Two principles shape an applied research field such as health services research. First, studies can be classified as methodological, descriptive, analytical, or experimental; and second, health services research is conducted at two levels, the doctor-patient level and the policy level. Twenty years of increasingly sophisticated description and analysis allow researchers now to measure better the appropriateness and outcome of care, both at the policy level (do cuts in medical coverage affect patients' health?) and at the individual level (does physicians' behavior affect patient compliance and thereby outcome?). Further, researchers have been able to develop increasingly useful measures of patient health status and function. Consequently, relations between the public and physicians and medical institutions will change further as patients receive more and better information about hospital mortality rates, quality-of-life measures, and similar results of health services research. Conducting good health services research is complicated, depends upon teams of physicians and social scientists working together, and produces complex, not simple, results, but it is worth the effort because the results make a difference to physicians and to the American people.