Validity of self-reported hypertension in the National Health and Nutrition Examination Survey III, 1988-1991.

BACKGROUND: The National Health and Nutrition Examination Survey (NHANES) is the main data source for hypertension surveillance. However, because of a gap of almost 10 years between each NHANES, self-reported data from annual surveys need to be examined as an alternative data source. This study analyzes the validity of self-reported hypertension in a national sample of non-Hispanic whites, non-Hispanic blacks, and Mexican-Americans.

METHODS: Sensitivity, specificity, and predictive values positive (PVP) and negative (PVN) of self-reported hypertension were calculated against two definitions of hypertension: the definition recommended by the Third Joint National Committee on Hypertension, JNC III (blood pressure > or = 140/90 and/or taking antihypertension medication) and a broader definition including control with lifestyle modifications. Data used come from the NHANES III, 1988-1991.

RESULTS: Overall test characteristics using the JNC III definition are sensitivity 71%, specificity 90%, PVP 72%, and PVN 89%. Test characteristics were consistently higher for the broad than for the JNC III definition. Validity of self-reported hypertension is higher among women than among men and among persons with a medical visit during the past year than among those with no visits: validity was lowest among Mexican-American men. Due to the similarity between sensitivity and PVP, the prevalence of self-reported hypertension is nearly equal to the prevalence of JNC III-defined hypertension.

CONCLUSIONS: Self-reported hypertension may be used for surveillance of hypertension trends, in the absence of measured blood pressure, among non-Hispanic whites and non-Hispanic black women and persons with a medical visit in the past year. Validation should be repeated with each NHANES.