Can human clinical rationales improve the performance and explainability of clinical text classification models?

TL;DR

This study evaluates whether incorporating human clinical rationales into training improves the performance and explainability of clinical text classification models, finding limited performance gains but some explainability benefits.

Contribution

It provides a large-scale analysis of the effectiveness of human-derived rationales as additional supervision in clinical NLP models, comparing them to simply adding more reports.

Findings

Rationales improve performance in high-resource settings

Models trained on rationales outperform those trained on reports in explainability

Using rationales yields smaller performance gains than adding more reports

Abstract

AI-driven clinical text classification is vital for explainable automated retrieval of population-level health information. This work investigates whether human-based clinical rationales can serve as additional supervision to improve both performance and explainability of transformer-based models that automatically encode clinical documents. We analyzed 99,125 human-based clinical rationales that provide plausible explanations for primary cancer site diagnoses, using them as additional training samples alongside 128,649 electronic pathology reports to evaluate transformer-based models for extracting primary cancer sites. We also investigated sufficiency as a way to measure rationale quality for pre-selecting rationales. Our results showed that clinical rationales as additional training data can improve model performance in high-resource scenarios but produce inconsistent behavior when resources are limited. Using sufficiency as an automatic metric to preselect rationales also leads to inconsistent results. Importantly, models trained on rationales were consistently outperformed by models trained on additional reports instead. This suggests that clinical rationales don't consistently improve model performance and are outperformed by simply using more reports. Therefore, if the goal is optimizing accuracy, annotation efforts should focus on labeling more reports rather than creating rationales. However, if explainability is the priority, training models on rationale-supplemented data may help them better identify rationale-like features. We conclude that using clinical rationales as additional training data results in smaller performance improvements and only slightly better explainability (measured as average token-level rationale coverage) compared to training on additional reports.