DR QIAO QIN WAN (Orcid ID : 0000-0001-5058-8437)

Article type : Original Research: Empirical research - quantitative

Accepted Article
Effects of work environment and job characteristics on the turnover
intention of experienced nurses: The mediating role of work engagement

1.Running head

Factors affecting nurses’ turnover intentions

2.Authors' names, professional and academic qualifications, positions and places of work.

Qiaoqin Wan, RN，Associate Professor，PhD，School of Nursing, Peking University, Beijing, China

Zhaoyang Li，RN，Master student，School of Nursing, Peking University, Beijing, China

Weijiao Zhou，RN，Master student，School of Nursing, Peking University, Beijing, China

Shaomei Shang，RN， Professor，MD，School of Nursing, Peking University, Beijing, China

3.Corresponding author address, and telephone and fax numbers and email address

Shaomei Shang, Email: mei916@263.net

This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/jan.13528
This article is protected by copyright. All rights reserved.
 4. Acknowledge
Accepted Article
The researchers would like to thank all RNs who participated in this study and all nurse leaders in

sampled hospitals who helped to coordinate the survey. And this project was funded by Hospital

Authority of Beijing, China.

5. Conflict of Interest statement

No.

6. Funding Statement

No.

ABSTRACT

Aim. To assess turnover intention among experienced nurses and explore the effects of work

environment, job characteristics and work engagement on turnover intention.

Background. The nursing shortage is an urgent concern in China. A high turnover rate of experienced

nurses could have serious effects on the quality of care, costs and the efficiency of hospitals. It is crucial

to explore the predictors of turnover intention and develop strategies tailored to experienced nurses.

Design. A descriptive, cross-sectional survey design.

Methods. A total of 778 experienced nurses from 7 hospitals was surveyed on their work engagement,

job characteristics, work environment and turnover intention in March-May 2017. Structural equation

modelling was used to test a theoretical model and the hypotheses.

This article is protected by copyright. All rights reserved.

 Results. The results showed 35.9% of experienced nurses had high-level turnover intention. The final

model explained 50% of the variance in experienced nurses’ turnover intention and demonstrated that:
Accepted Article
(1) work environment was positively associated with higher work engagement and lower turnover

intention and work engagement partially mediated the relationship between work environment and

turnover intention; and (2) job characteristics were positively related to higher work engagement and

lower turnover intention and work engagement fully mediated the relationship between job

characteristics and turnover intention.

Conclusions. The study confirms the intrinsic and extrinsic motivators on work engagement posited by

job demands-resources model. Theory-driven strategies to improve work environment, enhance job

characteristics and promote wok engagement are needed to address the nursing shortage and high

turnover intention among experienced nurses.

Key words: work engagement, job characteristics, work environment, turnover intention, nurse,

structural equation modelling

SUMMARY STATEMENT

Why is this research needed?

 Experienced nurses are valuable hospital resources, but few retention strategies are

aimed at them.

 The job demands-resources model proposes mechanisms on how work environment, job

characteristics and work engagement affect the turnover intention of nurses.

What are the key findings?

 The turnover intention of experienced nurses in China was found to be high.

This article is protected by copyright. All rights reserved.

  Work engagement fully mediated the effect of job characteristics on turnover intention

and partially mediated the effect of work environment on turnover intention.

Accepted Article
 Job characteristics and work environment were positively correlated, which was not an

element of the job demands-resources model that we tested.

How should the findings be used to influence research and practice?

 The findings confirm the intrinsic and extrinsic motivators on work engagement posited

by the job demands-resources model in nursing populations.

 Future studies should examine the dual mechanisms of work environment on turnover

intention through work engagement and burnout and the possible relationships among

job resources.

 The results provide support for theory-driven interventions to address the nursing

shortage and high turnover intention among experienced nurses.

INTRODUCTION

The nursing shortage has become an urgent concern in Mainland China in recent years. The

number of registered nurses in China was 3.24 million in 2015, representing a nurse-to-patient ratio of

2.36 nurses per 1,000 population (Ministry of Health of China, 2015). This ratio is much lower than the

9.8 ratio in the USA and the average 4.45 ratio in middle-income countries (World Health Organization

[WHO], 2011). Furthermore, the situation is exacerbated by the high turnover rate of nurses and the

surging demand for nursing services due to the growth and ageing of the population in China. Nurse

turnover jeopardizes health-care delivery in several ways. First, nurse turnover can negatively affect the

morale and increase the workload of the remaining nurses, which may lead to further turnover (Gauci

Borda & Norman, 1997). In addition, high nurse turnover can have a negative effect on patient

outcomes (Hayes et al., 2012; Aiken et al., 2012). Finally, nurse turnover can increase costs because of

reduced productivity and the need to recruit new nurses. The total cost of hiring a newly recruited nurse

This article is protected by copyright. All rights reserved.

 has been reported to be $15,825, on average and the cost of reduced productivity is estimated to range

from $6,245 to $15,102 (Waldman et al., 2004; Duffield et al., 2014).

