Pilot study of comparability of a smartphone blood pressure monitoring algorithm to conventional cuff-based blood pressure measurements

Abstract Background/Introduction There is a growing market for smartphone applications (apps), offering medical assessments such as blood pressure measurements (BPM). These apps have the potential to improve blood pressure (BP) control by making BPM broadly and easily accessible. Yet, to be suitable for clinical and diagnostic purposes, BPM measured with smartphone apps need to be comparable to conventional BPM. Purpose We sought to compare a novel photoplethysmographic BPM algorithm used in a smartphone app to conventional cuff-based BPM. Methods We included consecutive patients with an indication for ambulatory BPM. Office blood pressure measurements (OBPM) were taken with an oscillometric cuff-based device (Welch Allyn SureBP). The algorithm of the smartphone app detects the pulse wave in finger capillaries using the phone's camera and estimates BP based on the form of the pulse wave. Before estimating a BP value, the algorithm performs a quality assessment to automatically reject recordings with insufficient quality. On the first day (D1), we took 6 OBPM alternating with 5 smartphone BPM (TestBP). On the second day (D2), 4 OBPM alternating with 3 TestBP were measured. TestBP calibrated based on the first OBPM of D1. Each TestBP was then compared to its RefBP (defined as mean of the previous and following OBPM). Results 50 patients were included in the study, resulting in 50 TestBP values on D1 and 33 on D2. There was no difference at the 5% significance level between the TestBP and RefBP distributions on both days, and for both systolic and diastolic pressures. The mean ± standard deviation (SD) of the differences between TestBP and RefBP was 0.7±9.4 / 1.0±4.5 mmHg on D1 and 2.6±8.2 / 1.3±4.1 mmHg on D2 for systolic/diastolic values, respectively. The number of TestBP measurements within 5, 10 and 15 mmHg from RefBP are shown in Table 1. Bland-Altman plots depicting the agreement between TestBP and RefBP are shown in Figure 1. Conclusion This smartphone algorithm shows comparable values to oscillometric cuff-based especially diastolic values. Its differences between TestBP – RefBP have a good stability 1 day after calibration. Before clinical use, this algorithm needs to undergo formal validation against a reference BP method accepted by international standards (auscultatory or invasive methods). Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): Centre Suisse d'Electronique et de Microtechnique Table 1Figure 1