Skip Navigation LinksHome > July 2012 - Volume 54 - Issue 7 > Relationship Between Long Working Hours and Depression in Tw...
Journal of Occupational & Environmental Medicine:
doi: 10.1097/JOM.0b013e318250ca00
Original Articles

Relationship Between Long Working Hours and Depression in Two Working Populations: A Structural Equation Model Approach

Amagasa, Takashi MD, MPH; Nakayama, Takeo MD, PhD

Free Access
Article Outline
Collapse Box

Author Information

From the Department of Health Informatics (Drs Amagasa and Nakayama), Kyoto University School of Public Health, Konoe-cho, Yoshida, Sakyo-ku, Kyoto; and Department of Psychiatry (Dr Amagasa), Yoyogi Hospital, Sendagaya, Shibuya-ku, Tokyo, Japan.

Address correspondence to: Takashi Amagasa, MD, MPH, Department of Health Informatics, Kyoto University School of Public Health, Konoe-cho, Yoshida, Sakyo-ku, Kyoto 606-8501, Japan (amagasa_t@tokyo-kinikai.com).

The authors declare no conflicts of interest.

Collapse Box

Abstract

Objectives: To test the hypothesis that relationship reported between long working hours and depression was inconsistent in previous studies because job demand was treated as a confounder.

Methods: Structural equation modeling was used to construct five models, using work-related factors and depressive mood scale obtained from 218 clerical workers, to test for goodness of fit and was externally validated with data obtained from 1160 sales workers. Multiple logistic regression analysis was also performed.

Results: The model that showed that long working hours increased depression risk when job demand was regarded as an intermediate variable was the best fitted model (goodness-of-fit index/root-mean-square error of approximation: 0.981 to 0.996/0.042 to 0.044). The odds ratio for depression risk with work that was high demand and 60 hours or more per week was estimated at 2 to 4 versus work that was low demand and less than 60 hours per week.

Conclusions: Long working hours increased depression risk, with job demand being an intermediate variable.

In 1868, an 8-hour workday legislation was introduced in the United States for federal employees, and in 1919, the importance of the standard 8-hour workday was declared by the First Convention of the International Labour Organization. Nevertheless, long working hours remain an important issue for workers. In many developed countries, in particular, major depressive disorder is a pressing issue in occupational health.1 In Japan, where the International Labour Organization's First Convention was not ratified, suicides attributed to overwork have been a serious social problem since the late 1990s.2 Previous studies showed that more than 90% of workers who committed suicide had a mental illness or mood disorder, particularly major depressive disorder.3,4

On the basis of the job demand–control model of Karasek,5 a large body of research has shown that high job demand, low job control, or both increase risks associated with various aspects of physical and mental health.613 Nevertheless, a recent review indicated that only high-quality studies supported the hypothesis that high job strain increased the risk of psychological distress and the number of health complaints.12 There is mounting evidence that the combination of high job demand and low job control increases the risk of depression/depressive state.814 Furthermore, the effort–reward imbalance model15 and bullying16,17 have been verified as psychological risk factors of depression in workplaces.

Several studies have reported the impact of long working hours on various aspects of health,1823 such as mortality, cardiovascular disease, hypertension, diabetes, sickness absence, and subjective ill health. Some studies showed that long working hours have a negative effect on mental health1823; results of meta-analyses indicated small but statistically significant correlations between psychological health measures and working hours.19 Moreover, two of three studies found associations between extended hours and depression.20 Nevertheless, only one cohort study of women in Canada reported that long working hours increase the risk of major depressive disorder.23 As such, studies investigating the relationship between working hours and major depressive disorder/depressive state have produced inconsistent results.18,20,24 Some groups have pointed out that long working hours may have an indirect effect on health or through different pathways.1820 van der Hulst and Geurts21 suggested that the association between overtime work and psychological health may depend on working conditions, such as job rewards and company pressure to work long hours. Hobson and Beach25 suggested that if work stress factors, including overwork, are considered intermediate variables involved in psychological health impairment, overadjustment may occur if multivariate analyses are performed with the intermediate variable treated as a confounding factor. To investigate the causal relationship between exposure (long working hours) and outcome (depression), job demand should be treated as an intermediate variable rather than a confounding variable, as suggested by Rothman,26 if job demand is a consequence of the number of hours worked.

