Understanding adherence to hand

hygiene recommendations: The theory

of planned behavior
Carol A. O’Boyle, PhD, RNa
Susan J. Henly, PhD, RNb
Elaine Larson, PhD, RN, FAANc
Minneapolis, Minnesota, and New York, New York

Background: Most health care workers (HCWs) are aware of the rationale for hand hygiene procedures, yet failure to adhere to guidelines is com-
mon. Little is known about factors that motivate HCWs to practice hand hygiene.
Purpose: The purposes of this study were to (1) estimate adherence to hand hygiene recommendations; (2) describe relationships among motivation-
al factors, adherence, and intensity of nursing unit activity; and (3) test an explanatory model for adherence to hand hygiene guidelines based on the
theory of planned behavior (TPB).
Method: A longitudinal, observational design was used to collect data from 120 registered nurses employed in critical care and postcritical care
units. Nurses provided information about motivational factors and intentions and a self-report of the proportion of time they followed guidelines. At
least 2 weeks later, the nurses’ hand hygiene performance was observed while they provided patient care. Structural equation modeling was used to
test the TPB-based model.
Results: Rate of adherence to recommendations for 1248 hand hygiene indications was 70%. The correlation between self-reported and observed
adherence to handwashing recommendations was low (r = 0.21). TPB variables predicted intention to handwash, and intention was related to self-
reported hand hygiene. Intensity of activity in the nursing unit, rather than TPB variables, predicted observed adherence to hand hygiene recommen-
dations.
Conclusions: The limited association between self-reported and observed hand hygiene scores remains an enigma to be explained. Actual hand
hygiene behavior may be more sensitive to the intensity of work activity in the clinical setting than to internal motivational factors. (Am J Infect
Control 2001;29:352-60.)

Hand hygiene is accepted throughout the health
care community as a basic clinical procedure essential
for the prevention of infections in patients and health
care workers (HCWs) alike.1 However, studies consis-
tently show that HCWs practice hand hygiene incom-
pletely and infrequently, with rates of adherence to
From the Division of Disease Prevention and Control, Minnesota
Department of Healtha; the School of Nursing, University of
Minnesota, Minneapolisb; and the School of Nursing, Columbia
University, New York.c
Portions of this paper were presented at the 1998 annual meet-
ing of the Association for Professionals in Infection Control in
Baltimore, Md.
Supported in part by grants from the Georgetown University
School of Nursing, the 3M Enrich Program, and the Association
of Professionals in Infection Control Research Foundation.
Reprint requests: Carol A. O’Boyle, PhD, RN, Assistant Professor,
School of Nursing, 6-101 Weaver-Densford Hall, University of
Minnesota, 308 Harvard St, SE, Minneapolis, MN 55455-0342.
Copyright © 2001 by the Association for Professionals in
Infection Control and Epidemiology, Inc.
0196-6553/2001/$35.00 + 0 17/46/118405
doi:10.1067/mic.2001.118405
guidelines often lower than 70%.2 Although the hand
hygiene procedure itself is simple, HCW behavior relat-
ed to hand hygiene is a complex phenomenon that is
not easily understood, explained, or changed.
Knowledge about when, why, and how HCWs should
practice hand hygiene is widely available. According to
current guidelines from the Association for
Professionals in Infection Control, HCWs should prac-
tice hand hygiene when their hands are visibly soiled,
before and after patient contact, after contact with
body secretions or excretions or inanimate objects like-
ly contaminated, and after removing gloves.3
Appropriate technique includes use of soap and
water. The hand hygiene agent should be distributed to
all hand surfaces, and hands should be rubbed vigor-
ously for 10 to 15 seconds. An alcohol-based handrub
may be used instead of soap and water if the hands are
free of organic soiling.
Knowledge about hand hygiene, awareness of per-
sonal handwashing practices, types of hand hygiene
products, and accessibility of supplies have all been
recognized as factors that may influence HCW adher-
ence to hand hygiene recommendations. Education
and training, the most frequently implemented inter-
ventions designed to improve adherence, have had lim-