Accepted Article
Several recent studies report that the inclination of nurses to leave their jobs or the nursing

profession might be higher among older nurses than younger nurses (Leone et al., 2015; Sawatzky &

Enns, 2012). Older nurses possess extensive clinical skills and practical experiences and they are a

valuable source of knowledge as well as role models for younger nurses. The cost of replacing an

experienced nurse is also much higher than replacing a new nurse. It is obvious that high turnover

among experienced nurses can have serious effects on the quality of care, the costs and the efficiency of

hospitals. Accordingly, tailored strategies should be developed to tackle the problem of high turnover

intention among older and experienced nursing staff. How to reduce nurse turnover and tackle the

problem of nursing shortage has captured wide attention across the world. However, previous research

and retention strategies mainly focus on the general population or newly graduated nurses, whereas it is

essential to attend to the needs of experienced nurses (Bishop, 2013). Hence, this study aimed to

describe turnover intention among experienced nurses and explore the effects of work environment, job

characteristics and work engagement on turnover intention.

Background

The job demands-resources (JD-R) model describes two underlying different psychological

processes that occur in workplace and have critical effect on employees’ wellbeing. The first is a

motivational process, which results from access to adequate job resources and leads to work

engagement and positive job outcomes, such as low turnover and high performance. The second is a

health impairment process, which follows from excessive job demands and results in burnout and health

problems (Schaufeli & Bakker, 2004). This study mainly focused on the motivational process and used

the JD-R model as lens to link the relationships among work environment, job characteristics, work

engagement and turnover intention and proposes the theoretical model shown in Figure 1.

This article is protected by copyright. All rights reserved.

 Work engagement and turnover intention

This study considers nurse engagement as work engagement, which is defined as a positive
Accepted Article
work-related state of mind that is characterized by feelings of vigour, dedication and absorption. Vigour

is defined as a high level of energy and mental resilience while working, along with a willingness to

invest efforts in work and persistence when facing with difficulties. Dedication refers to being deeply

involved in one’s work with a sense of significance, enthusiasm, inspiration and challenge. Absorption

is characterized by a thorough interest and immersion in work, so that the sense of time is lost

(Schaufeli & Bakker, 2004). According to the JD-R model, engaged employees have high levels of

energy and are often enthusiastic about their work as well as fully immersed in their job, which

subsequently enhance positive person-level performance and organization-level performance (Bakker

& Demerouti, 2007). A growing body of research in various disciplines has further demonstrated the

association of work engagement with personal and organizational outcomes, including job satisfaction

(Giallonardo et al., 2010), turnover intention (Sawatzky & Enns, 2012; Walker & Campbell, 2013) and

work effectiveness (Laschinger et al., 2009). We, therefore, proposed Hypothesis 1 (see Table 1).

Work environment, job characteristics and work engagement

The JD-R model posits that work engagement is fostered by job resources because they play an

intrinsic motivational role by stimulating employees’ growth, learning and development, as well as

playing an extrinsic motivational role because they are instrumental in achieving work goals (Bakker &

Demerouti, 2007). The intrinsic motivational process is triggered by job characteristics. Every job has a

specific motivational potential that depends on the presence of five core job characteristics, namely,

skill variety, task identity, task significance, autonomy and feedback. The five job characteristics

enhance job attitudes by shaping three critical psychological states: an experience of job

meaningfulness, a sense of responsibility for work outcomes and knowledge of work results (Oldham &

Hackman, 1981). The extrinsic motivational process highlights the importance of a supportive work

environment that offers various resources to foster employees’ willingness to dedicate their efforts and

abilities to job tasks. Aiken et al. (2002) proposed five potentially modifiable environmental features

that provide supports to enhance nurses’ work attitudes and performance, including adequate staffing

This article is protected by copyright. All rights reserved.

 resources, nurse manager ability and leadership, nurse-physician relations, nurse participation in

hospital affairs and a nursing foundation to ensure quality of care. Studies have found the
Accepted Article
five-dimension work environment is associated with nurse engagement (Sullivan Havens et al., 2012;

Van Bogaert et al., 2013). For example, Sullivan Havens et al. (2012), who studied 747 nurses, found

that a supportive work environment was a significant predictor of nurse engagement. Despite the lack of

research confirming the effect of all five of the job characteristics on nurse engagement, Giallonardo et

al. (2010) found that autonomy had a positive impact on nurse engagement, which provides some

evidence for the intrinsic motivation process. Based on theoretical analyses and empirical supports, we

proposed Hypothesis 2 and Hypothesis 3 (see Table 1).

The meditating role of work engagement on the association of turnover intention with work

environment and job characteristics

There is an extensive body of hospital-based research showing that the work environment is linked

to turnover intention among nurses (Leone et al., 2015; Sawatzky & Enns, 2012). Although the

relationship of work environment and turnover intention has been widely investigated, the mechanism

underlying this association has been explored far less. To date, the five job characteristics and their

outcomes have seldom been tested in nursing research. The job characteristics model posits that the five

job characteristics motivate employees to exhibit positive job attitudes (i.e. job satisfaction and work

engagement) and job outcomes (i.e. high performance and low turnover); however, it is still unclear

whether job characteristics affect turnover intention directly or whether they are mediated by other

variables. The JD-R model provides some support for the mechanisms connecting work environment,

job characteristics and turnover intention. According to the JD-R model, work engagement plays a key

role in the association between job resources and positive outcomes. As job resources, a supportive

work environment and job characteristics may enhance an employee’s work engagement due to their

intrinsically and extrinsically motivating qualities and in turn, facilitate positive job outcomes, such as

low turnover intention. This means job characteristics and work environment may influence turnover

intention via the mediation of work engagement. Previous studies showed that work environment and

job characteristics could predict work engagement (Sullivan Havens et al., 2012; Laschinger et al.,

This article is protected by copyright. All rights reserved.

