Conway, Paul Maurice PhD; Hogh, Annie PhD; Rugulies, Reiner PhD; Hansen, Åse Marie PhD
* Discuss the limited previous evidence on the relationship between sickness presenteeism and mental health in general and depression in particular.
* Summarize the new findings on the association between days of presenteeism and subsequent depression risk.
* Outline the implications for workplace policies, if depression is among the health consequences of presenteeism.
Sickness presenteeism (SP) indicates the “phenomenon that people, despite complaints and ill-health that should prompt them to rest and take sick-leave, go to work in any case.”1(p958) Recent data from the Fifth European Working Condition Survey reveal that working while ill is widespread among European workers.2 This survey estimated an EU-level average prevalence of SP (defined as the proportion of employees reporting at least 1 day of SP during the previous 12 months) of 41% among men and 45% among women, with aggregated peaks of more than 50% observed in a cluster of countries including Denmark.
So far, research on employee attendance behaviors has focused primarily on causes and consequences of sickness absence.3 Nevertheless, growing evidence shows that SP may also significantly endanger individual health and work productivity.1,4–7 In prospective studies, SP has been found in association with an increased risk of various health problems, including coronary heart diseases,8 low–self-rated health,9 and poor mental well-being.10–12 Evidence has also been reported on SP as a predictor of future sickness absence.12–15 In these studies, the unique role played by SP in impairing health status is corroborated by the finding that the relationship between working while ill and future health remained significant even after adjustment for baseline health status. In other words, the higher risk for reduced health previously observed among employees who more often decided to show up for work despite sick could not be totally explained by the fact that their health status was lower than that of those employees with less frequent or no SP.1
To date, there are only few prospective studies investigating the relationship between SP and mental health problems.10–12 In particular, the potential role of SP as a risk factor for the onset of depression has been neglected so far. The link between SP and depression has been investigated only in studies focusing on the risk of reduced performance among employees going to work while depressed.6,16 Bridging this research gap may be important because of both the high prevalence of SP and the fact that depression represents the leading burden of disease in both middle- and high-income countries.17 An increased understanding of the possible environmental risk factors of depression, including those related to the workplace, is key to improve effectiveness in preventing the disorder.18
The previously mentioned prospective studies on the relationship between SP and mental health provided suggestive evidence for a causal relationship between working while ill and subsequent depression.10–12 In particular, Gustafsson and Marklund11 observed an association between SP and reduced well-being as measured with a general 10-item questionnaire that also covered depressive symptoms. A link between working while ill and depression seems plausible because SP often leads to inadequate recovery and to a cumulation of psychophysiological strain,8,12 which in turn, in accordance with the Allostatic Load hypothesis, may precipitate several kinds of diseases including depression.19,20 Despite these indications, however, the lack of direct evidence leaves open the question of whether a relationship actually exists between SP and the onset of depression.
It may be expected that the relationship between SP and subsequent depression differs between sexes. First, women are known to be at a higher risk for depression than men.21 Second, mainly because of their larger household responsibilities,22 women may have less opportunity than men to manage the potential strain associated with SP in an effective way, which is likely to result in more negative consequences for their mental health. In addition, SP may be more prevalent among women,1,4,13 mostly because the latter are more often employed in sectors, such as health and education, where the act of going to work despite illness tends to occur more frequently.1,4,23 Because previous studies in the field have treated sex as a confounder and not as a potential effect modifier, as yet the existence of sex-related differences in the prospective association between SP and mental health problems (including depression) remains an unexplored issue.
On the basis of a 2-year cohort study conducted on a sample of workers from different workplaces in Denmark, the aim of this study was, therefore, to investigate the impact of SP on the onset of depression, while adjusting for a number of relevant health-related variables and other potential confounders and also taking potential modification effects of sex into account. Possible mechanisms of the causal relationship between SP and depression are discussed in this article.
Study Design and Participants
This study is based on data collected by means of self-reported questionnaires in a two-wave prospective study on employees from several Danish workplaces (see, eg, Hogh et al24 for further details about the study). At baseline (fall of 2006), a total of 90 public and private workplaces were recruited through advertisements on the Web sites of employers' organizations and trade unions. Workplaces with fewer than 25 employees were excluded, leaving a total of 60 organizations (22 private and 38 public) invited to participate at baseline (n = 7358 employees). In the first round of assessment, 3363 participants (45.7% response rate) completed a paper-and-pencil questionnaire about characteristics of the psychosocial work environment and health-related variables. In the majority of cases, questionnaires were sent to the employees' home addresses and returned in sealed envelopes to the authors of this study; 3% of the responders chose to fill in the questionnaire electronically. At follow-up, all baseline responders were approached with a second questionnaire distributed in the fall of 2008. In all, 1664 employees (49.5% of the total baseline responders) took part in this second assessment. At both waves, all employees were informed that participation was voluntary and that all data would be treated confidentially.
