Multi-Scale Entropy of Nocturnal Breathing from Wearable Devices to Assess Post-Discharge Outcome Differences in ADHF Patients

The nighttime breathing of acute decompensated heart failure (ADHF) patients may be closely related to their health status and disease progression. However, there is currently no consistent and widely accepted approach for processing, analyzing, and extracting the potential value of respiratory signals obtained from wearable devices. In this study, a reference approach was proposed for ADHF patients, involving the processing and analysis of respiratory signals obtained from a wearable continuous physiological parameter monitoring system, as well as clinical application validation. The study included 62 ADHF patients, among whom 17 experienced readmission or death within six months after discharge, as well as 56 healthy subjects. The nighttime respiratory signals of ADHF patients and healthy subjects were selected and processed by removing motion artifacts, filtering, and quality assessment. High-quality respiratory segments were used to calculate parameters such as respiratory cycle, respiratory amplitude, chest-abdominal coordination, and respiratory entropy, validating the feasibility of quantifying respiratory features using wearable devices. A subgroup analysis was conducted to investigate the occurrence of adverse events after discharge in ADHF patients. Compared to the non-adverse event group, the group that experienced adverse events exhibited higher multiscale sample entropy of respiratory rate. This study has proposed a comprehensive approach for processing and quantifying respiratory signals obtained from wearable devices, improving the signal processing methods in the field of medical signal analysis. The study has also explored the potential clinical application value of these respiratory quantification measures, particularly in the context of home monitoring. This approach holds promise for widespread utilization and implementation in home-based monitoring settings.