CE: S.W.; JOEM-16-5754; Total nos of Pages: 6; JOEM-16-5754 ORIGINAL ARTICLE Chronic Disease Risks From Exposure to Long-Hour Work Schedules Over a 32-Year Period Allard E. Dembe, ScD and Xiaoxi Yao, PhD, MPH Objectives: This study aims at evaluating the chronic disease risk related to prolonged work in long-hour schedules for eight major chronic diseases: heart disease, non-skin cancer, arthritis, diabetes, chronic lung disease, asthma, chronic depression, and hypertension. Methods: The study used data from the National Longitudinal Survey of Youth, 1979 covering 32 years of job history (1978 to 2009) for 7492 respondents. Logistic regression analyses were performed to test the relationship between average weekly work hours, and the reported prevalence of those conditions for each individual. Results: Regularly working long hours over 32 years was significantly associated with elevated risks of heart disease, non-skin cancer, arthritis, and diabetes. The observed risk was much larger among women than among men. Conclusions: Working long-hour schedules over many years increases the risk for some specific chronic diseases, especially for women. T his study examines whether working long-hour schedules for many years increases the likelihood for contracting a chronic disease later in life. Previous studies have documented a variety of adverse health and safety outcomes related to working long hours, including fatigue, stress, sleep disorders, digestive problems, decreased work performance, and work-related injuries1–5 However, comparatively few studies have assessed the risk of long-term adverse outcomes that may develop gradually over many months or years as the result of working in long-hour schedules. Existing literature concerning the association between long work-hours and chronic disease is mixed. There have been several studies supporting the view that working long hours is associated with heart disease.6– 8 However, for other chronic conditions, the evidence is inconclusive and sometimes contradictory. For example, some studies suggest that long work hours leads to hypertension.9,10 Nakanishi et al,11 though, found that working long hours was negatively associated with hypertension among male Japanese white-collar workers. There is also inconsistent evidence concerning diabetes. Kawakami et al12 found that working long hours was a risk factor for type 2 diabetes, but Nakanishi et al13 found the opposite results. A recent prospective study found that working long hours is a risk factor for depression and anxiety among working women.14 However, other previous studies found no significant relationship in this area.15 Other studies have been constrained by focusing only on specific occupational subgroups, such as health care workers16, police officers17, and white-collar workers11, or employment confined to a single employer organization. To analyze the relationship between years of work in longhour schedules and outcomes that may develop gradually over From the Center for HOPES, College of Public Health, The Ohio State University, Columbus (Dr Dembe); and the Center for the Science of Health Care Delivery, Mayo Clinic, Rochester, Minnesota (Dr Yao). Funding for this study was provided by the U.S. Centers for Disease Control and Prevention, National Institute of Occupational Safety and Health (Grant number: R21-OH010323). The authors have no conflicts of interest. Address correspondence to: Allard E. Dembe, ScD, Professor of Public Health, The Ohio State University, 1841 Neil Avenue, Room 283, Columbus, OH 43210 (adembe@cph.osu.edu). Copyright ß 2016 American College of Occupational and Environmental Medicine DOI: 10.1097/JOM.0000000000000810 JOEM Volume XX, Number X, Month 2016 decades is methodologically challenging. Ideally, to perform such a study, it would be necessary to have a longitudinal database composed of a large number of workers from many occupations and industries, accurate job history and work-hour information, detailed outcomes assessment data, including date of onset, incidence by year and prevalence rates, as well as medical diagnosis and treatment information. From a practical standpoint, it is very difficult to perform such a study. In the U.S., there are few databases that collect work-hour and job history information over long periods of time. This difficulty is compounded because employees frequently change employers and occupations. Moreover, there are separate medical systems in the U.S. that individuals utilize to obtain medical care. Care for general medical disorders and chronic disease is most often obtained through employer-sponsored health insurance or public health care systems such as Medicaid and Medicare. Care for work-related conditions is typically provided under workers’ compensation