352
AJIC
Volume 29, Number 6
Fig 1. Initial TPB model.
ited long-term success.4-6 Interventions focused on the
organizational level have shown some evidence of
improving adherence.7 Yet “being too busy” is often
cited as an explanation for not practicing recommend-
ed hand hygiene, even though prevention of patient
infections is recognized as the most important reason
for adherence.8,9 Internal factors that motivate HCWs
to adhere to hand hygiene recommendations are large-
ly unstudied.
The theory of planned behavior (TPB)10-12 was pro-
posed to account for motivation to perform volitional
behaviors. The theoretical model pictured in Fig 1
shows how the TPB was used to develop an explanato-
ry model for adherence to hand hygiene recommenda-
tions. The TPB postulates that the immediate cause of
a planned behavior such as hand hygiene is intention
to perform the behavior. Intention is predicted directly
by 3 intermediate variables: attitude (feelings or affec-
tive regard for the behavior), subjective norm (a per-
son’s perception of the social pressure to perform or
not perform the behavior), and perceived behavioral
control (a person’s perception of the ease or difficulty
in performing the target behavior). These intermediate
variables are predicted by the strength of beliefs about
outcomes of the behavior, normative beliefs (the per-
son’s evaluation of specific other peoples’ expecta-
tions), and control beliefs (the person’s beliefs about
obstacles and resources), respectively.
The TPB provided a systematic framework for iden-
tification of antecedents to adherence to recommend-
ed hand hygiene practices. The purposes of this study
were to (1) estimate overall adherence to hand
hygiene recommendations; (2) describe relationships
among variables from the TPB, self-reported adher-
ence, observed adherence, and level of activity in the
nursing unit; and (3) test the TPB-based theoretical
model to explain self-reported and observed hand-
O’Boyle, Henly, and Larson 353
washing behavior of critical care and postcritical care
nurses.
METHODS
Setting
Intensive care and postintensive care units are set-
tings in which hand hygiene is especially important
because patients are vulnerable to nosocomial infec-
tions. Severity of their illnesses, presence of a high
number of invasive devices, and frequent physical con-
tact with caregivers are common risk factors. From
July 1996 through October 1997 data were collected on
hand hygiene practices and attitudes of nurses toward
hand hygiene. These nurses worked in adult
medical/surgical intensive care units (ICUs) and the
associated postintensive care units in 4 metropolitan
teaching hospitals in the Midwest. These units were
selected to include patient populations with compara-
ble nursing care requirements across hospitals. The
physical layout of the units consisted of rooms with 1
to 3 patient beds; each room contained at least 1 sink
with antimicrobial soap and paper towels. Sinks were
also available in halls or common work areas, such as
medication or utility rooms. During the study period
no alcohol-based solutions were supplied by the partic-
ipating hospitals.
Sample
Staff and charge nurses who provided direct patient
care, who were employed for at least 6 months, and
who worked at least 1 day per week on average were
eligible for participation. One hundred twenty nurses
(50 from critical care units and 70 from postcritical
care units) provided data for this analysis; of these, 20
(17%) from 1 hospital participated in a pilot phase of
the study that focused on refinement of the self-report

