Abstract

This review is based on a literature search made in January 2007 on request by the Danish National Board of Industrial Injuries. The search in PubMed, EMBASE, and PsycINFO resulted in more than 1,000 publications. This was reduced to 14 after the titles, abstracts, and papers were evaluated by using the following criteria: 1) a longitudinal study, 2) exposure to work-related psychosocial factors, 3) the outcome a measure of depression, 4) relevant statistical estimates, and 5) nonduplicated publication. Of the 14 studies, seven used standardized diagnostic instruments as measures of depression, whereas the other seven studies used self-administered questionnaires. The authors found moderate evidence for a relation between the psychological demands of the job and the development of depression, with relative risks of approximately 2.0. However, indication of publication bias weakens the evidence. Social support at work was associated with a decrease in risk for future depression, as all four studies dealing with this exposure showed associations with relative risks of about 0.6. Even if this literature study has identified work-related psychosocial factors that in high-quality epidemiologic studies predict depression, studies are still needed that assess in more detail the duration and intensity of exposure necessary for developing depression.

INTRODUCTION

Depression is a frequent mental disorder with severe consequences. The 1-year prevalence of major depression has been reported to be between 2 and 5 percent in several studies; however, prevalence rates as low as 0.6 percent and as high as 10.3 percent have also been found (1). Regarding lifetime prevalence, a recent review of 16 epidemiologic studies reported rates in the range of 3.0–24.4 percent, with a mean lifetime prevalence of 11.7 percent (2). It is unclear to what extent the substantial variation in both 1-year and lifetime prevalence of depression across studies reflects true differences in prevalence, cultural differences in reporting, or differences in the instruments that have been used to measure major depression (1). Milder forms of depression, also referred to as “minor depression” or “subthreshold depressive symptoms,” have been researched less but are thought to have even higher prevalence (3).

According to the World Health Organization, depression is one of the most disabling of disorders, ranking fifth place on the list of disorders with the highest disability-adjusted life-years score (4), and it is estimated that depression will move up to second place by the year 2020. It has been shown that the risk for days lost from work was approximately five times higher for patients with major depression compared with nondepressive persons, and even for patients with minor depression the risk was approximately 1.6 times higher (5). This association between severity of depression and days lost from work has been confirmed recently (6), and the impact of comorbidity with chronic medical conditions on work loss or quality of life is also reported (3, 7, 8).

The exact causes for the development of depression are still not known. It is widely assumed that the etiology is multifactorial involving genetic, biologic, and psychosocial factors (9). Studies have consistently shown that women have a higher risk for depression than men do, although the gender ratios differed considerably among the studies (1). Low socioeconomic status has also been identified as a risk factor for major depression in a recent meta-analysis (10). Family and twin studies suggest a genetic component in the etiology of major depression; however, the specific genes or gene-environment interactions have not been identified yet (11). There is also evidence that both anxiety disorders and minor depression are important risk factors for the onset of a major depression (3, 12).

Regarding psychosocial factors, research has focused mainly on adverse life events and lack of social support (9). Among life events, both the experience of severe physical diseases, such as myocardial infarction (13) or cancer (14), and the experience of loss, such as the death of a significant person, separation, or divorce (9), have predicted the risk of major depression.

A possible effect of exposure to adverse psychosocial working conditions on the risk of common mental disorders has been discussed for some time; however, most studies have been cross-sectional (15). Only recently have studies used longitudinal study designs in this research area (16). A review and meta-analysis by Stansfeld and Candy (16) on the prospective association between psychosocial working conditions and the risk of common mental disorders identified 38 studies, of which 11 were suitable for a meta-analysis. However, Stansfeld and Candy did not conduct subanalyses on the effect of psychosocial working conditions on specific common mental disorders, such as depression.

The present review is, on the one hand, an update of the review by Stansfeld and Candy (16) including several studies published after 2005 and, on the other hand, a specific focus on the effect of psychosocial working conditions on the risk of depression. The review is based on a report for the Danish Working Environment Research Fund on the relation between work-related stressors and the development of mental disorders other than post-traumatic stress disorder, which we finalized in September 2007. The report was requested by the Danish National Board of Industrial Injuries and the associated committee in connection with providing scientific evidence for the ongoing negotiations as to which disorders should be included in the Danish directory of occupational diseases.

EXPOSURE

For the purposes of this study, work-related psychosocial stress factors are defined as aspects of the job (i.e., work content, organization, relations, and so on) that can lead to a stress condition characterized by symptoms or impaired functioning and ill health. Physical as well as chemical, biologic, and psychological factors can lead to stress, but in the present context the focus is entirely on the so-called psychosocial stressors. Scientific studies that have examined the association between psychosocial stressors and disorder have generally used three different measures of exposure. 1) Stressors can be evaluated objectively (e.g., administrative data as number of working hours) (17) or by observation (18). 2) Stressors can be self-reported, measured by standardized questionnaires based on different models, such as the Job Strain Model of Karasek et al. (19). 3) Some studies use the so-called ecologic method, where the extent and type of psychosocial work-related factors are based on reports from people with particular types of jobs in order to avoid the common method bias, when both exposure and outcome are measured by self-report by the individuals (20). This method often implies the use of an environmental exposure matrix, where jobs are classified according to the degree of exposure.

An objective description of the exposure has been used in only a very few studies, as psychological stressors are either very difficult to measure or because documentation is unavailable (21, 22). On the other hand, it is fairly straightforward in scientific studies to ask people how much or how frequently they are exposed to different stressors in their work (23, 24). In this context, it is important to emphasize that some self-reported measures assess characteristics of the workplace (i.e., potential stressors), whereas others assess how people feel when they are exposed to stressors (i.e., stress reactions, reduced well-being). For the purpose of this review, we selected studies that assessed the characteristics of the workplace, because using stress reactions as a predictor variable when the outcome (depression) is thought to be a kind of stress reaction would constitute a “triviality trap.”

Psychological stressors in the work environment are highly varied and can be very different depending on the type of job. In order to compare the results of different types of studies, we used many studies of models for psychological strain that operationalize the most important stressors. Examples are the Job Strain Model developed by Karasek et al. (19) and the Effort-reward Imbalance Model developed by Siegrist (25). Both of these models have been used in a number of studies of health and psychological strain at work. The first model has been particularly dominant within occupational health research during the last 20 years.

The Job Strain Model uses two main dimensions: demands and decision latitude. The decision latitude dimension consists of two subdimensions: decision authority and skill discretion. By combining the two main dimensions, we present four stress conditions: Persons who experience high demands and have a high decision latitude are termed “active,” the combination of high demands and low decision latitude is termed “strained,” that of low demands and high decision latitude is termed “relaxed,” and the combination of both low demands and low decision latitude is termed “passive.” According to the model, people in strained jobs bear the highest risk for developing stress-related disorders.

Social support at work has been shown to modify the strain that might lead to stress, and in some studies social support is therefore used in combination with the job-strain model in a so-called isostrain model (26). In this context, it is decisive whether or not the social network provides real support in the handling of psychosocial strain. Most studies have used two measures of social support, one from coworkers and one from supervisors (19). However, social interactions at work can also be negative, and measures of conflicts and bullying at the workplace represent these more negative aspects (27).

In addition to the Job Strain Model, Siegrist et al. (28) have developed a stress model, called the Effort-reward Imbalance Model, that is based on the individual experience of the balance between the effort made and the reward received. According to this model, the most stressful condition is when the reward does not match the effort made. Reward includes financial rewards, esteem rewards, promotion prospects, and job security. Moreover, it is assumed in this model that people with a personal disposition called “work-related overcommitment” are at especially increased risk for poor health, when exposed to high effort and low reward.

OUTCOME

So far, the association between work-related psychosocial factors and depression with a clinical diagnosis by a physician or psychiatrist has been described in only a very few studies. The majority of scientific studies that have looked into the associations between psychosocial work-related factors and mental health have used different forms of more or less validated outcome measures, with several studies using general measures of distress (16). To a certain extent, this makes sense, especially in the light of recent considerations of group depression as a distress disorder in the forthcoming Diagnostic and Statistical Manual of Mental Disorders: DSM-V (29). However, for this review, we excluded studies that used measures of distress and included only studies that assessed depression by 1) a clinical diagnosis by a physician, 2) a validated diagnostic interview, conducted by a trained interviewer, or 3) a validated rating scale for depression.

