Introduction: Over 90% of all adolescent suicides occur in low- and middle-income countries (LMIC), yet the majority of suicide research has focused on primarily high-income countries (HIC).
Method: Using nationally representative data on 82,494 adolescents from thirty-four LMIC, this research employed machine learning to compare the predictive effects of multiple determinants of suicidal behaviors previously identified in the literature.
Results: Results indicate that distinct predictors are present for suicidal ideation, suicidal planning, and suicide attempts in youth living in LMIC as well as shared predictors common to all three behaviors.
Conclusion: These findings provide insights into the unique needs in global mental health policy and efforts within and across adolescents in LMIC.
Keywords: Random Forest; adolescents; low- and middle-income countries; machine learning; suicide.
© 2023 The American Association of Suicidology.