insurance by doctors who are not the workers’ usual source of care. The methods applied in this study address many of these concerns. We used a large longitudinal database, the National Longitudinal Survey of Youth 1979 (NLSY79), to collect information about job histories, work hours, and chronic disease status for cohort members who had reached at least 40 years of age. Disease status was ascertained for eight common chronic diseases, including heart disease, non-skin cancer, arthritis, diabetes, chronic lung disease, asthma, depression, and hypertension. A secondary aim of this study was to determine whether the relationship between working long hours and the risk of contracting chronic diseases differs for men and women. Previous research has indicated that women’s health may be affected by long working hours more than men’s. For example, a Swedish study found that working at least 10 overtime hours per week was related to increased hospitalization incidence for women, but a decreased hospitalization rate for men, after adjusting for age and other relevant cofactors.18 Another study found similar results; overtime work of more than 5 hours per week increased women’s mortality rate, but moderate overtime work (1 to 5 overtime hours per week) had a protective effect for men.19 To investigate this issue, we conducted additional stratified analyses to evaluate the differential effects of long work hours among male and female workers. METHODS This study used NLSY79 to determine individuals’ work hours and the prevalence of the eight chronic diseases. The NLSY79 is a nationally representative survey that was originally composed of 12,686 men and women who were 14 to 22 years old in 1979. Cohort members have been interviewed annually from 1979 to 1994 and biennially since 1996. The latest year of complete data for this study was drawn from the 2010 survey when the cohort members were 46 to 53 years old. The NLSY79 collects information on respondents’ demographic and socioeconomic characteristics as well as their detailed work histories. Beginning in 1998, additional survey questions, called the ‘‘40þ module,’’ were added to the NLSY79 to assess the health status and prevalence of chronic conditions among respondents upon turning 40 years old. A similar set of questions (50þ module) was added in 2008 as cohort members turned 50 years old. 1 Copyright © 2016 American College of Occupational and Environmental Medicine. Unauthorized reproduction of this article is prohibited CE: S.W.; JOEM-16-5754; Total nos of Pages: 6; JOEM-16-5754 JOEM Volume XX, Number X, Month 2016 Dembe and Yao The study’s primary independent variable was an individual’s self-reported average hours per week. The person’s average hours per week were summed across the entire 32-year study period to reflect a worker’s aggregate exposure to long-hour schedules. Only a worker’s full-time work experience was considered, where a ‘‘fulltime’’ schedule was defined as one containing at least 30 hours per week. If a person worked at two or more different jobs in a week (eg, 15 hours at one job and 15 hours at another), the sum total of work hours for all jobs was used to determine the individual’s full-time work status for that week. Because the study’s aim was to examine the effect of long work hours, weeks involving only part-time work (or no work) were not included. Overall, 73 of 7565 respondents never worked full time during the 32-year study period, and thus they were excluded from the study, resulting in an analytical sample of 7492 individuals. NLSY79 contained questions from the 40þ and 50þ modules asking respondents to self-report whether a doctor had ever diagnosed the individual with any of eight conditions: (1) ‘‘heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems’’; (2) ‘‘cancer or malignant tumor of any kind except skin cancer’’; (3) ‘‘arthritis or rheumatism’’; (4) ‘‘diabetes or high blood sugar’’; (5) ‘‘chronic lung disease such as chronic bronchitis or emphysema’’; (6) ‘‘asthma’’; (7) ‘‘depression’’ (the 40þ module asked about ‘‘depression, excessive worry or nervous trouble’’); and (8) ‘‘high blood pressure or hypertension.’’ Multivariable logistic regression was used to test the relationship between average hours worked per week and the occurrence of each of the eight chronic diseases. We adjusted for age, gender, race, education, family income, number of years worked, smoking status, and occupation in the regression model. After conducting the regression analyses within the overall study sample, we conducted additional stratified analyses by gender. We first performed the logistic regression among men and then repeated the analysis considering female cohort members only. Person-level survey weights were provided by NLSY79, in order to reflect the national distribution of Americans in this age range. SAS Statistical Software, Version 2.0 (SAS Institute Inc., Cary, NC) was used to calculate an odds ratio (OR) and associated 95% confidence interval for the relationship between a particular work-hour category and the reported prevalence of one or more of the aforementioned chronic diseases. In addition, because respondents completed the 40þ module as they turned 40 and completed the 50þ module after they turned 50, people over 50 had two opportunities to report diseases, whereas those under 50 only had one opportunity. We therefore controlled for whether individuals were aged 50 and over as of 2010 (who had two reporting opportunities) or aged 45 to 49 (who had one reporting opportunity). RESULTS Table 1 summarizes respondents’ key demographic and employment characteristics, as well as the reported prevalence of the eight chronic diseases. As of 2010, the mean age of the responding cohort was 49.6 years. Most of the respondents were age 50 years or older. The population consisted of slightly more men than women. The majority of respondents worked full-time for more than 20 years during the 32-year study period. The prevalence of reported chronic disease varied considerably by condition. For example, only 3.7% of the study population reported having been diagnosed with non-skin cancer, and 4.4% reported doctordiagnosed chronic lung diseases. By contrast, reporting of depression and arthritis was more common: 18.2% reported having doctor-diagnosed depression and 17.6% reported doctordiagnosed arthritis. For the purposes of analysis, working 30 to 40 hours per week (designated as ‘‘conventional work hours’’) was used as the referent group in the logistic regression. Among full-time workers, 27.7% 2 ß TABLE 1. Demographic Characteristics (N ¼ 7492) Age, yrs (as of 2010) 46–49 50–53 Gender Male Female Education High school or less College Graduate school Duration of work, yrs <10 10–20 >20 Average work hours per week 30–40 41–50 51–60 >60 Prevalence of diseases Heart disease Non-skin cancer Arthritis Diabetes Chronic lung disease Asthma Depression Hypertension Frequency Weighted Percentage (%) 3977 3515 48.95 51.05 3653 3839 51.06 48.94 3947 2749 796 49.09 37.78 13.13 1102 2115 4275 12.90 26.55 60.54 2242 4171 869 210 27.72 56.35 12.98 2.95 409 234 1257 712 319 738 1280 1239 5.81 3.68 17.59 8.88 4.43 9.44 18.24 15.09 worked an average of 30 to 40 hours per week, 56.4% worked an average of 41 to 50 hours per week, 13.0% worked an average of 51 to 60 hours per week, and 3.0% on average worked over 60 hours per week (Table 1). Table 2 summarizes the results of the multivariable logistic regression analyses comparing working long hours with conventional (30 to 40 per week) work hours. The analyses indicated that working long hours is significantly associated with elevated risks for four types of chronic conditions. People working 51 to 60 hours a week were found to have an elevated risk for heart disease, as did people working more than 60 hours per week. There was also a significant elevated risk for non-skin cancer. Likewise, the risk of arthritis was significantly elevated for cohort members who averaged more than 40 hours per week compared with workers with conventional hours. In addition, there was a statistically greater likelihood of reporting diabetes among people who worked more than 40 hours per week. There were no statistically significant findings (at P < 0.05) for an association between long hours and chronic lung disease, asthma, depression, or hypertension. The stratified analyses by gender showed that the effect of long work hours was comparatively much greater among women than among men. For men, long hour work appeared only to affect the risk of contracting arthritis. No adverse effects were found for other conditions. In fact, working moderately long hours (41 to 50 hours per week) was actually associated with less risk of contracting heart disease, chronic lung disease, or depression (Table 3). By contract, the results among female workers were striking. Women consistently showed a markedly strong relationship between long work hours and the prevalence of heart disease, non-skin cancer, arthritis, and diabetes. Moreover, the results for those four diseases among women provided very strong evidence of a consistent dose–response relationship (Table 4). Similar, but less strong associations were also found regarding women’s 2016 American College of Occupational and Environmental Medicine Copyright © 2016 American College of Occupational and Environmental Medicine. Unauthorized reproduction of this article is prohibited ß Nonskin Cancer — 1.50 (1.02–2.19)b 2.03 (1.16–3.54)b 2.83 (1.28–6.24)b Heart Disease — 1.19 (0.87–1.62) 1.68 (1.06–2.67)b 1.74 (0.93–3.24)a — 1.50 (1.25–1.80)b 1.99 (1.52–2.61)c 2.97 (1.98–4.44)c Arthritis 2016 American College of Occupational and Environmental Medicine — 1.46 (1.14–1.88)c 1.63 (1.12–2.36)b 1.62 (0.91–2.89)a Diabetes — 1.00 (0.71–1.40) 1.58 (0.95–2.63)a 1.62 (0.78–3.35) Chronic Lung Disease — 1.36 (0.46–3.99) 1.50 (0.43–5.26) 1.55 (0.30–7.97) — 0.57 (0.34–0.96)b 0.76 (0.40–1.43) 0.66 (0.26–1.68) — 1.34 (0.94–1.92) 1.85 (1.22–2.82)c 2.35 (1.34–4.12)c Arthritis Bolding based on P < 0.10 added to help show potential dose–response trends. a P < 0.10. b P < 0.05. c P < 0.01. Nonskin Cancer Heart Disease — 1.08 (0.70–1.67) 1.15 (0.67–1.97) 0.91 (0.41–2.04) Diabetes — 0.45 (0.25–0.81)c 0.65 (0.30–1.40) 0.63 (0.20–1.91) Chronic Lung Disease — 0.94 (0.61–1.45) 0.96 (0.54–1.69) 0.87 (0.33–2.28) Asthma — 1.23 (0.98–1.54)a 1.22 (0.86–1.75) 1.66 (0.97–2.84)a Asthma — 0.63 (0.45–0.88)c 0.81 (0.54–1.21) 0.50 (0.24–1.03)a Depression — 0.96 (0.80–1.15) 1.04 (0.79–1.38) 0.82 (0.50–1.35) Depression — 0.89 (0.65–1.21) 0.92 (0.63–1.34) 0.86 (0.47–1.57) Hypertension — 1.06 (0.88–1.27) 1.14 (0.87–1.51) 1.07 (0.66–1.72) Hypertension JOEM Volume XX, Number X, Month 2016 30–40 41–50 51–60 >60 Hours/Week TABLE 3. Logistic Regression Results Among All Men: Odds Ratios and 95% Confidence Intervals Bolding based on P < 0.10 added to help show potential dose–response trends. a P < 0.10. b P < 0.05. c P < 0.01. 30–40 41–50 51–60 >60 Hours/Week TABLE 2. Logistic Regression Results Among All Respondents: Odds Ratios and 95% Confidence Intervals CE: S.W.; JOEM-16-5754; Total nos of Pages: 6; JOEM-16-5754 Long-Hour Work Schedules and Chronic Disease Risk 3 Copyright © 2016 American College of Occupational and Environmental Medicine. Unauthorized reproduction of this article is prohibited CE: S.W.; JOEM-16-5754; Total nos of Pages: 6; JOEM-16-5754 JOEM Volume XX, Number X, Month 2016 4 — 1.13 (0.90–1.43) 1.47 (0.94–2.29)a 1.22 (0.51–2.93) risk for hypertension and asthma (among women working 51 to 60 hours per week). DISCUSSION — 1.28 (0.88–1.87) 2.54 (1.38–4.68)c 2.59 (0.97–6.91)a — 1.29 (0.99–1.67)a 1.28 (0.80–2.04) 2.89 (1.46–5.72)c — 1.12 (0.91–1.37) 1.03 (0.68–1.54) 1.18 (0.58–2.40) Our focus on the number of work hours per week as the main exposure measure is a very common approach, reflecting general practice within many employment settings. It has the advantage of encompassing weekend work and work performed at unusual hours. The long period of exposure assessment extending over 32 years helps to smooth out temporary fluctuations in employment history that may occur. This is particularly appropriate because the chronic disease outcomes in question generally develop over an extended period. It should be noted, though, that there are other potential approaches for defining long work-hour exposure. For example, alternative measures of exposure might include (1) the number of work hours per day, (2) the number of work hours per shift, (3) the total number of work hours in a year, and (4) the number of weeks working full-time during a year. This study utilized a large national longitudinal survey to assess the relationship between workers’ long-term exposure to extended work hours and their risk of developing eight major chronic diseases. The results of this study are limited by the available methodology for assessing long-term chronic disease using the NLSY79. In the NLSY79, respondents were asked to report chronic diseases only when they turned 40 (in the 40þ module) or 50 (in the 50þ module). As of 2010, all the respondents had turned 40 and thus were eligible for completing the 40þ module, but only half of the cohort members had yet reached age 50 years or older and thus eligible to complete the 50þ module. Consequently, the risk assessed in this study is not a lifetime risk, but rather the risk of ‘‘early-onset’’ chronic diseases that were identified upon reaching the age of 40 or, for some respondents, 50 years. These reporting practices may therefore result in some underestimation of true chronic diseases prevalence and thus bias results toward the null. Despite this limitation, our findings suggest a very strong and consistent relationship between long working hours and the risk of developing certain chronic diseases at a relatively young age (40 or 50 years). The early onset and identification of chronic diseases may not only reduce individuals’ life expectancy and quality of life but also increase health care costs in the long term. Another major finding of this study is that women’s health might be considerably more affected by long work hours than men. The observed effect among men was modest and confined to arthritis. However, among women, the results were striking, demonstrating high ORs and consistent dose–response relationships for heart diseases, non-skin cancer, arthritis, and