Globally, substance use is responsible for 11.8 million direct and indirect deaths per year. Substance use impacts low- and middle-income countries (LMICs) disproportionately and is exacerbated by a large treatment gap for people living with substance use disorders. Orphaned and Separated Children (OSC) are at an increased risk of developing substance use disorders. However, there is little known about what factors predict substance use in this population. This study explores potential predictors of substance use among OSC in LMICs. Data collected from the Positive Outcomes for Orphans (POFO) study will be used to estimate predictors of alcohol use and drug use. POFO is a longitudinal cohort study that randomly sampled OSC in both community-based care and residential care settings. A bivariable regression model will be used to identify predictors. Preliminary predictors of alcohol and drug use that will be evaluated are OSC status, sex, stigma, trauma, care setting, education, emotional wellbeing, and health status. If the preliminary predictor is found to have an association, it will be placed in a larger predictive model that will be evaluated through a log binomial regression. The findings will identify predictors of substance use among OSC in LMICs that can inform future health interventions to reduce the prevalence of substance use.