Wearable Patch Predicts Worsening Heart Failure

Josef Stehlik, MD

New research from the US Department of Veterans Affairs (VA) has found a removable, adhesive patch could help clinicians and caregivers predict worsening heart failure (HF) and the need for hospitalization several days before hospitalization occurred.

Results of the LINK-HF study, which examined the use of a disposable multisensor patch and a machine-learning algorithm to predict HF exacerbation, revealed the platform was able to detect precursors for exacerbation with more than 80% accuracy nearly a week before hospitalization.

“With the use of remote data from the sensor and through data analysis by machine learning, we have shown that we can predict the future. Next, we will look at whether we can change the future,” said lead investigator Josef Stehlik, MD, MPH, medical director of the Heart Transplant Program and co-chief of the Advance Heart Failure Program at the University of Utah Hospital and the Salt Lake City VA Medical Center, in a statement.

With more than 6 million Americans suffering from HF and 80% of the $30.7 billion spent on healthcare-related costs linked to hospitalizations, Stehlik and colleagues sought to determine if a noninvasive approach would allow for the same predictive accuracy seen with implantable cardiac sensors. The Multisensor Non-invasive Remote Monitoring for Prediction of Heart Failure Exacerbation (LINK-HF) study, was designed as an observational study with the aim of determining the accuracy of the machine learning algorithm in predicting HF readmission in veterans from 4 VA medical centers across the country.

Inclusion criteria for LINK-HF included being 18 years of age or older, a history of HF and NYHA functional class 2 or greater symptoms, and having been hospitalized for acute HF exacerbation. Patients with both HF with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF) were included in the study.

Wearable sensors, which were provided by Vital Connect in San Jose, CA—were worn by all subjects 24 hours a day for a minimum of 30 days and up to 90 days following discharge. The sensors provided were comprised of a disposable sensor patch with a disposable battery, and a reusable sensor electronics module. Measurements taken by the sensor included continuous ECG waveform, heart rate, heart rate variability, arrhythmia burden, respiratory rate, gross activity, sleep, body tilt, and body posture.

The sensor patch used was paired via Bluetooth with an android phone and data was continuously streamed from sensors to the phone. Sensors were equipped with storage for 10 hours in the event the subject was out of Bluetooth range.

A total of 100 participants were enrolled in the study between August 2015 and December 2016. The mean age of the population was 68.4 (10.2) years, 98% were male, 74% had HFrEF, and 26% had HFpEF. Investigators noted compliance was high, with 87 of the 100 individuals completing 30 days of monitoring and 74 of those 87 completed 90 days of monitoring—13 subjects died or became ineligible between days 30 and 90.

The success of the algorithm in predicting HF exacerbation was defined in 2 different ways. The first being a fixed positive window of 10 days before hospitalization and an event-specific positive window.

During the course of the study, a total of 49 hospitalizations occurred in 38 subjects and the median time from discharge to rehospitalization was 50.5 days. Of the 49 hospitalizations, 27 were HF hospitalizations and 40 were unplanned nontrauma hospitalizations.

A total of 12 subjects died during the study, including 6 that were identified as sudden cardiac death. None of the sudden cardiac deaths were preceded by an alert in the 10-day positive window.

Results of the investigators' analyses indicated the predictive algorithm was able to detect the risk of hospitalization for worsening of HF with 76% to 87.5% sensitivity and 85% specificity. Depending on the positive window method used and the type of hospitalization, the median time between the alert and admission ranged from between 6.5 and 8.5 days.

“If we can identify patients before heart failure worsens and if doctors have the opportunity to change therapy based on this novel prediction, we could avoid or reduce hospitalizations, improve patients’ lives and greatly reduce health care costs,” Stehlik added. “With the evolution of technology and with artificial intelligence statistical methods, we have new tools to make this happen.”

Stehlik and LINK-HF investigators noted multiple limitations within their study. These limitations included an overwhelming majority of males, a majority of patients had a single type of HF, and the need for additional research to determine if treatment changes based on alerts could improve outcomes.

This study, “Continuous Wearable Monitoring Analytics Predict Heart Failure Hospitalization,” was published in Circulation: Heart Failure.