Characteristics of Korean Employees Without Depression but Having Suicidal Ideation

Article information

Psychiatry Investig. 2023;20(7):644-654
Publication date (electronic) : 2023 July 7
doi : https://doi.org/10.30773/pi.2023.0035
1Department of Psychiatry, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
2Workplace Mental Health Institute, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
Correspondence: Sung Joon Cho, MD, PhD Department of Psychiatry and Workplace Mental Health Institute, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 29 Saemunan-ro, Jongno-gu, Seoul 03181, Republic of Korea Tel: +82-2-2001-2214, Fax: +82-2-2001-2211, E-mail: sjcho0812@hanmail.net
Correspondence: Sang Won Jeon, MD, PhD Department of Psychiatry and Workplace Mental Health Institute, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 29 Saemunan-ro, Jongno-gu, Seoul 03181, Republic of Korea Tel: +82-2-2001-2214, Fax: +82-2-2001-2211, E-mail: sangwonyda@hanmail.net
Received 2023 February 2; Revised 2023 April 8; Accepted 2023 April 21.

Abstract

Objective

The aim of this study was to identify the factors related to suicidal ideation targeting the risk group showing suicidal ideation despite the absence of depression in Korean workers.

Methods

The data of 14,425 participants who were employees aged of 18 to 75 years who attended a mental health checkup program at the Workplace Mental Health Institute, Kangbuk Samsung Hospital from June 2015 to October 2019 were analyzed. A self-report questionnaire consisting of sociodemographic factors, suicidal ideation, job stress, levels of depression and anxiety, and resilience was administered. A Hierarchical logistic regression model was used with suicidal ideation as dependent variable. Separate analyses were conducted according to depressive symptoms using the 20-item Center for Epidemiological Studies Depression (CES-D) scale.

Results

Being women, older, and having low resilience, more perceived stress, more severe anxiety and less sleeping hours were associated with suicidal ideation in no-depression group (CES-D <16). In the subcategories of job stress, lack of reward was significantly associated with suicidal ideation in no-depression group.

Conclusion

This study identified the characteristics of a group that has no depression but has suicidal ideation in Korean workers. Among job stress items, lack of reward is a clear characteristic to be considered with caution in this group.

INTRODUCTION

Suicidal ideation is not always accompanied by depression. A study reported that up to 4% of people who did not suffer from a mood disorder had suicidal ideation [1], and another reported that 3.2% of those who did not experience a depressive episode had suicidal ideation [2]. Suicide risk is usually addressed through depression screening, but predicting suicide with a depression scale alone has a limitation because not all who have suicidal ideation or plan or attempt suicide will experience depression clinically [1,3]. Previous studies have emphasized the importance of evaluating suicide risk not by solely using a depression screening test for clinicians [4]. Furthermore, the spectrum of suicide risk (suicidal ideation, plans, and attempts) should be dealt with independently and not as a symptom of major depressive disorder or borderline personality disorder [5]. It is also suggested that the absence of major depression should not be misunderstood as the absence of suicide risk and that efforts should be made to prevent it [6].

Suicide risk of workers is influenced by numerous clinical and psychological variables, such as individual vulnerabilities, external stressors, and working conditions. Previous studies on suicidal ideation targeting office workers have revealed that high levels of work stress were associated with suicidal behavior [7,8]. Previous studies also have emphasized that depression is a key risk factor for suicidal ideation [9,10]. Therefore it might be concluded that the individuals who have severe depression experience suicidal ideation. However, such an approach could ignore the individuals at risk without depression. One study found that the approximately one-third of adolescents who did not have depression reported their having suicide ideation [11]. In a psychological autopsy study of elderly from suicide, 27% of individuals who had no depression committed suicide [12]. These data support the possible risk of population without depression in the experience of suicidal ideation.

However most companies’ mental health screening programs use depression scales to set interest group, and as a result, the group having suicidal ideation without depression is not included in the interest group, so there is a limit to predicting suicide risk. If these risk groups are excluded from the interest group, it may increase the suicide rate in the workplace. This study compared those who had suicidal ideation without depression to those who had suicidal ideation with depression from the demographic, internal, extrinsic, and social perspective.

Moreover, the sociodemographic characteristics of the office workers showing suicidal ideation despite the absence of depression have not yet been investigated. If various factors affecting these can be identified and risk signals can be captured, those risk groups can be classified for prevention in advance, enabling follow-up management. Therefore, by confirming this, the risk of suicide in office workers can be evaluated for prevention at an early stage.

METHODS

Study participants

This study was part of the Kangbuk Samsung Workplace Mental Health Study, a cross-sectional study of men and women Koreans aged 18–75 years who attended a mental health checkup program, which was called Sim Care program, at the Workplace Mental Health Institute, Kangbuk Samsung Hospital, in Seoul, Republic of Korea. Participants comprised employees from 54 companies and local government organizations who voluntarily participated in mental health examinations upon invitation from their companies.

The present study was conducted between June 2015 and October 2019 with an initial 15,360 respondents. We excluded those with incomplete questionnaires or missing sociodemographic information. The study was approved by the Institutional Review Board of Kangbuk Samsung Hospital (IRB no. KBSMC 2019-01-042), which waived the requirement for informed consent, as we only used de-identified data routinely collected in the workplace mental health screening checkups.

Measures

Suicidal ideation

Suicidal ideation was assessed using a self-reported dichotomous question asking whether the respondent had ever seriously wanted to commit suicide during the past year (“Over the last year, have you ever felt you would be better off dead?”), with response options of “yes” or “no.” The question is a part of Korea National Health and Nutrition Examination Survey by the Korean government to investigate the level of public health [13].

Depressive symptoms

Depressive symptoms were assessed using the Korean version of the 20-item Center for Epidemiological Studies Depression (CES-D) scale. It is a self-reported questionnaire with responses measured on a 4-point Likert scale ranging from 0 to 3 points [14,15].

Anxiety

Anxiety was assessed with the Clinically Useful Anxiety Outcome Scale (CUXOS), a self-report with 45 items. The total scores were categorized into “minimal” (0–15), “mild to moderate” (16–44), and “severe” (45–80) [16,17].

