Work 40 (2011) 425435 DOI 10.

3233/WOR-2011-1254 IOS Press

425

Occupation, work organisation conditions and the development of chronic psychological distress
` Alain Marchanda,b, and Marie-Eve Blancb
a b

School of Industrial Relations, University of Montreal, Montreal, Quebec, Canada Institute of Public Health, University of Montreal, Montreal, Quebec, Canada

Received 12 January 2010 Accepted 4 April 2010

Abstract. Objective: The contribution of occupation and work organisation conditions to the development of chronic psychological distress in the Canadian workforce was examined. Participants: Longitudinal data came from the rst ve cycles (Cycle 1 = 19941995, Cycle 5 = 20022003) of Statistics Canadas National Population Health Survey. A sample of 5,500 workers nested in 1,300 neighbourhoods and aged 15 to 55 years. Methods: Data had a hierarchical structure and were analysed using multilevel logistic regression models. Results: The results showed 46.4% of workers reported one episode of psychological distress, 23.5% more than one, and 10.6% three or more. Psychological distress decreased over time and varied signicantly between individuals and neighbourhoods. Occupation was not signicant, but psychological demands, job insecurity, and social support in the workplace were important determinants. The results showed strong contributions of individual and outside work factors (family, social support, neighbourhood). Conclusions: Research in occupational mental health must expand theoretical and empirical frameworks in order to capture actual workplace dynamics and the effects on workers mental health. Interventions designed to reduce mental health problems should carefully evaluate not only the work environment itself, but also non-work factors and other individual characteristics. Keywords: Psychological, distress/chronicity/longitudinal, study/occupation/work, organisation, conditions

1. Introduction Over the years, a host of studies have been devoted to mental health problems in the workforce that manifest as psychological distress, depression, and burnout. In 2008, the World Health Organization identied depression as the third leading cause of disability [1] and many workers can develop chronic mental health problems. However, the study of the development of
Address for correspondence: Alain Marchand, School of Industrial Relations, University of Montreal, P.O. Box 6128, Station Centre-ville, Montreal (Que), H3C 3J7, Canada. Tel.: +1 514 343 6111, ext. 1288; Fax: +1 514 343 5764; E-mail: alain.marchand@ umontreal.ca.

chronic mental problems has been largely neglected in the scientic litterature, it has been estimated 19% of Canadian workers reported more than one episode of psychological distress over a six-year period [2]. The morbidity associated with mental health problems represents a signicant burden for both businesses and the whole of society, given the costs associated with mental illness (absenteeism, lower productivity, indemnity payments, and health care). In Canada, for example, related expenses were estimated at $51 billion in 2003 [3]. We analysed the contribution of occupation and work organisation conditions to the development of chronic psychological distress. According to many studies, type of occupation and work organisation condi-

1051-9815/11/$27.50 2011 IOS Press and the authors. All rights reserved

426

A. Marchand and M. Blanc / The development of chronic psychological distress

tions are important factors for explaining the genesis of mental health problems among individuals. Although studies have not specically addressed the problem of chronic psychological distress, the literature has identied occupations at greater risk for mental health problems, like white- and blue-collar workers, semiprofessionals, and supervisors [410]. Higher employment grades were also associated with distress [1113]. Mental health problems appeared widespread across the occupational structure and it was estimated that occupational type accounted for 1 to 3% of the variance in psychological distress [2,6]. Reviews of studies on work organisation conditions have identied factors related to skill utilisation, decision authority, psychological and physical demands, social support in the workplace, and gratications from work [14,15]. The Job Demand-Control Model [16] hypothesises decision latitude (high levels of skill utilisation and decision authority) moderates psychological demands, while the Demand-Control-Support Model [17] postulates low social support in the workplace leads to an increase in the effect of the interaction between decision latitude and psychological demands. While the contribution of the components (demand, control, social support) in these two models to workers mental health are well supported in longitudinal studies [18,19], the majority of empirical studies have failed to support the interactions between demands and control, and between demand-control and social support [5,1829]. Siegrists Effort-Reward Imbalance Model [30] posits the discrepancy between workplace demands and rewards leads to distress, and there is some evidence to support this hypothesis [27,2938]. The multilevel theoretical model of mental health determinants in the workforce is an integrated model postulating the role of stress promoted by constraintsresources embedded in macrosocial structures (occupational structure), structures of daily life (workplace, family, social networks outside the workplace, neighbourhood) and agent personality (demographics, physical health status, psychological traits, life habits, stressful childhood events). This is the more recent model and has received some support [2,15]. Since the impact is not limited to work factors, the multilevel model of mental health determinants in the workforce incorporates and recognises the role of marital, parental, and family economic status; marital and parental strains; social support outside the workplace,gender, age, physical health, self-esteem, locus of control and sense of cohesion; and physical activities, smoking, and life events occurring before the age of 18 years. All of

