Assessing syringe exchange programs

Anyone who has visited several syringe exchange programs (SEPs) is likely to be convinced that there are meaningful differences among the programs. Syringe exchange programs vary in their staffing, operating characteristics, characteristics of the participants and in their legal, cultural and economic environments [1‐3]. Epidemiological research indicates that some of these differences may be very important. Syringe exchange programs have been associated with preventing HIV epidemics among drug injectors [4‐7] and with reducing HIV incidence in high seroprevalence HIV epidemics [8,9], but there are also instances in which the presence of a syringe exchange program was not sufficient to prevent high incidence rates among local IDUs [10]. In this issue, Bluthenthal and colleagues [11] report on a comparison of three different US syringe exchange programs—in Hartford, Connecticut, Chicago, Illinois and Oakland, California. The major findings are certainly plausible: that police contact with drug injectors is associated positively with living in an area with no legal possession of syringes and that the lack of a cap on syringes to be exchanged is associated with less re-use of syringes. However, it is very important to recognize that comparisons of syringe exchanges in different cities involve many different dimensions and that this study had a sample of only three programs. For example, Hartford had the lowest rate of police contact with IDUs and also had the highest rate of receptive sharing (see Table 1 of Bluthenthal et al. in this issue). Additional research is needed urgently on the relationships between characteristics of syringe exchange programs, the participants in those programs, the legal/social/economic environment and actual HIV infection among IDUs in the area. This should include both risk behavior and actual HIV infection rates among program participants. There is a need for both additional case studies and quantitative syntheses of data from large numbers of programs. Statistical techniques of meta-analysis [12] and hierarchical linear modeling [13] may be very useful in quantitative syntheses of data from moderate to large numbers of programs. Qualitative studies, including case studies, can provide better understanding of complex or potentially anomalous findings, and enable development of new hypotheses.