Changes in the proportion of suicides attributed to work-related problems before, during, and after the COVID-19 pandemic in Japan

Abstract

Objectives

To examine changes in suicides attributed to work-related problems in Japan before, during, and after the COVID-19 pandemic, with attention to sex differences.

Methods

Monthly suicide data from January 2017 to February 2026 were obtained from the Ministry of Health, Labour and Welfare of Japan. The study period was divided into four phases: pre-emergency period (January 2017–March 2020), COVID-19 period (April 2020–December 2021), COVID-19 period after classification change (January 2022–May 2023), and post-COVID-19 period (June 2023–February 2026). Negative binomial regression models including the logarithm of monthly total suicides as an offset were used to examine the proportion of suicides attributed to work-related problems among all suicides according to period and sex. Supplementary analyses examined within-period temporal trends.

Results

During the study period, 188,178 suicides were recorded, including 20,822 recorded as involving work-related problems. In the offset model, among men, the proportion of suicides attributed to work-related problems was lower during the COVID-19 period than during the pre-emergency period (incidence rate ratio [IRR], 0.920; 95% confidence interval [CI], 0.859–0.986), but higher during the COVID-19 period after classification change (IRR, 1.389; 95% CI, 1.297–1.486) and the post-COVID-19 period (IRR, 1.261; 95% CI, 1.184–1.342). The relative change was more positive among women during the COVID-19 period and the post-COVID-19 period.

Conclusions

Work-related problems were more frequently represented in suicide statistics after the onset of the COVID-19 pandemic in Japan, although part of the increase after 2022 may reflect the revision of official classification.

Introduction

Suicide remains a major public health issue in Japan, and its overall pattern changed during the COVID-19 pandemic, with a greater impact observed among women^1,2,3,4). A recent study has suggested that the pandemic influenced not only the number of suicides but also the reasons associated with suicide, including work-related problems⁵⁾. Whereas international findings showed no uniform increase in suicide mortality during the early phase of the pandemic, the pattern observed in Japan may have been particularly marked among women^2,3,4,6,7).

Work-related mental health problems, including overwork-related suicide, are an important target of occupational health policy and practice in Japan⁸⁾. Furthermore, recent national discussions on workplace mental health have highlighted the need to strengthen preventive measures in small-scale enterprises, including the planned expansion of mandatory stress checks to workplaces with fewer than 50 employees⁹⁾.

Trends and sex differences in suicides identified as involving work-related problems may therefore serve as a useful indicator for considering occupational mental health responses to pandemic-related increases in suicide. Accordingly, this study examined monthly national suicide data in Japan from January 2017 to February 2026 to assess changes in suicides attributed to work-related problems before, during, and after the COVID-19 pandemic, with attention to sex differences.

Methods

Data source and study periods

In Japan, national suicide mortality can be described using two official statistics: vital statistics compiled by the Ministry of Health, Labour and Welfare and suicide statistics compiled by the National Police Agency. Because only the latter contain information on reasons for suicide, the present study used the suicide statistics published on the Ministry of Health, Labour and Welfare website. Monthly data from January 2017 to February 2026 were analyzed.

We analyzed monthly counts of suicides attributed to work-related problems and total suicides separately by sex. The study period was divided into four phases: pre-emergency period (January 2017 to March 2020), COVID-19 period (April 2020 to December 2021), COVID-19 period after classification change (January 2022 to May 2023), and post-COVID-19 period (June 2023 to February 2026). March 2020 was defined as the end of the pre-emergency period because the first state of emergency in Japan was declared in April 2020. January 2022 to May 2023 was treated as a separate period because the official method for classifying reasons for suicide was revised in 2022. June 2023 onward was considered the post-COVID-19 period to examine whether the observed patterns remained after the period most directly affected by both the pandemic and the revision of classification practice.

Before the 2022 revision, up to three ascertainable reasons for each suicide could be recorded only when documentary evidence, such as a suicide note or other material confirming the reasons, was available. From January 2022 onward, up to four ascertainable reasons could be recorded, and reasons could also be identified on the basis of testimony from family members or other informants even in the absence of documentary evidence. Because this revision may have affected the ascertainment of work-related problems, the period from January 2022 onward was treated separately in the analysis.

Statistical analysis

Descriptive statistics were calculated for each sex and period, including the monthly total number of suicides, the monthly number of suicides attributed to work-related problems, and the proportion of suicides attributed to work-related problems among all suicides. These descriptive statistics were presented separately for men and women. In addition, an interaction term between sex and period was included in the regression model to formally test whether period-related changes differed between men and women.