This study aimed to determine whether job demand should be treated as a confounding variable or intermediate variable to clarify the relationship between long working hours and depression. We constructed and tested multiple models based on structural equation modeling (SEM), also known as covariance structure analysis, on two working populations with different characteristics such as age, gender, and job type. One group was analyzed to test our hypothesis, and the other was used to check the external validity of our hypothesis.

Back to Top | Article Outline

MATERIALS AND METHODS

Populations and Design
Research Design

A cross-sectional study was conducted using existing unlinkable, anonymized data from self-administered questionnaires. Data sets were made available for potential research uses with the permission of each business.

Back to Top | Article Outline
Research Materials

Data were obtained from two businesses. The lead author (T.A.) was directly involved in carrying out the surveys. The sample from business A was used to determine the goodness of fit of five models. The sample from business B was used to check external validity of the models. Subjects of business A were all regular full-time clerical workers in Saitama Prefecture within the metropolitan area of Japan. Business A adopted the lifetime employment system with a seniority-based pay system, which is common among Japanese companies. Business B, on the contrary, adopted the lifetime employment system with a merit-based compensation system. Business B primarily conducts sales operations with branches throughout the country. Business B included both full-time and part-time workers, with more female workers than those in business A. Accordingly, the study population of business B was more heterogeneous than that of business A.

Back to Top | Article Outline
Business A

The occupational safety and health committee of business A, which operates branches throughout Saitama Prefecture, consulted the lead author regarding the implementation of a survey to improve the mental health of full-time clerical workers. The data sets analyzed in this study were obtained from this survey.

Self-administered questionnaires were distributed through the general affairs department of business A in December 1999.27 The general affairs department instructed each branch manager to remind employees repeatedly to complete the questionnaire. As a result, 218 workers with no sickness or other absences completed the questionnaire within a week (response rate, 96.9%), and the manager of each branch collected the sealed questionnaires. Questionnaire data were entered into a computer by the general affairs department. After anonymization, data were sent to the lead author (T.A.), who tallied the data and reported the results to the general affairs department.

Back to Top | Article Outline
Business B

The survey was implemented to understand factors that influenced work stress and workers' mental health with the gradual introduction of a merit-based compensation system. As in the case of business A, the lead author (T.A.) helped to design the survey, summarizing the data and making recommendations to each worksite and respondent. This survey was conducted between December 2007 and January 2008.

Subjects were workers at 26 branches whose branch managers agreed to participate in the survey. Branch managers solicited cooperation from the 28,072 sales workers (8007 full-time and 20,065 part-time workers), and 1160 workers (420 full-time and 699 part-time) responded; the response rate was 4.1% (5.3% and 3.5%, respectively). Therefore, this sample (based on convenient sampling) was not representative of the business B working population.

Back to Top | Article Outline
Models Tested

After a review of previous research, author T.A., a clinical psychiatrist with experience in identifying cases of suicide due to overwork, and T.N., an epidemiologist, constructed the following five models (Fig. 1):

Figure 1
Figure 1
Image Tools

Model 1 (multiple regression model) corresponded to linear models commonly used in previous studies. In this model, bidirectional arrows are drawn among job demand, job control, and work hours and from each of these factors to depression.

Model 2 (long working hours, parallel model) is a model in which paths are drawn from job demand, job control, and work hours to depression.

Model 3 (long working hours, overwork model) is a model in which a path is drawn from work hours to job demand, from job demand to depression, and from job control to depression.

Model 4 (long hour labor condition [LHLC] model) is a model in which a path is drawn from the constructive concept LHLC, which consists of survey items associated with work hours, to depression and from job control to depression.

Model 5 (LHLC overwork model) is a model in which a path is drawn from LHLC to depression, with job demand as an intermediate variable, and from job control to depression.

