Is efficacy the gold standard for quality assessment?

"Efficacy"-medical care that is achievable under optimal conditions-is generally considered to be the appropriate standard for quality assurance. In this paper I argue that what is needed is a broadened definition of efficacy that includes the appropriateness of an intervention as well as the level of technical skill used in its provision. To establish efficacy, so defined, I suggest that well-designed nonexperimental, as well as experimental, studies be undertaken that employ registries and data banks, as well as formal clinical trials, and that decision analysis be used to synthesize the results. This will require a significant effort and investment in outcome studies of medical and surgical technologies. With growing recognition by the public that quality can be highly variable-aided by the Health Care Financing Administration's publication of hospital mortality and morbidity data-the medical care system will ignore outcome studies and quality assurance at its own peril.