Although the three models differ in many ways, the primary distinguishing factors are as follows. The simplified circulation model focuses on obtaining estimates of averted HIV infections among program participants, based on a 1-year time frame, attributable to transmission from sharing contaminated needles. The behavior change model estimates averted infections among all injection drug users in the community as well as their sexual partners and offspring over a 5-year time interval, attributable to transmissions from both sharing and sexual behavior. The combined circulation and behavior change model incorporates characteristics of both models.

As depicted in Table 3.2, all models estimated that the hypothetical needle exchange program had a substantial impact on averted HIV infections and reported significantly lower costs per HIV infection averted than the $119,000 estimated lifetime HIV medical costs associated with treating a person infected with HIV (Hellinger, 1993). One noticeable disparity is the substantially different estimates in the number of infections averted over 5 years reported by the behavior change model compared with the combined model. This disparity is due mainly to the fact, as with the circulation model, that the combined model limits its estimates to infections averted among program participants only (participants are assumed to share only with each other).

These are obviously crude estimates of the cost-effectiveness of needle exchange programs and are of limited value in assisting health policy makers