From the 1664 participants of the follow-up sample, we excluded 107 responders showing depression at baseline (6.4%). We also excluded 242 participants (19%) with missing values on any of the study variables measured at baseline and 44 cases (2.6%) with missing values on depression at follow-up, leading to a final analytical sample of 1271 participants.
Table 1 shows the distribution of baseline characteristics among responders at baseline (n = 2479) and responders at follow-up (n = 1271), along with the bivariate associations between each baseline characteristic and being a nonparticipant versus a participant at follow-up. Dropout was not significantly related to SP. A higher risk of dropout was observed among men (odds ratio [OR], 1.31; 95% confidence intervals [CI], 1.11 to 1.55), among current smokers (OR, 1.50; 95% CI, 1.23 to 1.82), among participants doing no or light physical activity (OR, 1.23; 95% CI, 1.05 to 1.44), and among employees reporting a higher number of actual working hours per week (OR, 1.01; 95% CI, 1.00 to 1.02, for each additional weekly working hour). Dropout was less likely among older participants (OR, 0.98; 95% CI, 0.97 to 0.98, for each additional year of age) and among those with 3 to 4 years of theoretical/practical education than those with up to 3 years of education (OR, 0.60; 95% CI, 0.49 to 0.74).
TABLE 1-a. Distribut...Image Tools
TABLE 1-b. Distribut...Image Tools
Sickness presenteeism was measured using the following single-item question “How many working days have you gone to work even though you were ill during the last 12 months?” Answers were given using an open-ended fill-in-the-blank response format, with participants asked to report the number of days that they went to work despite illness. We divided SP days into three categories, that is, no SP (0 SP days), 1 to 7 SP days and 8 or more SP days, on the basis of a recent study by Taloyan et al.12
We measured depression by means of the well-established Major Depression Inventory (MDI).25 Major Depression Inventory consists of 10 questions capturing symptoms of depression according to the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) and the International Classification of Diseases, Tenth Revision diagnostic manuals. For each item, participants were asked to rate, on a response format ranging from “at no time” (0 points) to “all of the time” (five points), how frequently they experienced each depressive symptom in the last 2 weeks. The total MDI score may range between 0 and 50 points. To identify cases of depression, we used the previously validated cutoff of 20 or more points, indicating a probable depression.26
We considered a fixed set of possible confounders on the basis of their previously observed association with SP1,4,11,12,27–34 or depression.23,35–42 The selected confounders were subdivided into the following groups and entered hierarchically into the multivariate logistic regression analysis:
1. Background variables: Sex, age (continuous), family status (cohabiting and living alone), education (up to 3 years, 3 to 4 years theoretical/practical education, higher education, and other education), nationality (Danish and immigrant), hourly contract (part-time/full-time), and actual number of working hours per week (continuous).
2. Health-related variables: Self-rated health measured with a single-item question (“In general, would you say your health is: excellent, very good, good, fair, poor”) from the Short Form–36,43 smoking status (smoker vs nonsmoker), alcohol consumption (up to 7 drinks/wk vs more than 7 drinks/wk for women and up to 14 drinks/wk vs more than 14 drinks/wk for men44), physical activity (moderate/hard physical activity vs no/light physical activity), body mass index (less than 18.5, 18.5 to 24.9, 25 or more), and total number of sickness absence days in the last 12 months (continuous).
3. Work-related psychosocial variables: Job insecurity (Cronbach α = 0.78), measured with four items from the Copenhagen Psychosocial Questionnaire45 and subdivided into three categories (low, medium, and high) on the basis of tertiles of the sample distribution; job strain, assessed combining the two Copenhagen Psychosocial Questionnaire's scales for quantitative job demands (four items, Cronbach α = 0.81) and influence at work (four items, Cronbach α = 0.78) into two groups, that is, the high strain group, consisting of participants in the highest tertile of job demands and in the lowest tertile of job control, and the low strain group, including all the other participants.
