Suicide contributes to significant mortality in the United States, accounting for over 48,000 deaths in 2021 alone. There are multiple risk factors for suicide, including age, access to mental health care, geographic isolation, stigma, at-risk substance use, access to firearms, and socioeconomic factors. Many of these factors disproportionately affect people living in rural areas.

This project will examine the variation in suicide rates by geography and explore its driving factors among urban and rural areas in the United States from 2018 to 2021. The research design of this study will adopt a cross-sectional, retrospective approach, using secondary data.

Data from the National Vital Statistics System maintained by the Centers for Disease Control and Prevention will be used for this project. Age-adjusted suicide mortality data from multiple-causes-of-death mortality files for years 2018-2021, with the underlying cause-of-death codes will be used to define suicide. These data will be combined with other publicly available data sources for county-level sociodemographic, economic, health care characteristics, and health indices.

For county-level analysis, researchers will aggregate suicide rates from 2018-2021 and explore the spatial distribution across the U.S., and then compare these rates with demographic characteristics using multivariable linear and logistic regression. Researchers also plan to explore the utility of integrating an area-level measure of social vulnerability as part of the multivariable analyses. To further understand the spatial patterns and distribution, researchers will conduct spatial analyses such as hotspot analysis and bivariate mapping among counties in the U.S.

At the state level, researchers will analyze suicide rates per year (2018, 2019, 2020, and 2021) and by rurality (Rural-Urban Continuum Codes) using bivariate analyses. This will be followed by multivariable analyses that control for county-level characteristics. Similarly, researchers will compare these rates with demographic characteristics and examine the feasibility of incorporating an area-level measure of social vulnerability. These analyses will use multivariable linear and logistic regression and spatial regression.

Findings from this study will be of interest to policymakers, practitioners, researchers, and community leaders. Such applications of findings from this study will predominantly affect vulnerable communities with large rural residents.