Scholarly article on topic 'Burnout and Workload Among Health Care Workers: The Moderating Role of Job Control'

Burnout and Workload Among Health Care Workers: The Moderating Role of Job Control Academic research paper on "Psychology"

Share paper
Academic journal
Safety and Health at Work
OECD Field of science
{burnout / cynicism / exhaustion / "job control" / workload}

Abstract of research paper on Psychology, author of scientific article — Igor Portoghese, Maura Galletta, Rosa Cristina Coppola, Gabriele Finco, Marcello Campagna

Abstract Background As health care workers face a wide range of psychosocial stressors, they are at a high risk of developing burnout syndrome, which in turn may affect hospital outcomes such as the quality and safety of provided care. The purpose of the present study was to investigate the moderating effect of job control on the relationship between workload and burnout. Methods A total of 352 hospital workers from five Italian public hospitals completed a self-administered questionnaire that was used to measure exhaustion, cynicism, job control, and workload. Data were collected in 2013. Results In contrast to previous studies, the results of this study supported the moderation effect of job control on the relationship between workload and exhaustion. Furthermore, the results found support for the sequential link from exhaustion to cynicism. Conclusion This study showed the importance for hospital managers to carry out management practices that promote job control and provide employees with job resources, in order to reduce the burnout risk.

Academic research paper on topic "Burnout and Workload Among Health Care Workers: The Moderating Role of Job Control"

Accepted Manuscript

Burnout and workload among healthcare workers: the moderating role of job control

Igor Portoghese , Maura Galletta , Rosa Cristina Coppola , Gabriele Finco , Marcello Campagna

PII: S2093-7911(14)00041-9

DOI: 10.1016/

Reference: SHAW 46

To appear in: Safety and Health at Work

Received Date: 14 November 2013

Revised Date: 18 April 2014

Accepted Date: 29 May 2014

Please cite this article as: Portoghese I, Galletta M, Coppola RC, Finco G, Campagna M, Burnout and workload among healthcare workers: the moderating role of job control, Safety and Health at Work (2014), doi: 10.1016/

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Burnout and workload among healthcare workers: the moderating role of job control

Igor Portoghese, Maura Galletta*, Rosa Cristina Coppola, Gabriele Finco, Marcello Campagna Department of Public Health, Clinical and Molecular Medicine, University of Cagliari, SS 554

bivio Sestu, 09042, Monserrato (Ca), Italy

Correspondence address: Maura Galletta, University of Cagliari, Faculty of Medicine, SS 554 bivio Sestu, 09042, Monserrato (Ca), Italy; Cell: (+39) 3284869036; Institutional contacts: (+39) 070 675 4676; E-mail:

Running title: Burnout and workload: the moderating role of control

Sources of Funding: Authors declare that no funding has been received for this research.

Burnout and workload among healthcare workers: the moderating role of job control Abstract

Background: As health care workers face a wide of psychosocial stressors, they are at high risk of developing burnout syndrome, which in turn may affect hospital outcomes such as the quality and safety of provided care. The purpose of the present study was to investigate the moderating effect of job control on the relationship between workload and burnout.

Methods: 352 hospital workers from five Italian public hospitals completed a self-administered questionnaire including measures of exhaustion, cynicism, job control and workload. Data were collected in 2013.

Results: In contrast to previous studies, the results of this study showed support for the moderation of job control in the relationship between workload and exhaustion. Furthermore, the results found support for the sequential link from exhaustion to cynicism.

Conclusion: This study showed the importance for hospital managers to carry out management practices that promote job control and provide employee job resources in order to reduce the burnout risk.