these elements have been shown to be associated with mental health [14,15]. Despite numerous studies on workers mental health, the development of chronic psychological distress in the workforce and the extent to which occupation and work organisation conditions are related to this phenomenon have been neglected. Furthermore, the eld of occupational mental health has encountered considerable difculty integrating, both theoretically and empirically, the various elements that make up the social environment (work, family, social networks, local community-neighbourhood) of workers to explain the contribution of occupation and work organization conditions to psychological distress [15,18]. Research designs have not included a broad range of workplace conditions to which individuals are subjected, as just as they have not been to take into account both occupations and work organization conditions. Most of the research reviewed for the present study looked at specic occupations, limiting generalization from these results. To better understand how work is related to the development of chronic psychological distress, the research used a large Canadian prospective cohort study to investigate the role of occupation and work organisation conditions on repeated episodes of psychological distress, using the multilevel theoretical model of mental health determinants in the workforce. 2. Methods 2.1. Data Eight years of longitudinal data were derived from the ve cycles of Statistics Canadas National Population Health Survey (NPHS; Cycle 1: 1994-1995; Cycle 2: 19961997; Cycle 3: 19981999; Cycle 4: 2000 2001; Cycle 5: 20022003). The initial cohort included 17,276 subjects selected based on a two-stage sampling design. The rst stage was a sample of independent clusters selected from homogeneous strata of the Canadian territory. In the second stage, a sample of households was selected within each cluster, and one member was sampled from each household. Telephone and face-to-face interviews yielded high response rates for each cycle (80.6%93.6%), for an overall attrition rate of 27.4%. For this study, all respondents aged 15 to 55 years and working during Cycle 1 were selected (n = 7,338). After cases with missing values were deleted, the NPHS cohort analysed in this study comprised 5,500 workers nested in 1,300 neighbour-

A. Marchand and M. Blanc / The development of chronic psychological distress

427

hoods, with an average of 5.6 workers per neighbourhood (unbalanced design). In urban areas, neighbourhoods were dened according to Statistics Canada census tracts (CT). CTs are small geographic units having homogeneous socio-economic characteristics. CTs are found within census metropolitan areas and census agglomerations with urban core populations of 50,000. In rural areas, census subdivisions (CSD) were used. The CSD is the general term applied to small towns and municipalities. In the analysis, data were weighted according to selection probabilities, non-response rates for each cycle, and distribution by gender, age, and province in the 1996 Canadian Census. 2.2. Measures 2.2.1. Psychological distress Psychological distress was measured with the K6 questionnaire [39] based on the Composite International Diagnostic Interview [40], which yielded diagnoses according to criteria from the DSM-III-R and the 10th edition of the International Classication of Diseases (ICD-10). Respondents were to indicate, on a 5-point Likert scale (never/all the time), the frequency of six symptoms over the preceding month: feel so sad that nothing could cheer you up; nervous; restless or dgety; hopeless; worthless; everything was an effort. The psychological distress score was distributed between 0 and 24 (Cronbachs alpha = 0.77). The scale was further dichotomised to identify the presence/absence of psychological distress. Based on prior international studies [4145], the cut-off point ( 5) was set to correspond to the last quintile of the distribution of psychological distress in the general Canadian population in 1994-1995 [2]. Overall, taking repeated measurements of psychological distress in this study allowed us to analyse more precisely how individual patterns of psychological distress changed over time, thereby capturing the development of a chronic state. 2.2.2. Occupation Occupations were categorised according to the fourdigit codes of Statistics Canadas 1991 Standard Occupational Classication (SOC-1991) [46]. Overall, 471 occupations were rst merged into the 16 categories of the Pineo, Porter and McRobert classication of occupation [47]. To take care of the large number of categories and previous research showing greater risk for mental health problems in specic occupations, the 16 categories were further merged into six large groups: senior managers, managers, supervisors, professionals, white-collar workers, and blue-collar workers.

2.2.3. Workplace Skill utilisation, decision authority, physical and psychological demands, social support at work and job insecurity were derived from Statistics Canada brief version of the Karasek Job Content Questionnaire (JCQ) [48], which was based on a ve-point Likert scale (completely disagree/completely agree). Skill utilisation included three items (your job requires that you learn new things, your job requires a high level of skills, your job requires that you do things over and over [reversed coding]), and decision authority two items (your job allows you freedom to decide how you do your job, you have a lot to say about what happens in your job). Physical demands included one item (your job requires a lot of physical effort) and psychological demands were based on two indicators (your job is very hectic, you are free from conicting demands that others make (reversed coding). Job insecurity included one item (your job security is good) and social support included three items (you are exposed to hostility or conict from people you work with (reversed coding); your supervisor is helpful in getting the job done, the people you work with are helpful in getting the job done). Repeated measurements of these scales gave acceptable reliabilities, ranging from 0.68 to 0.86, and indicated that concept meanings were not changing over time [49]. Further, reliabilities of the mean of skill utilization, decision authority, physical and psychological demands across the 471 occupations of SOC-1991 were 0.88, 0.77, 0.88 and 0.56 respectively, which are quite similar to reliabilities of the complete version of the JCQ [50]. As for the other variables, hours worked were evaluated by summing the number of hours devoted to the main job and to other jobs (where applicable). Irregular work schedules was a dichotomous variable where 0 = normal shift and 1 = rotating, split, on call, other. 2.2.4. Family and social network Marital status distinguished between people living together as a couple (coded 1) and those with other marital situations (coded 0). Parental status was measured by the presence of children who lived with the respondent, grouped into three age categories: 5 years old and under, 611 years old, 1224 years old. Household income was determined using a ve-point ordinal scale (low/high) from Statistics Canada, which measured the level of income in relation to household size. Marital, parental, and economic status were measured for all cycles of the NPHS. Couple-related strain was based on an additive scale of three items (true/false)

