We have found that patients have distinct opinions about many components of hospital care, but that little recent work has identified those components important to today's patients in the United States. Investigators have constructed reproducible scales to obtain evaluations. Patient evaluations of the interpersonal features of hospital care are influenced by interventions that physicians or nurses identify as "higher" quality of care. We do not know if patient reports and ratings of specific aspects of care accurately reflect the effects of hospital care on health outcomes--the quality standard that public agencies and payers may consider the most important. However, patient evaluations of nursing care and medical care are independently related to patients' overall satisfaction, overall assessments of quality, and intentions to recommend and return to the same hospital. More studies are needed to clarify whether other components of hospital care also contribute to these ratings and intentions and, if so, how much. Nonresponse bias affects patient surveys, making patient ratings of care in each sample more favorable than the population mean; we do not know how this affects conclusions of surveys that compare hospitals or treatments. Few studies have compared how different methods affect the reproducibility and accuracy of patient reports and ratings. Practical issues may be the most important obstacles for users of patient ratings, particularly regarding whether potential users will be able to interpret results and accept them. Finally, no comprehensive instrument or survey method in the published literature has been tested enough to be recommended as a reproducible, accurate, and interpretable quality measure: a few do, however, appear worthy of further testing.