Marcus Dörr1, Stefan Weber2, Ralf Birkemeyer3, Licia Leonardi4, Clemens Winterhalder4, Christina J Raichle5, Noé Brasier6, Thilo Burkard7, Jens Eckstein8. 1. Department of Internal Medicine, University Hospital Greifswald, Germany; DZHK (German Centre for Cardiovascular Research), partner site Greifswald, Germany. 2. Department of Internal Medicine, University Hospital Regensburg, Germany. 3. Department of Internal Medicine, University Ulm, Germany. 4. Department of Internal Medicine, University Hospital Basel, Switzerland. 5. Departement of Gastroenterology, University Hospital Basel, Switzerland. 6. CMIO Office, University Hospital Basel, Switzerland. 7. Medical Outpatient Department and Hypertension Clinic, ESH Hypertension Centre of Excellence, University Hospital Basel, Switzerland; Department of Cardiology, University Hospital Basel, Switzerland. Electronic address: thilo.burkard@usb.ch. 8. Department of Internal Medicine, University Hospital Basel, Switzerland; CMIO Office, University Hospital Basel, Switzerland.
Abstract
BACKGROUND: The possibility to use built-in smartphone-cameras for photoplethysmographic (PPG) recording of pulse waves lead to the release of numerous health apps, claiming to measure blood pressure (BP) based on PPG signals. Even though these apps are highly popular, not a single one is clinically validated. Aim of the current study was to test systolic BP (sBP) estimation by a promising new algorithm in a large clinical setting. METHODS: The study was designed based on the European Society of Hypertension International Protocol Revision 2010. Each individual received 7 sequential BP measurements, starting with the reference device - an automated oscillometric cuff device - followed by the PPG recording at the patients' index finger. RESULTS: A total 1,036 subjects were recruited of which 965 could be included for final analysis leading to 2,895 pairs of comparison. Mean (±SD) error between test and reference device was -0.41 (±16.52) mmHg. Only 38.1% of all 2,895 BP comparisons reached a delta within ±5 mmHg, while 29.3% reached a delta larger than 15 mmHg. Bland-Altman plot showed an overestimation of smartphone sBP in comparison to reference sBP in low range and an underestimation in high sBP range. CONCLUSIONS: According to the European Society of Hypertension International Protocol Revision 2010 specifications the algorithm failed validation criteria for sBP measurement and was not commercialized. These findings emphasize that health apps should be rigorously validated according to common guidelines before market release as under- and/or overestimation of BP is potentially exposing persons at health risks in short and long term. TRIAL REGISTRATION: ClinicalTrials.gov, number NCT02552030.
BACKGROUND: The possibility to use built-in smartphone-cameras for photoplethysmographic (PPG) recording of pulse waves lead to the release of numerous health apps, claiming to measure blood pressure (BP) based on PPG signals. Even though these apps are highly popular, not a single one is clinically validated. Aim of the current study was to test systolic BP (sBP) estimation by a promising new algorithm in a large clinical setting. METHODS: The study was designed based on the European Society of Hypertension International Protocol Revision 2010. Each individual received 7 sequential BP measurements, starting with the reference device - an automated oscillometric cuff device - followed by the PPG recording at the patients' index finger. RESULTS: A total 1,036 subjects were recruited of which 965 could be included for final analysis leading to 2,895 pairs of comparison. Mean (±SD) error between test and reference device was -0.41 (±16.52) mmHg. Only 38.1% of all 2,895 BP comparisons reached a delta within ±5 mmHg, while 29.3% reached a delta larger than 15 mmHg. Bland-Altman plot showed an overestimation of smartphone sBP in comparison to reference sBP in low range and an underestimation in high sBP range. CONCLUSIONS: According to the European Society of Hypertension International Protocol Revision 2010 specifications the algorithm failed validation criteria for sBP measurement and was not commercialized. These findings emphasize that health apps should be rigorously validated according to common guidelines before market release as under- and/or overestimation of BP is potentially exposing persons at health risks in short and long term. TRIAL REGISTRATION: ClinicalTrials.gov, number NCT02552030.
Authors: D Konstantinidis; P Iliakis; F Tatakis; K Thomopoulos; K Dimitriadis; D Tousoulis; K Tsioufis Journal: J Hum Hypertens Date: 2022-03-23 Impact factor: 3.012