AI-powered wearable sensor accurately predicts worsening heart failure ahead of health crises
An artificial intelligence (AI)-powered, wearable sensor has been demonstrated as effective at detecting abnormal cardiovascular changes, which could help doctors remotely detect worsening heart failure ahead of a crisis or hospitalization, as implantable devices.
A novel, non-invasive wearable sensor, powered by artificial intelligence (AI) technology, has been demonstrated as effective at detecting abnormal cardiovascular changes in individuals discharged from hospital with heart failure, which could enable the remote detection of worsening heart failure ahead of a crisis or rehospitalization, as implantable devices.
Approximately 6.2 million individuals are affected by heart failure in the USA; up to 30% of individuals discharged from hospital with a heart failure diagnosis will be readmitted within 90 days, presenting with symptoms including shortness of breath, fatigue and fluid buildup. Such hospitalization can have a drastic, detrimental impact on individual’s quality of life.
In this novel study, researchers from University of Utah Health and the Veteran Affairs Salt Lake City Health Care System (both UT, USA) followed 100 individuals, over the ages of 68 years, who had been hospitalized and then discharged with a heart failure diagnosis. Following discharge, study participants wore a non-invasive, wearable patch on their chests, 24 hours a day, for up to 3 months.
The wearable sensors continuously measured individuals’ heart rates, rhythms, respiratory rates, body posture and various other ‘normal’ activities of subjects, via constant electrocardiogram and motion monitoring. ‘Normal’ baseline measurements of each parameter were established through the application of AI to the data; deviations from baseline measurements were taken as indicators of worsening heart failure.
Investigators determined the accuracy of the system at predicting imminent health crisis and/or rehospitalization to be approximately 80%; predictions occurred a median 6.5 days ahead of rehospitalization.
Contributing study author Biykem Bozkurt (Baylor College of Medicine; TX, USA) explained: “Those individuals who have repeated hospitalizations for heart failure have significantly higher mortality. Even if patients survive, they have poor functional capacity, poor exercise tolerance and low quality of life after hospitalizations. This patch…could potentially help us prevent hospitalizations and decline in patient status.”
“This study shows that we can accurately predict the likelihood of hospitalization for heart failure deterioration well before doctors and patients know that something is wrong,” concluded Josef Stehlik (University of Utah Health), lead study author.
Study limitations include the predominantly male and small study population, who largely presented with reduced ejection fraction heart failure, as well as the short observation time and exclusion of five events, which were not preceded by sufficient data transfer from the study subjects, in the analysis. Going forward, the researchers are planning to conduct a larger, prospective, randomized clinical trial to both test the accuracy and usability of the technology in a larger patient population, and determine if early intervention, based on sensor alerts, could result in fewer heart failure-related rehospitalizations.
Stehlik J, Schmalfuss C, Bozkurt B et al. Continuous wearable monitoring analytics predict heart failure hospitalization. Circ. Heart Fail. doi:10.1161/CIRCHEARTFAILURE.119.006513 (2020) (Epub ahead of print); https://healthcare.utah.edu/publicaffairs/news/2020/02/heart-failure.php