354 O’Boyle, Henly, and Larson
instrumentation and observational protocol. Of the
remaining 100 nurses, 40 (33%) were from the second
hospital, 40 (33%) were from the third hospital, and 20
(17%) were from the fourth hospital. Efforts were
made to recruit as many of the nurses assigned to the
study nursing units as possible. The 100 study-phase
nurses represented 21% of the total number of nurses
eligible to participate (100 of 474).
Procedures
The institutional review boards of each of the 4 par-
ticipating hospitals and the University of Minnesota
approved the protocol for this study. Nurses from crit-
ical and postcritical care units were recruited to partic-
ipate in the study by the principal investigator (C. O.)
at staff meetings, shift reports, and individual sessions.
Nurses who agreed to participate signed consent forms
and were provided with a copy of the Handwashing
Assessment Inventory (HAI) used for the study. The
HAI (described later) took approximately 10 minutes to
complete. Participants returned completed HAIs to the
principal investigator by mail or in person.
Approximately 2 weeks to 4 months after the comple-
tion of the HAI, all enrolled participants were observed
in the clinical setting for adherence to handwashing
recommendations with use of the Handwashing
Observation Instrument (HOI) protocol (described
later). Observation occurred in 1 continuous block of
time for 66% of the participants, whereas 34% were
observed on a noncontinuous basis due to unavoidable
changes in clinical assignment during the observation
period. Interrupted sessions were continued at a later
date until 120 minutes or 10 handwashing indications
were obtained. Observers recorded the number of indi-
cations for handwashing and the number of times
nurses washed their hands when the indications arose.
Information used to compute the nursing unit activity
score was obtained at the start of the observation peri-
od for each nurse.
Instruments
Adherence to handwashing recommendations.
Adherence to hand hygiene recommendations was
measured in 2 ways: observation and self-report. The
HOI was used to record actual hand hygiene indica-
tions and events for each nurse. Participants were
observed for 2 hours or until 10 indications for hand-
washing had occurred. Indications recorded with the
HOI were (1) before beginning care and/or resuming
care; (2) after completion of care; (3) before invasive
procedures; (4) moving from dirty to clean procedures;
(5) after removing gloves; (6) after contact with body
substances; and (7) before the nurse had contact with
his/her mouth, eyes, nose, and face (with contaminated
AJIC
December 2001
hands). Hand hygiene events were recorded when nurs-
es washed their hands with the handwashing agent and
water. Observed adherence scores, for specific indica-
tions and the overall total, were defined as the percent-
age of times a nurse practiced hand hygiene when indi-
cated. In a previous observational study, Larson et al13
reported that inter-rater agreement for HOI scores was
100%. Inter-rater reliability estimates for 2 sets of
raters were 0.94 and 0.98 during the pilot phase of this
study.2
Each nurse also provided a self-report of the per-
centage of time (from zero to 100%) that they practiced
hand hygiene when indicated (before care; when care
was interrupted; between patients; before performing
an invasive procedure; after contact with contaminated
material and before beginning a clean procedure on the
same patient; after removal of gloves; after direct con-
tact with body fluids; and before touching own mouth,
nose, eyes, and face with contaminated hands). The
score for each nurse was the average estimate across
the recommended indications. The internal consisten-
cy reliability estimate for the self-report of adherence
to handwashing recommendations was 0.87, suggest-
ing fairly consistent estimation of adherence across the
indications.9
Motivation for handwashing. The HAI9 was used to
measure the motivational schema for handwashing.
The HAI measured variables derived from the TPB for
use in explaining HCW hand hygiene behavior.
Table 1 lists definitions for each TPB variable, the
number of items in each scale, and the estimate of
internal consistency reliability (Cronbach alpha).
Scores for each of the TPB variables were computed by
summing the item responses and dividing by the num-
ber of items answered by each participant. Negatively
keyed items were reversed-scored before the score for
each HAI scale was computed. All scales were scored so
that high scores reflected a positive disposition toward
handwashing. As shown in Table 1, estimates of inter-
nal consistency reliability were adequate.
Intensity of activity. On the basis of the clinical expe-
rience of the investigators, intensity of nursing unit
activity was believed to be a factor in the participants’
opportunities to practice hand hygiene. Five indicators
were used to construct an index reflecting intensity of
activity in the nursing units. These indicators were type
of nursing unit (critical care vs postcritical care); time
of day (shift: day vs evening or night); amount of time
elapsed for 10 handwashing indications (> 1 hour vs ≤
1 hour); nursing unit census (> median vs ≤ median);
and patient-to-nurse ratio (> median vs ≤ median).
(Type of patient, time of day, and indications for hand-
washing have been reported in other studies to indicate
nursing unit activity.14) Points were assigned to critical
AJIC
Volume 29, Number 6 O’Boyle, Henly, and Larson 355

Table 1. Measures of TPB variables with use of the HAI

No. of Alpha
Variable Definition HAI items coefficient

Intention HCWs plan to adhere to hand hygiene recommendations in a 5 0.74

Perceived control
Control beliefs
Subjective norm
Normative beliefs
Attitude
Beliefs about outcomes
variety of clinical situations
Overall evaluation of the degree to which a HCW believes that 2 0.64
hand hygiene practices can be implemented as recommended
The extent to which a HCW believes that he or she possesses or 5 0.85
has access to resources required to adhere to hand hygiene
recommendations in a variety of patient care situations
Overall evaluation of the extent to which important people in the 1 n/a*
lives of HCWs are thought to support or endorse hand hygiene
as recommended
Beliefs about expectations that specific other people hold for the 7 0.89
HCWs’ personal hand hygiene behavior
Affective/cognitive evaluation of the hand hygiene procedure itself 8 0.83
Cognitive evaluation of the consequences of hand hygiene, 14 n/a*
including transmission of micro-organisms, control over the
environment, and professional behavior

HCW, Health care worker.

*Internal consistency reliability was not computed for these scales because the scale comprised only 1 item
(individual norm) or because items measured diverse outcomes (beliefs about outcomes).

