The 450,000 strong Fitbit Heart Study has found that Fitbit’s algorithm is pretty good at spotting undiagnosed signs of atrial fibrillation (Afib). Although not perfect, the tech could find use in spotting first signs of the condition, thus lowering the cost of diagnosing afib on a wide scale.
We wrote in May 2020 about the research study. Its aim was to determine whether Fitbit devices can be used for detecting irregular heart rhythm suggestive of atrial fibrillation (AFib) via heart rate sensors and an algorithm. This is the most common type of heart arrhythmia. It effects about 8 million people in the US and more than 30 million people world-wide. To study managed to enroll 455,669 device owners with no prior diagnoses of afib. Around 13% of these were over 65, so an age group more prone to risks of irregular heartbeat.
Essential reading: Top fitness trackers and health gadgets
The research analyzed wearers’ heart rhythms using the photoplethysmography (PPG) sensors that can be found on most of Fitbit’s wearables. The company has already had a successful clinical trial of an ECG feature. This study was looking to do the same for its PPG afib algorithm. The ultimate aim is to get FDA approval for its devices to be used for self-monitoring of the condition.
Volunteers who opted into the study were privy to the latest Fitbit afib algorithm which looked for irregularities in the blood flow from their wrist. If something is amiss, the user would get an alert.
The algorithm continuously samples the pulse data in 5-minute blocks, which by design are overlapping by 50%,” said Harvard Medical School Associate Professor of Medicine (and study investigator) Steven Lubitz.
If 11 out of 11 consecutive blocks, or tachograms, are irregular, an irregular heart rhythm detection occurs. By design, that means that at a minimum, the algorithm requires at least 30 minutes of an irregular rhythm to detect atrial fibrillation. The algorithm only operates if the participant is inactive, which is judged by accelerometers on the device. The algorithm resets with a normal 5-minute tachogram period. Unanalyzable periods, perhaps due to activity, are skipped by the algorithm.
The user than had the option to connect to a doctor for a virtual appointment free of charge. What’s more, device owners with suspected afib were also eligible to receive a ECG patch via mail to confirm the reading.
Apple has similar functionality in its smartwatches. The difference is that Fitbit is only using the heart rate monitor to spot irregularities. On an Apple Watch, users can take a reading on demand by launching the ECG app and resting their finger against the digital crown. This creates a closed circuit between the digital crown and the sensor on the back. Built in electrodes will then work to detect electrical impulses from the heart.
Interestingly, the Cupertino outfit conducted its own Apple Heart Study which enrolled some 400,000 participants. The results showed that around 0.5% of the study participants received irregular heart rhythm notification over 117 days of monitoring, and 3% of those over 65.
Early identification of AFib, is key to reducing the risk of a life-threatening event like stroke. Fitbit’s heart rate monitors work 24/7 so are in an ideal position to passively monitor for the condition and alert if abnormalities are spotted. But this is only one part of the picture.
The company is looking to provide users with both “long and short-term AFib assessment options”. The PPG-based heart rhythm looks for irregular rhythm episodes with no symptoms, while an ECG feature can record an actual ECG trace for evaluation by users and doctors.
During the period of observation, about 1% of the study participants (4,827 people) received an Afib alert. Around 4% of those were in the 65+ age group. Only around a third of participants (1,162 of which 1,057 had data that could be analyzed) took up the offer for a tele-health consultation and to wear the ECG patch (ePatch, BioTel) for a week. About one in three of these received a confirmation that they did have Afib.
On the one hand, this shows that there were quite a few false positives. But the study also demonstrates that Fitbit’s heart rate algorithm, when combined with a wearable ECG monitor patch, delivers a 98% positive predictive value. Fitbits could therefore be used to help enable mass screening for possible undiagnosed cases of afib, i.e. to catch the first signs of the condition.
Comparing with the Apple Heart Study, the positive predictive values in the Fitbit study were shown to be higher. Half as many people were notified in the Apple study. The positive predictive values totaled 84% in Aple’s research compared to 98% in Fitbit’s. The one-third positive Afib confirmation was similar between the two studies. This would imply that Apple and Fitbit devices are broadly comparable in terms of performance.
Studies just using the ECG patch have shown that less than 5% of individuals have signs of afib. Fitbits could be used as a cost saving tool, narrowing the field of potential patients to be screened for the condition. This would in turn reduce the overall costs in diagnosing Afib in the general population. Particularly as the algorithm operates in the background, so doesn’t require any specific interaction from the user.
The findings were presented at the annual scientific sessions of the American Heart Association. The US Food and Drug Administration (FDA) is currently reviewing Fitbit’s algorithm for clearance and widespread use.
Like this article? Subscribe to our monthly newsletter and never miss out!