Background: Automated devices (AD) for measuring blood pressure (BP) are gradually replacing mercury sphygmomanometers (MM) in clinical settings. However, the use of ADs in epidemiological surveys has not been established. We investigated the factors associated with measurement differences when using an MM and an AD.
Methods: Two trained observers took three BP measurements in 454 subjects as part of an epidemiological survey, alternately using an MM and an AD. BP measurement difference was defined as BPMM - BPAD. Alarm reactions (ARs) were calculated by subtracting the third systolic BP (SBP) measurement from the first SBP.
Results: The mean age of subjects was 50.7±15.4 years (n = 454). The mean BPs using the MM and the AD were 119.8±13.9 vs. 119.5±13.6mm Hg in males and 115.0±16.8 vs. 111.6±15.7mm Hg in females for SBP and 77.7±10.4 vs. 74.7±10.4mm Hg in males and 73.2±9.3 vs. 69.9±10.3mm Hg in females for diastolic BP (DBP). Age, gender, arm circumference, and AR were the factors related to the difference. The concordance correlation coefficients for SBP and DBP were 0.8914 (95% confidence interval [CI], 0.8727-0.9102) and 0.8207 (95% CI, 0.7920-0.8494). The kappa values for the diagnosis of hypertension and Joint National Committee 7 BP classification were 0.6538 (0.5436-0.7641) and 0.5703 (0.5055-0.6351), respectively. The diagnostic sensitivity for hypertension was 59.0%.
Conclusions: Age, gender, arm circumference, and AR were the factors related to the differences. Despite small differences in the mean values, the agreement and reliability were not good enough to recommend the A&D UA-767PC for adoption in epidemiological surveys of hypertension prevalence.
Keywords: blood pressure; diagnostic errors; epidemiological factors; hypertension; oscillometry..