 2009), which reflected the motivational process. In addition, the study by Sawatzky and Enns (2012)

found an effect of work engagement on turnover intention. Furthermore, a recent study by Santos et al.
Accepted Article
(2015) demonstrated that work engagement served as a mechanism linking job characteristics and

affective organizational commitment. Hence, we expected that work engagement would mediate the

effects of work environment and job characteristics on turnover intention. Thus, we proposed a partial

mediation model (Figure 1) and Hypothesis 4 - 7 (Table 1).

THE STUDY

Aims

This study aimed to describe turnover intention among experienced nurses and explore the effects

of work environment, job characteristics and work engagement on turnover intention.

Research hypotheses

We formulated seven hypotheses (able 1) with the guidance of the proposed theoretical model.

Design

The study was conducted in Beijing using a descriptive, cross-sectional survey design.

Participants

Participants were sampled through a two-stage sampling method. First, purposeful sampling was

used to select 7 hospitals, which were all tertiary public hospitals with more than 500 beds. Second,

random cluster sampling was used to select study participants. According to the total number of nurses

in the hospitals, 10 to 15 units were randomly sampled based on a list of all the units. To be eligible for

participation, the nurse had to: (1) have a registered nurse license; (2) be employed for at least 5 years;

and (3) be currently working as full-time staff.

Although there is no commonly agree on methods to calculate the sample size for testing structural

equation models (SEM), Klien (2005) suggests that a sample size of 200 or more is required to yield

This article is protected by copyright. All rights reserved.

 stable estimates. Fritz and MacKinnon (2007) proposed a sample size of 560 to estimate mediation

effects using bootstrapping. Hence, the required sample size was proposed to be 600 in this study.
Accepted Article
Data collection

After receiving institutional review board (IRB) approval, the investigators contacted the nursing

leaders of the sampled hospitals to explain the study’s procedures and determine the date of the survey.

Then, a written paper notice was delivered to the sampled units, which explained the aim of the study

and the principles of voluntary and anonymous participation. The study staff (trained graduate nursing

students) went to the units to distribute the survey packets to participants who consented to be in the

study. Considering an estimated non-response rate of 30%, 900 questionnaires were distributed, of

which 812 were retuned and 778 were valid, yielding a valid response rate of 86.4%.

Instruments

Farh’s turnover intention scale (Farh, 1998) was used to measure participants’ turnover intention.

Participants rated their agreement with items, such as ‘I often consider resigning from my current job’,

on a 5-point Likert scale. Items were averaged to compute a composite score. Nurses with scores less

than 3 were categorized as having low turnover intention, while those scored greater than 3 were

categorized as having high turnover intention. The Cronbach’s alpha of the scale was 0.81 (Tu et al.,

2013).

A Chinese version of the Utrecht Work Engagement Scale (C-UWES) was used to measure work

engagement, which includes three subscales: vigour, dedication and absorption (Zhang & Gan, 2005).

UWES contains 17 items with each item rated on a 7-point scale and higher scores representing greater

engagement. The scale was translated and back translated from English to Chinese by the professional

translators and bilingual researchers. The original UWES, which had been widely used, showed high

convergent and divergent validity and reliability (Bakker & Demerouti, 2007). The internal consistency

of C-UWES was consistently reported to be 0.70 or above (Zhang & Gan, 2005).

A Chinese version of the Job Diagnostic Survey (C-JDS) was used to measure nurses’ perception

on their job characteristics, which included five dimensions: skill variety, task identity, task

This article is protected by copyright. All rights reserved.

 significance, autonomy and job feedback (Dail & Yang, 2004). JDS was composed of 15 items (3 items

per subscale) that were measured on a 5-point Likert scale with higher scores indicating higher
Accepted Article
motivational job characteristics. A double translation procedure was used to translate the original JDS

into the Chinese language. The original scale possessed acceptable levels of validity and reliability

(Dail & Yang, 2004) and the intern reliability of C-JDS was reported to be high with Cronbach’s alpha

= 0.94 (Meng, 2011).

A Chinese version of the Practice Environment Scale (C-PES) of the Nursing Work Index (Wang

& Li 2011) was used to assess nurses’ perception on their work environment. PES was composed of

five subscales (31 items): staffing-resource adequacy (4 items), nurse manager ability and leadership (5

items), nurse-physician relations (3 items), nurse participation in hospital affairs (9 items) and nursing

foundation for quality of care (10 items). Each item was rated on a 4-point Likert scale and higher

scores indicated the presence of a more supportive practice environment. With the permission from the

original author, the scale was translated into Chinese by one professional translator and five Chinese

nursing experts. A back translation was conducted to ensure equivalency by two bilingual Chinese

nurses. As the most common scale to evaluate nurses’ work environment, the reliability and validity of

the original PES had been established in various settings (Lake, 2007). C-PES also showed high internal

reliability with Cronbach’s alpha = 0.91 and test-retest correlation coefficient = 0.84 (Wang & Li,

2011).

Ethical considerations

Ethical approval was obtained from IRB of Peking University (No. IRB0001052-17027). The

participants were informed that involvement was completely voluntary and anonymous.

