Our experience suggests that (1) comprehensive objective information is a valuable medical resource; (2) the systematic acquisition of objective information in a form suitable for computers is a necessary first step to building a useful data base; (3) when appropriately utilized, objective medical information may permit a study of medical practice, cost of care, and operational efficiency; (4) analysis of costs divorced from outcome and rehospitalization experiences may be misleading; (5) for some desired studies, the information required may be obtainable within a single institution. For others, however, community-wide studies may be the only possible way to obtain the information needed; (6) useful computer systems will be developed only if the health professionals become more knowledgeable (and demanding) and administrators are persuaded to provide effective support; and (7) effective hospital support for adequate medical information systems will probably come when criteria evaluating hospital performance are altered to reflect society's needs. Our experience with computer evaluation of care has been a sobering one. Still, it is obvious that the technology offers considerable hope that many inefficiencies in practice can be corrected. Considering all the current jeremiads about what we can afford in health care, it is reassuring to learn that good medicine is cost effective.