Abstract
Objectives
Suicide is a leading cause of death in the United States, and rates vary by race and ethnicity. An analysis of suicide across large US cities is absent from the literature. The objective of this study was to examine suicide rates among the total population, non-Hispanic Black population, and non-Hispanic White population in the United States and in the 30 largest US cities.
Methods
We used data from the National Vital Statistics System to calculate non-Hispanic White, non-Hispanic Black, and total age-adjusted suicide rates for the 30 largest US cities and for the entire nation during 2 periods: 2008-2012 and 2013-2017. We also examined absolute and relative differences in suicide rates among non-Hispanic White populations and non-Hispanic Black populations in each city.
Results
The overall age-adjusted suicide rate per 100 000 population in the United States increased significantly from 12.3 in 2008-2012 to 13.5 in 2013-2017. Total suicide rates were stable in most cities; rates increased significantly in only 1 city (Louisville), and rates decreased significantly in 2 cities (Boston and Memphis). The non-Hispanic White suicide rate was significantly higher—1.3 to 4.3 times higher—than the non-Hispanic Black suicide rate in 24 of 26 study cities during 2013-2017. From 2008-2012 to 2013-2017, non-Hispanic White suicide rates decreased significantly in 3 cities and increased significantly in 3 cities; non-Hispanic Black suicide rates increased significantly in 5 cities and decreased in none. Absolute differences in suicide rates among non-Hispanic White populations and non-Hispanic Black populations increased significantly in 1 city (Louisville) and decreased significantly in 2 cities (Memphis and Boston).
Conclusions
Our study may inform the use of evidence-based programs and practices to address population-level risk factors for suicide.
Suicide is a leading cause of premature mortality in the United States that resulted in more than 48 000 deaths in 2018. 1 It is the 10th leading cause of death across all age groups, the second leading cause of death among people aged 10-34, and the fourth leading cause of death among people aged 35-54. 2 Suicide also imposes a substantial economic cost on society, estimated at nearly $70 billion annually. 3 Suicide rates per 100 000 population increased from 10.4 in 2000 to 14.0 in 2017, a 35% increase. 4 This increase could be explained by a recent review of Healthy People 2020 data, which suggests that the nation’s mental health status declined during this time. 5
Several social determinants of health in urban areas put people at risk for suicide. 6,7 At the individual level, poor physical and mental health, job loss, substance abuse, and lack of social support are risk factors. 8 -11 Relationships and interactions with others may also be a risk factor for suicide. For example, online social networking can put adolescents at risk for self-harm and suicidal ideation. 12 Conversely, social connectedness may protect against suicide. Recent research shows that urban residents who live with family are less likely to commit suicide than urban residents who do not live with family, even after controlling for socioeconomic status. Furthermore, family connectedness protects against suicide in urban settings, even if people do not live with their family. 13,14
Structural factors, such as residential instability, economic inequality, and gender inequality, pose a risk for suicide. 15 A lack of access to mental health care 16,17 and access to firearms are also risk factors for suicide. 18,19 Data on access to firearms are nuanced; for example, Steelesmith and colleagues 10 found that the presence of gun shops has a stronger association with suicide in urban settings than in rural settings, possibly because of the prevalence of guns in rural counties. Although suicide rates are increasing fastest in rural settings, 20 examining suicide across urban contexts is important for understanding risk and protective factors, especially as suicide rates are projected to increase among all racial/ethnic groups aged 25-64. 21
Research that examines suicide in urban or metropolitan areas is also important because approximately 80% of the US population resides in these areas. 22 In these settings, age-adjusted suicide rates per 100 000 population are higher in the non-Hispanic White population than in the non-Hispanic Black population. 23 However, recent data show that the risk of suicide is increasing disproportionately for some racial groups. For example, from 2008 to 2012. among children aged 5-11, the suicide rate increased among non-Hispanic Black children and decreased among non-Hispanic White children. 24 Geographic differences in suicide rates exist across states, counties, and cities. 25 Residents of certain counties, such as Clark County, Nevada (the Las Vegas metropolitan statistical area), face an increased risk of suicide compared with residents in other counties across the United States. 26 In contrast, New York City has a suicide rate that is only half of the national average. 27 City efforts to address suicide require targeted, data-informed strategies. 28 Some large cities are implementing these strategies, focusing on particular groups, such as adolescents 29 and veterans. 30 These city efforts are essential for meeting national goals proposed in Healthy People 2030 that aim to reduce suicidal ideation and suicide attempts. 31
Analysis of suicide rates by city—and differences in suicide rates among racial populations—is critical for understanding the increase in suicide rates in the United States. Population-level research on other causes of death, such as breast cancer, in urban settings has led to improvements in policies and programs. 32,33 A more comprehensive documentation of suicide rates across large cities can support similar efforts. Systematic examination of suicide rates in large US cities and racial differences in these rates is absent from the literature. Therefore, the objective of this study was to examine suicide rates in the United States and in the 30 largest US cities. For each city and the nation, we examined total rates and rates among non-Hispanic White and non-Hispanic Black populations. In addition, we examined changes in rates from 2008-2012 to 2013-2017.