diabetes. Relatively similar effects were also observed among women with respect to the risk of chronic lung disease and asthma, albeit with lower effect sizes and statistical significance. With regard to arthritis, which was associated with working long hours among both men and women, the magnitude of the adverse effect was considerably larger among women. There are several possible reasons for this startling gender difference. For instance, numerous studies have shown that women are generally more likely to seek medical care for chronic disease diagnosis and treatment than men.20,21 In addition, research indicates that women generally assume greater family responsibilities and thus may be more likely to experience inter-role conflict and overload than men.22,23 Therefore, when women work long hours, they may experience more time pressure and stress than men, and their health consequently might be more effected by working long hours, especially when considered over a long timeframe. The observed gender difference may also be attributable to the fact that women are more likely to be exposed to negative psychosocial work Bolding based on P < 0.10 added to help show potential dose–response trends. a P < 0.10. b P < 0.05. c P < 0.01. — 1.62 (1.20–2.17)c 2.04 (1.18–3.54)b 3.20 (1.37–7.51)c — 1.53 (1.02–2.31)a 2.37 (1.22–4.57)a 3.52 (1.38–9.00)c 30–40 41–50 51–60 >60 — 1.61 (1.13–2.29)c 2.80 (1.48–5.31)c 3.67 (1.58–8.56)c — 1.55 (1.25–1.92)c 1.82 (1.22–2.72)c 3.93 (2.03–7.61)c Chronic Lung Disease Diabetes Arthritis Non-Skin Cancer Heart Disease Hours/Week TABLE 4. Logistic Regression Results Among Women: Odds Ratios and 95% Confidence Intervals Asthma Depression Hypertension Dembe and Yao ß 2016 American College of Occupational and Environmental Medicine Copyright © 2016 American College of Occupational and Environmental Medicine. Unauthorized reproduction of this article is prohibited CE: S.W.; JOEM-16-5754; Total nos of Pages: 6; JOEM-16-5754 JOEM Volume XX, Number X, Month 2016 characteristics, such as low substantive complexity and work control,24,25 lack of learning opportunities, and job monotony.26 Working long hours increases women’s exposure to these negative work characteristics, which might contribute to their overall burden of impaired health and chronic disease. We tried including marital status and number of children in our preliminary regression model, but found that those factors were not significant and did not exert any impact on chronic disease likelihood. One reason for the lack of effect may be that there was considerable churning of marital status and the number of children during the assessment period, with about 90% of people reporting having a spouse/partner at some point during the 32 years of followup and 92% reporting having at least one child during that time period. However, our data lack the nuance to further explain the complicated family dynamics and their interaction with work responsibilities. A methodological strength of our study is that, using 32 years of work history data, we were able to assess the influences of long work hours not only for workers’ current jobs but also for previous jobs that may be relevant to the development of chronic diseases. In addition, our study conducted the analyses on a general population sample that includes all the individuals who have worked full time during the 32-year period. This is an advantage compared with other studies that collect information only from individuals who are currently employed. Attrition of participants over the 32-year period was relatively modest averaging only 1.3% per year (12,686 À 7,565/ 12,686) Ä32 ¼ 1.26%. By comparison, the average annual attrition rate for the Health and Retirement Survey (1992 to 2008) was 2.38%, for the National Educational Longitudinal Survey, it was 3.1%, and for the American Housing Survey (1985 to 2009), it was 2.9%. Some of the observed attrition may be due to NLSY79’s decision to survey fewer people owing to a lack of funding, rather than respondents dropping out of the study. The NLSY79 survey has adjusted its sample weights accordingly, so the loss of follow-up should not unduly bias our results. Although the lack of diagnosis onset information did not permit the calculation of time trends in disease occurrence, we were able to calculate exposure-disease relationships in multiple categories, for example, an average coefficient of 40 to 50 hours of exposure per week, 50 to 60 hours per week, and more than 60 hours per week. For those purposes, we used the Wald test to evaluate the observed dose-relationship trend. The trend was significant for arthritis overall (P < 0.001), for men (P < 0.05), and for women (P < 0.05). In women, there was also a significant doserelationship trend for heart disease (P < 0.05) and chronic lung disease (P < 0.05). No other statistically significant trends for other chronic diseases were detected. Nevertheless, there are still a number of unanswered questions in this research field about the best ways of operationalizing the concept of long work-hour exposure. We chose to focus on the mean number of hours per week averaged over the entire study period as the primary measure. This has the advantage of capturing work-hour information from all the full-time work performed over the 32-year study period. However, this technique masks potentially useful information about the total duration of exposure. For example, two individuals with different patterns of work hours may have the same number of average weekly hours. Therefore, an individual who worked many hours at the beginning of the 32year period and later worked fewer hours can total the same number of average weekly hours as another person who worked fewer hours at the beginning and later worked long weeks. It is not known whether these two types of exposures have the same impact on workers’ health or not. Similarly, can somebody working 40 hours per week for 20 years be expected to have the same exposure as someone else ß Long-Hour Work Schedules and Chronic Disease Risk working 20 hours per week for 40 years? The methodology applied in this study would give greater weight to the person with the higher level of hourly work per week. But the true influence of work-hour intensity relative to the duration of exposure is not known. Another limitation of this study is that workers’ exposure to long-hour schedules accrued during the entire 32-year period. However, the onset of chronic disease may have occurred at an earlier date. The information provided in NLSY79, in most cases, was not sufficient to determine the precise date of disease onset (or diagnosis) relative to the dates of work-hour exposure. Thus, the exposure that accrued after the occurrence of the chronic disease might not be relevant. A Cox proportional hazards model potentially could take the time of disease onset into account and truncate any exposure following diagnosis. However, we were not able to perform such an analysis, because NLSY79 lacked information about the date of diagnosis for most of the eight conditions. For arthritis, NLSY79 data were collected about the date of diagnosis, and so, we were able to conduct a sensitivity test using survival analysis techniques. The calculated results using the truncated Cox proportional technique and the results using our primary methodology employing logistic regression analysis provided quite similar results. This provided some level of assurance that the observed effects are valid. Epidemiological studies have suggested that there may be an association between long work hours and smoking, which could affect long-term chronic disease risk. For that reason, we included smoking as a cofactor in the regression analysis. To further investigate the effect of smoking, we re-ran the analyses without smoking to assess its impact on the results. The sensitivity analysis indicated that the results remained virtually unchanged. Also, there is potential for a healthy worker effect (HWE) to bias the results of this study. In traditional versions of the HWE, healthy workers are more likely to stay employed and less healthy workers are more likely to stop being employed (or be employed less often) thereby leading to a selection effect whereby the remaining workers are more healthy than is to be expected. In this study, that would make it less likely to detect chronic disease, thus biasing the results to the null hypothesis and underestimating the true results. Our results are less sensitive to HWE bias than in other studies, because we retain all the exposure and chronic disease data even after an individual has stopping working due to illness. The method that the NLSY79 utilizes to collect health information might also create bias in the study. If a person was diagnosed with a disease after turning 40 years old, but had not yet turned 50 years old as of 2010, or was diagnosed as positive for a chronic disease after turning 50 years old, the reported disease status would be inaccurate. In addition, for asthma and depression, the interview questions only asked whether the respondent had the condition, rather than having a doctor-diagnosed condition. Therefore, the disease status for some respondents might be misclassified. In addition, NLSY data did not permit us to know whether the reason for people working long hours was ‘‘obligatory’’ (ie, mandatory) or discretionary. How much control a person has in their work hours is related to their socioeconomic characteristics (race, education, income, occupation, etc.), which were adjusted in our analytical