4. Signs of reduced psychological health: We used the cutoff of more than 9 on the MDI to identify participants with signs of reduced psychological health, indicating preclinical depression, because previous findings demonstrated an increased risk of long-term sickness absence using this cutoff.46
A series of bivariate logistic regression analyses was conducted to examine the prospective association between each baseline factor and incident depression (Table 2). We then performed a hierarchical logistic regression analysis (Table 3) to estimate the prospective association between SP and the incidence of depression in the sample of nondepressed participants at baseline (n = 1271). In model 1, we estimated the crude association between SP and subsequent depression. The previously mentioned blocks of confounders were additionally entered into the following sequence: background variables (model 2), health-related factors (model 3), work-related psychosocial variables (model 4), and signs of reduced psychological health (model 5). To examine whether there are significant differences between women and men in the prospective relationship between SP and depression, we carried out an interaction analysis (model 6) where the interactive term SP*sex was added to all the variables included in the previous model. Regression estimates were reported as OR and their 95% CI. All the analyses were conducted using the SPSS statistical package version 20 (SPSS, Inc, Chicago, IL).
Table 1 shows the baseline characteristics among the follow-up participants (n = 1271) included in this analysis (see column 2). About two-thirds of the participants were women (68.6%), mean age was 46.8 (SD = 9.6) years, and the majority (59%) had 3 to 4 years of theoretical/practical education. Most participants (90.6%) worked full-time and reported an average of 37.5 (SD = 8.7) work hours per week. With regard to employment sector (results not shown in Table 1), the majority of the sample was employed in the public sector (67.7%), including hospitals (25.4%), higher education (11.1%), public administration and services (8.5%), eldercare (7.7%), public schools (6.1%), doctors, dentists, and vets (3.2%), high schools (3.1%), childcare (1.6%), and special schools for mentally disabled children (0.9%). The remaining sample (32.3%) was employed in the private sector, including industries (8.8%), transportation (8.6%), trade unions and associations (6.1%), finance and business service (3.8%), construction (1.8%), agriculture, fishing, and mining (0.8%), storage (0.7%), and unemployment insurance (0.6%). Type of organization was unknown for 15 responders (1.2%). In all, 57.9% of the participants (n = 736) reported at least 1 day of SP in the past year; in more detail, 43.6% (n = 554) and 14.3% (n = 182) reported 1 to 7 days and 8 or more days of SP, respectively.
Table 2 shows the bivariate associations between baseline characteristics and depression at follow-up. The risk of depression was higher among participants who lived alone (OR 1.94, 95% CI, 1.09 to 3.46), reported “fair/poor” general health (OR, 5.22; 95% CI, 2.73 to 9.96), had a body mass index lower than 18.5 (OR, 3.82; 95% CI, 1.07 to 13.65), were in the high tertile of job insecurity (OR, 4.49; 95% CI, 2.38 to 8.46), showed signs of reduced psychological health (OR, 6.04; 95% CI, 3.57 to 10.23), and reported more days of sickness absence in the previous year (OR, 1.01; 95% CI, 1.01 to 1.02, for each additional sickness absence day).
Table 3 shows the results of the hierarchical logistic regression analysis carried out to examine the relationship between SP and incident depression. Among initially nondepressed participants, 62 (4.9%) developed depression at follow-up. The unadjusted association with depression (model 1) was statistically significant for both 1 to 7 days and 8 or more days of SP, as compared with no days of SP (OR, 2.32; 95% CI, 1.17 to 4.61, and OR, 5.99; 95% CI, 2.90 to 12.38, respectively). Adjustment for background variables (model 2) did not affect statistical significance for both 1 to 7 days and 8 or more days of SP (OR, 2.12; 95% CI, 1.05 to 4.25, and OR, 5.48; 95% CI, 2.62 to 11.48, respectively). Further adjustment for health-related variables (model 3) resulted in a reduced but still statistically significant association between 8 or more days of SP and future depression (OR, 3.43; 95% CI, 1.55 to 7.61). Nevertheless, in model 3, the relationship between 1 and 7 days of SP and depression became nonsignificant. The inclusion of adverse work-related psychosocial factors (model 4), while further decreasing the OR for 8 or more days of SP, did not affect its statistical significance (OR, 2.93; 95% CI, 1.30 to 6.65). The association between SP and incident depression also remained significant for 8 or more days of SP in the last step of the analysis (model 5), where we additionally adjusted for signs of reduced psychological health (OR, 2.45; 95% CI, 1.06 to 5.64). The interaction analysis (model 6, results not shown) revealed interactive effects between sex and both 1 to 7 days (P = 0.35) and 8 or more days of SP (P = 0.24) as not significant.
This 2-year prospective study has shown that, in a sample of initially nondepressed employees from different workplaces in Denmark, a self-reported prevalence of 8 or more days of SP in the previous 12 months was prospectively associated with an approximately 2.5-fold increase in the risk of developing depression at follow-up. The relationship between 8 or more days of SP and depression remained significant even after adjusting for several health-related variables and other potential confounders. Nevertheless, we did not find any statistically significant sex-related difference in the prospective association between SP and depression.
