Mental disorders are prevalent in the labour force.1 With the fast changes in the world economy driven by investment, global competition and industrial re-organisation, the work environment is becoming more competitive. There is also a growing importance of knowledge and brain-based skills.2 Such changes exist not only in Canadian companies but in multinational corporations.2 The changes may generate job insecurities for workers3 and demand more cerebral skills and mental performance.4 These work environmental factors may precipitate the onset of mental disorders.5^–7

To improve workers’ mental and physical health status and productivity, there is a need for primary prevention addressing work environmental risk factors. There are several widely accepted theoretical models in workplace health research, including the demand–control model, effort–reward balance model and work–family conflict.8 To this extent, the demand–control model9 10 has become a major theoretical framework of interest. The demand–control model posits that negative health outcomes, such as fatigue, depression, and other physical illnesses, result from situations in which one’s control over one’s work is low and the psychological demands imposed by one’s work are high.9 Studies have shown, however, that not all domains (e.g. decision authority and physical exertion) are associated with an increased risk of mental disorders.5 11 12 Some domains of the demand–control model seem to affect men and women differently in terms of the risk of mental disorders.5 11 12 This raises questions about the targets of workplace interventions and how they could be tailored for gender effects. Theoretically, men’s identity is tied more to their role at work and women’s to their roles at home.13 Results from the Whitehall II study have shown that low control at home affects women’s mental health and general health more than men’s.13 14 Therefore, it is possible that workplace psychosocial factors affect men and women differently. Gender roles may be mutable, however, varied by countries and changed over time.15

Mental disorders have a multifactorial aetiology. Workers’ mental health is not only related to the characteristics of the work environment, but also to the family with whom they live and the society within which they are embedded. In Canada, there is an increasing rate of participation by women in the labour force. Based on the 2006 Canadian national statistics,16 61.3% of women and 73.8% of men participated in the workforce (i.e. employed, looking for paid work, or available for work). The employment rate was 58.2% in women and 70% in men. Compared with men, women were more likely to have part-time employment (21.7% versus 6.4%).16 As a result of the changes in the demographics of the labour force, there has been considerable research interest in the work–family interface and how it affects workers’ health. One of the ways of investigating psychosocial effects could be through research into the effect of work–family and family–work conflicts on health.17 Work–family conflict occurs when efforts to fulfill the demands of the employee role interfere with the ability to fulfill the demands of the roles as a spouse, parent, or care provider.18 Conversely, family–work conflict may be an obstacle to meeting work-related demands and responsibilities successfully, thereby undermining a person’s ability to construct and maintain a positive work-related self-image. As both employee and family roles represent core components of adult identity, impediments to work and family-related identity formation and maintenance are likely to be experienced as stressful19 and may lead to adverse health outcomes.19^–21

Previous studies in workplace mental health mainly focused on depression11 12 or depressive/anxiety symptoms.13 22 The relationship between work stress and anxiety disorders has not been well studied. In addition, sophisticated diagnostic instruments for mental disorders have rarely been used in workplace health research. Previous research has shown that individual characteristics (being unmarried, low family income, chronic illnesses)23^–27 and family structure (single-parent family)26 27 are significant factors associated with mental health problems. Theoretically, these factors and the imbalance between work and family life may act as potential confounders in the relationship between work stress and mental disorders.

In this study, we analysed data from a large population-based mental health survey. The objective of this analysis was to estimate the gender-specific associations between work stress, major depression, anxiety disorders and any mental disorder, adjusting for the effects of demographic, socioeconomic, psychological and clinical variables.

METHODS

Study population

We used data from the Canadian Community Health Survey—Mental Health and Well-being (CCHS-1.2). The CCHS-1.2 was the first Canadian national mental health survey. The methodology of the CCHS-1.2 is described in detail by Gravel and Beland.28 Briefly, the CCHS-1.2 used a nationally representative sample of individuals aged 15 years and older (n  =  36 984). It was initiated and conducted by Statistics Canada between May and December 2002. The participants were selected using multistage, stratified random sampling procedures. The survey content included measures of a number of mental disorders, including major depression, mania, three anxiety disorders (panic disorder, agoraphobia, and social phobia) and alcohol and illicit drug dependence. The interviews were conducted face-to-face by interviewers trained by Statistics Canada. Informed consent was obtained before the interview. The response rate at the national level was 77%. In this analysis, participants who reported working any time in the past 12 months were included (n  =  24 277).

Measurements

In the CCHS-1.2, depressive and anxiety disorders were assessed using a modified version of the World Mental Health—Composite International Diagnostic Interview (WMH–CIDI),29 based on the Diagnostic and Statistics Manual for Mental Disorders, 4th edition, text revision (DSM-IV-TR) criteria.30 Previous versions of the WMH–CIDI have demonstrated reliability and validity,31 and a “clinical calibration” of the WMH–CIDI is under way, assessing its diagnostic consistency relative to the Structured Clinical Interview for the DSM-IV.32 In this analysis, we focused on mental disorders in the past 12 months before the interview.