Negative binomial regression models were used because the outcome variable was a monthly count and overdispersion was anticipated. In the primary model, the monthly number of suicides attributed to work-related problems was analyzed using the logarithm of the monthly total number of suicides as an offset term, so that the model estimated the proportion of suicides attributed to work-related problems among all suicides. Sex, period, and their interaction term were included as explanatory variables. The interaction term between sex and period was included to formally test whether period-related changes differed between men and women. Calendar month was included as a categorical variable to adjust for seasonality in suicide statistics. January was used as the reference category, and month-specific estimates were not the primary focus of the study. These models were intended to compare differences in average levels across periods rather than to attribute observed changes solely to the pandemic, because the classification of suicide reasons changed in 2022.

Period-specific analyses were used to account for major contextual changes, including the declaration of the state of emergency in 2020 and the revision of official reason classification in 2022, following the general logic of interrupted or segmented time-series approaches used in public health evaluation studies^10,11).

As a supplementary analysis, models including time, sex, period, and their interaction terms were fitted to examine whether within-period temporal trends differed by sex and whether such sex differences changed across periods. All analyses were conducted using Stata version 18 (StataCorp LLC, College Station, TX, USA). Incidence rate ratios (IRRs) and 95% confidence intervals (CIs) were estimated using robust standard errors.

Because this study used publicly available aggregated data without individual-level identifiable information, ethical review was not required.

Results

During the study period, a total of 188,178 suicides were recorded, including 20,822 recorded as involving work-related problems. A total of 110 monthly observations from January 2017 to February 2026 were included in the analysis. Descriptive statistics are shown in Table 1. Among women, both the monthly number of suicides attributed to work-related problems and their proportion among all suicides were higher in the later periods than in the pre-emergency period. Among men, these values were lower during the COVID-19 period than in the pre-emergency period, but higher in the later periods after 2022.

Table 1. Descriptive characteristics of monthly suicide data, January 2017 to February 2026.

	Overall (n=110)		Pre-emergency period (n=39)		COVID-19 period (n=21)		COVID-19 period after classification change (n=17)		Post-COVID-19 period (n=33)
			January 2017 to March 2020		April 2020 to December 2021		January 2022 to May 2023		June 2023 to February 2026
Men
Total suicides (Mean) (SD)	1162.9	115.0	1161.4	97.0	1167.8	89.9	1248.4	117.5	1117.4	125.9
Work-related-problem suicides (Mean) (SD)	161.3	37.4	143.1	23.6	133.0	20.5	214.5	36.5	173.3	24.3
Proportion of work-related-problem suicides (%) (SD)	13.9	2.8	12.4	2.1	11.4	1.3	17.1	1.9	15.6	1.8
Women
Total suicides (Mean) (SD)	547.8	73.9	517.9	47.1	607.0	91.8	588.9	58.8	524.4	64.0
Work-related-problem suicides (Mean) (SD)	28.0	10.0	20.3	6.3	27.2	8.6	36.3	9.0	33.4	8.2
Proportion of work-related-problem suicides (%) (SD)	5.1	1.6	3.9	1.2	4.4	1.0	6.1	1.3	6.4	1.3

Values are presented as monthly mean (standard deviation).

Proportions were calculated as suicides attributed to work-related problems divided by total suicides in each month.

The results of the negative binomial regression analysis are presented in Table 2. Among men, compared with the pre-emergency period, the proportion of suicides attributed to work-related problems among all suicides was lower during the COVID-19 period (incidence rate ratio [IRR], 0.920; 95% confidence interval [CI], 0.859–0.986; p=.018), but higher during the COVID-19 period after the classification change (IRR, 1.389; 95% CI, 1.297–1.486; p<.001) and the post-COVID-19 period (IRR, 1.261; 95% CI, 1.184–1.342; p<.001). Women had a lower baseline proportion than men (IRR, 0.317; 95% CI, 0.287–0.350; p<.001). However, the relative change was more positive among women than among men during the COVID-19 period (IRR for interaction, 1.241; 95% CI, 1.080–1.425; p=.002) and the post-COVID-19 period (IRR, 1.292; 95% CI, 1.140–1.465; p<.001), whereas the corresponding interaction during the COVID-19 period after the classification change was not statistically significant (IRR, 1.134; 95% CI, 0.979–1.313; p=.094).

Table 2. Negative binomial regression analysis of the proportion of suicides attributed to work-related problems among all suicides.

	IRR	(95%CI)	p value
COVID-19 period (vs pre-emergency period)	0.920	0.859–0.986	.018
COVID-19 period after classification change (vs pre-emergency period)	1.389	1.297–1.486	<.001
Post-COVID-19 period (vs pre-emergency period)	1.261	1.184–1.342	<.001
Female (vs male)	0.317	0.287–0.350	<.001
Female×COVID-19 period	1.241	1.080–1.425	.002
Female×COVID-19 period after classification change	1.134	0.979–1.313	.094
Female×Post-COVID-19 period	1.292	1.140–1.465	<.001
February (vs January)	1.009	0.927–1.098	.843
March (vs January)	1.080	1.012–1.153	.021
April (vs January)	1.031	0.957–1.111	.422
May (vs January)	1.052	0.973–1.138	.200
June (vs January)	0.985	0.919–1.056	.671
July (vs January)	1.039	0.959–1.125	.349
August (vs January)	0.994	0.900–1.096	.897
September (vs January)	1.034	0.963–1.109	.359
October (vs January)	1.034	0.949–1.125	.445
November (vs January)	1.135	0.997–1.292	.055
December (vs January)	1.059	0.935–1.199	.368

Abbreviations: IRR, incidence rate ratio; CI, confidence interval.