This study is the first to provide direct evidence of a significant impact of SP on future depression. This relationship was previously suggested by three prospective studies. In a representative sample of the Swedish workforce, Gustafsson and Marklund11 found that more than five self-reported episodes of SP in a year were significantly related to subsequent poor well-being as assessed with a general 10-item questionnaire that also included symptoms of depression. Demerouti et al10 observed, in a sample of staff nurses, a significant association between SP and subsequent emotional exhaustion, a factor potentially related to future depression.47 In another representative sample of the Swedish workforce, Taloyan et al12 found that emotional exhaustion fully mediated the prospective relationship between SP and two different outcomes, that is, self-rated health and sickness absence.
In this study, the relationship between SP and subsequent depression was attenuated but not affected, in terms of statistical significance, by the introduction of a number of potential confounders. In particular, by adjusting for several health-related factors at baseline (eg, self-rated health and sickness absence), we were able to isolate the effects produced by the exposure (ie, SP) from the effects exerted by the health problems themselves, which are, by definition, present in the event of SP.1 The OR for the relationship between high SP and depression remained significant also when signs of reduced psychological health were introduced as confounder. Because depression often presents a long preclinical insidious stage,48 employees with reduced psychological health at baseline might have been at a higher risk of both reporting SP and being a new case of depression at follow-up. As argued by Johns,3 employees with symptoms of poor psychological health may, indeed, see their problem as an illegitimate reason to stay at home, thus being at a higher risk of SP. With the inclusion of signs of reduced psychological health among the confounders, we could, thus, exclude a possible overestimation of the association between SP and future depression.
Possible Mechanisms of the Association Between SP and Depression
A first possible mechanism behind the relationship between SP and depression is grounded in the Effort-Recovery49 and the Allostatic Load50 models. Because working while ill is often accompanied by a reduction in work ability,3 the sick employee has to put in extra efforts to deal effectively with job demands and preserve his or her habitual performance levels.10 Nevertheless, a high effort expenditure may lead to increased stress levels, which in turn, according the Effort-Recovery model, may affect the individual capacity to unwind after work.51 The high levels of fatigue resulting from inadequate recovery may thus force the sick employee to further raise his or her efforts to meet job requirements, additionally burdening the recuperation process. Consistent with the Allostatic Load model, the prolonged psychophysiological overactivation produced by such vicious circle may ultimately lead to a corrupted functioning of different bodily systems that regulate human adaptation to external stimuli. In particular, an alteration of physiological stress responses may reflect the presence of allostatic load in psychiatric disorders, including depression.19,20 This health impairing process may be further exacerbated by the prolongation and/or aggravation of the health problems that result from SP–induced inadequate recovery.
It can be further expected that inadequate recovery and the consequent psychophysiological overactivation caused by SP are more severe when employees are exposed to a poor psychosocial work environment, increasing the risk of subsequent depression. Adverse work characteristics (such as high time pressure and poor influence at work) may, in fact, limit an employee's possibility to adjust efforts to his or her current health status, possibly resulting in higher fatigue and increased psychophysiological burden. In this study, we controlled for two important aspects of the psychosocial work environment, that is, job strain and job insecurity, for their previously observed association with both SP30,34 and depression.39,52 Nevertheless, future studies on larger samples should also examine the role of a range of work-related psychosocial factors as potential modifiers of the relationship between SP and depression.
A second possible explanation is that working while ill reflects a component of a general tendency of employees engaging in a lifestyle characterized by at-risk health behaviors. There is increasing evidence (see, eg, Lopresti et al53) that the adoption of risky lifestyles such as poor sleep, low physical activity, and poor diet is associated with the incidence of major depression through dysregulated physiological pathways.
A third and final mechanism could be that SP, being often associated with reduced performance,6,7 may in the long run jeopardize the quality of the relationships that the sick employee has with his or her supervisors and peers, ultimately leading to adverse psychological outcomes.54 In contrast to this explanation, however, there might also be circumstances where working while ill, even if linked with reduced performance, is seen with favor by coworkers, thereby improving rather than impoverishing one's working conditions. This may occur, for instance, in cases where employees are difficult to replace (eg, in knowledge-intensive jobs55) or in teamwork-based settings where a worker's sick leave may cause problems to colleagues.56 In such occasions, the negative effect of SP on psychological health can be compensated by the positive effect that this behavior may have on one's social work environment. To shed more light over these contrasting hypotheses, future research is needed, examining the possible modification effect of type of occupation and other organizational factors on the relationship between SP and depression.
