RJUA-MedDQA: A Multimodal Benchmark for Medical Document Question Answering and Clinical Reasoning

TL;DR

RJUA-MedDQA introduces a challenging multimodal benchmark for medical document question answering and clinical reasoning, emphasizing complex interpretation, numerical reasoning, and clinical inference, supported by an efficient annotation method and extensive evaluations.

Contribution

The paper presents a new comprehensive benchmark for medical document understanding, along with the ESRA annotation method that improves efficiency and accuracy, and evaluates current LMMs' capabilities and limitations.

Findings

Existing LMMs have limited overall performance.

LMMs are more robust to low-quality images than LLMs.

Reasoning across text and images remains challenging.

Abstract

Recent advancements in Large Language Models (LLMs) and Large Multi-modal Models (LMMs) have shown potential in various medical applications, such as Intelligent Medical Diagnosis. Although impressive results have been achieved, we find that existing benchmarks do not reflect the complexity of real medical reports and specialized in-depth reasoning capabilities. In this work, we introduced RJUA-MedDQA, a comprehensive benchmark in the field of medical specialization, which poses several challenges: comprehensively interpreting imgage content across diverse challenging layouts, possessing numerical reasoning ability to identify abnormal indicators and demonstrating clinical reasoning ability to provide statements of disease diagnosis, status and advice based on medical contexts. We carefully design the data generation pipeline and proposed the Efficient Structural Restoration Annotation (ESRA) Method, aimed at restoring textual and tabular content in medical report images. This method substantially enhances annotation efficiency, doubling the productivity of each annotator, and yields a 26.8% improvement in accuracy. We conduct extensive evaluations, including few-shot assessments of 5 LMMs which are capable of solving Chinese medical QA tasks. To further investigate the limitations and potential of current LMMs, we conduct comparative experiments on a set of strong LLMs by using image-text generated by ESRA method. We report the performance of baselines and offer several observations: (1) The overall performance of existing LMMs is still limited; however LMMs more robust to low-quality and diverse-structured images compared to LLMs. (3) Reasoning across context and image content present significant challenges. We hope this benchmark helps the community make progress on these challenging tasks in multi-modal medical document understanding and facilitate its application in healthcare.