428

A. Marchand and M. Blanc / The development of chronic psychological distress

developed by Wheaton [51]. Child-related strain was measured with an additive scale that included two items (true/false), also developed by Wheaton [51]. A reliability of 0.86 was estimated from repeated measurements of both couple-related strain and child-related strain. Social support outside the workplace was measured using an additive scale with four items that dealt with the presence of a condant; of someone to count on in a crisis situation; of someone to count on when making personal decisions; and of someone who makes one feel loved and cared for. Given its marked asymmetry, the scale was divided into two categories: low support (0 = 0, 1, 2, 3) and high support (1 = 4). 2.2.5. Personality of the agent Gender was a dichotomous variable coded 0 for men and 1 for women. Age was measured at the beginning of Cycle 1. Physical health status tallied the number of physical health problems affecting respondents using a list of 22 health problems (e.g., heart disease, cancer, arthritis). Self-esteem (alpha = 0.85) was measured by six ve-point items (disagreement/agreement) developed by Rosenberg [52]. The (internal) center of control (alpha = 0.76) was derived from the scale developed by Pearlin and Schooler [53], which included seven ve-point items (agreement/disagreement). The sense of coherence (alpha = 0.83) was based on the Antonovsky scale [54], which included 13 seven-point items (multiple choice). Tobacco use was calculated based on the number of cigarettes smoked per week. Physical activity was measured by the monthly frequency with which one or more physical activities were practiced for more than 15 minutes. Stressful childhood events quantied how many of seven items developed by Wheaton occurred before age 18 (e.g., two weeks or more spent in the hospital; parents divorcing; alcohol or drug abuse by parents) [51]. Table 1 presents summary descriptive data for each NPHS Cycle for the overall sample. 2.3. Analysis The data set had a hierarchical structure in which repeated measurements (level 1, n = 14,399) were nested in workers (level 2, n = 5,500), and workers were nested in their respective neighbourhoods (level 3, n = 1,300). Because the dependent variable was binary, we used repeated measures logistic multilevel regression models [55,56] to analyse the probability (risk) of developing repeated episodes of psychological distress over the eight-year period. Estimation using these

models does not require individuals be observed in each cycle of the survey. Assuming a random distribution of missing values, subjects not followed for all cycles afford less information for estimating the variation between levels of the data structure. Therefore, psychological distress is evaluated for individuals at work at each particular cycle of data. The general model used was of the form: Logit(ijk(t) ) = (t) + p0 Xpijk(t) +r0 Xrjk + 00k + 0 ijk(t) is the risk of a psychological distress episode i (level 1) for worker j (level 2) in neighbourhood k during the time interval t (t = 1..5). (t) represents model intercepts for each value of t (t = t1). p0 are slopes for independent worker variables varying with time (Xpij(t) ), and r0 are slopes for constant-time independent worker variables (Xrjk ). 00k and 0j are error terms for neighbourhood and workers, respectively. 00k and 0j are assumed to follow a normal dis2 2 tribution with variances and estimated from the data. The model accounts for the non-independence of the observations generated from repeated episodes of psychological distress in workers as well as for the clustering of observations at the neighbourhood level. The complex sampling design is thus taken into account. Parameters were estimated using the method of predictive quasi-likelihood (PQL) with a second-order Taylor expansion provided by the MlwiN 2.02 software [57]. Because data were weighted, robust sandwich estimators for standard errors were computed [56]. In all cases, the signicance of individual regression coefcients was evaluated using a bilateral Z test, and the probability of rejection of the null hypoth2 esis was set at P < 0.05. For random parameters ( , 2 ) Wald tests were applied with halved p-values (P < 0.05) [58]. The overall model contribution was evaluated with the Wald test. In the analysis, all independent variables were mean-centered for the period, with the exception of the binary variables. Occupations and work organisation conditions were analysed separately and jointly with other variables in order to evaluate the main effects and possible mediation-suppressive effects. Last, gender interactions with occupations and work organisation conditions were evaluated separately.

A. Marchand and M. Blanc / The development of chronic psychological distress Table 1 Descriptive statistics. NPHS 19941995/20022003, Canada, N = 5,500 Cycle 1 19941995 Mean/% Occupations Senior managers Directors Supervisors Professionals White-collar workers Blue-collar workers Workplace Skill utilization Decision authority Psychological demands Physical demands Social support Job insecurity Working hours (total) Work schedule (irregular) Family Marital status (couple) Child 05 Child 611 Child 1224 Family income Couple-related strain Child-related strain Social network Social support outside work (high) Agents Gender (female) Age Physical health status Self-esteem Control Cohesion Cigarettes (per day) Physical activities Stressful childhood events 1% 7% 4% 16% 49% 24% 7.09 5.38 4.53 2.07 7.98 1.55 43.35 23% 67% 22% 23% 32% 3.73 0.37 0.25 86% 45% 35.56 0.78 20.41 20.19 58.75 5.10 19.71 0.87 Cycle 2 19961997 Mean/% 1% 7% 5% 17% 47% 24% 7.09 5.38 4.53 2.07 7.98 1.55 44.22 21% 67% 21% 23% 30% 3.77 0.37 0.25 89% 45% 37.58 0.96 20.41 20.19 58.75 4.94 21.14 0.87 Cycle 3 19981999 Mean/% 1% 8% 5% 17% 46% 24% 7.09 5.38 4.53 2.07 7.98 1.55 45.12 21% 69% 19% 22% 30% 3.98 0.37 0.25 92% 45% 39.54 1.02 20.41 20.19 62.09 4.52 22.76 0.87 Cycle 4 20002001 Mean/% 1% 9% 6% 20% 41% 23% 7.18 5.42 4.25 1.86 7.93 1.62 42.48 19% 71% 17% 22% 28% 4.12 0.34 0.27 93% 45% 41.79 1.09 19.79 20.26 62.09 3.99 20.20 0.87 Cycle 5 20022003 Mean/% 1% 10% 6% 22% 40% 21% 7.16 5.45 4.18 1.81 7.91 1.75 42.29 20% 73% 15% 21% 27% 4.22 0.17 0.27 93% 45% 43.82 1.31 19.79 19.83 62.09 3.12 24.30 0.87

