated between assignments in different
industries.
Outcomes
All of the studies on job rotation, enlargement and Studies investigating job enlargement and
enrichment studied outcomes that can be broadly enrichment found a positive impact on RBSE454,458
classified in three categories: outcomes related to but a negative impact on workload, i.e. workload
customer service/satisfaction, outcomes related to increased.455
both individual and organisational performance/
productivity, and employee-centric outcomes, Knowledge enlargement was found to have a
including absenteeism, turnover, motivation, positive relationship with customer services.455
involvement, organisational commitment, job
satisfaction, effort, depression, role breadth self- Researchers investigating the impact of job
efficacy (RBSE) and other employee attitudes. enrichment on various outcomes were unanimous
in their conclusion that it is a desirable
• The studies on job rotation had physiological, intervention, which works quite well for some
psychophysiological and performance outcomes but has less straightforward results
outcomes as their predicted variables.456,457 for others. Of the nine studies that investigated
• There was only one study that investigated job satisfaction, eight found it to be positively
the impact of job enlargement (knowledge influenced by job enrichment235,455,459,460,462–465
enlargement) on customer service.455 and only one study had a no-effect result.461 Job
• The studies on job enrichment looked at a enrichment was also found to enhance employee
relatively large number of outcome variables. motivation and involvement by all the studies that
Job satisfaction was the most frequently investigated these links.460,462,467
studied, with nine studies investigating it as a
predicted outcome.235,450,459–465 Lack of enriched job has also been found to lead
• Employee turnover or intention to leave was to increased levels of psychological distress and
investigated by four studies.252,460,464,466 depression.470,471
• Individual performance, productivity,
motivation, absenteeism and commitment were Two studies found employee performance to be
studied by five studies each.34,235,455,460–464,466–469 positively influenced by job enrichment,455,464
One study34 examined the impact of job whereas two other studies failed to reach any
enrichment on organisational performance such conclusion.460,461 With regards employee
(company productivity and profitability). productivity, two research papers concluded that
• Two studies studied the impact of job it is positively related with job enrichment463,466
enlargement and job enrichment on whereas Orpen460 found productivity to be
RBSE.454,458 unaffected by job enrichment. The study by
• Two studies explored the impact of job Patterson et al.34 concluded that the extent of
enrichment on job involvement and positive empowerment within companies predicted the
employee attitudes.461,462 subsequent level of company performance (both
• Effort, psychological distress and depression productivity and profitability) controlling for prior
were investigated by one study each.461,470,471 performance, the association between enrichment
and profitability was mediated by productivity,
Results suggesting that enrichment affected profitability
Job rotation was found to have a clearly beneficial through its effects on productivity.
impact on certain psychophysiological stress
reactions, for example diastolic blood pressure and More definitive results were obtained for the
self-reported positive experience of the job. It also relationship between job enrichment and
had a positive impact on muscle activities in the turnover, with three studies favouring a positive
neck and shoulder. However, job rotation had no impact252,460,466 and one study failing to come to any
impact on other psychophysiological reactions of such conclusion.461
systolic blood pressure, heart rate and perceived
negative arousal.456 Blocher457 found a lapse in Of the three studies on absenteeism and job
performance over time for audit staff at senior enrichment, two found a positive impact of
150
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
151
(1993)234 longitudinal Role ambiguity; role conflict; role from 18 item index developed by After controlling for effects
USA 405 employees overload; autonomy; supervisory Brayfield and Rothe (1951)245 of other variables, role
3 months support; task significance; ambiguity, role conflict and
routinisation; pay role overload had no effect on
Moderators/mediators: job satisfaction
None
Moyle Prospective Practices: No Mental health and job Mental health: GHQ (Goldberg, Negative effect:
(1998)435 longitudinal Role ambiguity; demand and satisfaction 1978)239 Role ambiguity had a direct
UK 148 employees control; managerial support Job satisfaction: 10-item scale negative impact on job
12 months Moderators/mediators: adapted from Parkes (1993)449 satisfaction
None No effect:
Role ambiguity had no effect
on mental health
Saks and Prospective Practices: No Individual job performance Job satisfaction: Cammann et al. Positive effect:
Ashforth longitudinal Stressors: Role conflict, role details Adjustment to work: (1983)485 Stressors significantly
(2000)482 231 employees ambiguity and role overload Measured by the following Organisational commitment: Allen predicted all outcomes
USA 10 months Moderators/mediators: components: and Meyer (1990)486 Interaction effect:
Predispositions (GSE) Job satisfaction Organisational identification: Mael Role conflict interacted with
Organisational commitment and Ashforth (1992)487 GSE in a manner that person
Organisational Intentions to quit: Colarelli (1984)488 with high GSE had higher
identification Frustration: Ashforth (1989)489 organisational commitment and
organisational identification
Intentions to quit Stress symptoms: Patchen (1970)490 under conditions of high role
Frustration Performance: House et al. (1982)491 conflict, whereas the opposite
Stress symptoms was true for persons with low
GSE
Performance
Study, Design, sample Practice(s)/intervention(s);
country size, duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Bravo et al. Prospective Practices: No Immediate and intermediate Used items from two scales: Positive effect:
(2003)492 longitudinal Role conflict and role ambiguity; adaptive strategies Bachman et al. (1978)493 Role conflict had a positive
USA 661 employees institutional socialisation tactics; Career strategies inventory: Penley impact on immediate career-
DOI: 10.3310/hta14510
153
Impact of HRM practices in non-health settings
157
Impact of HRM practices in non-health settings
applicant attraction and intention once pass–fail socialisation (as opposed to individual, informal,
feedback had been given. This suggests that test random, variable, disjunctive and divestiture
outcome favourability (passing or failing) predicts socialisation, which characterise individualised
organisational outcomes more consistently than do socialisation tactics).
perceptions of procedural justice. The studies used
quasi-experimental design and the findings were Details of studies
replicated across both time and different samples. Nine studies met the inclusion criteria (Table 46).
All of the studies were published in English. Eight
Omissions studies were conducted in the USA and one in
One study used employee performance the UK. Three studies used quasi-experimental
(productivity) as an outcome measure of the methods and six had a prospective longitudinal
recruitment and selection techniques. All other research design. Most of the studies were
studies concentrated on employee-centric conducted over a relatively short duration, with the
outcomes. As most selection tools and techniques shortest time period being 4 weeks and the longest
are designed for, and aim to, select employees who being 2 years. The sample size of these studies
are best suited for a given job, it was expected that varied from 61 to 661.
more longitudinal studies would be available on
this theme. However, this review found a paucity of HRM practice/intervention
longitudinal research on such relationships. All of the studies investigated the benefits of some
form of induction or orientation programme
Summary implemented by various organisations to initiate
There are seven studies in this review that their new employees.
investigated how various recruitment and selection
techniques impact on a range of employee and All but one study510 used the classifications of either
organisational outcomes. All but one of the studies Van Maanen and Schein508 or Jones,509 as described
used employee-centric variables as outcomes; only above. Cooper-Thomas et al.510 investigated
one study investigated the impact of using referrals the efficacy of a computer-based orientation
from existing employees on the productivity of new programme to impact on later socialisation of
employees. The results from three of these seven employees, their job satisfaction and organisational
studies were particularly strong as they were based commitment.
on experimental and quasi-experimental research
designs. Outcomes
All but one study511 investigated the impact of
Induction/Orientation various socialisation tactics on employee-centric
Most organisations provide some form of outcomes, including level of later socialisation, job
induction, orientation, or socialisation programme satisfaction, organisational commitment, newcomer
to new recruits. These programmes may be adjustment, stress, intention to quit, intention to
conducted in various ways, depending on the remain and achievement of person–organisation
requirements of the position and the organisation. fit. Ashforth and Saks511 investigated the impact
Van Maanen and Schein508 propose that different of socialisation practices on self-appraised
socialisation tactics produce different results in performance.
newcomers’ adjustments to their new roles. They
identify six dimensions of socialisation tactics, with Results
each dimension consisting of a bipolar continuum: The results of the studies in this section largely
collective–individual, formal–informal, sequential– supported the conclusion that socialisation
random, fixed–variable, serial–disjunctive, and practices achieve their intended effects.
investiture–divestiture.
• Five studies found a positive impact of
Other authors have used this classification and institutionalised socialisation tactics on job
have proposed different combinations of these six satisfaction.499,509–512
dimensions. Jones509 grouped the six dimensions • Institutional socialisation tactics were
in two sets and classified them as ‘Institutionalized also found to have a positive impact on
socialisation tactics’ and ‘Individualized organisational commitment.492,509,510,512,513
socialisation tactics’. Organisations that scored • Wesson and Gogus514 found that a computer-
high on institutionalised tactics use orientation based orientation programme had a negative
programmes that emphasise a collective, formal, impact on job satisfaction, organisational
158 sequential, fixed, serial and investiture form of commitment and levels of later socialisation.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
• Studies by Waung,512 Jones509 and Chatman499 orientation programmes on given outcomes. Most
concluded that institutionalised socialisation of the studies had employee-centric outcomes
tactics had a positive impact on intention (e.g. job satisfaction, organisational commitment,
to remain with the organisation and, thus, person–organisation fit, and intention to remain)
reduced turnover. However, Waung512 found as the predicted variables of these socialisation
that providing relevant and potentially practices. Most of the research in this field
negative information about the job during centred on the classification of socialisation tactics
the process of orientation did adversely affect proposed by Van Maanen and Schein508 and
some new recruits, who left the organisation Jones.509 There was overwhelming support for the
in the first few weeks. Those who successfully positive impact of institutionalised socialisation
coped with the initial problems, and who used tactics over individualised tactics on various
the help provided during the orientation employee-centric outcomes.
programme, were more likely to remain with
the organisation. Work scheduling
• Person–organisation fit is another variable that Organisations use various forms of work scheduling
is related to socialisation tactics. Three studies to carry out their day to day activities. Some work
found vigorous socialisation, characterised schedules are driven by function, for example
by sequential, fixed, serial and investiture round-the-clock running of hospitals and factories,
characteristics, leads to better person– necessitating the use of shifts. Other forms of
organisation fit.499,510,515 work scheduling are intended to enhance the work
• Two studies researched the impact of experience of employees by, for example, making
institutionalised socialisation tactics on a more family friendly environment through HRM
later socialisation with other employees practices, such as flexitime and compressed work
in the organisation. Both found a positive weeks.
relationship between the two.493,512 The study
by Bravo et al.492 found that institutionalised Details of studies
socialisation significantly improved the Six studies met the inclusion criteria (Table 47). Five
relationship between new and old employees, studies were conducted in the USA and one was
leading to enhanced role clarity, and had a conducted in the UK. Three studies used quasi-
positive impact on workers’ career-enhancing experimental methods, and three had a prospective
strategies. longitudinal research design. The duration of the
• One study511 investigated the impact of studies varied from 2 weeks to 2 years. Sample sizes
socialisation tactics on employee performance varied from 68 to 1325 employees.
using self-rated performance as the dependent
variable rather than objective criteria. Ashforth HRM practice/intervention
and Saks511 found that although individualised All six studies focused on HRM practices that
socialisation tactics were not related to several govern employee work schedules. Practices covered
employee-centric outcomes, when contrasted in this section were the impact of flexitime,
with institutionalised socialisation tactics they compressed work week, monitoring the time of
did have a positive impact on self-perceived employees on job, temporary versus permanent
performance. employment and perception of procedural justice
of work schedule on various outcomes.
Omissions
The studies on induction and socialisation largely Outcomes
concentrated on employee-centric outcomes. There The various outcomes measured by the studies in
were no longitudinal studies that researched the this synthesis can be grouped into four categories:
impact of socialisation tactics on employee learning
and growth behaviour. There is also a dearth of • financial outcomes, including organisational
research on the impact of socialisation tactics on effectiveness and overtime costs
performance, with this review finding only one • customer or client outcomes, with one study
study based on self-rated employee performance researching quality of client service
that met the inclusion criteria. • employee performance, for example two
studies on work performance and one each on
Summary productivity and error rates
There are nine studies in this review investigating • employee outcomes, for example affective
the impact of various socialisation tactics and states of employees, sickness absence, personal
159
None
Klein and Quasi- Practices: No Level of later Level of socialisation: Chao et al. Positive effect:
Weaver experimental Orientation programme socialisation; (1994) 517 Orientation programme had a
(2000) 513 116 new hires organisational Organisational commitment: Allen positive impact on both level of
Moderators/mediators: commitment
USA 3 months and Meyer (1990)486 later socialisation and organisational
None commitment
Wesson Quasi- Practices: No Level of later Measure of socialisation: Chao et al. Negative effect:
and Gogus experimental Computer-based orientation socialisation; (1994) 517 Computer based orientation
(2005) 514 261 new hires programme organisational Job satisfaction: Hackman and programme led was negatively
USA commitment; job Oldham (1975)195 related to all of the three outcomes
5 months Moderators/mediators: satisfaction
None Organisational commitment: Allen
and Meyer (1990)486
Jones Prospective Practices: No Custodial role Scales for custodial role orientation Positive effect:
(1986) 509 longitudinal Socialisation tactics orientation; and intention to quit were devised Institutionalised socialisation has a
USA 102 new hires (institutionalised socialisation vs organisational by the authors positive impact on all four outcomes
individualised socialisation) commitment; Organisational commitment: Porter
5 months job satisfaction; Moderation effect:
Moderators/mediators: et al. (1974)244
intention to quit The effects are more pronounced
Self-efficacy Job satisfaction: Dunham and for individuals with lower self-
Herman (1975) 518 efficacy
Chatman Prospective Practices: No Person– Person–organisation fit: Positive effect:
(1991)499 longitudinal Socialisation organisation fit; Organisational culture profile Vigorous socialisation resulted in
USA 171 auditors intention to leave; (O’Reilly et al. 1991) 506 better person–organisation fit,
Moderators/mediators: job satisfaction
14 months Satisfaction: Faces Scale (Kunin low intention to quit and more job
None 1955) 507 satisfaction
Study, Design, sample Practice(s)/intervention(s);
country size, duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Ashforth Prospective Practices: No Self appraised 519 Positive effect:
Role innovation: West (1987)
and Saks longitudinal Socialisation tactics performance; role Role ambiguity and role conflict: Institutionalised socialisation had a
(1996) 511 222 new (institutionalised socialisation vs innovation; role Rizzo et al. (1970) 503 positive impact on all the outcomes
DOI: 10.3310/hta14510
161
Impact of HRM practices in non-health settings
leave and job satisfaction. Of the employee- The other scheduling and staffing issue
centric outcomes, there are three studies each investigated in studies was the perceived
on absence and job satisfaction. procedural justice of work schedule. Posthuma
and Campion526 found that perceived procedural
Results justice in work assignments and work schedule
Two studies of flexitime used quasi-experimental was positively related to employees giving
design. A study by Narayanan and Nath522 reported their permission for their names to be used in
that the effects of flexitime and introduction recruitment advertisements.
of time monitoring devices on employees were
moderated by the hierarchical position of Omissions
employees in the organisation. For employees at Although there were no notable omissions, it is
lower levels (non-exempt employees), flexitime had difficult in this synthesis to draw any evidence-
a beneficial impact on their social relationships and based conclusions about the impact of a given
enhanced their flexibility. However, there was no practice on a given outcome because the papers
impact of flexitime on professionals, except a trend considered are heterogeneous, focusing on
towards deterioration in superior–subordinate different aspects of the work schedule and on
relationships. The professionals resented the different outcomes. Therefore, the evidence cannot
introduction of punch cards and preferred to be synthesised into a single bottom-line conclusion.
negotiate work schedules with colleagues and Given the prevalence of the organisational use
managers on an informal basis. For managers, the of flexitime, compressed work week, and shift
introduction of flexitime had no impact on any work, it was anticipated that there would be more
of the outcomes. Job satisfaction was found to be longitudinal research on these variables. However,
unrelated to the use or non-use of flexitime in the this review found relatively few longitudinal studies
organisation. In contrast, the study by Dunham in this area.
et al.523 found that use of flexitime along with
compressed work week led to reports of greater Summary
organisational effectiveness, with employees The six studies in this review researched how
exhibiting positive reactions towards both practices. different forms of work schedules (flexitime,
compressed work week, shift work and perceived
Latack and Foster524 using quasi-experimental procedural justice of assigning work schedule)
design, suggested substantial organisational impact on various employee and organisational
gains including reduced sickness–absence costs, outcomes. Only flexitime and compressed work
overtime costs, and personal leave time, as a week were studied by two papers; all other
result of using compressed work weeks. However, practices were examined by only one study. A wide
use of a compressed work week had no effect on range of outcomes was investigated, including
either job satisfaction or error rates. The study financial outcomes in the form of overtime costs,
also investigated the interaction of participative customer outcomes such as quality of client
decision-making in the adoption of a compressed service, performance outcomes (e.g. error rates,
work week, and concluded that such participation productivity and work performance) and employee
has a positive impact on satisfaction with the outcomes in the form of affective employee states,
work schedule but not on other outcomes, such as job satisfaction and sickness–absence rates. The
sickness–absence costs, overtime costs and personal results of this section cannot be presented in an
leave. additive manner as the practices and outcomes
studied are varied. Hence, each study needs to
Markham et al.525 investigated the impact of shift be considered in isolation. More longitudinal
work on the absence behaviour of employees, research needs to be undertaken to arrive at a firm
concluding that all four factors (the shift and three conclusion as to how each of these practices can be
temporal factors of year, season and day of the reliably linked to achieving a given outcome.
week) combine to account for 17% of the variance
in absenteeism. Though no main effects were found Training and development
for the impact of shift time on absenteeism, there
were significant temporal effects on absenteeism. Studies in this section provide evidence on the
Also, temporal effects of year, season and day of various forms of training and development
the week interacted with shift timings to have practices in organisations. This review found the
significant effects on employee absenteeism. following subclusters in the longitudinal studies
conducted on these variables:
162
TABLE 47 Staffing (work scheduling)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
163
164
TABLE 47 Staffing (working schedule) (continued)
165
Impact of HRM practices in non-health settings
performance (e.g. productivity and profitability) satisfaction, ability to cope with the new place and
and reduced long-term costs.149,533,534 job performance.512,548,549
167
an investment leads to better organisational or The duration of the studies varied from 2 weeks
employee performance. to 20 years. The sample size of these 29 studies
varied from 38 to 3355 employees. Six studies
Summary reported data from organisations as a whole; one
There are 26 studies in this review on training and investigated four plants, one had data from five
development. Two studies used a RCT method, federal offices of the USA, one had data from 20
10 used a quasi-experimental design, 12 used a local and regional federal offices of the USA, two
prospective longitudinal design and two used a investigated five units of a single organisation, and
retrospective longitudinal design. The studies one used organisational data from 42 different
in this section can be grouped in five major firms.
categories; performance and skill enhancement
training, career or employee development training, HRM practice/intervention
cultural sensitivity training, socialisation training All studies used one of the four types of
to understand organisational culture, and TQM compensation and reward systems intended to
training. Each study used a distinct outcome achieve various types of outcomes mentioned
relevant to the training being provided. Overall, above.
the results indicated that training predicted desired
impacts, both at the individual and organisational • Eight studies on incentive plans investigated
level, and provides worthwhile return on how various types of financial incentives,
investment. However, the review found a lack of bonuses, ESOPs and gain-sharing could achieve
longitudinal research on the efficacy of the high the desired outcomes.
expenditure training programmes, for example • Nine longitudinal studies were found on merit
executive MBA, used by organisations to up-skill pay, performance-related pay, earnings-at-risk
their employees. pay plans and employees’ perceptions of fair
pay. These studies focused on the use of pay or
Compensation and rewards rewards based on employee performance, and
showed the relative merits and demerits of the
Organisations use various types of compensation various plan types.
and reward plans as instruments for influencing • There were four studies on the impact of
numerous objectives of vital organisational interest. promotion as an incentive for achieving the
In this review, the types of compensation and desired outcomes.
reward systems that have been investigated by the • Finally, this section also contains eight studies
researchers using longitudinal design are: that used a combination of various HRM
practices along with an incentive plan to
• various types of incentive plans, bonuses, achieve given outcomes.
employee stock option plans (ESOPs) and gain-
sharing Outcomes
• merit-based pay, performance-related pay, A wide range of outcomes were measured in these
earnings-at-risk plans and perceptions of pay 29 studies. Employee performance in one form or
fairness. the other was the most studied outcome, with 14
studies investigating job performance and three
Promotions studies concentrating on productivity gains as a
Combinations of compensation and rewards with result of compensation and incentive plans. Job
various other HRM practices. turnover or intention turnover was the other major
outcome of interest in this area. Five studies had
Details of studies job turnover as an outcome and two had intent to
Twenty-nine studies met the inclusion criteria turnover as the dependent variable. Absenteeism
(Table 49). Twenty-three of these studies were was studied in three research papers.
conducted in the USA and one each in France,
Hong Kong, Ireland, India, Israel and the Other dependent variables investigated were bonus
UK. One study used experimental design with or pay satisfaction, organisational commitment,
full randomisation, eight studies used quasi- occupational commitment, intrinsic motivation, job
experimental methods, thirteen had a prospective involvement, job satisfaction, job anxiety, physical
longitudinal research design and seven used a health and mental health.
retrospective analysis of longitudinal data sets.
168
TABLE 48 Training and development
USA engineers training, discussion group and no ratings) authors Results were more pronounced after
training group) 6–12 months of training
18 months
Eden et al. Seven field Pygmalion Leadership Training No Improvement in leadership Trainees rating of their No effect:
(2000) 540 experiments: six using Workshop (Pygmalion effect such that it will augment leader There was little evidence that
Israel experimental design being a special case of self- subordinates self-efficacy leadership training workshops
and one using quasi- fulfilling prophecy in which raising influenced either the leaders or their
experimental; sample leader expectations regarding followers
of each experiment was subordinate achievement produces
different and were very an improvement in performance)
diverse, for example
army personnel, youth
training camps, blue
collar supervisors,
school principles,
continued
Health Technology Assessment 2010; Vol. 14: No. 51
169
170
TABLE 48 Training and development (continued)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
171
172
TABLE 48 Training and development (continued)
7. Cooperation between
combining firms
Waung Quasi-experimental Practices: No Intention to remain; job Intentions to remain Positive effect:
(1995) 512 design Orientation programme satisfaction; organisational and job satisfaction The orientation programme had a
USA 61 new hires (experimental group received commitment were measured using positive impact on all the outcomes in
information warning of negative scales constructed by the long run
4 weeks the authors
aspects of job and information
about specific coping behaviours Organisational
plus training in cognitive commitment: Mowday
restructuring and positive self- et al. (1979) 516
talk and statements to bolster
self-efficacy as a part of the
orientation programme)
Moderators/mediators:
continued
Health Technology Assessment 2010; Vol. 14: No. 51
173
174
TABLE 48 Training and development (continued)
175
176
TABLE 49 Compensation and rewards
continued
Health Technology Assessment 2010; Vol. 14: No. 51
177
178
TABLE 49 Compensation and rewards (continued)
continued
Health Technology Assessment 2010; Vol. 14: No. 51
179
180
TABLE 49 Compensation and rewards (continued)
Intentions to quit
Performance: Organisational For external LOC individuals,
Performance rating records attitudes returned to baseline after
Absence frequency 6 months
Mixed effects:
Absenteeism and job performance
both decreased among promotees
Pergamit Retrospective Promotion No Job attachment; job Source: National Longitudinal Mixed effects:
and Veum longitudinal satisfaction Survey of Youth Promotion led to increased job
(1999) 575 3355 young men satisfaction There is no evidence that
USA and women promotion can lead to higher job
working in private attachment
sector
12 months
Pritchard et al. Quasi- ProMES: No Productivity ProMES Positive and interaction effect:
(1988) 576 experimental A system that developed Job satisfaction Job satisfaction: Seven items Results indicated that group-level
USA Five organisational productivity measurement in order Turnover intentions adapted from Minnesota feedback increased productivity on
units to provide feedback to measure Satisfaction Questionnaire average by 50%, which was raised to
productivity; comprising feedback, Morale (Weiss et al. 1967)473 76% when goal setting and incentives
2 years
goal setting, and incentives based Role clarity Turnover intention: Own were added to it
on achieved goals questionnaire Work attitudes for example job
Morale: Adapted from ISR satisfaction, turnover intentions,
instruments (Seashore et al. and morale were good or better
1983) 595 following intervention
Role clarity: Adapted from Rizzo No such changes were observed in
et al. (1970) 503 the control group
Study, Design, sample Practice(s)/intervention(s);
country size, duration moderators/mediators (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Pritchard et al. Prospective ProMES: No Productivity ProMES scale was used to Positive and Interactions effects:
(1989) 577 longitudinal measure productivity
DOI: 10.3310/hta14510
continued
Health Technology Assessment 2010; Vol. 14: No. 51
181
182
TABLE 49 Compensation and rewards (continued)
ISR, Institute of Social Research; ITQ, intention to quit ; JI, job involvement; JS, job satisfaction; LRD, Longitudinal Research Database; OCQ, Organizational Commitment
Questionnaire; ProMES, Productivity Measurement and Enhancement System; PSQ, Pay Satisfaction Questionnaire.
For National Longitudinal Survey of Youth, see Pergamit MR, Pierret CR, Rothstein DS, Veum JR. Data watch: the national longitudinal surveys. Journal of Economic Perspectives
2001;15:239–53.
For National Employees Survey (NES) data, see Cappelli P. The National Employer Survey: employer data on employment practices. Industrial Relations 2001;40:635–47.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
continued
185
186
TABLE 50 Communication (continued)
with by use of a restorative strategy using honest design, while four used a prospective longitudinal
communication with employees. design. The studies can be divided in to two
categories. In the first category were studies that
The outcome for the second category of dealt with communication by the organisation
communication papers discussed in this section was to create awareness of a given programme or
employee perception of RBSE. any impending changes. In the second category
were studies dealing with quality and quantity of
Results communication briefs given to employees and their
The results of the communication strategy on impact on RBSE. The results of these interventions
the desired outcomes were mixed. For the first are mixed. There was some support for the impact
category of studies, Schweiger and DeNisi596 of tailored communication strategies dealing with
found that realistic merger previews were a good specific issues as opposed to general organisational
tool for limiting the dysfunctional outcomes of a communications; however, this was based on only
merger, resulting in lower stress, job dissatisfaction, five eligible studies.
absenteeism and job turnover, and improved
commitment and employee performance. However, Family friendly
the study by Henessey et al.597 found that although
a benefit awareness programme had a positive The term ‘family friendly workplace’ (or ‘work
impact on satisfaction with benefits and the and family’) is one which recognises the non-
organisation, it had no effect on employee job workplace family responsibilities of its employees,
satisfaction. Similarly, the study by Kinnunen et and develops and implements policies that aim
al.598 found that a dedicated restorative strategy to help employees simultaneously fulfil work
using honest communication with employees, to and family commitments. Some authors extend
provide reassurance of job security and maintain the use of the term to include practices that are
a positive work atmosphere, did result in lower employee friendly and take care of the well-being
levels of job insecurity and improvements in of employees and their families (such as general
relationships with the colleagues. However, the employee assistance programmes or specific help
general communication system in the organisation for drug or alcohol addiction).
was completely unrelated to employees’ feelings of
job security. Details of studies
Four studies met the inclusion criteria (Table 51).
The studies by Parker454 and Axtell and Parker458 All studies were conducted in the USA. One study
both concluded that, although quality of had a quasi-experimental design, one study had a
communication briefings had a positive impact on prospective longitudinal design, and two studies
employees perception of their RBSE, the quantity used a retrospective longitudinal research design.
of such communication had no similar impact. The duration of the studies varied from 12 months
to 7 years. The sample size of the studies varied
Omissions from 207 employees to 60 sites encompassing
Management theory places great emphasis on 43,888 employees.
devising organisational communication plans for
internal and external stakeholders, yet it appears HRM practice/intervention
that little of that has translated in longitudinal Of the four studies in this section, three
peer-reviewed research. Five studies met the review investigated the impact of health promotion
criteria for HRM practice in this area, but they programmes or policies, elements of which
focused on a limited range of possible outcomes. included family friendly practice.603–605
There appears to be substantial scope for further
longitudinal research on the impact of different The fourth study by Hammer et al.606 investigated
communication practices. couples’ or individuals’ use of alternative work
arrangements and dependent care support.
Summary
The five studies in this section provided evidence Outcomes
on how various forms of communication can be The outcomes investigated these studies of family
used to achieve the employee outcomes of reduced friendly practices were absenteeism (two studies),
dysfunctional outcomes during a merger, feelings return on investment, medical claims made, job
of job security, job satisfaction and enhanced turnover, job satisfaction, and family–work/work–
RBSE. One study used a quasi-experimental family conflict.
187
Gilleskie and Retrospective Employer provided health No Employment transition Whether an individual has No effect:
Lutz (2002) 605 longitudinal insurance changed employer Employer provided health insurance
USA 4422 individuals (all Moderator: had no impact on married employees’
males) Married/unmarried decision to change their jobs and had
12 months a very small impact on the unmarried
males
Hammer et al. Prospective Utilisation of available No Work–family conflict; Work–family conflict: Positive effect:
(2005) 606 longitudinal workplace support (couples’ or family–work conflict; (Metemeyer et al. 1996) 607 Use of alternative work arrangements
USA 234 couples individuals’ use of alternative job satisfaction Job satisfaction: JDS and dependent care supports are
work arrangement; couples’ or (Hackmann and Oldham positively related to family–work
12 months individuals’ use of dependent care 1975)195 conflict in dual-earner couples; use of
support) workplace support is positively related
Moderator: to job satisfaction over time
None No effect:
Couple level utilisation of workplace
support does not appear to have a
strong impact on individuals’ report of
work–family conflict
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Practice(s)/intervention(s);
Study, Design, sample moderators/mediators
country size, duration (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Coch and Quasi-experimental Employee participation in the No Productivity; labour Company data Positive effect:
French design process of introducing changes turnover Participation of employees in the process
(1948) 608 46 employees to the job of making changes to their jobs reduces
USA 32 days Moderator/mediator: the resistance to change and increases
None their post change productivity. It also
helps in reducing employee turnover
Morgeson et al. Quasi-experimental Employee participation in job No Pay satisfaction PSQ (Horeman and Schwab No effect:
(2001) 609 168 employees evaluation process 1985) 615 Participation in the job evaluation
Impact of HRM practices in non-health settings
Kleinbeck and Quasi-experimental Participative productivity No Productivity Organisational productivity Positive effect:
Fuhrmann 43 employees management (PPM) system Group cohesion measure PPM led to improved productivity
(2000) 610 Group cohesion: Adapted
22 months No effect:
Germany from three scales – Widmeyer
et al. 1985;616 Luhtanen and PPM had no effect on group cohesion
Crocker (1992), 617 Wagner
and Zick (1993) 618
Pearson Quasi-experimental Participative goal setting Yes Performance; job Performance: Achievement of Positive effect:
(1987) 611 42 teams satisfaction set goals in the given time Participative goal setting has positive
Australia 12 months Job satisfaction: JDS (Hacman effects on employee performance and
and Oldham 1975)195 job satisfaction
Pearson Quasi-experimental Participative system for No Role ambiguity and 1. Rizzo et al. (1970):503 Role Positive effect:
(1991) 612 76 teams monitoring productivity role conflict; internal index Monitoring and feedback had a significant
Australia Feedback on performance work motivation; 2. JDS (Hackman and Oldham, effect on reducing role ambiguity,
5 months job satisfaction;
(productivity) 1975):195 Internal work increasing job satisfaction and increasing
productivity motivation productivity
3. Job satisfaction (Cammann
et al. 1979) 619
Practice(s)/intervention(s);
Study, Design, sample moderators/mediators
DOI: 10.3310/hta14510
country size, duration (if any) FM Outcomes Outcome measure(s) Results +ve/–ve
Kleingeld et al. Quasi-experimental PMS (ProMES): Participation in Yes Individual performance Performance data from Positive effect:
(2004) 613 182 employees implementation of a PMS data feedback reports generated Implementation of a ProMES was found
Netherlands by the PMS to be better in improving employee
18 months
performance than implementing the
system based on simply telling the
employees what was expected of them
Lee and Son Prospective Appraisal review: Employee No Satisfaction with Three-item scale made by the Positive effect:
(1998) 614 longitudinal get opportunity to participate review; performance authors Goal setting and career discussion of the
Korea 116 employees in discussion; goals are clearly Performance ratings appraisal review were positively related
set; career issues are discussed to employee satisfaction
6 months
No effect:
Overall, appraisal review had no effect
PMS, performance management system; PPM, Participative Productivity Management; ProMES, Performance Management System; PSQ, Pay Satisfaction Questionnaire.
Health Technology Assessment 2010; Vol. 14: No. 51
191
Impact of HRM practices in non-health settings
performance management and staffing, and are majority supported the positive impact of using
therefore discussed in those sections as well. employee participation when making key decisions
that affect their working. There appeared to be less
Outcomes emphasis in the literature on using longitudinal
The following outcomes were studied in the research design to study the impact of participation
articles reviewed in this section: overtime practices, such as collective bargaining and
costs, productivity (three papers), performance employee involvement in strategic decision-
(three papers), error rates, labour turnover, pay making.
satisfaction, group cohesion, job satisfaction
(three papers), role ambiguity and conflict, work Performance appraisal and
motivation and absenteeism. performance management
Results Studies in this section cover performance appraisal
Results from two studies focusing exclusively and performance management practices.
on employee participation arrived at different
conclusions, possibly due to the nature of the Details of studies
outcome measure and area of participation in Twenty-five studies met the inclusion criteria (Table
which the employees were involved (one being a 53). Nineteen of these studies were conducted
very specific intervention). The study by Coch and in USA, two in Australia and one each in the
French608 found that participation of employees Netherlands, Germany, New Zealand and Korea.
in the process of introducing changes to their jobs One study used a complete RCT experimental
led to less resistance to change, improved labour design, nine studies used quasi-experimental
productivity and helped reduce labour turnover. In methods and 15 had a prospective longitudinal
contrast, the study by Morgeson et al.609 found that research design. The duration of the studies varied
employee participation in a job evaluation process from 7 weeks to 6 years. The sample size of these
had no effect on employee pay satisfaction. 25 studies varied from five teacher–student dyads
(10 persons in all) to 4413 employees.
There were five studies on the use of participation
in performance/productivity management HRM practice/intervention
practices. All five studies supported the conclusion The studies were concerned with one of the
that participation and involvement in any form following:
of performance management practice positively
impacts on outcomes, for example productivity, job • feedback
performance, job satisfaction, role ambiguity and • goal setting, alone or in combination with
role conflict.610–614 feedback and/or incentives
• performance or productivity management
The eighth study was on the use of employee systems, alone or in combination with
participation in the implementation of a feedback, goal setting, and incentives
compressed work week. It concluded that those • performance appraisal.
involved in the decision to adopt a compressed
working week expressed greater satisfaction with The largest number of the studies was on the use
their work schedule than those not involved in the of feedback systems and their ability to influence
decision.524 desired outcomes. There were 13 studies that had
feedback as their main independent variable,
Omissions focusing on how rich feedback, characterised by
For this review, two studies had a primary focus on large amounts of specific and positive information
the effects of participation. The remaining studies on one’s work behaviour, can impact on
all included participation as part of the evaluation performance and employee outcomes.
of another HRM practice.
Goal setting, either alone or in combination with
Summary feedback and/or incentives, was investigated by six
There were eight studies in this category, with six studies. The focus of these studies was on Locke’s620
also being discussed in other sections, namely theory of goal setting, exploring how performance
performance management and staffing. Seven of and attitudinal outcomes can be influenced by
these studies used quasi-experimental design and the way in which goal setting is conducted and
one used a prospective longitudinal study. The implemented.
192
TABLE 53 Appraisal/performance management
continued
Health Technology Assessment 2010; Vol. 14: No. 51
193
194
TABLE 53 Appraisal/performance management (continued)
USA 3 months (Smith et al. 1969)242 satisfaction compared with goal setting alone
Campbell RCT Incentives: No pay; hourly pay; goal No Performance Achievement of set Positive effect:
(1984) 631 56 graduate contingent pay; goal setting targets Goal contingent payment with participative
USA students goal setting was better than hourly pay
6 weeks
Kleinbeck Quasi- PPM No Productivity; group Organisational Positive effect:
and experimental cohesion productivity measure PPM led to improved productivity
Fuhrmann 43 employees Group cohesion:
(2000) 610 No effect:
22 months Adapted from three
Germany scales: Widmeyer et PPM had no effect on group cohesion
al. 1985;616 Luhtanen
and Crocker 1992;617
Wagner and Zick 1993618
continued
Health Technology Assessment 2010; Vol. 14: No. 51
195
196
TABLE 53 Appraisal/performance management (continued)
(1998) 614 longitudinal opportunity to participate in review; performance made by the authors; Goal setting and career discussion of the
Korea 116 employees discussion performance ratings appraisal review were positively related to
6 months Goals are clearly set; career issues employee satisfaction
are discussed No effect:
Overall, appraisal review had no effect on
job performance
Westin Prospective Electronic monitoring No Employee perceptions Focus group and survey Negative effect:
(1992) 634 longitudinal of fairness; climate of Introduction of electronic monitoring of
USA 200 customer organisational trust or employees at workplace in place of self
service agents distrust regulated or supervisory monitoring had a
Not provided negative effect on employees’ perception of
fairness and led to decrease in a climate of
organisational trust
MBO, management by objetives; MSQ, Minnesota Employee Satisfaction Questionnaire; PMS, performance management system; ProMES, Productivity Measurement and
Enhancement System.
Health Technology Assessment 2010; Vol. 14: No. 51
197
Impact of HRM practices in non-health settings
There were six studies on how types of by Codding et al.627 reported that use of feedback
performance or productivity management systems can enhance the implementation integrity of an
are related to employee job performance or intervention programme. Feedback also helped
productivity. Three of these studies investigated to improve job satisfaction and reduce job
the impact of a productivity management tool turnover.612,628
[Productivity Measurement and Enhancement
System (ProMES)], on employee productivity and Goal setting, feedback and incentives
attitudinal outcomes. There were three studies on goal setting alone. Two
of these used a quasi-experimental design and one
There were two studies on how different types of used a prospective longitudinal research design.
performance appraisal or performance appraisal All three studies concluded that goal setting has
reviews can impact on given outcomes. a positive impact on job performance and job
satisfaction, and showed that participative goal
Of the remaining studies in this section, one looked setting is particularly beneficial.611,629 However, the
at electronic monitoring of employees at work, one study by Ivancevich630 found that the effects of goal
at recognition, and one at operant conditioning setting on job performance were not permanent
reinforcement schedules to boost performance or and tended to dissipate after 12 months.
other desired outcomes.
The other three studies were on goal setting in
Outcomes combination with feedback or financial incentives.
Sixteen of the 25 studies looked at job performance Two studies concluded that goal setting works
or productivity as one of their outcomes. Job best when it is participative and combined with
satisfaction was the second most investigated adequate feedback.624,625 Campbell631 used an
outcome variable, with eight studies including it as experimental design and reported that incentives
their dependent variable. Other variables that were linked to goal setting led to better performance
investigated in these studies are attendance, safety than when incentives are paid at an hourly rate.
behaviour and turnover.
Performance or productivity management systems,
Results feedback and incentives
Feedback All performance management systems used a
Of the 13 studies on feedback, eight were on combination of practices to achieve their desired
feedback alone and the rest were on feedback outcomes. Of the six studies in this category,
in combination with either goal setting or some five investigated how these systems influenced
form of performance or productivity management productivity or employee performance. All of these
system. Four studies investigated how feedback studies concluded that performance or productivity
impacts on job performance – all four used a management systems enhanced employee
prospective longitudinal design and indicated that performance or increased productivity.576,577,610,612,613
feedback leads to a significant improvement in job Three of these studies were on the use of ProMES
performance of employees. Two of these studies and all three concluded that, although feedback
used upward feedback (subordinate feedback alone improves productivity by around 50%
on supervisor performance) as an intervention when combined with goal setting and incentives,
and found that it enhanced managerial job productivity goes up by 76%.576,577 The sixth
performance.621,622 One study investigated study on the use of a performance management
how feedback sharing with raters and seeking system explored the impact of a newly installed
suggestions for improvement can influence the performance management system, along with
ratee’s performance. The study concluded that, pay for performance on employee attitudes. The
although there was a statistically significant study found that high and low performers reacted
difference between the ratees who shared their differently to use of a performance management
feedback and sought advice and those who did not, system involving pay for performance. For low
the effect size of improvement in performance was performers, the new performance management
very low and explained only 2% of the variance.623 system substantially increased ratings of satisfaction
Two studies on feedback, using quasi-experimental and cooperation with one’s supervisor. In
designs, found that a significantly positive impact contrast, high performers had a high baseline
on employee productivity.612,624 Two other studies, for these attitudes towards supervision, followed
one quasi-experimental and the other prospective by substantial drops immediately after receiving
longitudinal, concluded that feedback enhanced appraisal and bonus pay.632
198 the safety behaviour of employees.625,626 The study
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
‘supportive’ of this participation (Gittleman et al. it is this, and not the particular practices per se,
1998: 102).643 However, as Wood and Wall31 show, that is really yielding high performance.
the involvement aspect has increasingly been
neglected in the empirical work on HRM systems, Through our analysis of individual practices, which
at least in the stream of cross-sectional studies that has examined the connection between these and
developed in the 1980s following pioneering work outcomes, we have identified a set of practices that
by Arthur,14 MacDuffie,20 Huselid18 and others. As may be best in each domain. We could, therefore,
the resource-based theory has been increasingly create a complementary set on the basis of this.
used to justify the association of HRM systems with However, to be certain they are complementary, we
high performance, the seeds for a diminishing would need to ascertain that they have no negative
role for employee involvement were set relative impact on the effect of each other. For example,
to emphases on skills and knowledge, labour it is often argued that individual performance-
flexibility and extrinsic motivation. This theory related pay will detract from the positive effects of
stresses that organisations achieve competitive teamwork.
advantage through having distinctive, if not
unique, resources that are not easily imitated In our analysis of individual practices, we reported
or substitutable.644,645 It thus sees HRM systems any synergistic relationships between practices. For
as particularly important for generating the example, in our review of performance appraisal
competences of employees, and some even talk of studies we report two studies that show that goal
the HRM system itself becoming a unique asset of setting works best when it is participative and
the organisation, though this is not consistent with combined with adequate feedback.624,625 We have
the idea of there being a generic type (or types) of not examined, however, studies that have used
system(s) that should result of high performance. composite measures of HR practice or systems, as
those studies typically do not compare the effect of
The emphasis on the need for practices to cohere individual practices with those of their composite
if performance is to be optimised remains. The measure.
argument is, to use the words of Appelbaum et
al. (2000: 34) ‘that firms adopting a coherent In this section, we will introduce and report the
set of workplace practices designed to maximise results of studies that use composite measures
horizontal fit should have superior performance’. of the HR system. We will continue to use our
The authors take cohesion to mean that (1) the methodological selection criteria that limits our
workplace practices are complementary; (2) search to longitudinal, experimental or quasi-
synergies exist between the practices leading experimental studies.
to positive interaction effects on performance;
and (3) the practices form an integrated system. The studies are concerned with assessing the
However, these meanings are distinct and should impact of a particular system or subsystem of HRM.
not, as Appelbaum et al.13 and others do, be For example, Freeman and Kleiner649 concentrate
conflated. Following Delery646 and Wood and de on continuous flow production system, which
Menezes,647 these should be seen as distinct ways is centred on a time–rate compensation system
of conceiving a horizontal fit. As Wood and de complemented by other HR changes, for example
Menezes647 outline, a complement of practices a new safety programme. This is contrasted with
consists of all those practices that individually a piecework system in which payment is based
have a positive association with performance. As directly on output and the system of production
such, a complement of high-performance practices is highly individualised. Or in the case of Shipton
would be made up of the practices that are best et al.650 the central concept is the sophistication of
in each of the domains of HRM. Each would add HRM.
something unique and not detract from the effect
of any other. Synergistic practices are those that The studies are of two basic types: (1) ones that
enhance the effect on performance of another. A measure the system through a composite variable
high-performance synergistic set would be one in based on the extent of use of a set of practices
which all practices interact positively with each to capture a system – as in the case of Shipton et
other, so that the combined use of the practices al.’s650 sophistication of HRM – and (2) case studies
has ‘a greater effect on performance than the sum before and after the introduction of a new system,
of effects of the individual practices’ (Appelbaum as in Freeman and Kleiner’s649 study of a change
et al. 2000: 134).13 Finally, an integrated approach from piecework system to the new continuous flow
implies that the practices reflect an underlying production.
