Accuracy of self‐reported treated hypertension in the women's health initiative: Comparisons with medication inventories

Abstract

AbstractFew studies have reported on the accuracy of self‐reported hypertension history among older postmenopausal women, which was this study's objective. Participants were postmenopausal women enrolled in the Osteoporosis and Periodontal Disease (OsteoPerio) study, an ancillary investigation of the Women's Health Initiative Observational Study (WHI‐OS) at the Buffalo, New York, clinical site. Participants self‐reported their history of physician diagnosed hypertension treated with medication at WHI‐OS enrollment (1993–1998; n = 1342, mean age 63 years), then 3 years later at OsteoPerio enrollment (1997–2001; n = 1342), and again at OsteoPerio Year 5 follow‐up (2002–2005; n = 1020). At each time point, medication inventories were recorded and served as the criterion with which self‐report was compared in the present study. Physician diagnosed‐treated hypertension was also self‐reported annually on mailed health update questionnaires in the WHI‐OS and were compared against medication inventory at the subsequent clinic exam. Of those participants who self‐reported a history of hypertension at WHI enrollment, OsteoPerio enrollment, and OsteoPerio Year 5 follow‐up, 41.2%, 90.3%, and 94.4%, respectively, had anti‐hypertensive pills in their medication inventory. Across the three time points, sensitivity and specificity ranged from 0.72 to 0.98 and from 0.85 to 0.95, and kappa coefficients ranged from 0.52 to 0.79 when comparing self‐report with medication inventory. For self‐reported newly physician‐diagnosed and treated hypertension on the annual health update questionnaire, 88.4% and 95.2% of those reporting hypertension had anti‐hypertensive pills in the subsequent medication inventory. In general, sensitivity and kappa were lower in women aged ≥70 versus < 70 years and in those with history of cardiovascular disease and diabetes compared to those without these comorbidities. In this cohort of postmenopausal women, self‐reported physician diagnosed and treated hypertension demonstrated moderate to high accuracy when compared against anti‐hypertensive medication use documented by pill inventory, particularly for those who were younger and managing fewer comorbidities.