The sociodemographic factors

The sociodemographic factors collected for the purpose of this study were age, gender, educational level, marital status, sleeping hours, job grade, and years at the current job (job duration). Job grade was categorized into 3 groups: senior (executives, general managers, deputy general managers, and managers), staff (assistant manager, senior staff, and staff), and others.

Stress

Daily perceived stress was measured using the Korean version of the 10-item Perceived Stress Scale (PSS), which measures the degree to which situations in one’s life are appraised as stressful. It is a self-reported questionnaire with responses scored on a 5-point Likert scale ranging from 1=never to 5=very often [18]. The daily stress factor was also evaluated as a 5-point Likert scale named Daily Life Stressors Scale by borrowing the item that evaluates the cause of stress among the stress measurement tools of the National Health and Nutrition Survey [13].

Resilience

The Korean version of the Connor-Davidson Resilience Scale (K-CD-RISC) was used to measure resilience. The self-reported K-CD-RISC consists of 25 items measured on a 5-point Likert scale ranging from 0 to 4 points [19].

Alcohol use

We used the Alcohol Use Disorder Identification Test standardized for Koreans (AUDIT-K) [20]. It consists of 10 items comprising recent alcohol use, alcohol dependence, and alcohol related problems. The scores were categorized into “low-risk” (0–7), “medium-risk” (8–15), and “substantial to severe-risk” (16 and above 35) [20].

Occupational stress

Occupational stress was measured using the Korean Occupational Stress Scale (KOSS). This scale consists of 27 items in seven subcategories: job demands, insufficient job control, interpersonal conflict, job insecurity, organizational system, lack of reward, and organizational culture. Each item is scored from 1 to 4 [21].

Statistical analysis

The collected data were analyzed using PASW statistics for Windows (version 18.0; SPSS Inc., Chicago, IL, USA). Participants were divided into groups with and without suicide ideation, and demographic characteristics were analyzed using independent samples t-test and chi-square test. After that, the effect of each variable on suicidal ideation was analyzed through hierarchical logistic regression analysis. The effects on suicide ideation were analyzed through hierarchical logistic regression analysis using subitems of job stress, resilience, alcohol problems, familial, health, financial, interpersonal relationship problems, anxiety, perceived daily stress, and sleeping hours as independent variables and suicide ideation as dependent variables for each depression group and no-depression group.

CES-D cutoff point 16 has traditionally been used as the optimal cutoff point for detecting whether a person has depression [22]. Based on another study showing that top 20% of the total score of entire population suffers from clinically significant depressive symptoms, a cutoff of 16 points has been proposed for the top 20% of total scores for the population as a whole [23]. So 16 points of CES-D has been used to detect depressive symptoms for screening most effectively. But previous study conducted on Koreans, the cutoff point 25 was suggested that it is optimal for corresponding the clinical diagnosis of major depression rather than screening for everyday depressive symptoms [14]. Thus, in this study, CES-D ≥25 cutoff value was used to detect case-level depression. CES-D score 16 was used to rule out depression to perform a more accurate analysis, so who have the CES-D score lower than 16 was defined as no-depression group (CES-D <16). All significance of statistics were set at p<0.05.

In Model 1, the association between sociodemographic factors and suicidal ideation was tested. Resilience, alcohol problem, familial, health, financial, and interpersonal relationship problem which are characteristics that tend to be hard to change were added to Model 2, anxiety and daily perceived stress, sleeping hours to Model 3, and finally, each items of occupational stress in sort form of the KOSS (KOSS-SF) to Model 4 to test the associations between job stress and suicidal ideation.

RESULTS

Characteristics of the participants

Table 1 summarizes the characteristics of the participants according to depression. The mean age of the 14,425 participants was 43.20 years (standard deviation=9.46, range=18–75), and 58.6% were men. The age distribution was 18–29 years: 5.4%; 30–39 years: 34.8%; 40–49 years: 30.8%; 50–59 years: 23.7%; and ≥60 years: 5.2%. The no depression group comprised 92.0% of the participants and of those, 19.5% participants were with suicidal ideation. The mean sum of KOSS-SF scores was 61.36±9.48.

Sociodemographic and psychological characteristics of the study population

Compared to the participants without depression, participants with depression were more likely to be younger, women, not-married, have lower K-CD-RISC scores than those without depression. But educational level, job duration and position were not significantly associated with depression. Participants with depression have higher AUDIT-K score, more familial, health, financial, and interpersonal relationship problem than those without depression. They also showed more anxiety, higher PSS, shorter sleeping hours, and higher sum of KOSS-SF scores. But KOSS-SF score divided into each item was not significantly associated.

Hierarchical logistic regression analyses of depression group (CES-D ≥25) with suicidal ideation as the dependent variable

The results of the hierarchical logistic regression analyses of the association of suicidal ideation for depression group with sociodemographic factors, job related factors, resilience, alcohol problem, familial, health, financial, interpersonal relationship problems, anxiety, perceived daily stress, sleeping hours, and job stress factors are summarized in Table 2. In the first model, sociodemographic variables are not statistically significant. In the second model explained approximately 13.2% of variance of suicidal ideation. Identifying K-CD-RISC scores, AUDIT-K scores, familial problems, financial, and interpersonal relationship problems were associated with suicidal ideation beyond the effects of the sociodemographic factors. Anxiety, PSS, and sleeping hours tested in Model 3, explained an additional 1.3% of the variance in suicidal ideation. A positive association was found between PSS and suicidal ideation (p= 0.001). In the 4th model, the addition of each item of KOSS-SF explained an additional 1.1% of variance and among those, “occupational climate” was associated with suicidal ideation. The standardized coefficient (β) defines relative influence of the significant variables. In the analysis of all our models, KCD-RISC score (β=-0.199), AUDIT-K score (β=0.069), family problem (β=0.152), financial problem (β=0.108), interpersonal relationship problem (β=0.084), PSS (β=0.098), and “occupational climate” (β=-0.075) were statistically associated with suicidal ideation.