Table 2. Observed hand hygiene adherence

(n = 120)

Handwashing scores* M (%)† SD No. of participants Total No. of indications

Before care 62 34 91 352

After completing care 87 22 108 431
Moving from dirty to clean procedures 60 52 10 13
After removing gloves 80 32 103 317
Before invasive procedures 57 53 7 9
After direct contact with body substances 87 30 22 42
Before touching (own) eyes, mouth, nose 3 17 44 79
Total 70 22 120 1246

*Handwashing scores (expressed as a percentage) were computed by dividing the number of hand hygiene events by the number of hand hygiene indi-
cations and multiplying the result by 100.
†Mean percent.

care unit, day shift, 1 hour or less for 10 handwashing
indications, nursing unit census greater than the medi-
an, and patient-to-nurse ratio more than the median.
Scores were obtained by adding the points across the 5
indicators. Potential scores could range from zero to 5,
with high scores earned for observations that occurred
in busy units.
Data analysis
The data set available for analysis was complete;
each participant’s TPB scores, the intensity of nursing
unit activity at the time of observation, and each par-
ticipant’s hand hygiene score were available for each
participant (ie, all observations were specifically linked
to individual participants). Thus the HAI scores of an
individual (reflecting strength of motivation to adhere
to hand hygiene recommendations) were directly
linked, through use of the HOI, to his or her observed
adherence to recommendations.
The overall rate of adherence to hand hygiene was
computed by dividing the total number of hand
hygiene events by the total number of indications and
multiplying by 100. Pearson correlation coefficients
were computed to summarize linear relationships
among the TPB variables, intensity of nursing unit
activity, and the 2 measures of adherence to hand-
washing recommendations (observed and self-reported).
Structural equation modeling (SEM) was used to esti-
mate the TPB-based model explaining motivation for
 AJIC
356 O’Boyle, Henly, and Larson December 2001

Table 3. Descriptive statistics and correlations for motivational factors, intensity of unit activity, and hand hygiene
adherence
1 2 3 4 5 6 7 8 9 10

1. Beliefs about outcomes 1.00

2. Attitude 0.43† 1.00
3. Normative beliefs 0.33† 0.26† 1.00
4.Subjective norm 0.34† 0.21* 0.55† 1.00
5. Control beliefs 0.46† 0.50† 0.31† 0.26† 1.00
6. Perceived behavioral control 0.46† 0.45† 0.25† 0.29† 0.76† 1.00
7. 5. Intention 0.38† 0.46† 0.37† 0.38† 0.72† 0.60† 1.00
8. Intensity of activity in the unit 0.11 -0.09 0.08 0.19* -0.15 -0.11 -0.07 1.00
9. Self-report hand hygiene adherence 0.27† 0.31† 0.10 0.19* 0.41† 0.31† 0.39† 0.02 1.00
10. Observed hand hygiene adherence -0.06 0.11 -0.13 -0.10 0.04 -0.04 0.09 -0.32† 0.22* 1.00

N 120 120 120 120 120 120 120 120 120 120
Scale range 1-7 1-7 1-7 1-7 1-7 1-7 1-7 1-5 0-100 0-100
Mean 5.18 5.69 6.30 6.28 5.53 6.08 5.93 2.38 81.70 69.97
Standard deviation 0.52 0.95 0.70 0.97 1.10 0.99 0.76 1.05 15.68 21.88

*P < .05.
†P < .01.

handwashing. A model modification approach was Adherence to hand hygiene recommendations