Methods
Population Estimates
We used data from the 2013 US Census to identify the 30 most populous cities in the United States. Nashville/Davidson County, Tennessee; Indianapolis/Marion County, Indiana; and Louisville/Jefferson County, Kentucky, have consolidated their city and county governments; therefore, we used county data to calculate their city populations. To calculate denominator data for the total population and the non-Hispanic White population, we used American Community Survey (ACS) 5-year estimates 34 for 2010 (2008-2012) and for 2015 (2013-2017). The Decennial Census is the only source that offers population data for the non-Hispanic Black population. Therefore, we divided the non-Hispanic Black population by the total (that is, Hispanic and non-Hispanic) Black population in the 2010 Decennial Census to find the proportion of the Black population that was non-Hispanic in 2010. We then applied these proportions to the total Black population estimate from the ACS 5-year estimates for 2010 and 2015. We estimated the non-Hispanic Black population for the United States and each city by using the following formula:
where n is age group, NHB Pop is non-Hispanic Black population, Black Pop is total (Hispanic + non-Hispanic) Black population, US Census refers to the 2010 Decennial Census, and ACS refers to the 2010 or 2015 ACS.
The 2010 and 2015 population estimates (total population, non-Hispanic White population, and non-Hispanic Black population) were multiplied by 5 to estimate the population during the two 5-year periods (2008-2012 and 2013-2017).
Data
We extracted data from the National Vital Statistics System (NVSS) multiple cause-of-death data file 35 to calculate suicide rates. The NVSS provides official mortality data for all causes of death at the city level. We extracted data for 2008 through 2017 by using suicide-related International Classification of Diseases, Tenth Revision (ICD-10) codes U03, X60-X84, and Y87.0. 36 We extracted death data across race and ethnicity by age group: birth–4 years, 5-14, 15-24, 25-34, 35-44, 45-54, 55-64, 65-74, 75-84, and ≥85. Inclusion of all age groups enabled a comprehensive analysis of suicide mortality and an age-adjusted comparison with other causes of death. We aggregated data on suicide rates for all races, for the non-Hispanic White population, and for the non-Hispanic Black population for 10 years (2008-2012 and 2013-2017). For Nashville, Indianapolis, and Louisville, we summed the number of suicides in the entire county to match the ACS population denominator for the entire county (not the city).
Statistical Analysis
We calculated the overall, non-Hispanic White, and non-Hispanic Black age-adjusted suicide rates for each city and for the entire United States. We did not calculate mortality rates for cities with <20 suicides (for the overall, non-Hispanic White, or non-Hispanic Black population) during each 5-year period, because the rate would be unreliable. 37 Twenty-six cities had sufficient data for inclusion in the final analysis; we removed El Paso, Portland, and San Jose from the analysis because of insufficient data. In addition, we removed Las Vegas because health officials had misclassified death certificate data: some people who died had resided in unincorporated areas but were counted as Las Vegas deaths. The use of 5-year periods enabled the retention of the maximum number of cities in the final dataset. We used the standard US population (from the year 2000) to calculate age-adjusted rates per 100 000 population. To examine suicide disparities between non-Hispanic White and non-Hispanic Black populations, we calculated rate differences (an absolute measure of differences) by subtracting non-Hispanic Black rates from non-Hispanic White rates. We also calculated rate ratios (RRs)—relative disparities—by dividing the non-Hispanic White rate by the non-Hispanic Black rate.
We conducted z tests to assess the significance of the differences in mortality rates between groups (eg, races) from 2008-2012 and 2013-2017, using a Taylor series expansion technique, with a 95% CI. 38 We interpreted absolute z scores of ≥1.96 as significant (with a 95% CI). We conducted analyses in 2020 using SAS version 9.4 (SAS Institute, Inc). The Mount Sinai Hospital Medical Center Institutional Review Board determined this study was exempt because it used publicly available data without personal identifiers.
