Non-Invasive Sensor Predicts Heart-Failure Rehospitalization

By Reuters Staff

February 28, 2020

NEW YORK (Reuters Health) - A wearable sensor can predict rehospitalization in heart failure (HF) patients with an accuracy similar to that of implanted sensors, new research shows.

"The clinical efficacy and generalizability of this low-cost noninvasive approach to rehospitalization mitigation should be further tested," Dr. Josef Stehlik of VA Salt Lake City Health Care System and colleagues conclude in Circulation: Heart Failure.

Hospitalization accounts for 80% of HF costs, the authors note. Daily weight tracking and intrathoracic impedance monitoring have not been shown to reduce readmission in HF patients, they add, but more-invasive approaches have shown some promise in predicting exacerbations.

The Multi-Sensor Monitoring in Congestive Heart Failure (MUSIC) study, published in 2012, provided proof of concept for remote monitoring, but technological problems led to more than 40% of patients dropping out.

"More recent technological advances, including sensor miniaturization, improved battery life, and ubiquitous use of handheld devices, provide opportunities for more continuous telemonitoring," they add. "This is further amplified by advances in data science and artificial intelligence."

In the new study, patients were fitted with a wearable sensor (Vital Connect, San Jose, California) that streamed data continuously by Bluetooth to a smart phone. The data were then uploaded to a cloud analytics platform (PhysIQ, Chicago) after encryption.

Using a type of machine learning called similarity-based modeling (SBM), the platform used the patient's vital signs within the first 72 hours of discharge to create a baseline model, and then switched to surveillance mode. Vital signs were incorporated into a multivariate change index (MCI), with values less than 0 indicating improving health and values above 0 signifying worsening health.

MCI was determined every minute, and then converted to a daily average. Clinical alerts were triggered if the MCI exceeded a predetermined threshold.

Eighty-seven patients completed 30 days of monitoring, and 74 completed 90 days. During follow-up, 38 patients were hospitalized 49 times, including 27 HF hospitalizations.

The system issued clinical alerts 6.5 to 8.5 days before hospital admission, with 76.0%-87.5% sensitivity and 85% specificity.

"Since not all HF patients have an indication for a pacemaker or a defibrillator, and since implantation of a dedicated device presents procedural risks, noninvasive methods of monitoring may be more useful and cost-effective in patients temporarily at increased risk of HF-related hospitalization," Dr. Stehlik and colleagues note.

"These results provide a rationale for the next step, a prospective study, currently in planning, which will randomize patients to an active arm - remote monitoring with alerts communicated to the clinical team and clinicians following a standardized response algorithm, versus control - remote monitoring without alerts being generated," the authors add. "This study should provide important insights into the clinical efficacy of wearable analytics in improving HF outcomes."

The authors estimate that up to half of predicted hospitalizations due to heart failure could be prevented by prompt treatment intervention, so the platform could reduce rehospitalizations by about one-third.

The study did not have commercial funding. Several of the authors are employed by PhysIQ, and Dr. Stehlik consults for Abbott and Medtronic.

Dr. Stehlik was not available for an interview by press time.

SOURCE: Circulation: Heart Failure, online February 25, 2020.