Violence against children and substance abuse: A pooled study of 8 low- and middle-income countries

Abstract

Abstract Introduction Violence against children (VAC) has been linked to substance abuse later in life. However, most evidence of this association comes from high-income countries (HICs) and is understudied in the context of low- and middle-income countries (LMICs). Methods This study used the latest Violence Against Children Surveys (VACS) data from 2017-2019 across eight LMICs. The main exposure variable was defined as having experienced at least one type of violence (physical, sexual, or emotional) from any perpetrator before age 18. With the pooled dataset of 32,326 adolescents aged 13-24 years, a series of logistic regressions with country-fixed effects were estimated to examine the association between VAC and three substance abuse outcomes: smoking daily, binge drinking (i.e., more than four drinks a week) and use of illegal drugs in the last 30 days. Further, analyses based on the types of violence and whether the association differs by sex were performed. Results After adjusting for all covariates, any VAC experience was significantly associated with all three substance abuse outcomes: daily smoking (OR: 1.5, 95% CI 1.2,1.8), binge drinking (OR: 1.8, 95% CI 1.5, 2.2), and drug use (OR:1.9, 95% CI 1.5-2.3). Sexual violence, in particular, had the strongest association. Sex-stratified analyses revealed that women who experienced sexual violence in childhood were more likely (OR range 3.3-3.7) to engage in substance abuse than sexually abused men (OR range 1.2-2.1). Conclusions The findings highlight the need for interventions to prevent childhood violence and to address its consequences on substance abuse in LMICs. Key messages • Experience of any type of violence during childhood is significantly associated with daily smoking, binge drinking of alcohol, and drug use in later life. • Women who experience sexual violence in childhood are more likely to abuse substances than men who were sexually abused.