429

SD* 2.27 1.81 1.77 1.28 1.98 1.18 20.73 0.97 0.73 0.55 10.29 1.08 2.89 4.06 11.43 9.65 21.17 1.16

SD 2.27 1.81 1.77 1.28 1.98 1.18 19.57 0.91 0.73 0.55 10.26 1.20 2.89 4.06 11.43 9.28 21.78 1.16

SD 2.27 1.81 1.77 1.28 1.98 1.18 18.21 0.92 0.73 0.55 10.24 1.28 2.89 4.06 10.66 8.78 21.38 1.16

SD 1.99 1.47 1.57 1.12 1.61 1.08 14.04 0.90 0.72 0.56 10.24 1.31 2.69 3.72 10.66 8.26 18.86 1.16

SD 1.95 1.45 1.57 1.12 1.63 1.10 14.21 0.85 0.55 0.56 10.16 1.41 2.69 3.92 10.66 7.32 21.10 1.16

Notes: SD, standard deviation. Measurement ranges from low to high level for the following continuous variables: skill utilization, decision authority psychological demands, physical demands, social support, job insecurity, couple-related strains, child-related strains, physical health status, self-esteem, control, cohesion, physical activities, stressful childhood events.

3. Results Over the study period (1994/19952002/2003), 46.4% of workers experienced at least one episode of psychological distress, 23.5% had more than one episode, and 10.6% three or more. Overall, 23.0% had only one episode, 12.3% two, 6.1% three, 3.2% four, and 1.3% ve. Table 2 presents the results of the multilevel logistic regression analysis. Model 1 also showed signicant variations in psychological distress across workers and neighbourhoods. The computation of intraclass correlations attributed

35.9% and 5.6% of the logit variance to workers and neighbourhoods respectively. Model 2 revealed that the risk of psychological distress declined over time. The risk for Cycle 1 was 27.4%, 14.3% for Cycle 2, 14.4% for Cycle 3, 11.1% for Cycle 4, and 9.6% for Cycle 5. Model 3 estimated the effects of workplace variables and found skill utilisation, psychological demands, social support, and job insecurity were significant. Models 4 and 5 adjusted for family and social network. These variables did not change the contribution of workplace variables, with the sole exception of skill utilisation, which was no longer signicant when family variables were accounted for. Model 6 adjusted

430

A. Marchand and M. Blanc / The development of chronic psychological distress Table 2 Multilevel regression models of repeated episodes of psychological distress (Unstandardized regression coefcients). NPHS 1994 1995/20022003, Canada, N = 5,500 Constant Time Cycle 2 Cycle 3 Cycle 4 Cycle 5 Occupations Senior managers (ref) Directors Supervisors Professionals White-collar workers Blue-collar workers Workplace Skill utilization Decision authority Psychological demands Physical demands (high) Social support Job insecurity Working hours (total) Work schedule (irregular) Family Marital status (couple) Child 05 Child 611 Child 1224 Family income Couple-related strain Child-related strain Social network Social support outside work (high) Agents Gender (female) Age Physical health status Self-esteem Control Cohesion Cigarettes Physical activities Stressful childhood events Random part 2 Community 2 Individual Goodness-of-t 2 (df) Notes: P value < 0.05; P value < 0.01. 0.317 2.021 0.328 2.132 248.69 (4) 0.323 1.897 459.48 (17) 0.351 1.796 558.56 (24) 0.334 1.847 568.32 (18) Model 1 1.757 Model 2 0.973 0.814 0.811 1.107 1.270 Model 3 0.815 0.801 0.797 1.024 1.141 Model 4 0.013 0.755 0.661 0.898 0.694 Model 5 0.184 0.748 0.714 0.955 1.033 Model 6 4.876 0.822 0.399 0.605 0.566 Model 7 5.652 0.754 0.286 0.502 0.258

0.438 0.636 0.568 0.922 0.627 0.045 0.041 0.092 0.031 0.149 0.201 0.001 0.084

0.319 0.541 0.511 0.752 0.466 0.033 0.038 0.094 0.008 0.141 0.175 0.000 0.082 0.643 0.000 0.247 0.012 0.232 0.407 0.369

0.451 0.629 0.610 0.938 0.588 0.047 0.035 0.094 0.029 0.138 0.194 0.001 0.093

0.426 0.625 0.522 0.635 0.525 0.007 0.039 0.039 0.031 0.071 0.101 0.000 0.002

0.355 0.564 0.495 0.540 0.418 0.013 0.042 0.044 0.042 0.067 0.089 0.000 0.003 0.304 0.118 0.257 0.013 0.118 0.188 0.214

0.911 0.339 0.022 0.136 0.008 0.116 0.066 0.018 0.000 0.084 0.267 0.998 1169.58 (26)

0.571 0.362 0.028 0.125 0.009 0.107 0.062 0.016 0.000 0.068 0.287 0.987 1294.01 (34)

for agent personality and yielded substantial modications in workplace effects. Skill utilisation and psychological demands were no longer signicant, and effect sizes dropped sharply. Model 7 included all variables and revealed psychological demands, social support, and job insecurity

were still signicant. For family and social network, marital status, the presence of children 611 years old, family income, and couple-related and child-related strains, as well as social support outside the workplace, were signicant. Personality variables related to gender, age, physical health, locus of control, sense of co-