200 distinctive style648 or orientation to HRM, and that
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Perhaps the simplest concept used was the HR The core concept of Rauch et al. (2005: 683)645
practice of Wright et al.37 This was measured by was human development and utilisation, which
simply aggregating the use of nine practices: three is defined in terms of a set of ‘practices used
concerned with selection and staffing, one with the for enhancing employee skills through training
extent of training, three with pay for performance, and other forms of skill enhancement’. They
and two with participation – all measured by a included practices such as employee participation,
simple binary divide (have or have not). empowerment and communication in the other
forms of skill enhancement.
In a similar vein, HRM was the central concept
of Guest et al.27 It was again measured by an Ichniowski et al.21 assumed that there is a set of
index containing 48 items that covered nine innovative HR practices that are at least applicable
areas: recruitment and selection, training and to traditional manufacturing industries, for
development, appraisal, financial flexibility, job example steel. Innovative work practices are:
design, two-way communication, employment incentive pay, high screening in recruitment
security and the internal labour market, single and selection, teamwork, employment security,
status and harmonisation, and quality. flexible job assignments, high skills training and
information sharing, and line manager–worker
Shipton et al.650 considered HRM systems in meetings on production and other issues. Using
terms of their degree of sophistication, which, cluster analysis, Ichniowski et al.21 empirically
although not conceptually defined, was rated by identified four types of plants: at the two extremes,
an interviewer. Five dimensions were assessed: ‘innovative’ plants used all the labour practices,
performance management, recruitment and whereas ‘traditional’ plants used none of these
selection, induction, training, and strategy. A practices. The two intermediate types both had
high score on each represents sophistication teams, in contrast with the traditional system, but
and extensiveness, for example in the case of differed in the type of other practices they utilised.
performance management, not only is there a The more innovative of the two intermediate types
formal appraisal scheme, but also appraisals are differed from the high innovative system in that
frequent, appraisers are trained and it extends to it did not tend to have screening in recruitment,
all employees. employment security or job rotation, while the
other intermediate type tended only to have
Workman and Bommer654 focus on high- information sharing and line manager–worker
involvement work processes. Three types of meetings. It was having these that differentiated it
interventions, effectively measuring three degrees from the ‘traditional’ plants.
of high-involvement management, were designed.
The type that was the least level of involvement, Katz et al.652 reported a similar interplant study,
labelled ‘alignment job design’, entailed removing but studied only one firm and focused on a
impediments to effective motivation and the company-wide quality of work programme,
201
continued
Health Technology Assessment 2010; Vol. 14: No. 51
203
204
TABLE 54 Bundles (continued)
which, in Ichniowski et al.’s21 terms, was aimed at studies were economic; productivity was included
innovative practice. The practices used to measure in six and profitability or financial performance
the intensity of the quality-of-work programme in four. Workers’ compensation, a measure of the
were specific to the programme, ranging from wages bill over sales revenue, was more a measure
upward communication programmes to alcohol- or of efficiency than compensation. The other
drug-abuse programmes, and also included some economic outcomes used in the studies, mainly
procedural elements, for example the use of an in one only, were: quality, shrinkage, operating
outside consultant or quality-of-work specialist. expenses, employment growth, employment
stability, innovation in product, processes or
Schuster653 focused on unionised plants where the technology, and export performance. Workman
union and management had explicitly developed and Bommer654 used performance measures
a cooperative relationship. The concept of that were primarily specific to call centres, for
union–management cooperation, the impact of example percentage of calls escalated, but also
which Schuster653 assessed, reflects the notions of used a customer service measure. The other
Kochan about the changes in industrial relations in outcome measures used were concerned with HR
unionised settings in the early 1980s, also reflected or industrial relations measures – again mainly in
in Katz et al.’s652 study above, and culminated one study only – and included: labour turnover,
in Kochan’s655 book with Katz and MacKersie, absenteeism, organisational commitment, job
The Transformation of Industrial Relations. In satisfaction, industrial conflict or disruption,
Schuster’s653 study there were differences between contract demands, negotiation time, employment
the novel practices, which varied across the nine climate and grievance rate.
manufacturing plants, but they most commonly
involved plant-wide productivity-sharing plans, Results
labour-management committees and work redesign Wright et al.37 showed that HR practices lower
projects. operating expenses and increase profitability,
and are also associated with organisational
Freeman and Kleiner649 contrasted continuous commitment, which Wright et al.37 took to be
flow production with a piecework system, and suggestive of commitment mediating the link
conceptualised both primarily on the basis of their between HR practices and outcomes, although this
compensation systems, as these were, for them, is not formally tested.
‘a central element to any set of human resource
policies’ (ibid: 308). Nonetheless, the systems The results of Guest et al.’s27 study showed a
entailed complementary practices: piece rates positive association between HRM and profitability
were similar to supervisors who monitored the and a negative relationship between HRM and
quality output and set rates, while the change to labour turnover. It was not, however, related to
a continuous flow system included a change in productivity. Moreover, when prior profit was
the production system, teamwork, a new safety included in the profitability equation, profit was
programme, as well as time rates. Such a contrast no longer significantly associated with HRM.
in systems is most applicable to shoe and apparel This suggests that profits may well lead to the
manufacturing, where piece rates have continued introduction of HRM not vice versa, this conclusion
to be extensively used. being enhanced by the lack of association between
HRM and productivity.
Gomez-Mejia’s651 focus was specifically on
international HR strategy. The measure used was a Shipton et al.650 revealed a positive link between
seven-item scale that focused on how HR practice sophisticated HRM and innovation in product
is geared to international elements, for example and production technology, but not in process
international experience is highly desirable when technology.
managers are hired or promoted, whereas some
of the practices included reflect those in the other Workman and Bommer654 showed that the
studies, for example inventive pay, intensive introduction of alignment job design, high-
training. involvement work processes, and autonomous work
teams all increased job satisfaction. However, in
Outcomes the case of autonomous work teams, the effect was
The range of the outcomes used to test the impact moderated by the individual’s preference for group
of the various core HR constructs was equally large. working, so autonomous teams were only potent
The only two outcomes assessed in more than two for those with such a preference. In the case of
205
high-involvement work processes, this preference monitoring costs or costs involved in changing
intensified the positive impact of the processes on styles, were greater than those associated with the
job satisfaction. Hence, while the job satisfaction newly implemented time-based compensation and
of all individuals increased, the satisfaction of continuous flow method. These costs outweighed
those with a strong preference for group working the productivity gains to be made from piece work,
increased to a greater extent. However, only high- so, on balance, the new continuous flow system led
involvement work processes were associated with to an overall better performance.
organisational commitment. Tests of performance
showed that all three systems improved the four Gomez-Mejia’s651 study of international human
performance measures over and above the pre-test resources strategy showed that the extent to which
and control group measures, with the exception the firm had a strategy was positively associated
of one test, which showed that the number of with export performance.
problems resolved by employees was lower in
autonomous work groups than for any other groups Omissions
(including the control group). The main omission within the studies that directly
purport to examine the impact of a HR system is
The core concept of Rauch et al.,645 i.e. human that they did not specify, in any precise way, the
development and utilisation, was positively concept of HRM underlying their work nor did
associated with the only outcome measure they they develop a clear rationale for the inclusion
used, namely employment growth. The effect was of the practices by which HRM was measured.
also shown to be enhanced by the extent of the Consequently the main omission from the set of
human capital of employees in the firm, measured studies is research that systematically defines its
by the owner’s perception of whether employees core concept and in which the correspondence
were qualified to do the job. between this and its measure is precise. There
is also an absence of studies that first examine
Ichniowski et al.21 showed that the productivity the association between the practices before
of steel plants was progressively greater as they constructing measures and as part of the
moved through each of the four categories, from understanding of the reality of the HRM they are
traditional to innovative systems. Innovative studying.
systems were the most productive, while the limited
use of innovative practices had some pay-offs. There is also a lack of studies that test mediation
effects, for example whether organisational
The results of Katz’s et al.’s652 analysis of the commitment or job satisfaction mediates the
performance effects of ‘the quality of work life impact of HRM on performance.
programme’ revealed that the extent of the
programme’s adoption was related to product In the case studies, the practices are determined
quality but not to higher productivity. Its effects by the situation and in broad terms correspond to
on quality may have been partially mediated by the concept of the underlying programme that the
its impact on grievance and absence rates (the researcher is examining, although again in these
two main effects for HR outcomes), which were cases more articulation of this concept would have
negatively related to the extent of adoption. been useful.
Schuster653 showed that in six of the nine plants All studies are in the private sector and are mainly
studied, the introduction of the cooperative found in manufacturing settings. Studies of public
programme increased productivity without the sector and voluntary organisations were not
employment level being cut. The employment identified by this review.
level was reduced in only one of the plants, and,
unfortunately, this was the one plant where no Summary
productivity data were provided. The studies taken together showed that HR
practices do make a difference to performance.
In contrast, Freeman and Kleiner’s649 case study However, limitations in the methods and
showed no productivity effects in the change particularly the measures meant that, in many
from the traditional piece work system to a cases, the underlying concept was not clear
continuous flow method. However, profitability enough for one to say which HR system is (or
was improved. The costs of the complementary systems are) really more effective than others. The
practices associated with piece work, for example variety in the studies also limits generalisation.
206
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Case studies, such as Freeman and Kleiner’s649 examining the effects of different work schedules
offered convincing evidence of effects in one (such as flexitime, compressed work week and shift
situation but generalisation to either other sites or work) on various intermediate and final outcomes.
assumed related or similar interventions would be However, the results cannot be presented in an
premature. But even in some of the case studies additive manner due to the varied nature of the
there was insufficient detail of some practices to practices and outcomes studied.
be certain about what they entailed, most notably
Schuster’s653 insufficient detail of the changes in job The review of the health literature identified
design in his cases. the heterogeneity of the studies included in the
review and that finding is replicated here where
the diversity of research topics in terms of HRM
Conclusions practices and outcomes measured often make it
difficult to develop generalisable conclusions. This
The literature search to identify studies assessing point is well-illustrated by the categories of work
the impact of HRM practices on intermediate and design, staffing and training and development
final outcomes in non-health-sector settings found where the HRM practices clearly fell into distinct
evidence in nine of the ten categories of HRM subcategories.
practices established in Chapter 4. These are:
There are some exceptions to this finding, where
• work design small numbers of papers were sufficiently similar to
• staffing allow cumulative conclusions to be made.
• training and development
• compensation and rewards Within the work design literature, consistent
• communication evidence has been found for the positive impact
• family friendly of increased job control and negative impact of
• employee involvement, representation and high demands on employee outcomes, such as
participation job satisfaction and absenteeism. Additionally, a
• performance appraisal and performance number of concurring studies have established a
management positive link between autonomous team working
• integrated measures. and employee outcomes relating to individual/team
productivity and job satisfaction.
Work design has attracted most research attention
accounting for over 50 of the identified studies. Research investigating job rotation, job
Other areas that have been the frequent focus of enlargement and job enrichment has concentrated
research include staffing (23 studies), training more on the impact of job enrichment alone.
and development (25 studies), compensation and Enriched jobs have been shown to be linked
rewards (29 studies) and performance management to reduced turnover, and to have a positive
(28 studies). In contrast with this, no evidence was relationship with a number of intermediate
found in the area of single status/harmonisation. outcomes including job satisfaction, motivation,
This distribution of studies perhaps reflects the involvement and commitment. Studies also
fact that the practices studied are in the longer mostly supported a positive relationship between
established and more traditional HRM areas. job enrichment and employee performance and
reduced turnover.
Additionally, the majority of studies included in
this section measure individual level outcomes. Within the literature on staffing, the studies
Less research has looked at final outcomes (i.e. investigating differing recruitment and selection
organisational performance). techniques reported broadly positive impacts of
these tools on intermediate outcomes including
As noted in Chapter 7 (Results, Work design), there intention to leave, employee productivity, person–
are differences in the types of HRM practice that organisation fit and job satisfaction.
have been the focus of longitudinal research in
the health and non-health sectors. Less research The studies across five broad training areas
attention had been given here to the specific offer perhaps the best opportunity to develop
staffing practices of skill mix, staff ratios and generalisable conclusions. Although diversity was
working hours. In direct contrast with the health found in the types of training in this category
literature this chapter reports on six studies (five subcategories) within each of these there was
207
reasonable similarity between the studies. Studies of results support the positive impact of using
including performance and skill enhancement employee participation when making key decisions
training provided consistent evidence that this that affect them.
type of training is associated with a range of
positive outcomes. These are mostly measured at Despite these areas of commonality, the review is
the individual level and often reflect the specific characterised by diversity and a lack of cumulative
purpose of the training. Some of these studies data. Overall, there are no more than four or five
include limited evidence on final outcomes, such studies in any of the HRM practice categories that
as profitability and quality outcomes. Overall, the study the same intervention and outcomes. In
majority of studies present positive evidence on the addition, variations in organisational context in the
outcomes specific to the training. non-health literature make it difficult to know how
generalisable any findings are.
A group of studies that fall within both the
performance management and compensation These findings point to significant gaps in
and reward categories looked at the impact of longitudinal research in all of the HRM practice
goal setting and feedback and found consistently categories studied. These gaps exist on both the
positive results. Goal setting and feedback were types of interventions being studied under a given
found to enhance employee performance, practice and also the type of outcomes being
particularly when goal setting is participative. In investigated.
relation to compensation and rewards, several
studies suggest that the effects of goal setting and Despite these limitations, the findings do indicate
feedback are enhanced by incentives. some promising areas and give some indication
of where the general HRM evidence base could
The final area in which a consistent set of findings inform future practice, policy and research in the
was identified is that of employee participation, NHS.
representation and involvement. Here the majority
208
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Chapter 9
Intermediate and final outcomes:
correlations and effects
Introduction outcomes on health and non-health final
outcomes.
This chapter addresses Objectives 3 and 4 of the
review. It first considers findings from the literature The chapter first presents methodology and
on the relationship between intermediate outcomes findings from the meta-analyses in relation to
of HRM practices. It then goes on to look at correlations among intermediate outcomes and
what evidence exists for the causal link between between intermediate outcomes and employee
intermediate and final outcomes. The findings behaviours (Objective 3.1). The methodology
reported here specifically address: and findings for the longitudinal studies linking
intermediate with final outcomes are presented
• Objective 3: 3.1 What is the evidence on the below (see Findings for Objective 4.2 and Findings
intercorrelations of intermediate outcomes? for Objective 4.1).
• Objective 4: 4.1 What is the evidence for the
impact of intermediate outcomes on non-
health final outcomes? Review methodology
• Objective 4: 4.2 What is the evidence for the
impact of intermediate outcomes on patient No systematic reviews were identified in the
care outcomes? searches for Objective 3.1. The methodology in
relation to this objective therefore draws exclusively
Chapter 4 describes the intermediate and final on meta-analytic reviews. We next provide a brief
outcomes, identified from the HRM literature description of meta-analysis followed by details of
for inclusion in this review. In Chapter 4 the the strategy used to identify and assess relevant
importance of a third set of variables (salient meta-analyses.
employee behaviours) was also recognised. In
relation to Objectives 3 and 4 this means the Overview of meta-analysis
relationship between intermediate outcomes and
salient productivity related employee behaviours Meta-analysis is a powerful tool for summarising
(e.g. the association between job satisfaction and evidence of a relationship of interest across a
job performance). Employee behaviours were number of studies. Taking the example of job
therefore considered alongside intermediate and satisfaction and organisational commitment, it
final outcomes. These three groups of variables allows individual studies reporting correlations
form the basis for the searches undertaken for this between these two intermediate outcomes to be
chapter. aggregated to provide an estimate of their true
correlation. Meta-analyses are therefore an effective
Two sources of evidence were identified for this way of quantitatively reviewing the correlations
section, their use depending on the research among the intermediate outcomes and their
objective: relationships with employee behaviours. The
large number of studies reporting correlations
• Meta-analyses and systematic reviews were of interest to Objective 3.1 also points to the use
used to examine the correlations among of meta-analyses as the best source of evidence.
intermediate outcomes and the relationship For example, Judge et al.656 reporting on the
of intermediate outcomes with employee correlation between job satisfaction and job
behaviours. performance, carried out a meta-analysis on 254
• Individual longitudinal studies (in the studies, comprising a sample size of 54,471 for
absence of any systematic reviews) were used all the studies, offering a far better synthesis of
to investigate the impact of intermediate research than non-quantitative methods of review
would allow.
209
For inclusion in the review, meta-analyses had to The quality of all included meta-analyses in this
satisfy the following criteria: section of the review was evaluated using a checklist
developed for this study (see Appendix 10), based
• The study had to be a meta-analysis or on currently available checklists663,664 research on
systematic review providing an analysis of meta-analysis61,659–662,665–667 and piloted on meta-
‘the results of several independent … trials or analyses included in this review. The checklist was
studies’.658 This excludes articles that perform developed to assess the efforts made by selected
analysis on a primary sample collected for meta-analyses and reviews to limit publication bias,
the analysis, but that are described as a meta- selection bias and extractor bias, and to examine
analysis by its authors. the robustness of the results.659,662
• The dependent variable of the meta-analysis
or review had to be one of the selected Details of the assessment of bias of all meta-
intermediate outcome variables listed in analyses included in the review can be found
Chapter 4, such as organisational commitment at Appendix 11. For ease of interpretation, a
or job satisfaction. brief summary of the definitions of correlational
strength and assessment of bias is given below.
210
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
211
The findings for the review of Objective 3.1 are magnitude to suggest construct redundancy,
therefore described as follows: and it is reasonable to conclude that each of the
intermediate outcomes may contribute to efforts to
• summary of the correlations among the understand and manage employee behaviours.
intermediate outcomes
• summary of the correlations between the The meta-analyses provide the pattern of
intermediate outcomes and employee correlations that one might expect theoretically –
behaviours not only because some of the constructs share some
• correlations between climate, other conceptual space (e.g. both job satisfaction and
intermediate outcomes and employee organisational commitment tap into employees’
behaviours. affective reactions to their jobs and organisations),
but also because there are proposed causal linkages
Correlations among between the intermediate outcomes.
intermediate outcomes
The meta-analyses tend to classify variables studied
Table 56 summarises the corrected correlations in relation to the dependant variables in one of
between the intermediate outcomes reported in three ways:
the meta-analyses. Where there was more than
one meta-analysis reporting the relationship • antecedents variables considered to contribute
between two intermediate outcomes, the range of the development of the dependant variable
correlation values from the meta-analyses is shown. • correlates variables for which there is not
No meta-analyses of engagement or psychological consensus about the causal ordering with the
contract met the inclusion criteria. dependant variable
• consequences variables considered to be
Table 56 shows moderate to high correlations influenced by the dependant variable.
between all of the intermediate outcome variables
where data are available. A key objective of the So, for example, Rhoades and Eisenberger208 in
brief for this study was to identify whether each their meta-analysis of organisational support,
intermediate variable identified offered something proposed antecedents included organisational
different. The associations between, for example, justice; consequences of organisational support
job satisfaction and organisational commitment included job satisfaction, organisational
(0.65), although strong, are not of sufficient commitment, job involvement, and turnover. In
1 2 3 4 5 6 7 8
1. Motivation
212
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
the Meeyr et al.669 meta-analysis of organisational give systematically higher scores than their less
commitment, proposed antecedents included optimistic counterparts.
organisational justice and organisational support;
proposed consequences included turnover Although there is uncertainty regarding the causal
intentions; and proposed correlates included job ordering among the intermediate outcomes, all are
satisfaction, job involvement and occupational proposed to be implicated in the development of
commitment. However, the meta-analyses were salient organisational behaviours (job performance,
mostly conducted on cross-sectional research. OCB and withdrawal behaviours). We next turn
Although, the meta-analyses provide the patterns to the correlations between the intermediate
of correlations that one might predict, the fact that outcomes and individual employee behaviours.
we are dealing with correlations leaves uncertain
the causal ordering of the observed associations. Correlations between
For example, organisational justice is predicted intermediate outcomes and
to be a source of attitudinal outcomes such as individual employee behaviours
job satisfaction, but it is likely that perceptions
of justice are themselves an outcome of attitudes Meta-analyses were found for relationships between
towards work as well as being an antecedent.670 all the intermediate outcomes and at least one
There is a strong likelihood of reciprocal causal employee behaviour. The most comprehensive
influences between many, if not all, of these data examined relationships with job satisfaction
constructs. and organisational commitment, which is perhaps
unsurprising given the long history and enduring
A further limitation of the studies should be noted. popularity of these concepts in the research
Research on intermediate outcomes has consisted literature. Table 57 summarises the correlations
almost entirely of survey research using self-report between the intermediate outcomes and
measures. Thus, many of the relationships may performance-related employee behaviours.
be inflated due to common method variance.
It has become widely accepted that correlations Generally, the meta-analyses reported here support
between variables measured with the same method significant, small to moderate relationships,
(e.g. self-report surveys) are possibly inflated due between intermediate outcomes and individual
the propensity of the subject to answer similarly employee behaviours (positive relationships
to multiple items. For example, more optimistic with job performance and OCB, and negative
or organisationally committed employees rating relationships with withdrawal behaviours).
a number of intermediate outcomes might However, relationships between absence and
TABLE 57 Corrected correlations between intermediate outcomes and employee behaviours
Organisational
Organisational
Occupational
commitment
commitment
(distributive)
(procedural)
involvement
satisfaction
Motivation
Turnover
support
Justice
Justice
intent
Job
Job
213
intermediate outcomes are consistently small, and performance might be less about a causal
questioning the posited link between absence and relationship and more to do with the influence of
attitudes. Equally, the correlations between job other factors on both variables.
involvement and employee behaviours are all weak.
While many of the correlations point towards links These caveats to the satisfaction–performance
between intermediate outcomes and employee relationship generally apply to the other
behaviours, it is clear that caution is required in intermediate outcome-employee behaviour
interpreting these results. Some conceptual and correlations in Table 57. The correlations mostly
methodological problems are illustrated when demonstrate significant associations between
we consider, for example, the meta-analytic employee psychological states and important
correlations between job satisfaction and job employee behaviours but there is unresolved causal
performance. ambiguity.
The relationship between job satisfaction and job One further point, given the intermediate
performance has been described as the ‘Holy Grail’ outcomes are interrelated and therefore each will
of occupational and organisational psychologists671 not have a unique relationship with the employee
and has generated a considerable research behaviours, the question arises as to the degree
effort. As Table 57 shows, the strength of the to which the above associations with employee
correlations from the three meta-analyses reported behaviours are independent from one another. As
in this review vary from small to moderate. The already discussed in this section, the association
most recent meta-analysis, by Judge et al.656 is between job satisfaction and performance may be
superior in terms of methodological quality to partly accounted for by the association between
any of the other meta-analyses included in this organisational support and performance, or
report (see Appendix 11 for quality reviews of they may be linked in a causal chain, such
the meta-analyses). Examining several hundred that job satisfaction plays some part in linking
studies, Judge et al.656 found an average corrected organisational support and job performance. This
correlation between job satisfaction and job means that it is not possible to arrive at conclusions
performance of 0.3. Although a correlation of about whether an intermediate outcome has a
0.3 qualifies as a ‘moderate’ effect size using unique relationship with employee behaviours.
Cohen’s668 rule of thumb (see Critical appraisal,
Interpretation of results, above), this result suggests Correlations between climate,
the relationship is a significant one and should be other intermediate outcomes
treated as important, given the multitude of factors and employee behaviours
that may determine job performance.
We next look at two meta-analyses of climate that
However, although the correlations might have attempted to model more than one step in the
encourage an interpretation that job satisfaction causal chain between intermediate outcomes and
causes performance, we cannot determine from employee behaviours, and have sought to provide
these mostly cross-sectional investigations whether a more complete model of employees’ reactions to
this is the case. It appears quite plausible that their work environment.
there is an effect of performance on satisfaction.
Job satisfaction may be ‘a response to previous Employees’ perceptions of virtually every aspect
goal-achievement, job effectiveness, and associated of their work environment have been included in
personal rewards’ (Warr:672 410). Associated climate research, including the organisation as a
personal rewards may be, for example, increased whole, senior management, job characteristics,
pay, recognition and positive feedback, autonomy supervisors and co-workers. The picture is
or skill use, with possible ramifications for levels of further confused by the overlap of climate
job satisfaction. dimensions with other intermediate outcomes,
such as organisational support, and indeed with
An additional interpretation of the satisfaction– HRM practices. For example, job characteristics
performance relationship is that additional perceptions may be included in a climate measure
variables cause both job satisfaction and and also classified as measure of job design.
performance.672 For example, organisational
support might increase job satisfaction while We report here the results of two climate meta-
also creating an environment for effective job analyses which have attempted to meaningfully
performance. The correlation between satisfaction classify the array of climate dimensions. The
214
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
climate meta-analyses by Carr et al.673 and Parker Each meta-analysis adapted the model slightly,
et al.674 in addition to reporting various one-to-one depending on the specific climate facets each
relationships between climate and intermediate adopted in order to categorise the extensive range
outcomes and/or employee behaviours, each of climate dimensions they studied.
tested a series of models proposing that climate
affects individual work outcomes through affective Carr et al.673 categorised their wide variety of
and cognitive states, such as job satisfaction and climate dimensions according to the three facets of
motivation. This procedure is useful for comparing climate proposed by Ostroff;675 affective, cognitive
alternative explanations for relationships, but and instrumental. Affective climate is concerned
should be used with caution. Correlations that are with interpersonal and social relationships
being compared come from different samples and between employees, and includes the dimensions
populations and are not comparable in the same of participation, cooperation, warmth and social
way as correlations from the same sample. rewards. Cognitive climate relates to individuals
and their involvement in work activities, and
Both meta-analyses drew upon the model of HRM, includes the dimensions of growth, innovation,
climate and productivity proposed by Kopelman et autonomy and intrinsic rewards. Instrumental
al.154 This model proposes that HRM practices in climate is concerned with task involvement
part shape organisational climate (comprising goal and getting things done in the organisation,
emphasis, means emphasis, reward orientation, and comprises the dimensions of achievement,
task support and socioemotional support). Climate hierarchy, structure and extrinsic rewards.675 See
represents employees’ perceptions of what the Table 58 for the correlations between the climate
organisation is like in terms of its policies, practices facets and intermediate outcomes and employee
and procedures. The climate in turn impacts behaviours reported by Carr et al.673
on important individual- and organisational-
level outcomes (e.g. job performance, OCBs and Carr et al.673 then tested a series of conceptual
withdrawal behaviours) through the impact it models based on Kopelman et al.’s154 work using
has on cognitive and affective states, such as job meta-analytic path analysis. Figure 5 shows the final
satisfaction, work motivation and organisational path model that best represents the data.
commitment. The model suggests that HRM does
not directly impact on employee psychological This model shows differential relationship of
states; rather, HR practices are perceptually affective, cognitive and instrumental climate
filtered (i.e. climate) by employees. Subsequent perceptions with job satisfaction and organisational
cognitive and affective states then mediate the commitment. Affective climate has a stronger
relationship between climate and individual- and relationship with organisational commitment than
organisational-level outcomes. Furthermore, do the cognitive and instrumental facets, while
Kopelman et al.154 propose that different climate affective and instrumental climate perceptions
dimensions are related to different cognitive have a stronger relationship with job satisfaction
and affective states, which, in turn, impact on than does cognitive climate. Job satisfaction and
different outcomes. For example, the impact on organisational commitment each have a unique,
withdrawal behaviours occurs primarily through negative relationship with withdrawal (turnover),
job satisfaction and motivation, whereas the impact while job satisfaction also has a unique, positive
on job performance occurs primarily through relationship with both job performance and well-
motivation alone.154 being. This indicates that the relationships between
TABLE 58 Correlations between dimensions of climate, intermediate outcomes and employee behaviours
215
Cognitive and
Climate affective states Outcomes
Job
Affective
performance
0.29 0.18
Job
0.23 satisfaction
0.12 0.22
Psychological
Cognitive
well being
–0.34
0.15
0.26 Organisational
commitment
0.10 –0.10
FIGURE 5 Path model relating climate to outcomes (Carr et al. 2003673). Figures in the diagram are regression weights that
demonstrate the strength of the relationship on a scale between 0 and 1.
the three climate facets and job performance, cooperation, pride and warmth; and organisational
well-being and withdrawal are indirect, occurring attributes include innovation, management
through their relationship with job satisfaction and, awareness and openness of information. Parker
with withdrawal only, through their relationship et al.674 refer to these facets collectively as
with organisational commitment. Instrumental psychological climate perceptions. See Table 59 for
climate is the only facet to have a direct the correlations between the climate facets and
relationship with an outcome variable, having intermediate outcomes and employee behaviours
a direct, negative relationship with withdrawal reported by Parker et al.674
behaviours.
Parker et al.674 developed a path model in
The model infers causal links; however, it is an attempt to identify causal links between
important to remember that it is based on organisational climate and behavioural outcomes
correlational data and causal interpretations should (Figure 6). Like Carr et al.,673 the authors then
not be assumed. The majority of measures in the tested a series of conceptual models based on
model were obtained via self-report methods, Kopelman et al.’s154 work using meta-analytic path
thereby potentially inflating the relationships analysis. Figure 6 shows the final path model that
reported as a result of common method variance. best represents the data. Parker et al.674 found
evidence that a general climate factor accounted
In contrast with Carr et al.,673 Parker et al.674 for the relationship of the five climate dimensions
classified their wide range of climate dimensions with work outcomes. This is in line with James and
according to the five facets of climate posited by James’212 argument that a common judgement
Jones and James676 concerning work environment process underlies diverse climate dimensions. This
perceptions, namely job, role, leadership, work process is hypothesised to represent an assessment
group and organisational characteristics. Job of whether the work environment is viewed as
characteristics include autonomy, challenge and personally beneficial or detrimental to one’s well-
importance; role characteristics include ambiguity, being.
conflict and overload; leader characteristics
comprise goal emphasis, support and upward This model shows an indirect effect of
influence; work group characteristics, include psychological climate perceptions on performance,
216
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
through their relationship with job satisfaction, In this chapter, we have so far considered the
work attitudes and motivation. The five facets relationships among the intermediate outcomes
of climate have a stronger relationship with and the relationships between the intermediate
job satisfaction than with job involvement outcomes and employee behaviours that may
and commitment, while job involvement and contribute towards organisational performance.
commitment have a stronger relationship with However, organisational performance is not merely
motivation than does job satisfaction. However, an aggregate of individual performance. In this
only job satisfaction also has a direct, positive section and the next we review the evidence for
relationship with performance. This means that the impact of intermediate outcomes, such as
the relationship between psychological climate job satisfaction, on health and non-health final
perceptions and performance is mediated by job outcomes (Objectives 4.1 and 4.2).
satisfaction, work attitudes and motivation, i.e.
that the effects of climate are indirect, occurring
through its relationship with job satisfaction Findings for Objective 4.2
and work attitudes and, in turn, through their
Longitudinal research on
relationship with motivation.
intermediate outcomes and
As with the results of Carr et al.,673 and the other patient outcomes
meta-analyses reported in this chapter, causal Research question
relationships are not proven, and many of the Do intermediate outcomes, such as job satisfaction,
correlations are potentially inflated. have an impact on the patient care outcomes of
Job
satisfaction
0.18 0.23
0.61
General
0.15 Motivation 0.06 Performance
climate
0.43 0.28
Work attitudes:
Job involvement
and commitment
FIGURE 6 Path model relating climate to outcomes (Parker et al. 2003).674 Figures in the diagram are regression weights that
demonstrate the strength of the relationship on a scale between 0 and 1.
217
mortality rates, patient waiting times, patient safety, assessing the impact of the one on the other679
patient satisfaction and patient length of stay? or by considering a patient care outcome not of
interest, for example hospital admissions680 or
Method adherence to therapy.681 A previous review has also
A systematic search of the literature was conducted noted the lack of longitudinal studies examining
to identify potential studies to answer this research the impact of nurses’ job satisfaction on key patient
question. The following databases were MEDLINE, outcomes.682 This lack of evidence has not been
EMBASE, CINAHL, SSCI, PsycINFO, King’s Fund addressed in the intervening years.
database and IBSS. In order to be included, studies
have to satisfy the following criteria. Summary
There is therefore no longitudinal evidence to
Inclusion criteria assess whether intermediate outcomes, such as job
• Longitudinal study design. satisfaction or burnout impacts on the patient care
• Intermediate outcome related to health outcomes of mortality rates, patient waiting times,
professionals (e.g. job satisfaction among patient safety, patient satisfaction and patient
nurses or doctors). length of stay.
• One of the following patient outcomes:
mortality, waiting times, length of stay, patient
satisfaction, patient safety. Findings for Objective 4.1
• The study must test the causal relationship
Impact of intermediate
between the intermediate outcome and the
patient care outcome. outcomes on non-health final
outcomes
Exclusion criteria To identify evidence for the impact of intermediate
Unpublished dissertations. outcomes on final outcomes in the non-health
sectors, longitudinal studies were searched for to
Results and discussion provide stronger evidence of a causal link between
The literature search retrieved the following the two. The procedure (see Conclusions, below)
numbers of unique studies for each of the was repeated here, the only difference being
intermediate outcomes (Table 60). that the variables termed as final outcomes are:
performance, productivity or customer outcomes.
The titles and abstracts of all the retrieved studies
were independently assessed for inclusion by two Details of studies
reviewers. No studies were identified that satisfied Six studies meet the inclusion criteria (Table 61).
the inclusion criteria. Some studies satisfied all but Five of these studies were conducted in the USA
one of the criteria, and so had to be excluded. For and one in the UK. Five of these studies have
example, non-longitudinal design, by performing a prospective longitudinal research design and
data collection of two different types of data at two one uses a retrospective analysis of a longitudinal
different times (two cross-sectional studies rather data set. The duration of the studies varies from
than a single longitudinal study),677,678 studying the 12 months to 8 years. Two studies use cross-
impact of an intervention on both the intermediate industry sample, with Schneider et al.683 using data
outcome and the patient outcome, but not from 35 companies and Patterson et al.36 reporting
218
TABLE 61 The impact of intermediate outcomes on non-health final outcomes
Schneider et Prospective Service climate No Customer Survey designed by the Reciprocal causation:
al. (1998)217 longitudinal perceptions of authors Analysis examined the relationship between service
USA 134 branches of service quality climate and service quality, both at two points in time
a bank The results indicated that climate for service led to
subsequent customer perceptions of service quality
but that there was also a reciprocal effect for customer
perceptions and climate
Bernhardt et Prospective Employee satisfaction No Customer satisfaction Surveys designed by the Using a time-series analysis of the data, the study finds
al. (2000)684 longitudinal Restaurant authors a positive relationship between customer satisfaction
USA 382 restaurants performance (sales, Company records and organisational performance
of the same customer counts, No longitudinal analysis was conducted for the
company profitability) relationship between employee satisfaction and
customer satisfaction, though the cross-sectional
continued
Health Technology Assessment 2010; Vol. 14: No. 51
219
220
TABLE 61 The impact of intermediate outcomes on non-health final outcomes (continued)
EPS, earnings per share; OCB, Organisational Citizenship Behaviour; ROA, return on assets.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
data from 42 manufacturing firms. Four studies performance in cyclical and found stronger
investigate the same organisation but sample evidence of reverse causality, i.e. employee attitudes
across various subunits of these organisations. The are determined by organisational performance.
study by Bernhardt et al.684 uses data from 382 The analysis revealed that overall job satisfaction
restaurants; Koys685 reports data from 28 stores of and satisfaction with security were predicted by
the same organisation; Schneider et al.217 collected return on assets (ROA) and earnings per share
data from 134 branches of a bank; and Morrow and (EPS) more strongly than the reverse (although
McElroy686 investigated 31 subunits of a mortgage some of the reverse relationships were also
services company. significant); satisfaction with pay suggested a more
reciprocal relationship with ROA and EPS.683
Intermediate outcome variables studied
Four of the six studies have used employee A fourth study684 uses a longitudinal analysis
satisfaction as a predictor variable; other predictor only between the two final outcomes of customer
variables used by these studies are organisational satisfaction and restaurant performance; it reports
climate, service climate, organisational the link between job satisfaction and customer
commitment and employee turnover. satisfaction using a cross-sectional analysis.
prospective longitudinal research design and one we cannot determine whether one intermediate
study uses a retrospective analysis of a longitudinal outcome (e.g. high job satisfaction) promoted the
data set. Four studies have collected data from employee behaviour (e.g. good work performance),
various subunits of the same organisation and whether behaviours enhanced the intermediate
two have used cross-organisation data. Four of outcomes, whether both effects occurred, or
the six studies examine the relationship between whether third variables were causally more
job satisfaction and final outcomes; two studies important. Nevertheless, the data do point towards
explored the link between climate and final the importance of the intermediate outcomes for
outcomes, one looking at general organisational salient employee behaviours.
climate, the other at climate for service; the other
intermediate outcomes in this review have not been There were few meta-analytic data on motivation
subjected to longitudinal investigation in relation and none on the psychological contract. Both
to final outcomes. Given the few longitudinal constructs have considerable prominence in the
studies linking any given intermediate outcome HRM and organisational behaviour. In the case of
with a final outcome we cannot arrive at any clear the psychological contract this absence may reflect
conclusion about the causal connection between the concepts relative newness, but also because
these intermediate and final outcomes. it is an ‘umbrella construct capturing a range of
variables associated with trust, commitment and
the exchange of expectations and obligations’
Conclusions (Sparrow:687 15) and, consequently, it has been
operationalised in many different ways. However,
This chapter has covered a wide range of there is little doubt that the psychological contract
research, reviewing intercorrelations between the is seen by many commentators as a useful vehicle to
intermediate outcomes (Objective 3.1), and the help understand and predict employee behaviour.
impact of intermediate outcomes on final outcomes
in both the non-health literature (Objective 4.1) There were very few longitudinal studies to bring to
and the health literature (Objective 4.2). bear on Objective 4.1. In non-health settings there
was some limited evidence about organisations’
Many, but not all of, the intermediate outcomes average job satisfaction, climate and company
were covered by meta-analyses. While there performance, but very few studies have examined
remains debate over the value of measuring each of both intermediate and final outcomes on two or
the intermediate outcomes as a separate domain of more occasions.652 Schneider et al.683 did examine
assessment in relation to performance, the strength data at multiple time points, albeit on a very small
of the correlations does not indicate construct sample ranging from only 12 to 35 companies.
redundancy. The significant intercorrelations Their finding of reverse causality between
stem not only from some conceptual overlap and average satisfaction and financial performance
mostly likely common method variance, but also emphasises the need for caution in making causal
because there are causal linkages between the interpretations, and also the need for further, more
constructs. For example, employee perceptions of robust research.
the organisation, as embodied in such constructs
as climate, organisational support and justice, There were no relevant longitudinal studies
are generally proposed to shape motivation and examining intermediate outcomes to patient
employee attitudes, such as commitment and outcomes. It is therefore worth noting that there
satisfaction. However, the causal linkages between are cross-sectional studies on staff intermediate
these domains remain unclear. outcomes and patient outcomes. For example,
Robertson et al.688 found that nurse job satisfaction
Additional analyses reviewing the meta-analytic averaged at the ward level was related to observed
correlations between the intermediate outcomes standards of care delivery. Leiter et al.689 found
of HRM and salient employee behaviours that patients treated on wards where nursing staff
were mostly consistent in showing significant averaged greater job-related emotional exhaustion
associations, although none was large. However, (a dimension of burnout) were less satisfied with
to date most research has merely asked about the their nurses, doctors and average quality of care.
existence or otherwise of a significant association These investigations are clearly worth pursuing
between intermediate outcomes and employee with more substantial longitudinal data sets.
behaviours. From this correlational data alone
222
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Chapter 10
Measuring the intermediate outcomes
of HRM: the reliability of measures
Introduction general or an aspect of it) in the same way as the
other items in the scale. The most appropriate
This chapter focuses on objective two of the statistic to assess a scale’s reliability is known as
research. Specifically, it is concerned with Cronbach’s alpha and is expressed as a value
answering the following question: What is the between 0 and 1. Cronbach’s alpha is the average
evidence on the reliability of measures of the correlation of each item in a scale with all of the
intermediate outcomes of HRM practices covered other items in that scale. For a scale to have good
by this review? reliability (good internal consistency), as a rule of
thumb it should have a Cronbach’s alpha of at least
The chapter first considers the ways in which the 0.7. We have used this measure as the best indicator
properties of a scale or measure are evaluated. of scale reliability.
The process of selecting measures of intermediate
outcomes for the review is described and the Test–retest reliability
evidence used in the review is identified. Finally, Test–retest reliability concerns the extent to which
evidence about the reliability of selected measures a scale’s measurements are consistent over time.
is presented. It is not always appropriate but can be helpful in
assessing how well a measure performs. Hence
in situations where the work environment (for
What is measurement example) is expected to be stable, test–retest
reliability? reliability of a measure of work environment
(the correlation between the time 1 and time 2
This review is concerned with self-report measures measures) should be > 0.7. Equally, in situations
or scales used in survey instruments. Reliability where the work environment is expected to change
is an important aspect of the quality of such (e.g. in response to HRM practices) test–retest
measures. A scale’s reliability refers to how reliability can be low. Test–retest reliability is less
consistently the scale measures a psychological frequently reported in evaluation articles and is of
feature or work characteristic. Psychometrics is the less interest to the current review.
statistical process concerned with developing or
assessing the quality of measurement scales and Inter-rater reliability
scale reliability is one of the properties by which a A third form of reliability concerns the stability or
scale’s quality is rated psychometric. consistency of a measure when used by different
raters given the same experiences (inter-rater
Methods of measuring reliability reliability). Inter-rater reliability is important
when results are aggregated and analysed at
The focus of this part of the review is the reported the organisational level (such as using the mean
reliabilities of scales that measure the intermediate job satisfaction score across all employees in an
outcomes of HRM practices. Two basic forms of organisation to represent organisational level job
reliability exist for self-report instruments: internal satisfaction). It is usually measured by intraclass
consistency and test–retest reliability. correlation (ICC) of which there are two forms.
ICC1 measures the extent to which individuals in a
Internal consistency group agree (the basis for justifying aggregation to
The internal consistency of a scale indicates the the mean is that there is good general agreement
extent to which all of the items in the scale are within a group, so a mean score is not hiding
broadly measuring the same phenomenon as each high levels of variation between individuals).
other. For example, in a job satisfaction measure, ICC2 assesses the reliability of the mean score
the internal consistency would demonstrate the (approximately this measures the extent to
extent to which each item or question in the scale which the mean scores on a scale for two random
is actually measuring job satisfaction (either in subsamples of the same group correlate). While of 223
interest to the current review, inter-rater reliability performance chain. The intermediate outcomes
in the form of ICCs is only relevant to studies using included here were first identified in Chapter 4 as
organisational level aggregated data, which, as this having the most potential in explaining the HRM–
review has established, are few in number, so it performance link. Evidence on the causal links
has not been possible to report ICCSs for the data between HRM practices and these intermediate
presented in this chapter. outcomes was then reviewed in Chapters 5–7.
In the context of this review, those intermediate
Challenges in reviewing outcomes that have been demonstrated to be
reliability sensitive to HRM practices are deemed ‘valid’ to
measure:
In order to review the reliability of a scale, it is
ideal to identify several examples of scale use • motivation
and compare reliability scores from a number of • job satisfaction
different samples. Previous reviews of the reliability • organisational commitment
of measures have highlighted a number of • occupational/professional commitment
challenges inherent in extracting these data from • engagement
the social science research literature:49 • burnout
• job involvement
• inconsistent use of measures although a well- • turnover intentions
established measure is cited in the article, the • psychological contract
authors may have adapted or deleted items • organisational justice
from the scale or altered response categories • organisational support
(sometimes without stating this in the article) • climate.
