North America

Health eHeart Study's smartphone measurements offer population-level heart rate insights

After validating smartphone photoplethysmography against in-clinic ECG, UCSF researchers contextualized heart rate data from nearly 70,000 study participants.
By Dave Muoio
01:29 pm

As part of the extensive Health eHeart Study, University of California, San Francisco researchers found smartphone camera-based heart rate measurements often employed by apps to strongly correlate with in-clinic results collected via ECG.

From there, the researchers employed their app-based approach to record the heart rates nearly 70,000 individuals worldwide, and paired these with health and demographics data provided by the study participants. Doing so, they wrote in npj Digital Medicine, has yielded “the largest real-world norms for remotely obtained, real-world [heart rate] according to various strata [that] may help physicians interpret and engage with patients presenting such data.”


To start, the researchers validated their smartphone photoplethysmography approach in 50 participants seen at the UCSF general cardiology clinic. The heart rate values recorded using a smartphone and an ECG landed at 0.9 overall agreement (with all disagreements occurring among participants with irregular heart rhythms), and a median difference of 2.7 beats per minute (BPM).

From there, the team remotely collected 3,144,332 smartphone-derived heart rate data readings (33,344 of which were later excluded) from 66,788 study participants. Nearly half of these participants were considered to be healthy, with the most prevalent reported medical conditions including hypertension, hypercholesterolemia and arrhythmia.

When combined with other information supplied by the participants, the team recorded a mean real-world heart rate of 79.1 BPM. These rates greatly varied over the course of the day, the researchers wrote, with values at their lowest during the early morning and their highest throughout the day. Number of reported medical conditions, higher BMI, female gender and Hispanic ethnicity were each associated with higher blood pressure readings, while older participants trended lower.


Participants included in the validation exercise received simultaneous smartphone and ECG testing over a 10-second period. The former was measured using Azumio’s Instant Heart Rate app, which conducts users to place their finger over the smartphone’s camera.

The population study consisted of data collected from April 2014 to April 2018 through the Health eHeart Study, a global, longitudinal cohort of English-speaking adults with an email address. These participants submitted their health, physical activity and demographic information through online surveys, and recorded at least one smartphone heart rate measurement through the connected their Azumio account.


The Health eHeart has been underway for some time and has so far yielded a handful of findings. These have included validation for atrial fibrillation detection using Apple Watches paired with Cardiogram’s mobile app, and evidence that smartphone geofencing could be used to detect hospitalizations.

That being said, today’s publication is just the latest demonstration of mobile technology assisting with remote investigations of population health. In March, for instance, Stanford Medicine researches revealed major findings from the Watch-powered Apple Heart Study, while Fitbit’s devices are actively assisting the National Institutes of Health’s massive All of Us Research Program.


“As the use of smartphone sensors and wearable devices provides data on cardiovascular parameters such as [heart rate], physicians are increasingly expected to help patients interpret the results of these readings; however, existing norms derived from controlled, clinical settings may not reflect the range of [heart rate] values occurring in real-world conditions,” the researchers wrote. “In this study, we have shown that [heart rate photoplethysmography] measurements are valid, and our nomograms of [these] measurements obtained by patients remotely can now be interpreted by physicians, across a wide variety of patient phenotypes. These data can inform patients about physical fitness and could help providers offer counseling on lifestyle changes or provide overall encouragement and support based on these real-world [heart rate] norms.”

More regional news


The latest news in digital health delivered daily to your inbox.

Thank you for subscribing!
Error! Something went wrong!