In a recent article published in the journal Nature Medicine, researchers presented findings from the RAte control Therapy Evaluation in permanent Atrial Fibrillation (RATE-AF) trial, which aimed to assess the effectiveness of consumer wearable devices in monitoring heart rate control in older patients diagnosed with permanent atrial fibrillation and heart failure. The study specifically compared the impacts of digoxin and beta-blockers on heart rate, leveraging data collected through wearable technology over a 20-week period.
F1 scores combining precision and recall of each model are presented along with 95% CI for the prediction of NYHA functional class at the end of follow-up (mean 5 months); an F1 score of 0.35 (dashed line) is equivalent to chance. Derived from wearable sensor data from n = 41 individual patients. Image Credit: https://www.nature.com/articles/s41591-024-03094-4
Background
Atrial fibrillation is a common arrhythmia associated with increased morbidity and mortality, particularly in older adults. Effective heart rate control is crucial for managing symptoms and improving the quality of life in these patients. Traditional methods of monitoring heart rate often rely on periodic clinical assessments, which may not capture real-time fluctuations. The integration of wearable devices offers a novel approach to continuous monitoring, potentially enhancing patient management and treatment outcomes.
The Current Study
The RATE-AF trial was designed as a randomized controlled trial to evaluate the effectiveness of consumer-grade wearable devices in monitoring heart rate control among older patients with permanent atrial fibrillation and heart failure. The study enrolled a total of 53 participants, aged 65 years and older, with a diagnosis of permanent atrial fibrillation and heart failure.
Participants were randomly assigned to receive either digoxin or beta-blockers using a computer-generated randomization sequence. The dosage of digoxin was adjusted based on renal function and clinical response, while beta-blocker therapy was tailored to individual patient needs, with adjustments made as necessary throughout the study period.
Participants were equipped with a wrist-worn consumer-grade wearable device capable of continuously measuring heart rate and physical activity. The device was linked to a smartphone application that facilitated real-time data transmission to a secure cloud-based server. Data collection occurred over a 20-week period, during which the wearable devices recorded heart rate intervals (n = 143,379,796) and physical activity levels (n = 23,704,307) at regular intervals throughout the day.
Data preprocessing involved cleaning the raw data to address missing values and outliers. A convolutional neural network (CNN) was employed to analyze the wearable data, accounting for the inherent noise and variability associated with consumer-grade devices.
Statistical analyses were performed using appropriate software, with a focus on intention-to-treat principles. The primary outcome was the difference in heart rate between the two groups, analyzed using generalized linear models to account for repeated measures and potential confounding factors such as age, gender, and body mass index.
Receiver operating characteristic (ROC) curves were generated to evaluate the predictive performance of the wearable data compared to traditional clinical measures, such as electrocardiographic heart rate and the 6-minute walk test. The F1 score was calculated to assess the balance between precision and recall in the model's predictions. A p-value of less than 0.05 was considered statistically significant for all analyses. Confidence intervals (CIs) were calculated to measure the precision of the estimates.
Results and Discussion
During the 20-week monitoring period, a total of 143,379,796 heart rate intervals and 23,704,307 physical activity intervals were collected from the wearable devices. The analysis revealed that the mean heart rate for participants on digoxin was 72.4 beats per minute (bpm) (95 % CI 70.1 to 74.7), while those on beta-blockers had a mean heart rate of 71.2 bpm (95 % CI 68.9 to 73.5). The heart rate was found to be similar based on the regression analysis between the two groups, with a regression coefficient of 1.22 (95 % CI -2.82 to 5.27; P = 0.55). After adjusting for physical activity levels, the difference remained non-significant (adjusted P = 0.75).
The F1 score, which combines precision and recall, was calculated to evaluate the predictive performance of the wearable data. The F1 score obtained by the model was 0.56 (95 % CI 0.41 to 0.70) for predicting New York Heart Association (NYHA) functional class, which was comparable to that of 0.55 (95 % CI 0.41 to 0.68) obtained from traditional clinical measures such as electrocardiographic heart rate and the 6-minute walk test. The area under the receiver operating characteristic curve (AUC) for the wearable data was 0.77, indicating good predictive capability.
Conclusion
The RATE-AF trial underscores the feasibility of utilizing consumer wearable devices for real-time heart rate monitoring in older patients with atrial fibrillation and heart failure. Although the study did not reveal significant differences between the two pharmacological treatments tested, it paves the way for future research into integrating wearable technology into routine clinical care. The results highlight a shift towards more personalized and continuous patient monitoring, which has the potential to enhance management strategies and improve outcomes in cardiovascular care.
Journal Reference
Gill S.K., Barsky A., et al. (2024). Consumer wearable devices for evaluation of heart rate control using digoxin versus beta-blockers: the RATE-AF randomized trial. Nature Medicine. DOI: 10.1038/s41591-024-03094-4, https://www.nature.com/articles/s41591-024-03094-4