The authors performed a critical review of experimental studies on the outcomes for psychiatric patients of 1) alternatives to hospital admission, 2) modifications of conventional hospitalization, and 3) alternatives to continued long-term hospitalization. The internal validity of many of the studies was compromised by shortcomings in design and performance and generalizability limited by selection of patient populations. With these qualifications experimental alternatives to hospital care of patients have led to psychiatric outcomes not different from and occasionally superior to those of patients in control groups. This conclusion is best supported for alternatives to admission and for modifications of conventional hospitalization. The available studies do not permit firm conclusions regarding alternatives to continued long-term hospitalization of chronically ill patients or for a critical analysis of the optimal management of specific subpopulations of psychiatric patients. Satisfactory deinstitutionalization appears to depend on the availability of appropriate programs for care in the community.