Examining Heterogeneity in Adolescent Suicide Trends to Inform Suicide Prevention Policy

• In Event: Suicide and Mental Health Inequities in the U.S.: Evidence from Policy, Place, and Practice

Abstract

In recent years, concerns about mental health and suicide have intensified, driven by rising suicide death rates across various populations (Ivey-Stephenson et al., 2017). This is particularly troubling for adolescents, where suicide ranks as the second leading cause of death (Hink et al., 2022), and where key predictors of suicide-related mortality, past year suicidal ideation, are highly prevalent. Pediatric emergency department visits for mental health issues have doubled over the past decade, with an even larger increase in suicide-related visits, especially among adolescents (Bommersbach et al., 2023). For children aged 5-17, mental health hospitalization rates are notably higher in rural areas compared to urban ones (McDaniel et al., 2024). Despite some research examining trends on average or along one demographic dimension, most of the quantitative suicide-related research explores levels of outcomes and rarely examines heterogeneity of outcomes.

Our study is unique in several respects: (1) it uses repeated cross-sectional data to focus on trends in suicide outcomes for adolescents, as opposed to cross-sectional levels; (2) it assesses the extent to which there is heterogeneity in the trends across geographic regions of the U.S., as opposed to looking at only the national average; (3) it disaggregates the data by two key components identified in the suicide literature—sex and rurality; and (4) it uses two sources of population data to compare trends across a range of suicidality outcomes: active suicidal ideation, to suicide attempts, to death by suicide. 

More specifically, we use random effects within-between models (aka hybrid models), which allow us to understand how suicide risk changes over time and differentially by sex, both within geographic regions and between regions. For reference, the within-group mean centering of these models incorporates unit-level (i.e., geographic region) demeaning used in econometric fixed-effects analysis, but the use of random effects allows for both measuring the extent of variation remaining between geographic units as well as moderated relationships (Bell et al., 2019; Raudenbush & Bryk, 2002). 

Our research questions are as follows:  

1. How have trends in suicide risk evolved over time for male and female students/adolescents?
2. How do these trends vary across geographic regions and rurality?

We find that from 2007 to 2023, active suicidal ideation increased by 51.49% overall, with greater increases among female high school students (63.34%) than males (36.47%). Nonfatal suicide attempts increased by 26.81% overall, with disparate patterns by rurality: e.g., male attempts decreased by 20.38% in urban areas but increased by 99.93% in rural areas. Death rates were higher for males but increased more rapidly for females on a percentage basis, owing to the lower baseline death rate among females. Rural areas experienced steeper increases across all outcomes compared to urban areas, with males in rural areas showing the largest absolute increases in death rates. 

This work contributes to a growing literature on the mental health crisis among U.S. adolescents and points to specific subpopulations that are at elevated risk and could potentially benefit from targeted resources and policy interventions.

Author

• Sharmistha Chakrabarti, New York University
  Presenting Author