Data analysis

The SPSS 22.0 software package was used for descriptive analysis and correlation analysis and

AMOS 22.0 software package was used for SEM in this study. SEM is a set of statistical techniques that

combines regression analysis, path analysis and factor analysis, which simultaneously tests the

relationships between indicators and latent variables and among the latent variables themselves. It

This article is protected by copyright. All rights reserved.

 offers more reliable and more flexible estimations than traditional multivariate techniques.

Goodness-of-fit indices are used to evaluate the quality of model fit through comparing differences
Accepted Article
between the sample data set covariance matrix and the structural equation model’s theoretical

covariance matric structures (Wu, 2010; Iacobucci, 2010).

In this study, the hypotheses were analyzed through SEM procedures, with: (a) the dimension scores

for work engagement, job characteristics and work environment and the item scores for turnover

intention as measurement variables and (b) the composite scores as latent variables. The maximum

likelihood method was used in this study to estimate the covariance matrix and the bias-corrected

percentile bootstrap test was used to test direct and indirect effects with 2,000 bootstrap samples. In

addition, cross-validation was used to deal with post-hoc model fitting issues and Byrne’s invariance

testing strategy was used to cross-validate the model with the χ2 difference test (Wu, 2010; Byrne,

2010).

A combination set of fit indices was used to assess the model fit, including the Chi-square, the

Chi-square/degrees of freedom ratio (χ2/df), the Standardized Root Mean Square Residual index

(SRMR), the Root Mean Square Error of Approximation index (RMSEA), the Goodness-of-Fit Index

(GFI) and the Normed Fit Index (NFI). Regarding the chi-square, it is very sensitive to sample size.

Hence, the null hypothesis is often expected to be rejected in a model with a relatively large sample size

(Wu, 2010). Therefore, we used the following cut-off values as criteria for adequate fit: χ2/df < 5,

SRMR < 0.05, RMSEA < 0.08, GFI > 0.80 and NFI > 0.90 (Wu 2010).

Validity and reliability

The original hypothetical model (see Figure 1) proposed in this study included paths reflecting the

theoretical expectations derived from the JD-R model and previous empirical results. We used SEM to

examine the whole model. The analyses showed the model fit indices did not achieve acceptable levels

and the direct path of job characteristics to turnover intention was not significant (β = -0.05, p > .05).

Hence, we revised the model: (1) removing the direct path of job characteristics; and (2) adding the

correlation path between work environment and job characteristics, based on the suggested

modification indices and empirical evidence (Rafferty et al., 2001). Recognizing that the revisions

This article is protected by copyright. All rights reserved.

 could cause the results to be sample specific, the sample was randomly divided into a calibration sample

(n = 392) and a validation sample (n = 386), using Byrne’s invariance testing strategy to cross-validate
Accepted Article
the model. The test revealed no statistically significant differences between the calibration and

validation samples for the measurement weights (△χ2 = 13, △df = 5.95, p =.95), structural weights (△χ2

= 4, △df = 4.15, p = .39), or structural covariances (△χ2 = 3, △df = 5.41, p = .14). The final model was

shown in Figure 2, where all loadings and paths were statistically significant (p < .01). Regarding the

quality of fit, the indices suggested a satisfactory level of fit with χ2 = 545.30，df = 114, χ2/df = 4.78,

SRMR = 0.04, RMSEA = 0.070 (90% CI = 0.064-0.076), GFI = 0.92 and NFI = 0.95.

RESULTS

Descriptive results

There was a total of 778 valid questionnaires; 720 participants were female (98.0%) with an

average age of 36.35 (SD 7.83). The sample was reasonably well-balanced with respect to tenure:

34.8% of participants worked more than 20 years, followed by 34.4% who worked 10-19 years and

30.7% who worked less than 10 years. Most of the participants were married (83.9%), were permanent

employees (84.8%) and held a Bachelors degree or higher (57.6%). They worked on various units,

including the surgical department (27.0%), medical department (19.8%), intensive care unit (17.5%),

operating room (14.1%), emergency room (12.9%) and the outpatient clinic (8.7%).

Table 2 presents the descriptive statistics for turnover intention, work engagement, work

environment and job characteristics. The composite mean of turnover intention was 2.48 (SD= 0.80),

with a total of 279 nurses (35.9%) at the high level of turnover intention. The composite mean of work

engagement was 3.44 (SD 1.52), with vigour rated the lowest (Mean = 3.26, SD= 1.63). The composite

score of work environment was 3.07 (SD= 0.52), with nurse participation in hospital affairs having the

lowest rating (Mean = 2.89, SD= 0.65), whereas the composite mean of job characteristics was 3.94

(SD= 0.53), with autonomy having the lowest rating (Mean = 3.79, SD= 0.70).

This article is protected by copyright. All rights reserved.

 Hypothesis testing

Table 3-5 and Figure 2 present the results of SEM and hypothesis testing. As shown in Figure 2,
Accepted Article
the whole model explained 50% of the variance in turnover intention and 30% of the variance in work

engagement. As expected, the results revealed a significant negative relationship between work

engagement and turnover intention (β = -0.33, p < .01), which meant Hypothesis 1 should be accepted.

The results also indicated that the direct effect (β = -0.49, p < .01) and indirect effect (β = -0.06, p

< .05) of work environment on turnover intention were both significant as well as its direct effect on

work engagement (β = 0.17, p < .01). This meant that work environment was positively related to higher

work engagement and lower turnover intention and that work engagement partially mediated the

relationship between work environment and turnover intention. These results supported Hypotheses 2,

4 and 6.