Keywords: burnout, exhaustion, cynicism, healthcare workers, job control, workload


Stress in the workplace is globally considered a risk factor for workers' health and safety. More specifically, the health care sector is a constantly changing environment and hospitals are increasingly becoming demanding and stressful working contexts. According to the World Health Organization (WHO), "a healthy workplace is one in which workers and managers collaborate to use a continual improvement process to protect and promote the health, safety and well-being of all workers and the sustainability of workplace [...]" [1]. Despite the WHO's aim to promote and foster healthy work environments, in 2000 approximately 2 million work-related deaths occurred [2]. In the health care sector, several studies have shown that health care professionals are exposed to a variety of severe occupational stressors, such as time pressure, low social support at work, a high workload, uncertainty concerning patients treatment, and predisposition to emotional responses due to exposure to suffering and dying patients [3, 4]. In this sense, health care workers are at a high risk of experiencing severe distress, burnout, and both mental and physical illness. In turn, this could affect hospital outcomes, such as the quality of care provided by such institutions [4-7]. Particularly, in the last 35 years, the prevalence of stress-related illnesses such as burnout has increased significantly, affecting 19% -30% of employees in the general working population globally [8]. Burnout among healthcare workers, mainly medical staff, was becoming an occupational hazard, reaching rates between 25% and 75% in some clinical specialty [9]. Furthermore, it was reported that among the sources of occupational illnesses, burnout represents the 8% of the cases of occupational illnesses [10].

As defined by Leiter and Maslach [11, 12], burnout is a cumulative negative reaction to constant occupational stressors relating to the misfit between workers and their designated jobs. In this sense, burnout is a psychological syndrome of chronic exhaustion, cynicism, and inefficacy, and is experienced as a prolonged response to chronic stressors in the workplace [13]. Exhaustion is mainly related to an individual's experience of stress, which is in turn related to a decline in emotional and physical resources. According to Leiter and Maslach [14], "the experience of

exhaustion reduces workers initiative while progressively limiting their capacity for demanding work" (p. 50). Cynicism refers to detachment from work in reaction to the overload of exhaustion [13]. Cynicism pertains to the loss of enthusiasm and passion for one's work [14].The third component, perceived professional inefficacy, refers to feelings of ineffectiveness and lack of achievement and productivity at work. Perceived professional inefficacy refers to the loss of confidence in one's work [14]. Particularly, Maslach, Schaufeli and Leiter [15], hypothesized that the three dimensions of burnout develop as a result of varying sequential progression over time. Previous research on burnout has confirmed the sequential link from exhaustion to cynicism [15]. Specifically, researchers has found "that exhaustion occurs first, leading to the development of cynicism, which in turn leads to inefficacy. However, the subsequent link to inefficacy is less clear, with the current data supporting a simultaneous development of this third dimension rather than a sequential one" [15] (p. 406).

Job burnout has been associated with a multiplicity of health problems, such as hypertension, gastrointestinal disorders, and sleeplessness [15]. It has also been associated with performance-related issues [16], demonstrating its direct impact on workplace effectiveness. Regarding the etiology of burnout, researchers have mainly focused on the role played by an occupational context. Maslach and Leiter [17] provided a more comprehensive perspective by identifying six general areas of a working life considered as the most important antecedents of burnout: a manageable workload, job control, rewards, community, fairness, and values. According to this model, a mismatch between one's expectations and the structure or process within the occupational environment contributes toward burnout. These six areas have different relationships with the three dimensions of burnout [11, 18]. Mainly building on the Demand-Control theory of job stress described by Karasek and Theorell [19], authors assert that mismatches in workload and job control may aggravate exhaustion through excessive demands, by generating a general condition of anxiety. In contrast, a manageable workload sustains energy, thus contrasting the risk of burnout. A mismatch in workload means that workers feel overworked and\or do not have

enough time to do the job. Work overload is a major source of exhaustion that, in turn, is at the root of burnout [14], representing the basic individual stress component of burnout [20]. In addition, lack of job control means that employees' sense of autonomy and discretion are limited. As a result, their sense of control over what they do is limited or undermined, which also means that they do not have much of a say in what goes on in their work environments. In contrast, job control pertains workers' ability to take decisions regarding their work [11]. As described by Leiter and Maslach, control plays an important role in influencing, either directly or indirectly, workload and burnout among employees. In this sense, more control gives workers the opportunity to shape their work environment, such as reducing their workload accordingly. This is in line with the buffer hypothesis of job stress, where high job demands (mainly, high workload) coupled with low control lead to job strain. In this sense, it is central to clarify and control the variables involved in the job burnout process. This will enable the development of strategies aimed at protecting health care professionals from the risk of burnout [5].