Note: Negative binomial regression model including sex, period, and their interaction terms, adjusted for calendar month. The logarithm of the monthly total number of suicides was included as an offset term. Reference categories were the pre-emergency period (January 2017 to March 2020) and male sex. Periods were defined as follows: COVID-19 period, April 2020 to December 2021; COVID-19 period after classification change, January 2022 to May 2023; and post-COVID-19 period, June 2023 to February 2026. January was used as the reference category for calendar month. Robust standard errors were used.

In supplementary analyses, sex differences in within-period temporal trends were limited. A significant three-way interaction between sex, time, and period was observed only for the COVID-19 period (IRR, 0.980; 95% CI, 0.964–0.997; p=.018), suggesting attenuation of the stronger upward trend among women relative to the pre-emergency period. No significant three-way interaction was observed for the COVID-19 period after classification change (IRR, 0.989; 95% CI, 0.971–1.008; p=.262) or the post-COVID-19 period (IRR, 1.001; 95% CI, 0.991–1.011; p=.816).

Discussion

In this study, the proportion of suicides attributed to work-related problems among all suicides changed across periods and became higher in the later pandemic and post-COVID-19 periods. These findings suggest that work-related problems accounted for a larger share of suicides in the later pandemic and post-COVID-19 periods.

These findings are broadly consistent with previous studies showing that suicide patterns in Japan changed during the COVID-19 pandemic, with a greater impact among women^2,3,4). The present study extends this literature by suggesting that work-related problems were increasingly represented among suicides in the pandemic and post-pandemic periods⁵⁾. In supplementary analyses, the relative change was more positive among women during the COVID-19 period and post-COVID-19 period. However, sex differences in within-period temporal trends were not consistently observed, and the stronger upward trend among women was attenuated during the COVID-19 period relative to the pre-emergency period.

From an occupational health perspective, these findings are important because suicide prevention among workers has been recognized as an important issue in Japan, particularly in the context of the COVID-19 pandemic and its aftermath^1,8). Recent policy discussions have also emphasized the need to strengthen workplace mental health measures, including in small-scale enterprises⁹⁾. Monitoring suicides identified as involving work-related problems may therefore provide a useful indicator for considering occupational mental health responses to pandemic-related changes in suicide risk.

This study has several limitations. First, as described in the Methods, the 2022 revision expanded both the number of recordable reasons and the permissible sources of evidence; therefore, part of the increase observed from 2022 onward may reflect changes in ascertainment or coding practice rather than true change alone^5,8,12). Although we addressed this issue by analyzing the period after the classification change separately, the post-2022 findings should still be interpreted as reflecting an increase in suicides recorded as involving work-related problems. Second, this was an ecological analysis based on aggregated monthly data, and causal inference at the individual level is not possible. Nevertheless, the use of national monthly data enabled us to describe medium-term changes in work-related-problem suicides and their sex differences before, during, and after the COVID-19 pandemic.

Conclusion

Suicides recorded as involving work-related problems became more prominent in Japan after the onset of the COVID-19 pandemic. The relative increase was greater among women in some periods, suggesting the importance of considering sex differences when planning occupational mental health responses. Because the official classification of suicide reasons changed in 2022, the post-2022 findings should be interpreted with caution. Nevertheless, monitoring suicides identified as involving work-related problems may provide a useful indicator for occupational mental health practice and policy.

Author contributions

HE conceived and designed the study, collected and analyzed the data, interpreted the results, and drafted the manuscript. AN contributed to the study concept and participated in the discussion and interpretation of the findings. Both authors reviewed and approved the final manuscript.

Conflict of interest

There are no conflicts of interest to declare.

Study funding

This study was supported by a scholarship donation from Uchino Orthopedic Clinic and by a research grant (Grant Number JPMJRS22K1) from the Japan Science and Technology Agency (JST), Japan. The funders had no role in the design or conduct of the study; the collection, analysis, or interpretation of the data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Data availability statement

The data used in this study are publicly available from the Ministry of Health, Labour and Welfare of Japan website.

Declaration of generative AI use

During the preparation of this work, the authors used ChatGPT (OpenAI) to assist with language editing and improvement of the manuscript draft. All study design, data analysis, interpretation of the results, and final decisions regarding the content of the manuscript were carried out by the authors, who take full responsibility for the accuracy and integrity of the work.