Pulsewatch Study Reveals 91% Accuracy in Smartwatch Heart Rhythm Detection
A recent study highlights the potential of smartwatch AI in detecting heart rhythms with a 91% accuracy rate, while also noting significant trade-offs.
The study analyzed over 116,000 recordings from 72 participants aged 50 and older.
Random Forest model achieved the highest accuracy at 91.22%, outperforming other deep-learning models.
Future research will focus on improving detection of less common rhythms and testing on diverse populations.
A groundbreaking study from the Pulsewatch clinical trial has demonstrated that smartwatch AI can achieve an impressive 91.22% accuracy in detecting heart rhythms. This research involved 72 participants aged 50 and above, all with a history of ischemic stroke, who wore Samsung smartwatches alongside electrocardiogram patches to gather data from 116,313 recordings, each lasting 30 seconds.
The study, led by researchers from Babeş-Bolyai University in Romania, aimed to address the challenges of detecting irregular heartbeats, which often occur without noticeable symptoms. By leveraging the continuous monitoring capabilities of smartwatches, the researchers sought to differentiate genuine rhythm changes from distortions caused by movement or poor skin contact. The findings were published in the journal Frontiers in Artificial Intelligence.
Participants' heart rhythms were categorized into three groups: normal sinus rhythm, atrial fibrillation, and premature contractions. The analysis revealed that normal rhythms accounted for approximately 68% of the recordings, while atrial fibrillation and premature contractions made up 21% and 11%, respectively. The Random Forest model, which utilized 500 decision trees and various measurements, emerged as the most effective approach, achieving the highest overall accuracy and precision in detecting atrial fibrillation.
The implications of this study are significant for public health, particularly for older adults at risk of stroke. Accurate detection of heart rhythm abnormalities can lead to timely medical interventions, potentially reducing the incidence of strokes. However, the study also highlighted the challenges of false detections and missed abnormalities, which could complicate clinical assessments and monitoring.
Looking ahead, researchers plan to expand their studies to include more diverse populations and improve the detection of less common heart rhythms. They also aim to test the algorithms on actual smartwatches to assess their performance in real-world settings, including battery usage and heat generation during continuous monitoring. These advancements could pave the way for smartwatches to become reliable tools for everyday heart health screening.