The purpose of this study is to develop and test a conceptual model of the relationship between work environment (workload and job control) and burnout (cynicism and exhaustion) among Italian healthcare professionals. Specifically, the following working hypotheses were tested (Figure 1): Hypothesis 1a: workload is positively related to exhaustion Hypothesis 1b: job control is negatively related to exhaustion

Hypothesis 2: job control moderates the relationship between workload and exhaustion Hypothesis 3: exhaustion is positively related to cynicism

Hypothesis 4: exhaustion mediates the relationship between workload and cynicism

[Figure 1 here]

Materials and Methods

The study was performed in accordance with the code of ethics of the World Medical Association (Declaration of Helsinki). Participants and data collection

A cross-sectional survey was conducted. The study participants were recruited in January 2013 from five Italian Hospitals. A total of 352 hospital workers (nurses and other clinical professionals) voluntarily completed to a self-administered paper questionnaire distributed to 434 workers, representing a return rate of 81.1%. Researchers provided a briefing about the study objectives as well as statements guaranteeing both confidentiality and anonymity. The hospital workers were given 3 weeks to complete and return their questionnaires in locked boxes.

In total, the sample is composed of 352 health care workers with an average age varying between 40 and 46 years. Of these, 74.1% were woman and 61.1% have been working in the actual unit for more than 10 years. Ethical permission

Formal approval from the local ethical committee was not required according to national legislation in Italy.


The Exhaustion (5 items) and Cynicism (5 items) subscales of the Maslach Burnout Inventory-General Survey (MBI-GS) [13, 21] were used to measure burnout. Participants used a 7-point Likert scale ranging from 0 (never) to 6 (every day) to rate the extent to which they experience exhaustion and cynicism at work (e.g., "I feel burned out from my work"). In the present study, the internal reliability for each subscale was .87 for exhaustion and .77 for Cynicism.

Two subscales of the Areas of Worklife Scale (AWS) measure [11, 22] were used to measure workload (3) and job control (3). The items are worded as statements of perceived congruence or incongruence between oneself and the job. Thus, each subscale includes positively worded items of congruence, for example, "I have enough time to do what's important in my job" (workload), and negatively worded items of incongruence, for example, "Working here forces me to compromise my values" (values). Respondents indicate their degree of agreement with these statements on a 5-point Likert-type scale ranging from 1 (strongly disagree), through 3 (hard to decide), to 5 (strongly agree). The scoring for the negatively worded items is reversed.

The scale has yielded a consistent factor structure across samples with acceptable alpha levels: workload (.71), job control (.65). An indication of the subscales' construct validity is that when respondents were given an opportunity to comment on any issue in their worklives, the topics on which they wrote complaints corresponded with the areas of worklife that they evaluated negatively [11].

Data analysis

We tested our study hypotheses using the principles of structural equations modelling (SEM) techniques with the statistical package AMOS 19.0 software package (SPSS: An IBM Company, Chicago, IL, USA). Following the Anderson & Gerbing's two-step approach [23], we tested the measurement structure and the structural relationships in two separate steps. First, evaluation of the measurement model was carried out using exploratory factor analysis, complemented by confirmatory factor analysis. The second step tests the proposed model along with tests of rival alternative models.

