Concern about the spiraling cost of mental health care has increased the need for reliable data about the outcomes of such care. Treatment outcome studies, although difficult to design and carry out, are essential in demonstrating the efficacy of psychiatric treatment, rationalizing clinical decision making, and encouraging public support for the availability of appropriate, cost-effective care for the mentally ill. Ideally, outcome studies should focus not only on clinical symptomatology but also on patients' social, interpersonal, and occupational adjustment as well as on factors that, taken together, shape the quality of life. The authors discuss these issues and some of the impediments to collecting useful outcome data, including the complexity of the mental health service delivery system and problems related to patient selection, study design, and how treatment is defined.