Blood pressure measurements are unreliable to diagnose hypertension in primary care

Abstract

To evaluate the accuracy of blood pressure (BP) measurements performed by primary care physicians (PCPs), and to assess whether it improves following a short theoretical training. Observational study in 26 primary care practices in Geneva, Switzerland. The PCPs were asked to measure BP on 10 volunteers, within the usual context of their practice. Two trained research assistants repeated the measures immediately after the PCPs. The PCPs were then randomized to receive detailed training documentation on standardized BP measurement (group I: 14 doctors) or information about high BP (group II: 12 doctors). Measures were repeated a few weeks later. We computed accuracy and diagnostic categorization of high BP comparing the PCPs' measurements to the average value of four measurements by the research assistants (gold standard). T-tests were used to compare measurements between the two educational groups, both at baseline and following exposure to the educational material. BP measurements were performed on 257 volunteers at baseline and 251 after training. At baseline, the mean BP difference between PCPs and the gold standard were 23.0 mmHg (21.3-24.6) for systolic and 15.3 mmHg (14.3-16.3) for DBP. Following training, the mean difference remained high [group I: 22.3 mmHg (20.4-24.2) and 14.4 mmHg (12.6-16.2); group II: 25.3 mmHg (22.7-27.9) and 17.0 mmHg (15.3-18.7)]. As a result, 24-32% volunteers were misdiagnosed as having systolic hypertension and 15-21% as having diastolic hypertension. Though widely used in the assessment of hypertension in ambulatory settings, in-office BP measurements are highly inaccurate, even following training.