The results regarding job characteristics indicated that they had a significant indirect effect on

turnover intension (β = -0.14, p < .05) and a significant direct effect on work engagement (β = 0.47, p <

.01), but its direct effect on turnover intention was not significant. As mentioned above, this meant

motivational job characteristics were positively related to higher work engagement and lower turnover

intention and work engagement fully mediated the relationship between job characteristics and turnover

intention. These results supported Hypotheses 3 and 5, but not Hypothesis 7, which should be rejected.

DISCUSSION

The purpose of this study was to explore the state of turnover intention among experienced nurses

in China and examine the effects and the pathways of work environment, job characteristics and work

engagement on turnover intention. To our knowledge, this is the first study to examine the relationships

among these four variables in one structural model and to shed light on how work environment, job

characteristics and work engagement affect turnover intention. Based on the JD-R model and previous

research, we proposed a theoretical framework and specific hypotheses.

The results showed that 35.7% of experienced nurses had a high level of turnover intention, which

was in line with the results from other countries (Leone et al., 2015; Sawatzky & Enns, 2012). Hence,

This article is protected by copyright. All rights reserved.

 there was a general problem regarding the relatively high levels of turnover intention among

experienced nurses. The findings also showed that work environment had a significant direct effect on
Accepted Article
turnover intention and a significant indirect effect on turnover intention via work engagement. In

addition, a significant indirect effect of job characteristics on turnover intention via work engagement

was found, although the direct effect of job characteristics on turnover intention was not significant.

The findings were generally consistent with the proposed theoretical framework of the study and

supported all the study’s hypotheses except Hypothesis 7. These results had important theoretical and

practical implications.

Theoretical implications

This study yielded some meaningful theoretical findings. The most interesting findings were about

job characteristics, which received little attention in previous nursing research. The results found that

the association of job characteristics with turnover intention was fully mediated by work engagement.

The full-meditation model fitted well within the framework of the JD-R model, which featured work

engagement as the key mediator in the intrinsic motivational processes linking job characteristics to

positive outcomes (Bakker & Demerouti, 2007). The full-mediation effect indicated that work

engagement was necessary for nurses to transfer the benefits of motivational job characteristics to their

intention to stay in their job.

Though the results also confirmed the effect of work environment on turnover intention, this

association was only partially mediated by work engagement and the direct effect was much larger than

the indirect effect. Thus, the results only provided partial support for the premise of the JD-R model that

the work environment were involved extrinsic motivational processes; hence there might be other

important mechanisms that explained the pathway from work environment to turnover intention.

Several previous researches reported that high work environment was related to low nurse burnout and

in turn leaded to positive outcomes (e.g. low intention to leave) (Van Bogaert et al., 2009; Van Bogaert

et al., 2013), which was in line with the assumption that job resources might buffer the impact of job

demands on stress reaction and hence reduce burnout (Bakker & Demerouti, 2007). Accordingly, work

environment may act as a motivational factor to enhance positive psychological states (e.g. work

This article is protected by copyright. All rights reserved.

 engagement) as well as a protective factor to reduce negative states (e.g. burnout). Future research is

needed to confirm these dual mechanisms and the possibility of other mediators between work
Accepted Article
environment and turnover intention.

Additionally, the final model indicated that there was a positive relationship between job

characteristics and work environment. Although the relationship between job resources was not

described in the JD-R model, the modified job characteristics framework (Oldham & Hackman, 1981)

proposed that some features of work environment might influence employees’ perceptions of job

characteristics. The relationship was also supported by two studies in the nursing literature. A study by

Tummers et al. (2002) revealed that the degree of decision-making authority at the organizational level

influenced a sense of autonomy and a study by Rafferty et al. (2001) found that autonomy played a

central role in the nurse-physician relationship. Hence, it is suggested for future studies to prove the

relationship between job characteristics and work environment and examine that work environment

may work as a moderator of the relationships between job characteristics and work engagement.

In sum, this study furthered our knowledge by showing that work environment and job

characteristics could have an impact on turnover intention via work engagement and it lent empirical

support for the intrinsic and extrinsic motivational processes posited in the JD-R model. It also

suggested that future studies should explore other mediators that may link work environment to

turnover intention and examine the relationships among job resources.

Practical implications

The findings of this study provided some support for the applicability of the JD-R model to

experienced nurses in the context of Asian culture. In light of the finding that the final model, including

job characteristics, work environment and work engagement, explained 50% of the variance in

experienced nurses’ turnover intention, it is clearly essential for mangers to reduce the turnover

intention of experienced nurses through tailored strategies to improve their work environment, enhance

the motivational quality of job characteristics and promote nurses’ engagement.

This article is protected by copyright. All rights reserved.

 Furthermore, the results showed that job characteristics influenced turnover intention, indirectly

via work engagement rather than directly. In other words, employees must be engaged by motivational
Accepted Article
job characteristics to translate them into a low intention to leave. Thus, mangers should fulfil nurses’

needs for learning, growth and development by redesigning nursing work to foster their engagement

that subsequently keeps them in clinical practice. Considering that autonomy was rated the lowest job

characteristic by experienced nurses (see Table 2), managers should increase job autonomy to foster the

sense of responsibility and meaningfulness that nurses find in their work. Autonomy is a threshold issue

in nursing practice. As experienced nurses may have a high need for autonomy, nurse managers could

enhance their perception of autonomy by empowering them to determine the 4-T’s of task, time,

technique and team (Pink, 2009).

