Smartphone Blood Pressure Measurement to Screen for Hypertension

This is a single-center observational pilot study. Data will be collected prospectively on 100 participants (maximum of 120) with, ideally, 50 presumed healthy volunteers and 50 participants with a suspected diagnosis of hypertension. Presumably healthy volunteers will be recruited via flyers and participants presumed to be hypertensive will be recruited by the hypertension unit of the CHUV, where they are referred for the search for a diagnosis of hypertension.

To comply with the International Hypertension Societies guidelines for LMWH, participants will measure their blood pressure for 7 consecutive days, in the morning (before taking medications and food) and in the evening. (before dinner). At each measurement time point, participants will acquire three separate measurements, one minute apart. Participants will record signals by placing a fingertip of the right hand over a smartphone camera, while simultaneously acquiring a baseline blood pressure measurement using a validated cuff device (WatchBP O3 Ambulatory, Microlife, Widnau, Switzerland) on the left arm. Please note that the smartphone and armband will be provided by the site investigator. The data recorded using the smartphone will be processed and directly compared to the values obtained using the inflatable cuff device.

In accordance with international guidelines for LMWH (except those of the American Heart Association), the threshold for hypertension is set at systolic blood pressure values ≥ 135 mmHg and/or diastolic blood pressure values ≥ 85 mmHg.

The aim of this study is to evaluate the feasibility of classifying people whose blood pressure is in the hypertensive range or not, using a method based on machine learning. To do this, optical data will be recorded at the fingertip from a smartphone camera following the HBPM measurement protocol. The optical signals will then be processed by a machine learning method to obtain the classification. In other words, investigators will attempt to evaluate the reliability of the machine learning method compared to the automatic oscillometric brachial cuff and therefore evaluate the feasibility of using the smartphone as a potential future tool for diagnosis, screening and monitoring hypertension.

The methods of screening, diagnosis and monitoring of hypertensive pathologies are precise but require cumbersome equipment and are sometimes poorly available. Investigators hypothesize that smartphones can be used because of their availability and their capacity to generate quality optical signals, as a discriminating classifier in arterial hypertension thanks to the precision of a machine learning model hypothetically greater than a classification carried out by a random level.