As our hypotheses included both moderation and mediation effects, we had to apply different techniques. In order to investigate the moderation effect (Hypothesis 2), we followed the recommendations by Little et al. [24]. In particular, we used orthogonal centered product terms of the latent construct to model the interactions in our structural model. First, we multiplied an uncentered indicator of workload with an uncentered indicator of job control. This resulted in nine product terms. Then, we regressed each of the nine products on all indicators. The residual of this regression was saved in the data file. The nine residuals were used for the measurement of the latent product term variable. In the second step we included the nine orthogonalized product terms as indicators of a single latent interaction construct. For each latent variable (workload, job control, and the latent product (workload*jobcontrol), one factor loading was fixed to one to provide a scale for the respective latent variable. In addition, we specified error covariances between the residual variances of the interaction products. To better interpret the interactions, we also did a graphical plotting of the results and simple slope testing, as proposed by Aiken and West [25]. The significant

interaction effects were represented by plots. Independent lines of regression were generated from the regression equation to represent the relationship between the independent variable and the dependent variable, defining the high values and the low values of the moderator variable at relatively one SD above and below the mean. Following the recommendations of Aiken and West [25], simple effects tests were conducted to determine whether the slopes differed significantly from zero. The mediation effect (Hypothesis 4) was tested in SEM by comparing the mediation model to the baseline model and, following the recommendation by Cheung and Lau [26], we also applied bootstrapping procedures to test for the significance of the indirect effect. We tested our hypothesis by comparing models, using the Ax2 test with one or two degrees of freedom [27]. Following the recommendation by Bentler [28], in addition to %2 statistic, the overall fit was assessed using the comparative fit index (CFI), the incremental fit index (IFI), and the root mean square error of approximation (RMSEA). As a rule of thumb, applied cut-off values for the IFI and CFI are values > 0.90 [29] and < 0.06 for the RMSEA to indicate acceptable model fit [30].


Table 1 shows the means, standard deviations and zero-order correlations among the variables. The factorial structure of measures (burnout, workload and job control measures) was examined. All indicators loaded significantly on their corresponding latent constructs (p<.001) and the model showed a good fit to the data supporting the hypothesized structure, %2 (df=91)=205.9; IFI=.94; CFI =.94; RMSEA =.06. We also compared the measurement model to three alternative models (table 2). First, a three-factor model in which the job control and workload items loaded on one common factor (alternative model 1) had a significantly worse fit (Ax2(3)= 91.0; p < .001). Second, a three-factor model with the exhaustion and cynicism items loading on one factor (alternative model 2) was worse fitting (Ax2(3)= 72.8; p< .001). Finally, a one factor model (alternative model 3), with all items loading on one common factor was worse fitting (Ax2(6)= 231.6; p< .001).

In the second step of our analysis, we investigated the structural relationships specified in the hypothesized model. We specified paths from the control variables to all dependent study constructs.

[Table 1 here]

Latent variable model results

The overall results of the moderated-indirect model suggest that the fit model to the data is good: %2 (df=251) = 390.7 ; IFI= .96; CFI = .96; RMSEA = .04. All factor loadings were significant. A look at path coefficients revealed that all paths were significant (p<.05). To determine whether our model was parsimonious, the hypothesized model was compared to alternative models that added or dropped paths. As shown in Table 3, first, the alternative model 1 which allowed a path from workload to cynicism, showed a worse fit: A-/2( 1 )= 2.5, ns. Alternative model 2 restricted the effect of the interaction term to zero, showed a worse fit: Ax2(l)= 5.7, p < .05. Finally, the alternative model 3 which allowed a path from job control to cynicism, showed a worse fit: Ax2(l)= 3.0, ns. Overall, the hypothesized model was significantly better than the comepting models, confirming all our. As proposed in Hypotheses la and lb, workload (13 = .57, p<.001), and job control (13 = -.21, p<.001) had the hypothesized relationship to exhaustion.

