This paper highlights what the authors believe are the important issues and directions of change in the evaluation of mental health programs. The rationale for such evaluation is twofold. First, it provides a scientifically rigorous method of determining the therapeutic efficacy of the treatment or program, and secondly, these results can exercise a feedback into the system, modifying the clinical operations. The major types of evaluation studies are considered: intra-institutional studies, effect of relocation from one institution to another, alternative treatments upon admission to institutions, and alternatives to institutionalization. Further issues considered include contented vs. angry patients, patient deterioration, and covert deterioration which is less obvious to the observer. It is considered that the best hope for the future is a comprehensive mental health system with many flexible alternatives, integration of community and institution, and continuity of care. Extensive references are included. (Author/SES)