The CCHS-1.2 used a brief version (12 items) of the Job Content Questionnaire (JCQ)33 to measure work stress in the past 12 months in six dimensions: skill discretion, decision authority, psychological demand, job insecurity, physical exertion and social support from supervisor and/or coworkers. Specific questions can be found in the document from Statistics Canada34 and in other publications.7 11 21

In the CCHS-1.2, the abbreviated version of the JCQ had a moderate internal consistency (α  =  0.55). The Cronbach’s alpha was 0.50 for skill discretion, 0.65 for decision authority, 0.35 for psychological demand, and 0.44 for social support from supervisor/co-workers. The moderate internal consistency does not necessarily mean that the abbreviated version of the JCQ was insensitive. A moderate internal consistency may indicate that there is no redundancy in the measurement and that each item adds new information to the measure.35 If an instrument covers several dimensions, it is reasonable to expect a moderate internal consistency.36 Participants who worked in the past 12 months and who were under the age of 75 years were eligible for these work stress questions.

Other variables included in this analysis were age, marital status (married versus unmarried), personal income levels, single-parent family, long-term medical conditions, number of children in household under 12 years of age, average work hours per week and imbalance between work and family/personal life. In the CCHS-1.2, personal annual income was categorised into 11 levels from no income to $80 000 or over. For descriptive purposes, personal income levels were classified into three categories: low income (less than $30 000), middle income ($30 000 to less than $60 000) and high income ($60 000 and over). All currency figures are in Canadian dollars.

The CCHS-1.2 interview also included a record of 36 chronic medical conditions that had been diagnosed by a health professional. Long-term medical condition was dichotomised as none versus one or more long-term medical conditions, excluding schizophrenia, other psychosis, obsessive–compulsive disorder, dysthymia, post-traumatic stress disorder, autism, anorexia or bulimia and learning disability. The CCHS-1.2 participants were also asked “(During the past month) your life was well-balanced between your family, personal and professional activities?”. In this analysis, the answer to this question was used as an indicator of imbalance between work and family/personal life. On the basis of the results of preliminary analysis, we coded answers “never”, “rarely” and “half the time” as 1 and “frequently” and “almost always” as 0.

Analysis

Based on the theory of demand–control, work stress was classified into four categories: (1) low demand and high control; (2) low demand and low control; (3) high demand and high control; and (4) high demand and low control. Using the first group as a reference, the associations between demand–control and mental disorders were estimated. The relationships between other work stress components (job insecurity, physical exertion and social support from supervisor and co-workers) and mental disorders were analysed separately. The variables (job insecurity, physical exertion and social support from supervisor and co-workers) were analysed as continuous variables.

We first estimated the crude associations between work stress components and mental disorders using logistic regression by gender. Gender-specific associations between work stress, major depression, anxiety disorders and any depressive or anxiety disorder were estimated, controlling for the effects of other selected variables. Work stress components, age and personal income levels were analysed as continuous variables in the logistic regression models. In logistic regression modeling, groups (2), (3) and (4) based on work stress levels were compared with group (1) (the low demand and high control group), separately. The associations between job insecurity, physical exertion, social support from supervisor and co-workers and mental disorders were estimated, controlling for the effects of work stress levels (defined by demand–control) and other selected variables. The association was estimated in the form of odds ratios and associated 95% confidence intervals. The CCHS-1.2 used a complex study design, and therefore sampling weights provided by Statistics Canada were used to estimate the prevalence and odds ratios, accounting for the uneven sampling probability and for non-response. Bootstrap procedures were used to calculate accurate 95% confidence intervals, using the bootstrap weights developed for the CCHS-1.2 data, accounting for the design effects. This analysis was conducted using Stata version 8.0.37

RESULTS

The demographic, socioeconomic and psychosocial characteristics of male and female workers are presented in table 1. Men were more likely than women to be married, at the high personal income level, work more than 40 hours per week, report more stress in physical exertion and social support. Compared with women, men were less like to be a single parent, have one or more chronic conditions, report high demand–low control and stress in job insecurity. They were no different in age and perceived work/home conflict.

High stress in job insecurity and in social support was positively associated with mental disorders, irrespective of gender (table 2). Using the group of low demand–high control as a reference, it was found that both men and women who were in the high demand–low control group were more likely to have had mental disorders. Men in the two intermediate groups were also more likely to have had mental disorders; however, it was not found in women. Physical exertion was not associated with mental disorders in men or in women.

The results of multivariate logistic regression showed that only participants in the group of high demand–low control were more likely to have had mental disorders (table 3). The adjusted associations were, however, smaller than the crude associations, indicating confounding effects from other variables. High levels of work stress in job insecurity and in social support from supervisor/co-workers were associated with the prevalence of mental disorders among men and women.