Contrary to our expectations, the prospective association between SP and future depression did not result to be stronger among women than among men. We expected such a modification effect of sex mainly for two reasons. First, the prevalence of depression is known to be higher among women.21 Second, women could be expected to experience more negative health consequences as a result of SP because they may have less opportunity to effectively cope with the strain induced by working during illness. For instance, given their typically higher demands for fulfilling family obligations,22 women showing up for work despite sick may have less room for recuperation than men, which adds to the lack of recovery already present as a consequence of SP. Furthermore, there are indications that SP is more prevalent among women,1,4,13 mostly because the latter are more likely employed in sectors where SP tends to occur more frequently, such as health care, social work, and primary/secondary education.1,4,23 This is supported in our data, where there is a clear clustering of women in the previously mentioned occupational groups, which are massively concentrated in the public sector (data not shown).
Whereas our results did not support the hypothesized sex specificity in the relationship between SP and future depression, it must be pointed out that our reduced sample size might have played an important role by not providing sufficient power for detecting interactions that are, otherwise, present in the population. Also, in light of a lack of research investigating sex as a possible effect modifier of the association between SP and future mental health problems, we, therefore, recommend further studies on the basis of larger sample sizes to be carried out to investigate whether the prospective relationship between SP and depression differs between women and men.
While carrying out research on the health impact of SP, a crucial methodological issue is to isolate the unique effects on health produced by SP from the effects of the poor health status that is intrinsically associated with this behavior. To address this, a common procedure, which we observed in this study, was to introduce several health-related variables as potential confounders, allowing to compare rates of incident depression between groups exposed and unexposed to SP while adjusting for baseline differences in health status.9 Nevertheless, residual confounding because of health factors that we did not adjust for in our analysis may still operate influencing results of this and other studies. With regard to this study, other aspects of health associated with SP but not captured by the variables considered may have, indeed, contributed to the onset of depression. For instance, self-rated health might not be able to catch minor or isolated episodes of illness among, otherwise, healthy people.9 If these minor episodes were related to both the exposure and the outcome in this study, failing to control for them could have resulted in overestimating the relationship between SP and subsequent depression. In future research, it might be, therefore, important to evaluate on a more detailed level the type of illnesses (both minor and major) associated with SP to better account for the possible confounding effect of health-related factors.
Finally, controlling for signs of reduced psychological health could have introduced overadjustment bias. Subclinical depression may, indeed, serve as an intermediate variable, and not as a confounder, between SP and subsequent depression. Some participants may have, in fact, experienced subclinical psychological problems because of SP during the previous year, while developing clinically relevant depression only at a later stage. If this was true, the consequence would be an underestimation of the true effect of SP on future depression.
Study Strengths and Limitations
The major strengths of this study were the follow-up design, the inclusion of several potential confounders (including a number of health-related factors), and the use of the well-validated MDI instrument to assess depression.
Nevertheless, several limitations should also be taken into account while interpreting the results of this study. First, external validity may be questioned because we recruited workplaces through advertisements on the Web sites of trade unions and employers' organizations, implying that our results may not be representative of the Danish workforce. Furthermore, the low response rate at baseline (47.5%) may have introduced a selection bias that further affected generalizability. Nevertheless, strong selection seems unlikely because the point prevalence of depression (on the basis of the MDI cutoff of 20 or more) observed in this study (8.4% in the baseline sample) was near to the prevalence observed in the Danish working population (7.1%).26 With regard to SP, the point prevalence (57.9% of participants with 1 or more days of SP) was similar to that observed in the Fifth European Working Condition Survey, where more than 50% of workers in Denmark reported at least 1 day of SP in the previous 12 months.2
A second limitation is related to the considerable dropout of participants at follow-up (50.5%). Nevertheless, we found only small differences between the baseline and the follow-up samples in the distribution of most study variables.
A third limitation is the time lag between the two measurement points. In an interval of 2 years, it could be that some participants have developed clinically relevant depression during follow-up, but that they were in remission when the second round of measurement took place. If this has occurred and SP was significantly related to the incidence of new cases of depression being in remission at follow-up, the consequence would be an underestimation of the examined relationship.
