Evaluation of psychiatric consultation-liaison in primary care settings.

This article considers eight methodologic issues in the evaluation of psychiatric consultation-liaison services in primary care settings. What should be evaluated--standardized treatment regimens administered to highly selected patient groups or individualized treatments administered to heterogeneous patient groups? How can patient groups be selected to test for generic outcomes resulting from individualized treatments? When should randomized designs be employed? Should patients or physicians be randomized? How should diagnostic classification of study subjects be accomplished? What outcomes should be measured? How can study design be accommodated to patient flow in clinical settings? What resources are required to establish a productive program of research in liaison psychiatry?