An AI can predict heart failure, atrial fibrillation, and death up to ten years in advance using sleep test data. Developed by Emory University and others, the model shows promising results.
AI predicts heart risks from sleep data
An artificial intelligence has been developed that can estimate the risk of heart failure, atrial fibrillation, and death up to ten years in advance by analyzing heart rhythm data recorded during a sleep test. The AI was developed by a team led by Emory University in Atlanta, in collaboration with the Georgia Institute of Technology, hospitals in Boston, and Stanford University. The findings were published on October 9 in the journal SLEEP.Sleep test data used for analysis
The study utilized data from sleep tests, where heart activity is recorded throughout the night. These recordings are typically not used for further analysis. The researchers adapted an existing AI program, originally trained to detect rhythm disorders using 86,000 electrocardiograms, to analyze the heart traces from 15,809 patients at the Massachusetts General Hospital in Boston. The model analyzed the heart signals in fifteen-second segments, cross-referencing them with sleep phases.AI tested on large patient groups
The AI was tested on 9,810 patients from Emory University Hospital and 12,576 patients from the Beth Israel Deaconess Medical Center. The results were validated using medical records that tracked the patients over seven to ten years. For each increase in the AI score, the risk of atrial fibrillation was found to increase by 2 to 2.7 times, the risk of heart failure increased by 69 to 133 percent, and the risk of death also showed a significant increase.Results show promise but need confirmation
The study's results are promising but require further confirmation before they can be applied in medical practice. Researchers emphasize the need for additional studies to validate the AI’s accuracy and ensure its reliability in clinical settings.
