Two-Pronged Human Evaluation of ChatGPT Self-Correction in Radiology Report Simplification

TL;DR

This paper investigates the use of large language models with self-correction prompts to generate patient-friendly radiology report simplifications, validated through a novel evaluation involving radiologists and laypeople.

Contribution

It introduces a new evaluation protocol and demonstrates the effectiveness of self-correction prompting in improving report simplification quality.

Findings

Self-correction prompting enhances simplification quality

Radiologists verify factual correctness effectively

Laypeople find simplified reports more understandable

Abstract

Radiology reports are highly technical documents aimed primarily at doctor-doctor communication. There has been an increasing interest in sharing those reports with patients, necessitating providing them patient-friendly simplifications of the original reports. This study explores the suitability of large language models in automatically generating those simplifications. We examine the usefulness of chain-of-thought and self-correction prompting mechanisms in this domain. We also propose a new evaluation protocol that employs radiologists and laypeople, where radiologists verify the factual correctness of simplifications, and laypeople assess simplicity and comprehension. Our experimental results demonstrate the effectiveness of self-correction prompting in producing high-quality simplifications. Our findings illuminate the preferences of radiologists and laypeople regarding text simplification, informing future research on this topic.