Abstract
Monitoring injecting drug users' (IDUs) health is challenging because IDUs form a difficult to reach population. We examined the impact of recruitment setting on hepatitis C prevalence. Individual datasets from 12 studies were merged. Predictors of HCV positivity were sought through a multilevel analysis using a mixed-effects logistic model, with study identifier as random intercept. HCV prevalence ranged from 21% to 86% across the studies. Overall, HCV prevalence was higher in IDUs recruited in drug treatment centres compared to those recruited in low-threshold settings (74% and 42%, respectively, P < 0·001). Recruitment setting remained significantly associated with HCV prevalence after adjustment for duration of injecting and recent injection (adjusted odds ratio 0·7, 95% confidence interval 0·6-0·8, P = 0·05). Recruitment setting may have an impact on HCV prevalence estimates of IDUs in Europe. Assessing the impact of mixed recruitment strategies, including respondent-driven sampling, on HCV prevalence estimates, would be valuable.