Results
The overall age-adjusted suicide rate per 100 000 population in the United States increased significantly from 12.3 in 2008-2012 to 13.5 in 2013-2017 (P < .001; Table). In 2008-2012, the age-adjusted suicide rate was higher in 11 cities than nationally; the highest suicide rate was in Denver (17.7), the second highest rate in Jacksonville (15.5), and the third highest rate in Oklahoma City (14.6). The lowest rates were in New York City (5.8); Washington, DC (6.1); Chicago (6.7); and Boston (6.8).
Suicide rates and differences in rates among non-Hispanic White and non-Hispanic Black people in the United States and the 26 largest US cities, 2008-2012 and 2013-2017 a
Abbreviation: —, not applicable.
aData source: National Vital Statistics System, Multiple Causes of Death data file. 35 All values are age-adjusted rates per 100 000 population unless otherwise indicated.
bSignificant at P < .05 based on a z test for change in total mortality rate between periods. All other changes not significant.
From 2008-2012 to 2013-2017, the suicide rate decreased in 11 cities. Two decreases were significant: Boston had 1.5 fewer (P = .02) and Memphis had 1.7 fewer (P = .28) suicides per 100 000 population. The suicide rate increased in 16 cities, but the increase was significant only in Louisville, at 1.8 additional suicides per 100 000 population (P = 0.47). In 2013-2017, Denver (16.3) continued to have the highest suicide rate, followed by Louisville (16.0), Oklahoma City (15.6), and Indianapolis (14.9).
Mortality Rates by Race
From 2008-2012 to 2013-2017, the national suicide rate per 100 000 population increased significantly among non-Hispanic White populations (from 15.1 to 17.0) and non-Hispanic Black populations (from 5.8 to 6.5; Figure). From 2008-2012 to 2013-2017, the non-Hispanic White suicide rate decreased significantly in Detroit (−7.2), Houston (−3.7), and Boston (−2.8) and increased significantly in Louisville (+7.7), Indianapolis (+3.6), and Nashville (+3.4). Although the non-Hispanic Black suicide rate decreased in 12 cities, none of the decreases were significant. The non-Hispanic Black suicide rate increased significantly in Louisville (+10.5), Indianapolis (+9.3), Houston (+8.5), and New York City (+4.5).
Racial Differences Across Periods
Nationally, the non-Hispanic White suicide rate was more than 2½ times higher than the non-Hispanic Black suicide rate during 2008-2012 (RR = 2.6; 95% CI, 2.6-2.7), and the RR did not change in 2013-2017. Although the RR remained consistent, the rate difference increased slightly from 9.3 (2008-2012) to 10.5 (2013-2017). In 2008-2012, non-Hispanic White suicide rates were higher than non-Hispanic Black mortality rates in all 26 cities, and this difference was significant in all cities, except Seattle and Washington, DC. In 2013-2017, non-Hispanic White mortality rates remained higher than non-Hispanic Black mortality rates in all 26 cities, and this difference was significant in all but 3 cities (Boston, Seattle, and Washington, DC). In 2008-2012, the smallest RR was in Washington, DC (RR = 1.2; 95% CI, 0.9-1.7) and the biggest RR was in Dallas (RR = 4.2; 95% CI, 3.2-5.5). In 2013-2017, the smallest RR was in Boston (RR = 1.3; 95% CI, 0.9-1.9) and the biggest RR was in Fort Worth (RR = 4.3; 95% CI, 3.0-6.3). During the study period, the RR decreased significantly in New York (−0.3) and Houston (−1.7). Although the RR increased in 11 cities, the increase was not significant.
Discussion
This study analyzed suicide rates and racial (non-Hispanic White–Black) differences in suicide rates in the 30 largest US cities. Results showed that national suicide rates increased slightly, but significantly, from 2008-2012 to 2013-2017. Suicide rates remained stable in most cities; only 1 large city (Louisville) had a significant increase in its overall suicide rate, and 2 cities (Boston and Memphis) has significant decreases in suicide rates. In addition, the non-Hispanic White suicide rate was significantly higher (1.2-4.3 times higher) than the non-Hispanic Black suicide rate in 24 of the 26 study cities in 2013-2017. From 2008-2012 to 2013-2017, non-Hispanic White suicide rates decreased significantly in 3 cities and increased significantly in 3 cities. Non-Hispanic Black suicide rates increased significantly in 5 cities.
