The COVID-19 pandemic imposed substantial mental health stressors leading to concerns about an increased suicide risk. To investigate this issue, we investigated suicide mortality rates in the United States from March 1, 2020, through June 30, 2022, comparing them with data from the pre-pandemic period of January 2015 through February 2020. Suicide mortality in the United States was 3% below expected levels during the study period. However, there was an increased suicide incidence in adults ages 18–34 years. The concerns that the pandemic contributed to an overall marked increase in suicide risk is not supported by this analysis, but young adults did experience an increase.
Article
Suicide deaths during the COVID-19 pandemic in the United States, by region, gender and age, March 1, 2020-June 30, 2022
https://doi.org/10.21203/rs.3.rs-3093283/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Despite initial concerns about the potential increase in suicide deaths resulting from factors such as isolation, mourning, and economic difficulties during the early stages of the COVID-19 pandemic, empirical evidence indicated that suicide rates in the United States and other countries were lower than anticipated in 2020.1–4 Some pandemic-associated stressors may have been counteracted by increased access to mental health care (e.g., expanded telehealth), and decreased stigma in seeking out that care. However, the pandemic persisted longer than initially expected. To better understand the overall health effects of the pandemic, a more complete and granular examination of the suicide mortality as an indicator of accumulated stressors over time is warranted.5
Some reports suggest an increase in pediatric/adolescent (ages 5–17), and young adult (ages 18–34) suicide mortality. Also, substantial decreases in older but non-geriatric adults (ages 50–64) have been described.6–8 However, there is a need for further study because many of these reports were limited in scope and did not adequately consider suicide rates in the context of pre-pandemic trends. Also, some reports did not assess traditional age categories, and so were exploratory. Overall, an analysis that captures granular pandemic-associated changes in suicide death epidemiology—including changes by US region, age, by sex (and by sex within age groups), and whether changes in suicide deaths tracked with waves of Covid-19 mortality—remains lacking.
Accordingly, we conducted a comprehensive analysis of suicide mortality during the COVID-19 pandemic in the United States, encompassing regional, age group, and gender-specific variations from March 2020 through June 2022, in comparison with expected suicide mortality based on trends established by the pre-pandemic period of January 2015 through February 2020. In addition, temporal relationships between changes in suicide mortality and waves of Covid-19 mortality were tested.
Excess suicide mortality was defined as the number of observed minus the number of expected deaths. We used seasonal autoregressive integrated moving average (sARIMA) models and applied them to population statistics (2014–2020), all-cause and suicide mortality data (January 1, 2015-February 29, 2020) to generate expected monthly populations and suicide deaths for March 1, 2020-June 30, 2022.
We developed a component model based on the four Census Bureau regions, two genders and five age groups (5–17 years, 18–34 years, 35–49 years, 50–64 years and ≥ 65 years). Due to suppression constraints, data for ages 5–17 years were only available at the national level for both genders combined. We further adjusted the population during the pandemic period to account for pandemic-associated all-cause excess mortality. We divided the pandemic into three periods: March 2020-December 2020 (10 months), January 2021-December 2021 (12 months) and January 2022-June 2022 (6 months). In addition, we conducted sensitivity analyses for all component and summation models; in addition to the primary model, three additional models were deployed to generate alternate expected death projections by using a 4-year baseline period (January 1, 2016-February 29, 2020) a 3-year baseline periods (January 1, 2017-February 29, 2020), and estimates based on crude pre-pandemic averages using values three most recent pre-pandemic years for each month (2017–2019 for March through December, and 2018–2020 for January and February).
To assess whether changes from expected suicide mortality tracked COVID-19-specific mortality, deaths from suicide were graphed against COVID-19 mortality over time, to assess for correlation. Pearson correlation coefficients were calculated with corresponding p values.
95% confidence intervals (CI) were calculated for each month and period; boundaries were obtained directly from the sARIMA model in the auto.arima function in the R statistical software (version 4.0.2) via 5,000 simulation samples from the estimated sARIMA model for each component model. Statistical significance of excess deaths was defined as the 95% CI that did not include 0. Data were acquired from CDC WONDER. Data up to December 31, 2021 are finalized and data for January-June 2022 remain provisional. International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes U03, X60–X84 and Y87.0 were used to identify the underlying cause of death as suicide.9
This study used publicly-available data and was not subject to institutional review.