Fourth, self-reported assessment of SP could be subjected to recall bias, resulting in a possible misclassification of exposure. Using the previous 12 months as time frame of response (as common in this line of research) instead of a shorter recall interval (eg, 6 months) could have additionally contributed to inaccuracy in the measurement of SP. Nevertheless, an earlier study has shown a high 1-year test-retest reliability of an SP item resembling our measure.10
Fifth, the reduced analytical power due to the limited size of our sample may have contributed, along with the low number of depressed cases observed at follow-up, to the wide 95% CIs of the regression estimates. In particular, the fact that in the fully adjusted regression model the lower 95% CI was close to the unit might signal possible residual confounding. For instance, we were not able to control for income, a factor playing a potential role in the prediction of both SP and depression.3,29,57 Nevertheless, income is strongly related with education, a factor that, in this study, did not act as a significant confounder. Nonetheless, because the connection between education and income can be expected to be more substantial among men than among women, we are aware that income may have been adjusted for only partially in this study.
Although, in the last decade, research interest into the phenomenon of SP has increased, the vast majority of studies on employee attendance behaviors are still focused on investigating risk factors and consequences of sickness absence.13 Similarly, company policies are almost exclusively aimed at identifying ways to decrease employee sickness absence because it is believed that the latter accounts for almost all the costs related to illness and reduced on-the-job productivity.58 Nevertheless, recent evidence indicates that SP is a behavior deserving parallel attention for the serious implications it may have for both employee health and job performance. Indeed, as suggested in two recent studies,55,59 taking sickness absence as exclusive indicator of employee health may be insufficient or even misleading. For instance, there may be circumstances in which employees substitute sickness absence for SP, as during events inducing job insecurity (frequent in today's workplaces) such as downsizing.59 Two main implications follow from this reflection: (1) being present at work is not always reflective of good health; and (2) the calculation of the costs because of lost productivity should also account for the impact of performance cutbacks that are likely to occur among employees showing up for work despite sick. Therefore, from both the public health and the employers' perspectives, the total costs related to employees' attendance behaviors seem being better reflected by a combined measure including both sickness absence and SP as measures of health.55 In this respect, the indication provided by this study that SP significantly contributes to the development of depression may be particularly informative because the latter is among the illnesses with the highest economic impact, because of its high prevalence and comorbidity with other conditions.5–7
This study indicates that depression should be enumerated among the possible health consequences of SP, although more studies are needed to replicate our findings and clarify both underlying causal mechanisms and the role of possible effect modifiers. The growing evidence that SP has negative consequences on both physical and mental health should warn researchers and employers about the need to broaden their view concerning the possible links between employee attendance behaviors, health, and work performance.
1. Aronsson G, Gustafsson K. Sickness presenteeism: prevalence, attendance pressure factors, and an outline of a model for research. J Occup Environ Med. 2005;47:958–966.
2. Eurofound. Health and Well-being at Work: A Report Based on the Fifth European Working Conditions Survey. Dublin, Ireland: Eurofound; 2012:55–57.
3. Johns G. Presenteeism in the workplace: a review and research agenda. J Organ Behav. 2010;31:519–542.
4. Aronsson G, Gustafsson K, Dallner M. Sick but yet at work: an empirical study of sickness presenteeism. J Epidemiol Community Health. 2000;54:502–509.
5. Stewart WF, Ricci JA, Chee E, Hahn SR, Morganstein D. Cost of lost productive work time among US workers with depression. JAMA. 2003;289:3135–3144.
6. Burton W, Pransky GS, Conti D, Chen C, Edington D. The association of medical conditions and presenteeism. J Occup Environ Med. 2004;46:38–45.
7. Schultz AB, Chen CY, Edington DW. The cost and impact of health conditions on presenteeism to employers: a review of the literature. Pharmacoeconomics. 2009;27:365–378.
8. Kivimäki M, Head J, Ferrie JE, et al. Working while ill as a risk factor for serious coronary events: the Whitehall II study. Am J Public Health. 2005;95:98–102.
9. Bergström G, Bodin L, Hagberg J, Lindh T, Aronsson G, Josephson M. Does sickness presenteeism have an impact on future general health? Int Arch Occup Environ Health. 2009;82:1179–1190.
10. Demerouti E, Le Blanc PM, Bakker AB, Schaufeli WB, Hox J. Present but sick: a three-wave study on job demands, presenteeism and burnout. Career Dev Int. 2009;14:50–68.
11. Gustafsson K, Marklund S. Consequences of sickness presence and sickness absence on health and work ability: a Swedish prospective cohort study. Int J Occup Med Environ Health. 2011;24:153–165.
12. Taloyan M, Aronsson G, Leineweber C, Magnusson Hanson L, Alexanderson K, Westerlund H. Sickness presenteeism predicts suboptimal self-rated health and sickness absence: a nationally representative study of the Swedish working population. PLoS ONE. 2012;7:e44721.