LITERATURE SEARCH

The literature search was carried out by using the databases PubMed (1960-), EMBASE (1980-), and PsycINFO (1967-) up until January 28, 2007. Only longitudinal studies and studies published in English in peer-reviewed journals were included. The search words are listed in table 1. There were a total of 3,416 studies (2,291 in PubMed, 659 in EMBASE, and 466 in PsycINFO, some of which were duplicates) (figure 1). These were all checked and categorized according to title by one of the authors. A selection was made so that only longitudinal studies on work-related issues and mental disorder were considered. The result was 319 articles, the abstracts of which were read by two authors. After the exclusion of irrelevant articles following the same criteria, the number of relevant articles was 81. All of these were read in order to find the relevant exposures and outcome. The bibliographies of all selected articles, as well as the personal files and reference lists of the authors, were checked to identify further studies that might fulfill the eligibility criteria. In addition, the meta-analysis from Stansfeld and Candy (16) was searched for further qualifying studies.

FIGURE 1.

Search strategy on the relation between work-related psychosocial factors and depression.

FIGURE 1.

Search strategy on the relation between work-related psychosocial factors and depression.

TABLE 1.

Search terms in the literature search

Exposures Outcomes 
Workload Psychiatric 
Work conditions Psychiatric disorders 
Job conditions Mental 
Working hours Mental health 
Working time Substance use 
Night work Abuse 
Shift work Drug 
Stress* Alcohol 
 Benzodiazepine 
Psychosocial work environment Psychosis 
Effort reward Psychotic 
Emotional demands Paranoia 
Iso-strain† Paranoid disorder 
Job strain Mood disorders 
Job security or job insecurity Affective disorders 
Job control Bipolar 
Justice Depressive 
Meaning of work Depression 
Predictability of work Anxiety 
Psychosocial demands Anxiety disorder 
Bullying Generalized anxiety 
Mobbing Panic disorder 
Teasing Obsessive-compulsive disorder 
Exposures Outcomes 
Workload Psychiatric 
Work conditions Psychiatric disorders 
Job conditions Mental 
Working hours Mental health 
Working time Substance use 
Night work Abuse 
Shift work Drug 
Stress* Alcohol 
 Benzodiazepine 
Psychosocial work environment Psychosis 
Effort reward Psychotic 
Emotional demands Paranoia 
Iso-strain† Paranoid disorder 
Job strain Mood disorders 
Job security or job insecurity Affective disorders 
Job control Bipolar 
Justice Depressive 
Meaning of work Depression 
Predictability of work Anxiety 
Psychosocial demands Anxiety disorder 
Bullying Generalized anxiety 
Mobbing Panic disorder 
Teasing Obsessive-compulsive disorder 
*

Stress AND (job OR work OR occupation).

Job strain plus exposure to low social support at work (“isolated high strain” work).

The selection for inclusion in the review was made according to the following criteria:

  • The study should be longitudinal so that the outcome has been measured after the exposure.

  • Exposure should be work-related psychosocial factors.

  • Outcome should be a measure of depression.

  • The data analysis should be prospective.

  • There should be relevant statistical estimates of the associations tested.

  • The study population should be greater than 100.

  • The data from the study should not be published elsewhere. In the case of duplicate publication or publication of data from the same study with the same exposure measure or outcome, the paper with the most relevant follow-up period, analysis, and risk estimate was chosen.

RESULTS

Table 2 provides an overview of the 14 studies included in this review. In seven studies, the assessment of depression was based on psychiatric diagnoses, scales with diagnostic classifications, or antidepressant prescription, whereas in the seven other studies, depression was measured with a questionnaire.

TABLE 2.

Longitudinal studies on the relation between work-related psychosocial factors and depression