A. Marchand and M. Blanc / The development of chronic psychological distress Table 3 Odds Ratios and Condence Intervals for Model 7 of Table 2. NPHS 19941995/20022003, Canada, N = 5,500 OR Time Cycle 2 Cycle 3 Cycle 4 Cycle 5 Occupations Senior managers (ref) Directors Supervisors Professionals White-collar workers Blue-collar workers Workplace Skill utilization Decision authority Psychological demands Physical demands Social support Job insecurity Working hours (total) Work schedule (irregular) Family Marital status (couple) Child 05 Child 611 Child 1224 Family income Couple-related strain Child-related strain Social network Social support outside work (high) Agents Gender (female) Age Physical health status Self-esteem Control Cohesion Cigarettes Physical activities Stressful childhood events 0.47 0.75 0.61 0.77 95%CI 0.390.57 0.620.91 0.490.75 0.620.96

431

4. Discussion The goal of this study was to investigate the relationship between occupation, work organisation conditions, and the development of chronic psychological distress in Canadian workers. We used a multilevel theoretical model of mental health determinants in the workforce to hypothesise a role for stress promoted by constraints-resources embedded in macrosocial structures, structures of daily life, and agent personality. Our results suggest that certain work organisation conditions are associated with the outcome that go beyond the differential increments of distress associated with neighbourhoods, family situation, social support outside the workplace, and personal characteristics. However, individual and non-work factors appear to contribute more strongly to the development of chronic psychological distress. The results we obtained using the NPHS data set made it possible to estimate approximately 23% of workers experienced repeated episodes of psychological distress between 1994-1995 and 2002-2003. For more than 10% of workers, psychological distress was persistent at three occasion and more and thus seemed to manifest as a chronic condition. The risk of psychological distress varied considerably among individuals and decreased over time. Overall, the results support the multilevel theoretical model of mental health determinants in the workforce, since all components (workplace, family, neighbourhood, social network, personal characteristics) were linked to repeated episodes of psychological distress. Occupation per se did not appear to be signicantly related to the development of chronic psychological distress. Therefore, the major role of occupation reported in previous studies [4,5,7,8] was not conrm suggesting a limited role for occupation as an objective condition leading directly to psychological distress; a nding that is consistent with recent studies [2,6,9,15,59]. Still, workplace factors associated with work organisation conditions appear much more important than occupation. In keeping with the ndings of previous studies, psychological demands and job insecurity increased the risk of repeated episodes of psychological distress by 5% and 9% respectively for each additional point on the scales, whereas the level of social support given by coworkers and supervisors diminished the risk by 7%. The results also showed skill utilisation and decision authority were not signicantly related to psychological distress. These unexpected ndings refute those from previous studies that used more lim-

1.43 1.76 1.64 1.72 1.52 1.01 1.04 1.04 0.96 0.94 1.09 1.00 1.00 0.74 0.89 0.77 0.99 0.89 1.21 1.24 0.56 1.44 0.97 1.13 0.99 0.90 0.94 1.02 1.00 1.07

0.355.75 0.447.05 0.416.64 0.446.74 0.386.05 0.981.05 1.001.09 1.001.09 0.911.01 0.900.97 1.031.16 1.001.00 0.861.17 0.610.90 0.721.09 0.620.97 0.841.16 0.830.96 1.081.35 1.101.39 0.470.68 1.201.72 0.960.98 1.081.19 0.971.02 0.880.92 0.930.95 1.011.02 1.001.00 1.011.14

Notes: P value < 0.05; P value < 0.01.

herence, and stressful childhood life events maintained contributions. Overall, psychological distress still varied signicantly between workers and neighbourhoods. Model 7 explained 9.5% of the logit variance at the neighbourhood level and 51.2% at the worker level. Gender interactions were tested last and were found not signicant. Results give 2 = 6.14 df = 5 p = 0.30 for occupation and 2 = 7.83 df = 8 p = 0.45 for work organisation conditions. Table 3 reports odds ratios and 95% condence intervals for Model 7 as reported in Table 2.

432

A. Marchand and M. Blanc / The development of chronic psychological distress

ited conceptual frameworks and reported less psychological distress with high skill utilisation and decision authority [5,13,22,27,29,6064]. In addition to work factors, neighbourhood, family situation, social support outside the workplace, and individual characteristics (agent personality) had an important impact upon the risk of repeated episodes of psychological distress in this cohort of Canadian workers. These elements modied the strength of the associations, and family situation and individual characteristics mediated the inuence of skill utilisation. Neighbourhood accounted for 6% of the variance in repeated episodes of psychological distress. This interesting result shows that place matters, and gives support to studies of civil servant and assistant nurses reporting neighbourhoods with low socioeconomic status were associated with individuals experiencing more psychological distress and depressive symptoms [65,66]. Research into how neighbourhood living conditions provide constraints-resources that help or hinder the stress levels faced by workers on a day-today basis and how they relate to the promotion of mental health problems is needed. Marital status, parental status, economic status, and social support outside the workplace decreased the risk of the outcome by between 13% and 77%, whereas marital and child-related strains increased by 21% and 24% respectively the risk of repeated episodes of psychological distress. For agent personality, women were at higher risk for repeated episodes (44%), but gender did not exert a moderating effect on occupation, work organisation conditions, and distress. Each year of increased age reduced the risk by 3%, whereas the number of physical health problems was associated with a 13% rise in the risk of repeated episodes of psychological distress. As for psychological traits, each additional point increase on the internal locus of control and sense of cohesion scales led to 11% and 6% reductions in the development of chronic psychological distress. The impact of these two psychological traits opens the door for interventions directed toward individuals manifesting low internal control and low sense of cohesion. Finally, each cigarette smoked increases the risk of repeated psychological distress by 2%. Stressful childhood life events increase it by 7%. Overall, only physical activities and self-esteem were not signicantly related to the outcome. The limitations of this study should be considered carefully when interpreting the results. First, the relationships are not causal because the design is correlational. However, both dependant and independent

