MedEval: A Multi-Level, Multi-Task, and Multi-Domain Medical Benchmark for Language Model Evaluation

Abstract: Curated datasets for healthcare are often limited due to the need of human annotations from experts. In this paper, we present MedEval, a multi-level, multi-task, and multi-domain medical benchmark to facilitate the development of LLMs for healthcare. MedEval is comprehensive and consists of data from several healthcare systems and spans 35 human body regions from 8 examination modalities. With 22,779 collected sentences and 21,228 reports, we provide expert annotations at multiple levels, offering a granular potential usage of the data and supporting a wide range of tasks. Moreover, we systematically evaluated 10 generic and domain-specific LLMs under zero-shot and finetuning settings, from domain-adapted baselines in healthcare to general-purposed state-of-the-art LLMs (e.g., ChatGPT). Our evaluations reveal varying effectiveness of the two categories of LLMs across different tasks, from which we notice the importance of instruction tuning for few-shot usage of LLMs. Our investigation paves the way toward benchmarking LLMs for healthcare and provides valuable insights into the strengths and limitations of adopting LLMs in medical domains, informing their practical applications and future advancements.