Author(s), year, and reference Population Exposure measures Follow-up time, years Outcome Confounder adjustments Analysis Risk of depression 
Men Women Both genders 
Wieclaw et al., 2006 (20Among all employed Danes, 14,166 cases and 58,060 controls Occupation with exposure to threats and violence Depression diagnosed in psychiatric hospital Age, sociodemographics Baseline not clean† Violence: RR* = 1.5 (95% CI*: 1.2, 1.9) Violence: RR* = 1.5 (95% CI:* 1.3, 1.7)  
Threats: RR = 1.2 (95% CI: 0.9, 1.5) Threats: RR = 1.5 (95% CI: 1.3, 1.8)  
Plaisier et al., 2007 (301,529 employed men and 1,117 women JCQ* social support CIDI* by trained interviewer Age, gender, health, education Clean baseline; job strain not calculated  Risk for depression for women: RR = 1.8 compared with men Demands: RR = 3.5 (95% CI: 1.9, 6.3) 
 Social support: RR = 0.8 (95% CI: 0.7, 0.9) 
 Decision latitude and job insecurity: NS* 
Shields, 2006 (3112,011 employed Canadians aged 18–75 years JCQ measured coworker and supervisor support twice. 8 for job strain, and 2 for support CIDI in 2002 as a questionnaire Age, gender, marital status, education, occupation, lifestyle factors Clean baseline; detailed information on loss of follow-up Low coworker support: RR = 2.4 (95% CI: 1.7, 3.3) Low coworker support: RR = 1.8 (95% CI: 1.4, 2.4) Job strain in 1994 and 2000: RR = 3.4 (95% CI: 1.8, 6.4) 
Low supervisor support: RR = 1.7 (95% CI: 1.0, 2.7) Low supervisor support: RR = 1.6 (95% CI: 1.2, 2.2) No job strain in 1994 but in 2000: RR = 3.3 (95% CI: 1.8, 6.1) 
Virtanen et al., 2007 (321,704 women and 1,662 men all employed JCQ Antidepressant prescription; data from public register Age, marital status, occupational grade, mental disorder at baseline  Job strain: RR = 2.0 (95% CI: 1.0, 3.8) Job strain: RR = 1.2 (95% CI: 0.7, 2.0)  
Kawakami et al., 1990 (333,045 male workers in an electronics plant Job stress variables DSM-III* by psychiatrist Partly by comparing confounders Clean baseline; adequate analyses “Unsuitable job”: RR = 11.3 (95% CI: 2.0, 61.8)   
Job overload, overtime, human relations: NS   
Shields, 1999 (341,649 women and 2,181 men aged 25–54 years and employed ≥35 hours per week Weekly working hours self-reported CIDI in 2002 as a questionnaire Age, socioeconomics, education, income, occupation, self-employment No clean baseline ≥41 vs. 35–40 hours/week: RR = 0.6 (95% CI: 0.3, 1.3) ≥41 vs. 35–40 hours/week: RR = 2.2 (95% CI: 1.1, 4.4)  
Wang, 2005 (356,099 employed Canadians Work stress: 12 questions on skill discretion, decision authority, psychological demands, job insecurity, physical exertion, and social support Major depressive episode during last year by CIDI–Short Form for Major Depression Age, gender, marital status, income, race, education, medical illness, subsequent mental health service use Clean baseline   RR = 2.4 (95% CI: 1.5, 3.8) 
Paterniti et al., 2002 (362,790 women and 7,729 men working in the French National Electricity and Gas Company Demands, decision latitude (JCQ), social support CES-D* Age, education, marital status, stressful personal events, baseline CES-D score Adjusted for baseline illness p values of regression coefficients Regression coefficient p values  
    Demands: <0.001     Demands: <0.001  
    Decision latitude: <0.01     Decision latitude: NS  
    Social support: <0.01     Social support: <0.05  
de Lange et al., 2002 (38824 employed persons JCQ changes over time 1 × 4 CES-D Age, gender, education Changes in exposure and outcome   Increased job strain associated with depression 
Schonfeld, 2001 (39184 female teachers Episodic stressors (threats, confrontations) 3 × 4 CES-D Age, socioeconomics, race, marital status No clean baseline; no risk estimate  Regression coefficient p value: <0.01  
Moore et al., 2004 (401,235 employees Downsizing: layoffs CES-D Age, gender, education No clean baseline; no risk estimate   Prevalence of depression 5.2% among never layoffs; 10.4% among those with two layoffs (p < 0.001) 
Fullerton et al., 2004 (41116 persons exposed to rescue work after an airplane crash with deaths and survivors and 217 other rescue workers Previous disaster experience and disaster exposure to dead bodies, physical danger, and assistance of survivors  Zung scale Age, gender, marital status, education No clean baseline   Previous disaster experience: RR = 1.2 (95% CI: 0.3, 4.9) 
   Disaster exposure to dead bodies, etc.: RR = 3.5 (95% CI: 1.2, 10.6) 
Ylipaavalniemi et al., 2005 (424,278 women and 537 men, all hospital personnel Demands, control, job strain, bad team climate, low procedural injustice, low relational justice Self-reported “doctor-diagnosed depression” Age, gender, lifestyle Clean baseline   Demands: RR = 1.1 (95% CI: 0.8, 1.6) 
  Control: RR = 1.0 (95% CI: 0.7, 1.5) 
  Job strain: RR = 1.3 (95% CI: 0.9, 1.8) 
  Bad team climate: RR = 1.6 (95% CI: 1.1, 2.2) 
  Low procedural injustice: RR = 1.3 (95% CI: 0.9, 1.8) 
  Low relational justice: RR = 1.4 (95% CI: 1.0, 2.0) 
Godin et al., 2005 (43700 women and 836 men, all employees in four plants (Somstress Study) ERI* measured twice SCL90* for depression Age, education, job dissatisfaction, workplace instability Clean baseline    
    1) no-no ERI no-no compared with   
    2) yes-no     Yes-no: RR: NS   
    3) no-yes     No-yes: RR = 4.6 (95% CI: 2.3, 9.2)   
    4) yes-yes‡     Yes-yes: RR = 2.8 (95% CI: 1.3, 5.7)   
Author(s), year, and reference Population Exposure measures Follow-up time, years Outcome Confounder adjustments Analysis Risk of depression 
Men Women Both genders 
Wieclaw et al., 2006 (20Among all employed Danes, 14,166 cases and 58,060 controls Occupation with exposure to threats and violence Depression diagnosed in psychiatric hospital Age, sociodemographics Baseline not clean† Violence: RR* = 1.5 (95% CI*: 1.2, 1.9) Violence: RR* = 1.5 (95% CI:* 1.3, 1.7)  
Threats: RR = 1.2 (95% CI: 0.9, 1.5) Threats: RR = 1.5 (95% CI: 1.3, 1.8)  
Plaisier et al., 2007 (301,529 employed men and 1,117 women JCQ* social support CIDI* by trained interviewer Age, gender, health, education Clean baseline; job strain not calculated  Risk for depression for women: RR = 1.8 compared with men Demands: RR = 3.5 (95% CI: 1.9, 6.3) 
 Social support: RR = 0.8 (95% CI: 0.7, 0.9) 
 Decision latitude and job insecurity: NS* 
Shields, 2006 (3112,011 employed Canadians aged 18–75 years JCQ measured coworker and supervisor support twice. 8 for job strain, and 2 for support CIDI in 2002 as a questionnaire Age, gender, marital status, education, occupation, lifestyle factors Clean baseline; detailed information on loss of follow-up Low coworker support: RR = 2.4 (95% CI: 1.7, 3.3) Low coworker support: RR = 1.8 (95% CI: 1.4, 2.4) Job strain in 1994 and 2000: RR = 3.4 (95% CI: 1.8, 6.4) 
Low supervisor support: RR = 1.7 (95% CI: 1.0, 2.7) Low supervisor support: RR = 1.6 (95% CI: 1.2, 2.2) No job strain in 1994 but in 2000: RR = 3.3 (95% CI: 1.8, 6.1) 
Virtanen et al., 2007 (321,704 women and 1,662 men all employed JCQ Antidepressant prescription; data from public register Age, marital status, occupational grade, mental disorder at baseline  Job strain: RR = 2.0 (95% CI: 1.0, 3.8) Job strain: RR = 1.2 (95% CI: 0.7, 2.0)  
Kawakami et al., 1990 (333,045 male workers in an electronics plant Job stress variables DSM-III* by psychiatrist Partly by comparing confounders Clean baseline; adequate analyses “Unsuitable job”: RR = 11.3 (95% CI: 2.0, 61.8)   
Job overload, overtime, human relations: NS   
Shields, 1999 (341,649 women and 2,181 men aged 25–54 years and employed ≥35 hours per week Weekly working hours self-reported CIDI in 2002 as a questionnaire Age, socioeconomics, education, income, occupation, self-employment No clean baseline ≥41 vs. 35–40 hours/week: RR = 0.6 (95% CI: 0.3, 1.3) ≥41 vs. 35–40 hours/week: RR = 2.2 (95% CI: 1.1, 4.4)  
Wang, 2005 (356,099 employed Canadians Work stress: 12 questions on skill discretion, decision authority, psychological demands, job insecurity, physical exertion, and social support Major depressive episode during last year by CIDI–Short Form for Major Depression Age, gender, marital status, income, race, education, medical illness, subsequent mental health service use Clean baseline   RR = 2.4 (95% CI: 1.5, 3.8) 
Paterniti et al., 2002 (362,790 women and 7,729 men working in the French National Electricity and Gas Company Demands, decision latitude (JCQ), social support CES-D* Age, education, marital status, stressful personal events, baseline CES-D score Adjusted for baseline illness p values of regression coefficients Regression coefficient p values  
    Demands: <0.001     Demands: <0.001  
    Decision latitude: <0.01     Decision latitude: NS  
    Social support: <0.01     Social support: <0.05  
de Lange et al., 2002 (38824 employed persons JCQ changes over time 1 × 4 CES-D Age, gender, education Changes in exposure and outcome   Increased job strain associated with depression 
Schonfeld, 2001 (39184 female teachers Episodic stressors (threats, confrontations) 3 × 4 CES-D Age, socioeconomics, race, marital status No clean baseline; no risk estimate  Regression coefficient p value: <0.01  
Moore et al., 2004 (401,235 employees Downsizing: layoffs CES-D Age, gender, education No clean baseline; no risk estimate   Prevalence of depression 5.2% among never layoffs; 10.4% among those with two layoffs (p < 0.001) 
Fullerton et al., 2004 (41116 persons exposed to rescue work after an airplane crash with deaths and survivors and 217 other rescue workers Previous disaster experience and disaster exposure to dead bodies, physical danger, and assistance of survivors  Zung scale Age, gender, marital status, education No clean baseline   Previous disaster experience: RR = 1.2 (95% CI: 0.3, 4.9) 
   Disaster exposure to dead bodies, etc.: RR = 3.5 (95% CI: 1.2, 10.6) 
Ylipaavalniemi et al., 2005 (424,278 women and 537 men, all hospital personnel Demands, control, job strain, bad team climate, low procedural injustice, low relational justice Self-reported “doctor-diagnosed depression” Age, gender, lifestyle Clean baseline   Demands: RR = 1.1 (95% CI: 0.8, 1.6) 
  Control: RR = 1.0 (95% CI: 0.7, 1.5) 
  Job strain: RR = 1.3 (95% CI: 0.9, 1.8) 
  Bad team climate: RR = 1.6 (95% CI: 1.1, 2.2) 
  Low procedural injustice: RR = 1.3 (95% CI: 0.9, 1.8) 
  Low relational justice: RR = 1.4 (95% CI: 1.0, 2.0) 
Godin et al., 2005 (43700 women and 836 men, all employees in four plants (Somstress Study) ERI* measured twice SCL90* for depression Age, education, job dissatisfaction, workplace instability Clean baseline    
    1) no-no ERI no-no compared with   
    2) yes-no     Yes-no: RR: NS   
    3) no-yes     No-yes: RR = 4.6 (95% CI: 2.3, 9.2)   
    4) yes-yes‡     Yes-yes: RR = 2.8 (95% CI: 1.3, 5.7)   
*

RR, relative risk; CI, confidence interval; JCQ, Job Content Questionnaire; CIDI, Composite International Diagnostic Interview; NS, not significant; DSM-III, Diagnostic and Statistical Manual of Mental Disorders, Third Edition; CES-D, Center for Epidemiologic Studies Depression [Scale]; ERI, effort reward imbalance; SCL90, Symptom Check List-90.

Clean baseline (cases with depression at baseline are excluded from the analyses). The risk estimates are therefore based on incident cases.

Refer to text.

Psychological strain at work and outcome in the form of a psychiatric diagnosis or depression based on a diagnostic classification scale

Three of the seven studies used the job strain model as exposure measure (30–32). Three studies used questionnaire-based variables for psychological strain at work (33–35), whereas the last used the ecologic method where occupation was a proxy measure for exposure to threats and violence (20).

The Netherlands Mental Health Survey and Intervention Study (NEMESIS) is a prospective study of 2,646 working men and women aged between 18 and 65 years (30). The study is very well described insofar as there is a thorough description of the sampling procedures, nonrespondents, and methods of analysis. The Job Content Questionnaire (JCQ) was used as the exposure measure, but no job strain variable was calculated. The analysis unfortunately does not distinguish between male and female respondents, but it is stated that the relative risk of depression is 1.8 for women compared with men. The outcome was depression and anxiety disorders according to the Diagnostic and Statistical Manual of Mental Disorders: DSM-IIIR diagnosed by use of the diagnostic tool, the Composite International Diagnostic Interview (often referred to as “CIDI”). The data collection took place by use of trained interviewers. The relative risk of psychosocial strain at work, as measured by psychological demands, was significantly increased at 3.5 (95 percent confidence interval (CI): 1.9, 6.3) for depression. Social support was negatively associated with depression (relative risk (RR) = 0.8, 95 percent CI: 0.7, 0.9), whereas decision latitude and job insecurity were not significantly associated with depression.