Hierarchical logistic regression analyses of no-depression group (CES-D <25) with suicidal ideation as the dependent variable

Hierarchical logistic regression analyses of no-depression group (CES-D <25) with suicidal ideation as the dependent variable

The results of the hierarchical logistic regression analyses of the association of suicidal ideation for no-depression group with sociodemographic factors, job related factors, resilience, alcohol problem, familial, health, financial, interpersonal relationship problems, anxiety, perceived daily stress, sleeping hours, and job stress factors are summarized in Table 3. In the first model, sociodemographic variables explained 1.0% of the variance in suicidal ideation. Increasing age (p=0.048) and woman (p<0.001) were associated with suicidal ideation. In the second model explained approximately 10.4% of variance of suicidal ideation. Identifying all variables including K-CD-RISC scores, AUDIT-K scores, familial, health, financial, and interpersonal relationship problems were associated with suicidal ideation beyond the effects of the sociodemographic factors. Anxiety, PSS, and sleeping hours tested in Model 3, explained an additional 0.8% of the variance in suicidal ideation. Anxiety, PSS, and sleeping hours were associated with suicidal ideation (p<0.001). In the 4th model, the addition of each item of KOSS-SF also explained 11.2% of variance, adding no power to Model 3. There were no items statistically significant with suicidal ideation, but “lack of reward” was at the margin of statistical significance (p=0.083). The standardized coefficient (β) defines relative influence of the significant variables. In the analysis of all our models, age (β=0.028), being a woman (β=0.071), K-CD-RISC score (β=-0.152), AUDIT-K score (β=0.051), family problem (β=0.113), health problem (β=0.030), financial problem (β=0.079), interpersonal relationship problem (β=0.046), anxiety (β=0.057), PSS (β=0.052), and sleeping hours (β=-0.038) were statistically associated with suicidal ideation, while all KOSS-SF subitems were not associated.

Hierarchical logistic regression analyses of no-depression group (CES-D <25) with suicidal ideation as the dependent variable

Compared to depression group, no-depression group (CESD <25) in Table 3 were more likely to be older and woman. Health problem, anxiety, and sleep hours were not associated with suicidal ideation in depression group, but in no-depression group, they are related with suicidal ideation. Finally, “occupational climate” was associated with suicidal ideation in depression group, but in no-depression group, it was not significantly associated with suicidal ideation.

Hierarchical logistic regression analyses of no-depression group (CES-D <16) with suicidal ideation as the dependent variable

We performed same analysis for no-depression group (CES-D <16) and the results of the hierarchical logistic regression analyses of the association of suicidal ideation are summarized in Table 4. In the first model, sociodemographic variables explained 1.2% of the variance in suicidal ideation. Increasing age (p=0.010) and woman (p<0.001) were associated with suicidal ideation. In the second model explained an additional 15.0% of variance in suicidal ideation. Identifying all variables including K-CD-RISC scores, AUDIT-K scores, familial, health, financial, and interpersonal relationship problems were associated with suicidal ideation beyond the effects of the sociodemographic factors. In Model 3, anxiety, PSS, and sleeping hours were tested, and explained an additional 2.1% of the variance in suicidal ideation. Anxiety, PSS, and sleeping hours were associated with suicidal ideation (p<0.001). In Model 4, the addition of each item of KOSS-SF explained an additional 0.2% of variance. Among the KOSS-SF items, “lack of reward” (p=0.012) was slightly associated with suicidal ideation. The standardized coefficient (β) defines relative influence of the significant variables. In the analysis of all our models, age (β=0.026), being women (β=0.075), K-CD-RISC score (β=-0.169), AUDIT-K score (β=0.050), family problem (β=0.125), health problem (β=0.029), financial problem (β=0.068), interpersonal relationship problem (β=0.077), anxiety level (β= 0.048), PSS (β=0.067), sleeping hours (β=-0.042), and “lack of reward” (β=0.035) were statistically associated with suicidal ideation.

Hierarchical logistic regression analyses of no-depression group (CES-D <16) with suicidal ideation as the dependent variable

Compared to Tables 3 and 4 with no-depression group (CES-D <16) which was lowered the CES-D cutoff score showed that only “lack of reward” was different and it was significantly associated with suicidal ideation. Similar to comparing Tables 2 and 3, when comparing Tables 2 and 4, participants with nodepression group (CES-D <16) were likely to be older and woman. In no-depression group (CES-D <16), health problem, anxiety level, and sleep hours were associated with suicidal ideation while they are not associated with suicidal ideation in depression group. Finally, “occupational climate” was not associated with suicidal ideation but “lack of reward” was associated in no-depression group (CES-D <16).

DISCUSSION

Among 14,425 office workers and manufacturing workers at large corporations, this study investigated risk factors related to suicidal ideation in a population that did not have depression but had suicidal ideation. After dividing the presence or absence of depression using the CES-D scale, a separate analysis was performed to determine the factors related to suicidal ideation. Compared with the depressed group (CES-D ≥25), women, older age, anxiety, multiple health problems, and sleeping less were associated with suicidal ideation in both the probable no-depression group (CES-D <25) and the definite no-depression group (CES-D <16). None of the items among the subcategories of job stress exhibited a significant association with suicidal ideation in the probable no-depression group. In the definite no-depression group, “lack of reward” was significantly associated with suicidal ideation. When the study subjects were narrowed down with clearer criteria, significant values were found, suggesting that this item of “lack of reward” is a clear characteristic to be considered with caution in the no-depression group.

We noticed that the group that had suicidal ideation without depression included 2,813 people, accounting for 16.2%. Using the National Survey on Drug Use and Health from 2008 to 2012 in the US, an epidemiologic study showed suicidal ideation over the past 12 months among the population aged 18 years and over and reported that the population with suicidal ideation among the no-depressed group was 3.2% [2]. A 10-year follow-up study from 1994 conducted in the Baltimore Epidemiologic Area Study found that 4% of 753 participants without depression had suicidal ideation [1]. These differences are due to the fact that our study consisted of a group with relatively weak mental health because many companies and public institutions who felt a crisis in the mental health management of workers requested the Sim Care Program, an integrated mental health management system for employees. In addition, including only Korean workers and using different symptom measurement tools could be other reasons that influence the different distribution from previous studies.

