The purpose of this article is to suggest systematic observation of provider-patient interactions as part of the evaluation of the quality of patient care provided in managed-care settings. There have been many calls for the inclusion of patient satisfaction measures in outcome evaluations. Patient satisfaction is very important since it relates to appropriate service utilization, yet the measurement of patient satisfaction is not enough. An example of systematic observation, the "Clinic Observation Record" (COR), was used by trained observers in an early evaluation of federally-funded family planning clinics, along with a measure of patient satisfaction (PATSAT). Including time spent in the clinic, which was important but played a relatively minor role, the COR (in a regression analysis) accounted for 36% of the variance in the PATSAT score for patients in 34 clinics. Personal interactions with the doctor and nurse were the most important elements contributing to patient satisfaction. Improvements in the ways patients were addressed, introduced to staff, and treated, as well as in the clinic environment, were needed in many of the clinics observed.