record beyond a typically brief admitting history that may omit significant past events and illnesses. Patient satisfaction questionnaires may be distributed at the time of discharge, but are not usually part of the medical record. No information about the outcome of inpatient care is gathered routinely after discharge. Thus, assessment of the quality of care must depend on data collected during what is usually a short hospital stay.
In most moderate and large hospitals, patient care is very visible to peers because of the many practitioners who see both the patient and his or her medical record. Major procedures are generally observed by an array of physicians, nurses, and others, who are expected to use the incident and screening systems to report adverse occurrences. Teaching hospitals have substantial numbers of trainees and systems of supervision and accountability. There are multiple opportunities for consultation and patient care conferences, especially in the teaching hospital. The situation may be quite different in small rural hospitals relying on a handful of physicians.
Ideally, guidelines in the form of expert systems could be contemporaneous with care and help the practitioner in decision making. Some prototype interactive systems now provide computerized clinical reminders (McDonald et al., 1984; Tierney et al., 1986, 1988), predictive value of tests, and warnings of drug-drug interactions or misuse of antibiotics.5 Implementation of these systems is likely to occur first in the hospital environment.
Medical data systems tend to be more useful for detecting poor technical quality rather than overuse or underuse. Where data permit the development of performance profiles, results can be distributed to providers and performance followed over time. These data can figure prominently in decisions about clinical and staff privileges, which are the chief means of interventions exercised by hospital officials and directed at physicians who are considered to provide substandard care. Assuming that firm standards of care can be developed and applied for quality assessment, peer pressure from colleagues and pressure from chairpersons of departments and others with clinical authority can be brought to bear on those individuals not performing according to expectations.
Nevertheless, these arrangements can be weak instruments of problem correction, especially in smaller institutions where personal and social ties may be stronger than organizational procedures or leadership or where medical staff are intimidated by legal action from sanctioned practitioners. During site visits for this study, people responsible for hospital quality assurance repeatedly noted extreme difficulties in dealing with problem staff members