Consumer wearable devices may offer a scalable method for the daily monitoring of heart failure (HF) symptoms and predicting exacerbations. A new study has detailed how a deep learning model using Apple Watch data can estimate cardiopulmonary fitness and provide early risk discrimination for unplanned healthcare events in patients with HF.¹
Methodology
The Ted Rogers Understanding Exacerbations of Heart Failure (TRUE-HF; NCT05008692) study was a prospective, observational cohort study designed to assess the ability of consumer wearable data to predict clinical status in patients with HF.²
The study enrolled 217 patients with HF who were monitored in a free-living environment for a median of 94.5 days using an Apple Watch. A deep learning model was developed to predict daily peak oxygen uptake (pVO2) using the wearable data. The primary endpoint was the correlation of this wearable-derived pVO2 with in-clinic cardiopulmonary exercise testing (CPET).
A secondary analysis evaluated the association between drops in wearable-derived pVO2 and unplanned healthcare utilisation, defined as hospital admissions, unscheduled clinic visits, or intravenous furosemide treatment. The findings were then externally validated in an independent cohort from the All of Us Research Program.
Results
The deep learning model was trained with data from 154 patients and validated on a held-out set of 63 patients. The wearable-derived daily pVO2 correlated strongly with CPET-measured pVO2 (Pearson’s correlation = 0.85).¹
Each 10% drop in the wearable-derived daily pVO2 was associated with a 3.62-fold increased hazard ratio (HR) for unplanned healthcare events (95% confidence interval [CI], 1.37–9.55; P < 0.01). These events occurred at a median of 7.4 days after the first 10% drop in pVO2 was detected.
In the external validation cohort, which used a cross-platform model with reduced-sensor capacity, drops in wearable-derived daily pVO2 were also associated with an increased risk of unplanned healthcare utilisation (HR 1.32, 95% CI 1.03–1.69; P = 0.03). In this cohort, events occurred a median of 21 days after the first 10% drop was detected.
Interpretation
According to the study authors, “These results indicate that wearable-derived daily pVO2 provides earlier and improved risk discrimination compared with existing wearable fitness estimates and established clinical markers and offers a scalable and generalizable approach for longitudinal HF research and monitoring.”¹ The findings suggest that continuous remote monitoring may identify at-risk patients earlier than traditional static clinical assessments.
Next Steps
The study highlights the need for further research with larger validation sets to fully assess the independent prognostic value of this monitoring approach, particularly when adjusted for potential confounders.
This study was funded by the Ted Rogers Centre for Heart Research and the Natural Sciences and Engineering Research Council of Canada.
References
1. Gao Y, Moayedi Y, Foroutan F, et al. Remote monitoring of heart failure exacerbations using a smartwatch. Nat Med2026;32:924–933. https://doi.org/10.1038/s41591-026-04247-3
2. Moayedi Y, Foroutan F, Verma B, et al. Developments in digital wearable in heart failure and the rationale for the design of TRUE-HF (Ted Rogers Understanding of Exacerbations in Heart Failure) Apple CPET Study. Circ Heart Fail2025;18:e012204. https://doi.org/10.1161/CIRCHEARTFAILURE.124.012204
Disclaimer
The information presented in this article is for educational purposes only. Any quotes included reflect the opinions of the individual quoted, and do not necessarily reflect the views of the publisher. The publisher does not guarantee the accuracy or completeness of the content and accepts no responsibility for any errors, or any consequences arising from its use.