key: cord-0729799-cht8xf0x authors: Kurita, J.; Sugishita, Y.; Sugawara, T.; Ohkusa, Y. title: Excess Mortality in Suicide caused by COVID-19 in Japan date: 2021-02-18 journal: nan DOI: 10.1101/2021.02.13.21251670 sha: 9cd7e9ded6cab0fc9db47129142ec5b71bab66ee doc_id: 729799 cord_uid: cht8xf0x Background: Countermeasures against COVID-19 outbreak such as lockdown and voluntary restrictions against going out adversely affect human stress and economic activity. Particularly, this stress might lead to suicide. Object: We examined excess mortality attributable to suicide caused by COVID-19. Method: We applied the NIID model to suicide deaths from October 2009 through September, 2020 for the whole of Japan for both genders. Effects of the great earthquake that struck in eastern Japan on March 11, 2011 were incorporated into the estimation model. Results: Significant excess mortality in suicide was found in July, August and September in 2020 for both genders. It was greater among females than among males. In total, 810 excess cases of mortality were identified. Discussion and Conclusion: Excess mortality during the two months was 1.4 times greater than the number of COVID-19 deaths confirmed by PCR testing. Countermeasures against COVID-19 should be chosen carefully in light of suicide effects. The estimation procedure was almost identical to that used for an earlier study [1], except for some points. Excess mortality was defined as the difference between the actual number of deaths and an epidemiological threshold if the actual number of deaths exceeded an epidemiological threshold. The epidemiological threshold is defined as the upper bound of the 95% confidence interval (CI) of the baseline. The baseline is defined as the number of deaths that are likely to have occurred if an influenza outbreak had not occurred. Therefore, if the actual deaths were fewer than the epidemiological threshold, then excess mortality was not inferred. Additionally, we defined negative excess All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. where D t represents suicide deaths in month/year t, T t denotes the linear time trend, V t is a dummy variable for routine pneumococcal vaccination, C t is a dummy variable for effect of the great earthquake in eastern Japan in March, 2011, and M it is a dummy variable for the month. Also, V t takes a value of one after October, 2014; otherwise, it is zero. Similarly, C t takes a value of one in May, June, or July in 2011; otherwise, it is zero. M it is one if t is the i-th month; otherwise, it is zero. Moreover, ν t and ω t are stochastic variables as ν t ~N(0, μ 2 ) and ω t ~N(0,ξ 2 ); they are mutually independent. Although ν t represents stochastic disturbances, ω t denotes non-negative deaths All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The study area encompassed the entire nation of Japan. The study period for estimation was October 2009 through September 2020 because of data availability. Suicide is defined as X60-X84 in ICD10. We adopted 5% as the level at which significance was inferred for results. Figure 1 presents observed suicide deaths, the estimated baseline, and its threshold for males. Figure 2 shows corresponding data for females. These figures showed that males are more numerous than females, but consistently decline is noted in the period for both genders. The sharpest spike occurred around May 2011, probably because of the great earthquake that struck eastern Japan on March 11, 2011. More than two months passed after the event before suicide increased. The second sharpest increase occurred probably in the latest period. Because the estimated coefficients of the cross-term of routine vaccination initiation and time trend, earthquake dummy, and the cross-term of ICD10 revision and time trend were not found to be significant, we dropped these terms from the estimation equation (1). Table 1 presents estimation results without these variables. The estimated coefficients of the great earthquake in eastern Japan were 0.133 for males and 0.247 for females. That result implies that the numbers of suicide deaths jumped respectively by All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint 13% and 25% because of the earthquake. It is noteworthy that the effect on females was greater than on males, although the incidence of suicide death was much greater among males than among females. Figure 3 presents excess mortality attributable to suicide in 2019 and 2020. Among females, clearly large excess mortality was found in the last two months. Among males, these were not greater than among females, but the two highest degrees of excess mortality were shown since 2019. In the three months, total excess mortality attributable to suicide was 810. This study applied the NIID model to suicide deaths to detect excess mortality attributable to COVID-19. Results showed that significant excess mortality attributable to suicide was found in July to September 2020, especially in August. It was estimated as 810 for both genders. Females were infected more by the COVID-19 outbreak than males, although the number of suicide deaths among males has been usually much greater than among females. As described earlier, some excess mortality was found for August and October, 