When Cases Get Rare: A Retrieval Benchmark for Off-Guideline Clinical Question Answering

TL;DR

This paper introduces OGCaReBench, a retrieval-based benchmark for evaluating medical LLMs on rare, real-world clinical questions requiring open-ended reasoning beyond guidelines.

Contribution

It presents a new benchmark dataset for assessing LLMs' ability to answer complex, rare clinical questions with evidence-grounded reasoning.

Findings

GPT-5.2 correctly answers 56% of questions without retrieval.

Specialized models reach only 42% accuracy.

Retrieval augmentation improves performance to 82%.

Abstract

Across medical specialties, clinical practice is anchored in evidence-based guidelines that codify best studied diagnostic and treatment pathways. These pathways routinely fall short for the long tail of real-world care not covered by guidelines. Most medical large language models (LLMs), however, are trained to encode common, guideline-focused medical knowledge in their parameters. Current evaluations test models primarily on recalling and reasoning with this memorized content, often in multiple-choice settings. Given the fundamental importance of evidence-based reasoning in medicine, it is neither feasible nor reliable to depend on memorization in practice. To address this gap, we introduce OGCaReBench, a free-form retrieval-focused benchmark aimed at evaluating LLMs at answering clinical questions that require going beyond typical guidelines. Extracted from published medical case reports and validated by medical experts, OGCaReBench contains long-form clinical questions requiring free-text answers, providing a systematic framework for assessing open-ended medical reasoning in rare, case-based scenarios. Our experiments reveal that even the best-performing baseline (GPT-5.2) correctly answers only 56% of our benchmark with specialized models only reaching 42%. Augmenting models with retrieved medical articles improves this performance to up to 82% (using GPT-5.2) highlighting the importance of evidence-grounding for real-world medical reasoning tasks. This work thus establishes a foundation for benchmarking and advancing both general-purpose and medical LLMs to produce reliable answers in challenging clinical contexts.