variables are allowed to vary over time. Second, selfreport measures were used for both predictor and outcome variables. Third, the scales for skill utilisation, decision authority, psychological demands, and social support in the workplace have low internal consistency in Statistics Canadas NPHS [2], and the psychological demands scale was found to have questionable validity [67], implying the impact of these variables could have been underestimated. However, both psychological demands and social support were signicant here. Also, even with the full version of the Karaseks JCQ, skill utilisation and decision authority were not found signicant when the analysis included the family situation, the social network outside the workplace and a large spectrum of individual characteristics [15]. Fourth, the two-year hiatus between NPHS cycles leads to estimation errors for variations in psychological distress over time. In the present study, large variations in psychological distress over time were found, which suggests the need to examine this discrepancy further using shortened time intervals to better capture its dynamics. Fifth, the NPHS does not take into account workplace factors related to the physical environment (e.g., dust, noise, cold, heat, toxins), management and supervisory styles, health and safety resources, or elements in the work contract that allow employees to better balance work and family responsibilities. These elements could be strong determinants of quality of life and well-being at work and could be associated with the development of chronic psychological distress. Finally, while we did not nd any interaction between work variables and gender, other interactions between work and non work variables and work and personal variables have not been evaluated. In summary, this study found psychological demands, social support at work, and job insecurity were signicant workplace risk factors for the development of chronic psychological distress in workers. The effects remained after accounting for factors related to neighbourhood, family situation, social network outside work, and agent personality. Furthermore, the results presented here clearly show the contributions of individual factors and factors outside work. Future studies should incorporate these elements into analyses to arrive at a better understanding of how work promotes the development of chronic psychological distress. Further research in the eld of occupational mental health must expand theoretical and empirical frameworks to capture actual workplace dynamics and ultimately the effects on workers mental health. It also means interventions designed to reduce mental health

A. Marchand and M. Blanc / The development of chronic psychological distress

433

problems should carefully evaluate not only the work environment itself, but also outside work factors and other individual characteristics as they could be the principal determinants of problems faced by the worker. Psychological distress represents a signicant human and economic burden for both businesses and the whole of society [3,68]. Psychological distress appears persistent for non-trivial proportion of workers. The development of public health policies and intervention directed toward the regulation of mental health problems in the workplaces are undoubtedly needed and are surely one of the most important challenges ahead in the coming years. Acknowledgments This study was supported by a grant from the Social Sciences and Humanities Research Council of Canada (SSHRC, project no. 410-2006-2067). The authors wish to thank Statistics Canada for facilitating access to data, Nancy Beauregard for assisting with the literature review, and Robert Sullivan for revising the English. References
[1] [2] World Health Organization, The Global Burden of Disease 2004 Update. Geneva: World Health Organization, 2008. A. Marchand, A. Demers and P. Durand, Do occupation and work conditions really matter? A longitudinal analysis of psychological distress experiences among Canadian workers, Sociol Health Illn 27 (2005), 602627. K.L. Lim, P. Jacobs, A. Ohinmaa, D. Schopocher and C.S. Dewa, A new population-based measure of the economic burden of mental illness in Canada, Chronic Dis Can 28 (2008), 9298. A. Marchand, P. Durand and A. Demers, Work and Mental Health: The experience of the Quebec workforce between 1987 and 1998, Work 25 (2005), 135142. I. Niedhammer, M. Goldberg, A. Leclerc, I. Bugel and S. David, Psychosocial factors at work and subsequent depressive symptoms in the Gazel cohort, Scand J Work Environ Health 24 (1998), 197205. U. Bultmann, I. Kant, L.G. van Amelsvoort, P.A. van den Brandt and S.V. Kasl, Differences in fatigue and psychological distress across occupations: results from the Maastricht Cohort Study of Fatigue at Work, Journal of Occupational and Environmental Medicine 43 (2001), 976983. J.W. Grosch and L.R. Murphy, Occupational differences in depression and global health: results from a national sample of US workers, J Occup Environ Med 40 (1998), 153164. B. Sanne, A. Mykletun, A.A. Dahl, B.E. Moen and G.S. Tell, Occupational differences in levels of Anxiety and depression: The Hordaland health study, J Occup Environ Med 45 (2003), 628638.

[9]

[10] [11]

[12]

[13]

[14] [15]

[16] [17] [18] [19] [20]

[3]

[21] [22]

[4] [5]

[23]

[6]

[24] [25] [26]

[7] [8]