Back to Top | Article Outline
Data Anonymization

For both surveys, questionnaires were distributed from the survey office to subjects by their supervisors through the occupational safety and health committee of each business. Completed questionnaires were sealed by subjects, collected by supervisors, and returned to the survey office through the occupational safety and health committees.27 Data entry was performed and double-checked by the survey office. The lead author (T.A.) analyzed the anonymized data according to a predetermined plan, and results were reported to the survey office. Original questionnaires were later destroyed by the survey office. With permission from each business to use data for research purposes, the lead author (T.A.) retained the anonymized electronic data.

Back to Top | Article Outline
Measurements Used for Models
Hours Worked

At business A, hours worked per week, actual number of nonworking days per month, and number of days working overtime per month were measured with a scale that was used in a large-scale cohort study of Japanese businesses.28 Subjects were asked to answer the following questions: “About how many hours on average do you work in a week? (if you worked holidays/weekends or overtime, please include them regardless of whether or not you were paid)”; “Averaging the past 3 months, how many actual nonworking days have you had per month?”; and, “In a busy month, what is the number of days worked overtime excluding support group activities (including holiday/weekend work)?” At business B, hours worked per month, number of paid overtime hours, and unpaid overtime were measured with the following questions: “On average, what is the number of hours you worked per month over the past 1 year (including the so-called ‘unpaid overtime' as well as work at home)?”; “How do you feel about your current hours worked?”; and, “Do you ever perform the so-called ‘unpaid overtime' at your workplace?” (Table 1).

TABLE 1-a. Descripti...
TABLE 1-a. Descripti...
Image Tools
Back to Top | Article Outline
Job Demands and Control
TABLE 1-b. Descripti...
TABLE 1-b. Descripti...
Image Tools

At business A, job demand, control, and support were measured using the Japanese translation of the Job Content Questionnaire measurement scale based on the extended Karasek model,814 which was used in the previous Japanese study.28,29 Reliabilities (Cronbach α) for job demand (eg, “I feel that the amount of work is too large”: 1 = “usually,” 2 = “sometimes,” 3 = “never”) and job control (eg, “Does your work allow you to make independent decisions?”: 1 = “hardly,” 2 = “not so much,” 3 = “to some degree,” 4 = “freely”) were 0.85 and 0.58, respectively. At business B, job demand, control, and support were measured using the Brief Job Stress Questionnaire, which was developed on the basis of the Japanese version of Job Content Questionnaire and established as reliable and valid among Japanese workers through a nationwide survey sponsored by the Ministry of Health, Labour, and Welfare.30 The values of Cronbach α for job demand and control of this scale were both 0.72.

Back to Top | Article Outline
Depression

At business A, the General Well-being Scale was used as an index of depression.31 The reliability, validity, and acceptability of the Japanese version of this 18-item General Well-being Scale are established. Depressive symptoms were measured by a subscale “Depressive mood” consisting of the 6 items with a total score of 0 to 30 (eg, “Have you been bothered by nervousness or your ‘nerves' this month?”: 1 = “severely so,” 6 = “not at all”). The Cronbach α for this scale was 0.86. At business B, depression was measured using the Japanese version of the Center for Epidemiologic Studies Depression Scale, the reliability and validity of which have been established by Shima et al32 (eg, “In the past week, I did not eat or my appetite decreased”: 1 = “none,” 2 = “1 or 2 days,” 3 = “3 or 4 days,” 4 = “more than 5 days”). The Cronbach α for this scale was 0.85.