Our results are consistent with the results of previous studies that women and those with higher anxiety are more likely to experience suicidal ideation. Women in the workforce are at higher risk of suicide than men [3], and even though anxiety has high comorbidity with depression, anxiety disorder alone is known to be associated with suicidal ideation [24,25]. It is also known that suicidal ideation were associated with an increased number of health problems and age regardless of depressive symptoms, which is consistent with previous studies revealing that the probability of physical disease increases with increasing age, and that physical disease has a major influence on suicidal ideation [26]. The association between lack of sleep and suicidal ideation is consistent with previous studies [27].

On the other hand, differences in educational background, marital status, job position, and length of work were not related to suicidal ideation among office workers without depression. There are several contradictory studies on the effects of education [28,29] and working hours [30-32] on suicidal ideation in the general population, and studies about job position on suicidal ideation are still lacking.

In terms of job stress in definite no-depression group, “lack of reward” showed a stronger relationship with suicidal ideation than other subcategories of job stress, which is consistent with previous studies that lack of reward is a predictor that can harm mental and physical health due to high stress. Reward is a protective factor for anxiety and depression, and great external and internal efforts that are not rewarded appropriately generate stress and cause mental problems [33]. “Lack of reward” as subcategories of job stress evaluates whether the expected compensation level for work is appropriate or not. It also measures how much the individuals feel respected, have intrinsic motivation, and satisfy their expectations [21]. Previous studies reported that a lack of reward for effort could predict suicide [7,34]. One meta-analysis study collected six sources of epidemiological data, and the study population was mostly in their early mid-40s under the age of 50 years from Republic of Korea, Australia, China, and Germany. The gender ratio was close to 1:1, and more than two-thirds of the subjects were married. Most subjects had professional jobs with relatively high socioeconomic status. The probability of having suicidal ideation increased as the reward imbalance increased, and was twice as high as those with a low reward imbalance ratio. However, this study did not control other potential mediating factors associated with depression that could be considered predictors of suicide, such as excessive alcohol use, health problems such as chronic pain, and family conflict [7].

Another report found that a low reward relative to a high effort was associated with a risk of suicidal ideation. This prospective cohort study of 4,963 workers aged 50 years and over who had no suicidal ideation in 11 European countries was conducted, following up for eight years to determine the occurrence of suicidal ideation. The low reward for effort, suicidal ideation, and depressive symptoms were evaluated, and the mediating effect of depressive symptoms was tested. The high effort-low reward imbalance led to a significantly higher risk of suicidal ideation than the low effort-high reward combination. Furthermore, depression indirectly mediated between reward imbalance and suicidal ideation (indirect effect 14.4%) [35]. Another study found that higher occupational demands and lack of reward among male workers were associated with the risk of suicidal ideation in younger and middle-aged groups [36]. Repeated severe frustration of core rewards due to unequal and unfair treatment at work can lead to feelings of humiliation, lack of self-esteem, and hopelessness, and also can lead to depression, but even in the absence of depressive symptoms, exceeding personal tolerance can directly lead to more serious consequences, such as suicidal ideation.

This study has some limitations. First, it may lack a causal relationship as a cross-sectional study. In fact, in a previous cohort study of the general population, the population who did not have depression but had suicidal ideation displayed decreased concentration, feelings of unhappiness, and social withdrawal, which can be regarded as subclinical depression that does not meet clinically significant criteria [1,37].

Second, since this study only investigated the presence or absence of suicidal ideation, there was no evaluation of the severity of such ideation and likelihood of attempting suicide. Additional research is needed to determine the risk of suicide, including whether suicide is planned or attempted.

Third, since the evaluation was conducted using a self-report questionnaire, there may be faking-good responses, such as the study group underreporting suicidal ideation. Finally, the generalizability of the research results is limited as most of the subjects were office workers.

Despite these limitations, the strength of this study is that the number of participants is large, and various internal and external variables related to suicide were used, including work-related factors, sociostatistical factors, and psychopathological factors such as depression and anxiety. Lastly, since the data was obtained through the workplace health promotion testing program, the process, method, and confidentiality of the test were thoroughly conducted before, during, and after the tests, so the participation rate was high, and the insincere response rate was less than 3%.

Conclusion

Among workers, the group who had suicidal ideation even without depression displayed a significant association between woman, older age, increased anxiety, health problems, lack of sleep, and suicidal ideation. In the definite no-depression group with a CES-D score of less than 16 than the group with a CES-D score of less than 25, “lack of reward” among the subitems of job stress clearly showed a significant association with suicidal ideation. For mental health evaluation of workers, people with depression are the main target of evaluation, excluding many of the suicide risk groups if they do not have depression. To detect and manage such a specific population, it is necessary to broaden our concept of risk groups, and by understanding the factors related to suicide in such groups, the risk of suicide can be evaluated and prevented early.

Notes

Availability of Data and Material

The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.

Conflicts of Interest

The authors have no potential conflicts of interest to disclose.

Author Contributions

Conceptualization: Sung Joon Cho, Sang Won Jeon. Data curation: Hye Jeong Jeon. Formal analysis: Hye Jeong Jeon, Sang Won Jeon. Investigation: Hye Jeong Jeon, Sung Joon Cho. Methodology: Sung Joon Cho, Sang Won Jeon. Project administration: Young Chul Shin. Validation: Kang Seob Oh, Sung Joon Cho, Sang Won Jeon. Writing—original draft: Hye Jeong Jeon. Writing—review & editing: Kang Seob Oh, Young Chul Shin, Dong Won Shin, Sung Joon Cho, Sang Won Jeon.

Funding Statement

None

Acknowledgements

We are grateful to the workers who participated in this study and the doctors who worked together.