2020, respectively, as 12 and 104 cases. At least 12 cases of excess mortality among all causes in August might result from 304 cases of excess mortality attributable to suicide. At the same time, pneumonia deaths were recorded respectively as 178, 314, and 75 cases of negative excess mortality, in May, June, and July. Those figures were measured All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint as the gap separating the observation and lower bound of 95% CI for baseline if the former was lower than the latter in May, June, and July [11] . In August, negative excess mortality attributable to pneumonia was not found, although the observation was below the baseline. Therefore, excess mortality attributable to suicide was cancelled out partially by negative excess mortality attributable to pneumonia. Therefore, no excess mortality was found for all causes of death. However, in August, negative excess mortality attributable to pneumonia was less than before. Therefore, excess mortality in suicide led to overall excess mortality for all causes, even though it was a very small number. In this sense, excess mortality attributable to suicide might be large in October, although data were not available because excess mortality from all causes was reported as a larger number than in August. Throughout Japan, about 592 cases of mortality attributable to COVID-19 in July August and September 2020, as confirmed by PCR testing, were reported to the Ministry of Health, Labour, and Welfare (MHLW) officially [1]. Therefore, excess mortality caused by suicide was 1.4 times greater than that reported for COVID-19 death. Pneumonia deaths sometimes lead to negative excess mortality. Therefore, the target of countermeasures for COVID-19 should reduce excess mortality attributable to suicide, at least in Japan. In other words, economic activity should be less restricted if All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The present study has some limitations. First, our results reflect data obtained through September, 2020 when the COVID-19 outbreak was not so widespread. Subsequently, and especially in winter, a severer COVID-19 outbreak occurred. The second state of emergency was declared on January 7, 2021. It required shortened business hours for restaurants. In the first state of emergency declaration in April and May, 2020, restaurants were required to shut down. Even though requirements of the second state of emergency were milder than the first, it might have increased suicide deaths several months after. Continuous monitoring is expected to be necessary. Second, to ascertain the most appropriate countermeasures for COVID-19 in Japan, All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint cost-effectiveness analysis is necessary. At that time, loss of quality of life in should be counted among the costs of restriction of economic activity as a major part of countermeasures. It remains as a subject for our next research challenge. The obtained results show excess mortality in suicide in July, August and September, 2020. Continued careful monitoring of excess mortality attributable to suicide is expected to be necessary for evaluating countermeasures taken for the second state of emergency declaration. The present study is based on the authors' opinions: it does not reflect any stance or policy of their professionally affiliated bodies. We acknowledge Dr. Nobuhiko Okabe, Kawasaki City Institute for Public Health, Dr.Kiyosu Taniguchi, National Hospital Organization Mie National Hospital, and Dr.Nahoko Shindo, WHO for their helpful support. The authors have no conflict of interest to declare. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint All information used for this study was published on the web site of MHLW [12] . Therefore, no ethical issue is presented. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. μ 2 represents the variance of the disturbance term. "Earthquake" is a dummy variable for May-July, 2011, which was inferred to be caused by the great earthquake in eastern Japan occurred March 11, 2011. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Year Note: The blue line represents observations. The red line represents the estimated baseline. The green line shows its threshold. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint Year Note: The blue line represents observations. The red line represents the estimated baseline. The green line shows its threshold. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint 1 Figure 3 : Excess mortality in suicide by gender since 2019 by month in Japan. Note: The blue bars represents excess mortality in suicide among males. Orange bars indicate that among females. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted February 18, 2021. ; https://doi.org/10.1101/2021.02.13.21251670 doi: medRxiv preprint On the estimation of technical inefficiency in the stochastic frontier production function model Cost inefficiency in Washington Hospitals: A stochastic frontier approach using panel data Frontier Estimation: How useful a tool for health economics Managed care and technical efficiency Alternative methods to examine hospital efficiency: Data envelopment analysis and stochastic frontier analysis Negative Excess Mortality in