The analysis produced seven factors -- customer/prostitute sex, steady non-IDU ( non injection drug user) sex partner, injection frequency, non-steady sex partner, anal sex, needle sharing, and steady IDU sex partner -- which accounted for 69 percent of the variance. These factors and their intercorrelations showed them to be highly discrete, demonstrating the heterogeneity of the individuals engaging in these risk activities. However, the results suggest that, while drug injectors as a group engaged in a wide range of high-risk activities, individuals tended to restrict themselves to a more restrictive set of behaviors. This point was illustrated through the profiling of two individual respondents. The independence of risk factors implies that prevention messages must be highly individualized in order to relate to those involved in particular risk activities and promote behavioral change. The study proved the feasibility of assessing behavioral risk through an approach that did not focus on individual risk activities nor on a priori assumptions about the correlations of risk factors. 3 tables, 1 figure, and 43 references