J. Wieclaw, E. Agerbo, P.B. Mortensen and J.P. Bonde, Occupational risk of affective and stress-related disorders in the Danish workforce, Scand J Work Environ Health 31 (2005), 343351. K. Wilhelm, V. Kovess, C. Rios-Seidel and A. Finch, Work and mental health, Soc Psychiatry Psychiatr Epidemiol 39 (2004), 866873. R. Fuhrer, S.A. Stansfeld, J. Chemali and M.J. Shipley, Gender, social relations and mental health: prospective ndings from an occupational cohort (Whitehall study II), Soc Sci Med 48 (1999), 7787. S.A. Stansfeld, J. Head, R. Fuhrer, J. Wardle and V. Cattell, Social inequalities in depressive symptoms and physical functioning in the Whitehall II study: exploring a common cause explanation, J Epidemiol Community Health 57 (2003), 361 367. S. Paterniti, I. Niedhammer, T. Lang and S.M. Consoli, Psychosocial factors at work, personality traits and depressive symptoms: Longitudinal results from the GAZEL Study, Br J Psychiatry 181 (2002), 111117. S. Stansfeld, Work, personality and mental health, Br J Psychiatry 181 (2002), 9698. A. Marchand, A. Demers and P. Durand, Does work really cause distress? The contribution of occupational structure and work organization to the experience of psychological distress, Soc Sci Med 60 (2005), 114. R.A. Karasek, Job Demands, Job Decision Latitude, and Mental Strain: Implication for Job Redesign, Adm Sci Q 24 (1979), 285309. R.A. Karasek and T. Theorell. Healthy work: stress, productivity, and the reconstruction of the working life. New York: Basic Books, 1990. J.P.E. Bonde, Psychosocial factors at work and risk of depression: a systematic review of the epidemiological evidence, Occup Environ Med 65 (2008), 438445. S. Stansfeld and B. Candy, Psychosocial work environment and mental health-a meta-analytic review, Scand J Work Environ Health 32 (2006), 443462. M. Egan, C. Bambra, S. Thomas, M. Petticrew, M. Whitehead and H. Thomson, The psychosocial and health effects of workplace reorganisation. 1. A systematic review of organisationallevel interventions that aim to increase employee control, J Epidemiol Community Health 61 (2007), 945954. R. Bourbonnais, M. Comeau and M. Vezina, Job strain and evolution of mental health among nurses, J Occup Health Psychol 4 (1999), 95107. H. Mausner-Dorsch and W.W. Eaton, Psychosocial work environment and depression: epidemiologic assessment of the demand-control model, Am J Public Health 90 (2000), 1765 1770. M. Van der Doef and S. Maes, The Job Demand-Control (Support) Model and psychological well-being: a review of 20 years of empirical research, Work & Stress 13 (1999), 87 114. M. Vermeulen and C. Mustard, Gender differences in job strain, social support at work, and psychological distress, J Occup Health Psychol 5 (2000), 428440. K.-O. Park, Social support for stress prevention in hospital settings, J R Soc Promot Health 127 (2007), 260264. E. Clays, D. De Bacquer, F. Leynen, M. Kornitzer, F. Kittel and G. De Backer, Job stress and depression symptoms in middle-aged workersprospective results from the Belstress study, Scand J Work Environ Health 33 (2007), 252259.

434 [27]

A. Marchand and M. Blanc / The development of chronic psychological distress Mental illness in the United States: epidemiological estimates, al. DBe, ed., Praeger, New York, 1980, pp. 114132. L. Robins and D. Regier, editors. Psychiatric disorders in America: the epidemiologic catchment area study. New York: The free press, 1991. M. Tousignant, editor. Les origines sociales et culturelles des troubles psychologiques. Paris: Presses universitaires de France, 1992. Statistique Canada, editor. Classication type des professions 1991. Ottawa: Statistique Canada. Division des normes, 1993. P.C. Pineo, J. Porter and H.A. McRoberts, The 1971 Census and the Socioeconomic Classication of Occupations, Canad Rev Sociol Anthrop 14 (1977), 91102. R. Karasek, G. Gordon, C. Pietrokovsky, M. Frese, C. Pieper, J. Schwartz et al., Job Content Questionnaire: Questionnaire and Users Guide. Lowell: University of Massachusetts, 1985. E.G. Carmines and R.A. Zeller, Reliability and Validity Assessment. Beverly Hills, CA: Sage, 1979. A. Marchand, A. Demers, P. Durand and M. Simard, The moderating effect of alcohol intake on the relationship between work strains and psychological distress, J Stud Alcohol 64 (2003), 419427. B. Wheaton, Sampling the stress universe, in: Stress and Mental Health Contemporary Issues and Prospects for the Future, W.R. Avison and I.H. Gotlib, eds, Plenum Press, New York, 1994, pp. 77114. M. Rosenberg, Conceiving the Self, New York: Basic Books, 1979. L.I. Pearlin and C. Schooler, The Structure of Coping, J Health Soc Behav 19 (1978), 221. A.G. Antonovsky, Unraveling the Mystery of Health: How People Manage Stress and Stay Well, San Francisco: JosseyBass, 1987. M.H. Boyle and J.D. Willms, Multilevel modelling of hierarchical data in developmental studies, J Child Psychol Psychiatry 42 (2001), 141162. H. Goldstein, Multilevel Statistical Models, London, New York: Edward Arnold, Halstead Press, 1995. J. Rasbash, W. Brown, H. Goldtein, M. Yang, I. Plewis, M. Healey et al., A users guide to MlwiN. Version 2.1a. London: Multilevel Models Project, Institute of Education, University of London, 2000. T. Snijders, A.B and R. Bosker, J. Multilevel Analysis. An introduction to basic and advanced multilevel modeling. London: Sage Publications, 1999. C. Cohidon, G. Santin, E. Imbernon and M. Goldberg, Working conditions and depressive symptoms in the 200 decennial health survey: the role of the occupational category, Soc Psychiatry Psychiatr Epidemiol (2009), S.A. Stansfeld, R. Fuhrer, M.J. Shipley and M.G. Marmot, Work characteristics predict psychiatric disorder: prospective results from the Whitehall II study, Occup Environ Med 56 (1999), 302307. M. Van der Doef, S. Maes and R. Diekstra, An examination of the job demand-control-support model with various occupational strain indicators, Anxiety, Stress & Coping: an International Journal 13 (2000), 165185. U. Bultmann, I.J. Kant, P.A. Van den Brandt and S.V. Kasl, Psychosocial work characteristics as risk factors for the onset of fatigue and psychological distress: prospective results from the Maastricht Cohort Study, Psychol Med 32 (2002), 333 345. J.M. Grifn, R. Fuhrer, S.A. Stansfeld and M. Marmot, The importance of low control at work and home on depression

