Workers must adapt to rapidly changing working conditions and psychosocial factors in the labor market. Thus, appropriate and comprehensive approaches are needed to address job stress-related issues in the workplace. In addition, psychosocial factors (including job stress and mental health) have been important issues in the field of occupational health, especially in developed countries. In this context, researchers, management, health providers, trade unions, and policy makers were interested in productivity, sustainable employment, and the burden of social change. Furthermore, there was increasing evidence that job stress was significantly associated with mental disorders,123 coronary heart disease,45 stroke,6 diabetes mellitus,7 hypertension,89 and musculoskeletal disorders.10 Therefore, it is important to measure accurately the job stress-related burden of disease (BOD), in order to efficiently allocate limited resources and facilitate rational decision-making to promote workers' health.

The World Health Organization (WHO) and the Global Burden of Disease (GBD) study suggested that death and disability-adjusted life years (DALY) was a useful metric for quantifying BOD. DALY is defined as the sum of years of life lost (YLL: burden due to death) and years lived with a disability (YLD: burden due to disability). Another metric for quantifying the effect of job stress on BOD is the population attributable fraction (PAF). PAF or BOD for specific causes were rare in Korea but it is worthwhile to estimate the social impact or to decide policy priority. Thus, the job stress-attributable BOD can be calculated by multiplying DALY and PAF.1112

Psychiatric disorders account for a large part of the global BOD, and the WHO has predicted that unipolar depressive disorder will become the leading cause of global BOD by 2030 (6.2% of the total DALY, up from 4.3% in 2004).13 In addition, for several years during the 21st century, Korea has ranked highest in suicide rate among the members of the Organization for Economic Co-operation and Development.14 Thus, Korean companies became increasingly interested in the mental health of their workers. Furthermore, the Samsung Economic Research Institute reported that the number of mental illnesses patients increased by 50% from 1.58 million people in 2004 to 2.31 million people in 2010. Moreover, the estimated socioeconomic cost of mental illness in 2010 was approximately 23 trillion KRW (approximately 19 billion USD), which corresponds to approximately 2.01% of the Korean gross domestic product (GDP) in 2010. The indirect cost of lost productivity was estimated to be approximately 20 trillion KRW. This significant economic burden highlights the importance of job stress-related mental health issues (e.g., psychiatric disorders and suicide).15

Previous studies have revealed relatively consistent associations between job stress and several cardiovascular diseases, which were listed as occupational diseases in the Enforcement Decree of the Korean Industrial Accident Compensation Insurance Act. These cardiovascular diseases include intracerebral hemorrhage, subarachnoid hemorrhage, cerebral infarction, myocardial infarction, and dissecting aortic aneurysm with work-related causes (including psychological strain), rather than natural disease progression.16 Cardiovascular diseases and cerebrovascular diseases were the second and third leading causes of death in Korea, which suggests that these diseases are important targets when estimating the BOD from job stress in Korea. Interestingly, a French study from 2003 revealed that the attributable fractions of job stress were 25% for cardiovascular disease, 34% for anxiety and depression, and 27% for musculoskeletal disorders.1718

The present study aimed to evaluate the role of job stress in disease development by estimating PAF for job stress in selected health outcomes and job stress related DALY in Korea.

Approximately 21.2% of Korean workers were exposed to high job strain. When we evaluated the effects of job strain, we found that it was attributable for approximately 6.7% of IHD cases, 6.9% of stroke cases (4.8% for ischemic stroke and 2.1% for hemorrhagic stroke), 13.6% of MDD cases, and 4% of suicides (Table 1). The estimated prevalences per 100,000 people were 473 for IHD, 501 for stroke, and 3,100 for MDD. The deaths per 100,000 people were estimated to be 32 for IHD, 59 for stroke, and 67 for suicide. The total DALY values per 100,000 people were estimated to be 563 for IHD, 1,048 for stroke, 1,240 for MDD, and 1,102 for suicide. The total BOD per 100,000 people in terms of job stress-attributable DALY were estimated to be 38 for IHD, 72 for stroke, 168 for MDD, and 44 for suicide (Table 2). The detailed data that were used to calculate these findings are shown in Table 3.

As a result of the present study, the PAFs (%) for IHD, ischemic stroke, hemorrhagic stroke, MDD, and suicide were estimated to be 6.72, 4.84, 2.08, 13.6, and 4.00, respectively. Also, job stress-attributable BODs were estimated to be 38 years per 100,000 people for IHD, 72 years per 100,000 people for stroke, 168 years per 100,000 people for MDD, and 44 years per 100,000 people for suicide.

Sultan-Taïeb et al.1718 reported that 8.8%–10.2% (men) and 0.2%–1.1% (women) of coronary heart disease morbidity and 15.2%–19.8% (men) and 14.3%–27.1% (women) of mental disorders (defined as clinical depression) were attributable to job strain in France. This French study was the only one that estimated the PAF of job strain. Similar to the present study, the French job stress exposure data were calculated using the Karasek model,22 which defines job strain as a combination of high psychological demands at work (i.e., task quantity and complexity, and time pressure) and low decision-making latitude (i.e., decision authority and skill discretion). However, direct comparison with the French results was difficult for some reasons: 1) Unlike the French study, we did not consider gender differences, 2) The prevalence of job strain exposure was different (19.6% for men and 28.2% for women came from the national survey conducted in France in 2003), and 3) We used the RR values given in the studies of a meta-analysis, while they derived RRs from their own systematic review.

