Agencies in Portland, Ore., and Washington, D.C., which are experienced in providing case management to populations similar to the drug-involved and crime-involved study participants, were contracted to implement and evaluate a case management program for drug-involved arrestees in the two cities. Average caseload size was set at 30 per full-time case manager. An average of two face-to-face contacts and two telephone contacts per month was the recommended minimum level of service for each active client. Agencies were required to have formal referral arrangements with specified types of community service provided, including those offering drug treatment programs. Case management and other study staff were prohibited from providing to the criminal justice system information about any project participant. Participants were randomly assigned to one of three types of interventions: control, intermediate, or enhanced. Outcomes were evaluated by using formal assessment instruments to measure self-reported behavior at the start of the program and again at 3 months and 6 months; independent data from criminal justice and drug treatment systems were analyzed to gauge the validity of these self-reports. Findings show that the intervention, delivered for 6 months to drug-involved arrestees released after booking, significantly reduced drug use in one of the two cities and lowered criminal recidivism in both cities. Case management was no more successful than the other interventions in reducing drug injection-related and risky sexual behaviors implicated in the transmission of HIV, although the investigators recommend modifications to the case management model that might improve outcomes associated with risky behaviors.