Real-world morning home blood pressure in a personal health record application and the risk of cardiovascular events

Abstract We aimed to assess whether home blood pressure (BP) self-recorded through a personal health record (PHR) application in a real-world setting was associated with cardiovascular disease (CVD) events among individuals not receiving antihypertensive treatment. Using an insurer-based database that links annual health checkup records and home BP self-recorded in the “kencom” PHR application, 6654 participants (mean age: 51.8 ± 9.1 years, males: 68.0%) not receiving antihypertensive treatment were selected. Morning home BP was calculated as the average of morning values over the first 28 days. Office BP was derived from annual health checkup data. The primary outcome was total CVD identified from hospitalization claims. Cox models were used to assess hazard ratios (HRs). The average of home systolic/diastolic BP was 124.5 ± 13.7/80.9 ± 10.9 mmHg, and 80.6% entered their home BP manually. The correlation between home and office BP was r = 0.56 for systolic and r = 0.61 for diastolic BP. During a mean follow-up of 2.79 ± 1.76 years (18,592 person-years), 447 total CVD events occurred. The HRs for total CVD per 1-SD higher in BP were 1.31 for home systolic, 1.12 for office systolic, 1.32 for home diastolic, and 1.21 for office diastolic BP. Results were consistent for stroke and heart failure. When home and office BP were entered simultaneously into the same model, only home systolic/ diastolic BP was independently associated with total CVD risk. Morning home BP self-recorded by users in a PHR application under real-world conditions was more strongly associated with CVD than office BP, supporting the potential value of PHR-based home BP.