I. Niedhammer, J.-F. Chastang, S. David, L. Barouhiel and G. Barrandon, Psychosocial Work Environment and Mental Health: Job-strain and Effort-Reward Imbalance Models in a Context of Major Organizational Changes, Int J Occup Environ Health 12 (2006), 111119. [28] A.H. de Lange, T.W. Taris, M.A. Kompier, I.L. Houtman and P.M. Bongers, The very best of the millennium: longitudinal research and the demand-control-(support) model, J Occup Health Psychol 8 (2003), 282305. [29] J. de Jonge, H. Bosma, R. Peter and J. Siegrist, Job strain, effort-reward imbalance and employee well-being: a largescale cross-sectional study, Soc Sci Med 50 (2000), 1317 1327. [30] J. Siegrist, Adverse health effects of high-effort/low-reward conditions, J Occup Health Psychol 1 (1996), 2741. [31] H. Pikhart, M. Bobak, A. Pajak, S. Malyutina, R. Kubinova, R. Topor et al., Psychosocial factors at work and depression in three countries of Central and Eastern Europe, Soc Sci Med 58 (2004), 14751482. [32] A. Tsutsumi, K. Kayaba, T. Theorell and J. Siegrist, Association between job stress and depression among Japanese employees threatened by job loss in a comparison between two complementary job-stress models, Scandinavian Journal of Work and Environmental Health 27 (2001), 146153. [33] I. Godin, F. Kittel, Y. Coppieters and J. Siegrist, A prospective study of cumulative job stress in relation to mental health, BioMed Central Public Health 5 (2005), 67. [34] A. Shimazu and J. de Jonge, Reciprocal relations between effort-reward imbalance at work and adverse health: A threewave panel survey, Soc Sci Med 68 (2008), 6068. [35] N. Dragano, Y. He, S. Moebus, K.-H. J ckel, R. Erbel, J. o Siegrist et al., Two models of jobs stress and depressive symptoms. Results from a population-based study, Soc Psychiatry Psychiatr Epidemiol 43 (2008), 7278. [36] B.L. Janzen, N. Muhajarine, T. Zhu and I.W. Kelly, Effortreward imbalance, overcommitment, and psychological distress in Canadian police ofcers, Psychol Rep 100 (2007), 525530. [37] R. Bourbonnais, C. Brisson, A. Vinet, M. Vezina and A. Lower, Development and implementation of a participative intervention to improve the psychosocial work environment and mental health in an acute care hospital, Occup Environ Med 63 (2006), 326334. [38] N. van Vegchel, J. de Jonge, H. Bosma and W. Schaufeli, Reviewing the effort-reward imbalance model: drawing up the balance of 45 empirical studies, Soc Sci Med 60 (2005), 11171131. [39] R.C. Kessler, G. Andrews, L.J. Colpe, E. Hiripi, D.K. Mroczek, S.L.T. Normand et al., Short screening scales to monitor population prevalences and trends in non-specic psychological distress, Psychol Med 32 (2002), 959976. [40] World Health Organization, editor. Composite International Diagnostic Interview. Version 1.0. Geneva: World Health Organization, 1990. [41] C. Daveluy, L. Pica, N. Audet, R. Courtemanche, F. Lapointe, L. C t et al., Enqu te sociale et de sant 1998. Qu bec: oe e e e Institut de la statistique du Qu bec, 2000. e [42] G. Kaplan, R. Roberts, T. Camacho and J. Coyne, Psychosocial predictors of depression: prospective evidence from the human population laboratory studies, Am J Epidemiol 125 (1987), 206220. [43] B. Link and B. Dohrenwend, Formulation of hypotheses about the true prevalence of demoralization in the United States, in:

[44] [45] [46] [47] [48] [49] [50]

[51]

[52] [53] [54] [55] [56] [57]

[58] [59]

[60]

[61]

[62]

[63]

A. Marchand and M. Blanc / The development of chronic psychological distress and anxiety: do these effects vary by gender and social class? Soc Sci Med 54 (2002), 783798. M. Melchior, A. Caspi, B.J. Milne, A. Danese, R. Poulton and T.E. Moftt, Work stress precipitates depression and anxiety in young, working women and men, Psychol Med 37 (2007), 11191129. M. Stafford and M. Marmot, Neighbourhood deprivation and health: does it affect us all equally? Int J Epidemiol 32 (2003), 357366. C. Muntaner, Y. Li, X. Xue, T. Thompson, H. Chung and P. Ocampo, County and organizational predictors of depression

435

[64]

[67]

[65] [66]

[68]

symptoms among low-income nursing assistants in the USA, Soc Sci Med 63 (2006), 14541465. C. Brisson and B. Larocque, Validit des indices de dee mande psychologique et de latitude d cisionnelle utilis s dans e e lEnqu te nationale sur la sant de la population (ENSP) 1994e e 1995, Can J Public Health 92 (2001), 468474. U. Bultmann, R. Rugulies, T. Lund, K.B. Christensen, M. Labriola and H. Burr, Depressive symptoms and the risk of long-term sickness absence. A prospective study among 4747 employees in Denmark, Soc Psychiatry Psychiatr Epidemiol 41 (2006), 875880.

Copyright of Work is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.