Back to Top | Article Outline
Statistical Analysis

Data analyses were carried out using Microsoft Excel 2007, Amos Ver17.0, and Statics 17.0.33 After descriptive analyses, relationships among variables were analyzed using SEM.3335 Following the primary analysis of business A, the data sets from business B were examined to establish the external validity of each model. Multivariate logistic analysis incorporating work hours × job demand was performed in the case that model 2 was verified. Structural equation modeling was chosen because (1) it is capable of modeling both proximal and distal factors, both of which were necessary given the present study aim, whereas a conventional regression procedure might overlook the potential importance of the relationship of the distal factor (long working hours) with regard to the outcome (depression); (2) it can be used to test and compare competing models for relationships between various observed and latent variables because it can use several indices to judge the fit of our models, including the goodness-of-fit index (GFI) and root-mean-square error of approximation (RMSEA); (3) it can help to provide the causal direction of particular pathways for this type of cross-sectional research; and (4) observed variables showing similar trends due to the introduction of unobserved latent variables can be aggregated.12,13,34,35 As a result, better models can be constructed and the complexity of various variables can be more fully understood. Complete data and data sets without missing values were used for the analysis. The following criteria for model fitness were used in adopting a model: GFI more than 0.90, RMSEA less than 0.05, and P < 0.05 for estimates of standardized coefficients. For multivariate logistic analyses regarding business A data, depressive mood was binarized as a dependent variable by coding scores 10 or less (corresponding to 5% at the high end of depressive mood) as 1 and all other scores as 0. Job control was coded as 0 for scores above the mean and 1 for all other scores, and job demand was coded as 0 for scores below the mean and 1 for all other scores. Hours worked per week was initially coded as 0 for less than 60 hours, 1 for 60 hours or more, and then further coded as 3 for 60 hours or more and high job demand, 2 for 60 hours or more with low job demand, 1 for less than 60 hours with high job demand, and 0 for less than 60 hours and low job demand. These factors were incorporated into the model as independent variables. Center for Epidemiologic Studies Depression Scale data at business B were binarized as dependent variables, with scores of 16 or more (cutoff for screening occupational fields for Japanese data) coded as 1 and all others as 0.32 For hours worked per month, 240 hours or more was coded as 1 and all others as 0. Data were analyzed in the same way as in business A. In these analyses, α was set at 0.05.

Back to Top | Article Outline
Ethics Committee Approval

This study was approved by the medical ethics committee of Kyoto University (no. E-673).

Back to Top | Article Outline

RESULTS

Descriptive Statistics

Women accounted for about 20% of the 218 subjects from business A. Approximately 56% of subjects worked 60 hours or more per week. In contrast, women comprised a larger proportion of subjects from business B, and only 4.40% of respondents worked 240 hours or more per month, corresponding to 60 hours or more per week (Table 1).

Back to Top | Article Outline
Models Constructed From Observed Variables Alone: Models 1, 2, and 3

Results of SEM analysis of the five models are shown in Table 2. For business A, model 1 was rejected on the basis of the model fitness and P values for estimates of standardized coefficients. When business A data were used, the GFI/RMSEA was 1.000/0.172 for model 1, 0.400/0.234 for model 2, and 0.921/0.085 for model 3, respectively. Although model 3 was rejected for business A data (RMSEA >0.05), it was accepted for business B data (Table 2).

Table 2
Table 2
Image Tools
Back to Top | Article Outline
Models Incorporating the LHLC Constructive Concept: Models 4 and 5

To improve the model for business A data, model 4 incorporated the constructive concept LHLC, which consisted of survey items relating to hours worked per week, number of actual nonworking days, and number of days with overtime. The model was accepted on the basis of acceptance criteria (Table 2). The GFI was further improved for model 5, which treated job demand as an intermediate variable. For business B data, incorporating the LHLC constructive concept improved the GFI but not RMSEA.

Back to Top | Article Outline
Multivariate Logistic Analyses Suggested by Model 3

On the basis of results of testing model 3 with both data sets, multivariate logistic analysis was carried out with job control and long hours as explanatory variables and depression as the dependent variable (Table 3). Although the difference was not statistically significant for business A data, the odds ratio (OR) increased as overwork increased. On the contrary, the OR was significantly higher for less than 240 hours with high demand and 240 hours or more with high demand than that for less than 240 hours with low demand for business B. For both data sets, results of job control were consistent with previous studies.

Table 3
Table 3
Image Tools
Back to Top | Article Outline

DISCUSSION

Results of SEM supported our hypothesis that treating “job demand” as an intermediate variable rather than a confounding factor clarified the positive relationship between long working hours and the risk of depression.18,20,24 Using data obtained from employees of two businesses, model 1, which corresponds to the traditional model that treated job demand as a confounding factor, did not meet the acceptance criteria. Model fitness was improved for models that treated job demand as an intermediate variable (models 3 and 5).