• inaccurate reporting of reliability data reliability
statistics are given for the original or some Process for selecting specific
historical sample rather than the current intermediate outcome measures
sample in the article for review
• inconsistent reporting of reliability data in
articles although scales are used, their The aim is to identify key measures for each
reliabilities are not reported for the sample. intermediate outcome domain included in the
review. The strategy used is a four-step process:
These factors taken together mean that although
there might, on the surface, be a large number • First, for each of the above 12 intermediate
of articles that report using a specific measure, outcomes the specific measures used in the
there may in practice only be a limited amount of studies were identified. Where an intermediate
information available on the measure’s reliability. outcome had a single scale associated with it,
or if there were any scales predominantly used
A further complication for this review is the extent in the health literature, these would be selected
to which one-off scales or single items have been for the review of reliability. This exercise
used to measure concepts of interest. Where a scale revealed that for a number of intermediate
appears only once in the literature its reliability outcomes no single scale emerged as the main
is of relatively limited use in a review such as this. or preferred measure. In some instances, the
Single items cannot have their reliability assessed measure used was a one-off, designed by the
by the methods described here (as reliability is researchers or was simply not specified at all.
based on the relationships between items). For In other cases a number of specific measures
these reasons the review developed a pragmatic were each used by one study or a small number
strategy for identifying relevant evidence, which is of them. In these cases, expert opinion was
described below (see Conclusions). used to select the best-established measures.
• Second, where an intermediate outcome area
was not covered by the studies included in
Which intermediate this review (as is the case with psychological
outcomes is it valid to contract) or the measure identified in this
review was felt to be too specific for general
measure in this context? use (as with climate for creativity and
In line with the project brief, this review uses innovation) the subject experts on the research
validity to refer to the relevance of measuring team identified an appropriate measure for
224 certain intermediate outcomes in the HRM inclusion.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
• Third, for each of the included measures, a Established scales used in the review:
recent review of reliability data was sought. One
evidence source was invaluable in this: Taking
the Measure of Work – a guide to validated scales Warr et al. (1979):475 Intrinsic Job Motivation
for organisational research and diagnosis201
provides a comprehensive review, drawing on
all papers using validated measures published
in 15 top organisational research journals over
the period 1990–9. Scale description
• Fourth, for the intermediate outcomes not This is a six-item scale measuring intrinsic job
covered by Fields’ review201 a search was motivation, i.e. the degree to which an employee is
undertaken for reviews of reliability for scales motivated to perform well in his/her current job in
measuring the intermediate outcome. If no order to be intrinsically satisfied. The scale is taken
review of scale reliability could be identified from Warr et al.’s475 Work and Life Attitudes Survey.
then purposive searching was undertaken
for a meta-analysis or longitudinal study to Items and response category
supplement the data presented here. Meta- • ‘I feel a sense of personal satisfaction when I do
analyses were also used where conceptual this job well.’
uncertainty existed around a measure. The • ‘My opinion of myself goes down when I do
aim of this search was to identify reliability this job badly.’
data for the selected measures and to • ‘I take pride in doing my job as well as I can.’
highlight if there were other well-established • ‘I feel unhappy when my work is not up to my
measures for any of the outcomes that had not usual standard.’
already been covered in the review. In some • ‘I like to look back on the day’s work with a
instances intermediate outcomes had only sense of a job well done.’
been studied in cross-sectional research, so • ‘I try to think of ways of doing my job
no example scales were covered in this review. effectively.’
However, these intermediate outcomes had
been identified by the expert panel as being Each item is rated on seven-point Likert-type scale,
theoretically important and were thought where 1 = no, I strongly disagree, 2 = no, I disagree
to have demonstrable links with important quite a lot, 3 = no, I disagree just a little, 4 = I’m
employee behaviours such as job performance not sure about this, 5 = yes, I agree just a little,
and staff turnover. 6 = yes, I agree quite a lot, and 7 = yes, I strongly
agree.
The remainder of the chapter presents evidence
on scale reliabilities for measures of intermediate Reliability
outcomes of HRM practices identified earlier in A coefficient alpha value of 0.65 for test–
the report. For each intermediate outcome, the use retest reliability, based on 60 participants, was
of scales in the review is summarised and the main reported.475
scale used to measure that outcome is descibed.
Items and response categories are given and One review of measures of motivation was
reliability data reported. identified.690 This does not include the Warr et al.475
measure, nor does it provide scale reliabilities for
the measures it identifies. Eleven measures of work
Reliability of intermediate motivation are listed (including the JDS).
outcome measures
Job satisfaction
Motivation
Five studies reported using measures of motivation Job satisfaction was the most widely measured
as an outcome of HRM practices. Three studies intermediate outcome in the review. Over 50
each used a different established scale, one study studies reported data using 17 different measures.
used the researchers own scale and one study did Measures differ in the sense that they either
not specify the measure of motivation. Of the three measure overall satisfaction (e.g. with items such as
established measures, one was developed in the ‘in general, I like working here’) or they measure
UK475 and is presented here. A second measure, satisfaction with individual facets of a job (e.g.
which includes a subscale of work motivation (the pay, autonomy, supervision), which can then be
JDS), is reported below (see Job satisfaction). combined to give an overall satisfaction score. 225
The measures commonly used in studies in this • ‘My own feelings generally are not affected
review all measure satisfaction with facets of the much one way or the other by how well I do on
job to some degree. Additionally, some measures this job.’
distinguish between intrinsic satisfaction (job • ‘Most people on this job feel a great sense of
content) and extrinsic satisfaction (rate of pay and personal satisfaction when they do the job well.’
security).475 • ‘Most people on this job feel bad or unhappy
when they find that they have performed the
Four measures emerge as the most commonly used work poorly.’
in this type of research:
Responses for the items in the general and work
motivation subscales are scored on a seven-point
Likert scale: 1 = disagree strongly, 2 = disagree,
Hackman and Oldham (1974):195 Job Diagnostic 3 = disagree slightly, 4 = neutral, 5 = agree slightly,
Survey (JDS)
6 = agree, and 7 = agree strongly.
Smith et al. (1969):243 Job Descriptive Index (JDI)
Weiss et al. (1967):473 Minnesota Satisfaction Satisfaction with opportunities for
Questionnaire (MSQ) growth
Warr et al. (1979):475 Job Satisfaction Questionnaire • ‘The amount of personal growth and
development I get in doing my job.’
• ‘The feeling of worthwhile accomplishment I
get from doing my job.’
The most commonly used measure was JDS. The • ‘The amount of independent thought and
overall and facet-specific job satisfaction scales of action I can exercise in my job.’
the JDS form part of a larger questionnaire that • ‘The amount of challenge in my job.’
was first developed in the early 1970s. The full
JDS measures a range of different aspects of work Facet-specific items
including motivation, job characteristics. The job • Security
satisfaction and work motivation subscales are –– ‘the amount of job security I have’
reported here. –– ‘how secure things look for me in the
future in this organisation’.
Scale description • Compensation
This is a 25-item measure consisting of four –– ‘the amount of pay and fringe benefits I
subscales: overall satisfaction; internal work receive’
motivation; growth satisfaction; and satisfaction –– ‘the degree to which I am fairly paid
with specific facets of the job. Details of the for the amount I contribute to this
measure are taken from Fields201 ‘Taking the Measure organisation’.
of Work’. • Satisfaction with co-workers
–– ‘the people I talk to and work with on my
Items and response categories job’
Overall satisfaction –– ‘the chance to get to know other people
• ‘Generally speaking I am very satisfied with while on the job’
this job.’ –– ‘the chance to help other people while at
• ‘I am generally satisfied with the kind of work I work’.
do in this job.’ • Satisfaction with supervisor
• ‘I frequently think of quitting this job.’ –– ‘the degree of respect and fair treatment I
• ‘Most people in this job are very satisfied with receive from my boss’
the job.’ –– ‘the amount of support and guidance I
• ‘People on this job often think of quitting.’ receive from my supervisor’
–– ‘the overall quality of the supervision I
Internal work motivation receive in my work’.
• ‘My opinion of myself goes up when I do this
job well.’ The response categories for growth satisfaction
• ‘I feel a great sense of personal satisfaction and facet-specific satisfaction are: 1 = extremely
when I do this job well.’ dissatisfied, 2 = dissatisfied, 3 = slightly dissatisfied,
• ‘I feel bad and unhappy when I discover that I 4 = neutral, 5 = slightly satisfied, 6 = satisfied, and
have performed poorly on this job.’ 7 = extremely satisfied.
226
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Occupational/professional
Mowday et al. (1982):173 Shortened OCQ commitment
Meyer and Allen (1984):477 Affective Commitment One measure of professional or career commitment
was identified in the papers included in the review:
227
• ‘I am disappointed that I ever entered the ––‘When I get up in the morning, I feel like
nursing profession.’* going to work.’
• ‘I spend a significant amount of personal time • Dedication
reading nursing-related journals or books.’ –– ‘I am enthusiastic about my job.’
–– ‘My job inspires me.’
(*Indicates that the item is reverse scored.) –– ‘I am proud of the work that I do.’
• Absorption
The scale uses a five-point response category, which –– ‘I feel happy when I am working intensely.’
Blau691 describes as ranging from 1 = strongly –– ‘I am immersed in my work.’
disagree to 5 = strongly agree. –– ‘I get carried away when I am working.’
Example items
• Vigour
Maslach and Jackson (1981, 1984):693,694 Maslach
–– ‘At my work, I feel bursting with energy.’
Burnout Inventory (MBI)
–– ‘At my job, I feel strong and vigorous.’
228
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
A 22-item version of the MBI has been developed Other measures of burnout
and found to be reliabile.695,696 The 22-item scale is Research into burnout has largely been dominated
reported here. by use of the MBI. One review of measures of
burnout was identified697 in which two other scales
Items and response category are reviewed.
• Exhaustion subscale
–– ‘I feel emotionally drained from my work.’ Job involvement
–– ‘I feel fatigued when I have to get up in the
morning to face another day in the job.’ Four studies include a job involvement measure
–– ‘Working with people all day is really a as an outcome of HRM. Two studies reported
strain for me.’ using Lodahl and Kejner424 and one study used a
–– ‘I feel ‘burned out’ from my work.’ measure by Kanungo.593 The fourth study did not
–– ‘I feel frustrated by my job.’ specify the scale used.
–– ‘I feel I’m working too hard on my job.’
–– ‘Working directly with people puts too No reviews of the reliability of job involvement
much stress on me.’ measures were identified. A meta-analysis of job
–– ‘I feel like I’m at the end of my rope.’ involvement189 established that the two scales
–– ‘I feel used up at the end of the day.’ cited here tended to measure the same concept
• Personal accomplishment and that there were no substantive differences in
–– ‘I can easily understand how my clients feel relationships with other associated variables (e.g.
about things.’ job satisfaction). The more recent Kanungo593 scale
–– ‘I deal very effectively with the problems of is reported here.
my clients.’
–– ‘I feel I am a positive influence on other
peoples lives through my work.’ Kanungo (1982):593 Job Involvement Questionnaire
–– ‘I feel very energetic.’ (JIQ)
–– ‘I can easily create a relaxed atmosphere.’
–– ‘I feel exhilarated after working with my
clients.’ Scale description
–– ‘I have accomplished many worthwhile This is a 10-item scale to describe an employee’s
things in this job.’ specific beliefs regarding his/her relationship with,
–– ‘In my work I deal with emotional or involvement in, his/her current job.
problems calmly.’
• Depersonalisation Example items
–– ‘I feel I treat some clients as impersonal • ‘The most important things that happen to me
“objects”.’ involve my present job.’
229
• ‘To me, my job is only a small part of who I • ‘I will probably look for a new job in the next
am.’ year.’
• ‘I am very much involved personally with my
job.’ Response category
• ‘I live, eat and breathe my job.’ Item 1 is rated on a seven-point Likert-type scale,
• ‘Most of my interests are centred around my where 1 = not at all likely, 2 = somewhat likely,
job.’ 3 = quite likely, and 7 = extremely likely.
• ‘I have very strong ties with my present job
which would be very difficult to break.’ Items 2 and 3 are rated on a seven-point
• ‘Usually I feel detached from my job.’ Likert-type scale, where 1 = strongly disagree,
• ‘Most of my personal life goals are job- 2 = disagree, 3 = slightly disagree, 4 = neither agree
oriented.’ nor disagree, 5 = slightly agree, 6 = agree, and
• ‘I consider my job to be very central to my 7 = strongly agree.
existence.’
• ‘I like to be absorbed in my job most of the Reliability
time.’ A coefficient alpha value of 0.83 is given based on a
sample of more than 400 participants.619
Response category
Each item is rated on six-point ‘agree–disagree’ Psychological contract
Likert-type scale.
No studies measuring psychological contract were
Reliability found in the review of longitudinal studies. Equally,
Coefficient alpha values for the JIQ, as calculated no review of psychological contract scales was
by Kanungo,593 were 0.87 for internal consistency identified in the literature. The details of the scale
(based on 703 participants) and 0.85 for test–retest reported here are drawn from the Psychological
reliability (based on 63 participants). Contract Inventory Technical Report.698
Turnover intentions
Rousseau (2000):698 Psychological Contract
Over 50 of the studies in this review included some Inventory (PCI)
form of absence, turnover or turnover intention
measure. In 36 cases, a one-off or unspecified scale,
or company records were used. The remaining
studies cited 10 different scales, the most frequently Scale description
cited being the Intention to Quit subscale of the This scale measures four dimensions of the
Michigan Organisational Assessment Questionnaire psychological contract; relational (long-term or
(used in three studies). open-ended employment arrangements, based on
mutual trust and loyalty), balanced (dynamic, open-
No reviews of the reliability of turnover intention ended employment arrangements), transactional
scales were identified. (short-term employment arrangements, based
primarily on economic exchange), and transitional
(a cognitive state that reflects the consequences
Cammann et al. (1979):619 Intention to Turn Over of organisational change). All four dimensions
are considered from the perspective of both the
employee and the employer.
Scale description
This is a three-item subscale of the Michigan The dimensions of relational and transactional
Organizational Assessment Questionnaire, psychological contract are generally considered to
measuring an employee’s intentions to leave his/ underpin the concept and, therefore, only these
her job. two scales are reported below.
the general manager treats me with respect part. There is a shorter 17-item version of the
and dignity.’ scale, which is presented here. It consists of eight
–– ‘When decisions are made about my job, items about how employees perceive themselves
the general manager is sensitive to my to be valued by the organisation and nine items
personal needs.’ about the extent to which the organisation
–– ‘When decisions are made about my job, considers their well-being in the decisions that
the general manager deals with me in a it makes. Shorter versions have been used, but
truthful manner.’ the scale authors recommend that both facets of
–– ‘When decisions are made about my job, perceived organisational support are used in its
the general manager shows concern for my measurement.208
rights as an employee.’
–– ‘Concerning decisions about my job, the Items and response category
general manager discusses the implications ‘Listed below is a series of statements that represent
of the decisions with me.’ possible feelings that individuals might have
–– ‘The general manager offers adequate about the company or organisation for which they
justification for decisions made about my work. With respect to your own feelings about the
job.’ particular organisation with which you are now
–– ‘When making decisions about my job, the working [name of organisation] please indicate
general manager offers explanations that the degree of your agreement or disagreement
make sense to me.’ with each statement by checking one of seven
–– ‘My general manager explains very clearly alternatives below each statement.’
any decision made about my job.’
• ‘The organisation values my contribution to its
Responses are on a seven-point scale ranging from well-being.’
1 = strongly disagree to 7 = strongly agree. • ‘If the organisation could hire someone at a
lower salary to replace me it would do so.’*
Reliability • ‘The organisation fails to appreciate any extra
Fields201 reports Cronbach’s alpha coefficients for effort from me.’*
the scales as follows: distributive justice 0.72–0.74; • ‘The organisation strongly considers my goals
procedural justice 0.85 and interactive justice 0.92. and values.’
• ‘The organisation would ignore any complaint
Other measures of organisational justice from me.’
Fields201 identifies 18 additional validated measures • ‘The organisation disregards my bests interests
of organisational justice. The majority focus either when it makes decisions that affect me.’
on procedural justice or distributive justice. • ‘Help is available from the organisation when I
have a problem.’
Organisational support • ‘The organisation really cares about my well-
being.’
One measure of perceived organisational support • ‘The organisation is willing to extend itself in
was reported in the studies included in the review order to help me perform my job to the best of
and is reported here. my ability.’
• ‘Even if I did the best job possible the
organisation would fail to notice.’
Eisenberger et al. (1986):206 Survey of Perceived • ‘The organisation is willing to help me when I
Organistional Support need a special favour.’
• ‘The organisation cares about my general
satisfaction at work.’
Scale description • ‘If given the opportunity, the organisation
Perceived organisational support is a would take advantage of me.’*
unidimensional scale with two facets, valuation • ‘The organisation shows very little concern for
of employee contribution and concern about me.’*
employee well-being.208 The Survey of Perceived • ‘The organisation cares about my opinions.’
Organisational Support measures the employee’s • ‘The organisation takes pride in my
view of the extent to which his/her organisation accomplishments at work.’
values them, is concerned with their well-being • ‘The organisation tries to make my job as
and chooses to reward greater effort on their interesting as possible.’
232
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
(*Indicates that an item is reverse scored.) was felt to be too specific to be of general use so is
not included here.
Responses are scored on a seven-point scale,
running from 1 = strongly disagree to 7 = strongly The literature review identified two meta-analytic
agree. reviews of molar climate measures;673,674 two recent
studies of climate in health-care settings;701,702 and
Reliability one review of safety climate in health care.703 Both
Rhoades and Eisenberger208 describe the scale as of the studies in health-care settings support the
unidimensional and having high internal reliability. importance of safety climate in determining patient
This is confirmed in their review of more than 70 safety outcomes (including medication errors).
studies, most of which use the Survey of Perceived
Organisational Support. Fields, in his review of The review of Flin et al.703 examines the
the measure, reports coefficient alpha values from psychometric properties of 12 measures of safety
three studies, ranging from 0.74 to 0.95. climate and concludes that the Hospital Survey
on Patient Safety has undergone more rigorous
Other measures of organisational testing than other measures and met more of their
support psychometric criteria. The scale is reported here.
No other measures of organisational support were
identified.
Sorra and Nieva (2003):704 Hospital Survey on
Organisational climate Patient Safety
234
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
The same pattern was found in the reviews of (e.g. engagement). This is not a reflection on their
validated measures. The large numbers of scales theoretical importance.
identified in areas such as job satisfaction and
organisational justice are a reflection of both The aim here was to provide reliability information
the popularity of such concepts in the research on scales demonstrated as having causal links in
literature and their multidimensionality – often the HRM performance chain. Overall, validated
scales are developed to measure specific facets scales with published reliability data were found
of an intermediate outcome (e.g. procedural, for 10 of the 12 intermediate outcomes, the only
distributive or interactional justice). The lack of exceptions being psychological contract and
scales of other areas in longitudinal HRM studies engagement. This is perhaps a reflection of the
included in this review indicates either the relative relatively recent recognition of the potential offered
youth of the concepts in HRM research terms by these outcomes for explaining links in the HRM
(e.g. psychological contract) or the popularity of performance chain. The subject experts on the
cross-sectional as opposed to longitudinal research team identified promising scales for these areas.
235
Chapter 11
Conclusions and recommendations
Introduction such as those in many areas of medicine where
the methodology was first developed. Specifying
Identifying the links between HRM and individual and defining the interventions and outcomes
and organisational performance is a critical to be studied was particularly complicated, and
issue for policy-makers and practitioners. Since consequently developing inclusion and exclusion
the earliest studies in the USA by researchers criteria and a robust means of synthesis was not
such as Arthur14,648 and Huselid,18 many a straightforward matter. As a result, this report
commentators have concluded that the evidence presents an assessment of the evidence in relation
of the performance benefits of effective HRM to a wide range of HRM practices, rather than
is increasingly strong. As Marchington and an in-depth appraisal and synthesis of findings
Zagelmeyer705 note, a number of reports in the involving a few well-defined practices. There are
UK have stated that the economic and business two methodological principles that have guided the
case for good people management is now proven. review:
For example, in the UK, the CIPD706 reported
that ‘more than 30 studies carried out in the • The review is, in the main, concerned with
UK and the USA since the early 1990s leave no cause and effect and, therefore, mostly restricts
room to doubt that there is a correlation between itself to the type of evidence capable of
business management and performance, that the providing causal explanations.
relationship is positive, and that it is cumulative’ • The review adopts a broad and inclusive
(CIPD 2001:706 4). However, there is not unanimity approach to the assessment of evidence in
about the strength of the evidence and judgements order to reflect the complex and diverse nature
vary as to the extent to which the case is proven. of the HRM literature.
More considered reviews of the evidence suggest
that it is too early to be conclusive about the effects Given the multiple research questions this review
of HR practices on performance.30 Hence the sought to answer, along with the breadth of some
current review is a timely assessment of the nature of the research questions and the time limitations
and strength of the evidence linking HRM to on the project, we did not attempt a coverage of all
performance. literature relevant to HRM, such as that on culture,
strategy and leadership, but these concepts are
included where they are directly tied to the HRM
Systematic review practices reviewed (e.g. leadership training).
methodology
Previous reviews on the evidence base for the Identification of HRM
HRM performance link have been conducted practices
with varying degrees of adherence to systematic
principles, but none has followed the protocols of An immediate challenge for the research team was
systematic review methodology. There is no real that of marshalling the diverse/diffuse terminology
tradition of using this approach in the HRM field within HRM into categories that could be used
and their use more generally in organisational consistently throughout the review. Practices were
behaviour and management areas has only recently identified using three different methods:
emerged. Such reviews are still limited in number.
• a review of reviews
Systematic review methodology offers the most • expert opinion
sophisticated means of identifying and synthesising • all practices identified from the subsequent
large amounts of evidence from disparate literature searches were then mapped back
sources, as is the case in the HRM literature. on to the categories created by the first
The HRM topic itself, however, presented two methods, to test their coverage and
several challenges often not encountered in inclusiveness.
more limited and clearly demarcated reviews, 237
Two researchers independently undertook classification (SIC) codes that allows further
reciprocal translation on the HRM practice breakdown of the data by sector. It is recommended
terms in order to test the equivalence of themes that further analysis of this database is undertaken
across different research papers. This quickly to provide more detailed information about the use
demonstrated the diversity of approaches and of HRM in the health sector that this permits.
definitions within the HRM literature, and led us
to adopt a search strategy based on broader terms The categories of HRM practices, work design
such as HRM and personnel management. and staffing, training, employee involvement and
performance management, which we have found to
On the one hand, the implications for this review be most consistently associated with positive patient
are that by adopting a broad search strategy, the outcomes in the health sector have also been shown
evidence collected is an accurate reflection of the by the data in WERS to have good coverage in
diversity of research within HRM, but is more both health and non-health sectors. However, the
difficult to distil into a few key messages. On information on them is not as deep as one would
the other hand, as in any review, some relevant ideally like, as we do not have strong information
research studies may not have been identified, on the variety of forms that each HRM practice
particularly in this case if they did not include may take. Such variations may lead to differences in
HRM, personnel management, or an equivalent their effects and durability.
term in the title or abstract and were not identified
via citation searching or reference tracking. Additionally, there is a need for awareness
about the potential gap between intended HRM
practices and actual or implemented practices.
Range and fidelity of use of Social policy research has shown that the extent
HRM practices in the UK of implementation fidelity may affect whether
a practice has its desired impact. However, no
Objective 1 for the research is concerned with generic measure of implementation fidelity exists
identifying the main methods of HRM practices, and there is little consistency in its measurement in
the extent of their use in UK organisations, and the social policy area. Within HRM, some studies
the degree to which HRM policy is implemented. do gauge whether policies and practices have
Initial work to identify the main methods of been implemented as they should be (e.g. reduced
HRM identified 10 broad categories of practice/ hours working, and monitoring and feedback),
intervention. These were then searched for in high- but, on the whole, the fidelity with which HRM
quality HRM surveys in the UK. Several surveys practices are implemented is rarely measured in
were identified that examined the range of HRM the research. However, only by evaluating whether
practices being used in organisations in the UK. Of a practice has been implemented properly can we
these, the WERS in particular is an excellent source assess whether any lack of impact is due to poor
of data on the extent of use of practices, providing implementation or the inadequacies of the practice
some data on each of the broad categories of HRM itself. Without assessing fidelity, it is also unclear
practice. The majority of categories of practice whether any positive outcomes produced could
identified in this review are also reported in surveys be improved still further by increasing fidelity to
as present in over 70% of UK workplaces. However, the original model. Implementation fidelity has
within an individual workplace the extent of their then to be considered as a potential moderator
use may vary as they may not be used for all staff. of outcomes, i.e. a variable that can influence the
relationship between an HRM practice and its
Not all specific practices in each broad category outcome.
used in this review are measured. For example,
a survey might measure some aspects of work Consequently, implementation fidelity needs to
design, for example job autonomy, but not others, be measured and this report offers a framework
for example role breadth. Hence, the range of and guidance for understanding and evaluating
approaches to HRM practices within each broad the fidelity in practice. A major implication for
category of HRM activity is unknown. Separate the findings of this review is that it is unclear
information for the health sector is sparse and whether variations in outcome or inconclusive
points to the need for better monitoring of HRM findings are due to methodological weaknesses
practices used in the NHS and elsewhere in the in which researchers have failed to distinguish
health sector. WERS uses standard industrial between HRM policies as intended and HRM
238
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
practice as enacted. The recommendation here is commitment. Firm conclusions cannot be drawn
for better methodological rigour in the assessment from this research base.
of implementation fidelity within HRM and
performance research. Some constructs have emerged more recently (e.g.
engagement and psychological contract) and,
therefore, while theoretically important, may not
Evidence for the link yet have been the subject of longitudinal research.
between HRM practices and The research in health settings on burnout and job
satisfaction is perhaps unsurprising, as burnout has
intermediate outcomes in long been proposed as a potential adverse reaction
the health sector to care work, and job satisfaction is the most widely
researched intermediate outcome in organisational
Intermediate outcomes are the intended outcomes psychology and organisational behaviour research,
of HRM practices, such as attitudes that may act which chimes with how the layperson would first
as a link between HRM and final outcomes such as approach the assessment of work.
patient care. Longitudinal studies examining the
relationship between HRM and such outcomes in Overall, very limited evidence was found and
the health field were identified by two strategies. the main theoretical links have not been tested
First, a broad search for HRM and any outcome, longitudinally in the health sector. Considerably
with no limitations set in the search terms of type more longitudinal research is needed on a wider
of outcome or setting, was undertaken. Second, a range of variables to understand the impact of
more specific search of the link between HRM and HRM practices on final outcomes in the NHS.
patient outcomes was conducted. One quasi-longitudinal British study has appeared
subsequent to our review by West et al.,707 which
Longitudinal studies were found to exist for only uses aggregated data from a somewhat diverse set
5 of the 12 intermediate outcomes identified of practices to create an index of high-performance
for inclusion in the review. No longitudinal HRM system, and finds it is significantly related to
evidence was found for the impact of a large mortality rates in hospitals, controlling for prior
number of categories of HRM practices: these mortality.
were recruitment, training, communication,
employee involvement, family-friendly practices
and performance management. Additionally, Impact of HRM practices on
no longitudinal studies were identified that patient outcomes
reported on the links between HRM practices
and intermediate outcomes such as motivation, The longitudinal research suggests that three
job involvement, occupational commitment, types of HRM practices – employee involvement,
engagement, organisational justice, organisational training and performance management – have a
climate, psychological contract and perceived positive impact on the key patient-care outcomes:
organisational support. The review demonstrates mortality rates, length of stay, waiting times,
that longitudinal research linking HRM practices patient safety and patient satisfaction. More
and intermediate outcomes is underdeveloped in specifically, working in quality improvement teams,
the health field. training, and feedback were consistently and
strongly associated with reduced infection rates,
The findings of the research examining the impact although much of the evidence is based on studies
of HRM practices on intermediate outcomes in the USA. We might therefore target each of these
identified in this review were far from conclusive. practices for further research.
Positive relationships were found between team
working on morale and job satisfaction, job variety More than one-half of the studies involving
and turnover intentions, and job complexity and patient outcome identified in the review examined
job satisfaction. However, the numbers of studies the impact of different staffing models through
on these relationships is small, and, moreover, comparisons of cases using different grades and
other variables, such as job demands, job control, types of staff, different combinations of staff in
task reallocation, role ambiguity, pay and employee teams, or different lengths of shift. The research
involvement, were not found to have a consistent on the health sector is therefore very different from
impact on job satisfaction or organisational the non-health sector research, which has been less
concerned with staffing practices.
239
The mechanisms (i.e. the intermediate outcomes Third, some HRM practices have been the subject
or performance-related employee behaviours) of research in both the health and the non-health
by which these HRM practices impact on patient sectors. However, a major difference between the
outcomes have not been explored. This could research in the two sectors is the specific practices
be because the HRM practices that make up the that have been studied within each HRM category.
majority of studies in this area (i.e. staffing levels, For example, the studies in work design in non-
substitution) are not primarily conceptualised as health settings cover five main topics: team work,
having an impact via employee psychological states job complexity, job demands and control, job
such as satisfaction and motivation. For example, rotation, enlargement and enrichment, and role
they may affect the way work is done. Nonetheless, conflict or clarity. The work design studies in
we should not rule out the possibility that such health are more limited in scope. Within staffing,
issues as staffing levels may have important for example, health research has focused more
implications for employee outcomes such as on topics such as staffing levels and substitution,
burnout or commitment without further research. whereas the non-health research has looked more
at recruitment and selection, induction and flexible
The recommendation here is that future research working practices.
on HRM considers the impact of these practices
on important employee psychological states and Some other differences are to be expected. The
behaviours as well as at the patient outcomes level. final outcomes differ, between the health and non-
Studying a broader range of employee outcomes heath sectors, but also the intermediate outcomes
will help provide a better understanding of these and productivity-related employee behaviours
processes in an NHS setting. have received much greater attention in non-
health compared with health settings. Although
there is a large number of studies in the non-
Impact of HRM on health field, which have examined longitudinally
intermediate and final the relationship between some of HRM practices
and employee outcomes such as satisfaction,
outcomes in the non-health commitment and employee performance and
literature turnover, there are clear gaps in the HRM
literature. The coverage of intermediate outcomes
A comparison of the non-health literature that was such as engagement, psychological contract and
reported in Chapter 8 with the health literature climate are in non-health sectors just as neglected
reported in Chapters 6 and 7 reveals a number of as they are in health.
clear similarities. First, there is an imbalance in the
practices covered, so, in both areas certain domains The majority of studies in both areas present
of HRM are covered disproportionately more positive relationships between the practices
than in others. Practices in the categories of work studied. This may reflect a bias in the acceptance of
design and staffing are more prevalent, whereas papers by journal editors and referees’ comments.
practices in other categories such as family friendly, But its implication is that we do not have a very
communication and employee representation are firm basis for ruling out some practices as definitely
underexplored. There are no studies of single not contributing significantly or independently to
status in either area. The gaps in research are final or intermediate outcomes. An examination of
not dissimilar between the two sectors in terms of the cross-sectional studies, if there were sufficient
the broad categories of HRM practice for which on specific practice–outcome links, might be used
evidence was identified. to identify some of these.
Second, there is considerable heterogeneity in the There is some similarity among the studies in this
research within each of the HRM categories for review, although it is rare for more than three or
both the health and non-health areas, and very four studies to be sufficiently similar to compare
few replication studies exist. This limits what a directly, so many of the findings in this report are
systematic comparison of the studies in a particular based on only a small number of studies. However,
domain can achieve, as we are not comparing the it is possible to identify some broad themes
same practices or performance measures. underlying the findings.
240
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
the findings of Patterson et al.34 reported in this practices in terms of organisational performance.
review. Both studies found independent effects Limitations in the methods and particularly
for job enrichment (empowerment in the terms the measures means that, in many cases, the
of the Birdi et al.709 study) and extensive training underlying concept is not clear enough for one
on organisational performance, but not for the to say which HR system is (or systems) are really
operational management initiatives of total quality more effective than others or whether there
management, JIT and advanced manufacturing are some practices that account for most of the
technology. effect. Even in some of the case studies there was
insufficient detail of some practices to be even
Research effort expended on these promising certain what these entailed. Nor again can it be
areas may have the most pay-off but it would be ruled out that one or more practice is a critical
a mistake given the heterogeneity and limited ingredient and the others are of little consequence
scope of the studies in the HRM if research were to for performance. The variety in the studies also
overconcentrate on these for a number of reasons. limits generalisation. Case studies, such as that
First, other practices, especially those for which of Freeman and Kleiner,649 offered convincing
there is no research, may be important. Second, evidence of effects of a package of practices in one
while they may be critical ingredients in a package situation, but generalisation to either other sites or
of HRM practices, they may have synergistic assumed related or similar interventions would be
effects on other practices. For example, Birdi et al. premature.
found that team working was not independently
associated with productivity but it enhanced the Finally, we should not neglect practices outside
impact of both training and empowerment, as the HRM domain. As we have seen, their impact
well as operational practices such as JIT. This may be affected by the use of HRM practices. It
contrasts with the results of the longitudinal US may also be that their use is strongly related to
study by Capelli and Neumark149 included in the use of HRM practices, which means that we
our review, which tested for synergies between can not be sure that if they are not included in the
practices but found no strong ones The practices analysis any link between an HRM practice and
relating to work organisation and training were performance that is found is either spurious or is
limited, however, to team working, cross-training a mediator of the link between the other practice
and job rotation not enrichment or other training and performance. In contrast, if HRM practices are
measures. found to be significant and non-HRM practices are
found not to be then this strengthens any argument
Third, it may yet be that integrated approach about their importance.
to HRM in which practice use is a reflection of
a more fundamental orientation on the part of
management, be it high involvement or something Relationship between
else, is more significant for practice than any one intermediate outcomes and
individual practice. A cross-sectional study using
WERS9883 data shows that an integrated measure
final outcomes
of high-involvement management centred on The intermediate outcomes examined in this
work organisation practices such as team-working, review assessed a range of mental, emotional and
idea-capturing methods, and functional flexibility attitudinal states that are hypothesised to provide a
is related to the level and rate of change in bridge between HRM and performance. While we
productivity.647 Nonetheless, this high-involvement acknowledge that HRM practices can have a direct
management is shown to be discrete from job impact on the knowledge and skills of employees,
enrichment, as, for example, it is being used in which, in turn, may influence organisational
contexts where job autonomy is low as in mass performance, this review concentrated on the
production manufacturing firms. Moreover, job role of employee sociopsychological states.
enrichment is more strongly related to labour The aim of Objective 3.1 was to investigate the
productivity than high-involvement management. correlations between these intermediate outcomes
to assess whether they form discrete concepts. In
The studies of integrated measures of HRM that some cases the meta-analyses reporting strong
we reviewed in Chapter 8, while showing that HR intercorrelations between some of the variables,
systems do make a difference to performance, but not to the degree that there was construct
found no evidence of an optimal bundle of redundancy.
242
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
The review also explored, through meta-analytic the organisation’s climate are shaped, which, in
data, the correlations between the intermediate turn, determine employee attitudes and behaviours
outcomes and productivity-enhancing behaviours. supportive of organisational performance. HRM
The premise here is that the intermediate practice do not have a direct impact on employee’s
outcomes are determinants of salient employee attitudes and behaviours; rather HRM practices are
behaviours, which, in turn, help generate effective firstly filtered and interpreted by employees so they
organisational performance. The correlations arrive at a view of what the organisation is like and
between most of variables mostly were small to the goals that it pursues.155
moderate.
An understanding of the impact of HRM
Two limitations of the meta-analyses are important practices on how employees perceive their work
to note. First, their concentration on cross-sectional environment, and, in turn, the effect of this on
research means that it is not possible to draw their behaviours is then valuable and a promising
conclusions about the causal direction of any of the area for further research. However, rather than
relationships. Second, it is not possible to identify treat climate as a catch-all concept that measures
the unique relationship between each intermediate employee perceptions about organisation life,
outcome and employee behaviours. a more productive approach is to examine the
link between the strategic focus of particular
The studies of the impact of intermediate outcomes HRM practices and climate. For example, if an
on final outcomes were sparse. No longitudinal organisation’s policy is to promote employee
research was found that examined the impact involvement, voice and participation through
of intermediate outcomes on patient outcomes its HRM practices, to what extent do employees
(Objective 4.1). In the non-health field, a small perceive that the organisation is enacting this
number of longitudinal studies were identified that focus. Such an approach also acknowledges the
examined the impact of intermediate outcomes importance of assessing the implementation
(average employee job satisfaction and climate) on fidelity of HRM practices.
organisational performance (Objective 4.2). While
the studies in this review show associations, the
evidence on the casual direction of this relationship Measures of intermediate
is mixed. This relationship is a crucial link for outcomes
the premise that HRM influences final outcomes,
partially through its impact on employee outcomes In order to identify and report on the reliability
such as job satisfaction, and we clearly need more of intermediate outcome measures, the scales
substantial data sets for surer interpretation. in all included studies were identified. This
demonstrated that the number of measures of
Despite these caveats it is worth highlighting intermediate outcomes included in the studies
here a notable consistency in the findings on the varied between types of outcomes. In some
relationships between climate, employee attitudes areas with a long research tradition such as job
and performance at the individual level and satisfaction large numbers of different scales
the organisational level. At the individual level, were identified. In other, arguably more recent
two meta-analyses indicated that relationships areas of HRM research interest, no generic
of climate (employees’ perceptions of formal scales appropriate to an NHS setting could be
and informal policies, practices, procedures and identified (as was the case for climate, for which
rewards) with job performance were mediated measures vary widely in terms of subject matter
by employee attitudes such as job satisfaction. precluding the identification of any one measure
In other words, the studies suggest climate as particularly appropriate for HRM research)
perceptions affect individual-level performance or there was no evidence from longitudinal
through their impact on employee attitudes. This research examining the link (as was the case for
finding was replicated at the organisational level, psychological contract).
as the association between organisational climate,
measured by employees’ perceptions averaged Scales varied in the degree to which they were
for each company, and company productivity or established, with examples of one-off scales or
profitability was found to be mediated by average single-item measures. Where a well-established
level of job satisfaction. These studies provide scale could be identified for an intermediate
support for the idea that HRM practices are the outcome area, it was selected for inclusion in the
means through which employee’s perceptions of review. Where several comparable scales were
243
found or no scale could be identified the subject chain and, moreover, there is insufficient evidence
experts on the review team advised on scale that HRM causes changes in organisational
selection. A major review of measures201 was used performance. The conclusion from our analysis is
as the primary source of reliability data for scales that, on the one hand, it would be premature to say
in 5 of the 12 areas. For the remainder, purposive that we have a strong evidence base for the HRM
searching was undertaken for reviews of reliability, performance link; on the other hand, however,
meta- analyses or longitudinal studies using to there is some evidence supporting elements of
provide reliability data for the scale. each stage in the causal chain and theoretical
grounds for believing that an HRM system based
Scales were included for two intermediate outcomes on enhancing employee skills, attitudes and
– engagement and psychological contract – despite behaviours should be beneficial for organisational
the fact that these two outcomes were not amongst performance, and, specifically, in the case of the
those measured by studies in this review. The NHS, patient outcomes. However, there is a clear
rationale is that both are of burgeoning importance need for research that overcomes the weaknesses of
in HRM research and their conceptual importance the current studies.
in the area justified their inclusion.
This argument has implications for funding and sector is based on the Core Standards data. The
resource allocation. Funding for an individual HCC requires Trusts to declare their degree of
or research team, over just a few years, does not compliance with the Government’s core standards
provide the resources to deploy the recommended for health care. The core standards are not HR
methods. What is required, as Wall and Wood30 practices in themselves, but specify the need for
suggest, is much more intensive collaboration than systems to be in place within Trusts, for example
is the current norm among the many potentially to ensure that clinicians continuously update skills
interested parties, including academics, employer and techniques relevant to their clinical work (C5c,
organisations, employee organisations, professional HCC Annual Health Check, 2007/8). Empirical
bodies and government departments. In short we research could be undertaken to map core
need big science projects on management practices standards to HR practice areas and assess the range
and performance that build on the understanding of specific systems in place in Trusts to meet these
generated by such systematic reviews as this one. core standards, providing data on a limited range
of HR priorities within Trusts and offering the
The notion of ‘big science’ does not preclude scope to explore the differential impact of specific
detailed study of the nature of practices, the HR practices against a particular core standard.
intention behind them or the fidelity of their
implementation. Indeed, as noted earlier in the Change and intervention studies
chapter, smaller studies that examine intent and
the depth of practices might be useful. However, To date, most research conducted to investigate
attending to such detail and the diversity in the way the links between HRM and performance, and
in which the same generic practice is designed and the relationships with, and between, intermediate
used could be seen as amplifying the need for a outcomes and employee behaviours have been
large team of coordinated researchers as in natural cross-sectional and correlational. While in
science, for example the Human Genome Project. many cases the associations are as predicted,
we cannot verify causality. We concur with Iles
Using and supplementing and Sutherland710 (in their review of change
existing data management in the NHS) that we more need quasi-
experimental research, which is better suited to
There are several ways in which existing data detecting causal effects. The prevalence of changes
sources could be further used to inform this area in HRM policies and practices, especially in work
of investigation, some examples of which are given organisation, provide a good opportunity for
here: researchers and health-care managers to examine
the impact of changes in HR practice on employee
First, WERS,83 an example of a large survey and and patient outcomes. We need to use change
hence ‘big science’ by current standards, collects programmes to gather robust and comprehensive
data on the health sector and does, in fact, follow data to help unpick the multiple dimensions
up questions about use of HR practices with a and relationships that constitute HRM and
further layer of questions (e.g. is appraisal used performance. However, assessing the impact of any
for pay or development), but in fact there has been HRM initiative must be done with care, taking into
little or no use made thus far of these data. Further consideration its relationship with other practices,
bespoke analysis of the WERS data set could the organisational context in general, and using
provide more detailed information about the use of robust methodologies for studying employee
general HR practices in the health sector. responses and organisational outcomes.
246
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Acknowledgements
Contribution of authors All authors were involved in the development of
the methodology. Jo Rick was responsible for day-
Malcolm Patterson, Jo Rick, Stephen Wood and to-day project management, Christopher Carroll
Andrew Booth designed the review and drafted and Shashi Balain conducted the searches, and all
the protocol and were responsible for its ongoing authors were involved in sifting of articles and data
conduct. extraction. All authors contributed to the drafting
of the report and approved the final version.
247
References
1. Walton R. From control to commitment in the 14. Arthur JB. Effects of human resource systems on
workplace. Harvard Business Review 1985;63:76–84. manufacturing performance and turnover. Academy
of Management Journal 1994;37:670–87.
2. Pfeffer J. Competitive advantage through people:
unleashing the power of the workforce. Boston, MA: 15. Batt R. Managing customer services: human
Harvard University Press; 1994. resource practices, quit rates, and sales growth.
Academy of Management Journal 2002;45:587–98.
3. Kochan T, Osterman P. Mutual gains. Boston, MA:
Harvard Business School Press; 1995. 16. Becker BE, Huselid MA. High performance work
systems and firm performance: a synthesis of
4. Barney J. Looking inside for competitive advantage. research and managerial implications. In Ferris GR
Academy of Management Executive 1995;9:49–61. (ed.) Research in Personnel and Human Resource
Management, 16, Greenwich, CT: JAI Press; 1998.
5. Taylor S, Beechler S, Napier N. Toward an pp. 53–101.
integrative model of strategic international human
resource management. Academy of Management 17. Guthrie JP. High involvement work practices and
Review 1996;21:959–85. turnover and productivity: evidence from New
Zealand. Academy of Management Journal 2001;
6. Taylor S, Beechler S, Najjar M, Ghosh BC. A partial 4:180–90.
test of a model of strategic international human
resource management. Advances in International 18. Huselid MA. The impact of human resource
Comparative Management 1998;12:207–36. management practices on turnover, productivity,
and corporate financial performance. Academy of
7. Cua KO, McKone KE, Schroeder R. Relationships Management Journal 1995;38:635–72.
between implementation of TQM, JIT, and TPM
and manufacturing performance. Journal of 19. Lawler E, Mohrman SA, Ledford GE Jr. Creating
Operations Management 2001;19:675–94. high performance organizations. San Francisco, CA:
Jossey-Bass; 1995.
