Probability Models of Needle Exchange
Abstract
Needle exchange is an intervention for slowing HIV transmission among drug injectors. Most studies of needle exchange rely on changes in self-reported risky behaviors among program participants. This paper reports two models based on objectively observed operational data, such as needle distribution rates, client visit dates, needle circulation times, and the fraction of returned needles testing HIV positive. Application of the models to data from the New Haven needle exchange suggests that HIV transmission has been slowed significantly among program participants.

