There is a need for indicators of the outcome of health-care services against which the use of resources can be evaluated. From a previously published series of outcome indicators, which included diseases for which mortality is largely avoidable given appropriate medical intervention, causes were selected which were regarded as most amenable to medical intervention (excluding conditions whose control depends mainly on prevention) and for which there were sufficient numbers of deaths to allow an analysis of the variation in mortality rates among the 98 area health authorities of England and Wales. Considerable variation between AHAs was found in mortality from most of these diseases, and this variation remained even after adjustment for social factors. This substantial variation should be examined further in relation to health-service inputs and other factors. A finding of large variations in the quality of health-care delivery in different parts of the country would have important implications for resource allocation.