8. Flynn B, Sakakibara S, Schroeder R. Relationship
between JIT and TQM: practices and performances. 20. MacDuffie J. Human resource bundles and
Academy of Management Journal 1995;38:1325–60. manufacturing performance: organizational logic
and flexible production systems in the world auto
9. Sakakibara S, Flynn BB, Schroeder RG, Morris WT. industry. Industrial and Labor Relations Review
The impact of just-in-time manufacturing and its 1995;48:197–221.
infrastructure on manufacturing performance.
Management Science 1997;43:1246–57. 21. Ichniowski C, Shaw K, Prennushi G. The effects
of human resource management practices
10. Shah R, Ward PT. Lean manufacturing: context, on productivity. American Economic Review
practice bundles, and performance. Journal of 1997;87:291–313.
Operations Management 2003;21:129–49.
22. Guest DE. Human resource management and
11. Womack J, Jones D, Roos D. The machine that performance: a review and research agenda.
changed the world. New York, NY: Rawson; 1990. International Journal of Human Resource Management
1997;8:263–76.
12. Dean JW, Bowen D. Management theory and total
quality: improving research and practice through 23. Becker B, Gerhart B. The impact of human resource
theory development. Academy of Management Review management on organizational performance,
1994;19:392–418. progress and prospects. Academy of Management
Journal 1996;39:779–801.
13. Appelbaum E, Bailey T, Berg P, Kalleberg AL.
Manufacturing advantage: why high performance work 24. Delery JE, Doty DH. Modes of theorizing in
systems pay off. Ithaca, NY: Cornell University Press; strategic human resource management: tests of
2000. universalistic, contingency, and configurational
249
performance predictions. Academy of Management 37. Wright PM, Gardner T, Moynihan LM. The
Journal 1996;39:802–35. impact of HR practices on the performance of
business units. Human Resource Management Journal
25. Hoque K. Human resource management and 2003;1:21–36.
performance in the UK hotel industry. British
Journal of Industrial Relations 1999;37:419–43. 38. Guest D, Conway N. Peering into the black hole:
the downside of the new employment relations
26. Wood SJ, De Menezes L. High commitment in the UK. British Journal of Industrial Relations
management in the UK: evidence from the 1999;37:367–389.
Workplace Industrial Relations Survey, and
Employers: Manpower and Skills Practices Survey. 39. Peccei R, Wood S. Does total quality management
Human Relations 1998;51:485–515. make a difference to employee attitudes? (Special
edition on Total Quality Management edited
27. Guest DE, Michie J, Conway N, Sheehan M. Human by S Hill and A Wilkinson). Employee Relations
resource management and corporate performance 1995;1:55–65.
in the UK. British Journal of Industrial Relations
2003;41:291–314. 40. Coyle-Shapiro J. Changing employee attitudes: the
independent effects of TQM and profit sharing
28. West MA, Borrill C, Dawson J, Scully J, Carter M, on continuous improvement. Journal of Applied
Anelay S, et al. The link between the management Behavioral Science 2002;38:57–77.
of employees and patient mortality in acute
hospitals. International Journal of Human Resource 41. Parker S, Wall T, Jackson P. That’s not my job:
Management 2002;13:1299–1310. developing flexible employee work orientations.
Academy of Management Journal 1997;40:899–929.
29. Wood SJ. Getting the measure of the transformed
organization. British Journal of Industrial Relations 42. Huselid MA, Becker BE. Comment on
1999;37:391–419. ‘measurement error in research on human resources
and frim performance: how much error is there and
30. Wall TD, Wood SJ. The romance of human resource how much does it influence effect size estimates?’
management and performance, and the case for big Personnel Review 2000;5:835–54.
science. Human Relations 2005;58:429–62.
43. Gerhart B, Wright PM, McMahan GC, Snell SA.
31. Wood S, Wall T. Work enrichment and employee Measurement error in research on human resources
voice in human resource management – and firm performance: how much error is there and
performance studies. International Journal of Human how does it influence effect size estimates? Personnel
Resource Management 2007;18:1335–72. Psychology 2000;53:803–34.
32. De Menezes LM, Wood S. The reality of flexible 44. Briner R. Evidence-based human resource
work systems in Britain. International Journal of management. In Trinder L, Reynolds S (eds).
Human Resource Management 2006;17:1–33. Evidence-based practice – a critical appraisal. Oxford:
Blackwell Science; 2000. pp. 184–211.
33. Patterson MG, West MA, Lawthorn R, Nickell S.
Impact of people management practices on business 45. Levy P, Williams J. The social context of
performance. Issues in People Management. London: performance appraisal: a review and framework
Institute of Personnel and Development; 1997. for the future Journal of Management
2004;30:881–905.
34. Patterson MG, West MA, Wall TB. Integrated
manufacturing, empowerment, and company 46. Taris TW. Is there a relationship between burnout
performance. Journal of Organizational Behavior and objective performance? A critical review of 16
2004;2:641–65. studies. Work and Stress 2006;20:316–34.
35. Wood SJ. Human resource management and 47. NHS CRD. Undertaking systematic reviews of research
performance. International Journal of Management on effectiveness. CRD Report Number 4. York: CRD,
Reviews 1999;1:367–413. University of York; 2001.
36. Patterson MG, Warr PB, West MA. Organizational 48. Higgins J, Green S. Cochrane handbook for systematic
climate and company performance: the role reviews of interventions. 4.2.5 (Updated 2005).
of employee affect and employee level. Journal Oxford: The Cochrane Collaboration; 2005.
of Occupational and Organizational Psychology
2004;77:193–216. 49. Rick J, Briner RB, Daniels K, Perryman S, Guppy A.
A critical review of psychosocial hazard measures. HSE
250
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Report CRR 356/20,. London: Health and Safety 61. Egger M, Zellweger-Zahner T, Schneider M,
Executive; July 2001. Junker C, Lengeler C, Antes G. Language bias in
randomised controlled trials published in English
50. Rick J, Thomson L, Briner R, O’Regan S, Daniels K. and German. Lancet 1997;350:326–29.
Review of existing supporting scientific knowledge to
underpin standards of good practice for key work-related 62. Moher D, Fortin P, Jadad AR, Juni P, Klassen T, Le
stressors, Phase 1. HSE Research Report RR024. Lorier J, Liberati A, et al. Completeness of reporting
London: Health and Safety Executive; October of trials published in languages other than English:
2002. implications for conduct and reporting of systematic
reviews. Lancet 1996;347:363–66.
51. Petticrew M, Roberts H. Systematic reviews in the social
sciences. Oxford: Blackwell; 2006. 63. Booth A. Cochrane or cock-eyed? How should be
conduct systematic reviews of qualitative research?
52. Pawson R, Greenhalgh T, Harvey G, Walshe K. Qualitative Evidence-based Practice Conference,
Realist review – a new method of systematic review Taking a Critical Stance. Coventry University,
designed for complex policy interventions. Journal 14–16 May 2001. URL: www.leeds.ac.uk/educol/
of Health Services Research and Policy 2005;10:21–33. documents/00001724.htm (accessed 20 September
2006).
53. Greenhalgh T, Robert G, Bate P, Kyriakidou O,
Macfarlane F, Peacock R. How to spread good ideas. 64. Eakin J, Mykhalovskiy E. Reframing the evaluation
A systematic review of the literature on diffusion, of qualitative research: Reflections on a review of
dissemination and sustainability of innovations in health assessment guidelines in the health sciences. Journal
service delivery and organisation. London: NHS SDO; of Evaluation in Clinical Practice 2003;9:187–194.
2004.
65. Dixon-Woods M, Cavers D, Agarwal S, Annandale E,
54. Dixon-Woods M, Shaw RL, Agarwal S, Smith JA. Arthur A, Harvey J, et al. Conducting a critical
The problem of appraising qualitative research. interpretive synthesis of the literature on access
Quality and Safety in Healthcare 2004;13:223–25. to healthcare by vulnerable groups. BMC Medical
Research Methodology 2006;6.
55. Dixon-Woods M, Agarwal S, Young B, Jones D,
Sutton A. Integrative approaches to qualitative and 66. Straus SE, Richardson WS, Glasziou P, Haynes RB.
quantitative evidence. London: Health Development Evidence-based medicine: how to practice and teach EBM,
Agency; 2004. 3rd edn. New York, NY: Churchill Livingstone;
2005.
56. Tranfield D, Denyer D, Smart P. Towards
a methodology for developing evidence- 67. Rothman K, Greenland S. In Modern epidemiology.
informed management knowledge by means of 2nd edn. Philadelphia, PA: Lippincott-Raven; 1998.
systematic review. British Journal of Management
2003;14:207–22. 68. Rothman K. Epidemiology: an introduction. New York,
NY: Oxford University Press, 2002.
57. Dixon-Woods M, Fitzpatrick R, Roberts K. Including
qualitative research in systematic reviews: problems 69. Zacharatos A, Hershcovis MS, Turner N,
and opportunities. Journal of Evaluation in Clinical Barling J. Human resource management in the
Practice 2001;7:125–33. North American automotive industry: a meta-
analytic review. Personnel Review 2007;36:231–54.
58. Mays N, Roberts E, Popay J. Synthesising research
evidence. In Fulop N, Allen P, Clarke A, Black N 70. Grayson L, Gomersall A. A difficult business: finding
(eds). Studying the organisation and delivery of health the evidence for social science reviews. Swindon: ESRC
services. London: Routledge; 2001. UK Centre for Evidence Based Policy and Practice.
URL: www.evidencenetwork.org/Documents/wp19.
59. Dixon-Woods M, Sutton A, Shaw R, Miller T, pdf (accessed 23 August 2007).
Smith J, Young B, et al. Appraising qualitative
research for inclusion in systematic reviews: a 71. Grant M, Brettle A, Long A. Developing a review
quantitative and qualitative comparison of three question: a spiral approach to literature searching.
methods. Journal of Health Services Research and Policy Poster presentation. Third Symposium on Systematic
2007;12:42–7. Reviews: Beyond the Basics of Systematic Reviews.
Oxford, July 2000.
60. Bravata D, McDonald KM, Shojania KG,
Sundaram V. Challenges in systematic reviews: 72. Booth A. Who will appraise the appraisers?
synthesis of topics related to the delivery, The paper, the instrument and the user. Health
organization and financing of health care. Annals of Information and Libraries Journal 2007;24:72–6.
Internal Medicine 2005;142:1056–65.
251
73. Edwards A, Russell I, Stott N. Signal versus noise 86. ESWT Establishment Survey on Working Time
in the evidence base for medicine: an alternative 2004–2005. Luxembourg: European Foundation
to hierarchies of evidence. Family Practice for the Improvement of Living and Working
1998;15:319–22. Conditions; 2006. http://www.eurofound.eu.int/
pubdocs/2006/27/en/1/ef0627en.pdf (accessed 10
74. Soares HP, Daniels S, Kumar A, Clarke M, Scott C, August 2006).
Swann S, et al. Bad reporting does not mean bad
methods for randomised trials: observational study 87. CVTS Continuing Training in Enterprises in Europe.
of the randomised controlled trials performed by European Continuing Vocational Training Survey
the Radiation Therapy Oncology Group. British in Enterprises; 2001. URL: http://ec.europa.eu/
Medical Journal 2004;328:22–4. education/programmes/leonardo/new/leonardo2/
cvts/index_en.html (accessed 10 August 2006).
75. Dixon-Woods M, Agarwal S, Jones D, Young B,
Sutton A. Synthesising qualitative and quantitative 88. EPOC Employee Direct Participation in Organisational
evidence: a review of methods. Journal of Health Change (EPOC Project). European Foundation for the
Services Research and Policy 2005;10:45–53. Improvement of Living and Working Conditions;
1996. URL: www.eurofound.eu.int/publications/files/
76. Popay J, Roberts H, Sowden A, Petticrew M, Arai L, EF9803EN.pdf (accessed 10 August 2006).
Rodgers M, et al. Guidance on the conduct of narrative
synthesis in systematic reviews: a product of the ESRC 89. CIPD Reward management. Annual Postal Survey
methods programme (Version 1). Lancaster: Institute of Report 2006. London: Chartered Institute of
Health Research; 2006. Personnel and Development; 2006. URL: www.
cipd.co.uk/NR/rdonlyres/B7AA6A5C-0728–46FE-
77. Hyde P, Boaden RB, Cortvriend P, Harris C, A671–555950BD4BCF/0/rewmansr0206.pdf
Marchington M, Pass S, et al. Improving health through (accessed 10 August 2006).
HRM: mapping the territory. London: Chartered
Institute of Personnel and Development; 2006. 90. IPD Performance pay trends in the UK. IPD
Survey Report 9. London: Institute of Personnel
78. Boselie P, Dietz G, Boon C. Commonalities and and Development; 1999. URL: www.cipd.
contradictions in HRM and performance research. co.uk/NR/rdonlyres/972D73A3-B1DD-4499-
Human Resource Management Journal 2005;15:67–94. BCC1–66C8EB1C639C/0/topical_trend_survey.pdf
(accessed 10 August 2006).
79. Michie S, West M. Measuring staff management and
human resource performance in the NHS. London: 91. NESS National Employers Skills Survey 2004: main
Commission for Health Improvement, 2003. report. London: Department for Education and
Skills; 2005. URL: www.dfes.gov.uk/successforall/
80. National Institute of Clinical Studies Factors downloads/nationalemployersskillssurvey2004
supporting high performance in health care organisations. mainreport-176–268.pdf (accessed 10 August 2006).
melbourne, NSW: La Trobe University Health
Management Group. 92. CIPD Getting a kick out of work. Flexible working trends.
Postal Survey Report 2002. London: Institute of
81. Noblit GW, Dwight HR. Metaethnography:synthesising Personnel and Development; 2002. URL: www.cipd.
qualitative studies. London: Sage; 1988. co.uk/NR/rdonlyres/F7F79AF4–65F0–4236–888C-
6B6B8D4A4979/0/Flexible_Working.pdf (accessed
82. WERS Inside the workplace. Findings from the 2004 10 August 2006).
Workplace Employment Relations Survey. Abingdon:
Routledge; 2006. 93. CIPD Managing employee careers. Issues, trends
and prospects. Postal Survey Report 2003. London:
83. WERS Workplace Employee Relations Survey. First Chartered Institute of Personnel and Development;
findings; 1998. URL: www.dti.gov.uk/files/file11678. 2003. URL: www.cipd.co.uk/NR/rdonlyres/
pdf (accessed 10 August 2006). E96E8FE4–9403–4304–82AC-879D8BC51EDE/0/
manag_emp_car_Postal survey.pdf (accessed 10
84. WERS Inside the workplace. First findings from the August 2006).
Workplace Employee Relations Survey; 2004. URL:
www.dti.gov.uk/files/file11423.pdf (accessed 1 94. CIPD Flexible working: impact and implementation.
September 2006). An employer postal survey. Postal Survey Report
2005. London: Chartered Institute of Personnel
85. WERS Workplace Employee Relations Survey, 2004. and Development; 2005. URL: www.cipd.co.uk/
Cross-section management questionnaire (MQ). URL: NR/rdonlyres/257CE4EE-356B-43F5–8927–
www.wers2004.info/pdf/Vol%202%20(part%20 5C86203D7AA1/0/flexworksurv0205.pdf (accessed
2)%20-%20Cross-Section%20Management%20 10 August 2006).
Questionnaire.pdf (accessed 4 September 2006).
252
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
95. CIPD Performance management. Postal Survey Report 106. Godard J. Beyond the high-performance paradigm?
2005. London: Chartered Institute of Personnel An analysis of variation in Canadian managerial
and Development; 2005. URL: www.cipd.co.uk/ perceptions of reform programme effectiveness.
NR/rdonlyres/07BC1B70-E82A-455D-94F6- British Journal of Industrial Relations 2001;3:25–52.
AEEE3DEE0D4D/0/perfmansr0905.pdf (accessed
10 August 2006). 107. Cunningham I, James P, Dibben P. Bridging the
gap between rhetoric and reality: line managers and
96. CIPD Absence management. Postal Survey Report the protection of job security for ill workers in the
2006. London: Chartered Institute of Personnel modern workplace. British Journal of Management
and Development; 2006. URL: www.cipd.co.uk/ 2004;15:273–90.
NR/rdonlyres/22CB05C5-E52D-445B-891A-
9886C95FD90D/0/absmgmnt0706.pdf (accessed 10 108. Hutchinson S, Purcell J. Bringing policies to life, the
August 2006). vital role of front line managers in people management.
Executive Briefing. London: Chartered Institute of
97. CIPD Learning and development. Annual Postal Personnel and Development; 2003.
Survey Report 2006. London: Chartered Institute
of Personnel and Development; 2006. URL: www. 109. Whittaker S, Marchington M. Devolving HR
cipd.co.uk/NR/rdonlyres/97BE272C-8859–4DB1- responsibility to the line: threat, opportunity or
BD99–17F38E4B4484/0/lrnandevsurv0406.pdf partnership? Employee Relations 2003;25:245–61.
(accessed 10 August 2006).
110. Truss C. Complexities and controversies in linking
98. CIPD Recruitment, retention and turnover. Annual HRM with organizational outcomes. Journal of
Postal Survey Report 2006. London: Chartered Management Studies 2001;38:1121–49.
Institute of Personnel and Development;
2006. URL: www.cipd.co.uk/NR/rdonlyres/ 111. Mihalic S. The importance of implementation
A5316993-E9EB-413D-A673-D1D6A5063DD3/0/ fidelity. Emotional and Behavioural Disorders in Youth
recruitretntsurv06.pdf (accessed 10 August 2006). 2004;4:83–86 and 99–105.
99. Wood S, Stride C, Wall T, Clegg C. Revisiting the 112. Mihalic S. The importance of implementation
use and effectiveness of modern management fidelity. In (eds). Blueprints Violence Prevention
practices. Human Factors and Ergonomics in Initiaitive. Boulder, CO; IBS; 2002. pp. 1–2.
Manufacturing 2004;14:415–32.
113. Dane A, Schneider B. Program integrity in
100. CRANET CRANET Survey on Comparative Human primary and early secondary prevention: are
Resource Management. Executive Report. Cranfield: implementation effects out of control. Clinical
Cranfield University; 2003. Psychology Review 1998;18:23–45
101. CIPD Benefiting from a balanced life? CIPD Survey 114. Hulscher M, Laurant M, Grol R. Process evaluation
Report 10. London: Chartered Institute of Personnel on quality improvement interventions. Quality and
and Development; 1999. URL: www.cipd.co.uk/ Safety in Health Care 2003;12:40–6.
NR/rdonlyres/C930B528–55AA-4192–85BD-
888BEFA4345D/0/balancedsurvey.pdf (accessed 10 115. Blakely C, Mayer J, Gottschalk R,
August 2006). Schmitt N, Davidson W, Riotman D, et al. The
fidelity-adaptation debate: implications for
102. Newton J, Hunt J, Stirling J. Human resource the implementation of public sector social
management in general practice. British Journal of programs. American Journal of Community Psychology
General Practice 1996;46:81–85. 1987;15:253–68.
103. CHAI National NHS Staff Survey 2007. London: 116. Elliott D, Mihalic S. Issues in disseminating and
Commission for Healthcare Audit and Inspection; replicating effective prevention programs. Prevention
2008. Science 2004;5:47–53.
104. Dusenbury L, Brannigan R, Falco M, Hansen W. A 117. Forgatch M, Patterson G, DeGarmo D. Evaluating
review of research on fidelity of implementation: fidelity: predictive validity for a measure of
implications for drug abuse prevention in school competent adherence to the Oregon model of
settings. Health Education Research 2003;18:237–56. parent management training. Behaviour Theory
2005;36:3–13.
105. Dobson L, Cook T. Avoiding type III error in
proram evaluation: results from a field experiment. 118. Rinaldi M, McNeil K, Firn M, Koletsi M, Perkins R,
Evaluation and Program Planning 1980;3:269–76. Singh S. What are the benefits of evidence-based
253
supported employment for patients with 130. Whittaker JK, Schinke SP, Gilchrist LD. The
first-episode psychosis? Psychiatric Bulletin ecological paradigm in child, youth, and family
2004;28:281–4. services: implications for policy and practice. Social
Service Review 1986;60:483–503.
119. Arai L, Roen K, Roberts H, Popay J. It might work
in Oklahoma but will it work in Southampton? 131. Mueser KT, Drake RE, Bond GR, Resnick SG.
Context and implementation in the effectiveness Models of community care for severe mental
literature on domestic smoke detectors. Injury illness: A review of research on case management.
Prevention 2005;11:148–51. Schizophrenia Bulletin 1998;24:37–74.
120. Roen K, Arai L, Roberts H, Popay J. Extending 132. Steinhardt M. Physical education. In Jackson PW
systematic reviews to include evidence on (ed.). Handbook of research on curriculum. New York,
implementation: Methodological work on a review NY: Macmillan; 1992. pp. 961–1001.
of community-based initiatives to prevent injuries.
Social Science and Medicine 2006;63:1060–71. 133. Holland S. Measuring process in drug abuse
treatment research. In DeLeon G, Ziegenfuss JT
121. O’Brien R. Translating a research intervention into (eds). Therapeutic communities for addictions: readings
community practice: the nurse–family partnership. in theory, research, and practice. Springfield, IL:
Journal of Primary Prevention 2005;26:241–57. Charles C Thomas; 1986. pp. 169–81.
122. Resnick B, Inguito P, Orwig D, Yahiro J, Hawkes W, 134. Knutson NM, Forgatch MS, Rains LA. Fidelity of
Werner M, et al. Treatment fidelity in behaviour implementation rating system (FIMP): the training
change research: a case example. Nursing Research manual for PMTO. Eugene, OR: Oregon Social
2005;54:139–43. Learning Center; 2003.
123. Penuel W, Means B. Implementation variation and 135. Bond GR, Drake RE, Mueser KT, Latimer E.
fidelity in an Inquiry Science Program: analysis of Assertive community treatment for people with
GLOBE data reporting patterns. Journal of Research severe mental illness: critical ingredients and impact
in Science Teaching 2004;41:294–315. on patients. Diease Management and Health Outcomes
2001;9:141–59.
124. Cash S, Berry M. Measuring service delivery in a
placement prevention program: an application to 136. Noel P. The impact of therapeutic case management
an ecological model. Administration in Social Work on participation in adolescent substance abuse
2003;27:65–85. treatment. American Journal of Drug and Alcohol Abuse
2006;32:311–27.
125. Resnick B, Neale M, Rosenheck R. Impact of
public support payments, intensive psychiatric 137. Grol R, Dalhuijsen J, Thomas S, Veld C, Rutten G,
community care, and program fidelity on Mokkink H. Attributes of clinical guidelines that
employment outcomes for people with severe influence use of guidelines in general practice:
mental illness. Journal of Nervous and Mental Disease observational study. British Medical Journal
2003;191:139–44. 1998;17:858–61.
126. Herzog S, Wright P. Addressing the fidelity 138. Bellg A, Borelli B, Resnick B, Hecht J, Minicucci D,
of personal and social responsibility model Ory M, et al. Enhancing treatment fidelity in health
implementation. Research Quarterly for Exercise and behaviour change studies: best practices and
Sport 2005;76(Suppl.):A-76. recommendations from the NIH Behaviour Change
Consortium. Health Psychology 2004;23:443–51.
127. McGrew J, Griss M. Concurrent and predictive
validity of two scales to assess the fidelity of 139. Medical Research Council. A framework for the
implementation of supported employment. development and evaluation of RCTs for complex
Psychiatric Rehabilitation Journal 2005; 29:41–7. interventions to improve health. London: Medical
Research Council; 2000.
128. Faw L, Hogue A, Liddle H. Multidimensional
implementation evaluation of a residential 140. McGovern P, Gratton L, Hope-Haily V, Stiles P,
treatment program for adolescent substance use. Truss, C. Human resource management on the line?
American Journal of Evaluation 2005;26:77–93. Human Resource Management Journal 1997;7:12–29.
129. Bond GR, Becker DR, Drake RE, Vogler KM. 141. Protherough R, Pick J. Managing Britannia: culture
A fidelity scale for the Individual Placement and management in modern Britain. Exeter: Imprint
and Support model of supported employment. Academic; 2003.
Rehabilitation Counselling Bulletin 1997;40:265–84.
254
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
142. Van Veldhoven M. Financial performance and the 154. Kopelman RE, Brief AP, Guzzo R. The role
long-term link with HR practices, work climate of climate and culture in productivity. In
and job stress. Human Resource Management Journal Schneider B. (ed.). Organizational climate and culture.
2005;15:30–53. San Francisco, CA: Jossey-Bass; 1990. pp. 282–318.
143. Wright PM, Gardner TM, Moynihan LM, 155. Ostroff C, Bowen DE. Moving HR to a higher level:
Park HJ, Gerhart B, Delery JE. Measurement HR practices and organizational effectiveness. In
error in research on human resources and firm Klein KJ, Kozlowski SWJ (eds). Multilevel theory,
performance: additional data and suggestions research, and methods in organizations: foundations,
for future research. Personnel Psychology extensions, and new direction. San Francisco, CA:
2001;54:875–901. Jossey-Bass; 2000. pp. 211–67.
144. Hackman JR, Wageman R. Total quality 156. Sparrow PR. Developing diagnostics for high
management: empirical, conceptual and performance organization cultures. In The
practical issues. Administrative Science Quarterly international handbook of organizational culture and
1995;40:309–42. climate. Chichester: John Wiley and Sons Ltd; 2001.
pp. 85–106.
145. Khilji SE, Wang X. ‘Intended’ and ‘implemented’
HRM: the missing linchpin in strategic human 157. Iaffaldano MT, Muchinsky PM. Job satisfaction
resource management research. International Journal and job performance: a meta-analysis. Psychological
of Human Resource Management 2006;17:1171–89. Bulletin 1985;97:251–73.
146. Huselid M, Jackson S, Schuler R. Technical and 158. Locke EA. Job satisfaction. In Gruneberg M, Wall T
strategic human resource management effectiveness (eds). Social psychology and organizational behavior.
as determinants of firm performance. Academy of New York: Wiley; 1984. pp. 93–117.
Management Journal 1997;40:171–88.
159. Organ DW. Organizational citizenship behaviour: the
147. Currie G, Procter S. The interaction of human good soldier syndrome. Lexington, MA: Lexington
resource policies and practices with the Books; 1998.
implementation of teamworking: evidence from the
UK public sector. The International Journal of Human 160. Hoffman BJ, Blair CA, Meriac JP, Woehr DJ.
Resource Management 2003;14:581–99. Expanding the criterion domain? a quantitative
review of the OCB literature. Journal of Applied
148. Bacon N, Blyton P. The effects of co-operating or Psychology 2007;92:555–66.
conflicting over work restructuring: evidence from
employees. The Sociological Review 2006;54:1–19. 161. Taber TD, Alliger GM. A task-level assessment of
job satisfaction. Journal of Organizational Behaviour
149. Capelli P, Neurmark D. Do ‘high-performance’ work 1995;16:101–21.
practices improve establishment-level outcomes?
Industrial and Labor Relations Review 2001;54:737. 162. Organ DW, Ryan K. A meta-analytic review
of attitudinal and dispositional predictors of
150. Harachi TW, Abbott RD, Catalano RF, Haggerty KP, organizational citizenship behavior. Personnel
Fleming CB. Opening the black box: using process Psychology 1995;48:775–802.
evaluation measures to assess implementation and
theory building. American Journal of Community 163. Halbesleben KN, Jonathon RB, Petroski GF,
Psychology 1999;27:711–31. Meyer W. Relationship between management
philosophy and clinical outcomes. Health Care
151. Montgomery P, Donkoh C, Underhill K. Management Review 2007;32:128–39.
Independent living programs for young people
leaving the care system: the state of the evidence. 164. Becker B, Huselid M. Strategic human resources
Children and Youth Services Review 2006;28:1435–48. management: where do we go from here? Journal of
Management 2006;32:898–925.
152. Nutley S, Homel P. Delivering evidence-based policy
and practice: lessons from the implementation of 165. Combs J, Liu Y, Hall A, Ketchen D. How much
the UK Crime Reduction Programme. Evidence and do high-performance work practices matter? A
Policy 2006 2:5–26. meta-analysis of their effects on organizational
performance. Personnel Psychology 2006;59:501–28.
153. Becker B, Huselid M, Pickus P, Spratt M. HR
as a source of shareholdervalue: research and 166. Michie S, West M. Managing people and
recommendations. Human Resource Management performance: an evidence based framework applied
1997;36:39–47. to health service organisations. International Journal
of Management Reviews 2004;5/6:91–111.
255
167. Boxall P, Purcell J. Strategy and human resource 181. Maslach C, Schaufeli WB, Leiter MP. Job burnout.
management. London: Palgrave Macmillan; 2003. Annual Review of Psychology 2001;52:397–422.
168. Kanfer R, Ackerman PL. Individual differences 182. Schaufeli WB, Bakker AB. The measurement of
in work motivation: further explorations of a trait work engagement with a short questionnaire: a
framework. Applied Psychology: An International cross-national study. Educational and Psychological
Review 2000;49:470–82. Measurement 2007;66:701–16.
169. Arnold A. Work psychology: understanding human 183. Schaufeli WB, Bakker AB. Job demands, job
behaviour in the workplace. London: Pearson resources and their relationship with burnout
Education; 2005. and engagement: a multi-sample study. Journal of
Organizational Behavior 2004;25:293–315.
170. Locke EA. The nature and causes of job satisfaction.
In Dunnette MD (ed.). Handbook of industrial and 184. Schaufeli WB, Enzmann D. The burnout companion to
organizational psychology. Chicago, IL: Rand McNally; study and practice: a critical analysis. Philadelphia, PA:
1976. pp. 1297–349. Taylor & Francis; 1998.
171. Judge TA, Parker S, Colbert AE, Heller D, 185. Le Blanc PM, Hox JJ, Taris TW, Schaufelli WB,
Ilies R. Job satisfaction: a cross-cultural review. In Peeters MCW. Take care! The evaluation of a team-
Anderson N, Ones DS, Sinangil HK, Viswesvaran C based burnout intervention program for oncology
(eds). Handbook of industrial, work and organizational care providers. Journal of Applied Psychology 2007
psychology, Vol 2; 2001. pp. 25–52. 92:213–27.
172. Guest DE. Commitment. In Nicholson N, Audia PG, 186. Lee RT, Ashforth BE. A meta-analytic examination
Pillutla MM (eds). The Blackwell encyclopedia of of the correlates of the three dimensions
management: organizational behaviour, 2nd edn. of job burnout. Journal of Applied Psychology
London: Blackwell; 1995. pp. 44–6. 1996;81:123–33.
173. Mowday R, Porter L, Steers R. Employee– 187. Leiter MP, Maslach C. The impact of interpersonal
organizational linkages. New York: Academic Press environment on burnout and organizational
1982. commitment. Journal of Organizational Behavior
1988;9:297–308.
174. Meyer J, Allen N. A three-component
conceptualization of organizational commitment. 188. Schaufeli WB, Leiter MP, Maslach C, Jackson SE.
Human Resource Management Review 1991;1:61–89. Maslach Burnout Inventory – general survey.
In Maslach C, Jackson SE, Leiter MP (eds). The
175. Cohen A. Organizational commitment and Maslach burnout inventory – test manual, 3rd edn.
turnover: a meta-analysis. Academy of Management Palo Alto, CA: Consulting Psychologists Press; 1996.
Journal 1993;36:1140–57. pp. 19–26.
176. Wallace JE. Professional and organizational 189. Brown SP. A meta-analysis and review of
commitment: compatible or incompatible? Journal organizational research on job involvement.
of Vocational Behavior 1993;42:333–49. Psychological Bulletin 1996;120:235–55.
177. Lee K, Carswell JJ, Allen NJ. A meta-analytic review 190. Lawler EE III, Hall DT. Relationship of job
of occupational commitment: relations with person- characteristics to job involvement, satisfaction, and
and work-related variables. Journal of Applied intrinsic motivation. Journal of Applied Psychology
Psychology 2000;85:799–811. 1970;54:305–12.
178. Carson KD, Bedeian AG. Career commitment: 191. Lawler EE III. High-involvement management:
construction of a measure and examination of participative strategies for improving organizational
its psychometric properties. Journal of Vocational performance. San Francisco, CA: Jossey-Bass; 1986.
Behaviour 1994;44:237–62.
192. Hackman JR, Lawler EE III. Employee reactions
179. Schaufeli WB, Salanova M, Gonzalez-Roma V, to job characteristics. Journal of Applied Psychology
Bakker AB. The measurement of engagement and 1971;55:259–86.
burnout: a two-sample confirmatory factor analytic
approach. Journal of Happiness Studies 2002;3:71–92. 193. Kahn WA. Psychological conditions of personal
engagement and disengagement at work. Academy of
180. Mauno S, Kinnunen U, Ruokolainen M. Job Management Journal 1990;33:692–724.
demands and resources as antecedents of work
engagement: a longitudinal study. Journal of 194. Janssen O. Innovative behaviour and job
256
Vocational Behaviour 2006;70:149–71. involvement at the price of conflict and less
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
satisfactory relations with co-workers. Journal of the workplace. Westport, CT: Quorum; 1995. pp.
of Occupational and Organizational Psychology 149–65.
2003;76:347–64.
208. Rhoades L, Eisenberger R. Perceived organizational
195. Hackman JR, Oldham GR. Development of the support: a review of the literature. Journal of Applied
job diagnostic survey. Journal of Applied Psychology Psychology 2002;87:698–714.
1975;60:159–70.
209. Rousseau DM. The construction of climate
196. Morrell K, Loan-Clarke J, Wilkinson A. Unweaving in organizational research. In Cooper C,
leaving: the use of models in the management Robertson I (eds). International review of industrial
of employee turnover. International Journal of and organizational psychology. New York: Wiley; 1988.
Management Reviews 2001;3:219–44. pp. 139–158.
197. Rousseau DM. Psychological contracts in organizations: 210. Schneider B, Reichers AE. On the etiology of
understanding written and unwritten agreements. climates. Personnel Psychology 1983;36:19–39.
Thousand Oaks, CA: Sage; 1995.
211. Campbell J, Dunnette M, Lawler E, Weick K.
198. Rousseau DM, Schalk R. psychological contracts in Managerial behavior, performance, and effectiveness.
employment: cross-national perspectives. Thousand New York: McGraw-Hill; 1970.
Oaks, CA: Sage; 2000.
212. James LA, James LR. Integrating work environment
199. Cropanzano R, Greenberg J. Progress in perceptions: explorations into the measurement
organizational justice: tunnelling through the maze. of meaning. Journal of Applied Psychology
International Review of Industrial and Organizational 1989;74:739–51.
Psychology 1997;12:317–72.
213. James LR, Jones AP. Organizational climate: a
200. Greenberg J. Organizational justice: yesterday, review of theory and research. Psychological Bulletin
today, and tomorrow. Journal of Management 1974;8:1096–112.
1990;16:399–432.
214. Schneider B. The psychological life of
201. Fields DL. Taking the measure of work: a guide to organizations. In Ashkanasy NM, Wilderon CPM,
validated scales for organizational research and diagnosis. Peterson MF (eds). Handbook of organizational culture
Sage: London; 2002. and climate. Thousand Oaks, CA: Sage; 2000. pp.
xvii–xxi.
202. Cohen-Charash Y, Spector PE. The role of
justice in organizations: a meta-analysis. 215. Schneider B. Organizational climates: an essay.
Organizational Behavior and Human Decision Processes Personnel Psychology 1975;36:19–36.
2001;2:278–321.
216. Schneider B. The climate for service: an application
203. Bies RJ, Moag JF. Interactional justice: of the climate construct. In Schneider B (ed.).
communication criteria of fairness. Research on Organizational climate and culture. San Francisco, CA:
Negotiations In Organizations 1986;1:43–55. Jossey-Bass; 1990. pp. 383–412.
204. Greenberg J. Employee theft as a reaction to 217. Schneider B, White SS, Paul MC. Linking service
underpayment inequity: the hidden costs of pay climate and customer perceptions of service quality:
cuts. Journal of Applied Psychology 1990;75:561–68. tests of a causal model. Journal of Applied Psychology
1998;83:150–63.
205. Greenberg J. The social side of fairness:
interpersonal and informational classes of 218. Zohar D. A group level model of safety climate:
organizational justice. In Cropanzano R (ed.). Justice testing the effect of group climate on microaccidents
in the workplace: approaching fairness in human resource in manufacturing jobs. Journal of Applied Psychology
management. Hillsdale, NJ: Erlbaum; 1993. pp. 2000;85:587–96.
79–103.
219. Rentsch J. Climate and culture: interaction and
206. Eisenberger R, Huntington R, Hutchison S, Sowa D. qualitative differences in organizational meanings.
Perceived organizational support. Journal of Applied Journal of Applied Psychology 1990;75:668–81.
Psychology 1986;71:500–7.
220. Steel RP, Shane GS, Kennedy KA. Effects of social-
207. Shore LM, Shore TH. Perceived organizational system factors on absenteeism, turnover, and job
support and organizational justice. In performance. Journal of Business and Psychology
Cropanzano RS, Kacmar KM (eds). Organizational 1990;4:423–30.
politics, justice and support: managing the social climate
257
221. Mathieu JE, Hofmann DA, Farr JL. Job 234. Agho AO, Mueller CW, Price JL. Determinants of
perception–job satisfaction relations: an empirical employee job satisfaction: an empirical test of a
comparison of three competing theories. causal model. Human Relations 1993;46:1007–27.
Organizational Behaviour and Human Decision Processes
1993;56:370–87. 235. Bateman TS, Strasser S. A longitudinal analysis
of the antecedents of organisational commitment.
222. James LR, Tetrick LE. Confirmatory analytic tests Academy of Management Journal 1984;27:95–112.
of three causal models relating job perceptions
to job satisfaction. Journal of Applied Psychology 236. Blau G. Testing the longitudinal impact of work
1996;71:77–82. variables and performance appraisal satisfaction on
subsequent overall job satisfaction. Human Relations
223. James LR, Jones AP. Perceived job characteristics 1999;52:1099–113.
and job satisfaction: an examination of reciprocal
causation. Personnel Psychology 1980;33:97–135. 237. Lovgren G, Rasmussen BH, Engstrom B. Working
conditions and the possibility of providing good
224. De Cotiis TA, Summers TP. A path analysis of a care. Journal of Nursing Management 2002;10:201–9.
model of the antecedents and consequences of
organizational commitment. Human Relations 238. Sandman BA. The measurement of job stress:
1987;40:445–70. development of the Job Stress Index. In Cranny CJ,
Smith PC, Stone EF (eds). Job satisfaction: How people
225. Brown S, Leigh TW. A new look at psychological feel about their jobs and how it affects their performance.
climate and its relationship to job involvement, New York, NY: Lexington; 1992. pp. 241–254.
effort, and performance. Journal of Applied Psychology
1996;81:358–68. 239. Goldberg D. Manual of the General Health
Questionnaire. Windsor, ON: NFER-Nelson; 1978.
226. Pritchard RD, Karasick BW. The effects of
organizational climate on managerial job 240. Jones RA, Dougherty M, Martin S. Program
performance and satisfaction. Organizational evaluation of a unit reengineered for
Behavior and Human Performance 1973;9:126–46. patient-focused care. Holistic Nursing Practice
1997;11:31–46.
227. Ulrich D. Measuring human resources: an overview
of practice and a prescription for results. Human 241. Price JL, Mueller C. Absenteeism and turnover of
Resource Management Journal 1997;36:303–515. hospital employees. Greenwich, CT: JAI Press; 1986.
228. Parkes KR. Occupational stress among student 242. Marmot MG, Davey Smith G, Stansfeld SA, Patel C,
nurses: a natural experiment. Journal of Applied North F, Head J, et al. Health inequalities among
Psychology 1982;67:784–96. British civil servants: the Whitehall II study. Lancet
1991;337:1387–93.
229. Nielsen ML, Kristensen TS, Smith-Hansen L. The
Intervention Project on Absence and Well-being 243. Smith PC, Kendall LM, Hulin CL. The measurement
(IPAW): design and results from the baseline of a of satisfaction in work and retirement. Chicago, IL:
5-year study. Work and Stress 2002;16:191–206. Rand McNally; 1969.
230. Jones RAP, Dougherty M, Martin S. Program 244. Porter LW, Steers RM, Mowday RT, Boulian PV.
evaluation of a unit reengineered for Organizational commitment, job satisfaction and
patient-focused care. Holistic Nursing Practice turnover among psychiatric technicians. Journal of
1997;11:31–46. Applied Psychology 1974;59:603–9.
231. O’Connor S. Evaluating the implementation of 245. Brayfield AH, Rothe HF. An index of job
team-nursing in a clinical setting. British Journal of satisfaction. Journal of Applied Psychology
Nursing 1993;2:542–47. 1951;35:307–11.
232. Jung F, Pearcey L, Phillips J. Evaluation of a 246. Karasek R, Theorell T. Healthy work: stress,
program to improve nursing assistant use. Journal of productivity, and the reconstruction of working life. New
Nursing Administration 1994;24:42–7. York, NY: Basic Books; 1990.
233. De Jonge J, Dormann C, Janssen PPM, Dollard MF, 247. Rasmussen BH. Utveckling Av Ett Kliniskt Relevant
Landeweerd JA, Nijhuis FJN. Testing reciprocal Instrument För Att Mäta Arbetstillfredställelse I Vården.
relationships between job characteristics and Report. Umeå: Department of Nursing Umeå
psychological well-being: a cross-lagged structural University; 1996.
equation model. Journal of Occupational and
Organizational Psychology 2001;74:29–46.
258
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
248. Macdonald M, Bodzak W. The performance of a 259. Birnbaum ML, Robinson NE, Kuska BM, Stone HL,
self-managing day surgery nurse team. Journal of Fryback DG, Rose JH. Effect of advanced cardiac
Advanced Nursing 1999;29:859–68. life-support training in rural, community hospitals.
Critical Care Medicine 1994;22:741–9.
249. Browne G, Lam L, Giles H, McCaskill M,
Exley B, Fasher B. The effects of a seamless model 260. Gottlieb DJ, Parenti CM, Peterson CA, Lofgren RP.
of management on the quality of care for emergency Effect of a change in house staff work schedule on
department patients. Journal of Quality in Clinical resource utilization and patient care. Archives of
Practice 2000;20:120–6. Internal Medicine 1991;151:2065–70.
250. Pines AM, Aronson E, Kafry D. Burnout: from tedium 261. Fischer U, Vosshenrich R, Baum F, Schorn C,
to personal growth. New York, NY: Free Press; 1981. Funke M, Strasser G, et al. Qualitätsmanagement in
einer radiologischen Abteilung. Management in der
251. Ekvall G, Arvonen J, Waldenstrom-Lindblad I. Radiologie. Radiologe 2002;42:361–8.
Creative organizational climate. Construction and
validation of a measuring instrument, Report 2. 262. McMullen M, Alexander MK, Bourgeois A,
Stockholm: Swedish Council for Management and Goodman L. Evaluating a nurse practitioner service.
Work Life Issues; 1983 Dimensions of Critical Care Nursing 2001;20:30–4.
252. Krausz M, Koslowsky M, Shalom N, Elyakim N. 263. Lavis J, Davies H, Oxman A, Denis, J-L,
Predictors of intentions to leave the ward, the Golden-Biddle K, Ferlie E. Towards systematic
hospital, and the nursing profession: a longitudinal reviews that inform health care management and
study. Journal of Organizational Behavior policy-making. Journal of Health Services Research and
1995;16:277–88. Policy 2005;10(Suppl. 1):35–48.
253. Carson SS, Stocking C, Podsadecki T, 264. Healthcare Commission. NHS performance indicators.
Christenson J, Pohlman A, MacRae S, et al. Effects London: Healthcare Commission; 2005.
of organizational change in the medical intensive
care unit of a teaching hospital: a comparison of 265. Department of Health. Creating a patient-led NHS:
‘open’’and ‘closed’ formats. Journal of the American delivering the NHS Improvement Plan. London:
Medical Association 1996;276:322–8. Department of Health; 2005.
