Have Morbidity Surveys Been Oversold?

Abstract

Recent studies in Hunterdon County, N. J., Baltimore City, Md., and Pittsburgh, Penna., have demonstrated the unreliability of diagnostic information derived from household interviews when compared with physical examination findings. In a study by the National Health Survey in which prevalence information was compared with medical records, the results were almost equally disappointing. The Nashville Study showed a two-thirds discrepancy in the prevalence of the same diseases, reported in the household survey as compared with that given to a physician immediately prior to a physical examination. In a study of Kit Carson County, Colorado, it was shown that different interviewers reported widely different prevalences of disease in randomly assigned population groups. From certain other studies we are led to conclude that individuals when queried try to surmise the reasons for the inquiry and formulate answers in conformity with some self-divined purpose rather than according to the facts known to them. It is apparent, therefore, that diagnostic information from household interviews has little objective validity. To obtain morbidity data of greater reliability, it is recommended that use be made of diagnostic information from household surveys to stratify the sample to be examined clinically for appropriate diagnoses.