Applying a classification and regression tree approach to identify individual, socioenvironmental, and psychological interactions associated with suicidal ideation among Latinx adolescents

Document Type

Article

Publication Date

1-1-2021

Abstract

Introduction: The present study sought to uncover previously unknown interactions stemming from individual, socioenvironmental, and psychological factors associated with suicidal ideation among Latinx adolescents using a classification and regression tree (CART) modeling approach. Methods: This cross-sectional study included national data from a sample of 656 Latinx adolescents who participated in the 2018 National Survey on Drug Use and Health. Self-report data was collected on suicidal ideation, individual, socioenvironmental, and psychological factors. Results: CART identified several interactions that classified Latinx adolescents as endorsing suicidal ideation. The first split identified parental monitoring as a significant predictor of suicidal ideation, successfully classifying over 50% of the sample (n = 360). The ensuing interaction occurred between lower parental monitoring and depression symptoms, followed by increased parental monitoring and BMI. Other interactions included sex, self-rated health, household size, and peer relationships. CART results suggest that BMI was the most influential variable associated with suicidal ideation, followed by self-rated health and parental monitoring. Conclusion: Since suicidal thoughts and behaviors are malleable, utilizing an individual, socioenvironmental, and psychological framework would continue to provide comprehensive approaches to identifying data-driven interactions between risk and protective factors that may lessen the heightened risk of suicidal ideation among Latinx adolescents.

Publication Source (Journal or Book title)

Journal of Health and Social Sciences

First Page

527

Last Page

543

This document is currently not available here.

Share

COinS