254. Jeffery HE, Kocova M, Tozija F, Gjorgiev D, 266. Appleby J, Harison A, Devlin N. What is the real cost
Pop-Lazarova M, Foster K, et al. The impact of of patient choice? London: King’s Fund; 2003.
evidence-based education on a perinatal capacity-
building initiative in Macedonia. Medical Education 267. Appleby J, Devlin N. Measuring NHS success.
2004;38:435–47. London: King’s Fund; 2005.
255. Ogden J, Hoppe R. Changing practice nurses’ 268. Okamoto T, Shimozuma K, Katsumata N, Koike M,
management of obesity. Proceedings of the ASO Hisashige A, Tanaka K, et al. On behalf of the
and BDA symposium, 25 November 1997, St Task Force of the Japanese Breast Cancer Society
Bartholomew’s Hospital, London. Journal of Human for the Development of Guidelines for Quality of
Nutrition and Diet 1998;11:249–56. Life Assessment Studies of Breast Cancer Patients.
Measuring quality of life in patients with breast
256. Campbell JC, Coben JH, McLoughlin E, cancer: a systematic review of reliable and valid
Dearwater S, Nah G, Glass N, et al. An evaluation instruments available in Japan. Breast Cancer
of a system-change training model to improve 2003;10:204–13.
emergency department response to battered
women. Academic and Emergency Medicine 269. Gilbody SM, House AO, Sheldon T. Routine
2001;8:131–8. administration of health related quality of life
(HRQoL) and needs assessment instruments to
257. Barnason S, Merboth M, Pozehl B, Tietjen MJ. improve psychological outcome – a systematic
Utilizing an outcomes approach to improve pain review. Psychological Medicine 2002;32:1345–56.
management by nurses: a pilot study. Clinical Nurse
Specialist 1998;12:28–36. 270. Bunn F, Byrne G, Kendall S. The effects of
telephone consultation and triage on healthcare use
258. Harrison LL, Novak DA. Evaluation of a and patient satisfaction: a systematic review. British
gerontological nursing continuing education Journal of General Practice 2005;55:956–61.
programme: effect on nurses’ knowledge
and attitudes and on patients’ perceptions 271. Draper M, Cohen P, Buchan H. Seeking consumer
and satisfaction. Journal of Advanced Nursing views: what use are results of hospital consumer
1998;13:684–92.
259
surveys? International Journal for Quality in Health practitioners. American Journal of Critical Care
Care 2001;13:463–8. 2002;11:353–62.
272. Sitzia J. How valid and reliable are patient 284. Burns SM, Marshall M, Burns JE, Ryan B,
satisfaction data? An analysis of 195 studies. Wilmoth D, Carpenter R, et al. Design, testing,
International Journal for Quality in Health Care and results of an outcomes-managed approach
1999;11:319–28. to patients requiring prolonged mechanical
ventilation. American Journal of Critical Care
273. Sitzia J, Wood N. Patient satisfaction: a review of 1998;7:45–57.
issues and concepts. Social Science and Medicine
1997;45:1829–43. 285. Schweer LH, Cook BS, Fanta KB. Trauma
nurse practitioner: front line approach to
274. Gwadry-Sridhar FH, Flintoft V, Lee DS, Lee H, management of the pediatric trauma patient in
Guyatt GH. A systematic review and meta-analysis of an academic program. Journal of Trauma Nursing
studies comparing readmission rates and mortality 2004;11:157–63.
rates in patients with heart failure. Archives of
Internal Medicine 2004;164:2315–20. 286. Lau FL, Leung KP. Waiting time in an urban
accident and emergency department – a way to
275. Devereaux PJ, Choi PTL, Lacchetti C, Weaver B, improve it. Journal of Accident and Emergency Medicine
Schunemann HJ, Haines T, et al. A systematic review 1997;14:299–301.
and meta-analysis of studies comparing mortality
rates of private for-profit and private not-for-profit 287. George S, Read S, Westlake L, Williams B,
hospitals. Canadian Medical Association Journal Fraser-Moodie A, Pritty P. Evaluation of nurse triage
2002;166:1399–406. in a British accident and emergency department.
British Medical Journal 1992;304:876–8.
276. Bahtsevani C, Uden G, Willman A. Outcomes
of evidence-based clinical practice guidelines: a 288. Sparr LF, Ruud DH, Hickam DH, Cooney TG.
systematic review. International Journal of Technology The effect of house officer rotation on inpatient
Assessment in Health Care 2004;20:427–33. satisfaction and ward atmosphere: preliminary
findings. Military Medicine 1994;159:47–53.
277. Sinuff T, Kahnamoui K, Cook DJ, Luce JM,
Levy MM. Rationing critical care beds: a systematic 289. Barry TL, Davis DJ, Meara JG, Halvorson M. Case
review. Critical Care Medicine 2004;32:1588–97. management: an evaluation at Childrens Hospital
Los Angeles. Nursing Economics 2002;20:22–7.
278. West E. Management matters: the link between
hospital organisation and quality of patient care. 290. Sakr M, Kendall R, Angus J, Sanders A, Nicholl J,
Quality in Health Care 2001;10:40–8. Wardrope J. Emergency nurse practitioners: a
three part study in clinical and cost effectiveness.
279. Mitchell PH, Shortell SM. Adverse outcomes and Emergency Medicine Journal 2003;20:158–63.
variations in organization of care delivery. Medical
Care 1997;35:NS19–32. 291. Atwood J, Hinshaw A, Gerber R. Professional
Occupational Nurse Job Satisfaction Scale. Tucson, AZ:
280. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Tucson College of Nursing, University of Arizona;
Stroup DF. Improving the quality of reports of 1988.
meta-analyses of randomized controlled trials: the
QUOROM statement. Lancet 1999;357:1896–900. 292. Goldstone L, Ball J, Collier M. Monitor: an index
of the quality of nursing care for acute medical and
281. McEvoy GM, Cascio WF. Strategies for reducing surgical wards. North West Nurse Staffing Project
employee turnover: a meta-analysis. Journal of Report. Newcastle upon Tyne: Newcastle upon Tyne
Applied Psychology 1985;70:342–53. Polytechnic Products; 1984.
282. Burns SM, Earven S, Fisher C, Lewis R, Merrell P, 293. Speedling EJ, Morrison B, Rehr H, Rosenberg G.
Schubart J, et al. Implementation of an institutional Patient satisfaction surveys: closing the gap between
program to improve clinical and financial outcomes provider and consumer. Quality Review Bulletin
of mechanically ventilated patients: one-year 1983;9:224–8.
outcomes and lessons learned. Critical Care Medicine
2003;31:2752–63. 294. Moos RH. Evaluating treatment environments: a social
ecological approach. New York, NY: Wiley; 1974.
283. Russell D, Vorderbruegge M, Burns SM. Effect
of an outcomes-managed approach to care 295. Lankshear AJ, Sheldon TA, Maynard A. Nurse
of neuroscience patients by acute care nurse staffing and healthcare outcomes: a systematic
260
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
review of the international research evidence. 307. Beglinger J. Quantifying patient care intensity: an
Advances in Nursing Science 2005;28:163–74. evidence-based approach to determining staffing
requirements. Nursing Administration Quarterly
296. Tourangeau AE, Giovannetti P, Tu JV, Wood M. 2006;30:193–202.
Nursing-related determinants of 30-day mortality
for hospitalized patients. Canadian Journal of 308. Dimick JB, Swoboda SM, Pronovost PJ, Lipsett PA.
Nursing Research 2002;33:71–88. Effect of nurse-to-patient ratio in the intensive care
unit on pulmonary complications and resource use
297. Sovie MD, Jawad AF. Hospital restructuring and after hepatectomy. American Journal of Critical Care
its impact on outcomes: nursing staff regulations 2001;10:376–82.
are premature. Journal of Nursing Administration
2001;31:588–600. 309. Amaravadi R, Dimick JB, Pronovost PJ,
Lipsett PA. ICU nurse-to-patient ratio is
298. Dexter F, Epstein RH, Marcon E, de Matta R. associated with complications and resource use
Strategies to reduce delays in admission into a after esophagectomy. Intensive Care Medicine
postanesthesia care unit from operating rooms. 2000;26:1857–62.
Journal of Perianesthesia Nursing 2005;20:92–102.
310. Callaghan LA, Cartwright DW, O’Rourke P,
299. Brooks RG, Walsh M, Mardon RE, Lewis M, Davies MW. Infant to staff ratios and risk of
Clawson A. The roles of nature and nurture in mortality in very low birthweight infants. Archives
the recruitment and retention of primary care of Disease in Childhood Fetal and Neonatal Edition
physicians in rural areas: a review of the literature. 2003;88:F94–97.
Academic Medicine 2002;77:790–8.
311. Stilwell J, Szczepura A, Mugford M. Factors affecting
300. Baltes BB, Briggs TE, Huff JW, Wright JA, the outcome of maternity care. 1. Relationship
Neuman GA. Flexible and compressed workweek between staffing and perinatal deaths at the hospital
schedules: a meta-analysis of their effects on of birth. Journal of Epidemiology and Community Health
work-related criteria. Journal of Applied Psychology 1988;42:157–69.
1999;84:496–513.
312. Marcin JP, Rutan E, Rapetti PM, Brown JP,
301. Landrigan CP, Rothschild JM, Cronin JW, Rahnamayi R, Pretzlaff RK. Nurse staffing and
Kaushal R, Burdick E, Katz JT, et al. Effect of unplanned extubation in the pediatric intensive care
reducing interns’ work hours on serious medical unit. Pediatric Critical Care Medicine 2005;6:254–7.
errors in intensive care units. New England Journal of
Medicine 2004;351:1838–48. 313. Bourbonnais R, Vinet A, Meyer F, Goldberg M.
Certified sick leave and work load. A case referent
302. Poulose BK, Ray WA, Arbogast PG, Needleman J, study among nurses. Journal of Occupational Medicine
Buerhaus PI, Griffin MR, et al. Resident work 1992;34:69–74.
hour limits and patient safety. Annals of Surgery
2005;241:847–56. 314. Bailit JL, Blanchard MH. The effect of house
staff working hours on the quality of obstetric
303. Howard DL, Silber JH, Jobes DR. Do regulations and gynecologic care. Obstetrics and Gynecology
limiting residents’ work hours affect patient 2004;103:613–16.
mortality? Journal of General Internal Medicine
2004;19:1–7. 315. Bollschweiler E, Krings A, Fuchs K-H, Pistorius G,
Bein T, Otto U, et al. Alternative shift models and
304. Bolton LB, Aydin CE, Donaldson N, Brown DS, the quality of patient care. Langenbeck’s Archives of
Nelson MS, Harms D. Nurse staffing and patient Surgery 2001;386:104–9.
perceptions of nursing care. Journal of Nursing
Administration 2003;33:607–14. 316. Laine C, Goldman L, Soukup JR, Hayes JG. The
impact of a regulation restricting medical house
305. Potter P, Barr N, McSweeney M, Sledge J. staff working hours on the quality of patient
Identifying nurse staffing and patient outcome care. Journal of the American Medical Association
relationships: a guide for change in care delivery. 1993;269:374–8.
Nursing Economics 2003;21:158–66.
317. Burroughs TE, Davies AR, Cira J, Dunagan WC.
306. Davydov L, Caliendo G, Mehl B, Smith LG. Understanding patient willingness to recommend
Investigation of correlation between house-staff and return: a strategy for prioritizing improvement
work hours and prescribing errors. American Journal opportunities. Joint Commission. Journal on Quality
of Health-System Pharmacy 2004;61:1130–4. Improvement 1999;25:271–87.
261
318. Sibbald B, Shen J, McBride A. Changing the skill- impact trauma systems. Journal of Trauma: Injury
mix of the health care workforce. Journal of Health Infection and Critical Care 2005;58:917–20.
Services Research and Policy 2004;9:28–38.
330. Grouse A, Bishop R. Non-medical technicians
319. Fletcher KE, Underwood W III, Davis SQ, reduce emergency department waiting times.
Mangrulkar RS, McMahon LF Jr, Saint S. Effects of Emergency Medicine 2001;13:66–9.
work hour reduction on residents’ lives: a systematic
review. Journal of the American Medical Association 331. Aigner MJ, Drew S, Phipps J. A comparative study
2005;294:1088–100. of nursing home resident outcomes between care
provided by nurse practitioners/physicians versus
320. Lang TA, Hodge M, Olson V, Romano PS, physicians only. Journal of the American Medical
Kravitz RL. Nurse–patient ratios: a systematic Directors Association 2004;5:16–23.
review on the effects of nurse staffing on patient,
nurse employee, and hospital outcomes. Journal of 332. Barr M, Johnston D, McConnell D. Patient
Nursing Administration 2004;34:326–37. satisfaction with a new nurse practitioner service.
Accident and Emergency Nursing 2000;8:144–7.
321. Laurant M, Reeves D, Hermens R, Braspenning J,
Grol R, Sibbald B. Substitution of doctors by nurses 333. Green A, Edmonds L. Bridging the gap between
in primary care. The Cochrane Database of Systematic the intensive care unit and general wards – the
Reviews 2004;4:CD001271. ICU liaison nurse. Intensive Critical Care Nursing
2004;20:133–43.
322. Pronovost PJ, Angus DC, Dorman T, Robinson KA,
Dremsizov TT, Young TL. Caring for the critically 334. Person SD, Allison JJ, Kiefe CI, Weaver MT,
ill patient. Physician staffing patterns and clinical Williams OD, Centor RM, et al. Nurse staffing
outcomes in critically ill patients: a systematic and mortality for Medicare patients with acute
review. Journal of the American Medical Association myocardial infarction. Medical Care 2004;42:4–12.
2002;288:2151–62.
335. Li TC, Phillips MC, Shaw L, Cook EF,
323. Spisso J, O’Callaghan C, McKennan M, Holcroft JW. Natanson C, Goldman L. On-site physician staffing
Improved quality of care and reduction of housestaff in a community hospital intensive care unit.
workload using trauma nurse practitioners. Journal Impact on test and procedure use and on patient
of Trauma 1990;30:660–5. outcome. Journal of the American Medical Association
1984;252:2023–7.
324. Partovi S, Nelson B, Bryan ED, Walsh MJ. Faculty
triage shortens emergency department length of 336. Hinshaw AS, Atwood JR. A patient satisfaction
stay. Academic Emergency Medicine 2001;10:990–5. instrument: precision by replication. Nursing
Research 1982;31:170–5, 191.
325. Topeli A, Laghi F, Tobin MJ. Effect of closed unit
policy and appointing an intensivist in a developing 337. Hall LM, Doran D, Baker GR, Pink GH, Sidani S,
country. Critical Care Medicine 2005;33:299–306. O’Brien-Pallas L, et al. Nurse staffing models as
predictors of patient outcomes. Medical Care
326. Reynolds HN, Haupt MT, Thill-Baharozian MC, 2003;9:1096–109.
Carlson RW. Impact of critical care physician
staffing on patients with septic shock in a university 338. Lee T, Yeh M, Chen H, Lien G. The skill mix
hospital medical intensive care unit. Journal of the practice model for nursing: measuring outcome.
American Medical Association 1988;260:3446–50. Journal of Advanced Nursing 2005;51:406–13.
327. Bucheli B, Martina B. Reduced length of stay 339. Fontaine K, Rositani R. Cost, quality and satisfaction
in medical emergency department patients: with hospice after-hours care. The Hospice Journal
a prospective controlled study on emergency 2000;15:1–14.
physician staffing. European Journal of Emergency
Medicine 2004;11:29–34. 340. Barkell NP, Killinger KA, Schultz SD. The
relationship between nurse staffing models and
328. Pepe PE, Mattox KL, Duke JH, Fisher PB, patient outcomes: a descriptive study. Outcomes
Prentice FD. Effect of full-time, specialized Management 2002;6:27–33.
physician supervision on the success of a large,
urban emergency medical services system. Critical 341. Bolton LB, Jones D, Aydin CE, Donaldson N,
Care Medicine 1993;21:1279–86. Brown DS, Lowe M, et al. A response to California’s
mandated nursing ratios. Journal of Nursing
329. Christmas AB, Reynolds J, Hodges S, Franklin GA, Scholarship 2001;33:179–84.
Miller FB, Richardson JD, et al. Physician extenders
262
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
342. Oswanski MF, Sharma OP, Raj SS. Comparative 354. Arnow P, Allyn PA, Nichols EM, Hill DL,
review of use of physician assistants in a level I Pezzlo M, Bartlett RH. Control of methicillin-
trauma center. American Surgeon 2004;70:272–9. resistant Staphylococcus aureus in a burn unit: role
of nurse staffing. Journal of Trauma 1982;22:954–9.
343. Simmer TL, Nerenz DR, Rutt WM, Newcomb CS,
Benfer DW. A randomized, controlled trial of an 355. O’Connor AE, Lockney AL, Sloan PD, McGrail MR.
attending staff service in general internal medicine. Does the presence of an emergency physician
Medical Care 1991;2:JS31–40. improve access based quality indicators in a
rural emergency department? Emergency Medicine
344. Tenner PA, Dibrell H, Taylor RP. Improved survival Australasia 2004;16:55–8.
with hospitalists in a pediatric intensive care unit.
Critical Care Medicine 2003;31:847–52. 356. French D, Zwemer FL Jr, Schneider S. The effects
of the absence of emergency medicine residents in
345. Petersen LA, Brennan TA, O’Neill AC, Cook FE, an academic emergency department. Academic and
Lee TH. Does housestaff discontinuity of care Emergency Medicine 2002;9:1205–10.
increase the risk for preventable adverse events?
Annals of Internal Medicine 1994;121:866–72. 357. Porter JM, Ursic C. Trauma attending in the
resuscitation room: does it affect outcome? American
346. Fairchild DG, McLoughlin KS, Gharib S, Horsky J, Surgeon 2001;67:611–14.
Portnow M, Richter J, et al. Productivity, quality, and
patient satisfaction: comparison of part-time and 358. Multz AS, Chalfin DB, Samson IM, Dantzker DR,
full-time primary care physicians. Journal of General Fein AM, Steinberg HN, et al. A “closed” medical
Internal Medicine 2001;16:663–7. intensive care unit (MICU) improves resource
utilization when compared with an “open” MICU.
347. Hoffman LA, Tasota FJ, Zullo TG, Scharfenberg C, American Journal of Respiratory and Critical Care
Donahoe MP. Outcomes of care managed by an Medicine 1998;157:1468–73.
acute care nurse practitioner/attending physician
team in a subacute medical intensive care unit. 359. Baldock G, Foley P, Brett S. The impact of
American Journal of Critical Care 2005;14:121–32. organisational change on outcome in an intensive
care unit in the United Kingdom. Intensive Care
348. Terris J, Leman P, O’Connor N, Wood R. Making an Medicine 2001;27:865–72.
IMPACT on emergency department flow: improving
patient processing assisted by consultant at triage. 360. Ghorra S, Reinert SE, Cioffi W, Buczko G, Simms H.
Emergency Medicine Journal 2004;21:537–41. Analysis of the effect of conversion from open to
closed surgical intensive care unit. Annals of Surgery
349. Rudy EB, Davidson LJ, Daly JM, Clochesy JM, 1999;229:163–71.
Sereika SM, Baldisseri M, et al. Care activities and
outcomes of patients cared for by acute care nurse 361. Weitzel T, Robinson SB. A model of nurse assistant
practitioners, physician assistants, and resident care to promote functional status in hospitalized
physicians: a comparison. American Journal of Critical elders. Journal for Nurses in Staff Development
Care 1998;7:267–81. 2004;20:181–6.
350. McCausland MP, Persing RJ, Kiley M. Primary 362. Mayo PH, Weinberg BJ, Kramer B, Richman J,
nursing in a psychiatric setting. Nursing Economics Seibert-Choi O, Rosen MJ. Results of a program
1988;6:297–301. to improve the process of inpatient care of adult
asthmatics. Chest 1996;110:48–52.
351. Mark B, Harless D, McCue M. The impact of
HMO penetration on the relationship between 363. Davis D, Thomson O’Brien MA, Freemantle N,
nurse staffing and quality of care. Health Economics Wolf FM, Mazmanian P, Taylor-Vaisey A. Impact
2005;14:373–53. of formal continuing medical education. Do
conferences, workshops, rounds, and other
352. Mark B, Harless D, McCue M, Xu X. A longitudinal traditional continuing education activities change
examination of hospital registered nurse staffing physician behavior or health care outcomes? Journal
and quality of care. Health Services Research of the American Medical Association 1999;282:867–74.
2004;39:277–98.
364. Davis DA, Thomson MA, Oxman AD, Haynes RB.
353. Robert J, Fridkin S, Blumberg H, Anderson B, Evidence for the effectiveness of CME. A review
White N, Ray SM, et al. The influence of the of fifty randomised controlled trials. Journal of the
composition of the nursing staff on primary American Medical Association 1992;268:1111–17.
bloodstream infection rates in a surgical inensive
care unit. Infection Control and Hospital Epidemiology 365. Davis DA, Thomson MA, Oxman AD, Haynes RB.
2000;21:7–12. Changing physician performance: a systematic
263
review of continuing medical education strategies. 376. Rosenthal VD, Guzman S, Pezzotto SM, Crnich CJ.
Journal of the American Medical Association Effect of an infection control program using
1995;274:700–5. education and performance feedback on rates
of intravascular device-associated bloodstream
366. Hunter E, Nicol M. Systematic review: evidence of infections in intensive care units in Argentina.
the value of continuing professional development to American Journal of Infection Control 2003;31:405–9.
enhance recruitment and retention of occupational
therapists in mental health. British Journal of 377. Salahuddin N, Zafar A, Sukhyani L, Rahim S,
Occupational Therapy 2002;65:207–15. Noor MF, Hussain K, et al. Reducing ventilator-
associated pneumonia rates through a staff
367. Stajkovic AD, Luthans F. A meta-analysis of the education programme. Journal of Hospital Infection
effects of organizational behavior modification on 2004;57:223–7.
task performance, 1975–95. Academy of Management
Journal 1997;40:11–22. 378. Zack JE, Garrison T, Trovillion E, Clinkscale D,
Coopersmith CM, Fraser VJ, et al. Effect of an
368. Allen TD, Eby LT, Poteet ML, Lentz E, Lima L. education program aimed at reducing the occurence
Career benefits associated with mentoring for of ventilator-associated pneumonia. Critical Care
protegee: a meta-analysis. Journal of Applied Medicine 2002;30:2407–12.
Psychology 2004;89:127–36.
379. Lawrence SJ, Shadel BN, Leet TL, Hall JB,
369. Arthur J, Day EA, McNelly TL, Edens PS. A Mundy LM. An intervention to improve antibiotic
meta-analysis of the criterion-related validity of delivery and sputum procurement in patients
assessment center dimensions. Personnel Psychology hospitalized with community-acquired pneumonia.
2003;56:125–54. Chest 2002;122:913–19.
370. Fidler V, Nap R, Miranda DR. The effect of a 380. Saunders DG, Lynch AB, Grayson M, Linz D.
managerial-based intervention on the occurrence The inventory of beliefs about wife beating: the
of out-of-range-measurements and mortality construction and initial validation of a measure
in intensive care units. Journal of Critical Care of beliefs and attitudes. Violence and Victims
2004;19:130–4. 1987;2:39–57.
371. Lundstrom M, Edlund A, Karlsson S, Brannstrom B, 381. Soeken K, Parker B, McFarlane J, Lominak MC.
Bucht G, Gustafson Y. A multifactorial intervention The Abuse Assessment Screen: a clinical instrument
program reduces the duration of delirium, length to measure frequency, severity, and perpetrator of
of hospitalization, and mortality in delirious abuse against women. In Campbell J (ed.). Beyond
patients. Journal of the American Geriatric Society diagnosis: changing the health care response to battered
2005;53:622–8. women and their children. Newbury Park, CA: Sage;
1998. pp 195–203.
372. Cohran J, Larson E, Roach H, Blane C, Pierce P.
Effect of intravascular surveillance and education 382. Gosden T, Forland F, Kristiansen IS, Sutton M,
program on rates of nosocomial bloodstream Leese B, Giuffrida A. Impact of payment method on
infections. Heart and Lung 2006;25:161–4. behaviour of primary care physicians: a systematic
review. Journal of Health Services and Research Policy
373. Kelleghan SI, Salami C, Padilla S, McCord M, 2001;6:44–55.
Mermilliod G, Canola RN, et al. An effective
continuous quality improvement approach to the 383. Chaix-Couturier C, Durand-Zaleski I, Jolly D,
prevention of ventilator-associated pneumonia. Durieux P. Effects of financial incentives on medical
American Journal of Infection Control 1993;21:322–30. practice: results from a systematic review of the
literature and methodological issues. International
374. Rosenthal VD, Guzman S, Safdar N. Effect of Journal for Quality in Health Care 2000;12:133–42.
education and performance feedback on rates
of catheter-associated urinary tract infection in 384. Gosden T, Forland F, Kristiansen IS, Sutton M,
intensive care units in Argentina. Infection Control Leese B, Giuffrida A. Impact of payment method on
and Hospital Epidemiology 2004;25:47–50. behaviour of primary care physicians: a systematic
review. Journal of Health Services and Research Policy
375. Won S, Chou H, Hsieh W, Chen C, 2001;6:44–55.
Huang S, Tsou K, et al. Handwashing program
for the prevention of nosocomial infections in a 385. Gosden T, Pedersen L, Torgerson D. How should we
neonatal intensive care unit. Infection Control and pay doctors? A systematic review of salary payments
Hospital Epidemiology 2004;25:742–6. and their effect on doctor behaviour. Quartely
Jouurnal of Medicine 1999;92:47–55.
264
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
386. Steiner A, Robinson R. Managed care: US research 398. Kyriacou DN, Ricketts V, Dyne PL, McCollough MD,
evidence and its lessons for the NHS. Journal of Talan DA. A 5-year time study analysis of emergency
Health Services and Research Policy 1998;3:173–84. department patient care efficiency. Annals of
Emergency Medicine 1999;34:326–35.
387. Sempowski IP. Effectiveness of financial incentives
in exchange for rural and underserviced area 399. Leininger SM. One year later: did the quality
return-of-service commitments: systematic review circle of geriatric hip fracture care achieve quality
of the literature. Canadian Journal of Rural Medicine outcomes? Nursing Case Management 1999;4:263–7.
2004;9:82–88.
400. Smither JW, London M, Reilly RR. Does
388. Giuffrida A, Gosden T, Forland F, Kristiansen IS, performance improve following multisource
Sergison M, Leese B, et al. Target payments in feedback: A theoretical model, meta-analysis, and
primary care: effects on professional practice and review of empirical findings. Personnel Psychology
health care outcomes. Cochrane Database of Systematic 2005;58:33–66.
Reviews 1999;4:CD000531.
401. Rodgers R, Hunter JE. Impact of management by
389. De Groot T, Kiker DS. A meta-analysis of the non- objectives on organizational productivity. Journal of
monetary effects of employee health management Applied Psychology 1991;76:322–36.
programs. Human Resource Management
2003;42:53–69. 402. Kraiger K, Ford JK. The relation of job knowledge,
job performance, and supervisory ratings as
390. Stajkovic AD, Luthans F. Behavioral management a function of ratee race. Human Performance
and task performance in organizations: conceptual 1990;3:269–79.
background, meta-analysis, and test of alternative
models. Personnel Psychology 2003;56:155–94. 403. Heneman RL. The relationship between
supervisory ratings and results-oriented measures
391. Doucouliagos C. Worker participation and of performance: a meta-analysis. Personnel Psychology
productivity in labor-managed and participatory 1986;39:811–26.
capitalist firms: a meta-analysis. Industrial and Labor
Relations Review 1995;49:58–77. 404. Appelbaum E, Blatt R. The new American workplace.
Ithaca, NY, ILR Press; 1994.
392. Wiersma UJ. The effects of extrinsic rewards
in intrinsic motivation: a meta-analysis. Journal 405. Ichniowski C, Kochan T, Levine D, Strauss G. What
of Occupational and Organizational Psychology works at work: overview and assessment. Industrial
1992;65:101–14. Relations 1996;35:299–333.
393. Forsberg E, Axelsson R, Arnetz B. Financial 406. Goetz AM, Kedzuf S, Wagener M, Muder RR.
incentives in health care. The impact of Feedback to nursing staff as an intervention to
performance-based reimbursement. Health Policy reduce catheter-associated urinary tract infections.
2001;58:243–62. American Journal of Infection Control 1999;27:402–4.
394. Lee S, Cowie S, Slobodian P. Payment by salary or 407. Nettleman MD, Antoni T, Fredrickson M,
fee-for-service. Effect on health care resource use Pfaller M. Assigning responsibility: using feedback
in the last year of life. Canadian Family Physician to achieve sustained control of methicillin-resistant
1999;45:2091–6. Staphylococcus aureus. The American Journal of Medicine
1991;91:S228–32.
395. Armour BS, Friedman C, Pitts MM, Wike J, Alley L,
Etchason J. The influence of year-end bonuses on 408. Spierer M, Sims HW, Micklitsch CN, Lewis BE.
colorectal cancer screening. American Journal of Assessment of patient satisfaction as part of a
Management of Care 2004;10: 617–24. physician performance evaluation: the Fallon Clinic
experience. Journal of Ambulatory Care Management
396. Ruchholtz S, Waydhas C, Aufmkolk M, Tager G, 1994;17:1–7.
Piepenbrink K, Stolke D, et al. Interdisciplinary
quality management in the treatment of severely 409. Wagner J, Gooding R. Shared influence and
injured patients. Validation of a QM system for the organizational behavior: a meta-analysis of
diagnostic and therapeutic process in early clinical situational variables expected to moderate
management. Unfallchirurg 2001;104:927–37. participation–outcome relationships. Academy of
Management Journal 1987;30:524–41.
397. Forster DH, Krause G, Gastmeier P, Ebner W,
Rath A, Wischnewski N, et al. Can quality circles 410. Zottoli M, Wanous J. Recruitment source research:
improve hospital-acquired infection control? Journal current status and future directions. Human Resource
of Hospital Infection 2000;45:302–10. Management Review 2000;10:353–382.
265
411. Phillips J. Effects of realistic job previews on 425. Moorman RH, Blakely GL, Niehoff BP. Does
multiple organizational outcomes: a meta-analysis. perceived organizational support mediate the
Academy of Management Journal 1998;41:673–90. relationship between procedural justice and
organizational citizenship behavior? Academy of
412. Hackman JR, Oldham G. Motivation through the Management Journal 1998;41:351–7.
design of work: Test of a theory. Organizational
Behaviour and Human Performance 1976;16:250–79. 426. Coupland C, Blyton P, Bacon N. A longitudinal
study of the influence of shop floor work teams
413. Wall TD, Clegg CW. A longitudinal field study on expressions of ‘us’ and ‘them’. Human Relations
of group work redesign. Journal of Occupational 2005;58:1055–81.
Behavior 1981;2:31–49.
427. Campbell DJ. Task complexity: a review
414. Armstrong-Stassen M, Wagar TH, Cattaneo RJ. and analysis. Academy of Management Review
Work-group membership (in)stability and survivors’ 1988;13:40–52.
reactions to organizational downsizing. Journal of
Applied Social Psychology 2004;34:2023–44. 428. Schaubroeck J, Ganster DC, Kemmerer BE. Job
complexity, ‘type A’ behavior, and cardiovascular
415. Wall TD, Kemp NJ, Jackson PR, Clegg CW. disorder: a prospective study. Academy of Management
Outcome of autonomous workgroups: a long-term Journal 1994;37:426.
field experiment. Academy of Management Journal
1986;29:280. 429. Kalimo R, Pahkin K, Mutanen P,
Toppinen-Tanner S. Staying well or burning out at
416. Pearson CAL. Autonomous workgroups: an work: work characteristics and personal resources as
evaluation at an industrial site. Human Relations long-term predictors. Work and Stress 2003;17:109.
1992;45:905–36.
430. Elo A-L, Leppanen A, Lindstrom K,
417. Banker RD, Field JM, Schroeder RG, Ropponen T. OSQ. Occupational Stress Questionnaire:
Sintia KK. Impact of work teams on manufacturing user’s instructions. Helsinki: Institute of Occupational
performance: a longitudinal field study. Academy of Health; 1992.
Management Journal 1996;39:867.
431. Hackman JR, Oldham G. Work redesign. Reading,
418. Cordery JL, Mueller WS, Smith LM. Attitudinal and MA: Addison-Wesley; 1980.
behavioral effects if autonomous group working:
a longitudinal field study. Academy of Management 432. House JS. Occupational stress and the mental and
Journal 1991;34:464. physical health of factory workers. Ann Arbor, MI:
Institute for Social Research; 1980.
419. Kirchmeyer C. Demographic similarity to the work
group: a longitudinal study of managers at the 433. Karasek R. Job demands, job decision latitude,
early career stage. Journal of Organizational Behavior and mental strain: implications for job redesign.
1995;16:67–83. Administrative Science Quarterly 1979;24:285–307.
420. Goldberg DP. The detection of psychiatric illness by 434. De Langeff AH, Taris TW, Kompier MAJ,
questionnaire. Maudsley Monograph No. 21. Oxford: Houtman ILD, Bongers PM. The relationships
Oxford University Press; 1972. between work characteristics and mental health:
examining normal, reversed and reciprocal
421. Clegg CW, Wall TD. A note on some new scales for relationships in a 4-wave study. Work and Stress
measuring psychological well-being at work. Journal 2004;18:149–66.
of Occupational Psychology 1981;54:221–5.
435. Moyle P. Longitudinal influences of managerial
422. Hackman JR, Oldham GR. The Job Diagnostic Survey: support on employee well-being. Work and Stress
an instrument for the diagnosis of jobs and the evaluation 1998;12:29–49.
of job redesign projects. Technical Report No. 4. New
Haven, CT: Department of Administrative Sciences, 436. Moyle P, Parkes K. The effects of transition stress:
Yale University; 1974. a relocation study. Journal of Organizational Behavior
1999;20:625.
423. Jick TD. Process and impacts of a merger: individual and
organizational perspectives. Unpublished PhD thesis, 437. Smulders PGW, Nijhuis FJN. The Job Demands–
Cornell University, Ithaca , NY. Job Control Model and absence behaviour: results
of a 3-year longitudinal study. Work and Stress
424. Lodahl T, Kejner M. The definition and 1999;13:115–31.
measurement of job involvement. Journal of Applied
Psychology 1965;49:24–33.
266
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
438. Nielsen ML, Rugulies R, Christensen KB, with job demands: a longitudinal study. Academy of
Smith-Hansen L, Bjorner JB, Kristensen TS. Impact Management Proceedings 2000;OB:1–16.
of the psychosocial work environment on registered
absence from work: a two-year longitudinal 451. Smith CS, Tisak J, Hahn SE, Schnieder RA. The
study using the IPAW cohort. Work and Stress measurement of job control. Journal of Organizational
2004;18:323–35. Behavior 1997;18:225–37.
439. De Croon EM, Sluiter JK, Blonk RWB, 452. Maslach C, Jackson SE, Schwab RL. Maslach
Broersen JPJ, Frings-Dresen MHW. Stressful work, Burnout Inventory*/Educators Survey (MBI-
psychological job strain, and turnover: a 2-year ES). In Maslach C, Jackson SE, Leiter MP (eds).
prospective cohort study of truck drivers. Journal of Maslach Burnout Inventory. 3rd Edn. Palo Alto, CA:
Applied Psychology 2004;89:442–54. Consulting Psychologists Press; 1996.
440. Holman D, Wall TD. Work characteristics, 453. Parker SK, Wall TD. Job and work design: organizing
learning-related outcomes, and strain: a test of work to promote well-being and effectiveness. London:
competing direct effects, mediated, and moderated Sage Publications; 1998.
models. Journal of Occupational Health Psychology
2002;7:283–301. 454. Parker SK. Enhancing role breadth self-efficacy: the
roles of job enrichment and other organizational
441. Taris TW, Kompier MAJ, De Lange AH, interventions. Journal of Applied Psychology
Schaufeli WB, Schreurs PJ. Learning new behaviour 1993;83:835–52.
patterns: a longitudinal test of Karasek’s active
learning hypothesis among Dutch teachers. Work 455. Campion MA, McClelland CL. Follow-up and
and Stress 2003;17:1. extension of the interdisciplinary costs and benefits
of enlarged jobs. Journal of Applied Psychology
442. Johnson JV, Stewart W, Hall EM, Fredlund P, 1993;78:339–51.
Theorell T. Long-term psychosocial work
environment and cardiovascular mortality among 456. Rissen D, Melin B, Sandsjo L, Dohns I,
Swedish men. American Journal of Public Health Lundberg U. Psychophysiological stress reactions,
1996;86:324–31. trapezius muscle activity, and neck and shoulder
pain among female cashiers before and after
443. O’Brien GE. Psychology of work and unemployment. introduction of job rotation. Work and Stress
Chichester: Wiley; 1986. 2002;16:127–37.
444. Axtell C, Holman D, Wall T, Waterson P, 457. Blocher E. Performance Effects of different audit
Harrington E, Unsworth K. Shopfloor innovation: staff assignment strategies. Accounting Review
facilitating the suggestion and implementation 1979;54:563–73.
of ideas. Journal of Occupational and Organisational
Psychology 2000;73:265–85. 458. Axtell CM, Parker SK. Promoting role breadth self-
efficacy through involvement, work redesign and
445. Warr PB. The measurement of well-being and other training. Human Relations 2003;56:113.
aspects of mental health. Journal of Occupational
Psychology 1990;63:193–210. 459. Laschinger HKS, Finegan JE, Shamian J, Wilk P.
A longitudinal analysis of the impact of workplace
446. Setterlind S, Larsson G. The stress profile: a empowerment on work satisfaction. Journal of
psychosocial approach to measuring stress. Stress Organizational Behavior 2004;25:527–45.
Medicine 1995;11:85–92.
460. Orpen C. The effects of job enrichment on
447. Daniels K, Guppy A. Occupational stress, social employee satisfaction, motivation, involvement, and
support, job control, and psychological well-being. performance: a field experiment. Human Relations
Human Relations 1994;479:1523–44. 1979;32:189.
448. Goldberg D, Williams P. A user’s guide to the General 461. Hall DT, Goodale JG, Rabinowitz S, Morgan MA.
Health Questionnaire: GHQ. Windsor, ON: NFER- Effects of top-down departmental and job change
Nelson; 1988. upon perceived employee behavior and attitudes: a
natural field experiment. Journal of Applied Psychology
449. Parkes KR. Human factors, shift work, and alertness in 1978;63:62–72.
the offshore oil industry. Report OTH 92–389. London:
HMSO; 1993. 462. Bhagat RS, Chassie MB. Effects of job changes
in job characteristics on some theory-specific
450. Jia LX, Schaubroeck J, Lam SSK. Psychological and attitudinal outcomes: results from a naturally
immunological effects of worker control on coping
267
occurring quasi-experiment. Human Relations 475. Warr P, Cook J, Wall T. Scales for the measurement
1980;33:297. of some work attitudes and aspects of psychological
well-being. Journal of Occupational Psychology
463. Griffin RW. Objective and social sources of 1979;52:129–48.
information in task redesign: a field experiment.
Administrative Science Quarterly 1983;28:184–200. 476. Van Veldhoven M, Meijman T. [Measuring
psychosocial workload by means of a questionnaire.
464. Griffin RW. Effects of work redesign on employee Questionnaire on the Experience and Evaluation of
perceptions, attitudes, and behaviors: a long- Work]. Amsterdam: NIA; 1994.
term investigation. Academy of Management Journal
1991;34:425–435. 477. Meyer JP, Allen NJ. Testing the “side-bet theory” of
organizational commitment: some methodological
465. Morgeson FP, Campion MA. Minimizing considerations. Journal of Applied Psychology
tradeoffs when redesigning work: evidence from a 1984;69:372–78.
longitudinal quasi-experiment. Personnel Psychology
2002;55:589–612. 478. Derogatis LR. The SCL–90–R. Baltimore, MD:
Clinical Psychometrics; 1975.
466. Locke EA, Sirota D, Wolfson AD. An experimental
case study of the successes and failures of job 479. Barnett RC, Marshall NL, Raudenbush S,
enrichment in a government agency. Human Brennan RT. Gender and the relationship between
Relations 1976;32:701–11. job experiences and psychological distress: a study
of dual-earner couples. Journal of Personality and
467. Houkes I, Janssen PP M, de Jonge J, Bakker AB. Social Psychology 1993;64:794–806.
Specific determinants of intrinsic work motivation,
emotional exhaustion and turnover intention: 480. Rooijen L. van Widows’ bereavement: stress
a multisample longitudinal study. Journal and depression after 1.5 years. In Sarason IG,
of Occupational and Organizational Psychology Spielberger CD (eds). Stress and anxiety. Vol. 6.
2003;76:427–50. Washington, DC: Hemisphere; 1979.
468. Meyer JP, Bobocel DR, Allen NJ. Development 481. Lubin B. Adjective checklists for the measurement
of organizational commitment during the first of depression. Archives of General Psychiatry
year of employment: a longitudinal study of pre- 1965;12:57–62.
and post-entry influences. Journal of Management
1991;17:716. 482. Saks AM, Ashforth BE. The role of dispositions,
entry stressors, and behavioral plasticity theory in
469. Hammer TH, Landau JC, Stern RN. Absenteeism predicting. Journal of Organizational Behavior 2000;
when workers have a voice: the case of employee 21:43.
ownership. Journal of Applied Psychology
1981;66:561–73. 483. Fried Y, Slowik LH, Shperling Z, Franz C,
Ben-David HA, Avital N, et al.. The moderating
470. Barnett RC, Brennan RT, Raudenbush SW, effect of job security on the relation between role
Pleck JH, Marshall NL. Change in job and marital clarity and job performance: a longitudinal field
experience and change in psychological distress: a study. Human Relations 2003;56:787–805.
longitudinal study of dual-earner couples. Journal of
Personality and Social Psychology 1995;69:839–50. 484. Saks and Ashforth
471. Taris TW, Bok IA, Calje DG. On the relation 485. Cammann C, Fichman M, Jenkins GD Jr,
between job characteristics and depression: a Klesh JR. Assessing the attitudes and perceptions
longitudinal study. International Journal of Stress of organizational members. In Seashore SE,
Management 1998;5:157–67. Lawler EE III, Mirvis PH, Cammann C (eds).
Assessing organizational change: a guide to methods,
472. Sims HP Jr, Szilagyi AD. Job characteristic measures, and practices. New York, NY: Wiley; 1983.
relationships: individual and structural moderators. pp. 71–138.
Organizational Behavior and Human Performance
1976;17:211–30. 486. Allen NJ, Meyer JP. The measurement and
antecedents of affective, continuance and normative
473. Weiss DJ, Dawis RV, England GW, Lofquist LH. commitment to the organization. Journal of
Manual for the MSQ. Minneapolis, MN: University of Occupational Psychology 1990;63:1–18.
Minnesota; 1967.
487. Mael F, Ashforth BE. Alumni and their alma
474. Seashore SE, Lawler EE III, Mirvis P, Cammann C. mater: a partial test of the reformulated model
Observing and measuring organizational change: a guide of organizational identification. Journal of
268
to field practice. New York: Wiley; 1982. Organizational Behavior 1992;13:103–23.
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
488. Colarelli SM. Methods of communication and 501. Maertz J, Bauer TN, Mosley J, Posthuma RA,
mediating processes in realistic job previews. Journal Campion MA, Talya NBT. Do procedural justice
of Applied Psychology 1984;69:633–42. perceptions in a selection testing context predict
applicant attraction and intention toward the
489. Ashforth BE. The experience of powerlessness in organization? Journal of Applied Social Psychology
organizations. Organizational Behavior and Human 2004;34:125–45.
Decision Processes 1989;43:207–42.
502. Porter LW, Smith FJ. The etiology of organizational
490. Patchen M. Participation, achievement and involvement commitment. Unpublished manuscript, University of
on the job. Englewood Cliffs, NJ: Prentice-Hall; 1970. California at Irvine, Irvine, CA; 1970.
