Two trained reviewers screened abstracts from each of the 2,695 unduplicated references for specific criteria. To avoid over-estimation of population size, we excluded studies likely to have disproportionately high proportions of PWID (e.g., studies among HIV or HCV infected persons; studies focused on populations who are mentally ill, incarcerated, or in drug treatment; and studies that selected for PWID and non-PWID separately). We also excluded school-based samples (including university students) as these may under-estimate the PWID population proportion. We included studies if they reported the proportion of PWID in the sample. We screened conference abstracts and dissertations but did not include them, as they provided too few details. A total of 2,702 abstracts met the key terms; 2,695 were excluded in first-level screening. Full reports were obtained for the 7 abstracts that met the Phase 1 selection criteria.