References

1. Joo J, Hwang S, Gallo JJ. Death ideation and suicidal ideation in a community sample who do not meet criteria for major depression. Crisis 2016;37:161–165.
2. Han B, Compton WM, Gfroerer J, McKeon R. Prevalence and correlates of past 12-month suicide attempt among adults with past-year suicidal ideation in the United States. J Clin Psychiatry 2015;76:295–302.
3. Harmer B, Lee S, Duong TVH, Saadabadi AS. Suicidal ideation Treasure Island, FL: StatPearls Publishing; 2022.
4. Mournet AM, Smith JT, Bridge JA, Boudreaux ED, Snyder DJ, Claassen CA, et al. Limitations of screening for depression as a proxy for suicide risk in adult medical inpatients. J Acad Consult Liaison Psychiatry 2021;62:413–420.
5. Ahrens B, Linden M, Zäske H, Berzewski H. Suicidal behavior--symptom or disorder? Compr Psychiatry 2000;41(2 Suppl 1):116–121.
6. Batterham PJ, Christensen H. Longitudinal risk profiling for suicidal thoughts and behaviours in a community cohort using decision trees. J Affect Disord 2012;142:306–314.
7. Loerbroks A, Cho SI, Dollard MF, Zou J, Fischer JE, Jiang Y, et al. Associations between work stress and suicidal ideation: individual-participant data from six cross-sectional studies. J Psychosom Res 2016;90:62–69.
8. Milner A, Witt K, LaMontagne AD, Niedhammer I. Psychosocial job stressors and suicidality: a meta-analysis and systematic review. Occup Environ Med 2018;75:245–253.
9. Bonde JP. Psychosocial factors at work and risk of depression: a systematic review of the epidemiological evidence. Occup Environ Med 2008;65:438–445.
10. Takada M, Suzuki A, Shima S, Inoue K, Kazukawa S, Hojoh M. Associations between lifestyle factors, working environment, depressive symptoms and suicidal ideation: a large-scale study in Japan. Ind Health 2009;47:649–655.
11. Sihvola E, Keski-Rahkonen A, Dick DM, Pulkkinen L, Rose RJ, Marttunen M, et al. Minor depression in adolescence: phenomenology and clinical correlates. J Affect Disord 2007;97:211–218.
12. Alexopoulos GS. Psychological autopsy of an elderly suicide. Int J Geriat Psychiatry 1991;6:45–50.
13. Lee ES, Shin HC, Yang YJ, Cho JJ, Ahn KY, Kim SH. Development of the stress questionnaire for KNHANES: report of scientific study service Cheongju: Korea Centers for Disease Control and Prevention; 2010.
14. Cho MJ, Kim KH. Use of the Center for Epidemiologic Studies Depression (CES-D) Scale in Korea. J Nerv Ment Dis 1998;186:304–310.
15. Park JH, Kim KW. A review of the epidemiology of depression in Korea. J Korean Med Assoc 2011;54:362–369.
16. Zimmerman M, Chelminski I, Young D, Dalrymple K. A clinically useful anxiety outcome scale. J Clin Psychiatry 2010;71:534–542.
17. Jeon SW, Han C, Ko YH, Yoon S, Pae CU, Choi J, et al. A Korean validation study of the Clinically Useful Anxiety Outcome Scale: comorbidity and differentiation of anxiety and depressive disorders. PLoS One 2017;12:e0179247.
18. Lee J, Shin C, Ko YH, Lim J, Joe SH, Kim S, et al. The reliability and validity studies of the Korean version of the perceived stress scale. Korean J Psychosom Med 2012;20:127–134.
19. Baek HS, Lee KU, Joo EJ, Lee MY, Choi KS. Reliability and validity of the Korean version of the Connor-Davidson Resilience Scale. Psychiatry Investig 2010;7:109–115.
20. Lee BO, Lee CH, Lee PG, Choi MJ, Namkoong K. Development of Korean version of Alcohol Use Disorders Identification Test (AUDIT-K): its reliability and validity. J Korean Acad Addict Psychiatry 2000;4:83–92.
21. Chang SJ, Koh SB, Kang D, Kim SA, Kang MG, Lee CG, et al. Developing an Occupational Stress Scale for Korean employees. Korean J Occup Environ Med 2005;17:297–317.
22. Roberts RE, Rhoades HM, Vernon SW. Using the CES-D scale to screen for depression and anxiety: effects of language and ethnic status. Psychiatry Res 1990;31:69–83.
23. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977;1:385–401.
24. Norton PJ, Temple SR, Pettit JW. Suicidal ideation and anxiety disorders: elevated risk or artifact of comorbid depression? J Behav Ther Exp Psychiatry 2008;39:515–525.
25. Thibodeau MA, Welch PG, Sareen J, Asmundson GJ. Anxiety disorders are independently associated with suicide ideation and attempts: propensity score matching in two epidemiological samples. Depress Anxiety 2013;30:947–954.
26. Szanto K, Gildengers A, Mulsant BH, Brown G, Alexopoulos GS, Reynolds CF 3rd. Identification of suicidal ideation and prevention of suicidal behaviour in the elderly. Drugs Aging 2002;19:11–24.
27. Pigeon WR, Pinquart M, Conner K. Meta-analysis of sleep disturbance and suicidal thoughts and behaviors. J Clin Psychiatry 2012;73:e1160–e1167.
28. Pompili M, Vichi M, Qin P, Innamorati M, De Leo D, Girardi P. Does the level of education influence completed suicide? A nationwide register study. J Affect Disord 2013;147:437–440.
29. Phillips JA, Hempstead K. Differences in U.S. suicide rates by educational attainment, 2000–2014. Am J Prev Med 2017;53:e123–e130.
30. Choi B. Job strain, long work hours, and suicidal ideation in US workers: a longitudinal study. Int Arch Occup Environ Health 2018;91:865–875.
31. Lee HE, Kim I, Kim HR, Kawachi I. Association of long working hours with accidents and suicide mortality in Korea. Scand J Work Environ Health 2020;46:480–487.
32. Takeuchi A, Sakano N, Miyatake N. Combined effects of working hours, income, and leisure time on suicide in all 47 prefectures of Japan. Ind Health 2014;52:137–140.
33. de Jonge J, Bosma H, Peter R, Siegrist J. Job strain, effort-reward imbalance and employee well-being: a large-scale cross-sectional study. Soc Sci Med 2000;50:1317–1327.
34. Byun J, Kim HR, Lee HE, Kim SE, Lee J. Factors associated with suicide ideation among subway drivers in Korea. Ann Occup Environ Med 2016;28:31.
35. Zhuo LB, Yao W, Yan Z, Giron MST, Pei JJ, Wang HX. Impact of effort reward imbalance at work on suicidal ideation in ten European countries: the role of depressive symptoms. J Affect Disord 2020;260:214–221.
36. Kim SY, Shin YC, Oh KS, Shin DW, Lim WJ, Cho SJ, et al. Association between work stress and risk of suicidal ideation: a cohort study among Korean employees examining gender and age differences. Scand J Work Environ Health 2020;46:198–208.
37. Joiner TE Jr, Brown JS, Wingate LR. The psychology and neurobiology of suicidal behavior. Annu Rev Psychol 2005;56:287–314.