Last but not least, based on our results (a) that the work environment contributed to both nurse

engagement and turnover intention and (b) that the total effect of work environment on turnover

intention was stronger than the effects of the other two independent variables in this study, it was

extremely important to provide experienced nurses with adequate environmental resources to help them

feel engaged and thereby reduce their intention to leave. Given that experienced nurses were the least

satisfied with nurses’ participation in hospital affairs (see Table 2), mangers should allow experienced

nurses to be involved in organizational decisions about policies and practices, which also was a

commonly recognized characteristic of a supportive and effective administration (Lake, 2007).

Taken together, the findings of this study underscored the importance for nurse mangers to foster

nurse engagement and reduce turnover intention through enhancing the motivational quality of job

characteristics and creating a healthy work environment, especially in terms of autonomy and

participation in hospital affairs.

Limitations

Some limitations of our study must be acknowledged. First, this study used the self-report method

and common method variance could bias our results. However, Harman’s single-factor test (Podsakoff

et al., 2003) was used and it revealed that common variance was not a serious deficiency of this study.

In addition, despite a large sample from 7 hospitals, the data we obtained were only from tertiary public

This article is protected by copyright. All rights reserved.

 hospitals, so it would be interesting to replicate this study with a more diverse sample. Finally, this

study used a cross-sectional design, which did not allow an examination of causal relationships, hence,
Accepted Article
longitudinal studies would be needed to confirm our findings and examine causal relationships among

the variables.

Conclusion

Our results provided support for the links between work environment, job characteristics and

turnover intention, which were partially and fully mediated by work engagement, which leaded to

theory-driven interventions to address the nursing shortage and high turnover intention among

experienced nurses. Nursing leaders have suggested ways to retain experienced nurses by developing

strategies to promote wok engagement through creating a supportive work environment and enhancing

job characteristics, especially in terms of participation in hospital affaires and autonomy. Accordingly,

a self-governance model should be adopted to provide experienced nurses with the opportunity and

responsibility to (a) develop their work schedule, (b) exert authority over the content and context of

nursing care through input into policies, practices and nursing protocols and (c) provide input into

hiring decisions (Hess, 2004), which is beneficial to enhance nurses’ perception of autonomy and being

involved in hospital affairs decision.

This research confirmed the intrinsic motivation effect of job characteristics and the extrinsic

motivation effect of work environment on work engagement and turnover intention. Future research is

recommended to construct a complex model to examine the dual mechanism of work environment on

turnover intention via work engagement and burnout, the possible relationships among job resources

and the interaction of job demands and job resources.

This article is protected by copyright. All rights reserved.

 Author Contributions:

All authors have agreed on the final version and meet at least one of the following criteria
Accepted Article
(recommended by the ICMJE*):

1) substantial contributions to conception and design, acquisition of data, or analysis and

interpretation of data;

2) drafting the article or revising it critically for important intellectual content.

* http://www.icmje.org/recommendations/

References

Aiken, L. H., Clarke, S. P., & Sloane, D. M. (2002). Hospital staffing, organization and quality of care:

Cross-national findings. International Journal for Quality in Health Care, 14(1), 5-13. doi:

10.1093/intqhc/14.1.5

Aiken, L. H., Sermeus W., Van den Heede, K., Sloane, D. M., Busse, R., McKee, M., …Kutney-Lee,

A. (2012). Patient safety, satisfaction and quality of hospital care: cross sectional surveys of nurses and

patients in 12 countries in Europe and the United States. BMJ, 344, e1717. doi: 10.1136/bmj.e1717

Bakker, A. B., & Demerouti, E. (2007). The job demands-resources model: state of the art. Journal of

Management Psychology, 22, 309-328. doi: 10.1108/02683940710733115

Bishop, M. (2013). Work engagement of older registered nurses: the impact of a caring-based

intervention. Journal of Nursing Management, 21, 941-949. doi: 10.1111/jonm.12182

Byrne, B. M. (2010). Structural Equation Modelling With AMOS: Basic Concepts, Applications and

Programming (2th ed.). New York: Routledge.

Dail L.F., & Yang Z.P. (2004). Job evaluation: organizational diagnosis and research scales. Beijing,

China: China light Industry Press.

This article is protected by copyright. All rights reserved.

 Duffield, C. M., Roche, M. A., Homer, C., Buchan, J., & Dimitrelis, S. (2014). A comparative review of

nurse turnover rates and costs across countries. Journal of Advanced Nursing, 70, 2703-2712. doi:
Accepted Article
10.1111/jan.12483

Farh, J. L., Tsui, A. S., Xin, K., & Cheng, B. S. (1998). The influence of relational demography and

guanxi: The Chinese case. Organization Science, 9, 471-488. doi: doi.org/10.1287/orsc.9.4.471

Fritz, M. S., & Mackinnon, D. P. (2007). Required sample size to detect the mediated effect.

Psychological science, 18, 233-240. doi: 10.1111/j.1467-9280.2007.01882.x

Gauci Borda, R., & Norman, I. J. (1997). Factors influencing turnover and absence of nurses: a research

review. International Journal of Nursing Studies, 34, 385-394. doi:10.1016/S0020-7489(97)00031-X

Giallonardo, L. M., Wong, C. A., & Iwasiw, C. L. (2010). Authentic leadership of preceptors: predictor

of new graduate nurses’ work engagement and job satisfaction. Journal of Nursing Management, 18,

993-1003. doi: 10.1111/j.1365-2834.2010.01126.x

Hayes, L. J., O’Brien-Pallas, L., Duffield, C., Shamian, J., Buchan, J., Hughes, F., … Stone PW.

