Keywords: Causal inference; Respondent Driven Sampling (RDS); clinical trials; heroin; opioid substitution therapy; people who inject drugs (PWID); polysubstance use; prescription drug abuse; propensity score matching.