Exploratory study evaluates a noninvasive multiparameter monitoring device for detecting decompensated heart failure, revealing high specificity but low sensitivity.
Key Points
The optimal model achieved 97.2% specificity but only 5.3% sensitivity for detecting decompensated heart failure.
Seventeen patients were monitored using a wrist-worn device, measuring heart rate and other parameters for three months.
Data quality was variable, as only 72% of energy expenditure and 79% of activity counts were deemed high quality.
Future studies must address class imbalance and enhance data fidelity to improve predictive performance.