The follow-up time was 2 years, while in a similar Canadian National Population Health Survey study, it was 8 years (31). In the Canadian study, more than 12,000 employed people were followed, and exposure was measured by using the JCQ and social support as in the Dutch study. However, it has to be noted that only a shortened form of the JCQ was used and that the reliability of the exposure measures was regarded as problematic. The studies both used the same diagnostic instrument. The Canadian study also had a relevant confounder control, and the design ensured a clean baseline; for example, only incident cases were included in the analyses. Self-reported job strain in 1994 and 2000 caused a relative risk of depression of 3.4 (95 percent CI: 1.8, 6.4). Similarly, for those who did not report job strain in 1994 but did in 2000, the relative risk of depression was 3.3 (95 percent CI: 1.8, 6.1) when measured 2 years later, in 2002. The study does not state the exact numbers for each gender, but in one figure it is shown that the 2-year incidence for depression for men with high job strain is three times greater compared with that for men with low job strain. This is the same for women, where the incidence is twice as high for those with high job strain compared with the incidence for those with low job strain. As in the Dutch study, there was an association between social support at work for both men and women.

On the basis of data from the Canadian National Population Health Survey cohort, Wang (35) published data from a 2-year follow-up from 1994/1995 to 1996/1997. Job strain in this study was analyzed together with other exposure variables, creating a work-stress index. The relative risk of depression during the 2-year follow-up period was 2.4 (95 percent CI: 1.5, 3.8), controlled for relevant confounders, but gender-specific estimates were not reported. It is not stated why the work-stress index was used as an independent variable in the analysis instead of a strain variable derived from the JCQ.

A part of the Canadian National Population Health Survey cohort was analyzed with a focus on the association between weekly working hours and depression (34). The follow-up period was 2 years, but the actual analysis is unclear compared with the later analyses based on this cohort. Furthermore, it is not stated whether the study refers to incident cases of depression or whether the risk estimates are also based on cases of earlier depression. Women working more than 40 hours per week compared with women working 35–40 hours per week were found to have a significantly increased risk of depression of 2.2 (95 percent CI: 1.1, 4.4). There was no such increased risk for men.

In a Japanese study of more than 3,000 male industrial workers, 35 cases of depression diagnosed by a psychiatrist were found during a 3-year period (33). According to baseline exposure measures, it was found that “unsuitable jobs” gave a relative risk of depression of more than 11 (95 percent CI: 2.0, 61.8). A nested case-control design was applied. “Human relations” also predicted significant depression, but in an overall multivariate analysis this variable was not significant. The exposure measure seems to be an ad hoc instrument, but the design and outcome measures are strong, indicating that a mismatch between working conditions and personal resources might increase the risk for developing depression.

In a Finnish register-based study, Virtanen et al. (32) used the JCQ as the exposure measure and antidepressant prescription during the following 3 years as the outcome. The relative risk for an antidepressant prescription was 2.0 (95 percent CI: 1.0, 3.8) for job strain for men and 1.2 (95 percent CI: 0.7, 2.0) for women. An antidepressant prescription is a rather rough proxy measure of depression as the drugs are sometimes also used for other conditions. However, the results from this study were supported by similar results regarding a 12-month prevalence of depressive disorders.

A Danish study, including more than 14,000 people who were admitted to a psychiatric ward and nearly 60,000 controls, examined the association between violence and threats at work on the one side and affective and stress-related diagnoses on the other (20). Women who worked in jobs that had, on average, a high level of exposure to violence and threats had an increased odds ratio of 1.5 (95 percent CI: 1.3, 1.8) for depression. For men, violence was associated with depression of the same magnitude, whereas exposure to threats did not reach significant association. The exposure is not self-reported as the ecologic method is used, whereby the job title the year before hospital admission was used as a proxy measure for the exposure. We therefore regarded the results from this study as longitudinal results, although the study itself was a case-control study.

Psychological strain at work and outcome measured with depression scales

Four studies used the Center for Epidemiologic Studies Depression (CES-D) Scale. The most convincing is the French study of nearly 10,000 employees of Electricité de France–Gaz de France, known as the “GAZEL Study” (36). This cohort was followed for 3 years. The baseline exposure used the JCQ and a measure for social support at work. Confounder adjustment was carried out and controlled for psychiatric symptoms at baseline. In the paper, no risk estimates, that is, relative risks, are published, but regression coefficients showed that for men as well as women high psychological demands and low social support had a significant positive association with subsequent depression measured by means of the Center for Epidemiologic Studies Depression Scale. For men, there was also a positive association between depression and low decision latitude, but this did not apply to women. For high levels of psychological demands, the risk estimates for depression in the first year of follow-up were 1.8 (95 percent CI: 1.6, 2.0) for men and 1.4 (95 percent CI: 1.1, 1.7) for women. The odds ratio for low decision latitude was 1.4 (95 percent CI: 1.2, 1.6) for both genders (37).

In the Study on Musculoskeletal Disorders, Absenteeism, Stress, and Health (known as the “SMASH Study”) involving more than 800 Dutch employees, four measurements of job strain were made at 1-year intervals, showing a significant association between a high score on the depression scale and increased job strain (38). Unfortunately, no risk estimates were made.

A small study of 184 female teachers (39) who were followed for 9 months showed an association between “episodic stressors” (stress and confrontations) and depression. In a study of companies that were in a down-sizing process lasting 2 years, Moore et al. (40) showed a prevalence of depression that was twice as high for those employees who had experienced two layoffs or more, compared with those who had never experienced one. Indirect threats to employment also increased the risk of depression, but the analyses reported that only direct layoff contacts seem to increase the risk significantly.

Fullerton et al. (41) studied the development of depression among rescue workers who were exposed to dead bodies and physical danger and who gave assistance to survivors in disaster situations. The study observed the rescue workers for 1 year by use of the Zung depression scale and found a relative risk of developing depression of 3.5 (95 percent CI: 1.2, 10.6). Previous experience of a disaster did not affect the development of depression during the follow-up period.

A Finnish study of nearly 5,000 mainly female employees at a hospital showed an association between “team climate” and self-reported, physician-diagnosed depression in the 2-year follow-up period (42). No significant association between job strain and depression was found (RR = 1.3, 95 percent CI: 0.9, 1.8). Neither were demands (RR = 1.1, 95 percent CI: 0.8, 1.6) nor control (RR = 1.0, 95 percent CI: 0.7, 1.5) associated with depression. However, a low degree of relational justice at the workplace was associated with increased risk of depression (RR = 1.4, 95 percent CI: 1.0, 2.0).

The effort/reward imbalance model was tested in a Belgian study known as “the Somstress Study” (43), which showed that an increasing effort/reward imbalance during the course of 1 year had a strong statistical association with depression and anxiety symptoms in both men and women. In the analyses, the reference group was those who did not experience effort/reward imbalance at either baseline or time 2, the follow-up (in table 2 labeled “no-no”). Only those reporting an effort/reward imbalance at time 2 had increased risk of depression, especially the “no-yes” group who went from a balanced to an imbalanced situation during the follow-up period. The measure of exposure at time 2 was, however, done at the same time as the outcome evaluation. For this reason, the risk estimates might be biased, as the two sets of variables cannot be considered independent. A closer look into the data, however, justifies the conclusion, that effort/reward imbalance is associated with outcome as the relative risks would be approximately 2.0 if those with imbalance at time 1 were compared with those without imbalance at the same time.

Taken as a whole, the studies using the Job Strain Model did not show consistent results, but there is a certain support for the proposition that psychological demands increase the risk of depression.

An overview of results from studies that have calculated risk estimates on the relation between work-related psychological strain and depression is shown in figure 2. Only nine of the 14 studies reported calculations or relative risks, which could be used in a graphical expression. The relative risks in most studies are seen to be significantly higher than 1 regarding strain and demands.

FIGURE 2.

Relative risk and 95% confidence interval in longitudinal studies on the relation between work-related strain and development of depression. The studies are grouped according to exposure.

FIGURE 2.

Relative risk and 95% confidence interval in longitudinal studies on the relation between work-related strain and development of depression. The studies are grouped according to exposure.

Social support and depression

Social support as an exposure variable has been included in four of the studies (30, 31, 37, 42). The studies, however, used different, not comparable, measures of social support. In the NEMESIS (30), a broad definition of emotional social support including private life support was applied. A high degree of social support was negatively associated with depression (RR = 0.8, 95 percent CI: 0.7, 0.9). The effect was stronger for men compared with women. Social support did not buffer the unfavorable effect of working conditions.