The exposure prevalence of job strain among Korean workers was estimated to be 21.2%, based on data from a cross-sectional questionnaire survey of 7,856 Korean workers.23 As the PAF is very sensitive to changes in this value, care should be taken when selecting the data. We tried to reflect the Korean working environment and find the latest value at the same time; however, we had no choice but to select the result. This value would be representative, but there might be a gap with the current situation because it was from 2006. Several other studies have used different epidemiological data and reported similar findings: 19% among Finnish men and 23% among Finnish women,1934 18.6% among Australian men and 25.5% among Australian women,35 and 23.2% among the general French population (19.6% for men and 28.2% for women).36 Although we were unable to provide sex-specific Korean estimates, our data provide useful perspective regarding the Korean working environment.

In the present study, we estimated the prevalence of IHD and stroke using 2011 claims data from the Health Insurance Review and Assessment Service, and also estimated the MDD prevalence based on data from the Korean suicide prevention center. According to the 2011 Korea National Health and Nutrition Examination Survey report, the prevalence of physician-diagnosed disease were 2.4% for angina or myocardial infarction among > 30-year-old individuals, 3.8% for stroke among > 50-year-old individuals, and 4.5% for depression among > 30-year-old individuals.37 However, these values were noticeably different from our estimates: 0.6% for IHD among > 30-year-old individuals, 1.3% for stroke among > 50-year-old individuals, and 3.9% for MDD among > 30-year-old individuals. This result was dependent on the respondents' reports and did not cover detailed diagnosis that we would like to estimate. On the other hand, our results would be more relevant in that they contained detailed and accurate diagnoses, and especially for stroke, we analyzed the claims data entered by doctors, not survey data. However, since the health insurance data was for the general population rather than workers, strictly speaking, the estimates do not reflect the prevalence among workers. Thus, it is possible that the prevalence was overestimated. We could not estimate the accurate prevalence among Korean workers by analyzing the claims data from the Health Insurance Review and Assessment Service. Self-employed workers, who accounted for 23.1% of the economically active population in 201138 and laborers are classified as district subscribers, which means they could not be distinguished from the general population.

In the GBD report, Korea was included in a group of countries with DALY values of 600–1,200 per 100,000 person-years for stroke and 0–300 per 100,000 person-years for IHD.39 The present study estimated a relatively high DALY value for IHD, although the DALY value was similar for stroke. However, it was difficult to guess why the DALY for IHD in this study was high, because we could not figure out which data was used to estimate the DALYs. Similarly, Hong et al.40 have reported that the Korean DALY for ischemic stroke was 483 per 100,000 person-years. This result was obtained from using the incidence of ischemic stroke by analyzing the health insurance data of 2004. Unlike these studies, we used the prevalence, not the incidence that was commonly used in estimating BOD. It was also impossible to measure the portion of job stress-attributable BOD among the total DALY.

The assessment of job stress-related health effects should be performed based on an understanding of the specific working environments and situations in each country. However, there were limited Korean studies regarding this topic, which made it difficult to estimate job stress-related BOD using domestic data. Inevitably, the present study has several limitations that should be considered. First, we did not consider other diseases that are relevant to job strain, such as musculoskeletal disorders, because it was difficult to access prevalence data and reasonable RR values of many types of musculoskeletal disorders to calculate DALY. In addition, musculoskeletal disorders were excluded from the analysis because of possible reverse causation. Second, the data were not analyzed according to sex. However, it is reasonable to estimate the BOD with DALY as it is obtained by summation and there were no available study results to measure job stress exposure proportion by sex. Third, our result may be overestimated, and not accurately reflect the Korean working environment, as it was impossible to access data on prevalence or death rate targeting the Korean working population. Fourth, we replaced the RR with an OR for depression and with a hazard ratio for stroke. However, the prevalences of diseases were so low that the results obtained by using OR were not expected to make a large difference. Fifth, the proportion of workers exposed to job stress may not reflect the current working environment as there is a gap between the time of data collection and the present. Despite these limitations, the present study provides a useful perspective regarding the Korean working environment, and was designed to estimate the outcomes in a way that closely reflected this working environment. In particular, we tried to increase the validity of the calculated value by using representative data such as claim data from the Health Insurance Review and Assessment Service.

Despite several limitations due to data, the present study would be meaningful in that studies on PAF or BOD for specific causes were rare in Korea, and especially since studies on the calculation of BOD from job stress were also rare globally. However, in order to estimate the BOD more accurately, further studies would be necessary to: 1) Estimate the disease prevalence according to the employment type by using nationally representative data such as data of the National Health Insurance Service, 2) Calculate accurate RR values through a large-scale cohort study of workers, and 3) Evaluate the proportions of workers exposed to other job stressors including job strain.

BOD could be used to quantitatively estimate major health impacts and could be applied as the main outcome of public health policy, unlike other estimates of health impact such as mortality or morbidity. This is the first study to attempt to estimate the Korean BOD from job stress using the latest and best available data. As society becomes more concerned about suicides or deaths due to cerebrovascular disease in the working-age population in relation to burden at the workplace, estimating the BOD from job stress is essential in determining the priority of the national policy on prevention.