used to estimate, evaluate, and improve the initial
model. With use of this process, a tentative, theoreti-
cally defensible model is postulated and fit to the data.
If the initial model does not fit, it is modified and test-
ed again with the same data. The Lisrel7 program with
maximum likelihood estimation was used for the sta-
tistical analysis.16
SEM is a versatile methodologic tool that allows
straightforward translation of scientific theory to a
statistical model. SEM results in a quantitative snap-
shot of a dynamic relationship among variables. The
structure of underlying correlations among theoreti-
cally associated variables is formalized as a model or
models that reflect and summarize relationships
among the variables. By isolating the direct and indi-
rect effects of 1 variable on another, SEM places rela-
tionships among variables in focus so that important
relationships are identified and unimportant relation-
ships are absent. SEM gauges the impact of indepen-
dent x variables on the dependent y variables, the
impact of dependent variables on other dependent
variables, and the impact of random factors on depen-
dent variables by estimating path coefficients and
unexplained variances.17
RESULTS
Hand hygiene observation was completed for all
enrolled participants (70 critical care and 50 post-
critical care nurses). Nurses were observed on all
shifts: 69 (57%) during the day, 26 (22%) during the
evening, and 25 (21%) at night. Observation time for
the entire study totaled approximately 215 hours;
most observation periods (n = 80; 66%) lasted 61 to
120 minutes.
Table 2 lists indications for hand hygiene and observed
hand hygiene events. The most frequently occurring
indications for hand hygiene were after completion of
care and before beginning care, followed closely by
after removing gloves; these 3 indications accounted
for 1100 of the 1246 observed indications (88%).
Adherence was highest for hand hygiene after comple-
tion of care (87.08%) and after direct contact with body
substances (87.12%).
Nurses were least likely to wash their hands before
touching their own face, hair, eyes, teeth, nose, or
mouth with contaminated hands (3.4% of the time, n =
79 indications). Of the 120 nurses, 23 were observed to
touch their face, eyes, mouth, or teeth with contami-
nated hands during the observation period. A small
number of nurses were observed to touch their facial
areas with contaminated hands more than 5 times dur-
ing the observation period. Face touching was more
likely to occur during problem-solving discussions and
grooming behavior. For some participants, face touch-
ing seemed to be a repeated (habitual) gesture.
Across the 1246 indications for hand hygiene, the
mean observed adherence was 70%, with a range of
61% to 74%; there was no significant difference in the
observed adherence rate among the 4 hospitals. In con-
trast, the average self-reported rate of adherence was
82%, with a range of 71% to 89%; the differences in the
self-reported rates among the hospitals were signifi-
cant (P < .001).
Motivation, adherence, and activity levels
Correlations among the TPB motivational factors
were positive and moderately high (Table 3). The TPB
AJIC
Volume 29, Number 6
Fig 2. Model generation history.
motivational factors were associated with intention
and self-reported hand hygiene adherence but not with
observed hand hygiene adherence. The association
between self-reported and observed hand hygiene
adherence was positive but low (r = 0.22). Intensity of
activity in the nursing unit at the time of the observa-
tion was significantly and negatively associated with
observed adherence to hand hygiene recommendations
(r = -0.32).
Explanatory model for hand hygiene adherence
The TPB model pictured in Fig 1 was estimated and
subsequently modified as shown by the dashed arrows
in Fig 2 to improve fit. Paths were added from control
beliefs to intention and from control beliefs to attitude.
With these modifications, the Goodness-of-Fit index,
adjusted Goodness-of-Fit index, and Root Mean
Squared residual all suggested a close fit of the model
to the data. The final model is pictured in Fig 3, with
parameter estimates noted and estimates significant at
P < .05 (indicated by asterisks in the illustration).
The model shows that many relationships predicted
by the TPB were supported. Motivational factors pre-
dicted intention, and intention was related to self-
report of adherence to hand hygiene recommenda-
tions, confirming the importance of intention as a
mediating variable. However, the control beliefs vari-
able displayed a greater impact than suggested by the
theoretical model; empirical adjustment showed that
control beliefs had both an indirect effect (through atti-
tude) and a direct effect on intention to practice hand
hygiene. The TBP variables, including intention to
practice hand hygiene, did not predict actual, observed
hand hygiene. Instead, intensity of activity in the nurs-
ing unit at the time of observation was significantly
O’Boyle, Henly, and Larson 357
negatively associated with observed adherence to hand
hygiene recommendations (that is, adherence was
lower when the nursing unit was busier).
DISCUSSION
Hand hygiene is acknowledged to be a fundamental
practice in providing safe health care to patients; how-
ever, the motivational issues involved in adherence to
guidelines have not been elucidated. Progress in
explaining and understanding variation in adherence
to hand hygiene recommendations has been impeded
by lack of a theoretical perspective. Intervention stud-
ies with a cognitive focus designed primarily to
increase HCW knowledge about handwashing have
generally been unsuccessful at improving adherence.
Recent attempts to improve hand hygiene behavior
have met with success when administrative, organiza-
tional, and environmental factors have been incorpo-
rated into the work setting.7,18 These multifaceted stud-
ies are congruent with the theoretical approach that
served as a framework for this work. The TPB was
selected as the guiding framework for the study
because it has been used successfully to understand
other health-related behaviors.19,20
In this study, correlations among the motivational
variables of the TPB, intention to practice hand
hygiene, and self-reported adherence suggested a well-
structured motivational schema; however, the TPB
variables did not predict actual adherence. Motivations
were largely unrelated to observed hand hygiene prac-
tices. Observed handwashing behavior was predicted
only with the addition of a variable that reflected and
quantified an aspect of the situational context: activity
of the nursing unit. The TPB was developed to under-
stand human volitional behavior, and it includes the