491. House RJ, Levanoni E, Schuler RS. An empirical 503. Rizzo JR, House RJ, Lirtzman SI. Role conflict and
examination of the construct validity of the rizzo, ambiguity in complex organizations. Administrative
house and lirtzman role scales: toward a clarification Science Quarterly 1970;15:150–63.
of the nature of role conflict. Paper presented at
the Annual Mid-West Meeting of the Academy of 504. Macan TH, Avedon MJ, Paese M, Smith DE. The
Management, Columbus, OH, 1982. effects of applicants’ reactions to cognitive ability
tests and an assessment center. Personnel Psychology
492. Bravo MJ, Peiro JM, Rodriguez I, Whitely WT. Social 1994;47:715–38.
antecedents of the role stress and career-enhancing
strategies of newcomers to organizations: a 505. Smither JW, Reilly RR, Millsap RE,
longitudinal study. Work and Stress 2003;17:195–217. Pearlman K, Stoffey RW. Applicant reactions
to selection procedures. Personnel Psychology
493. Bachman P, O’Maley P, Johnston J. Adolescence to 1993;46:49–76.
adult change and stability in the lives of young men. Ann
Arbor, MI: University of Michigan Press; 1978. 506. O’Reilly CA, Chatman J, Caldwell DF. People
and organizational culture: a Q-sort approach
494. Penley L, Gould S. Measuring career strategies: the to assessing person-organization fit. Academy of
psychometric characteristics of the Career Strategies Management Journal 1991;34:487–516.
Inventory. San Antonio, TX: Centre for Studies
in Business, Economics and Human Resources, 507. Kunin T. The construction of a new type of attitude
University of Texas; 1981. measure. Personnel Psychology 1955;8:65–78.
495. Bauer TN, Dolen MR, Maertz J, Campion MA. 508. Van Maanen J, Schein EH. Towards a theory of
Longitudinal assessment of applicant reactions to organizational socialization. In Staw BM (ed.).
employment testing and test outcome feedback. Research in organizational behavior. Greenwich, CT:
Journal of Applied Psychology 1998;83:892–903. JAI Press; 1979.
496. Castilla EJ. Social networks and employee 509. Jones GR. Socialization tactics, self-efficacy, and
performance in a call center. American Journal of newcomers’ adjustments to organizations. Academy of
Sociology 2005;110:1243–83. Management Journal 1986;29:262–79.
497. Meglino BM, DeNisi A, Youngblood SA, 510. Cooper-Thomas HD, Van Vianen A, Anderson N.
Williams KJ. Effects of realistic job previews: Changes in person-organization fit: the impact of
a comparison using an enhancement and a socialization tactics on perceived and actual P-O fit.
reduction preview. Journal of Applied Psychology European Journal of Work and Organizational Psychology
1988;73:259–66. 2004;13:52–78.
498. Hinrichs JR. An eight-year follow-up of a 511. Ashforth BK, Saks AM. Socialization tactics:
management assessment. Journal of Applied longitudinal effects on newcomer adjustment.
Psychology, 1978;63:596–601. Academy of Management Journal 1996;39:149.
499. Chatman JA. Matching people and organizations: 512. Waung M. The effects of self-regulatory coping
selection and socialization in public accounting orientation on newcomer adjustment and job
firms. Administrative Science Quarterly survival. Personnel Psychology 1995;48:633.
1991;36:459–84.
513. Klein HJ, Weaver NA. The effectiveness of an
500. Browne JH, Warnock SH, Boykin NJ. Predicting organizational-level orientation training program
success of police officer applicants using weighted in the socialization of new hires. Personnel Psychology
application blanks. Journal of American Academy of 2000;53:47–66.
Business, Cambridge 2005;6:26–31.
269
514. Wesson MJ, Gogus CI. Shaking hands with a development and validation. Academy of Management
computer: an examination of two methods of Journal 1986;29:170–82.
organizational newcomer orientation. Journal of
Applied Psychology 2005;90:1018–26. 528. Pierce JL, Newstrom JW. Employee responses
to flexible work schedules: an interorganization,
515. Cable DM, Parsons CK. Socialization tactics inter-system comparison. Journal of Management
and person–organization fit. Personnel Psychology 1982;8:9–25.
2001;54:1–23.
529. Parker SK, Griffin MA, Sprigg CA, Wall TD.
516. Mowday RT, Steers EF, Porter LW. The Effect of temporary contracts on perceived work
measurement of organizational commitment. characteristics and job strain: a longitudinal study.
Journal of Vocational Behavior 1979;14:224–47. Personnel Psychology 2002;55:689–719.
517. Chao GT, O’Uary-Kelly AM, Wolf S, Klein HJ, 530. Caplan RD, Cobb S, French JRP Jr, Van
Gardner PD. Organizational socialization: its Harrison RV, Pinneau SR Jr. Job demands and
content and consequences. Journal of Applied worker health: main effects and occupational differences.
Psychology 1994;79:730–43. Washington DC: US Department of Health,
Education and Welfare; 1975.
518. Dunham R, Herman J. Development of female
faces scale for measuring job satisfaction. Journal of 531. Adam J. Operations change interactions in a service
Applied Psychology 1975;60:629–31. environment: attitudes, behaviors, and profitability.
Journal of Operations Management 1981;2:63–73.
519. West MA. A measure of role innovation at work.
British Journal of Social Psychology 1987;26:83–5. 532. Price JL, Mueller CW. A causal model of turnover
for nurses. Academy of Management Journal
520. Cable D, Judge TA. Interviewers’ perceptions 1981;24:543–65.
of person–organization fit and organizational
selection decisions. Journal of Applied Psychology 533. D’Arcimoles CH. Human resource policies and
1997;82:562–77. company performance: a quantitative approach
using longitudinal data. Organization Studies
521. Saks AM, Ashforth BE. Organizational socialization: 1997;18:857.
making sense of the past and present as a prologue
for the future. Journal of Vocational Behavior 534. Vogus TJ, Welbourne TM. Structuring for high
1997;51:234–79. reliability: HR practices and mindful processes
in reliability-seeking organizations. Journal of
522. Narayanan VK, Nath R. Hierarchical level Organizational Behavior 2003;24:877–903.
and the impact of flextime. Industrial Relations
1982;21:216–30. 535. Ivancevich JM. Longitudinal study of the effects
of rater training on psychometric error in ratings.
523. Dunham RB, Pierce JL, Castaneda MB. Alternative Journal of Applied Psychology 1979;64:502–8.
work schedules: two field quasi-experiments.
Personnel Psychology 1987;40:215–42. 536. Cash RW, Vellema CK. Conceptual versus
competency approach in human relations
524. Latack JC, Foster LW. Implementation of training programs. Personnel and Guidance Journal
compressed work schedules: participation and job 1979;58:91.
redesign as critical factors for employee acceptance.
Personnel Psychology 1985;38:75. 537. Rautalinko E, Lisper HO. Effects of training
reflective listening in a corporate setting. Journal of
525. Markham SE, Dansereau J, Alutto JA. On the use Business and Psychology 2004;18:281–99.
of shift as an independent variable in absenteeism
research. Journal of Occupational Psychology 538. Leivo A. Field study of the effects of a self-
1982;55:225–31. implemented feedback program on housekeeping
performance. International Journal of Industrial
526. Posthuma RA, Campion MA. When do multiple Ergonomics 2005;35:471–85.
dimensions of procedural justice predict agreement
to publicly endorse your employer in recruitment 539. Green P, Skinner D. Does time management
advertisements? Journal of Occupational and training work? An evaluation. International Journal of
Organisational Psychology 2005;78:431–45. Training and Development 2005;9:124–39.
527. Dunham RB, Pierce JL. Attitudes toward work 540. Eden D, Gelter D, Gewirtz A, Gordon-Terner R,
schedules: construct definition, instrument Inbar I, Liberman M, et al. Implanting Pygmalion
270
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
leadership style through workshop training: seven psychology. Homewood, IL: Dorsey Press; 1967. pp.
field experiments. Leadership Quarterly 2000;11:171. 335–43.
541. Jacob BA, Lefgren L. The impact of teacher 554. Cook J, Wall T. New work attitude measure of trust,
training on student achievement. Journal of Human organizational commitment and personal need
Resources 2004;39:50–79. non-fulfilment. Journal of Occupational Psychology
1980;53:39–52.
542. Yarnall J. Line managers as career developers:
rhetoric or reality? Personnel Review 1998;27:378. 555. Sanchez JI, Brock P. Outcomes of perceived
discrimination among Hispanic employees: is
543. Sturges J, Guest D, Conway N, Davey KM. What diversity management luxury or a necessity?
difference does it make? A longitudinal study of Academy of Management Journal 1996;39:704–19.
the relationship between career management and
organizational commitment in the early years at 556. Ashford SJ, Lee C, Bobko P. Content, causes, and
work. Academy of Management Proceedings 2001:B1– consequences of job insecurity: a theory-based
B6. measure and substantive test. Academy of Management
Journal 1989;32:803–29.
544. Martin G, Pate J, McGoldrick J. Do HRD
investment strategies pay? Exploring the 557. Chatterjee S, Lubatkin MH, Schweiger DM,
relationship between lifelong learning and Weber Y. Cultural differences and shareholder
psychological contracts. International Journal of value in related mergers: linking equity and
Training and Development 1999;3:200. human capital. Strategic Management Journal
1992;13:319–34.
545. Kossek EE, Roberts K, Fisher S, DeMarr B. Career
self-management: a quasi-experimental assessment 558. Lubatkin M, Schweiger DM, Weber Y. Top
of the effects of a training intervention. Personnel management turnover in related M&As: an
Psychology 1998;51:935–62. additional test of the theory of relative standing.
Journal of Management 1999;25:55–67.
546. Sanchez JI, Medkik N. The effects of diversity
awareness training on differential treatment. Group 559. McIntire SM, Levine EL. Task specific self-esteem:
and Organization Management 2004;29:517–36. an empirical investigation. Journal of Vocational
Behavior 1984;25:290–303.
547. Schweiger DM, Goulet PK. Facilitating acquisition
integration through deep-level cultural learning 560. Lawler EE. Pay and organizational development.
interventions: a longitudinal field experiment. Reading, MA: Addison-Wesley; 1981.
Organization Studies 2005;26:1477–99.
561. House RJ, Levanoni E, Schuler RS. An empirical
548. Tannenbaum SI, Mathieu JE, Salas E, examination of the construct validity of the rizzo,
Cannon-Bowers JA. Meeting trainees’ expectations: house and lirtzman role scales: toward a clarification
the influence of training fulfillment on the of the nature of role conflict. Paper presented at
development of commitment, self-efficacy, the Annual Mid-West Meeting of the Academy of
and motivation. Journal of Applied Psychology Management, Columbus, OH, 1982.
1991;76:759–69.
562. Fairbrother G, Hanson KL, Friedman S, Butts GC.
549. Saks AM. The relationship between the amount and The impact of physician bonuses, enhanced
helpfulness of entry training and work outcomes. fees, and feedback on childhood immunization
Human Relations 1996;49:429–51. coverage rates. American Journal of Public Health
1999;89:171–75.
550. Coyle-Shapiro J. The impact of a TQM intervention
on teamwork: a longitudinal assessment. Employee 563. Petty MM, Singleton B, Connell DW. An
Relations 1995;17:63–74. experimental evaluation of an organizational
incentive plan in the electric utility industry. Journal
551. Murray B, Raffaele GC. Single-site, results-level of Applied Psychology 1992;77:427–36.
evaluation of quality awareness training. Human
Resource Development Quarterly 1997;8:229–45. 564. Buchko AA. Employee ownership, attitudes, and
turnover: an empirical assessment. Human Relations
552. Carkhuff RR. The development of human resources. 1992;45:711–33.
New York, NY: Holt, Rinehart and Winston; 1971.
565. Arthur JB, Jelf GS. The effects of gainsharing on
553. Smith PC. The development of a method of grievance rates and absenteeism over time. Journal
measuring job satisfaction: the Cornell Studies. In of Labor Research 1999;20:133–45.
Fleishman EA (ed.). Studies in personnel and industrial
271
566. Bhattacherjee D. The effects of group incentives 579. Welbourne TM, Cable DM. Group incentives and
in an Indian firm: evidence from payroll data. pay satisfaction: understanding the relationship
LABOUR: Review of Labour Economics and Industrial through an identity theory perspective. Human
Relations 2005;19:147–73. Relations 1995;48:711–26.
567. Yukl GA, Latham GP. Consequences of 580. Heneman HG. Pay satisfaction. Research in Personnel
reinforcement schedules and incentive magnitudes and Human Resources Management 1985;3:115–39.
for employee performance: problems encountered
in an industrial setting. Journal of Applied Psychology 581. Rusbult CE, Farrell D. A longitudinal test of the
1975;60:294–8. investment model: the impact on job satisfaction,
job commitment, and investments. Journal of Applied
568. Begley T, Lee C. The role of negative affectivity in Psychology 1983;68:429–38.
pay-at-risk reactions: a longitudinal study. Journal of
Applied Psychology 2005;90:382–8. 582. Quinn R, Shepard L. The 1973–74 quality of
employment survey. Descriptive statistics. Ann Arbor,
569. Lazear EP. Performance pay and productivity. MI: Survey Research Center of the Institute for
American Economic Review 2000;90:1346–61. Social Research; 1974.
570. Brown KA, Huber VL. Lowering floors and raising 583. Rusbult CE. Commitment and satisfaction in
ceilings: a longitudinal assessment of the effects romantic associations: a test of the investment
of an earnings-at-risk plan on pay satisfaction. model. Journal of Experimental Social Psychology
Personnel Psychology 1992;45:279–311. 1980;16:172–86.
571. Brown S. Dynamic implications of absence 584. Asch BJ. Do incentives matter? The case of Navy
behaviour. Applied Economics 1994;26:1163. recruiters. Industrial and Labor Relations Review
1990;43:89–106.
572. Johnston MW, Griffeth RW, Burton S, Carson PP.
An exploratory investigation into the relationship 585. Rosen C, Klein KJ, Young KM. Employee ownership
between promotion and turnover: a quasi- in America: the equity solution. Lexington, MA: DC
experimental longitudinal study. Journal of Heath; 1986.
Management 1993;19:33.
586. Pearce JL, Perry JL. Federal merit pay: a
573. Schwarzwald J, Koslowksy M, Shalit B. A field longitudinal analysis. Public Administration Review
study of employees’ attitudes and behaviors after 1983;43:315.
promotion decisions. Journal of Applied Psychology
1992;77:511–14. 587. Pearce JL, Stevenson WB, Perry JL. Managerial
compensation based on organizational
574. Lam SSK, Schaubroeck J. The role of locus of performance: a time series analysis of the effects
control in reactions to being promoted and to of merit pay. Academy of Management Journal
being passed over: a quasi experiment. Academy of 1985;28:261–78.
Management Journal 2000;43:66–78.
588. Kahn LM, Sherer PD. Contingent pay and
575. Pergamit MR, Veum JR. What is a promotion? managerial performance. Industrial and Labor
Industrial and Labor Relations Review Relations Review 1990;43:107–20.
1999;52:581–601.
589. Heneman HG, Schwab DP. Pay satisfaction: its
576. Pritchard RD, Jones SD, Roth PL, Stuebing KK, multidimensional nature and measurement.
Ekeberg SE. Effects of group feedback, goal setting, International Journal of Psychology 1985;20:129–41.
and incentives on organizational productivity.
Journal of Applied Psychology 1988;73:337–58. 590. Shaw JD, Gupta N. Pay fairness and employee
outcomes: exacerbation and attenuation effects
577. Pritchard RD, Jones SD, Roth PL, Stuebing KK, of financial need. Journal of Occupational and
Ekeberg SE. The evaluation of an integrated Organizational Psychology 2001;74:299–320.
approach to measuring organizational productivity.
Personnel Psychology 1989;42:69–115. 591. Bluedorn A. A unified model of turnover from
organizations. Human Relations 1982;35:135–53.
578. Knight D, Durham CC, Locke EA. The relationship
of team goals, incentives, and efficacy to strategic 592. House R, Rizzo J. Role conflict and ambiguity
risk, tactical implementation, and performance. as critical variables in a model of organizational
Academy of Management Journal 2001;44:326–38. behavior. Organizational Behavior and Human
Performance 1972;7:467–505.
272
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
593. Kanungo RN. Measurement of job and work dual-earner couples’ utilization of family-friendly
involvement. Journal of Applied Psychology workplace supports on work and family outcomes.
1982;67:341–9. Journal of Applied Psychology 2005;90:799–810.
594. Hoppock R. Job satisfaction. New York, NY: Harper; 607. Netemeyer RG, Boles JS, McMurrian R.
1935. Development and validation of work–family conflict
and family–work conflict scales. Journal of Applied
595. Seashore SE, Lawler EE III, Mirvis PH, Psychology 1996;81:400–10.
Cammann C. Assessing organizational change: a guide
to methods, measures, and practices. New York, NY: 608. Coch L, French JRP. Overcoming resistance to
Wiley; 1983. change. Human Relations 1948;1:512–532.
596. Schweiger DM, DeNisi A. Communication with 609. Morgeson FP, Campion MA, Maertz CP.
employees following a merger: a longitudinal Understanding pay satisfaction: the limits of a
field experiment. Academy of Management Journal compensation system implementation. Journal of
1991;1:110–35. Business and Psychology 2001;16:133–50.
597. Henessey J, Perrewe PL, Hochwarter WA. Impact of 610. Kleinbeck U, Fuhrmann H. Effects of a
benefit awareness on employee and organizational psychologically based management system on work
outcomes: a longitudinal field examination. Benefits motivation and productivity. Applied Psychology: An
Quarterly 1992;8:90–6. International Review 2000;49:596.
598. Kinnunen U, Mauno S, Natti J, 611. Pearson CAL. Participative goal setting as a strategy
Happonen M. Organizational antecedents and for improving performance and job satisfaction:
outcomes of job insecurity: a longitudinal study a longitudinal evaluation with railway track
in three organizations in Finland. Journal of maintenance gangs. Human Relations 1987;40:
Organizational Behavior 2000;21:443. 473–88.
599. Cohen S, Kamarck T, Mermelstein R. Journal of 612. Pearson CAL. An assessment of extrinsic feedback
Health and Social Behavior 1983;24:385–96. on participation, role perceptions, motivation,
and job satisfaction in a self-managed system for
600. House JS, Wells JA, Landerman LR, McMichael AJ, monitoring group achievement. Human Relations
Kaplan BH. Occupational stress and health among 1987;44:517–37.
factory workers. Journal of Health and Social Behavior
1979;20:139–60. 613. Kleingeld A, van Tuijl H, Algera JA. Participation
in the design of performance management
601. Schweiger DM, Ivancevich JM, Power FR. Executive systems: a quasi-experimental field study. Journal of
actions for managing human resources before and Organizational Behavior 2004;25:831–51.
after acquisition. Academy of Management Executive
1987;1:127–38. 614. Lee M, Son B. The effects of appraisal review
content on employees’ reactions and performance.
602. Greenhalgh L, Rosenblatt Z. Job insecurity: towards International Journal of Human Resource Management
conceptual clarity. Academy of Management Review 1998;9:203–14.
1984;9:438–48.
615. Heneman HG, Schwab DP. Pay satisfaction: its
603. Bertera RL. The effects of workplace health multidimensional nature and measurement.
promotion on absenteeism and employment costs International Journal of Psychology 1985;20:129–41.
in a large industrial population. American Journal of
Public Health 1990;80:1101–5. 616. Widmeyer WN, Brawley LR, Carron AV. The
measurement of cohesion in sport teams: the group-
604. Astrup SJ, McGovern PM, Kochevar LK. The environment questionnaire. London, ON: Sports
relationship between wellness participation and Dynamics; 1985.
health care benefit utilization. Benefits Quarterly
1992;8:41–54. 617. Luhtanen R, Crocker J. A collective self-
esteem scale: self evaluation of one’s social
605. Gilleskie DB, Lutz BF. The impact of employer- identity. Personality and Social Psychology Bulletin
provided health insurance on dynamic 1992;18:302–18.
employment transitions. Journal of Human Resources
2002;37:129–62. 618. Wagner U, Zick A. [Self definition and inter-
group relations: The social identity approach].
606. Hammer LB, Neal MB, Newsom JT, Brockwood KJ, In Poerzgen B, Witte EH (eds). Selbstkonzept und
Colton CL. A longitudinal study of the effects of Identitat. Braunschweig: Braunschweiger Studien
273
zur Erziehungs und Sozialwissenschaft; 1993. pp. 631. Campbell DJ. The effects of goal-contingent
109–29. payment on the performance of a complex task.
Personnel Psychology 1984;37:23–40.
619. Cammann C, Fichman M, Jenkins GD Jr,
Klesh JR. Assessing the attitudes and perceptions 632. Taylor PJ, Pierce JL. Effects of introducing a
of organizational members. In Seashore SE, performance management system on employees’
Lawler EE III, Mirvis PH, Cammann C (eds). subsequent attitudes and effort. Public Personnel
Assessing organizational change: a guide to methods, Management 1999;28:423.
measures, and practices. New York, NY: Wiley; 1983.
pp. 71–138. 633. Murray RS. Managerial perceptions of two appraisal
systems. California Management Review 1981;23:92.
620. Locke EA. Toward a theory of task motivation
and incentives. Organizational Behavior and Human 634. Westin AF. Two key factors that belong in a
Performance 1968;3:157–89. macroergonomic analysis of electronic monitoring:
employee perceptions of fairness and the climate of
621. Reilly RR, Smither JW. A longitudinal study organizational trust or distrust. Applied Ergonomics
of upward feedback. Personnel Psychology 1992;23:35–42.
1996;49:599–612.
635. Markham SE, Scott KD, McKee GH. Recognizing
622. Walker AG, Smither JW. A five-year study of upward good attendence: a longitudinal, quasi-
feedback: what managers do with their results experimental field study. Personnel Psychology
matters. Personnel Psychology 1999;52:393. 2002;55:639–60.
623. Smither JW, London M, Reilly RR, Flautt R, 636. Fellows C, Mawhinney TC. Improving
Vargas Y, Kucine I. Discussing multisource feedback telemarketers’ performance in the short-run using
with raters and performance improvement. Journal operant concepts. Journal of Business and Psychology
of Management Development 2004;23:456–68. 1997;11:411–24.
624. Kim JS, Hamner WC. Effect of performance 637. Kinicki AJ, Prussia GE, Bin J, Kee-Ryan FM.
feedback and goal setting on productivity and A covariance structure analysis of employees’
satisfaction in an organizational setting. Journal of response to performance feedback. Journal of Applied
Applied Psychology 1976;61:48–57. Psychology 2004;89:1057–69.
625. Ludwig TD, Geller ES. Assigned versus participative 638. Porac JF, Nottenburg G, Eggert J. On extending
goal setting and response generalization: managing Weiner’s attributional model to organizational
injury control among professional pizza deliverers. contexts. Journal of Applied Psychology 1981;66:124–6.
Journal of Applied Psychology 1997;8:253–61.
639. LeBaron CW, Chaney M, Baughman AL, Dini EF,
626. Komaki JL, Collins RL, Penn P. The role of Maes E, Dietz V, et al. Impact of measurement and
performance antecedents and consequences in feedback on vaccination coverage in public clinics,
work motivation. Journal of Applied Psychology 1988–1994. Journal of the American Medical Association
1982;67:334–40. 1997;278:631–5.
627. Codding RS, Feinberg AB, Dunn EK, Pace GM. 640. Hackman JR, Oldham GR. Development of the
Effects of immediate performance feedback on Job Diagnostic Survey. Journal of Applied Psychology
implementation of behaviour support plans. Journal 1975;60:158–70.
of Applied Behavior Analysis 2005;38:205–19.
641. US Department of Labor. High performance work
628. Parsons CK, Herold DM, Leatherwood ML. practices and firm performance. Washington, DC: US
Turnover during initial employment: a longitudinal Department of Labor; 1993.
study of the role of causal attributions. Journal of
Applied Psychology 1985;70:337–41. 642. Datta DK, Guthrie JP, Wright PM. Human
resource management and labor productivity: does
629. Arvey RD, Dewhirst HD, Brown EM. A longtitudinal industry matter? Academy of Management Journal
study of the impact of changes in goal setting 2005;48:135–45.
on employee satisfaction. Personnel Psychology
1978;31:595–608. 643. Gittleman M, Horrigan M, Joyce M. “Flexible”
workplace practices: evidence from a national
630. Ivancevich JM. Effects of goal-setting on representative survey. Industrial and Labor Relations
performance and job-satisfaction. Journal of Applied Review 1998;52:99–115.
Psychology 1976;61:605–12.
274
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
644. Barney JB. Firm resources and sustained 657. Montori VM, Wilczynski NL, Morgan D,
competitive advantage. Journal of Management Haynes RB. Optimal search strategies for retrieving
1991;17:99–120. systematic reviews from Medline: analytical survey.
British Medical Journal 2005;330:68–73.
645. Rauch A, Frese M, Utsch A. Effects of human capital
and long-term human resources development and 658. Huque M. Experiences with meta-analysis in NDA
utilization on employment growth of small-scale submissions. Proceedings of the Biopharmaceutical
businesses: a causal analysis. Entrepreneurship: Theory Section of the American Statistical Association
and Practice 2005;29:681–98. 1988;2:28–33.
646. Delery JE. Issues of fit in strategic human resource 659. Cook D, Sackett D, Spitzer W. Methodologic
management: implications for research. Human guidelines for systematic reviews of randomized
Resource Management Review 1998;8:289–309. control trials in health care from the Potsdam
Consultation on Meta-analysis. Journal of Clinical
647. Wood SJ, De Menzes L. Comparing perspectives on Epidemiology 1995;48:167–71.
high involvement management and organizational
performance across the British economy. 660. LaValley M. A consumer’s guide to meta-analysis.
International Journal of Human Resource Management Arthritis Care and Research 1997;10:208–13.
2008;19:639–83.
661. Geller N, Proschan M. Meta-analysis of
648. Arthur J. The link between business strategy and clinical trials: a consumer’s guide. Journal of
industrial relations systems in American steel Biopharmaceutical Statistics 1996;6:377–94.
minimills. Industrial and Labor Relations Review
1992;45:488–506. 662. Felson D. Bias in meta-analytic research. Journal of
Clinical Epidemiology 1992;45:885–92.
649. Freeman RB, Kleiner MM. The last American
shoe manufacturers: decreasing productivity and 663. CASP. Critical Appraisal Skills Programme. 10
increasing profits in the shift from piece rates to questions to help you make sense of reviews.
continuous flow production. Industrial Relations Oxford: Public Health Resource Unit; 2006.
2005;44:307–30.
664. MOOSE Group. Meta-analysis of observational
650. Shipton H, Fay D, West M, Patterson M, Birdi K. studies in epidemiology: a proposal for reporting.
Managing people to promote innovation. Creativity Journal of the American Medical Association
and Innovation Management 2005;14:118–28. 2000;283:2008–12.
651. Gomez-Mejia LR. The role of human resources 665. Ankem K. Approaches to meta-analysis: a guide
strategy in export performance: a longitudinal for LIS researchers. Library and Information Science
study. Strategic Management Journal 1988;9:493–505. Research 2005;27:164–74.
652. Katz HC, Kochan TA, Gobeille KR. Industrial 666. Egger M, Smith GD. Meta-analysis: bias in location
relations performance, economic performance, and and selection of studies. British Medical Journal
QWL programs: an interplant analysis. Industrial 1998;316:61–6.
and Labor Relations Review 1983;37:3–17.
667. Whitener E. Confusion of confidence intervals
653. Schuster M. The impact of union-management and credibility intervals in meta-analysis. Journal of
cooperation on productivity and employment. Applied Psychology 1990;75:315–21.
Industrial and Labor Relations Review
1983;36:415–430. 668. Cohen J. Statistical power analysis for the behavioral
sciences, 2nd edn. Hillsdale, NJ: Lawrence Erlbaum
654. Workman M, Bommer W. Redesigning computer Associates; 1988.
call center work: a longitudinal field experiment.
Journal of Organizational Behavior 2004;25:317–37. 669. Meyer JP, Stanley DJ, Herscovitch L, Topolnytsky L.
Affective, continuance and normative commitment
655. Kochan TA, Katz HC, McKersie RB (eds) The to the organization: a meta-analysis of antecedents,
transformation of American industrial relations. Ithaca, correlates and consequences. Journal of Vocational
NY: Cornell University Press; 1994. Behavior 2002;61:20–52.
656. Judge TA, Thoresen CJ, Bono JE, Patton GK. The 670. Colquitt JA, Conlon DE, Wesson MJ, Porter CO,
job satisfaction-job performance relationship: a Ng KY. Justice at the millennium: a meta-analytic
qualitative and quantitative review. Psychological review of 25 years of organizational justice research.
Bulletin 2001;127:376–407. Journal of Applied Psychology 2001;86:425–45.
275
671. Landy FJ. Psychology of work behaviour. Pacific Grove, 683. Schneider B, Hanges PJ, Smith DB,
CA: Brooks/Cole; 1989. Salvaggio AN. Which comes first: employee
attitudes or organizatioal and financial market
672. Warr P. Work, happiness, and unhappiness. London: performance? Journal of Applied Psychology
Laurence Erlbaum Associates; 2007. 2003;88:836–51.
673. Carr JZ, Schmidt AM, Ford JK, DeShon RP. Climate 684. Bernhardt KL, Donthu N, Kennett PA. A
perceptions matter: a meta-analytic path analysis longitudinal analysis of satisfaction and profitability.
relating molar climate, cognitive and affective Journal of Business Research 2000;47:161–71.
states, and individual level work outcomes. Journal
of Applied Psychology 2003;88:605–19. 685. Koys DJ. The effects of employee satisfaction,
organizational citizenship behaviour, and
674. Parker CP, Baltes BB, Young SA, Huff JW, turnover on organizational effectiveness: a unit-
Altmann RA, Lacost HA, et al. Relationships level, longitudinal study. Personnel Psychology
between psychological climate perceptions and 2001;54:101–14.
work outcomes: a meta-analytic review. Journal of
Organizational Behaviour 2003;24:389–416. 686. Morrow PC, McElroy JC. Efficiency as a mediator
in turnover-organizational performance relations.
675. Ostroff C. The effects of climate and personal Human Relations 2007;60: 827–49.
influences on individual behaviour and attitudes in
organizations. Organizational Behaviour and Human 687. Sparrow P, West M. Psychology and organizational
Decision Processes 1993;56:56–90. effectiveness. In Robertson IT, Callinan M,
Bartram D (eds). Organizational effectivevness: the role
676. Jones AP, James LR. Psychological climate: of psychology. Chichester: John Wiley & Sons; 2002.
dimensions and relationships of individual and
aggregated work environment perceptions. 688. Robertson A, Gilloran A, McGleew T, McKee K,
Organizational Behaviour and Human Performance McInley A, Wight D. Nurses’ job satisfaction and
1979;23:201–50. the quality of care receivevd by patients in
psychogeriatric wards. International Journal of
677. Tourangeau AE, Doran DM, McGillis Hall L, Geriatric Psychiatry 1995;10:575–84.
O’Brien Pallas L, Pringle D, Tu JV, et al. Impact
of hospital nursing care on 30-day mortality for 689. Leiter MP, Harvie P, Frizzell C. The correspondence
acute medical patients. Journal of Advanced Nursing of patient satisfaction and nurse burnout. Social
2007;57:32–44. Science and Medicine 1998;47:1611–17.
678. Grembowski D, Paschane D, Diehr P, Katon W, 690. Mayer J, Faber M, Xu X. Seventy-five years of
Martin D, Patrick DL. Managed care, physician job motivation measures (1930–2005): a descriptive
satisfaction and the quality of primary care. Journal analysis. Motivation and Emotion 2007;31:83–103.
of General Internal Medicine 2005;20:271–7.
691. Blau G. The measurement and prediction of career
679. McGilton K, Irwin-Robinson H, Boscart V, commitment. Journal of Occupational Psychology
Spanjevic L. Communication enhancement: nurse 1985;58:277–88.
and patient satisfaction outcomes in a complex
continuing care facility. Journal of Advanced Nursing 692. Schaufeli WB, Bakker A, Salanova M. The
2006;54:35–44. measurement of work engagement with a short
questionnaire: a cross-national study. Educational
680. Priebe S, Fakhoury W, White I, Watts J, and Psychological Measurement 2006;20:1–16.
Bebbington P, Burns T, et al. Characteristics of
teams, staff and patients: associations with outcomes 693. Maslach C, Jackson SE. The measurement of
of patients in assertive outreach. British Journal of experienced burnout. Journal of Occupational
Psychiatry 2004;185:306–11. Behaviour 1981;2:99–113.
681. DiMatteo MR, Sherbourne CD, Hays RD, 694. Maslach C, Jackson SE. Burnout in organizational
Ordway L, Kravitz RL, MGlynn EA, et al. Physicians’ settings. Applied Social Psychology Annual
characteristics influence patients’ adherence 1984;5:133–53.
to medical treatment: results from the Medical
Outcomes Study. Health Psychology 1993;12:93–102. 695. Koeske GF, Koeske RD. Construct validity of the
Maslach Burnout Inventory: a critical review and
682. Goodell TT, Coding HVE. Outcomes of nurses’ reconceptualization. Journal of Applied Behvioral
job satisfaction. Journal of Nursing Administration Science 1989;25:131.
1994;24:36–41.
276
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
696. Cordes CL, Dougherty TW. A review and an on company productivity: a longitudinal study.
integration of research on job burnout. Academy of Personnel Psychology 2008;61:467–501.
Management Review 1993;18:621–56.
710. Iles V, Sutherland K. Managing change in the NHS:
697. Arthur NM. The assessment of burnout: a review of organisational change, a review for health care managers,
three inventories useful for research and counseling. professionals and researchers. London: SDO; 2001.
Journal of Counseling and Development 1990;69:186–9.
711. Godard J, Delaney JT. Reflections on the “high
698. Rousseau DM. Psychological contract inventory technical performance” paradigm’s implications for industrial
report [online]. Pittsburgh, PA. Carnegeie Mellon relations. Industrial and Labor Relations Review
University. URL: www.andrew.cmu.edu/user/ 2000;53:482–502.
rousseau/0_reports/PCI.pdf (accessed 21 November
2007). 712. Wood S, Wall T. Human resource management
and business performance. In Warr P (ed.). The
699. Gilliland SW. The perceived fairness of selection Psychology of Work. Harmondsworth, UK: Penguin;
systems: an organizational justice perspective. 2002. pp. 351–74.
Academy of Management Review 1993;18:694–734.
713. De Menezes LM, Wood S. The reality of flexible work
700. Niehoff BP, Moorman RH. Justice as a mediator of Systems in the UK. London: Cass Business School;
the relationship between methods of monitoring 2003.
and organizational citizenship behaviour. Academy of
Management Journal 1993;36:527–56. 714. Wright MA, Gardner TM. The human resource–
firm performance relationship: methodological
701. Hofman DA, Mark B. An investigation of the and theoretical challenges. In Holman D, Wall TD,
relationship between safety climate and medication Clegg CW, Sparrow P, Howard A (eds). The new
errors as well as other nurse and patient outcomes. workplace: a guide to the human impact of modern
Personnel Psychology 2006;59:847–69. working practices. London: Wiley; 2003. pp. 271–90.
702. Katz-Navon T, Naveh E, Stern Z. Safety climate 715. Department of Health. Mental Health Information
in health care organizations: a multidimensional Strategy. London: Department of Health; 2001.
approach. Academy of Management Journal
2005;48:1075–89. 716. Bartram D. Review of personality assessment instruments
(Level B) for use in occupational settings. Leicester:
703. Flin R, Burns C, Mearns K, Yule S, Robinson EM. British Psychological Society; 1995.
Measureing safety climate in health care. Quality and
Saftey in Healthcare 2006;15:109–15. 717. Michaels JA, Booth A. Pragmatic system for the
grading of evidence and recommendations in
704. Sorra J, Nieva VF. Saftey culture assessment: a clinical guidelines. Journal of Clinical Excellence
tool for improving patient safety in health care 2001;3:139–43.
organizations. Quality and Saftey in Health Care
2003;12(Suppl. 2):17–23. 718. Woolf SH, Battista RN, Anderson GM, Logan AG,
Wang E. Assessing the clinical effectiveness of
705. Marchington M, Zagelmeyer S. Linking HRM and preventive maneuvers: analytic principles and
performance – a never-ending search? Human systematic methods in reviewing evidence and
Resource Management Journal 2005;15:3–8 developing clinical practice recommendations.
A report by the Canadian Task Force on Periodic
706. CIPD Performance through people: the new people Health Examination. Journal of Clinical Epidemiology
management. Postal survey report 2001. URL: www. 1990;43:891–905.
cipd.co.uk/subjects/corpstrtgy/busiperfm/perfthrpeo.
htm (accessed 20 December 2007). 719. Sackett DL. Rules of evidence and clinical
recommendations. Canadian Journal of Cardiology
707. West MA, Guthrie JP, Dawson JF, Borrill CS, 1993;9:487–9.
Carter MR. Reducing patient mortality in hospitals:
the role of human resource management. Journal of 720. Harris RP, Helfand M, Woolf SH, Lohr KN,
Organizational Behavior 2006;27:983–1002. Mulrow SM, Teutsch SM, et al. Current methods
of the third US Preventive Services Task
708. Walton RE. From ‘control’ to ‘commitment’ in the Force. American Journal of Preventive Medicine
workplace. Harvard Business Review 1985;63:77–84. 2001;20(Suppl. 3): 21–35.
709. Birdi K. Clegg C, Patterson MG, Robinson A, 721. Guyatt G, Drummond R. Users’ guide to the medical
Stride C, Wall T, et al. The impact of human literature: essentials of evidence-based clinical practice.
resource and operational management practices Chicago, IL: AMA Press; 2001.
277
722. Muir Gray J. Evidence-based healthcare. Edinburgh: 738. Judge TA, Ilies R. Relationship of personality to
Churchill-Livingstone; 2001. performance motivation: a meta-analytic review.
Journal of Applied Psychology 2002;87:797–807.
723. Easterby-Smith M, Thorpe R, Lowe A. Management
research: an introduction. London: Sage; 1991. 739. Faragher EB, Cass M, Cooper CL. The relationship
between job satisfaction and health: a meta-
724. Popay J, Rogers A, Williams G. Rationale and analysis. Occupational and Environmental Medicine
standards for the systematic review of qualitative 2005;62:105–12.
literature in health services research. Qualitative
Health Research 1998;8:341–51. 740. Hellman CM. Job satisfaction and intent to leave.
Journal of Social Psychology 1997;137:677–89.
725. Oxman A. Systematic reviews: checklists for review
articles. British Medical Journal 1994;309:648–51. 741. Brown SP, Peterson RA. Antecedents and
consequences of salesperson job satisfaction: meta-
726. Goldstein IL. Training in organizations, 3rd edn. analysis and assessment of causal effects. Journal of
Pacific Grove, CA: Brooks-Cole; 1993. Marketing Research 1993;30:63–77.
727. Gerhart B, Milkovich GT. Employee compensation: 742. Blegen MA. Nurses’ job satisfaction: a meta-
research and practice. In Dunnette MD, Hough LM analysis of related variables. Nursing Research
(eds). Handbook of industrial and organizational 1993;42:36–41.
psychology, Vol. 3, 2nd edn. Palo Alto, CA:
Consulting Psychologists Press; 1992. pp. 484–569. 743. Witt LA, Nye LG. Gender and the relationship
between perceived fairness of pay or promotion
728. Hartin W. Employment law. Corporate Management and job satisfaction. Journal of Applied Psychology
1994;46:75–6. 1992;77:910–17.
729. Work and family. Government Perspective 1994;6:2. 744. Dalal RS. A meta-analysis of the relationship
between organizational citizenship behavior and
730. Bakker T. Doing well by doing good: the bottom line on counterproductive work behavior. Journal of Applied
workplace practices. Washington, DC: Economic Policy Psychology 2005;90:1241–55.
Institute; 1999.
745. LePine JA, Erez A, Johnson DE. The nature and
731. Delaney JT, Goddard J. An industrial relations dimensionality of organizational citizenship
perspective on the high-performance behavior: a critical review and meta-analysis. Journal
paradigm. Human Resource Management Review of Applied Psychology 2002;87:52–65.
2001;11:395–429.
746. Harter JK, Schmidt FL, Hayes TL. Business-unit-
732. Van Buren ME, Werner JM. High performance work level relationship between employee satisfaction,
systems. Business and Economics Review 1996;12: employee engagement, and business outcomes:
15–22. a meta-analysis. Journal of Applied Psychology
2002;87:268–79.
733. Egger M, Smith GD, Phillips AN. Meta-analysis:
principles and procedures. British Medical Journal 747. Carsten JM, Spector PE. Unemployment, job
1997;315:1533–37. satisfaction, and employee turnover: a meta-analytic
test of the Muchinsky model. Journal of Applied
734. Higgins J, Thompson SG, Deeks JJ, Altman DG. Psychology 1987;72:374–81.
Measuring inconsistency in meta-analysis. British
Medical Journal 2003;32: 557–60. 748. Irvine DM, Evans MG. Job satisfaction and turnover
among nurses: integrating research findings across
735. Hardy R, Thompson S. Detecting and describing studies. Nursing Research 1995;44:246–53.
heterogeneity in meta-analysis. Statistics in Medicine
1998;17:841–56. 749. Tett RP, Meyer JP. Job satisfaction, organizational
commitment, turnover intention, and turnover:
736. Hunter J, Schmidt F. Methods of meta-analysis: path analyses based on meta-analytic findings.
correcting error and bias in research findings. Beverley Personnel Psychology 1993;46:259–93.
Hills, CA: Sage; 1990.
750. Farrell D, Stamm CL. Meta-analysis of the
737. Hunter J, Schmidt F, Jackson G. Meta-analysis: correlates of employee absence. Human Relations
cumulating research findings across studies. Beverley 1988;41:211–27.
Hills, CA: Sage; 1982.
751. Petty MM, McGee GW, Cavender JW. A meta-
analysis of the relationships between individual job
278
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
752. Steel RP, Ovalle NK. A review and meta-analysis 760. Randall DM. The consequences of organizational
of research on the relationship between behavioral commitment: methodological investigation. Journal
intentions and employee turnover. Journal of Applied of Organizational Behavior 1990;11:361–78.
Psychology 1984;69:673–86.
761. Ferris KR, Aranya N. A comparison of two
753. Scott KD, Taylor GS. An examination of conflicting organizational commitment scales. Personnel
findings on the relationship between job satisfaction Psychology 1983;36:87–98.
and absenteeism: a meta-analysis. Academy of
Management Journal 1985;28:599–612. 762. Meyer JP, Allen NJ. Commitment in the workplace:
theory, research and application. Thousand Oaks,
754. Cooper-Hakim A, Viswesvaran C, Cooper-Hakim A, CA: Sage Publications; 1997.
Viswesvaran C. The construct of work commitment:
testing an integrative framework. Psychological 763. Meyer JP, Allen NJ, Smith CA. Commitment to
Bulletin 2005;131:241–59. organizations and occupations: extension and test
of a three-component conceptualization. Journal of
755. Colquitt JA, Lepine JA, Noe RA. Toward an Applied Psychology 1993;78:538–51.
integrative theory of training motivation: a meta-
analytic path analysis of 20 years of research. Journal 764. Wright TA, Bonett DG. The moderating effects
of Applied Psychology 2000;85:678–707. of employee tenure on the relation between
organizational commitment and job performance:
756. Cohen A, Gattiker UE. Rewards and organizational a meta-analysis. Journal of Applied Psychology
commitment across structural characteristics: a 2002;87:1183–90.
meta-analysis. Journal of Business and Psychology
1994;9:137–57. 765. Riketta M. Attitudinal organizational commitment
and job performance: a meta-analysis. Journal of
757. Cohen A, Hudecek N. Organizational commitment– Organizational Behavior 2002;23:257–66.
turnover relationship across occupational groups: a
meta-analysis. Group and Organization Management 766. Melchior ME, Bours GJ, Schmitz P, Wittich Y.
