One function of contemporary psychiatric emergency services is to divert patient admissions from state hospitals. Underlying this mandate are a series of untested assumptions about the positive effects of admission diversion. The author examines these assumptions using data on inpatient admissions from a crisis intervention service. Although the service was successful in preventing first admissions to the state hospital, the rate of recidivist admissions increased. Inpatient treatment in general or private hospitals did not result in shorter lengths of stay or fewer bed days than state hospital treatment. Because patients could be sent to any of several hospitals, some located far from the catchment area, continuity of care and treatment in the local community were not advanced by diverting admissions from the state hospital.

Integrated mental health and substance abuse treatment within an assertive community treatment (ACT) approach was compared to that within a standard case management approach for 223 patients with dual disorders over three years. ACT patients showed greater improvements on some measures of substance abuse and quality of life, but the groups were equivalent on most measures, including stable community days, hospital days, psychiatric symptoms, and remission of substance use disorder.

This article addresses the ethical and legal duties that must be confronted in any study of the risk of interpersonal violence in the community. Ongoing research--the MacArthur Risk Assessment Study--on the markers of violence among released mental patients is taken as illustrative. Methods by which the researchers are discharging their legal and ethical duties are described and justified. Strategies center around the duty to protect research subjects from their own violence, and the duties to protect research staff and third parties from subjects' violence. By airing these rarely discussed issues, the authors hope to initiate a professional dialogue on crucial ethical and legal aspects of the research process.