Event detection for post lung transplant based on home monitoring of spirometry and symptoms

Abstract

The goal of this dissertation research was to develop, implement, and test an automated decision system to provide early detection of actual acute bronchopulmonary events in a population of lung transplant recipients following a home monitoring protocol. Decision rules were developed using wavelet analysis of spirometry and symptom signal data collected daily at home by the lung transplant recipients, and transmitted weekly to our study data center. Rules were developed based on a learning set of patient home data, and validated with an independent set of patients. Using either FEV1 or symptom-based home data monitoring, the detection algorithm can capture the majority of events (sensitivity > 80%) at an acceptable level of false alarms. Detection occurs 6.6 to 10.8 days earlier than the corresponding events recorded in the patient's clinical records. Combining rules using the Dempster-Shafer theory of evidence incrementally improves performance over a single variable. This framework can be readily implemented as an automatic event detection tool to aid medical discovery and diagnosis of acute pulmonary events.