Article information Continued

Table 1.

Sociodemographic and psychological characteristics of the study population

Variables Total (N=14,425 [100%]) No depression group (N=13,278 [92.05%]) Depression group (N=1,147 [7.95%]) p
Age (yr) 43.20±9.46 43.43±9.48 40.61±8.85 <0.001*
Age distribution <0.001*
18–29 yr 775 (5.4) 675 (5.1) 100 (8.7)
30–39 yr 5,024 (34.8) 4,536 (34.2) 488 (42.5)
40–49 yr 4,447 (30.8) 4,099 (30.9) 348 (30.3)
50–59 yr 3,423 (23.7) 3,240 (24.4) 184 (16.0)
≥60 yr 756 (5.2) 728 (5.5) 28 (2.4)
Gender <0.001*
Men 8,455 (58.6) 7,956 (59.9) 499 (43.5)
Women 5,969 (41.4) 5,321 (40.1) 648 (56.5)
Education level 0.149
No bachelor’s degree 4,279 (31.4) 3,917 (31.3) 362 (33.4)
Bachelor’s degree 9,329 (68.6) 8,607 (68.7) 722 (66.6)
Marital status <0.001*
Unmarried 4,601 (33.5) 4,084 (32.3) 517 (47.4)
Married 8,806 (64.1) 8,280 (65.5) 526 (48.3)
Others (divorced, widowed) 324 (2.4) 277 (2.2) 47 (4.3)
Job duration (yr) 15.03±9.72 15.05±9.75 14.75±9.37 0.603
Job position 0.887
Senior 9,954 (69.0) 9,157 (69.0) 797 (69.5)
Junior 4,040 (28.0) 3,722 (28.0) 318 (27.7)
Other 431 (3.0) 399 (3.0) 32 (2.8)
K-CD-RISC score 63.63±15.88 64.72±15.33 51.04±16.71 <0.001*
AUDIT-K score 7.96±6.38 7.96±6.38 9.55±7.99 <0.001*
Familial problem 1.60±0.83 1.55±0.77 2.20±1.16 <0.001*
Health problem 1.59±0.78 1.54±0.73 2.17±1.02 <0.001*
Financial problem 1.70±0.86 1.64±0.81 2.29±1.13 <0.001*
Interpersonal relationship problem 1.40±0.70 1.35±0.63 1.96±1.05 <0.001*
CUXOS for Anxiety <0.001*
Minimal 9,393 (65.1) 9,014 (67.9) 379 (33.1)
Mild to moderate 4,596 (31.9) 4,024 (30.3) 572 (49.9)
Severe 435 (3.0) 240 (1.8) 195 (17.0)
PSS score 16.64±5.28 16.3±5.18 20.56±4.82 <0.001*
Sleeping hours 6.02±1.11 6.06±1.08 5.50±1.22 <0.001*
KOSS-SF score
Job demand 9.97±2.42 9.98±2.42 9.93±2.41 0.529
Insufficient job control 9.92±2.20 9.92±2.20 9.93±2.16 0.992
Interpersonal conflict 6.33±1.59 6.33±1.59 6.34±1.58 0.919
Job insecurity 3.93±1.47 3.92±1.47 3.95±1.50 0.692
Organizational system 9.62±2.12 9.61±2.12 9.69±2.16 0.342
Lack of reward 9.34±2.14 9.34±2.14 9.36±2.07 0.738
Occupational climate 6.26±1.73 6.26±1.73 6.25±1.71 0.941
Physical environment 5.98±1.68 5.98±1.68 5.99±1.68 0.724
KOSS-SF sum 61.36±9.48 60.59±9.07 70.22±9.60 0.004*
Suicidal ideation <0.001*
SI 2,819 (19.5) 2,148 (16.2) 671 (58.5)
No SI 11,605 (80.5) 11,129 (83.8) 476 (41.5)

Values are presented as mean±standard deviation or number (%).

*

p<0.05;

for categorical covariates, the p-value was generated using the chi-square test. For continuous covariates, p-values were generated from the analysis of variance.

K-CD-RISC, Korean version of the Conner-Davidson Resilience Scale; AUDIT-K, Alcohol Use Disorder Identification Test standardized for Koreans; CUXOS, Clinically Useful Anxiety Outcome Scale; PSS, Perceived Stress Scale; KOSS-SF, sort form of the Korean Occupational Stress Scale; SI, participants with suicidal ideation; No SI, participants without suicidal ideation

Table 2.

