The prevalence of diabetes mellitus in Oxford and related epidemiologic problems.

Abstract
The results of the Oxford, Massachusetts, pioneer diabetes survey of 1946-47 are discussed in relation to the ideal requirements for such studies. Reconstruction of the original study population is described, and problems related to the accuracy of the population definition are explored. Disease definition and its significance for epidemiologic studies are examined. Both the choice of criteria and the uniformity with which they are applied were found to supersede the importance of refinements in population definition. Several suggestions regarding future epidemiologic studies of diabetes mellitus and frequency distributions of the original data are presented to facilitate further inter-study comparisons.