For business A data, the best fit model was model 5, in which the constructive concept of LHLC was incorporated and job demand was treated as an intermediate variable. Nevertheless, for business B, the best fit models were those with hours worked monthly incorporated as an observed variable and job demand as an intermediate variable.

Terms related to long working hours, such as overtime, overtime work, overwork, and working long hours, are used with different meanings and definitions. For instance, the mean commuting time for workers residing in the Tokyo metropolitan area is approximately 1 hour. These workers may take commuting time into account when reporting long hours in the workplace. In the present study, data from two businesses were analyzed using a construct (LHLC) as a proxy. The work-related factors measured by LHLC differed between the two data sets, probably due to different characteristics of the two businesses. This suggests that the country, region, or business may influence the items and methods needed to measure accurately long working hours, as defined by the workers, and may reflect the more homogeneous working population in business A than in business B. Working hours were also measured differently between the two businesses. In business A, hours were presented as working hours per week, whereas in business B, hours were presented as average working hours per month over the past year. This would suggest that average working hours over a longer period may be a parameter that more clearly relates to depression and suggests that the definition of working hours adopted is crucial for assessing its effects (eg, on depression). In addition, the model in which job demand was associated with depression through LHLC, and which treated hours worked as an intermediate variable, was tested but statistically rejected.

Given that long working hours do not directly increase the risk of depression but have an effect through increased job demand, the variable “long hours overworked,” consisting of hours worked and job demand, was analyzed with a multivariate logistic regression model. The results suggested that the OR for depression was 2 to 4 times higher with long hours and high job demand than with short hours and low job demand. Thus, the combination of long hours and high job demand was moderately associated with depression.

Some differences were observed between the two businesses from which we obtained data. Jobs in business A were mainly clerical, whereas those in business B were mainly sales related. The male:female ratio in workplaces in Japan is 6:4,36 and about 30% of employees worked 50 hours or more in 2000.37 Nevertheless, the proportion of female employees was relatively small in business A and many worked long hours. In contrast, business B had a larger proportion of female workers and fewer employees working long hours. Nevertheless, results were similar despite these differences in measured variables.

In 2005, the Industrial Safety and Health Act was amended in Japan. Since 2008, all business must provide face-to-face guidance by physicians for those working long hours.38 Specifically, when the amount of work beyond the standard weekly 40 hours exceeds 100 hours per month and fatigue is observed, the business must provide physician guidance upon worker's request. In addition, when the amount of work beyond the standard weekly 40 hours exceeds 80 hours per month and fatigue is observed or there is a health concern as a result, the business is required to provide face-to-face guidance by a physician upon worker's request.

Nevertheless, our findings suggest that exceeding 40 hours per week by more than 80 hours per month is associated with an OR of about 2 to 4 for depression. This implies that the law may not effectively prevent mental health problems in workers. Even with less than 80 hours of overtime per month, the risk of depression may increase if high job demand exists. In other words, depression cannot be managed solely by reducing the hours worked.

Our findings also suggest that, based on the established role that job demand plays in this association, work hours alone do not lead to depression and that the impact of work hours on depression is rather small. In fact, the standardized total effect of working hours in model 3 for business A is 0.10 whereas that of job demand is 0.26 (data not shown). This implies that reducing job demands, rather than working hours, may be more beneficial for employers, managers, and occupational health care professionals in reducing the risk of depression and developing a more effective workplace mental health program.

In conclusion, our hypothesis that long working hours increase the risk of depression when “job demand” is treated as an intermediate variable, but not a confounder, was supported by both the primary analysis performed on business A and an external validity test of the data set from business B. This result should be considered in reevaluating previous data and planning future studies to elucidate the effects of work hours on depression.