During the pandemic period (March 1, 2020-June 30, 2022), 114,363 suicide deaths were expected in individuals ages ≥ 5 years in the United States and 110,959 occurred (-3,404 suicide deaths; 95% CI: -5,705 to -1,103) (Table 1). This result corresponds to an observed-to-expected ratio of 0.97 (95% CI: 0.95–0.99). The first ten months of the pandemic in 2020 saw significantly lower suicide mortality (-2,402; 95% CI: -3,535 to -1,269) than expected levels; however, excess suicide mortality in 2021(-849; 95% CI: -2,099 to 402) and the first six months of 2022 (-153; 95% CI: -981 to 675) were lower than expected, but not statistically significant (Table 1).
Regional analysis. Examining deviations from expected suicide mortality by region revealed heterogeneity over the pandemic period (Fig. 1). All regions except the Midwest saw a statistically significant decrease in suicide mortality in 2020 (Table 1). In 2021, only the West had significantly lower suicide mortality than projected. In 2022, the Midwest had a statistically significant increase in suicide mortality and the West saw a statistically significant decrease; the South and Northeast suicide mortality were similar to what was projected.
Age analysis. The youngest age group studied, individuals ages 5–17 years, did not demonstrate a significant deviation from expected suicide rates overall (-750 suicide deaths, 95% CI -1,537 to 37) (Fig. 2). However, numerical (not statistically significant) decreases were observed in 2020 and 2021, and a statistically significant decrease was observed in 2022 (Supplemental Appendix Table 2). Excess suicide mortality was observed among individuals ages 18–34 years, with 1,537 (95% CI 900 to 2,174) excess suicide deaths observed during the overall pandemic study period. All other age groups had lower than expected suicide mortality; in particular, among individuals ages 50–64 years, suicide deaths were 2,452 lower (95% CI -3,935 to -969) than what would have been expected (Fig. 2, Supplemental Appendix Table 2).
Gender analysis. Changes in suicide mortality varied by gender. As during the pre-pandemic period, males continued to comprise a majority of suicide mortality during the study period (Supplemental Table 3). In individuals ages 18–34 years, increases in suicides were concentrated among males, but statistically significant excess suicide mortality was observable in both genders by the end of the study (Fig. 3, Supplemental Table 3). In adults ages 50–64, decreases from expected suicide counts were highly concentrated among males, though a decreased rate among females was also significant (Fig. 3, Supplemental Table 3).
COVID-19 temporal analysis. Changes in suicide deaths from expected values were generally not temporally correlated with waves of COVID-19 mortality. Adults ages 18–34 were the only age demographic in whom a statistically significant correlation between changes from expected suicide deaths and waves of COVID-19 deaths was detected (r = 0.38, p = 0.05) (Fig. 4).
Sensitivity analysis. Results from the sensitivity analyses using alternative pre-pandemic time periods to model pandemic-era expected suicide deaths revealed that the selection of the model could alter the bandwidth of the uncertainty corresponding to the reported point estimates, but no qualitative reversals were found (Supplemental Appendix Table 2). Adults ages 50–64 were the only age demographic in which a significant decrease in suicide mortality during the overall study period was found in the main analysis (modeling based on a 5-year pre-pandemic baseline period). The first (4-year baseline period) and third (crude averages) sensitivity analyses confirmed this finding, but the second sensitivity analysis (3-year baseline period) did not, owing to wide confidence intervals inherent in the shorter baseline period (Supplemental Appendix Table 2).
We found that suicide deaths during the COVID-19 outbreak in the US (March 1, 2020-June 30, 2022) were statistically lower than expected. This change was most pronounced in the West, where statistically significant decreased suicide mortality persisted during the entire study period. Contrary to other reports, we found that suicide deaths in children and adolescents ages 5–17 were unchanged during the study, and a statistical decrease was observed in the first half of 2022. Increases were seen in younger adults ages 18–34 (which moderately correlated with contemporaneous COVID-19 waves), while decreases were seen in adults ages 50–64, driven by marked decreases among males in this age group (who otherwise continued to comprise the higher rates of suicide than females).
Our findings contextualize suicide deaths in the United States (and elsewhere), finding little change in suicide death rates during the pandemic period, despite rising death rates from alcohol poisonings, accidents, and other non-medical deaths.2,10 In some instances, our data confirm other previously published findings. However, our findings contradict reports which concluded that increases in suicide mortality in pediatric/adolescent populations occurred.6–8
This study adds to the existing literature in several ways. First, we provide increased granularity with respect to demographics and time periods. Second, we account for pre-pandemic trends in suicide mortality and provided several sensitivity analyses, both of which are important in strengthening or weakening conclusions regarding epidemiologic data with relatively low event rates.