358 O’Boyle, Henly, and Larson
Fig 3. Structural equation modeling: TPB variables and adherence to hand hygiene. Path coefficients
(n = 120).
concept that the necessary opportunities to perform
the behavior of interest must be present. Performance
of a behavior depends not only on the internal motiva-
tional factors but also on the available resources.
Perceived control can predict behavior only to the
extent to which it reflects actual control. Actual behav-
ioral control refers to the extent to which the resources
are available for the person to perform the desired
behavior. Thus the variable used to reflect the intensity
of nursing unit activity may not be “actual behavioral
control” but may represent factors that influence actu-
al behavioral control.10,11,21
Certain limitations dictate caution in the interpreta-
tion of these findings. The study used a convenience
sample; all study participants were volunteers who
were informed that the study involved having their
nursing practice observed. This process allowed for
self-selection by those nurses who had a sense of comfort
and possibly confidence about their nursing practice.
Also volunteers may have tended to be more conscien-
tious about their work, including greater adherence to
hand hygiene recommendations.
The sample size (n = 120) was relatively small for a
SEM problem.22 Thus the model fit to these data may
be a simpler representation of the motivational schema
for practicing hand hygiene than actually exists.23
Nevertheless, the theory-based statistical model pro-
vides a useful summary of a previously unstudied prob-
lem.
Prospective study protocols that first measure inten-
tion and later observe behavior may unwittingly influ-
ence the behavior of interest. Studies that include
methods to correlate intentions (measured by a survey
instrument) with actual behavior risk influencing the
AJIC
December 2001
behavior of interest (such as hand hygiene) by admin-
istering the survey before observation of the behavior.
Conversely, if the behavior is observed first, the risk is
that the survey may be influenced by the respondents’
need for consistency in beliefs and behavior. The time
period between the administration of the survey, HAI,
and observed behavior ranged from 2 weeks to 4
months. The longer periods may have reduced the
influence of the survey on actual hand hygiene behav-
ior. However, other factors may have intervened and
produced changes in motivational factors and perfor-
mance.
The only health care providers represented in the
sample were nurses. Three of the 4 study hospitals
were undergoing management reorganizations and/or
mergers when the protocol was implemented; it was
not possible to control any potentially confounding
effects associated with concurrent organizational
change.
The finding that high levels of nursing activity or
“understaffing” negatively influence hand hygiene was
also reported by Pittet and colleagues.14 The singular
importance of the intensity of nursing unit activity in
the prediction of observed adherence to hand hygiene
guidelines suggests recommendations for practice and
directions for future study. The following recommen-
dations span theoretical and applied work.
First, a theoretical perspective on HCW hand
hygiene behavior must be expanded from a focus on
the individual to a focus on the individual in situation-
al context. Perceived control over a behavior that
reflects both past experiences and environmental limi-
tations has been found to correlate strongly with actu-
al performance of a behavior.10 The critical role that
AJIC
Volume 29, Number 6
intensity of nursing unit activity played in the modified
TPB model underscores the need for such theory. The
success of the TPB at explaining intentions might be
extended if a “situational” theory were proposed as a
context for individual behavior. For example, perceived
lack of time to adhere to hand hygiene recommenda-
tions is a frequently expressed concern of HCWs.8,24
Inadequate time to adhere to infection control recom-
mendations may reflect HCW awareness of the hierarchy
of patient needs and their perception of the realities of
clinical practice. In this study, intention to practice
hand hygiene was lowest for patient crisis situations.
Insufficient time and competing patient needs may
also have been factors in the low score for the HAI item
related to emergency patient situations. In view of the
limited time to perform hand hygiene, as reported by
the nurses in this investigation and as suggested by
Voss and Widmer, HCWs could appropriately replace