13. Hansen CD, Andersen JH. Sick at work a risk factor for long-term sickness absence at a later date? J Epidemiol Community Health. 2009;63:397–402.
14. Bergström G, Bodin L, Hagberg J, Aronsson G, Josephson M. Sickness presenteeism today, sickness absence tomorrow? A prospective study on sickness presenteeism and future sickness absenteeism. J Occup Environ Med. 2009;51:629–638.
15. Janssens H, Clays E, De Clercq B, De Bacquer D, Braeckman L. The relation between presenteeism and different types of future sickness absence. J Occup Health. 2013;55:132–141.
16. Wang JL, Schmitz N, Smailes E, Sareen J, Patten SB. Workplace characteristics, depression and health-related presenteeism in a general population sample. J Occup Environ Med. 2010;52:836–842.
17. World Health Organization. The Global Burden of Disease: 2004 Update. Geneva, Switzerland: World Health Organization; 2008.
18. McTernan WP, Dollard MF, LaMontagne AD. Depression in the workplace: an economic cost analysis of depression-related productivity loss attributable to job strain and bullying. Work Stress. 2013;27:321–338.
19. McEwen BS. Protection and damage from acute and chronic stress: allostasis and allostatic overload and relevance to the pathophysiology of psychiatric disorders. Ann N Y Acad Sci. 2004;1032:1–7.
20. McEwen BS. Protective and damaging effects of stress mediators: central role of the brain. Dialogues Clin Neurosci. 2006;8:367–381.
21. Ustun TB. Cross-national epidemiology of depression and gender. J Gend Specif Med. 2000;3:54–58.
22. Grönlund A, Öun I. Rethinking work–family conflict: dual earner policies, role conflict and role expansion in Western Europe. J Eur Soc Policy. 2010;20:179–195.
23. Elstad JI, Vabø M. Job stress, sickness absence and sickness presenteeism in Nordic elderly care. Scand J Public Health. 2008;36:467–474.
24. Hogh A, Hansen AM, Mikkelsen EG, Persson R. Exposure to negative acts at work, psychological stress reactions and physiological stress response. J Psychosom Res. 2012;73:47–52.
25. Olsen LR, Jensen DV, Noerholm V, Martiny K, Bech P. The internal and external validity of the Major Depression Inventory in measuring severity of depressive states. Psychol Med. 2003;33:351–356.
26. Olsen LR, Mortensen EL, Bech P. Prevalence of major depression and stress indicators in the Danish general population. Acta Psychiatr Scand. 2004;109:96–103.
27. Johansson G, Lundberg I. Adjustment latitude and attendance requirements as determinants of sickness absence or attendance. Empirical tests of the illness flexibility model. Soc Sci Med. 2004;58:1857–1868.
28. Hansen CD, Andersen JH. Going ill to work—what personal circumstances, attitudes and work-related factors are associated with sickness presenteeism? Soc Sci Med. 2008;67:956–964.
29. Agudelo-Suarez AA, Benavides FG, Felt E, Ronda-Perez E, Vives-Cases C, Garcıa AM. Sickness presenteeism in Spanish-born and immigrant workers in Spain. BMC Public Health. 2010;10:791.
30. Heponiemi T, Elovainio M, Pentti J, et al. Association of contractual and subjective job insecurity with sickness presenteeism among public sector employees. J Occup Environ Med. 2010;52:830–835.
31. Leineweber C, Westerlund H, Hagberg J, Svedberg P, Luokkala M, Alexanderson K. Sickness presenteeism among Swedish police officers. J Occup Rehabil. 2011;21:17–22.
32. Leineweber C, Westerlund H, Hagberg J, Svedberg P, Alexanderson K. Sickness presenteeism is more than an alternative to sickness absence: results from the population-based SLOSH study. Int Arch Occup Environ Health. 2012;85:905–914.
33. Janssens H, Clays E, Kittel F, De Bacquer D, Casini A, Braeckman L. The association between body mass index class, sickness absence, and presenteeism. J Occup Environ Med. 2012;54:604–609.
34. Jourdain G, Vézina M. How psychological stress in the workplace influences presenteeism propensity: a test of the Demand-Control-Support model. Eur J Work Organ Psychol. 2013. Available at: http://dx.doi.org/10.1080/1359432X.2012.754573
. Accessed October 18, 2013.
35. Bildt C, Michelsen H. Gender differences in the effects from working conditions on mental health: a 4-year follow-up. Int Arch Occup Environ Health. 2002;75:252–258.