Pediatric/Adolescent population
The impact of COVID-19 interventions and the pandemic itself on the potential suicide deaths in school-aged children has been discussed in public forums but studied comparatively little. The present analysis suggests that school-aged children did not experience an increase in suicide deaths during the pandemic, despite fears.11
Some other groups have reported transient increases in suicide deaths among children and adolescents during the pandemic.7,12 Those studies reported either descriptive increases (i.e., lacked modeling, which fails to account for secular pre-pandemic trends) or failed to provide sensitivity analyses with multiple models, a crucial step for low-event rates. For example, our primary model found statistically lower suicide mortality in some groups during some periods, but other models (based on projections calculated from 3- or 4-year pre-pandemic baseline periods) had uncertainty estimates which call the statistical significance of the findings into question. In addition, the use of certain selective age groupings raises the possibility of a positive finding in non-traditional age groupings that would not be present if typical groupings had been used, which would undermine the validity of any generalized conclusions about the data. Genuine epidemiologic changes should be robust to small changes in demographic boundaries, and our findings suggest that statistical differences found in some previous studies may hinge on the age group choices.
Indeed, our findings come in contrast with a recent published study which reported increases in suicide deaths in some youth age groupings, and there are likely reasons for this discrepancy.7 First, their general summation of 5–24 included mostly older suicides (18–24) due to the increasing frequency of suicides by age. Our grouping (5–17 years) removes young adults from the analysis, making the finding more reflective of the school-aged population. Second, in their age group analysis, data suppression likely led to falsely lower point estimates for expected deaths (and thus overestimates of excess suicide mortality). For example, in the 5–12 years age group, suppression occurred in 16 of 62 months in the pre-pandemic period (all prior to 2018). This likely led to an underestimate of expected deaths (did owing to failure to account for a nonstationary upward trend). By contrast, their analysis of adolescents aged 13–17 years did not have this problem—and which would make up the bulk of suicides in ages 5–17; the observed numbers of suicide in that age group were not different from to the expected numbers. Third, our modeling shows that pre-pandemic trends anticipated a continued steady rise in suicide mortality in this (and other) demographics, meaning that some increases described in other studies were not related to the pandemic but, rather, reflected existing systemic issues. Studies that simply divided 2021 suicides rates by rates from 1–2 years prior (including 2020, a time in which suicides are known to have been down in the US) were not equipped to capture this finding.6,8
Still, there have been reports of increases in suicidal thoughts and suspected attempts in adolescents during the pandemic, and an increase in the share of emergency visits owing to mental health conditions.13–15 We note that the increases in suicidal thoughts and suspected attempts occurred mainly in 2021, after the initial pandemic period, and corresponded to schools re-opening in the United States, comporting with prior research.16,17 In addition, the increase in the share of mental health visits in emergency departments in younger persons likely reflected decreases in medical-related visits, rather than an increase in mental-health related visits. Unlike other age groups, adolescent suicide is highly associated to the academic calendar, with more suicides occurring during school months than non-school months, making predictions during adolescent suicide rates during school closures difficult. Therefore, our research need not be seen as entirely contradictory to the above reports–suicidal deaths are rare outcomes of suicidal thinking or attempts in youth, and influences on what results in deaths versus ideation or attempt are likely distinct.
Young adults
Comparatively less media and research focus has been given to suicide in youth of transitional age (18–24 years) and young adults (24–34 years). Our findings of excess suicides based upon trends in the ages 18–34 years group implies the need to target interventions to these ages, in whom statistically significant increases in suicide mortality (beyond what was expected during the period) were observed here. Indeed, the excess suicide mortality in those aged 18–34 during the pandemic represents an augmentation of a pre-pandemic trend seen. This finding is in keeping with data reporting that this age group as having had the largest increase in depression and anxiety early in the pandemic.18 However, this increase may also be attributable to increases in substance use; fatal substance use disorders are known to have increased during the pandemic and precise cause of death attribution is not always attainable. Studies have shown that adults ages 18–34 were highly vulnerable for mental distress during the COVID-19 pandemic, with lower resilience relative to other groups, owing to high levels of loneliness and stress.19 Additionally, poorer mental health in this demographic may be attributable to higher career demands, financial concerns and parental responsibilities, as noted in previous investigations in the oncology literature.20
Non-geriatric older adults
The reversal of a general increasing trend of suicide during the pandemic for adults 50–64 years of age is of particular interest. Despite a higher overall risk for COVID-19, older adults showed a more positive mental health profile compared with younger adults which could in part explain the finding of lower from expected suicide mortality during the COVID-19 pandemic period in this group.19,21,22 Alternative explanations include the benefits of short-term economic support (government-supplied paycheck relief) or early retirement, which could have reduced related stressors associated with suicide mortality in this age group.23–25
Overall findings/historical context
There are several possible explanations for the overall absence of increased suicide mortality during the COVID-19 pandemic, despite initial concerns raised by mental health professionals. An early “pulling together effect” may have given people a shared sense of purpose, which correlated to a period of decreased suicide death overall.2 Later, mental health stressors may have mounted but increased care demands were counterbalanced by newly increased access to psychiatric resources previously unavailable, including telehealth and, in 2022, access to a universal national suicide hotline (988).26 Additionally, mental health received more media attention during the pandemic, possibly decreasing stigma for seeking care, thereby leading to patient engagement with psychiatric resources, rather than suicidal behavior.27
Decreases in suicide mortality has been observed in prior pandemics, possibly related to the “pulling-together” effect described during previous periods of disaster; lower than expected suicide deaths appear to have been most pronounced when shelter-in-place orders were in effect in most US jurisdictions.28,29 Moreover, in our study, we mostly did not see a correlation with Covid-19 deaths and suicide deaths; in fact, our data confirmed previous modeling showing decreased suicide mortality during the early shelter-in-place period in the US, all of which implies that suicidal ideation and suicide completion may be influenced by markedly different stressors. In addition, there was not increased suicide in the Northeast, where mitigation was longer lasting (i.e., more social isolation), reflecting the multifactorial nature of suicide deaths.
