A multilevel approach for assessing the variability of hepatitis C prevalence in injection drug users by their gathering places

Abstract

The aim of this study was to investigate the variation in hepatitis C virus (HCV) prevalence in injection drug users (IDUs) by their gathering places, using a multilevel approach. IDUs recruited from their gathering places were invited to respond to a questionnaire on demographics, drug use history, injection behaviors, and methadone treatment. Dried blood spots were collected for HCV antibody testing by ELISA. Factors associated with the anti-HCV test result were explored by linear logistic regression, followed by the evaluation of heterogeneity between gathering places by multilevel analysis. A total of 622 respondents from 19 gathering places in Hong Kong, recruited between August and September 2011, were evaluated. Anti-HCV seroprevalence was 81.7% (95% confidence interval 78.6-84.7%), ranging from 67% to 100% by gathering place. HCV infection was associated with current practice of injection, needle-sharing, and midazolam injection. On multilevel analysis, there was a modest but significant variation in HCV antibody prevalence by gathering place, adjusted by midazolam injection (adjusted odds ratio (AOR) 3.91) and current injection (AOR 2.88) or injection over a long duration (AOR 3.17). There was heterogeneity in HCV antibody prevalence in IDUs by gathering place, while the influence of injection behaviors varied, suggesting interactivity between factors at the individual and group levels.