In the GAZEL Study (31), the risk for depression by low social support at work was 1.6 (95 percent CI: 1.4, 1.8) for men and 1.3 (95 percent CI: 1.1, 1.6) for women. In this study, the measure of social support dealt only with coworker support. In the Canadian National Population Health Survey study (37), both coworker and supervisor support were taken into account. The risk for depression was significantly increased with lack of both, most pronounced for coworker support (table 2). Finally, a Finnish study (42) used “team climate” covering whether the interpersonal atmosphere was supportive, including support for innovation, a shared vision, and commitment to excellence as the exposure measure. A low degree of team climate increased the risk of depression to 1.6 (95 percent CI: 1.1, 2.2).

In conclusion, in light of these studies, there is clear evidence showing an association between low social support at work and increased depression.

Publication bias and feasibility of a meta-analysis

The feasibility of a formal meta-analysis was discussed after a search was performed and studies were included. Exposure measures were somewhat diverse and, for some of the studies, findings were reported independently for men and women. Thus, two funnel plots were made (plotting the point estimates against their variance), one plot using a broad definition of “job strain” and one main finding from each study and another plot where the studies were regrouped according to more narrow definitions. Each funnel plot was supplemented with a test for publication bias by using Kendall's tau, assessing the rank correlation between variance and point estimate.

The first funnel plot included the main finding from each of the 10 studies judged to include a broad conceptualization of “work-related strain” as the exposure variable (not shown); the plot clearly indicated publication bias (Kendall's tau, 0.74; p < 0.01). The studies were then regrouped into studies using the strict definition of job strain according to Karasek et al. (19) (figure 2, topmost part); studies reporting only one or two of the dimensions from the strict definition, demand (figure 2, upper middle part) and decision latitude (figure 2, lower middle part); and “other” operationalizations of work-related strain (figure 2, lowest part). The first three groups each included too few studies to make assessment of possible publication bias meaningful. Visual inspection of the plot for the last group indicates strong publication bias (Kendall's tau, 0.97; p < 0.01). Thus, we have not pooled the point estimates from the individual studies by a formal statistical meta-analysis.

DISCUSSION

This literature review has identified occupational psychosocial factors that, in longitudinal epidemiologic studies, were associated with the development of depression. The associations were strongest and most consistent with regard to high psychological demands and low degree of social support. The risk estimates for men and women were slightly different in the various studies, but there was no general tendency for the risk to be higher in one gender compared with the other.

Demands have, however, been measured in different ways, and not all studies have published the exact items used. This is especially a problem as the JCQ has changed over time and the nature of jobs has changed too. The main problem is to what extent demands are measured by variables indicating speed and tempo at work or troubles and other more cognitive burdens. Work pace and time pressure might be relevant in industrial settings, while psychological pressure is more important in the health-care sector. This issue has been recently discussed in detail (16). This might be the reason for the negative findings in some of the studies dealing with employees in the public sector.

The results regarding decision latitude are contradictory. The term “decision latitude” reflects the degree of control one experiences over one's working condition. The degree of control is normally associated with social status and, as pointed out by Griffin et al. (44), control has different levels of importance for prediction of mental illness in different social groups. Another point made by this group is that there is a spillover effect from work to private life and vice versa regarding control, and this effect is different for men and women. Again, this dimension might be more useful to apply in studies in industrial settings or in working environments where other psychological exposures are not so dominant. The lack of homogeneous results regarding decision latitude might partly be a result of this.

Social support has been measured in different ways, most often as a combination of coworker and supervisor support. In the NEMESIS, the measures even included social support outside work. The studies dealt with in this review, however, do not clarify what kind of social support is important in preventing depression. Neither answers the question if social support is more important for men than for women or vice versa.

The epidemiologic studies do not allow any conclusions to be drawn with regard to the duration or the intensity of the exposure. A limited number of the studies measured exposure several times before the outcome measure, but it was not clear whether an increase in the exposures measured caused an increased risk of developing depression more than a long-term exposure did. The studies document neither the overall extent of the exposure over time nor the onset of the depression in relation to the exposure.

The majority of the epidemiologic studies included in this review measured exposure only a few years prior to the outcome. This means that the studies cannot state whether or not the exposure was present at the onset of the depression, which means that there is a strong possibility that the exposure might have changed from weak to strong or vice versa before the onset of the disorder. For this reason, there is likely to be an underestimation of the statistical association between exposure and disorder. Although the studies seem to conclude that there is evidence for an association between certain types of psychosocial strain at work and the development of depression, the studies cannot give answer to the duration or intensity of exposure needed to increase the risk of depression. This applies to precise information about the method of diagnosing the depression and about the severity of the depression. In addition, it is impossible to conclude anything with regard to competing causes of the disorder, except in those cases where the appropriate adjustments for alcohol consumption and other lifestyle factors have been made. In those studies where adjustment for confounding has been carried out, the risk estimate has not been affected to any great extent.

The assessment of publication bias is slightly hampered by the facts that some of the included studies have published independent risk estimates for men and women, some studies report several risk estimates with various confounder controls, and the delimitation of studies with “other” operationalizations of “job strain” is somewhat subjective. Nevertheless, the indication of publication bias in relation to job strain seems very strong with a correlation coefficient of almost one for the large fraction of studies using nonstandard definitions of work-related strain. Among all the studies, only one study reported only nonsignificant associations (42). One might suspect that this nonsignificant finding was published mainly because the social support variable in the same study turned out as statistically significant and, thus, the job strain finding was published alongside this other finding irrespective of its nonsignificance. In this context, it is of interest to note that a study of predictors of false positive findings among epidemiologic studies of various exposures and various types of cancers (45) found that one of the strongest predictors was “fishing expeditions,” that is, testing post hoc hypotheses in a data set at hand. Because all the risk factors studied in this paper have their origin in relation to diseases other than depression, post hoc testing and publication bias are almost to be expected. For controlled trials, the largest studies have been considered the most reliable in situations with publication biases, but there is no reason to assume that this would also be true for observational studies.

FUTURE RESEARCH

In future research, it will be important that exposure measures are both more broad and more precise. To date, assessment of psychosocial working conditions in epidemiologic studies is mainly focused on the components of the demand/control/support model and, to a lesser extent, on the effort/reward imbalance model. Whereas these models have great merits, other psychosocial working conditions, for example, emotional demands, interpersonal conflicts, quality of management, or predictability at work, have been suggested as health-hazardous factors. However, these factors have rarely been investigated in epidemiologic studies.

In future research, more precise exposure measures are needed. Most studies until now have no evaluation of the duration or the intensity of a given exposure. Such evaluation cannot be obtained by means of postal or Internet-based questionnaires alone. In addition, interviews in depth, although standardized, can be applied in order to gather more information on the quality and quantity of stressors. This method, however, contains a risk for information bias, especially as cases of depression might have their experiences colored by their mental illness condition. Using information from other sources might be a way to validate the exposure. Yet, in the case of emotional strain, other sources might be lacking. The exposure assessment could also be improved by characterization of specific working conditions that might be responsible for the increased risk for development of mental disorders.

All this can be obtained by using a nested case-control design. Cases have to be incident cases well characterized diagnostically, and gathering a sufficient number of cases for a case-control design is the only feasible way. The cases could be derived from a cohort study with exposure assessment several times prior to the follow-up as done in the SMASH Study in order to estimate the duration of exposure and the relation between exposure and onset of depression. A screening instrument might be applied in such studies, followed by a more direct diagnostic evaluation of each case by use of the Composite International Diagnostic Interview. Ongoing cohort studies might use the outcome measures as screening instruments and add psychiatric evaluation of cases by standardized methods in order to improve the validity of the outcome.

Confounding factors must include personality and private life stressors, in addition to the confounders included in many of the reviewed studies. The analyses should be stratified by gender instead of just adjusting for gender as done in most studies. As the quality regarding the design of several of the referred studies is high, adjustments, as described above, of measures applied in these studies might to some extent improve the validity of these studies and thereby give further contribution to our knowledge about the relation between work-related stressors and the development of mental disorders. In order to minimize publication bias, detailed protocols for future analyses in the research area should be published before follow-up data are collected. Just a few such studies will be much more reliable than many selective publications based on fishing expeditions.

CONCLUSION

A total of 14 longitudinal epidemiologic studies on the relation between work-related psychological stress factors and the development of depression have been traced and evaluated. Psychological strain at work seems to be associated with future depression. Moderate evidence for a relation between psychological demands in the job and development of depression was found. The relative risk estimates were approximately 2.0. However, indication of publication bias weakens the evidence. Social support at work seems to decrease the risk for future depression, as all four studies dealing with this exposure showed associations with relative risks to be about 0.6.

Even if this literature study has identified work-related psychosocial factors that in high-quality epidemiologic studies predict depression, we still need studies that assess in more detail the duration and intensity of exposure needed for developing major depression. For other common mental disorders, such as anxiety and somatoform disorders, studies based on clinical diagnostic measures are strongly needed. Furthermore, transient and less severe reactions as adjustment disorder may be even more relevant to study as these probably are more prevalent and overall result in more sick leave than the more severe mental disorders. Attention in this context must be drawn to the fact that work-related psychosocial factors might have different impacts in different occupational settings.