1993;18:188–213. Burnout in psychiatric nursing: a meta-analysis of
related variables. Journal of Psychiatric and Mental
758. Cohen A. Career stage as a moderator of the Health Nursing 1997;4:193–201.
relationships between organizational commitment
and its outcomes: a meta-analysis. Journal of 767. Rosenthal R. Meta-analytic procedures for social research
Occupational Psychology 1991;64:253–68. (revised edition). Newbury Park, CA: Sage; 1991.
279
Appendix 1
Taken from original proposal to NCCSDO
Introduction morale to the modernisation of the NHS and its
The research will be conducted by a team from models of patient care such as Protocol Based Care
the University of Sheffield, based at The Institute (NHS Modernisation Agency 2003). In the words
of Work Psychology (IWP) and the School of of the Department of Health’s website ‘Successful
Health and Related Research (ScHARR): Malcolm human resources management improves efficiency
Patterson, Jo Rick, Stephen Wood, Andrew Booth, through a culture that supports and develops
Tony Smith, Malcolm Whitfield and two research its staff, allowing the workforce to share in the
associates. In line with the invitation to tender organisation’s objectives’.
(ITT), a panel of around eight expert advisers
will be recruited. The proposed project seeks to The desire to base the modernisation on evidence
maximise the strengths of the two Institutes and is confirmed by the Department’s attempt to
to build particularly on the strong contribution support its overall orientation with research
that the IWP has made to the human resource findings. It thus, for example, offers the finding
management (HRM) field, empirically and that patient mortality rates improve in acute
theoretically, and ScHARR and Jo Rick have hospitals where an HR director is on the Trust
made to the development of systematic literature Board in support of its view that ‘Human resources
reviewing generally and particularly in the management is vital to the NHS’. In this way, the
medical and health-related areas. The research Department of Health and associated bodies are
design aims to achieve the objective of delivering mirroring the more general emphasis on employee
a critical review of the evidence for the validity development and empowerment as the key to
and reliability of measures of HRM, through an maximising the human resource’s contribution
examination of the five objectives outlined in the to strategic objectives of organisations. It is this
tender document. human capital approach to HRM that has been
at the centre of the literature on HRM in the past
This section sets out the following: decade. In the light of the centrality given it in the
tender document, it will form the fulcrum of our
• the research context and background to the research. First it will be a core part of the literature
study, including policy relevance and related reviewed and it will form the starting point for
research identifying the other literatures and avenues that
• our understanding of the aims and objectives we need to include in the review. For example, the
of the research and the associated research core of the HRM performance literature focuses on
issues aggregates of practices, but we will need to extend
• our proposed methodological approach and to that which looks at single practices, for example
analytic methods to be used payment systems.
• our proposed timetable for the research
and the outputs we would produce from the In some overviews29,646,711,712 of the HRM
research. performance literature there has been a stress on
how the research is demonstrating a link between
Background to the study human capital approach and key organisational
outcomes. If the evidence were perhaps as strong
Appropriate HRM policies and practices are widely as these reviews or the authors of some individual
seen as vital if the NHS is to change the process studies might suggest, there would be little need
of health care and improve outcomes for patients. for this research project. The studies are not so
The key issue is to identify what are appropriate, homogeneous, consistent in their results or without
and the evidence on which this can be based. methodological weaknesses for us to conclude this.
The Department of Health statement and the Indeed several recent reviews29 have questioned
NHS’s Modernisation Agency have repeatedly the evenness of the results and highlighted the
stressed the vital role of human resources, skill methodological issues in this area. The implication
development, involvement, teamworking and is that the field is young and there is a need to
281
take stock of the work so far before a further big that will, regardless of the context, outperform
investment is made in the area and particularly all others, either eternally or at this particular
if the general models are to be considered in the historical conjuncture. Outside of the human
under-researched service sectors like health care. resource (HR) and organisational behaviour
The core background factors to this application arenas, in the strategy literature, the resource-
are then (1) the salience being given to HRM as a based theory of the firm5,6 has also highlighted the
major determinant of organisational performance; way in which human resources and the knowledge
(2) the conclusions that research is supporting a embodied in them may be decisive for the long-
link between HRM and performance; while, on term competitive advantage of firms, while in the
the other hand (3), that there is a questioning of operations management literature7–10 the emphasis
the homogeneity and robustness of the research is on how HR practices and the knowledge and
evidence; and (4) the timeliness of taking a broad- skills of workers are decisive for the successful
ranging stock-take of the HRM field and its exploitation of lean production11 and other modern
implications for health-care policy. We shall now management methods (see, for example, Dean and
expand on each of these factors. Bowen12 on total quality management).
The salience given to HRM in both Research supports the core thesis
health care and as a universal model The second background factor is the widely
The key background to the study is then the expressed view that the main research studies in the
growing attention over the past 20 years to HRM as HRM area have confirmed that high-involvement
a major contributor to organisation performance, management does positively affect organisational
even in capital or technological intensive performance. The basic hypothesis that underlies
industries. It is assumed to contribute both in its much of this work is that: high-involvement HR
own right and as a support for such other modern systems will have positive effects on organisational
management practices as total quality management performance. As academics have sort to move
and just-in-time (JIT). So, in the health context beyond advocacy of high-involvement systems, they
HRM adds to the human capital of the workforce have focused on testing this hypothesis. A spate
and helps support new modes of organisation, of research studies in the 1990s were motivated by
technology and patient delivery. Its importance this. Many concluded that high-involvement HR
derives from an assumption that some forms of systems do perform best.13–20 Several overviews
personnel management have greater effects on the of the studies21,22 concluded on the basis of the
development, skills and motivation of employees first few studies that the universalistic hypothesis
than others. Forms that are thought to have such is supported. The increasing use of the term
effects, as the tender specification acknowledges, ‘high-performance’ model to describe the set
are indeed often labelled HRM. Alternative labels of practices might be seen as implying that the
are the human capital, high-commitment, high- matter is indeed settled, much as a drug might be
involvement and high-performance approaches. named by the disease it is known to cure. Overviews
For ease of presentation we shall refer to the have tended to present the studies in a rather
specific model as the high-involvement model and homogeneous way, largely assuming that it can
use HRM when a potentially more all-embracing be taken for granted they have been studying the
concept of personnel management is being same phenomenon and the results are broadly the
referred to. same. These reviews have led to the conclusion that
the main issue now is to explain the link between
Much of the attention given to the high- high-involvement management performance, i.e.
involvement or -commitment HRM system has assess the mechanisms between them or get inside
been advocatory. Contrasting it with a traditional the black box between the system and the outcome.
Taylorist or Fordist approach, academics (e.g. A second issue might be the extent to which the
Pfeffer,2 Kochan and Osterman,3 Walton1) link will be found in all contexts, and particularly
have been at the forefront of publicising high- in service industries such as health care. The
involvement management as a progressive form majority of the early studies were in manufacturing,
of management that can remove past restrictions and it is significant that one of the few early
to economic efficiency and achieve high-quality studies in services24 found evidence of a contingent
performance, and thus sharpen the competitive relationship between the human capital approach
edge of Western organisations and improve public and performance. In the past few years there have
sector delivery. The argument is that there is a been more studies that have either been in the
unique set of practices or approach to management service sector or included it in their study15,25–27 and
282
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
the results have been mixed. Studies specifically on practices as supports to help motivate and
health care are emerging, For example, West et al.’s equip people to work in a new more flexible
study28 has investigated the link between HRM and and proactive way. On the other hand are those
patient mortality. who do not prioritise the task system but treat
high-involvement management as entailing
The debate is far from over the use of the whole gamut of sophisticated
A third background to the study is the increasing personnel management methods and conceive
realisation that the portrayal of the studies as them as primarily operating through people’s
providing conclusive evidence for a universal link commitment and sense of whether they are
between HRM and performance is premature, and being treated fairly and consistent with their
that there is a clear need to delve more deeply psychological contract. The two approaches
both in to the studies and the reality of high- need not affect greatly the practices researchers
involvement management and its links to other include in their studies, but the first would
aspects and methods of management. The reviews not treat them in an undifferentiated way.
so far have not used the systematic review methods Nonetheless, the inclusion of job design and
we propose to adopt, hence the timeliness of the work organisation practices is one major source
project. of diversity between the studies. The other is
incentive systems with some studies treating
There are eight reasons for extending the debate: them as part of high-involvement management
or at least performance management, others as
1. The results of the various studies are neither contradicting it.
as clear-cut nor as uniform as some have 4. The vast majority of studies are based on
concluded (see Wood29 and particularly his reported use of a practice and do not assess its
table 1 for summary of the main results). effectiveness or differentiate between its mere
Within studies, there is unevenness in the existence and its effective implementation.
findings between performance measures. While For example, appraisals may be conducted
some results point to universal effects, others but the extent to which they are done well
do not. Moreover, in many studies only the or whether they are mainly for control or
universal hypothesis has been tested, so one payment purposes rather than development
cannot rule out the contingency argument that or employee involvement purposes is not
the effects of high-involvement management considered.
on performance are contingent on a third 5. It is not always clear whether high involvement
factor, either a dimension of the context or or performance management is being defined
the strategy of the organisation, even when simply as the combined used of the best
a positive link between high-performance practice in each domain of human resources,
management and performance has been a synergistic set of practices, or by a more
found. fundamental managerial orientation.713 In fact
2. A variety of other hypotheses have been tested, the discussion so far has begged the question
including the contingency hypotheses that of how we define and identify a system. Most
high-involvement management works best studies have defined an HR system a priori,
when used in conjunction with other quality often with little theoretical justification, as
or lean practices. The results of these in some the emphasis has been on testing the high
cases question the universal thesis. involvement–performance link rather than first
3. There is considerable variability across studies investigating the relationship between practices
in the practices included either in the model or the nature of any systems. Researchers have
of high-involvement management or the set of measured HRM on the basis of differentiating
practices used to test its link to performance. organisations by their usage of a set of
Some of the differences are terminological, practices, the precise means varying between
perhaps reflecting disciplinary biases or studies. They then correlate the measure
a quest to differentiate ones wares. But a with performance measures. The problem of
core difference within the literature can be simply aggregating practices to provide an
identified. On the one hand are those who see overall measure is that each practice is treated
the core of high-involvement management implicitly as if it were equally important.
as changes in work organisation, job design With only a few exceptions (e.g. Patterson
and employee involvement methods (and et al.33 forthcoming), studies have provided
particularly idea capturing) with the other insufficient information about which individual
283
or subset of practices has the strongest effects on intermediate HR outcomes (e.g. labour
and which may be marginal or even irrelevant. turnover in Huselid18) or concentrated on
In one case13 where this was done and it was commitment and satisfaction (e.g. Patterson et
shown that only some of the practices affect al.,36 Wright et al.37). Several authors hint at, but
performance, and, moreover, that their do not explore in depth, other mechanisms, for
strength varied between industrial context, the example Guest and Conway38 imply procedural
authors still framed their conclusions in terms justice. The emphasis in statements on the
of an overall (high-performance work) system modernisation of health care, in addition to
yielding superior performance. Aggregating morale and staff development, is on achieving
practices does not allow one to test whether the what might be termed the strategic integration
effects are synergistic and begs the question of individuals into the organisation so that
of whether the pattern of association between they both understand and share the goals that
practice use reflects a single underlying achieving patient care entail and consequently
orientation [which, in statistical terms, would orientate their behaviour towards common
be implied if a one-factor model could be fitted visions. Studies outside the narrow confines
to the data (see, for example, Wood and de of the HRM studies, particularly building
Menezes26 and Wood29]. on the job redesign and lean production
6. Despite the differences in focus and measures, literature, point to the importance of the
most of the studies share a basic research kinds of attitudes that advocates of total
design, which has a number of weaknesses. quality management and other modern
These include: management methods highlight, for example
i. The use of cross-sectional designs – quality consciousness (e.g. Peccei and Wood39)
limiting causal inference (i.e. do HRM continuous improvement orientation (Coyle-
practices promote performance, or does Shapiro40) or flexible work orientation (Parker
better performance encourage greater et al.41).
investment in HRM?). 8. The majority of studies have concentrated
ii. Reliance on a single data source (e.g. on a narrow range of HRM practices in
CEO or HRM Director) for information isolation, the exception being those that
on HRM practices and performance have included TQM in their analysis (e.g.
– resulting in measures of unknown MacDuffie,20 Patterson et al.,33 de Menezes
reliability and susceptible to common- and Wood713), but there is a general failure to
method bias (see, for example, the debate assess the effectiveness of HRM and related
between Huselid and Becker42 and Gerhart practices relative to both other key elements of
et al.43 on the reliability of measures based management such as leadership and practices
on single respondents). further afield, such as R&D expenditure.
iii. Many but not all include small samples,
low response rates and a concentration Taking stock of the HRM performance
on a limited range of sectors – limiting area
generalisability. The fourth background factor is the clear need
7. There has been insufficient theoretical to appraise the area in depth prior to developing
discussion or empirical exploration of the further studies. Despite the diversity in the
mechanisms or intervening variables that studies and the limitations of their methodology,
might explain any link between HRM and the studies thus far have provided a foundation
performance. Labels such as high commitment from which broader and more methodologically
and involvement imply certain mechanisms appropriate studies can be designed. There is
that are concerned with the extent to sufficient promise in the body of evidence to
commitment or engagement on the part consider major investments in the area. However,
of the work force, which is mirrored in the it is timely to invest in detailed assessment of what
importance given to morale in statements we know and can take from the studies so far. If an
about the modernisation of health care. Much exercise is treated as the foundation for both good
of the theoretical justification that precedes HRM practice and the next stage of research, it
the empirical studies, however, implies that will need, however, to go beyond simply assessing
skills, knowledge and learning might be as the studies. As we have implied, they have been
important, independent of attitude change. limited. Assessing their precise limits will be part of
While some studies have sought to include the analysis of them. But, as we have implied, some
intervening variables these have either centred of the limits are clear:
284
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
1. The focus has been on the practices– the likely literature available to provide relevant
performance link to the neglect of both evidence. We then consider the main research
theoretical discussion and research on the issues and associated implications for the design of
nature of the association between the practices. the methodology.
2. The mechanisms or intervening variables that
might explain any links have been insufficiently Objectives for the research
explored, theoretically or empirically, and what
discussion there has been has been limited. This research aims to deliver an up-to-date,
3. The precise performance variables that one comprehensive and authoritative critical review
would expect particular HR practices to have of scientific evidence in relation to the reliability
been linked to have not been spelt out. and validity of measures of HRM. Specifically, the
4. Research studies have not tended to be research comprises the following five subsections
designed to test multiple competing for which the research must make explicit the
hypotheses. evidence in order to achieve its objectives.
5. How the high-involvement management
model links to other elements of management Describe HRM methods and measure
practice and theory and particularly leadership the fidelity with which they may be
corporate governance has been neglected. implemented (Objective 1)
We understand this objective of the research is
In order to overcome these weaknesses any concerned with identifying the main methods of
taking-stock exercise should be broad ranging, HRM, how widespread is their use, the fidelity with
well beyond the narrow confines of the recent which they are implemented (i.e. the degree to
HRM performance studies. There is also a need which HRM policy is put into practice), and how
to examine in more depth the applicability of that fidelity can be assessed. This is potentially a
practices to all settings and how they may have to very wide literature with different approaches to
be adapted in specific service organisations. Health the definition and measurement of specific HRM
care provides an excellent example to aid this. methods. The implications for this research are
discussed below (see The scale of the relevant
The importance for policy of finding out if there literature).
is a link between a given model of HRM and
performance is self-evident. But the above implies Review the reliability of measures of
that the key issues are: is any relationship general particular intermediate outcomes
or specific? Is it an underlying orientation towards (intended outcomes of HRM that may
development and involvement or is it certain key effect patient care indirectly, e.g.
practices that is key? If the latter, are there a few morale, absences from work, stress)
critical ingredients or is there a holistic set that (Objective 2)
react together synergistically? Is HRM largely a We understand this research objective is concerned
support for operational practices? If there is a link, with the robustness and utility of measures of a
which are the key performance measures that are number of intermediate outcomes, i.e. the results
most affected by HRM? And, above all, why and of HRM methods that can impact on patient care.
what are the mediating variables? If the crucial Such an exercise inevitably involves the cataloguing
thing is internal consistency between practices and of intermediate outcomes. These are likely to be
ensuring policy is realised through them, is HRM wide-ranging and, in some instances (e.g. stress
simply one of a number bundle that works, as the measures), command a substantial literature in
configuration approach implies? In the health- their own right. Additionally, as we have argued
care case, the additional question is the extent to above, we will consider as broad a range of
which ideas like lean production can be readily intermediate variables are possible. We extend
translated to the health-care context, and where from the current concern with commitment and
there has been a conscious attempt how successful involvement to other individual-level outcomes,
has it been. All these issues are entailed in the five such as stress, morale and procedural fairness,
objectives of the study as proposed in the tender as well as newer ‘outcomes’, such as proactivity.
document. Clearly, dealing with such a potentially vast and
uneven literature and deciding what it is feasible,
Research objectives and issues pragmatic and desirable to include in relation to
the current study has major implications for the
The next section (Objectives for the research) research methodology and these are discussed
considers each of the research objectives and below (see The state of the relevant literature). 285
Review of the literature on the See below (The scale of the relevant literature) for
correlation of intermediate outcomes to a description of our proposed approach to rating
each other (Objective 3) the evidence on quality criteria and the conclusions
We understand this objective of the research to be that can be drawn from different types of evidence.
concerned with establishing the extent to which
individual measures of intermediate outcomes Investigation of the sensitivity of
contribute unique information. An important issue intermediate outcomes to individual
to note here is that whilst the general literature HRM methods alone or in conjunction
may indicate a high degree of correlation between with each other (Objective 5)
certain intermediate outcome measures, it is As with Objective 4 of the research, we understand
possible that these measures will behave differently this element of the study to be concerned with
when considered in relation to specific HRM evidence both of the association between the
methods or final outcomes (relevant to research presence of HRM methods and intermediate
Objectives 4 and 5). outcomes, as well as with evidence of any causal
link between changes in HRM methods and
Interrogation of the literature intermediate outcomes. That is the extent to which
investigating the correlation of there is evidence that, for example, changes in
intermediate outcomes with final HRM practices cause reductions in self-reported
outcomes (be these business or patient stress. As for the previous subsection of the
centred) (Objective 4) research, there is a large body of cross-sectional
We understand this objective of the research to research detailing evidence of the co-existence of
require evidence on two levels: specific HRM methods and intermediate outcomes.
However, research of the type that can provide
• the correlation (i.e. level of association) evidence of the extent to which HRM methods
between the presence of certain intermediate bring about specific intermediate outcomes is much
outcomes (e.g. improvements in morale) and more limited and likely to amount to only a few
the presence of certain final outcomes (e.g. studies in total. Once again, the same implications
improvements in clinical quality standards), for the research are identified and these are
and, where possible, discussed more fully below (see The scale of the
• evidence of a causal link between intermediate relevant literature).
outcomes and final outcomes (i.e. evidence
that improvements in morale led to improved Research issues and
clinical quality standards). methodological solutions
Some evidence exists (of a cross-sectional nature) Thinking through the research objectives in the
that demonstrates associations between, for ITT there are a number of technical and scientific
example, job satisfaction (intermediate outcome) issues that this research will need to overcome:
or self-reported stress (intermediate outcome)
and organisational performance (final outcome). • nature of evidence sought
However, limited evidence about the impact of • scale of the relevant literature
both types of outcome can be drawn from cross- • state of the relevant literature
sectional studies. The research concerned with • ensuring work is free from bias.
cause and effect is likely to be a relatively small
literature with only a limited number of studies of Each of these research issues is discussed in
a longitudinal design and sufficiently robust as to turn and the implications for the design of the
demonstrate the intermediate–final outcome link methodology are highlighted.
causally.
The nature of the evidence sought
The issues for the research to address are thus: The five inter-related objectives of this project
call for a comprehensive review of the literature
• how to ensure that all such longitudinal in relation to a number of different questions.
research is identified, and Answering the research questions posed in the ITT
• how to clearly distinguish the level of evidence demands evaluations of different types of evidence
identified and the type of conclusions that can from different sources. Table 62 summarises the
be drawn in the report. information/evidence needs for all five subsections
of the research.
286
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Italics indicate that an awareness of the issues in these areas is implicit, if not specified in the tender document.
The range of evidence needs has several One potential strategy for dealing with the scale of
implications for the research design: the literature, as recommended in the ITT, is the
use of existing systematic reviews of the literature.
• a process that keeps the review firmly focused Where they exist, we will seek to use such sources;
on the research objectives however, it is inevitable that in some areas there
• clear and efficient strategies for identifying will be no such reviews, or that they will be out of
relevant evidence within the literature date or of poor quality and it will be necessary to
• clear rational for including or excluding studies resort to original sources. A key requirement of
from the review the methodology for this research, therefore, is
• clarification about the type of evidence being the need for expertise in the most sophisticated
drawn into the analysis and the conclusions systems for data retrieval and synthesis of evidence.
that can be drawn.
A further consideration in dealing with a
The scale of the relevant literature literature of this scale is to have sophisticated and
The scale of the literature pertaining to this study transparent rationale for the scoping and targeting
is immense. Searches on a single database for the of content relevant literature. In order to deliver
term ‘HRM and performance’ alone would typically this we propose to use an approach based on the
produce in excess of 500 peer-reviewed references. medical model of systematic reviews,49,50 which
The grey literature (i.e. relevant articles that have members of the team have successfully used in
not undergone a peer review process, or individual similar previous studies. Implicit in the proposed
case studies) may also be significant. approach is the use of a standard proforma to
apply relevance criteria to each potential source of
Additionally, many of the relevant intermediate evidence. This is an efficient and reliable method
outcomes for this research (e.g. stress, job of rapidly identifying the most important articles
satisfaction, morale) command extensive literatures from a large body of literature. The approach
in their own right. A search on the relevant is described fully in the methodology (see
databases for each intermediate outcome area Methodological approach, below).
would be likely to generate a list of references,
in some cases, reaching the thousands. For The state of the relevant literature
example, a recent review of Psychosocial Hazard The literature relating to the research objectives,
(Stressor) Measures,49 which looked specifically at particularly Objectives 2–5, is likely to be very
the reliability and validity of measures of stress, uneven in terms of size, depth, availability
identified in excess of 3000 original sources of of reviews and type of questions covered. For
potential relevance, which, when checked against example, in certain areas, recent, comprehensive
criteria for content, elicited in excess of 150 usable systematic reviews of the reliability of intermediate
articles. outcomes measures (Objective 2) exist (e.g. Stressor
287
measures, Rick et al.50). Equally, cross-sectional Our judgement is that it will be possible to provide
studies reporting on the presence of specific HRM easily accessible information about the quality
practices, such as job design or training, and and quantity of evidence in relation to each of the
intermediate outcomes, such as job satisfaction, main relationships of interest within each research
are relatively plentiful (Objective 5). In contrast, subsection through the development of quality
longitudinal studies demonstrating causality ratings for the review. This is an approach that
between, for example, organisational commitment has been used previously to aid the assimilation
and final outcomes, such as, for example labour of wide-ranging evidence about different subjects
productivity, are relatively scarce (Objective 4) as (see, for example, the Department of Health715
would be studies of specific HRM practices and publication on selection of appropriate mental
‘new’ intermediate outcomes such as proactivity health interventions and the British Psychological
(Objective 5). Society review of Level B measures, Bartram et
al.716). We feel that this will be particularly useful
A further consideration here is that, dependent for Objectives 4 and 5 of the research, where
on the availability of evidence, the research the ITT specifies that the intended audiences
subsections call for the comparison of very different are general (and will not necessarily be research
types of evidence. For example, the methodology literate).
will need to make consistent comparisons across
systematic reviews, meta-analyses, traditional The research team (e.g. Michaels and Booth717) are
reviews, empirical peer-reviewed articles, case aware of the existence of numerous such schemes
studies and grey literature. A further complication that have been used to classify the literature in this
inherent in the nature of HRM research is that the way.717–722 Rather than invent yet another variation,
approaches often combine very different types of we would attempt to identify an appropriate
data. For example, studies show the use of different existing classification of strength of evidence,
measures of individual HRM practices, variation in based on the type of question being asked and the
HRM system composition across studies, variation type of evidence being used to address it. Critical
in the techniques researches use to combine elements of such a classification would include
HRM practices into coherent groups, different whether the research is theoretical, observational
levels of analysis (e.g. HRM and organisational or experimental, the overall robustness of the study
performance data at workplace, business and design (where experimental) and whether there
corporate level, and intermediate data at individual is homogeneity or heterogeneity of results. Such
and organisational level) and a combination of a grading system would be agreed on completion
objective and self-report measures of, for example, of the criteria and identification of the key
performance (Wall and Wood;712 Wright and relationships of interest. However, for the purposes
Gardner714). of illustration, the system might (for illustrative
purposes only) resemble the following:
In order to ensure consistency in approach when
comparing such different types of evidence and • theoretical knowledge suggests X (*)
data, we will develop a system for recording the • one cross-sectional study or several with
various types of research design, types of data varying results (**)
collected and the quantity of data identified in • several cross-sectional studies with the same
each source of evidence. This will be based on results or one longitudinal study (***)
existing classifications, but be tailored to the issues • several longitudinal studies with conflicting
and complexities of the HRM literature. Such an results (****)
approach is implicit within the systematic review • several longitudinal studies with consistent
process. results (*****)
• randomised controlled trial(s) (******).
Aside from proforma for reviewing each individual
source of evidence, a primary concern of this Selection of an appropriate classification system
review is to establish what can be said about each of might well vary according to the different levels
the five subsections of the research overall on the of evidence sought in relation to the research
basis of theoretical and empirical knowledge. This subsections specified in the ITT.
should not only reflect the state of the literature
overall, but should identify any uncertainties in the Ensuring the research is free from bias
evidence base, drawing attention to areas where A major task for the researchers undertaking this
the evidence is strong or weak, or where there is no review is to demonstrate that the work is conducted
288 reliable evidence. in a way that is free from bias. Previous research by
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
members of this team in the area of reviews (e.g. of evidence around a given research question
the psychosocial hazards research by Rick et al.49), and making recommendations for practice. This
has highlighted the need for explicit, transparent approach involves a number of stages:
criteria to be developed for literature reviews. These
criteria should reflect generally agreed standards • formulating specific, answerable questions for
within research and should be developed in each of the research subsections
line with the specific aims and objectives of the • devising the most parsimonious search
project. It must also be possible to demonstrate strategies for identifying relevant data
that these criteria are applied by the researchers • developing inclusion and exclusion criteria
in a systematic way (e.g. through checks including (based on relevance to the research questions
inter-rater reliability). This ensures that the review and minimum quality standards) for sifting all
process is as free from bias as it is possible to make identified sources of evidence
it, and that any other researchers conducting the • developing review proforma for assessing the
same task with the same criteria would achieve quality of evidence in each data source (passing
similar results. This further reinforces the the inclusion/exclusion criteria sift) in relation
arguments for adopting a systematic review process to the specific review questions
in line with the approach that we will propose here. • synthesising the assessments for each specific
review question
In addition to using a systematic approach to the • summarising and drawing conclusions on the
review process, we propose to establish a panel evidence in relation to each.
of experts to contribute to specific stages of the
research. Expert advisers would be appointed Each of these stages is described fully below. A
on the basis of their expert knowledge of the potential difficulty of this approach is that often
HRM literature and/or their knowledge of the the available evidence about any question is small
health sector and/or the contribution they can in quantity, of poor quality or inconsistent. For this
make (through experience) to developing the reason the review will demonstrate a sensitivity to
appropriate processes for conducting this type of the particular requirements of systematic reviews of
review. The rationale is twofold: the management literature.56,723 This will recognise
the need to consider the best available evidence,
• It provides an efficient way of assessing leading even where rigorous experimental studies are not
edge knowledge on HRM methods and available, and, if necessary, where no empirical
literature, the context for this research and the evidence exists, resorting to the best theoretical
best procedural approaches for conducting this approaches. In doing so, the collaborating team
type of review. will draw on their considerable experience of
• Given the immature nature of the evidence different types of integrative synthesis of evidence-
base, it seeks to achieve consensus in the based approaches to general management
processes for identifying the best evidence topics.49,50 They will draw on systematic reviews
available in relation to the research questions of the health management literature (National
specified in the ITT. Institute of Clinical Studies80) in order to inform an
appropriate methodology.
Such a consultative approach should ensure
that the processes developed for conducting the Research stages
research are as free from bias as possible, and as
relevant as possible to the focus of this research. We envisage a methodology comprising the
following stages.
The next section describes the proposed
methodological stages for this research. Establish expert advisor panel
One of the first steps in the project will be to
Methodological approach establish the expert adviser panel, which will have
Systematic literature reviews an number of roles:
The literature review will be conducted according 1. helping to identify experts for the interviews
to established evidence-based principles developed 2. contributing to the list of HRM practices
in medicine and allied health fields, as a systematic 3. suggesting additions to the identified literature
way of pulling together and assessing the quality for each subsection of the research
289
• the macro level (for terms such as HRM, HR e-mail, introducing the project, informing them
practices, personnel, staff recruitment, staff of publications already identified and asking them
development) for the references for other relevant published or
• the micro level for specific HRM methods unpublished evidence. This request will be followed
as identified in Part I of the research (e.g. up by up to two reminders as necessary.
worker participation, training, job security,
promotion, teamworking, incentives); Towards the end of the project a brief
intermediate outcomes (e.g. morale, stress, supplementary search for recent journal articles
proactivity, commitment); and final outcomes and publications will be conducted to ensure an up-
(e.g. performance, organisational performance, to-date bibliography and evidence base.
turnover, productivity, retention, clinical
standard, patient care). As the need is for transparent criteria, it is
essential that our methodology in identifying and
These computer searches will be supplemented reviewing evidence is both explicit and objective.
by citation searching (to detect research studies We therefore propose to conduct this review by
that share a common ancestry) and related-article adhering strictly to predetermined criteria for
features to allow the generation of related concepts. assessing published work. Such criteria will reflect
Hand-searching and examination of relevant generally agreed standards within research,
journals (examples include: Academy of Management developed to meet the specific aims and objectives
Journal, European Journal of Innovation Management, of the project (e.g. Popay et al.,724 Oxman725).
Health Care Management Review, Health Services In demonstrating that these criteria have been
Research, International Journal of Human Resource applied in a systematic way this will ensure that the
Management, Journal of Healthcare Management, review process is as free from bias as is possible,
Journal of Management Development, Journal of ensuring reproducibility for complementary
Managerial Psychology and Journal of Organizational projects or subsequent work.
Behavior) will be supplemented by tightly focused
searches of the internet using advanced search The methodology for the literature review process
features. will involve four main stages, each building on
the previous one. Unlike many reviews where
New systematic reviews the search processes are conducted early in the
We anticipate that Objectives 3 and 4 of the project and then retrieved materials passed on
research in particular will require in-depth to the review team, the iterative nature of the
appraisal and analysis of existing reviews together project combined with the task dependencies
with the conduct of a new review. The theoretical arising from Objectives 1–5, will require ongoing
framework for the new review will divide the interaction with the expert advisers needing to be
mechanisms identified into those where the consulted at each stage. For example, the compiled
evidence base derives from within the health bibliography will be circulated to the expert
literature, those where the literature originates panel on a number of occasions with a request for
from general management and those where both additional evidence omitted from the list. Further
literatures are represented. Each item of evidence items will thus be added to the database, and the
identified will be assessed both for the strength of updated bibliography circulated again asking the
its signal regarding application to HRM in health experts to check for omissions.
organisations and its noise regarding limitations of
research design.73 A systematic literature review will Summary of stages of review for
be conducted using accepted systematic literature Objectives 2–4
search techniques for the years 1995–2004. Online For each substantive review, the following stages
searches of relevant computerised bibliographic will be undertaken:
databases will be undertaken using comprehensive
permutations of key words representing • formulation of specific questions to be
mechanisms and key words depicting outcomes. answered
Databases to be searched will again include: • development of inclusion and exclusion criteria
MEDLINE, PsycInfo, the Emerald databases and based on the relevance of the articles to the
Web of Science. review questions and minimum quality criteria
• development of review criteria for assessing the
The expert advisors will identify key researchers quality of evidence
in the area. These researchers will be contacted by • training of review team and piloting of criteria
291
• conduct of review and data extraction within the team and resolved prior to completion
• synthesis of data and analysis of results. of the reviews. While this approach necessarily
deviates from idealised conditions of quantitative
Bibliographic software review, whereby two reviewers extract data for every
The reference management software package included paper, we believe that this approach is
reference manager will be used for this review. This justified by the nature of the evidence types being
package allows direct transfer of references from considered. In contrast with formal meta-analysis,
the electronic databases. References will be coded there is little empirical methodological research
according to source of the reference, whether the to indicate incremental benefits for dual data
paper is for inclusion, and, if so, to which of the extraction and analysis where non-quantitative
objectives it will likely contribute. This system of research is being examined. However, all issues or
coding will ensure that lists of publications can be queries arising from the extraction or analysis will
easily retrieved according to topic area or purpose be referred for a second opinion (either to another
within the review. researcher or a member of the expert panel), and
where a candidate homogeneous quantitative
Objective 5: to assess the subset is identified these papers will be subject to
sensitivity of intermediate dual extraction and analysis.
outcomes to individual HRM
Initially, evidence tables containing such details as
practices citation, setting, study type, level, scope/variables,
By this stage of the review it is anticipated that outcomes/results and comments will be produced
the team will have identified points of theoretical by the research associates. These will then be
saturation. For each combination of methods and subjected to detailed analysis and quantitative/
outcomes (intermediate and final) they will identify qualitative review by the experts on the review
methods where there is: team.
TABLE 63 Timetable
Journal and SDO Newsletter. We will also undertake Occupational Psychology Conference, European
presentations of research plans and objectives, Academy of Occupational Health Psychology,
intermediate findings and overall final findings at European Association of Work and Organisational
conferences, seminars and professional meetings, Psychology, and the Academy of Management
such as SDO conferences, Public Health Association Conference.
Conference, British Psychological Society
293
Appendix 2
HRM practice categories and definitions
HR category Definition
Work design Organisations have particular functions to accomplish if they are to meet their objectives.
Those functions comprise a number of tasks, which are then grouped to form jobs undertaken
by individuals. Job incumbents typically are trained to carry out their prescribed tasks, and
given a certain degree of discretion over how they do so. ‘Job design’ refers to the outcome of
this process and may be defined as the specification of the content and methods of jobs
Other terms often used as synonymous for job design include ‘work design’ and ‘job’ or ‘work
structuring’. ‘Work organisation’ is also frequently used to encompass job design, but usually
signifies broader perspective linking jobs more explicitly to their organisational context
Within the field of organisational behaviour job design has specifically focused on properties of
the job, referred to as ‘job characteristics’, such as the variety of tasks in jobs and the amount
of discretion job incumbents have in completing those tasks
Staffing: Recruitment/ Traditionally, recruitment has been conceptualised as the organisational process that precedes
selection/placement/ selection. More specifically, recruitment has been described as the process of generating
working hours/staffing pools of applicants for job vacancies; selections as the process of choosing from among those
levels/skill mix applicants. Recruitment has been regarded as an attraction activity, selections as a screening or
winnowing activity
More recently, recruitment and selection have come to be regarded as less distinct activities
Training and development When learning events are planned in a systematic fashion and are related to events in the
work environment, they are called training programmes. From this point of view, the training
process is defined as the systematic acquisition of skills, rules, concepts or attitudes, which
results in improved performance in the work environment (Goldstein 1993726)
Most training programmes have the following stages:
Need assessment: Consists of a series of analyses that assess the organisation, the job and
the person performing the job in order to provide input for the design and evaluation of the
training system. This stage requires determination of which tasks are required on the job and
which knowledge, skills and abilities are necessary to learn to perform those tasks
Training environment: Once the tasks, knowledge, skills and abilities and objectives have been
specified, the next step is designing the environment to achieve the objectives. Training is a
delicate process that requires a supportive learning environment. The training process must be
designed to facilitate the learning of the knowledge, skills and abilities required to perform the
tasks that the trainee needs for successful job performance
Training evaluation: The number of different types of objectives that organisations hope training
programmes can achieve varies widely. However, the true efficacy of any training programme
lies in how the programme is evaluated against pre-determined objectives. The evaluation
process centres around two procedures – establishing measures of success (criteria) and using
research designs to determine what changes have occurred during the training and transfer
process. Criteria must be established for both the evaluation of trainees at the conclusion of
the training programme and the evaluation of on-the-job performance
Compensation and Pay can be in cash or benefits, such as health care, a paid vacation or a company car. Payment
rewards systems vary within and across organisations in their mix. Some organisations pay all
employees a base salary whereas others use variable pay where some portion of employees’
pay is uncertain. For example, performance-related pay refers to a number of pay programmes
that link pay to the individual, group and organisation level performance measures. Pay
programmes that influence pay mix are merit pay, incentive pay, gainsharing, profit-sharing and
stock options. Pay can also be described in terms of its level, i.e. how much organisations pay
for specific jobs, and in terms of its structure which refers to the nature of pay differentials
within an organisation. Payment systems can influence attitudes and behaviours of employees,
and which employees are attracted to the organisation and which decide to leave
295
HR category Definition
Communication Through communication, organisations and their members exchange information, form
understandings, coordinate activities, exercise influence, socialise, and generate and maintain
systems of beliefs, symbols and values. Communications has been called the ‘ “nervous system”
of an organised group’ and the ‘glue’ that holds organisations together
Due to the complexity of the organisational communication process and the many levels
at which communication occur, there is no generally agreed theory of organisational
communication. Different positions have been advanced on several issues
There are at least two positions on the role of communication in organisations. One regards
communication as a subprocess that plays an important role in other organisational processes.
For example, communication serves as a channel for the exercise of leadership or for the
maintenance of interorganisational linkages. The other position argues that communication is
the process that constitutes the organisation and its activities. Rather than being a subsidiary
to key phenomena such as leadership, communication is regarded as the medium through
which these phenomena and, more generally, organisations are created and maintained
Organisations have two distinct communication systems: formal and informal. The formal
communications system is a part of the organisational structure and includes supervisory
relationships, work groups, permanent and ad hoc committees, and management information
systems. The informal communication system emerges from day-to-day interaction among
organisational members. Ties in the informal network are based on proximity, friendship,
common interests and political benefits more than formal job duties. The informal system
includes the ‘grapevine’ and the ‘rumour mill’
Employee/family friendly The term ‘work and family’ or the ‘family friendly workplace’ is one that recognises the non-
workplace family responsibilities of its employees, and develops and implements policies
that allow employees to simultaneously fulfil work and family responsibilities. Different
commentators propose different definitions, usually accompanied by a list of policies that
allegedly meet their criteria. Hartin (1994: 76),728 for example, describes ‘family-friendly’
policies as those ‘designed to minimise the impact of work on family life’ and identifies the
‘family-friendly’ policies currently in force as varieties of leave for maternity or paternity,
sickness, emergencies and compassionate reasons; career breaks and extended leave;
flexitime, part-time work and rostered days off. Gary Johns, Assistant Minister for Industrial
Relations in 1994 in Australia, identified family-friendly work policies ‘as flexible working
arrangements, permanent part-time work, job-sharing, career break schemes, paid or unpaid
family leave, and assistance with child care and elder care responsibilities that can help workers
balance their work and family responsibilities’ (Work and Family 1994729)
Single status/status Single status refers to the provision of common terms and conditions of employment to all
harmonisation/status employees of a single enterprise. It means, essentially, that all workers of whatever grade
differentials will share the same set of employment benefits (pension, sick pay, annual leave, parental and
maternity provisions) and be paid through the same mechanism (e.g. annual salary). Single
status contrasts with the traditional ‘status divide’ between hourly paid manual workers and
salaried staff, and may arise either from trade union negotiating pressure or from a desire on
the part of the employer to foster teamwork and commitment. The achievement of single
status can be costly and for this reason there may be incomplete harmonisation of employment
conditions across the workforce hierarchy
Harmonisation is referred to as movement away from separate terms and conditions of
employment for manual hourly paid and white-collar salaried employees, towards single status,
i.e. common basic terms and conditions of employment. Harmonisation most frequently
affects holiday and leave entitlement, redundancy pay and occupational pensions. It may also
have a symbolic element involving the use of common restaurant and car-parking facilities or a
standard company uniform
Employee representation/ Employee involvement has been used to describe a wide range of practices in organisations.
involvement/participation Common to all these practices is the attention paid to increasing employee influence over how
their work is carried out or over other areas of organisational policy and practice
HR category Definition
Higher-level concept Organising the work process so that non-managerial employees have the opportunity to
(HPWP/HPWS) contribute discretionary effort is the central feature of an HPWP. HPWPs emphasise the
devolution of the gathering and processing of information to the level of non-managerial
employees, who act on and use the information for problem-solving and decision-making.
No one has consistently defined, or even uniformly named, HPWPs (Bakker 1999;730 Becker
and Gerhart 1996;23 Delaney and Goddard 1997;731 Wood 199935). They have been called
HPWSs, alternate work practices, and flexible work practices (Delaney and Goddard 2001731).
Despite the name variances, many of these programmes share common elements, including
rigorous recruitment and selection procedures, incentives based upon performance, and
extensive training programmes focused on the needs of the business (Becker et al. 1997153).
Essentially HPWPs require heavy investment in human capital that is intended to enhance
employee skills, knowledge, motivation and flexibility with the expectation that the employer is
providing employees the ability and the opportunity to provide input into workplace decisions
(Van Buren and Werner 1996732). Companies expect this empowerment to enable employees
to adapt quickly and readily to rapidly changing product and labour market conditions, and to
improve operational efficiency and firm performance (Becker and Huselid 1998;16 Cappelli and
Neumark 2001149)
297
Appendix 3
HRM practice categories and associated
terms from the literature
HRM practice categories Example terms for HRM practices from the literature
1 Work design Team working
Job design
Job rotation
Work characteristics, job control, job demands
2 Staffing, recruitment and Computer-based orientation programme
selection Employee referrals as a hiring practice
Use of WABs
Orientation programme
Selection processes; employee perceptions of selection processes
3 Training and development Cultural learning interventions
Time management training
Feedback training
Cultural diversity awareness training
Training in reflective communication
Teacher training in probation schools
Lifelong learning and training
Quality awareness training
4 Compensation and reward Labour–management cooperation and trust
Compensation system, merit pay, bonus plans
Group incentives
Economic incentives
Incentives, performance awards; special act or service awards; quality step
increases; time-off awards
Promotion
Contingent rewards
Incentives: Financial bonus; enhanced fees
Gain-sharing
5 Communication Communication briefings
Communication briefs
Communication skills
Contact among team members
Feedback
Goal communication
Opportunity to participate/communication
Reinforcement schedule and feedback function
Two-way communication
6 Family friendly/work–life WIF, FIW
balance Utilisation of available workplace support
Employer-provided health insurance
Work–life balance
299
HRM practice categories Example terms for HRM practices from the literature
7 Single status/harmonisation None
8 Employee involvement Employee participation in the job evaluation process
PPM system
9 Appraisal/performance Performance feedback
management Performance appraisal feedback
PMS; participation in implementation of a PMS
Goal setting, goals and personal projects, goal setting and appraisal
Supervision: No supervision; direct human supervision; and computer monitoring
Positive attendance improvement programme; recognition programme: Personal
attention; public celebration; use of mementoes; time-framed recognition; clarity
of expectations; source of acknowledgement
Performance management
Appraisal; participation; goals
Reinforcement schedule and feedback function
10 Higher-level concept (HPWP/ Opportunity to participate/communication; skills; incentives; appraisal systems;
HPWS) PRP; flexible employment practices
FIW, family interfering with work; HPWP, high-performance work practice; HPWS, high-performance work system;
PMS, performance management system; PPM, participative productivity management; PRP, performance-related pay;
WAB, weighted application blank; WIF, work interfering with family.
300
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Appendix 4
Guidance on monitoring
implementation fidelity
There should be some
T his guidance highlights the necessary
components and steps that need to be satisfied
or followed if a policy is to be implemented as it
monitoring of compliance
• Monitoring to ensure that those responsible for
should be, and so have its intended effect. the delivering the practice are doing so.
• Monitoring to ensure that the practice being
There should be a policy delivered is in accordance with the written
document describing the policy in terms of those covered by it, its
practice frequency and duration.