We also obeserved the expected moderation effect of workload and job control on exhaustion (13 = . 14, p < .05; Hypothesis 2). In order to interpret the form of interaction, the equation at the high and low levels of job control was plotted according to the procedure that was proposed by Aiken and West [24]. The results showed that the form of the interaction was in the expected direction. An increase in the workload was significantly associated with higher job exhaustion; this relationship was attenuated by high job control (Figure 2). The health care workers were more exhausted in response to higher levels of workload when they had low job control. Thus, hypothesis 2, about the moderating effect of job control, also was supported.

[Figure 2 here]

We also observed the expected relationship between exhaustion and cynicism (B =.64, p <. 001) confirming the hypothesis 3. Finally, since both the relationship between workload and exhaustion as well as the relationship between exhaustion and cynicism were significant, we carried out bootstrapping procedures as proposed by Cheung and Lau [26] to test Hypothesis 4. The results from 1000 bootstrapping samples showed a significant indirect relation between workload and cynicism via the transmission of exhaustion (B = .45; 95% confidence intervals .35 ~ .39) and thus it confirmed the full mediation hypothesis.

2 2 This final model (Figure 3) accounted for R =40% of the variance in exhaustion and R =42% of

variance in cynicism.

[Figure 3 here]

[Table 2 here]


Burnout among health care workers is associated with high turnover ratesand absenteeism due to sickness, relative ineffectiveness in the workplace as well as low job satisfaction [15, 31]. In view of this, it is important to identify organizational stressors that are related to job burnout in order to promote and facilitate strategies aimed at its prevention and reduction. The key finding of this study was the noteworthy moderation effect of job control on the relationship between workload and exhaustion. This interaction is considered one of the most controversial aspects of Karasek et al.'s theory. However, previous studies have shown that workload contributes towards the prediction of employee exhaustion [11, 32-33] thus indicating incompatibility with Karasek et al.'s interaction hypothesis [19]. Recently, Taris [34] showed that, of the 90 studies in which this interaction was tested, only nine provided support for the hypothesized interaction. Building on this result, we found a positive association between workload

and exhaustion and this relationship was strongest when job control was lower. In this sense, both workload and job control play an important role in improving working conditions. In turn, improved working conditions are demonstrated by a low workload and exhaustionlevels, which can also be attributed to an increase in job control. In this manner, job control seems to protect workers from exhaustion when the workload increases. Our findings showed that high workload does not pose major concerns when workers have sufficient job control.

Furthermore, with regard to the burnout dimension, exhaustion was found to be positively associated with cynicism. This result confirmed the well-known sequential progression from exhaustion to cynicism [13], as defined in the theoretical model [13, 35]. Specifically, the results are in line with literature, which indicates that exhaustion occurs first while cynicism occurs as a reaction to excessive exhaustion [15, 36-37].

The well-being of health care workers depends on the quality of their work environment. In the last 30 years, many scholars have examined factors contributing to job burnout. However, remains to continue studying this phenomenon, in order to build and sustain healthy work environments. The results of this research show the importance of developing organizational management practices that enable job control and provide employee with the resources to mitigate the risk of burnout. Maslach [12] and Maslach and Leiter [11] have argued that organizational' interventions aimed at reducing the risk of burnout should be framed according to the dimension considered (exhaustion, cynicism, and sense of efficacy). These authors developed the Areas of Work life Model [11], proposing that organizational interventions consider policies and practices that are capable of shaping the six key areas of work life (manageable workload, job control, reward, community, fairness, and values). This suggests that organizations would benefit from interventions aimed at reducing the workload and fostering job control. There is a large agreement about the fact that "preventing burnout is a better strategy than waiting to treat it after it becomes a problem" [38]. In fact, beyond the impact to individual workers' health, burnout also poses risks to others, in the form of workplace accidents, injuries, and fatalities [39]. Furthermore, there is an unexplored issue

relating to the crossover effect of burnout. Specifically, it concerns the interpersonal process that occurs when job stress, as experienced by one person, affects the level of strain experienced by another person in the same social environment [40-41]. Thus, the crossover effect describes the burnout contagion effect among professionals in the same work environments [41]. While the study findings provide support to the proposed hypotheses, the research design and sample are subject to limitations. First, participants were not randomly selected from the national health-care system. This may create a selection bias and limit the generalizability of the results. The study must be replicated by analyzing a larger and more representative sample of healthcare workers in order to validate the model further. A second limitation is with regard to the fact that this study was a cross-sectional; thus, no hard conclusions can be drawn with regard to causation. Burnout is a process and longitudinal data will be necessary to establish causality among the relationships studied.

