To decrease adolescent morbidity and mortality and improve the quality of life, a violence-prevention consultation is offered to hospitalized victims of nondomestic violence. The context is a violence-prevention team approach to patient assessment, treatment, and follow-up. Psychoeducational counseling emphasizes the individual through a cognitive behavioral approach and also recognizes the individual in the proximal social setting through referrals to community resources. The in-hospital component draws on the health beliefs model, self-efficacy, the theory of reasoned action and their synergy with cognitive mediation theory as expressed in developmental psychology. The target group for the intervention is adolescents (12-17 years of age) who have been victims of violent assaults severe enough to warrant treatment at a Level One trauma center. The six steps in the intervention are to (1) review and assess the incident, (2) review the patient's conflict-resolution strategies and introduce nonviolent alternatives, (3) provide information on the prevalence of violence/homicide and determine the patient's risk status, (4) explore the patient's coping skills and support system, (5) develop a plan to stay safe, and (6) refer patient to services for follow-up activities. Approximately 15 study participants are identified each month, half of whom are randomly assigned to receive the intervention. Over the 12-month recruitment interval, approximately 180 adolescent patients will be identified. Baseline data are collected through hospital intake procedures and chart reviews. A battery of standardized measures supplemented by a brief structured, closed-ended interview is collected four months after the youths leave the hospital. Preliminary baseline data for 39 youths are reported. The "typical" youth is a 16-year-old African-American male. Even though nearly one third of victims had been shot, the typical patient was injured in a fight during which he was kicked, bitten, or beaten with or without a blunt instrument. The majority of incidents involved only one attacker who was known to the victim. Nearly half the injuries were precipitated by an argument or fight. No statistically significant differences between intervention subjects and nonintervention controls in terms of baseline variables have been observed. For inner-city adolescent victims of violent assaults, a hospital-based intervention offers a unique opportunity for reduction of the incidence of reinjury. We describe the elements of the intervention, including the theoretical basis and implementation; detail the overall evaluation design including modifications; and present preliminary analyses of baseline data.