Abbreviations

    Abbreviations
  • CI

    confidence interval

  • JCQ

    Job Content Questionnaire

  • NEMESIS

    Netherlands Mental Health Survey and Intervention Study

  • RR

    relative risk

The Danish Work Environmental Fund has supported this review (grant 2-2006-09).

Conflict of interest: none declared.

References

1.
Horwath
E
Cohen
R
Weissman
M
Tsuang
M
Tohen
M
Epidemiology of depressive and anxiety disorders
Textbook in psychiatric epidemiology
 , 
2002
New York, NY
Wiley-Liss
(pg. 
389
-
426
)
2.
Goodwin
R
Jacobi
F
Bittner
A
, et al.  . 
Stein
D
Kupfer
D
Schtzberg
A
Epidemiology of mood disorders
Textbook of mood disorders
 , 
2006
Washington, DC
American Psychiatric Publishing
(pg. 
33
-
54
)
3.
Fogel
J
Eaton
WW
Ford
DE
Minor depression as a predictor of the first onset of major depressive disorder over a 15-year follow-up
Acta Psychiatr Scand
 , 
2006
, vol. 
13
 (pg. 
36
-
43
)
4.
Murray
CJ
Lopez
AD
Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study
Lancet
 , 
1997
, vol. 
349
 (pg. 
1436
-
42
)
5.
Broadhead
WE
Blazer
DG
George
LK
, et al.  . 
Depression, disability days, and days lost from work in a prospective epidemiologic survey
JAMA
 , 
1990
, vol. 
264
 (pg. 
2524
-
8
)
6.
Demyttenaere
K
Bruffaerts
R
Posada-Villa
J
, et al.  . 
Prevalence, severity, and unmet need for treatment of mental disorders in the World Health Organization World Mental Health Surveys
JAMA
 , 
2004
, vol. 
291
 (pg. 
2581
-
90
)
7.
Bech
P
Lucas
R
Amir
M
, et al.  . 
Association between clinically depressed subgroups, type of treatment and patient retention in the LIDO study
Psychol Med
 , 
2003
, vol. 
33
 (pg. 
1051
-
9
)
8.
Kessler
RC
Greenberg
PE
Mickelson
KD
, et al.  . 
The effects of chronic medical conditions on work loss and work cutback
J Occup Environ Med
 , 
2001
, vol. 
43
 (pg. 
218
-
25
)
9.
Harris
T
Recent developments in understanding the psychosocial aspects of depression
Br Med Bull
 , 
2007
, vol. 
57
 (pg. 
17
-
32
)
10.
Lorant
V
Deliege
D
Eaton
W
, et al.  . 
Socioeconomic inequalities in depression: a meta-analysis
Am J Epidemiol
 , 
2003
, vol. 
157
 (pg. 
98
-
112
)
11.
Sullivan
PF
Neale
MC
Kendler
KS
Genetic epidemiology of major depression: review and meta-analysis
Am J Psychiatry
 , 
2000
, vol. 
157
 (pg. 
1552
-
62
)
12.
Wittchen
HU
Kessler
RC
Pfister
H
, et al.  . 
Why do people with anxiety disorders become depressed? A prospective-longitudinal community study
Acta Psychiatr Scand Suppl
 , 
2000
, vol. 
406
 (pg. 
14
-
23
)
13.
Schleifer
SJ
Macari-Hinson
MM
Coyle
DA
, et al.  . 
The nature and course of depression following myocardial infarction
Arch Intern Med
 , 
1989
, vol. 
149
 (pg. 
1785
-
9
)
14.
Fann
JR
Thomas-Rich
AM
Katon
WJ
, et al.  . 
Major depression after breast cancer: a review of epidemiology and treatment
Gen Hosp Psychiatry
 , 
2008
, vol. 
30
 (pg. 
112
-
26
)
15.
Tennant
C
Work-related stress and depressive disorders
J Psychosom Res
 , 
2007
, vol. 
51
 (pg. 
697
-
704
)
16.
Stansfeld
S
Candy
B
Psychosocial work environment and mental health—a meta-analytic review
Scand J Work Environ Health
 , 
2006
, vol. 
32
 (pg. 
443
-
62
)
17.
Tarumi
K
Hagihara
A
Morimoto
K
A prospective observation of onsets of health defects associated with working hours
Ind Health
 , 
2003
, vol. 
41
 (pg. 
101
-
8
)
18.
Greiner
BA
Ragland
DR
Krause
N
, et al.  . 
Objective measurement of occupational stress factors—an example with San Francisco urban transit operators
J Occup Health Psychol
 , 
1997
, vol. 
2
 (pg. 
325
-
42
)
19.
Karasek
R
Brisson
C
Kawakami
N
, et al.  . 
The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics
J Occup Health Psychol
 , 
1998
, vol. 
3
 (pg. 
322
-
55
)
20.
Wieclaw
J
Agerbo
E
Mortensen
PB
, et al.  . 
Work related violence and threats and the risk of depression and stress disorders
J Epidemiol Community Health
 , 
2006
, vol. 
60
 (pg. 
771
-
5
)
21.
Griffin
JM
Greiner
BA
Stansfeld
SA
, et al.  . 
The effect of self-reported and observed job conditions on depression and anxiety symptoms: a comparison of theoretical models
J Occup Health Psychol
 , 
2007
, vol. 
12
 (pg. 
334
-
49
)
22.
Leitner
K
Resch
MG
Do the effects of job stressors on health persist over time? A longitudinal study with observational stressor measures
J Occup Health Psychol
 , 
2005
, vol. 
10
 (pg. 
18
-
30
)
23.
Frese
M
Zapf
D
Cooper
C
Payne
R
Methodological issues in the study of work stress: objective vs subjective measurement of work stress and the question of longitudinal studies
Causes, coping and consequences of stress at work
 , 
1988
Chichester, England
John Wiley & Sons
(pg. 
375
-
411
)
24.
Kasl
SV
Jones
B
Quick
J
Tetrick
L
An epidemiological perspective on research design, measurement, and surveillance strategies
Occupational health psychology
 , 
2003
Washington, DC
American Psychological Association
(pg. 
379
-
98
)
25.
Siegrist
J
Adverse health effects of high-effort/low-reward conditions
J Occup Health Psychol
 , 
1996
, vol. 
1
 (pg. 
27
-
41
)
26.
Johnsson
JV
Hall
EM
Theorell
T
Combined effects of job strain and social isolation on cardiovascular disease morbidity and mortality in a random sample of the Swedish male working population
Scand J Work Environ Health
 , 
1989
, vol. 
15
 (pg. 
271
-
9
)
27.
Zapf
D
Einarsen
S
Fox
S
Spector
P
Mobbing at work. Escalated conflicts in organizations
Counterproductive workplace behavior: investigations of actors and targets
 , 
2005
Washington, DC
American Psychological Association
(pg. 
237
-
70
)
28.
Siegrist
J
Starke
D
Chandola
T
, et al.  . 
The measurement of effort-reward imbalance at work: European comparisons
Soc Sci Med
 , 
2004
, vol. 
58
 (pg. 
1483
-
99
)
29.
Watson
D
Rethinking the mood and anxiety disorders: a quantitative hierarchical model for DSM-V
J Abnorm Psychol
 , 
2005
, vol. 
114
 (pg. 
522
-
36
)
30.
Plaisier
I
de Bruijn
JGM
de Graff
R
, et al.  . 
The contribution of working conditions and social support to the onset of depressive and anxiety disorders among male and female employees
Soc Sci Med
 , 
2007
, vol. 
64
 (pg. 
401
-
10
)
31.
Shields
M
Stress and depression in the employed population
Health Rep
 , 
2006
, vol. 
17
 (pg. 
11
-
29
)
32.
Virtanen
M
Honkonen
T
Kivimaki
M
, et al.  . 
Work stress, mental health and antidepressant medication findings from the Health 2000 Study
J Affect Disord
 , 
2007
, vol. 
98
 (pg. 
189
-
97
)
33.
Kawakami
N
Araki
S
Kawashima
M
Effects of job stress on occurrence of major depression in Japanese industry: a case-control study nested in a cohort study
J Occup Med
 , 
1990
, vol. 
32
 (pg. 
722
-
5
)
34.
Shields
M
Long working hours and health
Health Rep
 , 
1999
, vol. 
11
 (pg. 
33
-
48
)
35.
Wang
J
Work stress as a risk factor for major depressive episode(s)
Psychol Med
 , 
2005
, vol. 
35
 (pg. 
865
-
71
)
36.
Paterniti
S
Niedhammer
I
Lang
T
, et al.  . 
Psychosocial factors at work, personality traits and depressive symptoms: longitudinal results from the GAZEL Study
Br J Psychiatry
 , 
2002
, vol. 
181
 (pg. 
111
-
17
)
37.
Niedhammer
I
Bugel
I
Goldberg
M
, et al.  . 
Psychosocial factors at work and sickness absence in the Gazel cohort: a prospective study
Occup Environ Med
 , 
1998
, vol. 
55
 (pg. 
735
-
41
)
38.
de Lange
AH
Taris
TW
Kompier
MA
, et al.  . 
Effects of stable and changing demand-control histories on worker health