Back to Top | Article Outline
Limitations

The items measured in the two initial surveys are not identical; therefore, there are some limitations in the between-model comparisons. First, “job control” in business A was less reliable because it might be composed of only three items. Given that it was not the main pathway of a causal model, its influence on the interpretation might be limited. Second, given that subjects for business B were from a convenient sampling, the verified models may not be generalizable to the entire population. Careful interpretation of the results is required because studies with data sets with variable qualities should not be treated equally, that is, results from business B. Third, because the present study had a cross-sectional design, causal relationships are inconclusive. Finally, rather than being assessed objectively (eg, time cards), long working hours were assessed on the basis of the response to self-administered questionnaires.

Back to Top | Article Outline
Recommendations for Future Research

To clarify the relationship between working hours and the risk of depression, more longitudinal studies are needed, with identical items measured to treat job demand as an intermediate variable rather than a confounding factor. In these studies, researchers should consider “evaluation criteria,” such as the design (in which all study variables are measured at all points), time lags (there is little information available about the influence of the effect variable), measures (standardized measures should be used to calculate a reliability score for one's own data and the definition of working hours should be carefully considered), method of analysis (multiple regression analysis and SEM are preferable), and nonresponse analysis (in which possible selectivity of the response on baseline and follow-up measurements should be examined).12

Back to Top | Article Outline

REFERENCES

1. World Health Organization and International Labour Organization. Mental Health and Work: Impact, Issues and Good Practices. Geneva, Switzerland: World Health Organization and International Labour Organization; 2000:56.

2. Amagasa T, Nakayama T, Takahashi Y. Karojisatsu in Japan: characteristics of 22 cases of work-related suicide. J Occup Health. 2005;47:157–164.

3. Isometsä ET. Psychological autopsy studies—a review. Eur Psychiatry. 2001;16:379–385.

4. Promotion of preventive measures based on the actual state of suicides. In: Kawakami N, Takeshima T, Takahashi Y. Research on a Pilot Study of Psychological Autopsy: Analysis of Suicide-Related Factors in a Case-Control Study (in Japanese). FY2006 Health, Welfare, and Labour Sciences Research, Research on Psychiatric and Neurological Diseases and Mental Health Project. Tokyo: Ministry of Health, Labour, and Welfare; 2007.

5. Karasek RA. Job demands, job decision latitude, and mental strain: implications for job redesign. Adm Sci Q. 1979;24:285–308.

6. Kivimäki M, Leino-Arjas P, Luukkonen R, Riihimäki H, Vahtera J, Kirjonen J. Work stress and risk of cardiovascular mortality: prospective cohort study of industrial employees. BMJ. 2002;19;325:857–860.

7. Kivimäki M, Virtanen M, Elovainio M, Kouvonen A, Väänänen A, Vahtera J. Work stress in the etiology of coronary heart disease—a meta-analysis. Scand J Work Environ Health. 2006;32:431–442.

8. Michie S, Williams S. Reducing work related psychological ill health and sickness absence: a systematic literature review. Occup Environ Med. 2003;60:3–9.

9. Stansfeld SA, Fuhrer R, Shipley MJ, Marmot MG. Work characteristics predict psychiatric disorder: prospective results from the Whitehall II Study. Occup Environ Med. 1999;56:302–307.

10. Niedhammer I, Goldberg M, Leclerc A, Bugel I, David S. Psychosocial factors at work and subsequent depressive symptoms in the Gazel cohort. Scand J Work Environ Health. 1998;24:197–205.

11. Mausner-Dorsch H, Eaton WW. Psychosocial work environment and depression: epidemiologic assessment of the demand-control model. Am J Public Health. 2000;90:1765–1770.

12. de Lange AH, Taris TW, Kompier MA, Houtman IL, Bongers PM. “The very best of the millennium”: longitudinal research and the demand-control-(support) model. J Occup Health Psychol. 2003;8:282–305.

13. de Lange AH, Taris TW, Kompier MA, Houtman IL, Bongers PM. The relationships between work characteristics and mental health: examining normal, reserved, and reciprocal relationships in a 4-wave study. Work Stress. 2004;18:149–166.

14. Kawakami N, Haratani T. Epidemiology of job stress and health in Japan: review of current evidence and future direction. Ind Health. 1999;37:174–186.

