and not inconsistent with the proportion estimated in the controls in this study. This suggests that the effect of selection bias is probably minor.
In sum, the choice of controls for the study of hepatitis B and C virus infections and needle exchange in Tacoma is intuitively open to criticism. A broader cross section of approaches to sample injection drug users might add confidence in the inferences drawn. However, on closer inspection, there is evidence from other studies to suggest that (1) the sampling approach is not obviously biased; (2) the investigators excluded those who might have biased the sample (referrals); (3) the sample estimates and population estimates for use of needle exchange are not dissimilar; and (4) no mechanism for differential bias could be identified. Taken together, these suggest that, whereas it is prudent to be cautious about the results of any case-control study, the effect of selection bias in this situation is likely to be minimal.
The lack of any consistent alternative explanation for the Tacoma findings leads the panel to accept the primary hypothesis that implementation of the needle exchange program led to substantial risk reduction among local injection drug users, which then led to reduced transmission of blood-borne viruses (both HIV and hepatitis). Although different study designs were used, there was consistency among the results; under these circumstances, it is less likely that the findings are due to bias (Rothman, 1986). Thus, the panel concludes that the Tacoma studies show that a needle exchange program can lead to behavior change and reduced transmission of blood-borne viruses.
The Tacoma and New Haven studies and previous reviews had few, if any, empirical data on the impact of needle exchange programs on drug initiation or drug abuse among noninjection drug users. This leaves any conclusion regarding the impact of such programs on nondrug users to be theoretical at best. Although few empirical data are available on how these programs may impact overall drug use in communities that choose to implement such programs, a sizable literature does exit on the etiology of drug use. In this section we first briefly summarize what factors have been found to influence drug-use initiation and progression to drug abuse. Based on that knowledge base we attempt to foresee how needle exchange programs may potentially impact drug-use initiation and/or progression to drug abuse.
Some research suggests that the reasons people begin using drugs are different from the reasons they continue or escalate their use, which is to say, the factors that influence initiation are different from those that influence progression to more serious use (e.g., injection drug use). Several