Can Attention Be Used to Explain EHR-Based Mortality Prediction Tasks: A Case Study on Hemorrhagic Stroke

TL;DR

This paper introduces an attention-based transformer model for early hemorrhagic stroke mortality prediction that emphasizes interpretability and fidelity, aiming to improve upon traditional scoring systems.

Contribution

It presents a novel interpretable transformer model utilizing attention mechanisms to enhance explainability in EHR-based mortality prediction tasks.

Findings

The model achieves higher interpretability scores compared to traditional methods.

Attention-based explanations align well with clinical feature importance.

The approach improves trustworthiness of mortality predictions.

Abstract

Stroke is a significant cause of mortality and morbidity, necessitating early predictive strategies to minimize risks. Traditional methods for evaluating patients, such as Acute Physiology and Chronic Health Evaluation (APACHE II, IV) and Simplified Acute Physiology Score III (SAPS III), have limited accuracy and interpretability. This paper proposes a novel approach: an interpretable, attention-based transformer model for early stroke mortality prediction. This model seeks to address the limitations of previous predictive models, providing both interpretability (providing clear, understandable explanations of the model) and fidelity (giving a truthful explanation of the model's dynamics from input to output). Furthermore, the study explores and compares fidelity and interpretability scores using Shapley values and attention-based scores to improve model explainability. The research objectives include designing an interpretable attention-based transformer model, evaluating its performance compared to existing models, and providing feature importance derived from the model.