Whether changes in healthcare delivery during the pandemic may in part explain the present findings warrant consideration. Telehealth accounted for less than 1% of all ambulatory care prior to the COVID-19 pandemic. At its peak, this shifted to encompass 40% of mental health and substance use outpatient visits (March-August 2020). As the pandemic progressed, in-person care returned with an overall decrease in telehealth visits but notably, for mental health and substance use treatment, telehealth has persisted—representing 36% of these visits.26 Additionally, during the COVID-19 pandemic, many US states expanded coverage of mental health services provided via telehealth through their Medicaid programs including expansion of types of services and the allowance of increased provider types to be reimbursed for these services. Additionally, some private insurers have also removed pre-pandemic telehealth coverage restrictions, thereby expanding access to care.
This study has several limitations. One confounding variable may be manner of death adjudication; some deaths assigned to unintentional overdose may have actually been suicides, and vice versa, as previously.30 However, baseline rates of misattribution are unlikely to have changed during the pandemic. We were unable to assess suicide by race and ethnicity, due to our model’s reliance on 5 years of pre-pandemic data (the CDC no longer publishes race/ethnicity data in categories used until 2020). Also, while some data used in this study (2022) are provisional, they are unlikely to change dramatically.
In sum, suicide deaths are lower in the US since the COVID-19 pandemic began. Future work on suicide mortality should take both pre-pandemic and pandemic trends into account.
- Isumi, A., Doi, S., Yamaoka, Y., Takahashi, K. & Fujiwara, T. Do suicide rates in children and adolescents change during school closure in Japan? The acute effect of the first wave of COVID-19 pandemic on child and adolescent mental health. Child Abuse & Neglect 110, 104680 (2020).
- Faust, J. S. et al. Mortality From Drug Overdoses, Homicides, Unintentional Injuries, Motor Vehicle Crashes, and Suicides During the Pandemic, March-August 2020. JAMA 326, 84–86 (2021).
- Pirkis, J. et al. Suicide numbers during the first 9-15 months of the COVID-19 pandemic compared with pre-existing trends: An interrupted time series analysis in 33 countries. eClinicalMedicine 51, (2022).
- Ehlman, D. C. Changes in Suicide Rates — United States, 2019 and 2020. MMWR Morb Mortal Wkly Rep 71, (2022).
- Gunnell, D. et al. Suicide risk and prevention during the COVID-19 pandemic. The Lancet Psychiatry 7, 468–471 (2020).
- Curtin, S. & Garnett, M. Suicide and Homicide Death Rates Among Youth and Young Adults Aged 10–24: United States, 2001–2021. https://stacks.cdc.gov/view/cdc/128423 (2023) doi:10.15620/cdc:128423.
- Bridge, J. A. et al. Youth Suicide During the First Year of the COVID-19 Pandemic. Pediatrics 151, e2022058375 (2023).
- Stone, D. M. Notes from the Field: Recent Changes in Suicide Rates, by Race and Ethnicity and Age Group — United States, 2021. MMWR Morb Mortal Wkly Rep 72, (2023).
- National Center for Health Statistics. CDC WONDER. https://wonder.cdc.gov/.
- Angus, C., Buckley, C., Tilstra, A. M. & Dowd, J. B. Increases in ‘deaths of despair’ during the COVID-19 pandemic in the USA and UK. Public Health (2023) doi:10.1016/j.puhe.2023.02.019.
