Quality issues of self-report of hypertension: analysis of a population representative sample of older adults in Taiwan.

The study was to evaluate the quality of self-report of hypertension and examine the factors associated with under- and over-reporting of hypertension in older Taiwanese. Data for this analysis were from the Social Environment and Biomarkers Study in Taiwan 2000, which involved a national sample of 1021 Taiwanese over 54 years of age. We performed binary classification tests to compare the prevalence rates of self-reported vs. clinically measured hypertension according to World Health Organization (WHO) (blood pressure ≥ 160/95 mm Hg or on hypertension medication) and JNC-6 (140/90 mm Hg or on hypertension medication) definitions. Logistic regression analysis was performed to analyze the potential factors associated with under- or over-reporting of blood pressure status. Results showed the test characteristics of self-reports were: sensitivity 73%, specificity 93%, and kappa = 0.68 (p < 0.001) based on the WHO definition; and sensitivity 51%, specificity 95% and kappa = 0.43 (p < 0.001) based on the JNC-6 definition. Old age was associated with over-reporting whereas having no health checkup during the past 12 months was associated with under-reporting. The relatively low agreement between self-reports and clinically measured hypertension (JNC-6 definition) was mainly due to the lack of a well-defined hypertension practice guideline and the failure of clinicians to clearly inform patients of their diagnoses. The consistency of hypertension practice guidelines and the effectiveness of informing the patients of their diagnoses are two main factors impacting the quality of self-report of hypertension in elderly Taiwanese. Better self-reports of health data can improve the efficiency of public health surveillance efforts.