Cluster-based Feature Importance Learning for Electronic Health Record Time-series

The recent availability of Electronic Health Records (EHR) has allowed for the development of algorithms predicting inpatient risk of deterioration and trajectory evolution. However, prediction of disease progression with EHR is challenging since these data are sparse, heterogeneous, multi-dimensional, and multi-modal time-series. As such, clustering is used to identify similar groups within the patient cohort to improve prediction. Current models have shown some success in obtaining cluster representation of patient trajectories, however, they i) fail to obtain clinical interpretability for each cluster, and ii) struggle to learn meaningful cluster numbers in the context of the imbalanced distribution of disease outcomes. We propose a supervised deep learning model to cluster EHR data based on the identification of clinically understandable phenotypes with regard to both outcome prediction and patient trajectory. We introduce novel loss functions to address the problems of class imbalance and cluster collapse, and furthermore propose a feature-time attention mechanism to identify cluster-based phenotype importance across time and feature dimensions. We tested our model in over 100,000 unique trajectories from hospitalised patients with Type-II respiratory failure to predict five different outcomes. Our model yielded added interpretability to cluster formation and outperformed benchmarks by at least 5% in mean AUROC.