To reduce the risk of burnout, intervention programs should be aimed at reducing worker's experience of stressors and, subsequently, , should be directed towards both individuals and organizations [42]. Following Leiter and Maslach's [14] approach, in controlling the risk of burnout, health care managers should devise strategies aimed at reducing workers' workload and increasing their sense of control. Firstly, reducing workers' workload when job resources are limited could rpose major challenges for health care managers. However, in instances where it is difficult to hire new employees due to economic and regulatory constraints, managers could provisionally reduce the workload by providing employee with flexible schedule, such as floating workforce (primarily applicable to nurses). Health care managers may improve workers'sense of control by promoting their autonomy in the workplace. In fact, job autonomy is considered an important coping strategy in decreasing job strain [19, 43].

Finally, an interesting avenue for future research is the investigation into the contagious nature of burnout by considering the workplace-related crossover effect among health care professionals in a multilevel research design. In this manner, an investigation of crossover as a unit level factor could

expand the current boundaries of burnout models, giving the consideration of a unit burnout level and its effect on individual burnout. Conflict of interest

There is no conflict of interest to be declared. References

[1] Burton J. World Health Organization Healthy Workplace Framework and Model: Background and supporting literature and practices. 2010, p.3.

[2] World Health Organization. Global strategy on occupational health for all: The way to health at work. 2008. Available from:

[3] McVicar A. Workplace stress in nursing: a literature review. J Adv Nurs 2003;44:633-42.

[4] Marine A, Ruotsalainen JH, Serra C, Verbeek JH. Preventing occupational stress in healthcare workers. Cochrane Database of Syst Rev 2006; 4 Art. No.: CD002892.

[5] Wu S, Li H, Zhu W, Lin S, Chai W, Wang X. Effect of Work Stressors, Personal Strain, and Coping Resources on Burnout in Chinese Medical Professionals: A Structural Equation Model. Ind Health 2012; 50:279-287.

[6] Nieuwenhuijsen K, Bruinvels D, FringsDresen M. Psychosocial work environment and stress related disorders, a systematic review. Occup Med (Lond) 2010; 60:277-86.

[7] Stansfeld S, Candy B. Psychosocial work environment and mental health—a meta-analytic review. Scand J Work Envi-ron Health 2006;32:443-462.

[8] Finney C, Stergiopoulos E, Hensel J, Bonato S, Dewa CS. Organizational stressors associated with job stress and burnout in correctional officers: a systematic review. BMC Public Health 2013; 13:82.

[9] Laschinger HKS, Wong C, Greco P. The impact of staff nurse empowerment on person-job fit and work engagement/burnout. Nurs Adm Q 2006; 30:358-367.

[10] Sundin L, Hochwa'lder J, Bildt C, Lisspers J. The relationship between different work-related sources of social support and burnout among registered and assistant nurses in Sweden: a questionnaire survey. Int J of Nurs Stud 2006; 44: 758-769.

[11] Leiter MP, Maslach C. Areas of worklife: A structured approach to organizational predictors of job burnout. In P. Perrewé & D. C. Ganster, (Eds.), Research in occupational stress and well being (Vol. 3, pp. 91-134). Oxford, UK: Elsevier; 2003.

[12] Maslach C. Job burnout: New directions in research and intervention. Curr Dir Psychol Sci, 2003;12:189-192.

