Improving the assessment of discharge fitness in an acute medical unit using wearable sensor data through machine learning

Assessing the physiological discharge fitness of acute medical unit (AMU) patients is feasible using machine learning with electronic medical record (EMR) data, including intermittently measured vital signs. Wearable sensor data including continuous monitored vital signs may improve this assessment. Therefore, the aim was to investigate the value of sensor data in assessing the physiological discharge fitness of AMU patients. EMR and sensor data of AMU patients ( n =145) who participated in a randomized controlled trial, conducted at a large teaching hospital in the Netherlands from December 2021 to October 2023, were used. EMR features were extracted from patient characteristics, vital signs, nursing assessments and laboratory results. Statistical features and neural features, derived from convolutional neural networks, were extracted from sensor data. Logistic regression (LR), random forest (RF) and extreme gradient boosting (XGB) were used as algorithms. For each, three models were optimized and assessed using nested cross validation: based on EMR, EMR with sensor, or only sensor data. The mean area under the receiver operator characteristic curve (AUROC) with 95%-confidence interval (95% CI) was used as the primary metric. For LR, RF and XGB, the mean AUROC (shown with 95% CI), obtained by models based on EMR with sensor data (0.73 (0.70 – 0.77), 0.73 (0.69 – 0.77) and 0.69 (0.64 – 0.73), respectively) was slightly better than or as good as obtained by models only considering EMR data (0.69 (0.65 – 0.73), 0.70 (0.66 – 0.73) and 0.69 (0.66 – 0.71), respectively) and as obtained by models only considering sensor data (0.70 (0.66 – 0.75), 0.71 (0.67 – 0.75) and 0.67 (0.63 – 0.71), respectively). Wearable sensor data could slightly improve the assessment of physiological discharge fitness of AMU patients compared to only using EMR data.