HOME / PUBLICATIONS /

Population size estimation of people who inject drugs using capture-recapture analysis of multiple respondent-driven sampling rounds: Implications for HIV/HCV burden and harm reduction service planning in Thessaloniki, Greece

Citation:

Roussos S, Tsirogianni E, Goulis I, Kalamitsis G, Hatzakis A, Sypsa V. Population size estimation of people who inject drugs using capture-recapture analysis of multiple respondent-driven sampling rounds: Implications for HIV/HCV burden and harm reduction service planning in Thessaloniki, Greece. Int J Drug Policy. 2025;147:105078.

Abstract:

BACKGROUND: Accurate population size estimation of people who inject drugs (PWID) is essential for evidence-based drug policy and service planning, yet it remains challenging. An emerging HIV outbreak in Thessaloniki, Greece's second-largest city, highlighted the urgent need for evidence-based population size estimates. METHODS: We applied capture-recapture analysis to five respondent-driven sampling (RDS) rounds conducted during 2019-2021 to estimate PWID population size in Thessaloniki for the 2019-2021 period. These RDS rounds were part of a community-based program aimed at increasing HIV/HCV testing and linkage to care among PWID. We treated each RDS round as a capture source and used log-linear models to estimate PWID population size (past 12 months and past 30 days), accounting for potential dependencies between rounds through interaction terms. We then estimated HIV/HCV disease burden and assessed prevention and harm reduction service coverage against international standards (HIV testing, OAT, NSP). RESULTS: Based on data from 1093 unique participants across five rounds (53.9% currently injecting, 20.3% currently in OAT), capture-recapture analysis estimated 1512 PWID (95% confidence interval (CI): 1345-1741) who had injected drugs in the past 12 months. The estimated prevalence of injecting drug use was 0.22% (95% CI: 0.20-0.25) among adults aged 18-64 years. We estimated 106 people living with HIV (95% uncertainty interval (UI): 83-130) and 945 HCV-antibody-positive individuals (95% UI: 815-1077) among PWID. Needle and syringe program coverage was 36 (95% CI: 31-40) syringes per PWID in 2021. CONCLUSION: Based on this community-based population size estimate, the prevalence of injection was nearly double the official national Greek average. The annual distribution of syringes should increase by 5.6 times to reach the WHO target (≥200 syringes/PWID/year). These findings demonstrate how community-based programs with multiple RDS rounds can also yield population estimates essential for evidence-based drug policy interventions.

Notes:

1873-4758Roussos, SotiriosTsirogianni, EffrosyniGoulis, IoannisKalamitsis, GeorgiosHatzakis, AngelosSypsa, VanaJournal ArticleNetherlandsInt J Drug Policy. 2025 Nov 27;147:105078. doi: 10.1016/j.drugpo.2025.105078.