Hierarchical logistic regression analyses of no-depression group (CES-D <25) with suicidal ideation as the dependent variable

Variables Model 1
Model 2
Model 3
Model 4
B beta T p B beta T p B beta T p B beta T p
Age -0.021 -0.040 -1.023 0.307 -0.007 -0.013 -0.367 0.713 -0.005 -0.009 -0.244 0.807 -0.006 -0.011 -0.299 0.765
Gender, women 0.020 0.020 0.591 0.555 0.047 0.047 1.448 0.148 0.040 0.040 1.238 0.216 0.038 0.038 1.168 0.243
Education level -0.031 -0.029 -0.954 0.341 -0.004 -0.004 -0.142 0.887 -0.014 -0.014 -0.463 0.643 -0.007 -0.007 -0.234 0.815
Marital status, not married 0.010 0.012 0.320 0.749 0.005 0.006 0.177 0.860 0.001 0.002 0.044 0.965 0.003 0.003 0.093 0.926
Job duration 0.002 0.042 1.346 0.179 0.003 0.050 1.706 0.088 0.003 0.053 1.820 0.069 0.003 0.056 1.927 0.054
Job position, higher 0.012 0.013 0.422 0.673 0.025 0.026 0.911 0.362 0.023 0.024 0.849 0.396 0.029 0.031 0.999 0.318
Resilience (K-CD-RISC) -0.006 -0.190 -6.424 <0.001* -0.006 -0.198 -6.673 <0.001* -0.006 -0.199 -6.700 <0.001*
Alcohol (AUDIT-K) 0.005 0.082 2.754 0.006* 0.004 0.071 2.382 0.017* 0.004 0.069 2.307 0.021*
Family problem 0.068 0.160 5.092 <0.001* 0.068 0.158 5.061 <0.001* 0.065 0.152 4.870 <0.001*
Health problem 0.001 0.002 0.077 0.938 -0.009 -0.018 -0.568 0.570 -0.005 -0.010 -0.315 0.753
Financial problem 0.049 0.113 3.615 <0.001* 0.046 0.105 3.366 0.001* 0.047 0.108 3.454 0.001*
Interpersonal relationship problem 0.041 0.088 2.815 0.005* 0.040 0.084 2.694 0.007* 0.040 0.084 2.693 0.007*
Anxiety (CUXOS) 0.020 0.028 0.958 0.338 0.019 0.027 0.909 0.363
Perceived daily stress (PSS) 0.010 0.097 3.256 0.001* 0.010 0.098 3.286 0.001*
Sleeping hours -0.017 -0.042 -1.426 0.154 -0.017 -0.041 -1.420 0.156
Job demand 0.011 0.054 1.724 0.085
Insufficient job control 0.007 0.031 0.934 0.351
Interpersonal conflict 0.012 0.040 1.147 0.252
Job insecurity -0.011 -0.034 -1.113 0.266
Organizational system 0.006 0.028 0.662 0.508
Lack of reward -0.004 -0.017 -0.394 0.693
Occupational climate -0.022 -0.075 -2.133 0.033*
Physical environment -0.011 -0.036 -1.140 0.254
Statistics of the model F=35.408, p<0.001, R2=0.014 F=154.327, p<0.001, R2=0.130 F=12.044, p<0.001, R2=0.139 F=105.187, p<0.001, R2=0.140
R2 change=0.014 R2 change=0.117 R2 change=0.009 R2 change=0.001

Model 1: Sociodemographic factors. Model 2: Resilience, alcohol problem, familial, health, financial, and interpersonal relationship problems. Model 3: Anxiety, daily perceived stress, and sleeping hours. Model 4: Occupational stress (KOSS-SF).

*

p<0.05;

for categorical covariates, the p-value was generated using the chi-square test. For continuous covariates, p-values were generated from the analysis of variance.

CES-D, the 20-item Center for Epidemiological Studies Depression; K-CD-RISC, Korean version of the Conner-Davidson Resilience Scale; AUDIT-K, Alcohol Use Disorders Identification Test–Korea; CUXOS, Clinically Useful Anxiety Outcome Scale; PSS, Perceived Stress Scale; KOSS-SF, sort form of the Korean Occupational Stress Scale

Table 3.

Hierarchical logistic regression analyses of no-depression group (CES-D <25) with suicidal ideation as the dependent variable

Variables Model 1
Model 2
Model 3
Model 4
B beta T p B beta T p B beta T p B beta T p
Age -0.008 -0.022 -1.974 0.048* 0.010 0.026 2.488 0.013* 0.010 0.028 2.648 0.008* 0.010 0.028 2.666 0.008*
Gender, women 0.069 0.092 9.732 <0.001* 0.060 0.079 8.411 <0.001* 0.053 0.071 7.462 <0.001* 0.053 0.071 7.425 <0.001*
Education level -0.003 -0.004 -0.437 0.662 0.010 0.013 1.489 0.136 0.004 0.005 0.621 0.534 0.004 0.005 0.611 0.541
Marital status, not married -0.001 -0.001 -0.097 0.923 -0.006 -0.008 -0.817 0.414 -0.006 -0.008 -0.745 0.456 -0.006 -0.008 -0.775 0.438
Job duration 0.000 0.003 0.372 0.710 0.000 0.002 0.212 0.832 0.000 0.002 0.257 0.797 0.000 0.002 0.266 0.791
Job position, higher 0.001 0.001 0.115 0.908 0.003 0.004 0.428 0.669 0.002 0.004 0.419 0.675 0.005 0.007 0.751 0.452
Resilience (K-CD-RISC) -0.004 -0.151 -16.726 <0.001* -0.004 -0.152 -16.554 <0.001* -0.004 -0.152 -16.550 <0.001*
Alcohol (AUDIT-K) 0.003 0.059 6.620 <0.001* 0.003 0.050 5.640 <0.001* 0.003 0.051 5.639 <0.001*
Family problem 0.059 0.123 12.757 <0.001* 0.054 0.113 11.657 <0.001* 0.054 0.113 11.659 <0.001*
Health problem 0.022 0.044 4.771 <0.001* 0.015 0.030 3.183 0.001* 0.015 0.030 3.191 0.001*
Financial problem 0.039 0.086 8.974 <0.001* 0.035 0.079 8.178 <0.001* 0.036 0.079 8.177 <0.001*
Interpersonal relationship problem 0.030 0.052 5.450 <0.001* 0.026 0.046 4.745 <0.001* 0.026 0.046 4.741 <0.001*
Anxiety (CUXOS) 0.041 0.057 6.460 <0.001* 0.041 0.057 6.463 <0.001*
Perceived daily stress (PSS) 0.004 0.052 5.656 <0.001* 0.004 0.052 5.617 <0.001*
Sleeping hours -0.013 -0.038 -4.434 <0.001* -0.013 -0.038 -4.436 <0.001*
Job demand 0.000 0.001 0.120 0.904
Insufficient job control 0.000 0.000 0.001 0.999
Interpersonal conflict 0.001 0.006 0.586 0.558
Job insecurity -0.003 -0.011 -1.196 0.232
Organizational system -0.002 -0.009 -0.703 0.482
Lack of reward 0.004 0.022 1.734 0.083
Occupational climate -0.001 -0.003 -0.329 0.742
Physical environment 0.000 -0.002 -0.237 0.813
Statistics of the model F=35.408, p<0.001, R2=0.014 F=154.327, p<0.001, R2=0.130 F=128.124, p<0.001, R2=0.139 F=105.187, p<0.001, R2=0.140
R2 change=0.014 R2 change=0.117 R2 change=0.009 R2 change=0.001