36. Wieclaw J, Agerbo E, Mortensen PB, Bonde JP. Occupational risk of affective and stress-related disorders in the Danish workforce. Scand J Work Environ Health. 2005;31:343–351.
37. Hasin DS, Goodwin RD, Stinson FS, Grant BF. Epidemiology of major depressive disorder: results from the National Epidemiologic Survey on Alcoholism and Related Conditions. Arch Gen Psychiatry. 2005;62:1097–1106.
38. Rugulies R, Bultmann U, Aust B, Burr H. Psychosocial work environment and incidence of severe depressive symptoms: prospective findings from a 5-year follow-up of the Danish work environment cohort study. Am J Epidemiol. 2006;163:877–887.
39. Bonde JP. Psychosocial factors at work and risk of depression: a systematic review of the epidemiological evidence. Occup Environ Med. 2008;65:438–445.
40. Andersen I, Thielen K, Nygaard E, Diderichsen F. Social inequality in the prevalence of depressive disorders. J Epidemiol Community Health. 2009;63:575–581.
41. Boden JM, Fergusson DM, Horwood LJ. Cigarette smoking and depression: tests of causal linkages using a longitudinal birth cohort. Br J Psychiatry. 2010;196:440–446.
42. Amagasa T, Nakayama T. Relationship between long working hours and depression: a 3-year longitudinal study of clerical workers. J Occup Environ Med. 2013;55:863–872.
43. Ware JE, Snow KK, Snow K, Kosinski M, Gandek B. SF-36® Health Survey Manual and Interpretation Guide. Boston, MA: New England Medical Center, The Health Institute; 1993.
45. Kristensen TS, Hannerz H, Hogh A, Borg V. The Copenhagen Psychosocial Questionnaire—a tool for the assessment and improvement of the psychosocial work environment. Scand J Work Environ Health. 2005;31:438–449.
46. Hjarsbech PU, Andersen RV, Christensen KB, Aust B, Borg V, Rugulies R. Clinical and non-clinical depressive symptoms and risk of long-term sickness absence among female employees in the Danish eldercare sector. J Affect Disord. 2011;129:87–93.
47. Hakanen JJ, Schaufeli WB. Do burnout and work engagement predict depressive symptoms and life satisfaction? A three-wave seven-year prospective study. J Affect Disord. 2012;141:415–424.
48. Kolstad HA, Hansen AM, Kaergaard A, et al. Job strain and the risk of depression: is reporting biased? Am J Epidemiol. 2011;173:94–102.
49. Meijman TF, Mulder G. Psychological aspects of workload. In: Drenth PJD, Thierry H, eds. Handbook of Work and Organizational Psychology. Vol. 2: Work Psychology. Hove, England: Psychology Press/Erlbaum; 1998:5–33.
50. McEwen BS. Protective and damaging effects of stress mediators. N Engl J Med. 1998;338:171–179.
51. Geurts SAE, Sonnentag S. Recovery as an explanatory mechanism in the relation between acute stress reactions and chronic health impairment. Scand J Work Environ Health. 2006;32:482–492.
52. Netterstrøm B, Conrad N, Bech P, et al. The relation between work-related psychosocial factors and the development of depression. Epidemiol Rev. 2008;30:118–132.
53. Lopresti AL, Hood SD, Drummond PD. A review of lifestyle factors that contribute to important pathways associated with major depression: diet, sleep and exercise. J Affect Disord. 2013;148:12–27.
54. Dormann C, Zapf D. Social stressors at work, irritation, and depressive symptoms: accounting for unmeasured third variables in a multi-wave study. J Occup Organ Psychol. 2002;75:33–58.
55. Aronsson G, Gustafsson K, Mellner C. Sickness presence, sickness absence, and self-reported health and symptoms. Int J Workplace H Manage. 2011;4:228–243.
56. Grinyer A, Singleton V. Sickness absence as risk-taking behaviour: a study of organizational and cultural factors in the public sector. Health Risk Soc. 2000;2:7–21.
57. Lorant V, Deliege D, Eaton W, Robert A, Philippot P, Ansseau M. Socioeconomic inequalities in depression: a meta-analysis. Am J Epidemiol. 2003;157:98–112.
58. Biron C, Brun J, Ivers H, Cooper CL. At work but ill: psychosocial work environment and well-being determinants of presenteeism propensity. J Public Ment Health. 2006;4:26–37.
59. Caverley N, Cunningham JB, MacGregor JN. Sickness presenteeism, sickness absenteeism, and health following restructuring in a public service organization. J Manag Stud. 2007;44:304–319.
Copyright © 2014 by the American College of Occupational and Environmental Medicine