[13] Maslach C, Jackson SE, Leiter MP. Maslach burnout inventory manual (3rd edn.). Palo Alto, CA: Consulting Psychologists Press; 1996.

[14] Leiter MP, Maslach C. A mediation model of job burnout. In Antoniou, A. S. G., & Cooper, C. L. (Eds.), Research companion to organizational health psychology (pp. 544-564). Cheltenham, UK: Edward Elgar; 2005.

[15] Maslach C, Schaufeli WB, Leiter MP. (2001). Job burnout. Annu Rev of Psychol 2001; 52: 397-422.

[16] Leiter MP, Harvie PL, Frizzell C. The correspondence of patient satisfaction and nurse burnout. Soc Sci Med 1998; 37:1-7.

[17] Maslach C, Leiter MP. The truth about burnout. San Francisco: Jossey-Bass; 1997.

[18] Maslach C, Leiter MP. (2008). Early predictors of job burnout and engagement. J of Appl Psychol 2008; 93:498-512.

[19] Karasek R, Theorell T. Stress, productivity, and the recon-struction of working life. New York: Basic Books; 1990.

[20] Maslach C. Understanding job burnout. In A. M. Rossi, P. Perrewe, and S. Sauter (Eds.), Stress and quality of working life: Current perspectives in occupational health (37-51). Greenwich, CT: Information Age Publishing; 2006.

[21] Schaufeli WB, Leiter MP, Maslach C, Jackson SE. The Maslach Burnout Inventory—General Survey. In C Maslach, SE Jackson, MP Leiter (Ed.), MBI manual (3rd ed.). Palo Alto, CA: Consulting Psychologists Press; 1996.

[22] Leiter MP, Maslach C. Preventing burnout and building engagement: A complete program for organizational renewal. San Fran-cisco: Jossey Bass; 2000b.

[23] Anderson JC, Gerbing DW. Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach. Psychol Bull 1988; 103: 411-423.

[24] Little TD, Bovaird JA, Widaman KF. On the merits of orthogo-nalizing powered and product terms: Implications for modeling latent variable interactions. Struct Equ Modeling 2006; 13:479519.

[25] Aiken LS, West SG. Multiple Regression: Testing and Interpreting Interactions. Newbury Park, CA: Sage; 1991.

[26] Cheung GW, Lau RS. Testing mediation and suppression effects of latent variables: Bootstrapping with structural equation models. Organ Res Methods 2008; 11(2): 296-325.

[27] Kline RB. Principles and Practice of Structural Equation Modeling (2nd Edition ed.). New York: The Guilford Press; 2005.

[28] Bentler PM. Comparative Fit Indexes in Structural Models. Psychol Bull 1990; 107(2):238-46.

[29] Hu L, Bentler PM. Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychol Methods 1998; 3:424-453.

[30] Browne MW, Cudeck R. Alternative ways of assessing model fit. In: KA Bollen & JS Long (Eds.), Testing structural equation models (pp. 136-162). Beverly Hills, CA: Sage; 1993.

[31] Schaufeli WB, Leiter MP, Maslach C. Burnout: 35 years of research and practice. Career Development International 2008; 14:204-220.

[32] Leiter MP, Shaughnessy K. The areas of worklife model of burnout: Tests of mediation relationships. Ergonomia: An International Journal 2006; 28:327-341.

[33] Leiter MP, Maslach C. Burnout and workplace injuries: A longitudinal analysis. In A. M. Rossi, J. C. Quick, & P . L. Perrewe (Eds.), Stress and quality of work-ing life: The positive and the negative (pp. 3-18). Green-wich, CT: Information Age; 2009.

[34] Taris TW. Bricks without clay: On urban myths in occupational health psychology. Work Stress 2006; 20:99-104

[35] Maslach C. Burnout: A multidimensional perspective. In W. B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional burnout: Recent developments in theory and research(pp. 19-32). Washington, DC: Taylor & Francis; 1993.

