Last in a series of studies assessing medical care for the New Mexico Medicaid population and evaluating the impact of a physician peer review organization (a PSRO prototype). This study developed validated algorithms for creating "episodes" of care from computerized insurance claims, compiled comprehensive age/diagnosis-specific "profiles" for judging quality of care, and investigated differences in quality by physicians' professional characteristics. Quality of care for the six respiratory infections was often inadequate; a few "outliers" accounted for much of the poor care. Generally, specialty-board-certified physicians gave better care than noncertified; medical doctors better care than osteopaths; and group practices with board-certified members gave the best care. The peer review organization substantially curtailed the use of inappropriate injectable antibiotics, especially among physicians with the poorest practice initially. The implications of the methodologic advances and the empirical findings for quality assessment and the Professional Standards Review Organization program are discussed.