models. The results of this study suggest that measures should be taken by employers, employees, and safety officials to become familiar with the potential risks associated with working excessively long hours on the job. The greatest risk involves women working more than 40 hours per week. The risk appears to be proportional to the number of hours worked. Averaged over many years, women working more than 60 hours per week have a nearly three-fold risk of chronic disease, including heart disease, non-skin cancer, arthritis, and diabetes. According to the U.S. Bureau of Labor Statistics (as of 2012 data), 1.8% of women regularly work 60 or more hours 2016 American College of Occupational and Environmental Medicine 5 Copyright © 2016 American College of Occupational and Environmental Medicine. Unauthorized reproduction of this article is prohibited CE: S.W.; JOEM-16-5754; Total nos of Pages: 6; JOEM-16-5754 JOEM Volume XX, Number X, Month 2016 Dembe and Yao per week and 6.2% of women work 50 or more hours per week.27 That translates to an estimated 1,083,000 American women working 60 or more hours per week, and 3,837,000 women working 50 or more hours per week. The primary way to decrease risk is to adopt work schedules with fewer hours. Population surveys have found that the majority of American workers desire working fewer hours. The gap between the ideal and actual work hours is largest among those who work very long weeks (ie, more than 60 hours per week).28 Considering the adverse effects of long working hours, reducing employees’ excessive hours potentially can benefit both workers and employers. On the one hand, reducing work hours can lessen workers’ stress and improve their health. At the same time, employers can also benefit from increased productivity and enhanced performance by healthy workers, as well as fewer accidents and errors related to worker fatigue. Providing employees with more flexibility, support, and control over their work schedules and work arrangements might also help alleviate the adverse effects of long work hours.29–31 Because women’s health may be disproportionately affected by working long hours, employers should consider specific familyfriendly policies to help women better manage the demands from work and family, and thereby help maintain good health status. Some employers have created special on-site health promotion programs customized for female workers.32 Increased research and greater attention to the needs of women working long hours is needed. Workplace wellness and health promotion programs provide an option by which employers can address and mitigate the adverse health effects of long working hours. Some employers have established or financially supported chronic disease management programs to help employees deal with existing chronic disease issues. Major components of successful workplace programs include health education, worksite screening of risk factors, and modification of health behaviors.33 Employers can benefit from investing in high-quality workplace wellness programs. It has been demonstrated that by modifying long-hour work schedules and conducting effective wellness programs, workers can be more productive, have less sickness absence, and decrease medical expenditures.34 – 36 REFERENCES 1. Caruso CC. Possible broad impacts of long work hours. Ind Health. 2006;44:531–536. 2. Johnson JV, Lipscomb J. Long working hours, occupational health and the changing nature of work organization. Amer J Ind Med. 2006;49:921–929. 3. Sparks K, Cooper C, Fried Y, Shirom A. The effects of hours of work on health: a meta-analytic review. J Occup Organ Psychol. 1997;70:391–408. 4. Kodz J, Davis S, Lain D, et al. Working long hours in the U.K.: a review of the evidence: Volume 1 - Main Report. Employment Relations Research Series ERRS16. Brighton, UK: Department of Industry and Trade, The Institute for Employment Studies; October 2003. 5. Dembe AE, Erickson JB, Delbos RG, Banks SM. The impact of overtime and long work hours on occupational injuries and illnesses: new evidence from the United States. Occup Environ Med. 2005;62:588–597. 6. Liu Y, Tanaka H. Overtime work, insufficient sleep, and risk of non-fatal acute myocardial infarction in Japanese men. Occup Environ Med. 2002;59:447–451. 7. Virtanen M, Heikkila K, Jokela M, Ferrie JE, Batty GD, Vahtera J, et al. Long working hours and coronary heart disease: a systematic review and metaanalysis. Am J Epidemiol. 2012;176:586–596. 8. Landsbergis P. Long work hours, hypertension, and cardiovascular disease. Cad Sau´de Pu´blica. 2004;20:1746–1748. 9. Hayashi T, Kobayashi Y, Yamaoka K, Yano E. Effect of overtime work on 24hour ambulatory blood pressure. J Occup Environ Med. 1996;38:1007–1011. 