some of the soap-and-water hand hygiene episodes
with a waterless hand hygiene process.25
Second, refined conceptualization and measurement
of environmental factors are needed. The challenge will
be to identify and quantify the influence of environ-
mental conditions both from within and outside the
immediate setting since the interplay between persons
and environmental forces may profoundly influence
behavior.26
In this study, an index that reflected level of activity
in patient care units was constructed. Indicators were
similar to those used in a study of adherence to infec-
tion control recommendations in a Swiss hospital,14
and the index performed well in prediction of observed
adherence. Future research identifying and quantifying
the situational or contextual variables influencing clin-
ical practice could establish critical thresholds at
which recommended practices by HCWs are less likely
to occur.
Third, many studies of hand hygiene adherence rely
on self-report measurement techniques. The correla-
tion between observed adherence and self-report in this
study was 0.22. Findings here and elsewhere27-29
emphasize the inadequacy of the self-report approach.
Unless future work identifies conditions under which
self-report is an accurate representation of actual hand
hygiene behavior, the practice should be abandoned.
Hand hygiene adherence should be measured by actu-
al observation, despite the incremental cost of the
observation procedure.
Fourth, previous studies have documented an asso-
ciation between nurse staffing levels and patient out-
comes such as respiratory and urinary tract infec-
tions,30 staphylococcal infections in a neonatal unit,
and bloodstream infections from central venous
catheters.31,32 In this study, increased activity levels and
O’Boyle, Henly, and Larson 359
lower adherence to hand hygiene recommendations
were linked. In combination, the findings suggest that
adherence may be an important variable that mediates
the relationship between staffing levels and patient out-
comes. Studies designed to elucidate this postulated
mechanism would be helpful.
Fifth, interventions to improve adherence to hand
hygiene and other recommended infection control prac-
tices traditionally focus on the educational or cognitive
dimension. The underlying principle seems to be that if
the correct information is possessed by the HCW, the cor-
rect behavior will follow. Descriptive studies to identify
reasons for lack of adherence to recommended practices
typically focus on the level of information HCWs pos-
sesses about the recommended practice. Recently, orga-
nizationally focused interventions showed promise.7,14
Thus a multidimensional, theoretically grounded
approach that includes attention to motivational and
situational factors might prove more effective. For
example, education about the benefits of handwashing
hygiene could be augmented by establishment of feed-
back systems that articulate adherence as the norm
and establish or reinforce the cultural context of work
safety practices.33
The frequent face touching by some of the partici-
pants raises several interesting questions about the rela-
tionship between contaminated hands, the health of
HCWs, and the risk of transmission of micro-organisms
and subsequent colonization of HCWs with hospital
flora. Are HCWs aware of their facial touching? Is there
a greater risk for these frequent face touchers to be
nasal carriers of hospital micro-organisms? What
interventions would reduce the frequency of face
touching with contaminated hands?
Another focus of potential contamination of HCWs
was the inanimate environment. Nurses were observed
to touch numerous objects in the environment with
gloves contaminated with patient secretions or excre-
tions. These objects in the patient environment were
then subsequently handled by other HCWs (without
gloves) who may have had little appreciation for the
potential contamination. In this study, some HCWs
were observed to move from dirty to clean procedures
on patients without changing contaminated gloves or
move to the common desk area and touch items with
contaminated gloves. The impact of glove use on hand-
washing practices needs to be carefully evaluated.
In this study, intensity of activity rather than internal
motivational factors influenced adherence to hand
hygiene recommendations. Future research that is
focused on the relationship between work intensity and
the ability of HCWs to function at safe levels and
accomplish appropriate patient care activities is war-
ranted. Additional studies should also focus on eco-