Although recent studies documented the differences in suicide rates by race and ethnicity, 23,39,40 no study to our knowledge has examined suicide rates across entire populations of large US cities. The consistently higher suicide rate among non-Hispanic White populations compared with non-Hispanic Black populations and the large variation in RRs aligns with suicide research at the state level, which showed that social capital from activities such as engagement with community organizations or clubs may protect against suicide; evidence of its protective effects on non-Hispanic White populations, however, were mixed. 41,42 At the county level, some evidence suggests that social capital may be more protective among non-Hispanic Black populations than among non-Hispanic White populations. 41 Investigations of suicide at the city level conducted outside the United States showed that social support protects against suicide, whereas income inequality is associated with a higher risk for suicide. 43 However, US-based studies that examine social support and income inequality at the city level do not exist.
Research also shows a link between negative life events and suicidal ideation and attempts. 44 Although the non-Hispanic Black population disproportionately faces negative life events ranging from income inequality to racism, non-Hispanic White suicide rates are higher. 45,46 Cultural factors such as religiosity among the non-Hispanic Black population may protect against suicide; however, the exact protective mechanisms of suicide among non-Hispanic Black populations are unclear. 46
Previous research that examined year-by-year comparisons found an approximate 35% increase in the national age-adjusted suicide rate from 1999 to 2018, with an average annual rate of 1.5 suicides per 100 000 population—similar to the findings in our study. 47,48 With these slight increases across large US cities, it is important to consider segments of the population who are at risk for suicide. 24 Future research should explore factors that contribute to differences in suicide rates among additional racial/ethnic categories across and within these cities. Because most suicide deaths involve guns, 23 policies that promote safe gun storage and prevent suicidal people from obtaining guns also show promise at the population level. 19,49 The Centers for Disease Control and Prevention has assembled a technical report of suicide prevention strategies across the social ecology—from individual-level interventions (eg, social–emotional programs for children and adolescents), to relationship-based approaches (eg, crisis intervention telephone lines), to community-level efforts (eg, reducing access to lethal means of suicide). 50 Data from our study may help city leaders communicate the importance of resources for these types of strategies, to address suicide in their cities.
Limitations and Strengths
Our study had several limitations. First, our mortality data came from the NVSS. Although this system uses a single, standardized dataset and a consistent methodology, a recent study showed that suicides among the non-Hispanic Black population were twice as likely as suicides among the non-Hispanic White population to be misclassified as nonsuicides; this misclassification may explain in part the disparity in suicide rates between the 2 races. 25 However, validation of NVSS data was beyond the scope of our study. A second limitation involved the removal of 3 cities from our analysis (El Paso, Portland, and San Jose) because they documented <20 suicides for the overall or race-specific populations; we also removed Las Vegas because of concerns about the reliability of death certificate data. El Paso and San Jose have small non-Hispanic Black populations, which may account for the low numbers of non-Hispanic Black suicides across the study period.
We examined 5-year periods to allow for the inclusion of the largest number of cities; small sample sizes prevented a year-by-year comparison of suicide rates. We focused on systematically examining suicide rates in large cities—an approach that is absent from the literature—however, we did not attempt to analyze causes of suicide. An assessment of policies and programs that protect against suicide in large urban areas—through the reduction of poverty and the promotion of social integration—is another important focus for future research. 51 Future research should also examine suicide mortality and disparities by characteristics such as sex and race (beyond non-Hispanic White–Black differences) across large cities, because most research that examines suicide by demographic characteristics focuses on national, state, or rural–urban differences. 48,52 Our study also suggests an intersection between place (ie, city) and race (ie, non-Hispanic White, non-Hispanic Black). Although some evidence has described the role of geographic location in suicides (eg, clustering of suicides among children and adolescents; spatial clustering of fatal and nonfatal suicide), 53,54 research that examines suicide by place and race at the city level is absent from the literature.
One primary strength of our study was that we examined a decade of data (2008-2017), during which the suicide rate increased nationally. Although other studies examined suicide across states, 48,52 to our knowledge, our study is the first study to systematically examine population-level suicide and non-Hispanic White–Black differences across 26 large cities. Recent data showed a 57% increase in suicide from 2007 to 2018 among children, adolescents, and young adults aged 10-24, and this increase prompted calls for research and intervention. 55,56 The data from our study can provide city leaders and key interested parties with an objective assessment of suicide in their city and a comparative assessment with similar cities.
Conclusion
Our study examined suicide rates in the 30 largest cities in the United States, highlighting the need for city-specific intervention efforts. Although suicide rates are increasing at the national level, most large US cities had stable rates. Racial differences in suicide rates show higher rates among non-Hispanic White populations but more consistent increases among non-Hispanic Black populations. Data from our study can clarify our understanding of suicide rates in large urban areas and can assist city leaders and other interested parties in their efforts to prevent suicide and its harms.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