10. Yang H, Schnall PL, Jauregui M, Su T-C, Baker D. Work hours and selfreported hypertension among working people in California. Hypertension. 2006;48:744–750. 6 ß 11. Nakanishi N, Yoshida H, Nagano K, Kawashimo H, Nakamura K, Tatara K. Long working hours and risk for hypertension in Japanese male white collar workers. J Epidemiol Community Health. 2001;55:316–322. 12. Kawakami N, Araki S, Takatsuka N, Shimizu H, Ishibashi H. Overtime, psychosocial working conditions, and occurrence of non-insulin dependent diabetes mellitus in Japanese men. J Epidemiol Community Health. 1999;53:359–363. 13. Nakanishi N, Nishina K, Yoshida H, Matsuo Y, Nagano K, Nakamura K, et al. Hours of work and the risk of developing impaired fasting glucose or type 2 diabetes mellitus in Japanese male office workers. Occup Environ Med. 2001;58:569–574. 14. Virtanen M, Ferrie JE, Singh-Manoux A, Shipley MJ, Stansfeld SA, Marmot MG, et al. Long working hours and symptoms of anxiety and depression: a 5year follow-up of the Whitehall II study. Psychol Med. 2011;41:2485–2494. 15. Michelsen H, Bildt C. Psychosocial conditions on and off the job and psychological ill health: depressive symptoms, impaired psychological wellbeing, heavy consumption of alcohol. Occup Environ Med. 2003;60: 489–496. 16. Poissonnet CM, Veron M. Health effects of work schedules in healthcare professions. J Clin Nurs. 2000;9:13–23. 17. Vila B. Impact of long work hours on police officers and the communities they serve. Am J Ind Med. 2006;49:972–980. 18. Alfredsson L, Spetz C-L, Theorell T. Type of occupation and near-future hospitalization for myocardial infarction and some other diagnoses. Int J Epidemiol. 1985;14:378–388. 19. Nylen L, Voss M, Floderus B. Mortality among women and men relative to unemployment, part time work, overtime work, and extra work: a study based on data from the Swedish twin registry. Occup Environ Med. 2001;58:52–57. 20. Kroenke K, Spitzer RL. Gender differences in the reporting of physical and somatoform symptoms. Psychosom Med. 1998;60:150–155. 21. Bertakis KD, Azari R, Helms LJ, Callahan EJ, Robbins JA. Gender differences in the utilization of health care services. J Fam Pract. 2000;49: 147–152. 22. Hall EM. Gender, work control, and stress: a theoretical discussion and an empirical test. Int J Health Serv. 1989;19:725–745. 23. Roxburgh S. Gender differences in work and well-being: effects of exposure and vulnerability. J Health Soc Behav. 1996;37:265–277. 24. Matthews S, Hertzman C, Ostry A, Power C. Gender, work roles and psychosocial work characteristics as determinants of health. Soc Sci Med. 1998;46:1417–1424. 25. Denton M, Prus S, Walters V. Gender differences in health: a Canadian study of the psychosocial, structural and behavioural determinants of health. Soc Sci Med. 2004;58:2585–2600. 26. Melamed S, Ben-Avi I, Luz J, Green MS. Objective and subjective work monotony: effects on job satisfaction, psychological distress, and absenteeism in blue-collar workers. J Appl Psychol. 1995;80:29–42. 27. U.S. Bureau of Labor Statistics (BLS). Women in the Labor Force: A Databook. Report 1049. Washington, DC: BLS; May, 2014. 28. Jacobs JA, Gerson K. The Time Divide: Work, Family, and Gender Inequality. Cambridge, MA: Harvard University Press; 2004. 29. Butler AB, Grzywacz JG, Ettner SL, Liu B. Workplace flexibility, selfreported health, and health care utilization. Work Stress. 2009;23:45–59. 30. Grzywacz JG, Carlson DS, Shulkin S. Schedule flexibility and stress: linking formal flexible arrangements and perceived flexibility to employee health. Community Work Fam. 2008;11:199–214. 31. Ala-Mursula L, Vahtera J, Kivima¨ki M, Kevin MV, Pentti J. Employee control over working times: associations with subjective health and sickness absences. J Epidemiol Community Health. 2002;56:272–278. 32. Campbell MC, Tessaro I, DeVellis B, et al. Effects of a tailored health promotion program for female blue-collar workers: health works for women. Prev Med. 2002;34:313–323. 33. Linnan L, Bowling M, Childress J, et al. Results of the 2004 National Worksite Health Promotion Survey. Am J Public Health. 2008;98:1503– 1509. 34. Baicker K, Cutler D, Song Z. Workplace wellness programs can generate savings. Health Aff. 2010;29:304–311. 35. Merrill RM, Aldana SG, Garrett J, Ross C. Effectiveness of a workplace wellness program for maintaining health and promoting healthy behaviors. J Occup Environ Med. 2011;53:782–787. 36. Ozminkowski RJ, Ling D, Goetzel RZ, Bruno JA, Rutter KR, Isaac F, et al. Long-term impact of Johnson & Johnson’s Health & Wellness Program on health care utilization and expenditures. J Occup Environ Med. 2002;44: 21–29. 2016 American College of Occupational and Environmental Medicine Copyright © 2016 American College of Occupational and Environmental Medicine. Unauthorized reproduction of this article is prohibited