• By preference, this monitoring should consist
There is a greater chance of high implementation of independent observation of the policy
fidelity if the description is simple, but the in practice, but may also include employee
document should provide as many of the following perceptions and experience of the practice.
details of the practice as possible:
There should be strategies
• whom the practice is to cover (all staff or only in place to ensure continued
certain staff) implementation in the long
• who is to deliver it or be responsible for its
implementation (e.g. line manager, HR staff)
term
• when, how often and for how long the practice • Training for new members with responsibility
should be delivered. for delivering the practice.
• Ongoing support for and monitoring of those
There should be strategies to delivering the practice.
help implementation of the • Long-term monitoring to ensure that the
practice practice being delivered is in accordance with
the written policy in terms of those covered by
• Training for those delivering the practice. it, its frequency and duration.
• Documents, manuals or guidance for those
delivering the practice.
• Ongoing support for those delivering the
practice.
301
Appendix 5
Implementation fidelity checklist
303
If so, fill in a separate list for each practice relating to that policy
(eg. different types of appraisal, training, payment systems)
PRACTICE
1. According to this written policy, how often should the practice take place?
3. According to this policy, what are the various elements of this practice?
(eg. appraisal: to be by line manager; to be 1 hour; to offer employee opportunity for
feedback)
304
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
4. Which of these elements of this practice are ‘core’, that is, are considered
essential to the practice having its desired effect?
4.a. How have the ‘core’ elements of the practice been decided?
3. Were guidelines or a manual provided for people responsible for delivering the
practice?
4. Were the people affected by the practice involved in its development or their
opinion asked about the practice?
305
Details
5. Was the quality of delivery by people responsible for the practice monitored?
Details
(eg. Feedback, performance monitoring)
5. Was the quality of delivery by people responsible for delivering the practice
5. Was the quality of delivery by people responsible for the practice monitored?
monitored?
(eg. Feedback, performance monitoring)
(eg. feedback, performance monitoring)
5. Was the quality of delivery by people responsible for delivering the practice
Yes No Don’t know
monitored?
performance monitoring)
(eg. feedback,
Details Yes No Don’t know
Details
6. Was there any ongoing support provided to the people responsible for delivering
the practice?
(eg. support services, helplines, technical help)
6. Was there any ongoing support provided to the people responsible for delivering
Yes No Don’t know
the practice?
(eg. support
services, helplines, technical
help)
Details Yes No Don’t know
Details
7. Did the practice require any new facilities or materials, and if so, were they made
available?
306
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
8. Is there formal training for new members of staff responsible for delivering the
practice?
9. Are guidelines or a manual provided for new members of staff responsible for
delivering the practice?
307
If a practice is only called upon irregularly, how many eligible employees have called
on the practice when they wished or were required to do so? (eg. parental leave,
working from home)
4. Are the other elements of the practice, as described in Question C3, being
implemented?
1 Yes No
2 Yes No
3 Yes No
4 Yes No
5 Yes No
5. Have the designated ‘essential’ elements of the practice been implemented?
1 Yes No
2 Yes No
2 Yes No
F. Results
Very many Quite a lot Some Very few None Don’t know
308
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
309
Appendix 6
Patient outcomes search strategy
311
Appendix 7
Typical HRM and longitudinal search filters
Database: Ovid MEDLINE 2. longitudinal$.tw. (71367)
<1966 to November Week 3 3. exp Retrospective Studies/ (248824)
2005> 4. retrospective study.tw. (39390)
5. exp Prospective studies/ (209208)
Search strategy 6. prospective study.tw. (56189)
1. exp Personnel Management/ (84372) 7. quasi-experimental.tw. (1827)
2. (human resource$ or HRM).tw. (2180) 8. (follow-up$ or follow$ up).tw. (346566)
3. ((HR and (organi?ation$ or staff$ or employ$)) 9. repeat$ measure$.tw. (12546)
not (hazard$ or ratio$)).tw. (2020) 10. (post-test or posttest or (pre adj5 post)).tw.
4. ((high performance or high involvement or (24056)
high commitment) adj2 work).tw. (29) 11. (T1 or T2 or T3).tw. (73071)
5. or/1–4 (87742) 12. baseline.tw. (148159)
13. (Over time or time period).tw. (55619)
Database: Ovid MEDLINE 14. interrupted time series.tw. (234)
<1966 to May Week 1 2006> 15. predict$ design$.tw. (64)
16. exp Case Control/
Search strategy 17. exp Cohort studies/
1. exp Longitudinal Studies/ (539439) 18. or/1–17
313
Appendix 8
Example Business Source Premier
and PsycInfo search filters
Business Source Premier 2. HRM* or Human resource manag* or Human
(Ebsco) resourc* or high perf* high commitment*
S1 (DE “PERSONNEL management”- exploded) (4538 records)
(154567 records) HRM* or Human resource manag* or Human
S2 TI ( human resource* or HRM ) Or AB ( resourc* or high perf* high commitment* or
human resource* or HRM ) (27272 records) personnel managem* (14831 records)
S3 TI HR Or AB HR (13945 records) HRM* or Human resource manag* or Human
S4 TI ( high commitment work or high resourc* or high perf* high commitment*
involvement work or high performance work or explode personnel management (17336
) Or AB ( high commitment work or high records)
involvement work or high performance work ) 3. explode ‘Longitudinal-Studies’ in DE (14229
(209 records) records)
S5 TI (organisation* or organisation* or 4. ‘Repeated-Measures’ in DE (260 records)
staff* or employ*) Or AB (organisation* or explode ‘Prospective-Studies’ in DE (264
organisation* or staff* or employ*) (603395 records)
records) 5. Retrospective stud* (1702 records)
S6 (S1 OR S2 OR S3 OR S4 OR S5) (813114 explode ‘Retrospective-Studies’ in DE (243
records) records)
S7 TI ( longitudinal* or quasi-experimental or 6. ‘Followup-Studies’ in DE (12282 records)
retrospective study or prospective study or 7. (explode ‘Longitudinal-Studies’ in DE) or
followup or follow-up ) Or AB ( longitudinal* (‘Repeated-Measures’ in DE) or (explode
or quasi-experimental or retrospective study ‘Prospective-Studies’ in DE) or (Retrospective
or prospective study or followup or follow-up ) stud*) or (explode ‘Retrospective-Studies’
(11694 records) in DE) or (‘Followup-Studies’ in DE) (28266
S8 TI ( follow* up or repeat* measure* or baseline records)
or post-test or posttest ) Or AB ( follow* up or 8. (HRM* or Human resource manag* or Human
repeat* measure* or baseline or post-test or resourc* or high perf* high commitment*
posttest ) (9196 records) or explode personnel management) and
S9 TI ( T1 or T2 or T3 or over time or time ((explode ‘Longitudinal-Studies’ in DE) or
period or interrupted time series or predict* (‘Repeated-Measures’ in DE) or (explode
design ) Or AB ( T1 or T2 or T3 or over time ‘Prospective-Studies’ in DE) or (Retrospective
or time period or interrupted time series or stud*) or (explode ‘Retrospective-Studies’ in
predict* design ) (18559 records) DE) or (‘Followup-Studies’ in DE) (213 records)
S10 (S7 OR S8 OR S9) (39449 records) 9. 1 or 2 or 3 or 4 (25082 records)
S11 (S6 AND S10) (1323 records) 10. 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 (57459)
11. 13 and 14 (289 records)
PsycINFO
1. explode ‘Personnel-Management’ in DE (14118
records)
315
Appendix 9
Discussion of bias in meta-analyses
Publication bias Extractor bias
The tendency for academic journals to publish The extraction of data for the meta-analysis should
only studies with significant results means that be ‘explicit, unbiased and reproducible’ (Cook
any meta-analysis based on a limited search of et al.659) If only a single reviewer is involved in
the literature may be open to publication bias coding or extracting data from studies for a meta-
because smaller studies demonstrating no effect, analysis or review, or if no assessment is made of
or a negative effect, may be being omitted.662 the consistency between two or more reviewers
This form of bias is the most researched.61,666 The performing coding or data extraction (e.g. inter-
failure to conduct a comprehensive multimethod rater reliability score), then there is an increased
literature search of various sources, including so- likelihood of bias or error entering the analysis
called ‘grey literature’, i.e. unpublished studies because the data on which it is based may not have
and research published in sources other than been recorded consistently.662 The potential degree
peer-reviewed journals, increases the likelihood of of this so-called extractor bias is therefore assessed
this publication bias. The potential for relevant, by the checklist.
published literature or evidence to be missed is
also increased. If this occurs, the meta-analysis Quality appraisal
may be affected by so-called publication bias. This
is because a limited literature search may only A major source of bias also resides in the
find those studies listed in one or two of the most methodological quality of included studies. Giving
popular databases or catalogued with a particular equal weight to studies of very different quality
keyword or term. Consequently, other, equally may have an adverse effect on the validity of the
relevant and potentially important studies listed findings of a meta-analysis; this is emphasised in
in other databases and catalogued using slightly all of the literature evaluating meta-analyses.659–662
different terms may therefore be missed. The If there is variation in the methodological quality
findings of any analysis based on such a limited of included studies then there is an increased
literature search would therefore be compromised likelihood of a ‘false-positive’ conclusion or Type
because it would not represent all of the relevant II error (that there is an effect, when in fact
and available evidence on the topic. The checklist there is not), if, for example, less rigorous or
therefore makes an evaluation of the degree of methodologically robust studies have provided
potential search or publication bias in the included the larger effect sizes in the meta-analysis. On the
meta-analyses and reviews. other hand, these less rigorous studies might also
potentially obscure the results of the higher-quality
Selection bias studies and give a ‘false-negative’ conclusion, or
Type I error (indicating no effect, when there
In order to guard against inconsistency in the is one).725 The failure to conduct some form of
selection of studies, meta-analyses need to develop quality assessment means that either these less
and apply explicit inclusion criteria. The more rigorous studies cannot be excluded or sensitivity
specific and detailed the criteria, the more likely analyses cannot be performed to gauge the effect
it is that only primary research studies satisfying on the overall effect size found by the analysis of
these criteria will be included, and therefore the studies of different quality. The checklist developed
more consistent and homogeneous the sample for this report therefore evaluates whether a
included in the analysis.662 The checklist therefore quality assessment of included studies had been
includes an evaluation of stated inclusion criteria in performed, in order to assess whether potential
order to assess the potential degree of any selection bias or error may have been introduced into the
bias in the analysis. meta-analysis or review from this source. However,
no meta-analysis in this field conducted such an
317
evaluation, so it exists as a potential source of bias this model nor some other appropriately selected
in every meta-analysis or review included in this random effects model has been used.
paper.
Moderator or sensitivity analysis
Analysis
Sensitivity analysis (sometimes called ‘moderator
Random effects models give a more conservative, analysis’) is used to test assumptions about the
cautionary, and therefore reliable, estimate of findings of the analysis by exploring the effect of
effects sizes, that is the size of the relationship excluding certain studies from the analysis, or by
between two variables. Fixed effects models should comparing the results for different subgroups. For
only be used for meta-analyses where there is clear example, by excluding studies with the smallest
homogeneity between effects sizes. A meta-analysis sample sizes, it is possible to determine whether
should therefore test for the homogeneity or the analysis is subject to publication bias.733 Or,
heterogeneity of effect sizes in order to determine by excluding studies of poorer quality, it can
the most appropriate effects model.665,667,733 If a be assessed how far study quality affects the
Q-statistic/chi-squared statistic or credibility interval results. The outcome is a more robust analysis
indicates heterogeneity (i.e. if the Q-statistic or that has tested assumptions that might affect its
chi-squared statistic is significant or the credibility findings.659,662,733 For meta-analyses relating to
interval includes 0) then the analysis needs to use intermediate outcomes, a sensitivity analysis might
a random effects model. In the absence of any involve either comparing the results of overall
such test, a random effects model should be used. correlations with those for particular measures of a
However, it is always preferable to use a random concept only. For example, Lodahl and Kejner’s424
effects model because this test is only capable of measure of job involvement has 20- and six-item
detecting heterogeneity when the heterogeneity scales, and a sensitivity or moderator analysis
of effect sizes is large, and the number of studies would examine results for overall and more specific
is neither too small nor too large.733–735 The failure measures of job involvement.189 The performance
to use the appropriate model might mean that the of such an analysis permits a more rigorous
overall effect size reported is more precise than the assessment of the findings of the meta-analysis.
data can actually support. However, almost all of The meta-analytic process advocated by Hunter
the meta-analyses included in this report used the and Schmidt and Hunter et al.736,737 requires the
random effects model of Hunter and Schmidt736 performance of such moderator analyses, so many
and Hunter et al.737 The checklist therefore records of the included meta-analyses have tested the
whether this model has been used, but reports the robustness of their results in this way.
potential presence of potential bias only if neither
318
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Appendix 10
Checklist to assess bias in meta-analyses
Study
319
Study
320
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Appendix 11
Assessment of bias for included meta-analyses
321
al.739 and Blegen742 were of high quality as they • Judge et al.656 carried out a meta-analysis of 254
only had one potential source of serious bias, while studies comprising 312 independent samples,
the potential for other types of bias was only low. and reported data on correlations between job
Faragher et al.739 was subject to low publication satisfaction and job performance.
and selection bias, and was only vulnerable to high • Organ and Ryan162 conducted a meta-analysis
extractor bias in the coding of studies. Moderator of 55 studies (55 samples), reporting data
analyses were performed by testing the effect on on correlations between satisfaction and five
the results of using both fixed and random effects dimensions of OCB (altruism, civic virtue,
models. Blegen742 had relatively high publication courtesy, sportsmanship and compliance, but
bias because a specific search was conducted to not conscientiousness).
identify unpublished studies, but details of all other • Brown and Peterson,741 as mentioned in
aspects of the literature search are extremely vague. the section on job satisfaction correlations
However, potential selection and extractor bias was with other intermediate outcomes, above,
low; a test for homogeneity was performed and a performed a meta-analysis of 59 studies and,
random effects model was used; and moderator in addition to the correlations between job
analyses were performed by testing the effect of satisfaction and the intermediate outcome
setting and publication date on the results. variables of interest, reported corrected
correlation coefficients for the relationships
The meta-analysis of Brown and Peterson741 was of between job satisfaction and both performance
similar, relatively high quality: it had low selection and turnover.
and extractor bias, but there was potentially a • Carsten and Spector747 conducted a meta-
moderate amount of publication bias. This meta- analysis of 47 studies examining the
analysis used a fixed effects model rather than a relationship between job satisfaction and
random effects model, testing for homogeneity, employee turnover.
which revealed consistency across effect sizes; • The meta-analyses performed by Irvine and
although they did perform moderator analyses by Evans,748 comprising 11 studies, and by Tett
testing for the effect of different populations and and Meyer,749 comprising 155 studies (178
different measures. samples), both examined the relationship
between job satisfaction and turnover.
Finally, the quality of the meta-analysis by • The meta-analysis of Farrell and Stamm750
Hellman740 was very low: publication, selection, was based on 72 studies and reported
and extractor bias were all high (no details were correlations between job satisfaction and two
given of any of these processes) and no information absence variables (total time lost and absence
was given about the effects model used, although frequency).
moderator analyses were performed. • Iaffaldano and Muchinsky157 reported a
meta-analysis of correlations from 74 studies
Correlations with behaviours on various aspects of job satisfaction and its
Studies relationship to performance. Petty et al.751
The literature search identified 121 studies, 14 of performed a similar meta-analysis on the same
which met the inclusion criteria: variables based on 31 studies, but focused
specifically on individual level job satisfaction
• Dalal744 conducted a meta-analysis of 38 studies and performance.
with 49 samples, reporting data on correlations • Steel and Ovalle752 conducted a meta-analysis
between job satisfaction and an overall score of 34 studies and reported data on correlations
for OCB. between job satisfaction and turnover, while the
• LePine et al745 conducted a meta-analysis of meta-analysis of Scott and Taylor753 examined
37 studies (37 samples) and reported data on the relationship between job satisfaction and
correlations between satisfaction and an overall absenteeism in 23 studies (114 correlations).
score for OCB.
• Harter et al.746 conducted a meta-analysis of 42 Critical appraisal of meta-analyses
studies across 36 organisations conducted by The methodological quality of the meta-analyses
the Gallup Organization. They reported data included in this report ranged from very good
on correlations between job satisfaction and to very poor. One meta-analysis was of very high
turnover (measured as a dimension of business quality,656 six meta-analyses were of relatively
unit performance). high,157,162,741,744,745,749 one was of moderate quality747
and five were of only low quality.748,750–753
322
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
The rigorous meta-analysis carried out by Judge et Sensitivity analyses were also performed to test the
al.656 was superior to the other analyses involving robustness of the findings by evaluating the impact
job satisfaction in this section. Although only on effect sizes of different outcome measures.
one electronic database was searched, potential
publication bias was low because searches of The meta-analysis of Brown and Peterson741 was of
unpublished literature, hand-searches of 21 relatively high quality (for details, see Correlations
journals and reference tracking of previous meta- with other intermediate outcomes, above).
analyses took place. Furthermore, the potential for
publication bias is acknowledged and a test carried The meta-analysis of Laffaldano and Muchinsky157
out by Judge et al.656 suggests limited effects on the was also of relatively high quality. There was
results. Potential selection bias was low as inclusion moderate potential publication bias because only
criteria were clear, although no information about one database and no sources of grey literature
the number of reviewers involved or about inter- were searched, although supplementary literature
rater reliability was given. Extractor bias was also searching methods were used. Selection and
low, with inter-rater reliability reported for the extractor bias were both low because inclusion
coding process. The random effects model of criteria were specific and detailed, and the
Hunter and Schmidt736 was used and a test for reliability of the coding was tested and an
the homogeneity of effect sizes was performed. inter-rater reliability score reported. No test of
Moderator analyses, again superior to those of the homogeneity of effect sizes was performed, but a
other meta-analyses included, were also performed, random effects model was used. Sensitivity analyses
testing for the effects not only of different measures were performed to assess the effect on the results
of satisfaction and performance, but also of study of year of publication and the different measures
design and quality. In particular, these moderator of job satisfaction and performance used by the
analyses tested for the effects of longitudinal versus included studies.
cross-sectional studies and of high-ranked versus
low-ranked journals or sources; no other meta- Moderate quality meta-analyses
analyses distinguish these factors. The meta-analysis of Carsten and Spector747 was
of moderate quality. Potential publication bias
The meta-analyses of Dalal744 and Organ and was moderate as there were no sources of grey
Ryan162 were both good methodologically. In literature and only one database was searched,
both cases, there was moderate publication bias although supplementary literature searching
because only a small number of databases or grey methods were used. Selection bias was low because
literature sources were searched, but otherwise inclusion criteria were specific and detailed, but
selection and extractor bias were low. Both studies no details were provided about the coding process
also performed sensitivity analyses. LePine et al.745 so potential extractor bias was high. A test of
had moderate potential publication bias because homogeneity of effect sizes was performed, and a
grey literature was searched and relevant authors random effects model was used, but the sensitivity
contacted, but had low selection bias because analysis by occupational group alone was limited.
inclusion criteria were clear. LePine et al.745 also
performed a test for homogeneity, used a random Low quality meta-analyses
effects model for the analysis, and performed The meta-analysis of Petty et al.751 had a number
moderator analyses by testing for the effect on the of substantial weaknesses. The literature search
results of using different measures of OCB and was extremely limited, consisting only of hand-
measuring different dimensions of this concept. searching a number of journals, so the potential for
Only extractor bias was high. publication bias was high. Extractor bias was also
high: no information was given about the coding
The meta-analysis of Tett and Meyer749 was of of the included studies. Potential selection bias,
moderate to high quality. The literature search however, was low because clear inclusion criteria
could have been more comprehensive, as only were given. There was no test for homogeneity of
one bibliographical database was searched, but effect sizes, but a random effects model was used
multiple methods were used, a specific search was and moderator analyses were performed based on
made for unpublished studies, and the potential the different measures of job satisfaction, such as
for some publication bias in the included sample the JDI.
of studies was acknowledged. Potential selection
bias was low. A test for homogeneity was performed The meta-analysis of Irvine and Evans748 was also of
and a random effects model used for the analysis. poor quality. A test for homogeneity was performed
323
and a random effects model was used, but there motivation. The high-quality meta-analysis of
was a high degree of publication and extractor bias, Farragher et al.739 found a large, positive correlation
and a moderate degree of selection bias. Although between low job satisfaction and burnout (0.48).
sensitivity analyses were performed, these results There were medium to large negative correlations
were not related to job satisfaction. with turnover intentions in two studies, although
one of these meta-analyses was of very low
The analysis of Scott and Taylor753 was of similar quality.740
low quality. The potential for publication bias
was high because the details of the literature A relatively large number of meta-analyses
search consisted of no more than the statement reporting correlations between job satisfaction
that, ‘A comprehensive review of the … literature and employee behaviours were included in this
uncovered 23 studies’. Inclusion criteria were review, therefore only those of high or moderate
limited, and potential extractor bias was high. quality will be included here. Three meta-analyses
However, although no test for homogeneity was reported small to medium positive relationships
performed, a random effects model was used and between job satisfaction and overall measures of
sensitivity analyses were performed, based on OCB, and three reported small to medium negative
different measures of absenteeism. associations between job satisfaction and turnover.
The rigorous meta-analysis by Judge et al.656 found
The meta-analyses by Harter et al.746 Farrell a moderate positive relationship between job
and Stamm750 and Steel and Ovalle752 were satisfaction and performance (0.30), while two
methodologically weaker. In Harter et al.746 other studies found positive but small correlations.
publication bias was high because studies were only
included if they were undertaken by the Gallup Organisational commitment
Organization. Some selection criteria were applied
to the small pool of available studies, although Correlations with other intermediate
these were limited, and no information was given outcomes
on coding of data, so extractor bias may be high Studies
also. However, the form of analysis was good: a The literature search identified 40 studies, 10 of
random effects model was used and subgroup which met the inclusion criteria.175,176,754–761 An
analyses were performed. additional meta-analysis that satisfied these criteria
was identified by reference-tracking of included
In the other two analyses, extractor bias was high meta-analyses.669 The definition of organisational
in both because no information was given on how commitment used by these studies is within the
the coding process was performed. Publication broad definition outlined in Chapter 4:
bias was moderate in Farrell and Stamm750 because
databases were searched and other information • The various meta-analyses performed by
retrieval methods, such as reference-tracking and Cooper-Hakim and Viswesvaran754 were
hand-searching of journals, were used, but no based on findings from 997 articles, which
grey literature sources were searched. Publication included correlations between organisational
bias was high in Steel and Ovalle752 because the commitment and career commitment, job
literature search performed was even more limited. satisfaction and turnover intentions. The
However, selection bias was low in Farrell and primary studies used many different measures
Stamm750 and Steel and Ovalle752 because inclusion to measure these correlates.
criteria were given. Neither analysis tested for • Meyer et al. 669 conducted a meta-analysis
homogeneity of effect sizes, although both did use of 155 samples from both published and
a random effects model. Both analyses performed unpublished studies. The meta-analysis
sensitivity analysis based on the occupation status reported correlations between organisational
of the sample, but Steel and Ovalle752 conducted an commitment and the following intermediate
additional analysis based on time to turnover. outcomes of interest: justice, occupational
commitment, job satisfaction and pay
Results satisfaction. The analysis only included data
Generally, the high-quality meta-analyses reporting from studies that used one or more of the
correlations with intermediate outcomes found commitment measures (affective, continuance
job satisfaction to have a moderate to high or normative) developed by Allen and Meyer,741
positive association with perceptions of fairness, Meyer and Allen174,477,762 or Meyer et al.763
organisational commitment, job involvement and
324
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
325
bias were both low, and each study also performed The analyses of Cohen and Hudecek757 and
moderator analyses. Cohen175 essentially reported the same data.
Results Burnout
The meta-analyses of both Lee et al.177 and
Wallace176 found large positive correlations between Correlations with other intermediate
professional commitment and organisational outcomes
commitment. Lee et al.177 also reported a strong Studies
positive association with job satisfaction (0.44) and The literature search identified 37 studies, two
a strong negative association with burnout (–0.44), of which met the inclusion criteria.186,766 The
although it is important to acknowledge that definition of burnout used by these studies is within
these data came from a single meta-analysis only, the broad definition outlined in Chapter 4:
328
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
• Lee and Ashforth186 conducted a meta-analysis had only a single author so extractor bias was high.
of 61 studies (61 samples). They reported A random effects model was used, but, again, there
data on correlations between the three was no test for homogeneity of effect sizes and no
dimensions of burnout, namely emotional moderator analyses were performed.
exhaustion, depersonalisation and personal
accomplishment (based on the Maslach Results
Burnout Inventory, MBI), and the intermediate The meta-analyses of both Melchior et al.766 and
outcomes of job satisfaction, organisational Lee and Ashforth186 reported medium to large
commitment, job involvement and turnover negative relationships between the burnout
intentions. (overall burnout and the dimensions of emotional
• Melchior et al.766 conducted a meta-analysis exhaustion and depersonalisation for each study,
of nine studies. The small number of studies respectively) and job satisfaction. Although only
is due to the inclusion criteria: the studies reported by Lee and Ashforth,186 the correlations
included in this meta-analysis had to be between burnout (emotional exhaustion and
composed of nurses engaged in psychiatric depersonalisation) and organisational commitment
patient care. Further, the only variable of were negative and of moderate strength, while
interest in this meta-analysis with reported positive moderate to strong relationships with
data on correlations was job satisfaction, and turnover intentions and negative very small to
the mean corrected correlation statistic for this very large associations with job involvement were
variable was based on three studies only. Two of also reported. Taris46 found weak to moderate
these studies used the MBI as their measure of negative associations between two dimensions of
burnout, and one used the Tedium Measure of burnout and OCB. Nevertheless, the conclusions
Pines et al.250 drawn must be treated with caution, as these meta-
analyses had multiple sources of bias.
Critical appraisal of meta-analyses
The meta-analyses of Melchior et al.766 and Lee and Job involvement
Ashforth186 were both of low-to-moderate quality.
Both used a random effects model, although Correlations with other intermediate
neither tested for homogeneity of effect sizes. outcomes
The potential degree of publication, selection Studies
and extractor bias in both studies was moderate The literature search identified eight studies, two
only: some effort was taken to identify all relevant of which met the inclusion:189,754
studies, to select relevant studies, and extract
data with limited bias or error, but more could • Brown189 conducted a meta-analysis of 212
have been done. Also, neither study performed a studies (249 independent samples). The meta-
moderator or sensitivity analysis, so the robustness analysis reported correlations between job
of the findings was not tested. involvement and the following intermediate
outcomes of interest: organisational
Correlations with behaviours commitment, turnover intentions, job
Studies satisfaction and pay satisfaction. The primary
The literature search identified 37 studies, one of studies used many different measures to
which met the inclusion criteria:46 measure these correlates.
• Cooper-Hakim and Viswesvaran754 carried
• Taris46 conducted a meta-analysis of 16 studies, out multiple meta-analyses based on findings
reporting correlations between two of the from 997 articles. Correlations were reported
dimensions of burnout (emotional exhaustion between job involvement and organisational
and depersonalisation) and OCB. commitment, career commitment, job
satisfaction and turnover intentions. The
Critical appraisal of meta-analyses primary studies used many different measures
The meta-analysis of Taris46 was of low quality. to measure these correlates.
There was a high degree of potential publication
bias because only a single database was searched, Critical appraisal of meta-analyses
and no other methods of information retrieval were The meta-analyses that satisfied the inclusion
used. However, potential selection bias was low criteria for this review were all were of reasonable
because clear inclusion criteria were given, but no and comparable methodological quality. The
information was given on coding and the analyst analysis of Brown189 was of slightly higher quality
329
than the two other meta-analyses: there was a and comparable methodological quality. The
moderate degree of publication bias because analysis of Brown189 was of slightly higher quality
only one database of peer-reviewed journals and than the two other meta-analyses (see previous
no grey literature sources were searched, but section for critical appraisal details).
the search strategy was good and both hand-
searching of journals and reference tracking were Farrell and Stamm750 had moderate publication
performed. The inclusion criteria were also limited, bias because several techniques were used to
introducing moderate potential selection bias, but find the evidence and databases were searched,
coding was rigorous so the potential for extractor although database names were not given. However,
bias was low. Brown,189 unlike Cooper-Hakim no searches of formal grey literature sources were
and Viswesvaran,754 also performed a moderator reported. Inclusion criteria were given, so selection
analysis (to test the effect of different measures bias was low, but extractor bias was high because no
of job involvement) and, therefore, the findings information was provided on the coding process.
of Brown189 are the most robust in this group of A random effects model was used, and some
studies. moderator analyses were performed, based on the
occupational status of the sampled population.
The potential for selection bias in the meta-
analysis of Cooper-Hakim and Viswesvaran754 In the meta-analyses of Cooper-Hakim and
was moderate because the inclusion criteria were Viswesvaran,754 both publication and extractor bias
limited and unspecific, whereas both publication were low, as data collection and coding were of high
and extractor bias were low because data collection quality, and the potential for selection bias was low
and coding were of high quality. Although no test to moderate because the inclusion criteria were
for homogeneity was performed, an appropriate limited and none specific. No test for homogeneity
random effects model was used. was performed, but an appropriate random
effects model was used. Moderator analyses were
Correlations with behaviours not performed beyond the focus on the different
Studies dimensions of the measurement of organisational
The literature search identified eight studies, commitment.
two of which met the inclusion criteria.189,754 A
further meta-analysis was identified by reference- Results
tracking:750 Large positive correlations were found between
job involvement and overall organisational
• Brown,189 as mentioned in the previous commitment, while medium to large positive
section, conducted a meta-analysis of 212 correlations were found between job involvement
studies (249 independent samples), reporting and satisfaction. Turnover intentions were reported
correlations between job involvement and the to have a moderate negative association with job
following organisational behaviours of interest: involvement.
performance, absence and turnover.
• The various meta-analyses performed by Small negative relationships were reported with
Cooper-Hakim and Viswesvaran754 were based turnover and small to large negative relationships
on findings from 997 articles, which reported, were found with absence. Brown189 also reported
in addition to the correlations mentioned a very small positive correlation between job
previously in relation to intermediate involvement and performance. However,
outcomes, correlations between job conclusions drawn from these findings must be
involvement and turnover. treated with caution, as all meta-analyses were
• The meta-analysis of Farrell and Stamm750 was subject to multiple biases.
based on 72 studies and reported correlations
between job involvement and the absence Turnover intentions
variables of total time lost and absence
frequency. No meta-analyses of turnover intentions satisfied
the inclusion criteria for intermediate outcomes or
Critical appraisal of meta-analyses organisational behaviours.
The meta-analyses that satisfied the inclusion
criteria for this review were all were of reasonable
330
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
332
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
between each facet and job satisfaction leadership, work group and organisation and,
organisational commitment and psychological in addition to the correlations mentioned
well-being. The primary studies used many above, between these dimensions and
different measures to determine these performance.
correlates.
• Parker et al.674 conducted a meta-analysis of 94 Critical appraisal of meta-analyses
studies (121 samples), reporting correlations Both meta-analyses were of moderate quality (see
between the five facets of psychological climate previous critical appraisal section for details).
perceptions proposed by James et al. (e.g. Jones
and James676) – job, role, leadership, work Results
group and organisation – and between the five The meta-analysis by Carr et al.673 found large
facets and job satisfaction, job involvement, positive relationships between affective climate and
commitment, psychological well-being and job satisfaction (0.46) and instrumental climate
motivation. and job satisfaction (0.44), and a medium positive
correlation between cognitive climate perceptions
Critical appraisal of meta-analyses and job satisfaction (0.33). Medium positive
This meta-analysis of Carr et al.673 was of moderate correlations were reported between the three
quality. Potential selection and extractor bias were climate facets and organisational commitment
both low. The potential degree of publication bias (affective, 0.34; cognitive, 0.28; instrumental, 0.26).
was moderate because although hand-searching A very weak relationship was reported between
of specific journals and reference-tracking was cognitive climate perceptions and well-being
performed, only one electronic database and no (0.07), while weak relationships were reported
grey literature were searched, and no information between affective and instrumental climate and
was provided on the search strategy used. A well-being (0.17 and 0.11, respectively). Job
fixed effects model was used, after testing for performance was found to have very weak positive
homogeneity of effect sizes. Only path analysis, and associations with climate (affective, 0.9; cognitive,
not moderator analysis, was performed. 0.5; instrumental, 0.5), while turnover had a very
weak negative association with cognitive climate
This meta-analysis of Parker et al.674 was also of (–0.07) and moderate negative associations with
moderate quality. Potential selection and extractor affective (–0.28) and instrumental (–0.33) climate
bias were both low, as specific inclusion criteria and perceptions. For further analysis and interpretation
a high consistency score between five raters were of the findings, see Chapter 9 (Correlations
reported. Potential publication bias was moderate between climate, other intermediate outcomes and
because although clear search terms were included employee behaviours).
and reference-tracking was performed, only one
electronic database, and no grey literature, was The meta-analysis by Parker et al.674 reported
searched. The random effects model of Hunter and large, positive correlations between the climate
Schmidt736 was used, but no test for homogeneity facets of leader, work group and organisation and
was conducted and only path analyses, and not job satisfaction (0.41, 0.48, 0.42, respectively),
moderator analyses, were performed. a medium, positive correlation between role
characteristics and job satisfaction (0.28),
Correlations with behaviours and a small, positive correlation between job
The literature search identified two studies that characteristics and satisfaction (0.21). Job
met the inclusion criteria:673,674 involvement and commitment were grouped
together to form the dimension of ‘job attitudes’.
• As previously described, Carr et al.’s673 meta- Job attitudes was found to have a small, positive
analysis of 51 studies (70 samples) reported relationship with role characteristics (0.22), and
correlations between affective, cognitive and medium, positive relationships with job (0.26),
instrumental climate perceptions and between leader (0.32), work group (0.27) and organisation
each of these climate dimensions and job (0.36) characteristics. A strong, positive association
performance and withdrawal (turnover). was reported between leader characteristics
• Parker et al.674 conducted a meta-analysis of 94 and well-being (0.44), while moderate, positive
studies (121 samples), reporting correlations associations were reported between role, job,
between the climate dimensions of job, role, work group and organisation characteristics and
333
well-being (0.26, 0.35, 0.35, 0.29, respectively). characteristics (0.09), and small correlations with
The relationship between role characteristics and job (0.20), leader (0.21), work group (0.20), and
performance was found to be very weak (0.06), organisation (0.22) dimensions. See Table 59 for a
while the relationships between job, leader, work summary of these correlations. For further analysis
group and organisation characteristics with and interpretation of the findings, see Chapter 9
performance were all reported to be small (0.10, (Correlations between climate, other intermediate
0.16, 0.12, 0.13, respectively). Motivation was outcomes and employee behaviours).
found to have a very weak correlation with role
334
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
No. 9 No. 13
No. 4
Implications of socio-cultural contexts
A cost–utility analysis of interferon beta Choosing between randomised and
for the ethics of clinical trials.
for multiple sclerosis. nonrandomised studies: a systematic
A review by Ashcroft RE, Chadwick
By Parkin D, McNamee P, Jacoby A, review.
DW, Clark SRL, Edwards RHT, Frith L,
Hutton JL. Miller P, Thomas S, Bates D. By Britton A, McKee M, Black N,
McPherson K, Sanderson C, Bain C.
No. 10 No. 5
A critical review of the role of neonatal Effectiveness and efficiency of methods No. 14
hearing screening in the detection of of dialysis therapy for end-stage renal Evaluating patient-based outcome
congenital hearing impairment. disease: systematic reviews. measures for use in clinical trials.
By Davis A, Bamford J, Wilson I, By MacLeod A, Grant A, Donaldson A review by Fitzpatrick R, Davey C,
Ramkalawan T, Forshaw M, Wright S. C, Khan I, Campbell M, Daly C, et al. Buxton MJ, Jones DR. 335
No. 7 No. 16
No. 26
The clinical effectiveness and cost- Screening for fragile X syndrome: a Can randomised trials rely on existing
effectiveness of routine dental checks: literature review and modelling. electronic data? A feasibility study to
a systematic review and economic By Song FJ, Barton P, Sleightholme explore the value of routine data in
evaluation. V, Yao GL, Fry-Smith A. health technology assessment.
By Davenport C, Elley K, Salas By Williams JG, Cheung WY,
C, Taylor-Weetman CL, Fry-Smith A, No. 17 Cohen DR, Hutchings HA, Longo MF,
Bryan S, et al. Russell IT.
Systematic review of endoscopic sinus
surgery for nasal polyps.
No. 8 No. 27
A multicentre randomised controlled By Dalziel K, Stein K, Round A,
Garside R, Royle P. Evaluating non-randomised
trial assessing the costs and benefits intervention studies.
of using structured information and By Deeks JJ, Dinnes J, D’Amico R,
analysis of women’s preferences in the No. 18 Sowden AJ, Sakarovitch C, Song F, et al.
management of menorrhagia. Towards efficient guidelines: how to
By Kennedy ADM, Sculpher MJ, monitor guideline use in primary care. No. 28
Coulter A, Dwyer N, Rees M, Horsley S,
By Hutchinson A, McIntosh A, A randomised controlled trial to assess
et al.
Cox S, Gilbert C. the impact of a package comprising a
patient-orientated, evidence-based self-
No. 9
No. 19 help guidebook and patient-centred
Clinical effectiveness and cost–utility consultations on disease management
of photodynamic therapy for wet Effectiveness and cost-effectiveness
and satisfaction in inflammatory bowel
age-related macular degeneration: of acute hospital-based spinal cord
disease.
a systematic review and economic injuries services: systematic review.
By Kennedy A, Nelson E, Reeves D,
evaluation. By Bagnall A-M, Jones L, Richardson Richardson G, Roberts C, Robinson A,
By Meads C, Salas C, Roberts T, G, Duffy S, Riemsma R. et al.
Moore D, Fry-Smith A, Hyde C.
No. 20 No. 29
No. 10
Prioritisation of health technology The effectiveness of diagnostic tests for
Evaluation of molecular tests for
assessment. The PATHS model: the assessment of shoulder pain due
prenatal diagnosis of chromosome
methods and case studies. to soft tissue disorders: a systematic
abnormalities.
By Townsend J, Buxton M, review.
By Grimshaw GM, Szczepura A,
Hultén M, MacDonald F, Nevin NC, Harper G. By Dinnes J, Loveman E, McIntyre L,
Sutton F, et al. Waugh N.
No. 21
No. 11 No. 30
Systematic review of the clinical
First and second trimester antenatal effectiveness and cost-effectiveness of The value of digital imaging in diabetic
screening for Down’s syndrome: tension-free vaginal tape for treatment retinopathy.
the results of the Serum, Urine and of urinary stress incontinence. By Sharp PF, Olson J, Strachan F,
Ultrasound Screening Study (SURUSS). Hipwell J, Ludbrook A, O’Donnell M,
By Cody J, Wyness L, Wallace S, et al.
By Wald NJ, Rodeck C, Hackshaw Glazener C, Kilonzo M, Stearns S, et al.
AK, Walters J, Chitty L, Mackinson AM.
No. 31
No. 22 Lowering blood pressure to prevent
No. 12
The effectiveness and cost-effectiveness The clinical and cost-effectiveness of myocardial infarction and stroke: a new
of ultrasound locating devices for patient education models for diabetes: preventive strategy.
central venous access: a systematic a systematic review and economic By Law M, Wald N, Morris J.
review and economic evaluation. evaluation.
By Calvert N, Hind D, McWilliams By Loveman E, Cave C, Green C, No. 32
RG, Thomas SM, Beverley C, Royle P, Dunn N, Waugh N. Clinical and cost-effectiveness of
Davidson A. capecitabine and tegafur with uracil for
No. 23 the treatment of metastatic colorectal
No. 13 The role of modelling in prioritising cancer: systematic review and economic
A systematic review of atypical and planning clinical trials. evaluation.
antipsychotics in schizophrenia. By Ward S, Kaltenthaler E, Cowan J,
By Chilcott J, Brennan A, Booth A,
By Bagnall A-M, Jones L, Lewis R, Brewer N.
Karnon J, Tappenden P.
Ginnelly L, Glanville J, Torgerson D,
et al. No. 33
No. 24 Clinical and cost-effectiveness of new
No. 14 Cost–benefit evaluation of routine and emerging technologies for early
Prostate Testing for Cancer and influenza immunisation in people localised prostate cancer: a systematic
Treatment (ProtecT) feasibility study. 65–74 years of age. review.
By Donovan J, Hamdy F, Neal D, By Allsup S, Gosney M, Haycox A, By Hummel S, Paisley S, Morgan A,
Peters T, Oliver S, Brindle L, et al. Regan M. Currie E, Brewer N. 341
357
Observers
Ms Kay Pattison, Dr Morven Roberts,
Section Head, NHS R&D Clinical Trials Manager,
Programme, Department of Medical Research Council
Health
359
Observers
Dr Tim Elliott, Dr Catherine Moody, Dr Ursula Wells,
Team Leader, Cancer Programme Manager, Principal Research Officer,
Screening, Department of Neuroscience and Mental Department of Health
Health Health Board
Observers
Ms Christine McGuire, Ms Kay Pattison Dr Caroline Stone,
Research & Development, Senior NIHR Programme Programme Manager, Medical
Department of Health Manager, Department of Research Council
Health
360
Current and past membership details of all HTA programme ‘committees’ are available from the HTA website (www.hta.ac.uk)
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
Observers
Dr Phillip Leech, Ms Kay Pattison Dr Morven Roberts, Dr Ursula Wells
Principal Medical Officer for Senior NIHR Programme Clinical Trials Manager, MRC, PRP, DH, London
Primary Care, Department of Manager, Department of London
Health , London Health
361
Pharmaceuticals Panel
Members
Chair, Dr Peter Elton, Dr Dyfrig Hughes, Dr Martin Shelly,
Professor Imti Choonara, Director of Public Health, Reader in Pharmacoeconomics General Practitioner, Leeds,
Professor in Child Health, Bury Primary Care Trust and Deputy Director, Centre and Associate Director, NHS
University of Nottingham for Economics and Policy in Clinical Governance Support
Professor Robin Ferner, Health, IMSCaR, Bangor Team, Leicester
Deputy Chair, Consultant Physician and University
Dr Lesley Wise, Director, West Midlands Centre Dr Gillian Shepherd,
Unit Manager, for Adverse Drug Reactions, Dr Yoon K Loke, Director, Health and Clinical
Pharmacoepidemiology City Hospital NHS Trust, Senior Lecturer in Clinical Excellence, Merck Serono Ltd
Research Unit, VRMM, Birmingham Pharmacology, University of
Medicines & Healthcare East Anglia Mrs Katrina Simister,
Products Regulatory Agency Dr Ben Goldacre, Assistant Director New
Research Fellow, Division of Professor Femi Oyebode, Medicines, National Prescribing
Mrs Nicola Carey, Psychological Medicine and Consultant Psychiatrist Centre, Liverpool
Senior Research Fellow, Psychiatry, King’s College and Head of Department,
School of Health and Social London University of Birmingham Mr David Symes,
Care, The University of Service User Representative
Reading Dr Bill Gutteridge, Dr Andrew Prentice,
Medical Adviser, London Senior Lecturer and Consultant
Mr John Chapman, Strategic Health Authority Obstetrician and Gynaecologist,
Service User Representative The Rosie Hospital, University
of Cambridge
Observers
Ms Kay Pattison Mr Simon Reeve, Dr Heike Weber, Dr Ursula Wells,
Senior NIHR Programme Head of Clinical and Cost- Programme Manager, Principal Research Officer,
Manager, Department of Effectiveness, Medicines, Medical Research Council Department of Health
Health Pharmacy and Industry Group,
Department of Health
Observers
Ms Kay Pattison Dr Morven Roberts, Professor Tom Walley, Dr Ursula Wells,
Senior NIHR Programme Clinical Trials Manager, MRC, HTA Programme Director, Policy Research Programme,
Manager, Department of London Liverpool DH, London
Health
362
Current and past membership details of all HTA programme ‘committees’ are available from the HTA website (www.hta.ac.uk)
DOI: 10.3310/hta14510 Health Technology Assessment 2010; Vol. 14: No. 51
363