Scand J Work Environ Health
 , 
2002
, vol. 
28
 (pg. 
94
-
108
)
39.
Schonfeld
IS
Stress in 1st-year women teachers: the context of social support and coping
Genet Soc Gen Psychol Monogr
 , 
2001
, vol. 
127
 (pg. 
133
-
68
)
40.
Moore
S
Grunberg
L
Greenberg
E
Repeated downsizing contact: the effects of similar and dissimilar layoff experiences on work and well-being outcomes
J Occup Health Psychol
 , 
2004
, vol. 
9
 (pg. 
247
-
57
)
41.
Fullerton
CS
Ursano
RJ
Wang
L
Acute stress disorder, posttraumatic stress disorder, and depression in disaster or rescue workers
Am J Psychiatry
 , 
2004
, vol. 
161
 (pg. 
1370
-
6
)
42.
Ylipaavalniemi
J
Kivimaki
M
Elovainio
M
, et al.  . 
Psychosocial work characteristics and incidence of newly diagnosed depression: a prospective cohort study of three different models
Soc Sci Med
 , 
2005
, vol. 
61
 (pg. 
111
-
22
)
43.
Godin
I
Kittel
F
Coppieters
Y
, et al.  . 
A prospective study of cumulative job stress in relation to mental health
BMC Public Health
 , 
2005
, vol. 
5
 (pg. 
67
-
77
)
44.
Griffin
JM
Fuhrer
R
Stansfeld
SA
, et al.  . 
The importance of low control at work and home on depression and anxiety: do these effects vary by gender and social class?
Soc Sci Med
 , 
2002
, vol. 
54
 (pg. 
783
-
98
)
45.
Swaen
GMH
Teggeler
O
Van Amelsvoort
LGPM
False positive outcomes and design characteristics in occupational cancer epidemiological studies
Int J Epidemiol
 , 
2001
, vol. 
5
 (pg. 
948
-
54
)
46.
de Lange
AH
Taris
TW
Kompier
MAJ
, et al.  . 
The relationships between work characteristics and mental health: examining normal, reversed and reciprocal relationships in a 4-wave study
Work Stress
 , 
2004
, vol. 
18
 (pg. 
149
-
66
)
47.
Kawakami
N
Hayashi
T
Masumoto
T
, et al.  . 
Perceived job-stress and major depression in industry: a case-control study
J UOEH
 , 
1989
, vol. 
11
 (pg. 
467
-
73
)
48.
Kawakami
N
Haratani
T
Araki
S
Effects of perceived job stress on depressive symptoms in blue collar workers of an electrical factory in Japan
Scand J Work Environ Health
 , 
1992
, vol. 
18
 (pg. 
195
-
200
)
49.
Marchand
A
Demers
A
Durand
P
Social structures, agent personality and workers' mental health: a longitudinal analysis of the specific role of occupation and of workplace constraints-resources on psychological distress in the Canadian workforce
Hum Relat
 , 
2006
, vol. 
59
 (pg. 
875
-
901
)
50.
Moore
S
Grunberg
L
Greenberg
E
A longitudinal exploration of alcohol use and problems comparing managerial and nonmanagerial men and women
Addict Behav
 , 
2003
, vol. 
28
 (pg. 
687
-
703
)
51.
Niedhammer
I
Tek
ML
Starke
D
, et al.  . 
Effort-reward imbalance model and self-reported health: cross-sectional and prospective findings from the GAZEL cohort
Soc Sci Med
 , 
2004
, vol. 
58
 (pg. 
1531
-
41
)
52.
Schonfeld
IS
A longitudinal study of occupational stressors and depressive symptoms in first-year female teachers
Teach Teacher Educ
 , 
1992
, vol. 
8
 (pg. 
151
-
8
)
53.
Bourbonnais
R
Brisson
C
Vinet
A
, et al.  . 
Effectiveness of a participative intervention on psychosocial work factors to prevent mental health problems in a hospital setting
Occup Environ Med
 , 
2006
, vol. 
63
 (pg. 
335
-
42
)
54.
Buck
D
Gregson
BA
Bamford
CH
, et al.  . 
Psychological distress among informal supporters of frail older people at home and in institutions
Int J Geriatr Psychiatry
 , 
1997
, vol. 
12
 (pg. 
737
-
44
)
55.
Dormann
C
Zapf
D
Social support, social stressors at work, and depressive symptoms: testing for main and moderating effects with structural equations in a three-wave longitudinal study
J Appl Psychol
 , 
1999
, vol. 
84
 (pg. 
874
-
84
)
56.
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
, vol. 
75
 (pg. 
33
-
58
)
57.
Firth-Cozens
J
Individual and organizational predictors of depression in general practitioners
Br J Gen Prac
 , 
1998
, vol. 
48
 (pg. 
1647
-
51
)
58.
Kawakami
N
Araki
S
Kawashima
M
, et al.  . 
Effects of work-related stress reduction on depressive symptoms among Japanese blue-collar workers
Scand J Work Environ Health
 , 
1997
, vol. 
23
 (pg. 
54
-
9
)
59.
Macleod
J
Davey
SG
Heslop
P
, et al.  . 
Psychological stress and cardiovascular disease: empirical demonstration of bias in a prospective observational study of Scottish men
BMJ
 , 
2002
, vol. 
324
 (pg. 
1247
-
51
)
60.
Michelsen
H
Bildt
C
Psychosocial conditions on and off the job and psychological ill health: depressive symptoms, impaired psychological wellbeing, heavy consumption of alcohol
Occup Environ Med
 , 
2003
, vol. 
60
 (pg. 
489
-
96
)
61.
Olstad
R
Sexton
H
Sogaard
AJ
The Finnmark Study. A prospective population study of the social support buffer hypothesis, specific stressors and mental distress
Soc Psychiatry Psychiatr Epidemiol
 , 
2001
, vol. 
36
 (pg. 
582
-
9
)
62.
Sanderson
K
Andrews
G
Common mental disorders in the workforce: recent findings from descriptive and social epidemiology
Can J Psychiatry
 , 
2006
, vol. 
51
 (pg. 
63
-
75
)
63.
Shields
M
Shift work and health
Health Rep
 , 
2002
, vol. 
13
 (pg. 
11
-
33
)
64.
Taris
TW
The mutual effects between job resources and mental health: a prospective study among Dutch youth
Genet Soc Gen Psychol Monogr
 , 
1999
, vol. 
125
 (pg. 
433
-
50
)
65.
Tyssen
R
Vaglum
P
Gronvold
NT
, et al.  . 
The impact of job stress and working conditions on mental health problems among junior house officers. A nationwide Norwegian prospective cohort study
Med Educ
 , 
2000
, vol. 
34
 (pg. 
374
-
84
)
66.
Carayon
P
A longitudinal test of Karasek's Job Strain Model among office workers
Work Stress
 , 
1993
, vol. 
7
 (pg. 
299
-
314
)
67.
Carayon
P
Quick
JC
Murphy
LR
Hurrell
J
Jr
A longitudinal study of job design and worker strain: preliminary results
Stress & well-being at work: assessments and interventions for occupational mental health
 , 
1992
Washington, DC
American Psychological Association
(pg. 
19
-
32
)
68.
Gopal
R
Glasheen
JJ
Miyoshi
TJ
, et al.  . 
Burnout and internal medicine resident work-hour restrictions
Arch Intern Med
 , 
2005
, vol. 
165
 (pg. 
2595
-
600
)
69.
Maguen
S
Litz
BT
Wang
JL
, et al.  . 
The stressors and demands of peacekeeping in Kosovo: predictors of mental health response
Mil Med
 , 
2004
, vol. 
169
 (pg. 
198
-
206
)
70.
McKnight
JD
Glass
DC
Perceptions of control, burnout, and depressive symptomatology
J Consult Clin Psychol
 , 
1995
, vol. 
63
 (pg. 
490
-
4
)
71.
Parker
SK
Longitudinal effects of lean production on employee outcomes and the mediating role of work characteristics
J Appl Psychol
 , 
2003
, vol. 
88
 (pg. 
620
-
34
)
72.
Rydstedt
LW
Devereux
J
Furnham
AF
Are lay theories of work stress related to distress? A longitudinal study in the British workforce
Work Stress
 , 
2004
, vol. 
18
 (pg. 
245
-
54
)
73.
Schaefer
JA
Moos
RH
Effects of work stressors and work climate on long-term care staff's job morale and functioning
Res Nurs Health
 , 
1996
, vol. 
19
 (pg. 
63
-
73
)
74.
Vaananen
A
Kalimo
R
Toppinen-Tanner
S
, et al.  . 
Role clarity, fairness, and organizational climate as predictors of sickness absence: a prospective study in the private sector
Scand J Public Health
 , 
2004
, vol. 
32
 (pg. 
426
-
34
)
75.
van der Ploeg
E
Kleber
RJ
Acute and chronic job stressors among ambulance personnel: predictors of health symptoms
Occup Environ Med
 , 
2003
, vol. 
60
 