(2012). Nurse turnover: a literature review – an update. International Journal of Nursing Studies, 49,

887–905. doi: 10.1016/j.ijnurstu.2005.02.007

Hess, R. G. Jr. (2004). From bedside to boardroom - Nursing shared governance. The Online Journal of

Issues in Nursing, 9(1), Retrieved March 20, 2017, from

http://www.nursingworld.org.proxy.bc.edu/ojin/topic23/tpc23_1.htm

Iacobucci, D. (2010). Structural equations modelling: Fit Indices, sample size and advanced topics.

Journal of Consumer Psychology, 20, 90-98. doi: 10.1016/j.jcps.2009.09.003

Klien, F. D. (2005). Beyond significance testing: Reforming data analysis methods in behavioural

research. The American Journal of Psychiatry, 2005, 162, 643-644. doi: 10.1176/appi.ajp.162.3.643-a

Lake, E. T. (2007). The nursing practice environment. Medical Care Research and Review, 64,

104S-122S. doi: 10.1177/1077558707299253

This article is protected by copyright. All rights reserved.

 Laschinger, H.K.S., Wilk, P., Cho, J., & Greco, P. (2009). Empowerment, engagement and perceived

effectiveness in nursing work environments: does experience matter? Journal of Nursing Management,
Accepted Article
17, 636-646. doi: 10.1111/j.1365-2834.2008.00907.x

Leone, C., Bruyneel, L. anderson, J. E., Murrells, T., Dussault, G., Henriques de Jesus, É., … Rafferty,

A. M. (2015). Work environment issues and intention to leave in Portugueses nurses: a cross-sectional

study. Health Policy, 119, 1583-1592. doi: 10.1016/j.healthpol.2015.09.006

Meng, H., Song, J. W., Sun, Z. Q., & Wang, W. (2011). Transformational leadership affects the

employee's work outcomes: a mediation-mediator effect analysis. Journal of Psychological Science, 5,

1167-1173. doi: 10.16719/j.cnki.1671-6981.2011.05.010

Ministry of Health of China. (2015). China Health Statistical Yearbook 2015. Beijing, China: China

Union Medical University Press.

Oldham, G. R., & Hackman, J. R. (1981). Relationships between organizational structure and employee

reactions: comparing alternative frameworks. Administrative Science Quarterly, 26, 66-83. doi:

10.2307/2392600

Pink, D. H. (2009). Drive: The Surprising Truth About What Motivates Us. NY: Riverhead Books.

Podsakoff, P. M., Mackenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common Method Biases in

Behavioural Research: a critical review of the literature and recommended remedies. Journal of

Applied Psychology, 88, 879-903. doi: 10.1037/0021-9010.88.5.879

Rafferty, A. M., Ball, J., & Aiken, L. H. (2001). Are teamwork and professional autonomy compatible

and do they result in improved hospital care? Quality in Health Care, 10, ii32-ii37. doi:

10.1136/qhc.0100032

Santos, A., Chambel, M. J., & Castanheira, F. (2015). Relational job characteristics and nurses'

affective organizational commitment: the mediating role of work engagement. Journal of Advanced

Nursing, 72, 294-302. doi: 10.1111/jan.12834

This article is protected by copyright. All rights reserved.

 Sawatzky, J. V., & Enns, C. L. (2012). Exploring the key predictors of retention in emergency nurses.

Journal of Nursing Management, 20, 696-707. doi: 10.1111/j.1365-2834.2012.01355.x

Accepted Article
Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources and their relationship with

burnout and engagement: A multi-sample study. Journal of Organizational Behavior, 25, 293-315. doi:

10.1002/job.248

Sullivan Havens, D., Warshawsky, N.E., & Vasey, J. (2013). RN work engagement in generational

cohorts: the view from rural US hospitals. Journal of Nursing Management, 21, 927-940. doi:

10.1111/jonm.12171

Tu, L. J., Kong, W. L., & Zhang, P. (2013). A study on the relationship between the organizational

justice and turnover intension of newly-employed nurses of grade III – A general hospitals: as

intermediary with job satisfaction. Journal of Qilu Nursing, 19(15),15-17. doi:

10.3969/j.issn1006-7526.2013.15.006

Tummers, G. E. R., Van Merode, G. G., & Landeweerd, J. A. (2002). The diversity of work:

differences, similarities and relationships concerning characteristics of the organisation, the work and

psychological work reactions in intensive care and non-intensive care nursing. International Journal of

Nursing Study, 39, 841-855. doi: 10.1016/S0020-7489(02)00020-2

Van Bogaert, P., Clarke, S., Willems, R., & Mondelaers, M. (2013). Nurse practice environment,

workload, burnout, job outcomes and quality of care in psychiatric hospitals: a structural equation

model approach. Journal of Advanced Nursing, 2013, 68,1515-1524. doi: 10.1111/jan.12010

Van Bogaert, P., Meulemans, H., Clarke, S., Vermeyen, K. & Van de Heyning, P. (2009). Hospital

nurse practice environment, burnout, job outcomes and quality of care: test of a structural equation

model. Journal of Advanced Nursing, 65, 2175-2185. doi: 10.1111/j.1365-2648.2009.05082.x

Waldman, J. D., Kelly, F., Arora, S., & Smith, H. L. (2004). The shocking cost of turnover in health

care. Health Care Management Review, 29, 2-7. doi: 10.1097/00004010-200401000-00002

This article is protected by copyright. All rights reserved.

