Early Warning of Acute Decompensation in Heart Failure Patients Using a Noncontact Measure of Stability Index.

Abstract

Hospitalization due to heart failure (HF) continues to be a major clinical and economic challenge. To reduce hospitalization, this paper proposes a novel home telemonitoring system for an early warning of acute decompensation in patients with chronic stable New York Heart Association class II-IV HF. It is based on deriving nocturnal respiratory related time series using a noncontact radio-wave bio-motion sensor. The system generates an alert when there is a change in the underlying probability distribution of the time series which can be regarded as a surrogate marker of patient stability. The system's performance is evaluated using dual-site longitudinal data collected from 104 HF patients over 12-24 months. The system reported an average sensitivity of 0.62 to detect a change during an episode of acute decompensation and an average specificity of 0.67 on the blind validation set, when the frequency of alert was four weeks. The system also performed well to predict acute decompensation with an average sensitivity of 0.55 and specificity of 0.73 on the validation set, where an event window was defined as three weeks preceding an event. These results demonstrate that the design and implementation of such a system is a positive step toward developing noncontact systems capable of preventing acute decompensation, reducing readmissions to hospital and ensuring better quality of life for HF patients.