Accuracy of a health interview survey in measuring chronic illness prevalence.

The accuracy of a household health interview survey in measuring the prevalence of chronic illness was tested by comparing survey responses with the same respondents' medical records. Results indicate that the survey is far from a perfect discriminator of recorded medical status. The false-positive rate was 25 percent, and the rate of false negatives was nearly 40 percent. False positives were greater for Mexican-Americans and other nonwhites than for white respondents. The household health interview survey is nonetheless an important planning tool since it reflects respondents' perceived health levels, which may be used to predict utilization of primary medical services.