[36] Leiter MP, Jackson NJ, Shaughnessy K. Contrasting burnout, turnover intention, control, value congruence, and knowledge sharing between boomers and generation X. J Nurs Manag 2008; 16: 100-109.

[37] Gascon S, Leiter MP, Pereira JP et al. The role of aggressions suffered by healthcare workers as predictors of burnout. J Clin Nurs, 2013; 22:3120-3129.

[38] Maslach C. Engagement research: Some thoughts from a burnout perspective. Eur J Work Org Psychol 2011; 20: 47-52. (p.45)

[39] Nahrgang JD., Morgeson FP, Hoffmann DA. Safety at Work: A Meta-Analytic Investigation of the Link Between Job Demands, Job Resources, Burnout, Engagement, and Safety Outcomes. J Appl Psychol 2011; 96(1): 71-94.

[40] Bolger N, DeLongis A, Kessler RC, Wethington E. The contagion of stress across multiple roles. J Marriage Fam 1989a; 51:175-183.

[41] Westman M, Bakker AB. Crossover of burnout among health care professionals. In J. Halbesleben (Ed.), Stress and burnout in health care (pp. 111-125). Hauppauge, NY: Nova Science; 2008.

[42] Awa W, Plaumann M, Walter U. Burnout prevention: A review of intervention programs. Patient Educ Couns 2010; 78:184-190.

[43] Hausser JA, Mojzisch A, Niesel M,Schulz-Hardt S. Ten years on: A review of recent research on the job demand-control (-support) model and psychological well-being. Work Stress 2010; 24(1): 1-35.

Means, standard deviations, and correlations of variables (N=352)

M SD 1 2 ^ 3 4

1. Workload 3.14 1.04 1

2. Job control 3.28 0.88 -0.02 1

3. Exhaustion 2.69 1.50 0.42* -0.17* 1

4. Cynicism 1.76 1.35 0.23* -0.19* 0.53* 1

Note. * p < 0.01

Fit Indices for Measurement Models

Model X2 df Ax2 Adf P IFI CFI RMSEA

Four-independent-factor measurement model 205.9 91 .94 .94 .06

Alternative model 1 (job control and workload as one factor) 296.9 94 91.0 3 .001 .90 .90 .08

Alternative model 2 (exhaustion and cynicism as one factor) 278.7 94 72.8 3 .001 .91 .91 .08

Alternative model 3 (one-factor model) 437.5 97 231.6 6 .001 .83 .83 .10

Note: N = 352. A %2 different test was assessed in contrasting measurement model against three nested models.

Fit Statistics for All Models Tested

Model X2 df Ax2 Adf P IFI CFI RMSEA

Hypothesized model 390.7 251 .96 .96 .04

Alternative model 1 (allowed path workload —> cynicism) 388.2 250 2.5 1 ns .96 .96 .04

Alternative model 2 (rescticted to zero the interaction effect) 396.2 252 5.5 1 .05 .96 .96 .04

Alternative model 3 (allowed path job control —> cynicism) 387.7 250 3 1 ns

Note: N = 352. A %2 different test was assessed in contrasting measurement model against three nested models.

Figure 1. Hypothesized model. H, Hypothesis. Hypothesis 1a: workload is positively related to exhaustion; Hypothesis 1b: Job control is negatively related to exhaustion; Hypothesis 2: Job control moderates the relationship between workload and exhaustion; Hypothesis 3: exhaustion is positively related to cynicism; Hypothesis 4: exhaustion mediates therelationship between workload and cynicism;

5.0 -I


HLJJI-C-eTlCSI Leu r-Owvtrol


-0.2 05


—i— Is

Figure 2. Two-way interaction effect of workload and control on exhaustion.

Figure 3. Results of the structural equation modelling analysis of the hypothesized model with standardized path coefficients for mediating and moderating effects. (*P < 0.05, two-tailed).