suppl 1
(pg. 
i40
-
6
)
76.
Verhaeghe
R
Mak
R
Maele
V
, et al.  . 
Job stress among middle-aged health care workers and its relation to sickness absence
Stress Health
 , 
2003
, vol. 
19
 (pg. 
265
-
74
)
77.
Crum
RM
Muntaner
C
Eaton
WW
, et al.  . 
Occupational stress and the risk of alcohol abuse and dependence
Alcohol Clin Exp Res
 , 
1995
, vol. 
19
 (pg. 
647
-
55
)
78.
Freels
SA
Richman
JA
Rospenda
KM
Gender differences in the causal direction between workplace harassment and drinking
Addict Behav
 , 
2005
, vol. 
30
 (pg. 
1454
-
8
)
79.
Head
J
Stansfeld
SA
Siegrist
J
The psychosocial work environment and alcohol dependence: a prospective study
Occup Environ Med
 , 
2004
, vol. 
61
 (pg. 
219
-
24
)
80.
Romelsjo
A
Hasin
D
Hilton
M
, et al.  . 
The relationship between stressful working conditions and high alcohol consumption and severe alcohol problems in an urban general population
Br J Addict
 , 
1992
, vol. 
87
 (pg. 
1173
-
83
)
81.
Muntaner
C
Anthony
JC
Crum
RM
, et al.  . 
Psychosocial dimensions of work and the risk of drug dependence among adults
Am J Epidemiol
 , 
1995
, vol. 
142
 (pg. 
183
-
90
)
82.
Reed
PL
Storr
CL
Anthony
JC
Drug dependence enviromics: job strain in the work environment and risk of becoming drug-dependent
Am J Epidemiol
 , 
2006
, vol. 
163
 (pg. 
404
-
11
)
83.
Ala-Marsula
L
Vahtera
J
Pentti
J
, et al.  . 
Effect of employee worktime control on health: a prospective cohort study
Occup Environ Med
 , 
2004
, vol. 
61
 (pg. 
254
-
61
)
84.
Bultmann
U
Kant
IJ
van den Brandt
PA
, et al.  . 
Psychosocial work characteristics as risk factors for the onset of fatigue and psychological distress: prospective results from the Maastricht Cohort Study
Psychol Med
 , 
2002
, vol. 
32
 (pg. 
333
-
45
)
85.
Ferrie
JE
Martikainen
P
Shipley
MJ
, et al.  . 
Employment status and health after privatisation in white collar civil servants: prospective cohort study
BMJ
 , 
2001
, vol. 
322
 (pg. 
647
-
51
)
86.
Ferrie
JE
Head
J
Shipley
MJ
, et al.  . 
Injustice at work and incidence of psychiatric morbidity: the Whitehall II study
Occup Environ Med
 , 
2006
, vol. 
63
 (pg. 
443
-
50
)
87.
Johnson
JV
Hall
EM
Ford
DE
, et al.  . 
The psychosocial work environment of physicians: the impact of demands and resources on job dissatisfaction and psychiatric distress in a longitudinal study of Johns Hopkins Medical School graduates
J Occup Environ Med
 , 
1995
, vol. 
37
 (pg. 
1151
-
9
)
88.
Kivimaki
M
Virtanen
M
Vartia
M
, et al.  . 
Workplace bullying and the risk of cardiovascular disease and depression
Occup Environ Med
 , 
2003
, vol. 
60
 (pg. 
779
-
83
)
89.
Stansfeld
SA
Bosma
H
Hemingway
H
, et al.  . 
Work characteristics predict psychiatric disorders: prospective results from the Whitehall II Study
Occup Environ Med
 , 
1999
, vol. 
56
 (pg. 
302
-
7
)
90.
Swaen
GMH
Bultmann
U
Kant
I
, et al.  . 
Effects of job insecurity from a workplace closure threat on fatigue and psychological distress
J Occup Environ Med
 , 
2004
, vol. 
46
 (pg. 
443
-
9
)
91.
Kivimaki
M
Elovainio
M
Vahtera
J
, et al.  . 
Organisational justice and health of employees: prospective cohort study
Occup Environ Med
 , 
2003
, vol. 
60
 (pg. 
27
-
33
)
92.
Bourbonnais
R
Mondor
M
Job strain and sickness absence among nurses in the province of Quebec
Am J Ind Med
 , 
2001
, vol. 
39
 (pg. 
194
-
202
)
93.
Cheng
Y
Kawachi
I
Coakley
EH
, et al.  . 
Association between psychosocial work characteristics and health functioning in American women: prospective study
BMJ
 , 
2000
, vol. 
320
 (pg. 
1432
-
6
)
94.
Eriksen
W
Tambs
K
Knardahl
S
Work factors and psychological distress in nurses' aides: a prospective cohort study. (Electronic journal)
BMC Public Health
 , 
2006
, vol. 
6
 pg. 
290
 
95.
Stansfeld
SA
Fuhrer
R
Head
J
, et al.  . 
Work and psychiatric disorder in the Whitehall II study
J Psychosom Res
 , 
1997
, vol. 
43
 (pg. 
73
-
81
)
96.
Stansfeld
SA
Bosma
H
Hemingway
H
, et al.  . 
Psychosocial work characteristics and social support as predictors of SF-36 health functioning: the Whitehall II study
Psychosom Med
 , 
1998
, vol. 
60
 (pg. 
247
-
55
)
97.
Rugulies
R
Bultmann
U
Aust
B
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
, vol. 
163
 (pg. 
877
-
87
)
98.
Suwazono
Y
Okubo
Y
Kobayashi
E
A follow-up study on the association of working conditions and lifestyles with the development of (perceived) mental symptoms in workers of a telecommunication enterprise
Occup Med
 , 
2003
, vol. 
53
 (pg. 
436
-
42
)
99.
Orth-Gomer
K
Leineweber
C
Multiple stressors and coronary disease in women. The Stockholm Female Coronary Risk Study
Biol Psychol
 , 
2005
, vol. 
69
 (pg. 
57
-
66
)
100.
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
, vol. 
75
 (pg. 
252
-
8
)
101.
Bourbonnais
R
Comeau
M
Vezina
M
Job strain and evolution of mental health among nurses
J Occup Health Psychol
 , 
1999
, vol. 
4
 (pg. 
95
-
107
)
102.
Bromet
EJ
Dew
MA
Parkinson
DK
, et al.  . 
Predictive effects of occupational and marital stress on the mental health of a male workforce
J Organ Behav
 , 
1988
, vol. 
9
 (pg. 
1
-
13
)
103.
Holman
DJ
Wall
TD
Work characteristics, learning-related outcomes, and strain: a test of competing direct effects, mediated, and moderated models
J Occup Health Psychol
 , 
2002
, vol. 
7
 (pg. 
283
-
301
)
104.
Babazono
A
Mino
Y
Nagano
J
, et al.  . 
A prospective study on the influences of workplace stress on mental health
J Occup Health
 , 
2005
, vol. 
47
 (pg. 
490
-
5
)
105.
Shigemi
J
Mino
Y
Ohtsu
T
, et al.  . 
Effects of perceived job stress on mental health. A longitudinal survey in a Japanese electronics company
Eur J Epidemiol
 , 
2000
, vol. 
16
 (pg. 
371
-
6
)
106.
Hogh
A
Henriksson
ME
Burr
HA
5-year follow-up study of aggression at work and psychological health
Int J Behav Med
 , 
2005
, vol. 
12
 (pg. 
256
-
65
)
107.
Mausner-Dorsch
H
Eaton
WW
Psychosocial work environment and depression: epidemiologic assessment of the demand-control model
Am J Public Health
 , 
2000
, vol. 
90
 (pg. 
1765
-
70
)
108.
Petterson
IL
Hertting
A
Hagberg
L
, et al.  . 
Are trends in work and health conditions interrelated? A study of Swedish hospital employees in the 1990s
J Occup Health Psychol
 , 
2005
, vol. 
10
 (pg. 
110
-
20
)