Prediction of health status in individuals with congestive heart failure using a home-based telecare system.

Telecare is increasingly used to remotely monitor long-term conditions such as congestive heart failure (CHF) and provide interventions based upon the data collected. In order to improve health care efficiency, there remains a need for decision support tools to automate this monitoring function and help guide interventions; this study sought to develop such a tool. Data was obtained from 45 elderly individuals with CHF who participated in a telecare trial for an average duration of 18 months. Physiological data along with subjective health perspectives and symptoms were reported. Clinicians responded to abnormalities in the data resulting in 154 key medical interventions/events. A multivariate logistic regression model was developed to predict these medical interventions/events. The developed model correctly predicted key medical events in 75% of cases with a specificity of 74% and an overall cross-validated accuracy of 74% [68-80%, 95% confidence interval]. Key predictors included: number of system alerts, self-rated mobility, self-rated health, and self-rated anxiety, strongly suggesting the utility of subjective measures in addition to physiological ones for prediction of health status. Overall this study demonstrates the potential of a multivariate decision-support model to enhance predictions of medical need in CHF patients using home-based telecare systems.