15. Tsutsumi A, Kawakami N. A review of empirical studies on the model of effort–reward imbalance at work: reducing occupational stress by implementing a new theory. Soc Sci Med. 2004;59:2335–2359.

16. Kivimäki M, Virtanen M, Vartia M, et al. Workplace bullying and the risk of cardiovascular disease and depression. Occup Environ Med. 2003;60:779–783.

17. 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;84:874–884.

18. Spurgeon A, Harrington MJ, Cooper LC. Health and safety problems associated with long work hours: a review of the current position. Occup Environ Med. 1997;54:367–375.

19. Sparks K, Cooper C. The effects of hours of work on health: a meta-analytic review. J Occup Organ Psychol. 1997;70:391–408.

20. van der Hulst M. Long work hours and health. Scand J Work Environ Health. 2003;29:171–188.

21. van der Hulst M, Geurts S. Associations between overtime and psychological health in high and low reward jobs. Work Stress. 2001;15:227–240.

22. International Labour Office. Working Time: Its Impact on Safety and Health. Geneva, Switzerland: International Labour Organization; 2003:42–44.
23. Shields M. Long work hours and health. Health Rep. 1999;11:33–48.

24. Fujino Y, Horie S, Hoshuyama T, Tsutsui T, Tanaka Y. A systematic review of working hours and mental health burden (in Japanese). Sangyo Eiseigaku Zasshi. 2006;48:87–97.

25. Hobson J, Beach JR. An investigation of the relationship between psychological health and workload among managers. Occup Med (Lond). 2000;50:518–522.

26. Rothman KJ. Epidemiology: An Introduction. New York, NY: Oxford University Press; 2002:108.

27. Nakayama T, Amagasa T. Special reference to employee knowledge about depression and suicide: baseline results of a workplace-based mental health support program. Psychiatry Clin Neurosci. 2004;58:280–284.

28. Stress Disease Worker's Compensation Insurance Study Group. “Stress and Health” General Survey Research Group. Stress and Health in Japanese Industrial Workers General Survey Report 1 (in Japanese). Japan: Stress Disease Worker's Compensation Insurance Study Group; 1991.

29. Sugisawa A, Uehata T, Sekiya E, et al. Mental health and its related factors among middle-aged male workers: an 18-month follow-up study (in Japanese with an English abstract). Jpn J Ind Health 1994;36:91–101.

30. Shimomitsu T, Odagiri Y. Brief Job Stress Questionnaire (in Japanese). Sangyo Seishin Hoken. 2004;12:25–35.

31. Nakayama T, Toyoda H, Ohno K, Yoshiike N, Futagami T. Validity, reliability and acceptability of the Japanese version of the General Well-Being Schedule (GWBS). Qual Life Res. 2000;9:529–539.

32. Shima S, Kano T, Kitamura T, Asai M. New self-rating scales for depression (in Japanese). Seishin Igaku. 1985;27:717–723.

33. IBM Japan Inc. http://www.ibm.com/jp/ja. Accessed April 5, 2012.

34. Toyoda H, ed. Covariance Structure Analysis [Amos Ed.] Structural Equation Modeling (in Japanese). Tokyo, Japan: Tosho; 2007.

35. Toyoda H. Covariance Structure Analysis [Introduction]. Structural Equation Modeling (in Japanese). Tokyo, Japan: Asakura Publishing; 1998

36. Ministry of Internal Affairs and Communication, Statistics Bureau home page. World statistics. Chapter 12. Labour and wages. http://www.stat.go.jp/data/sekai/12.htm#h12-06. Accessed April 5, 2012.

37. Lee S. Working-hour gaps: trends and issues. In: Messenger JC, ed. Working Time and Workers' Preference in Industrialized Countries: Finding the Balance. London, England: Routledge Taylor & Francis; 2004:29–59.

38. Ministry of Health, Labour, and Welfare. Amended Industrial Safety and Health Act (in Japanese). http://law.e-gov.go.jp/htmldata/S47/S47HO057.html. Enacted April 1, 2006. Accessed April 5, 2012.

©2012The American College of Occupational and Environmental Medicine

Login

Article Tools

Images

Share