360 O’Boyle, Henly, and Larson
nomic issues to determine whether it is cost-effective to
staff patient care units at levels at which recommended
practices are less likely to occur. If future research con-
firms the finding of this investigation and those report-
ed by Pittet et al14 regarding the relationship between
intensity of nursing activity and adherence to hand
hygiene recommendations, consideration must be
given to ethical issues implicit in the relationship
between health care organizations and patients, in
which safe care is an expected component.
We wish to thank Laura Duckett, PhD, RN, for her invaluable support
and guidance. We also thank Sharon Ridgeway, PhD, RN, and Nancy
Van Drunen, BS, RN, for assistance with data collection.
References
1. Larson E. A causal link between handwashing and risk of infec-
tion? Examination of the evidence. Infect Control Hosp
Epidemiol 1988;9:28-36.
2. O’Boyle C. Variables that influence health care workers’ adher-
ence to recommended handwashing practices [dissertation].
Minneapolis, Minn: University of Minnesota, Minneapolis;
1998:165.
3. Larson EL, the 1992 and 1994 APIC Guidelines Committee. APIC
guideline for handwashing and hand antisepsis in health care set-
tings. Am J Infect Control 1995;23:251-69.
4. Conly JM, Hill S, Ross J, Lertzman J, Louie TJ. Handwashing
practices in an intensive care unit: the effects of an educational
program and its relationship to infection rates. Am J Infect
Control 1989;17:330-9.
5. Dubbert PM, Dolce J, Richter M, Miller M, Chapman SW.
Increasing ICU staff handwashing: effects of education and group
feedback. Infect Control Hosp Epidemiol 1990;11:191-3.
6. Mayer JA, Dubbert PM, Miller M, Burkett PA, Chapman SW.
Increasing handwashing in an intensive care unit. Infection
Control 1986;7:259-62.
7. Larson EL, Early E, Cloonan P, Sugrue S, Parides M. An organi-
zational climate intervention associated with increased hand-
washing and decreased nosocomial infections. Behav Med
2000;26:14-22.
8. Larson E, Killien M. Factors influencing handwashing behavior
of patient care personnel. Am J Infect Control 1982;10:93-9.
9. O’Boyle C, Henly SJ, Duckett LJ. Nurses’ motivation to wash their
hands: a standardized measurement approach. Appl Nurs Res
2001;14:136-45.
10. Ajzen I. Attitudes, personality and behavior. Chicago, IL: Dorsey;
1988.
11. Ajzen I. A theory of planned behavior. In: Kuhl J, Beckmann J,
editors. Action-control: from cognition to behavior. Heidelberg:
Springer; 1985.
12. Ajzen I, Fishbein M. Understanding attitudes and predicting
social behavior. Englewood-Cliffs, NJ: Prentice Hall; 1980.
13. Larson EL, Bryan JD, Adler LM, Blane C. A multifaceted
approach to changing handwashing behavior. Am J Infect Control
1997;25:3-10.
AJIC
December 2001
14. Pittet D, Mourouga P, Perneger TV, Compliance with handwash-
ing in a teaching hospital. Ann Intern Med 1999;130:126-30.
15. Joreskog KG. Testing structural equation models. In: Bollen KA,
Long JS, editors. Testing structural equation models. Newbury
Park, CA: Sage; 1993. p. 294-316.
16. Joreskog KG, Sorbom D. Lisrel7. Moorsville, IN: Scientific
Software Inc; 1989.
17. Bollen KA. Structural equations with latent variables. New York:
Wiley; 1989.
18. Pittet D, Hugonnet S, Harbarth S, et al. Effectiveness of a hospi-
tal-wide programme to improve compliance with hand hygiene.
Lancet 2000;356:1307-12.
19. Chan DK, Fishbein M. Determinants of college womens’ inten-
tions to tell their partners to use condoms. J Appl Psychol
1993;23:1455-70.
20. Duckett LJ, Henly SJ, Avery M, et al. A theory of planned behav-
ior-based structural model for breastfeeding. Nurs Res
1998;47:325-36.
21. Ajzen I. Nature and operation of attitudes. Annu Rev Psychol
2001;52:27-58.
22. Tanaka JS. How big is big enough: sample size and goodness of
fit in structural equation models with latent variables. Child Dev
1987;58:134-46.
23. Cudeck R, Henly SJ. Model selection in covariance structures,
analysis and the “problem” of sample size: a clarification. Psychol
Bull 1991;109:512-19.
24. Williams COB, Henry K, Campbell S, Collier P. Variables influ-
encing worker compliance with Universal Precautions in the
emergency department. Am J Infect Control 1994;22:138-48.
25. Voss A, Widmer AF. No time for handwashing!? Handwashing
versus alcoholic rub: can we afford 100% compliance. Infect
Control Hosp Epidemiol 1997;18:205-8.
26. Bronfenbrenner U. The ecology of human development experi-
ments by nature and design. Cambridge, Mass: Harvard
University Press; 1979. p. 39.
27. Larson E, McGinley KJ, Grove GL, Leyden JJ, Talbot GH.
Physiologic, microbiologic, and seasonal effects of handwashing
on the skin of health care personnel. Am J Infect Control 1986;
14:51-9.
28. Ronk LL, Girard NJ. Risk perception, Universal Precautions com-
pliance. Assoc Oper Room Nurses 1994;59:253-66.
29. Alvaran MA, Butz A, Larson E. Opinions, knowledge, and self-
reported practices related to infection control among nursing
personnel in long-term care settings. Am J Infect Control
1994;22:367-70.
30. Blegen MA, Goode CJ, Reed L. Nurse staffing and patient out-
comes. Nurs Res 1998;47:43-50.
31. Haley RW, Bregman DA. The role of understaffing and over-
crowding in recurrent outbreaks of staphylococcal infection in a
neonatal special-care unit. Infect Dis 1982;145:875-85.
32. Robert J, Fridkin S, Blumberg H, White N, Ray S, Chan J, Jarvis
W. The influence of the composition of the nursing staff on pri-
mary bloodstream infection rates in a surgical intensive care unit.
Infect Control Hosp Epidemiol 2000;21:12-7.
33. Saari J. On strategies and methods in company safety work: from
informational to motivational strategies. Occup Accidents
1990;12:107-77.