The early history of QA is one of stunning achievements. Men and women led by a vision of what hospital health care should be brought about fundamental changes in the structure of hospitals and medical schools. These successes required lifetimes of work and enormous capital (Flexner's war chest would have amounted to billions of 1990 dollars). A second wave of reform included the creation of credentialing committees, tissue committees, and infection control efforts. Although less dramatic in their impact, these efforts have had measurable influence on the outcome of health care and, under the joint administration of local hospitals and the JCAHO, continue to guarantee excellence in health care. The most recent attempts at quality assurance, driven at least in part by a federal mandate to control costs, have been much more modest in their success. Committed groups working within a sound theoretical framework have had great difficulty monitoring and evaluating centrally a process as decentralized as health care. The regularity with which new fashions in QA have appeared underscores the frustration felt with this approach. If the quality of health care is to be monitored centrally, reliable measures of quality will be required. No one knows if such measures actually exist. The absence of objective evidence that quality has been improved by these efforts suggests that little has been accomplished, perhaps because all easily attainable improvements had already been implemented. The basic concept of a centrally monitored hospital structure within which provision is made for ongoing observation and innovation by those actually involved in the care of patients retains the advantages of central monitoring and local invention. It is a model that will be hard to improve on. We should persist in our efforts, but fundamental advances are unlikely. In the end, quality is only doing everything the best it can be done. The best quality assurance requires an excellent hospital in which a well-trained and committed staff has a central mandate and local purview to do the best possible job. Two thousand years after Hammurabi's minions busied themselves determining the appropriate punishment for various medical misadventures, the Greek Hippocrates advocated a less complex approach to the same problem. He suggested that the quality of health care could be best guaranteed by administering an oath to physicians, exhorting them simply to do what was best for their fellow men. This approach has seen us through two millennia and, however variable its results, may not be easily improved on.