With major depression as a dependent variable, men in the group of high demand–low control and those who reported high stress in job insecurity and in social support from supervisor/co-workers were more likely to have had major depression (table 4), whereas only perceived high stress in social support was associated with major depression in women.

Levels of psychological demand and control at work were not associated with anxiety disorders (table 5). High stress in job insecurity and in social support was positively associated with having an anxiety disorder, regardless of gender.

DISCUSSION

The CCHS-1.2 data revealed that the associations between work stress and mental disorders varied by gender and by disorders. High demand–low control was significantly associated with major depression in men, but not in women; it was not associated with anxiety disorders, regardless of gender. The 95% confidence intervals related to the non-significant associations were, however, relatively wide. Therefore, interpretations of the non-significant results should be made with caution. On the other hand, these findings were also plausible.

Female single parents were more likely to have had mental disorders; women with one or more children under 12 years were less likely to have had mental disorders. Single-parent status and having children under the age of 12 years were not associated with mental disorders in men. On the other hand, in the multivariate analyses, high demand–low control and job insecurity were strongly associated with major depression in men, but not in women. Domestic psychosocial factors may thus have stronger effects on the health of women compared with men, whereas workplace psychosocial factors may have stronger effects on the mental health of men compared with women.38

Previous research has shown that workers exposed to a high level of psychological demand are more likely to have and to develop major depression.11 12 Psychological demand may have more impact on men than on women with regard to major depression.11 Re-designing job and organisational structure to address the disparities between demand and control, especially reducing the levels of psychological demand, may thus have a significant impact on major depression in men. It is possible that such policy changes in the workplace will improve women’s mental health in general and reduce the prevalence of minor depression. It may have a long-term impact on women’s mental health, i.e. lowering the incidence of major depression in women, because previous research indicates that the gender difference in the incidence of major depressive episodes disappears, after controlling for the effects of work stress and psychosocial factors.7

In multivariate analyses, imbalance between work and family/personal life was the strongest factor associated with mental disorders. As we have not found a study examining the interrelationship between work stress and work/family life imbalance and how it conjunctively affects the risk of having mental disorders, this particular finding could not be compared with those of previous research. The data suggested that balance between work and family/personal life was equally important for men and women in terms of their mental health. This could be a result of the fact that, whereas women still retain primary responsibility for the majority of housework and childcare, men are more involved than previous generations in their family roles.15 Conceptually, work/family conflict defines a link between the workplace, family and a broader society. Therefore, occupational health research should consider psychosocial factors both inside and outside of the workplace to understand the determinants and risk factors for workplace mental health problems fully. Balance between work and family life should be an important component in interventional studies.

Findings from the Whitehall study reported no gender differences in the association between job insecurity and the risk of psychiatric symptoms.38 39 In our analysis, we found that job insecurity increased the likelihood of having both mood and anxiety disorders in men. Among women, job insecurity was associated with anxiety disorders, not with major depression. One explanation may be that it is a reality that men have been the breadwinners in families. Achievements at work are part of identity formation for men.13 The differences between our analysis and the Whitehall study may result from the fact that in the Whitehall study psychiatric symptoms (including both depressive and anxious symptoms) were measured using the General Health Questionnaire40 and the Whitehall study was a longitudinal study in one organisation. The CCHS-1.2 was a population-based cross-sectional survey, assessing mental disorders based on the DSM-IV criteria.

Another interesting finding of this analysis was that there was no gender difference in the relationship between high stress in social support from supervisor/co-workers and mental disorders. Low social support from supervisor/co-workers was associated with mental disorder in both men and women. Research indicates that women are more likely to rely on social support to cope with stress and to solve problems.41 The CCHS-1.2 data suggest that, in the workplace, social support may be equally important for men and women in terms of maintaining good mental health.

This study had several limitations. First, the CCHS-1.2 is a cross-sectional survey. Therefore, the temporal relationship between work stress and mental disorders could not be established. Work stress and work/family imbalance may lead to an increased risk of mental disorders. Conversely, individuals who have a mental disorder may perceive their work environment as more stressful. The interrelationships among psychosocial factors inside and outside of the workplace and their impacts on workers’ mental health should be investigated in longitudinal studies. Second, the CCHS-1.2 data were self-reported. Reporting and recall biases were thus a possibility. Third, work–life balance was assessed using one question. Ideally, a fully structured instrument should be used to assess work–family conflict and family–work conflict explicitly.42 Finally, the JCQ used in the CCHS-1.2 was an abbreviated version. Ideally, a full version of the JCQ should be used.

In spite of the limitations, the CCHS-1.2 was a population-based study using a large representative sample. It assessed mental disorders using a state-of-the-art diagnostic instrument. The results of this analysis highlighted the importance of workplace psychosocial factors and work/family life balance for workers’ mental health. Longitudinal studies incorporating important theoretical models (demand–control, effort–reward imbalance, work–family conflict) are needed to delineate the causal relationships between work stress, effort–reward imbalance, work–family conflict and mental disorders in the whole social context.