- Manzar, Md. D., Albougami, A., Usman, N. & Mamun, M. A. Suicide among adolescents and youths during the COVID-19 pandemic lockdowns: A press media reports-based exploratory study. Journal of Child and Adolescent Psychiatric Nursing 34, 139–146 (2021).
- Curtin, S. C. & Ahmad, F. B. Provisional Numbers and Rates of Suicide by Month and Demographic Characteristics: United States, 202.
- Yard, E. Emergency Department Visits for Suspected Suicide Attempts Among Persons Aged 12–25 Years Before and During the COVID-19 Pandemic — United States, January 2019–May 2021. MMWR Morb Mortal Wkly Rep 70, (2021).
- Holland, K. M. et al. Trends in US Emergency Department Visits for Mental Health, Overdose, and Violence Outcomes Before and During the COVID-19 Pandemic. JAMA Psychiatry (2021) doi:10.1001/jamapsychiatry.2020.4402.
- Bommersbach, T. J., McKean, A. J., Olfson, M. & Rhee, T. G. National Trends in Mental Health–Related Emergency Department Visits Among Youth, 2011-2020. JAMA 329, 1469–1477 (2023).
- Hansen, B. & Lang, M. Back to school blues: Seasonality of youth suicide and the academic calendar. Economics of Education Review 30, 850–861 (2011).
- Black, T. Children’s Risk of Suicide Increases on School Days - Scientific American. https://www.scientificamerican.com/article/childrens-risk-of-suicide-increases-on-school-days/.
- Twenge, J. M. et al. Worldwide increases in adolescent loneliness. Journal of Adolescence 93, 257–269 (2021).
- Na, L., Yang, L., Mezo, P. G. & Liu, R. Age disparities in mental health during the COVID19 pandemic: The roles of resilience and coping. Social Science & Medicine 305, 115031 (2022).
- Arndt, V., Merx, H., Stegmaier, C., Ziegler, H. & Brenner, H. Restrictions in quality of life in colorectal cancer patients over three years after diagnosis: A population based study. European Journal of Cancer 42, 1848–1857 (2006).
- Carstensen, L. L., Shavit, Y. Z. & Barnes, J. T. Age Advantages in Emotional Experience Persist Even Under Threat From the COVID-19 Pandemic. Psychol Sci 31, 1374–1385 (2020).
- Zach, S., Zeev, A., Ophir, M. & Eilat-Adar, S. Physical activity, resilience, emotions, moods, and weight control of older adults during the COVID-19 global crisis. European Review of Aging and Physical Activity 18, 5 (2021).
- Bureau, U. C. Pandemic Disrupted Labor Markets but Had Modest Impact on Retirement Timing. Census.gov https://www.census.gov/library/stories/2022/09/did-covid-19-change-retirement-timing.html.
- Jo, A., Jeon, M. & Oh, H. Age-differentiated Risk Factors of Suicidal Ideation among Young and Middle-aged Korean Adults. Osong Public Health Res Perspect 8, 201–210 (2017).
- Qin, P. et al. Midlife suicide: A systematic review and meta-analysis of socioeconomic, psychiatric and physical health risk factors. Journal of Psychiatric Research 154, 233–241 (2022).
- Lo, J., Panchal, N., Mar 15, B. F. M. P., & 2022. Telehealth Has Played an Outsized Role Meeting Mental Health Needs During the COVID-19 Pandemic. KFF https://www.kff.org/coronavirus-covid-19/issue-brief/telehealth-has-played-an-outsized-role-meeting-mental-health-needs-during-the-covid-19-pandemic/ (2022).
- Beckjord, E. Contributor: COVID-19 and Stigma About Mental Health—A Pandemic Silver Lining? AJMC https://www.ajmc.com/view/contributor-covid-19-and-stigma-about-mental-health-a-pandemic-silver-lining- (2022).
- Bastiampillai, T., Allison, S. & Looi, J. C. L. Spanish Flu (1918–1920) Impact on US Suicide Rates by Race: Potential Future Effects of the COVID-19 Pandemic. Prim Care Companion CNS Disord 23, 38058 (2021).
- Faust, J. S. et al. Suicide Deaths During the COVID-19 Stay-at-Home Advisory in Massachusetts, March to May 2020. JAMA Network Open 4, e2034273 (2021).
- Bohnert, A. S. B. & Ilgen, M. A. Understanding Links among Opioid Use, Overdose, and Suicide. New England Journal of Medicine 380, 71–79 (2019).
Table 1. Expected and observed suicide deaths in the US by region, and period.
There is NO Competing Interest.