SAVEHR: Self Attention Vector Representations for EHR based Personalized Chronic Disease Onset Prediction and Interpretability

TL;DR

SAVEHR is a self-attention model for EHR data that predicts chronic disease onset with high accuracy, generalizes well across populations, and offers interpretability, advancing personalized healthcare predictions.

Contribution

The paper introduces SAVEHR, a novel self-attention architecture for EHR data that outperforms baselines in multi-condition prediction, generalizes across populations, and enhances interpretability.

Findings

Achieves >0.51 AUC-PR and >0.87 AUC-ROC in predicting four conditions 15 months early.

Performs better than ten baseline models across all evaluation axes.

Transfers effectively to new populations with high performance.

Abstract

Chronic disease progression is emerging as an important area of investment for healthcare providers. As the quantity and richness of available clinical data continue to increase along with advances in machine learning, there is great potential to advance our approaches to caring for patients. An ideal approach to this problem should generate good performance on at least three axes namely, a) perform across many clinical conditions without requiring deep clinical expertise or extensive data scientist effort, b) generalization across populations, and c) be explainable (model interpretability). We present SAVEHR, a self-attention based architecture on heterogeneous structured EHR data that achieves $>$ 0.51 AUC-PR and $>$ 0.87 AUC-ROC gains on predicting the onset of four clinical conditions (CHF, Kidney Failure, Diabetes and COPD) 15-months in advance, and transfers with high performance onto a new population. We demonstrate that SAVEHR model performs superior to ten baselines on all three axes stated formerly.