Comparison of automated clinical and research blood pressure measurements: Implications for clinical practice and trial design

Abstract

Discrepancies between clinic and research blood pressure (BP) measurements lead to uncertainties in translating hypertension management guidelines into practice. We assessed the concordance between standardized automated clinic BP, from a primary care clinic, and research BP, from a randomized trial conducted at the same site. Mean single-visit clinic BP was higher by 4.4/3.8 mm Hg (P = 0.007/<0.001). Concordance in systolic BP (SBP) improved with closer proximity of measurements (difference = 2.5 mm Hg, P = 0.21 for visits within 7 days), but not averaging across multiple visits (difference =5.1(9.2) mm Hg; P < 0.001). This discrepancy was greater among female participants. Clinic-based difference in SBP between two visits was more variable than research-based change (SD = 19.6 vs 14.0; P = 0.002); a 2-arm trial using clinic measurements would need 95% more participants to achieve comparable power. Implementation of a bundled standardization intervention decreased discrepancies between clinic and research BP, compared to prior reports. However, clinic measurements remained higher and more variable, suggesting treatment to research-based targets may lead to overtreatment and using clinic BP approximately halves power in trials.