 Walker, A., & Campbell, K. (2013). Work readiness of graduate nurses and the impact on job

satisfaction, work engagement and intention to remain. Nurse Education Today, 33, 1490-1495. doi:
Accepted Article
10.1016/j.nedt.2013.05.008

Wang, L., & Li, L. Z. (2011). Validity and reliability of Chinese version of nursing practice

environment scale. Chinese Journal of Nursing, 46(2), 121-123. doi:

10.3761/j.issn.0254-1769.2011.02.002

World Health Organization. World health statistics 2011. Retrieved November 28, 2016 from

http://www.who.int/gho/publications/world_health_statistics/2011/en/

Wu, M. L. (2010). The structural equation model: AMOS operation and application. Chongqing, China:

Chongqing University Press.

Zhang, Y. W., & Gan, Y. Q. (2005). The Chinese Version of Utrecht work engagement scale: A

Examination of Reliability and Validity. Chinese Journal of Clinical Pychology, 13, 268-272.

doi:10.3969/j.issn.1005-3611.2005.03.005

This article is protected by copyright. All rights reserved.

 Table 1 Hypotheses

1 Nurses’ engagement is negatively associated with their turnover intention

Accepted Article
2 A supportive work environment is positively associated with nurse engagement

3 Job characteristics are positively associated with nurse engagement

4 The relationship between a supportive work environment and turnover intention

among nurses is mediated by work engagement

5 The relationship between job characteristics and turnover intention among nurses is

mediated by work engagement

6 A supportive work environment has a directly negative association with nurses’

turnover intention

7 Job characteristics have a directly negative association with nurses’ turnover intention

This article is protected by copyright. All rights reserved.

 Table 2 Descriptiveness of turnover intention, work engagement, work environment, and job

characteristics
Accepted Article
Maximum Mean SD

turnover intention 5 2.48 0.80

work engagement 6 3.44 1.52

dedication 6 3.73 1.66

absorption 6 3.44 1.52

vigor 6 3.26 1.63

work environment 4 3.07 0.52

nursing foundation for care quality 4 3.22 0.49

nurse-physician relations 4 3.22 0.54

nurse manager ability and leadership 4 3.14 0.56

staffing-resource adequacy 4 2.90 0.65

nurse participation in hospital affairs 4 2.89 0.65

job characteristics 5 3.94 0.53

task significance 5 4.08 0.59

skill variety 5 4.02 0.62

job feedback 5 4.00 0.56

task identity 5 3.84 0.69

autonomy 5 3.79 0.70

This article is protected by copyright. All rights reserved.

 Table 3 Correlations between the study variables

Variables 1 2 3
Accepted Article
1 turnover intention

2 work engagement -0.52**

3 work environment -0.59** 0.45**

4 job characteristics -0.50** 0.50** 0.68**

** p < .01

Table 4 Loadings of the measurement models

Loadings

turnover intention

often consider resigning from my current job 0.85

have long-term career development in the current organization 0.76

haven’t a promising prospect in the current organization 0.82

plan to turn over in next 12 months 0.79

work engagement

dedication 0.91

absorption 0.92

vigor 0.94

work environment

nursing foundation for care quality 0.92

nurse-physician relations 0.91

This article is protected by copyright. All rights reserved.

 nurse manager ability and leadership 0.88

staffing-resource adequacy 0.88

Accepted Article
nurse participation in hospital affairs 0.80

job characteristics

task significance 0.70

skill variety 0.82

job feedback 0.73

task identity 0.82

autonomy 0.85

This article is protected by copyright. All rights reserved.

 Table 5

Total effects, direct effects and indirect effects of work engagement, work environment, job
Accepted Article
characteristics on turnover intention

△
BC 95% CI
Estimate B β S.E.
Lower Upper

Total Effects

Work engagement  Turnover intention -0.22 -0.33 0.03 -0.28 -0.18

Work environment  Turnover intention -1.16 -0.54 0.09 -1.34 -0.10

Job characteristics Turnover intention -0.3 -0.14 0.06 -0.42 -0.20

Direct effects

Work engagement  Turnover intention -0.22 -0.33 0.03 -0.28 -0.18

Work environment  Turnover intention -1.04 -0.49 0.09 -1.22 -0.88

Work environment  Work engagement 0.52 0.17 0.17 0.18 0.84

Job characteristics Work engagement 1.32 0.47 0.18 0.98 1.67

Indirect effects

Work environment  Turnover intention -0.12 -0.06 0.04 -0.20 -0.05

Job characteristics Turnover intention -0.3 -0.14 0.06 -0.42 -0.20

△
: That a 95% bias corrected bootstrap confidence interval does not include zero means significant effects.

This article is protected by copyright. All rights reserved.

 Figure 1

Theoretical model and hypotheses

Accepted Article
-
+

+
-
H7

Figure 2

Final model and standardized model paths

TI: turnover intention; WE: work engagement;

JC: Job characteristics; NWE: nurse work environment.

This article is protected by copyright. All rights reserved.