Peer Review Report For: Socio-demographic and social support factors related to substance use in South African in-school adolescents: Insights from the Girls Achieve Power (GAP Year) trial in three peri-urban settings [version 1; peer review: 2 approved]

Background: Substance use is a significant public health problem worldwide, with consequences including violence, risky behaviours, and even death. Substance use amongst adolescents is increasing in South Africa, and limited research on frequency, risk and protective factors means that prevention interventions are difficult to design. This paper aims to describe and discuss factors associated with substance use among school-going adolescents in three peri-urban South African settings. Methods: A cross-sectional analysis was conducted using baseline data from participants in the Girls Achieve Power (GAP Year) trial. Grade 8 learners (N=2383), aged 11-18, were recruited from 26 lowest quintile public high schools in three townships: Soweto and Thembisa in Gauteng Province, and Khayelitsha in Western Cape Province. A baseline survey gathered demographic and behavioural data. Questions relevant to substance use and social support were used for this analysis. Multivariate logistic regression analyses were conducted to identify factors associated with substance use. The final variables were included in an unadjusted and adjusted logistic regression for current substance use, and a multinomial logistic regression for frequency of substance use. Results: A total of 22.5% (534) of participants indicated they had ever used substances. Being male was strongly associated with substance use (P<0.001), and less strongly with frequency of substance use. Age significantly predicted substance use, with older adolescents being more likely to engage in substance use (P<0.001); having a parent/guardian employed was negatively associated with substance use (P=0.021). Family-related social support variables were predictive of substance use. Being able to count on friends when things went wrong was predictive of lower frequency of substance use (P=0.019). Conclusions: These results can inform the targeting of prevention interventions to males and younger learners, as well as ensuring youth interventions build family and peer support to make substance use less likely and less frequent.