Validity of self-reported hypertension is inversely associated with the level of education in Brazilian individuals

Background: Self-reported hypertension is an important piece of information for public health that is available in epidemiological studies. For proper use of this information, such studies should be validated. Objective: To validate self-reported hypertension and associated factors in adults and elderly individuals in Sao Paulo, Brazil. Methods: Participants were selected from the sample of a population-based cross-sectional health survey carried out in Sao Paulo (ISA Capital-2008). Their age was 20 years or older, they were from both genders, and had their blood pressure measured (n = 535). Hypertension was defined as blood pressure >= 140/90 mmHg and/or use of medication for hypertension. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and Kappa coefficient were calculated. Poisson regression was used to identify factors associated with sensitivity of self-reported hypertension. Results: Sensitivity of self-reported hypertension was 71.1% (95%CI: 64.8 to 76.9), specificity 80.5% (95%CI: 75.6 to 84.8), PPV 73.7% (95%CI: 67.4 to 79.3), and NPV 78.5% (95%CI: 73.5 to 82.9). There was moderate agreement between self-reported hypertension and hypertension as diagnosed by blood pressure measurement (kappa = 0.52, 95%CI: 0.45 to 0.59). Body mass index and level of education were independently associated with sensitivity (body mass index >= 25 kg/m(2): PR = 1.42, 95% CI: 1.15 to 1.76; schooling >= 9 years: PR = 0.71 95% CI: 0.54-0.94). Conclusion: Self-reported hypertension was shown to be valid in adults and the elderly in the city of Sao Paulo, and is thus an appropriate indicator for the surveillance of hypertension prevalence in the absence of blood pressure measurement. Overweight was positively associated with validity of self-reported hypertension. Further studies are needed to elucidate the inverse association between the validity of self-reported hypertension and level of education. (Arq Bras Cardiol. 2013;100(1):52-59)