Mental Health
The role of alcohol in sexual orientation-related disparities in youth suicidal thoughts and behavior Sarah McKetta* Sarah McKetta Ran Barzilay
Background: Youth who identify as lesbian, gay, or bisexual (LGB) experience markedly higher rates of suicidal ideation (SI) than their heterosexual peers. They additionally report higher alcohol use, which is itself a critical contributor to suicide. Despite extensive evidence linking both LGB identity and alcohol use to suicide risk, no research to date has explored the contributing role of alcohol consumption in LGB youth SI.
Methods: The purpose of this study was to evaluate the role of alcohol consumption in LGB youth SI disparities using the Adolescent Brain Cognitive Development (ABCD) Study, a national cohort of US youth followed annually since 2016 (eligible N=11,669 unique respondents with N=38,578 observations). Using general estimating equations, we examined interactions in disparities in SI by LGB identity and alcohol consumption, i.e., whether they had ever consumed a full drink of alcohol, adjusting for age.
Results: In bivariate models, the age-adjusted odds ratio (AOR) for SI was 4.91 (95% CI 1.42-5.46) for LGB youth compared to heterosexual youth. The AOR for alcohol use (vs. none) was 4.42 (95% CI 2.93-6.66). Interaction models showed that alcohol use increased the risk of SI for LGB youth: in the absence of alcohol use, the AOR was 4.84; in the presence of alcohol use, the AOR was 7.06. The relative excess risk due to interaction was 14.11 (95% CI 0.58-27.23), demonstrating supra-additivity, and the proportion of SI attributable to interaction was 0.67 (95% CI 0.45-0.91).
Conclusions: Disparities in SI among LGB youth are dramatically increased in the presence of alcohol use. These results underscore the public health urgency of addressing youth alcohol use for suicide prevention efforts.
The CI on this AOR was extremely blown out, so we’ll want to decide how we show this information. I’ve chosen to report AOR w/o CI in text and show predicted probabilities in figure, but for the paper we’ll need a better strategy