Model 1: Sociodemographic factors. Model 2: Resilience, alcohol problem, familial, health, financial, and interpersonal relationship problems. Model 3: Anxiety, daily perceived stress, and sleeping hours. Model 4: Occupational stress (KOSS-SF).

*

p<0.05;

for categorical covariates, the p-value was generated using the chi-square test. For continuous covariates, p-values were generated from the analysis of variance.

CES-D, the 20-item Center for Epidemiological Studies Depression; K-CD-RISC, Korean version of the Conner-Davidson Resilience Scale; AUDIT-K, Alcohol Use Disorders Identification Test–Korea; CUXOS, Clinically Useful Anxiety Outcome Scale; PSS, Perceived Stress Scale; KOSS-SF, sort form of the Korean Occupational Stress Scale

Table 4.

Hierarchical logistic regression analyses of no-depression group (CES-D <16) with suicidal ideation as the dependent variable

Variables Model 1
Model 2
Model 3
Model 4
B beta T p B beta T p B beta T p B beta T p
Age -0.011 -0.031 -2.564 0.010* 0.009 0.023 2.013 0.044* 0.010 0.026 2.231 0.026* 0.010 0.026 2.256 0.024*
Gender, women 0.076 0.101 9.754 <0.001* 0.064 0.084 8.216 <0.001* 0.056 0.075 7.262 <0.001* 0.056 0.075 7.250 <0.001*
Education -0.002 -0.003 -0.295 0.768 0.015 0.019 1.988 0.047* 0.008 0.010 1.031 0.302 0.008 0.010 1.030 0.303
Marital status, not married -0.004 -0.005 -0.426 0.670 -0.006 -0.008 -0.681 0.496 -0.005 -0.007 -0.620 0.536 -0.005 -0.007 -0.633 0.527
Job duration 0.000 0.003 0.330 0.741 0.000 0.003 0.353 0.724 0.000 0.005 0.508 0.611 0.000 0.005 0.518 0.604
Job position, higher 0.006 0.009 0.861 0.389 0.007 0.010 1.078 0.281 0.006 0.009 1.010 0.312 0.011 0.015 1.523 0.128
Resilience (K-CD-RISC) -0.004 -0.166 -16.880 <0.001* -0.004 -0.169 -17.017 <0.001* -0.004 -0.169 -17.014 <0.001*
Alcohol (AUDIT-K) 0.003 0.060 6.187 <0.001* 0.003 0.050 5.190 <0.001* 0.003 0.050 5.208 <0.001*
Family problem 0.065 0.136 12.940 <0.001* 0.059 0.125 11.833 <0.001* 0.059 0.125 11.807 <0.001*
Health problem 0.022 0.045 4.389 <0.001* 0.014 0.029 2.795 0.005* 0.014 0.029 2.815 0.005*
Financial problem 0.035 0.077 7.319 <0.001* 0.031 0.068 6.491 <0.001* 0.031 0.068 6.488 <0.001*
Interpersonal relationship problem 0.049 0.086 8.183 <0.001* 0.043 0.077 7.296 <0.001* 0.043 0.077 7.249 <0.001*
Anxiety (CUXOS) 0.036 0.048 5.041 <0.001* 0.035 0.048 5.032 <0.001*
Perceived daily stress (PSS) 0.005 0.067 6.764 <0.001* 0.005 0.067 6.715 <0.001*
Sleeping hours -0.014 -0.042 -4.489 <0.001* -0.014 -0.042 -4.467 <0.001*
Job demand 0.000 0.002 0.225 0.822
Insufficient job control 0.000 0.001 0.073 0.942
Interpersonal conflict 0.000 0.000 0.012 0.990
Job insecurity -0.004 -0.016 -1.587 0.113
Organizational system -0.001 -0.006 -0.427 0.670
Lack of reward 0.006 0.035 2.516 0.012*
Occupational climate -0.003 -0.016 -1.365 0.172
Physical environment 0.000 -0.001 -0.098 0.922
Statistics of the model F=35.408, p<0.001, R2=0.014 F=154.327, p<0.001, R2=0.130 F=128.124, p<0.001, R2=0.139 F=105.187, p<0.001, R2=0.140
R2 change=0.014 R2 change=0.117 R2 change=0.009 R2 change=0.001

Model 1: Sociodemographic factors. Model 2: Resilience, alcohol problem, familial, health, financial, and interpersonal relationship problems. Model 3: Anxiety, daily perceived stress, and sleeping hours. Model 4: Occupational stress (KOSS-SF).

*

p<0.05;

for categorical covariates, the p-value was generated using the chi-square test. For continuous covariates, p-values were generated from the analysis of variance.

CES-D, the 20-item Center for Epidemiological Studies Depression; K-CD-RISC, Korean version of the Conner-Davidson Resilience Scale; AUDIT-K, Alcohol Use Disorders Identification Test–Korea; CUXOS, Clinically Useful Anxiety Outcome Scale; PSS, Perceived Stress Scale; KOSS-SF, sort form of the Korean Occupational Stress Scale