in the number of improperly discarded used needles. Another possibility is that the issuance of injection equipment, condoned by government, will "send a message" undercutting efforts at combatting illegal drug use and will promote more drug use (with more attendant HIV incidence). A third possible negative outcome is that needle exchanges will lower the perception of risk of injection drug use and thus attract more users to inject drugs and to other forms of illegal drug use.

Appraisal of the success of a needle exchange program, then, should also attend to measures of these possible negative outcomes. An observed pattern of negative outcomes would weigh against the idea that needle exchange programs are beneficial.

Assessment of the effects of needle exchanges involves the simultaneous consideration of a number of intended positive and unintended negative outcomes (Table 7.1). Ideally, a successful exchange program would reduce the risk of new infection among injection drug users without increasing drug use and health risks to the public.

The purpose of this section is to clearly explain the panel's perspective in assessing the effectiveness of needle exchange programs. We first briefly review the basis for the traditional review procedures. We then outline the argument for a different approach—one that examines the patterns of evidence in order to assess effectiveness.

Traditional reviews of experimental analysis search for studies with well-controlled research designs. By well-controlled designs, we mean ones