Abstract
The concept of risk is integral to the field of epidemiology. Risk can be defined as the probability of an event (including drug use) occurring in a population of interest. If drug prevention and intervention programs are to be successful, they must be targeted at high-risk population groups. This article proposes a two-step, multilevel, multivariate analysis strategy to estimate high-risk population groups in high-risk geographic areas. Its successful use could result in better data for decision makers to make informed judgments about developing intervention programs.