Med-CoReasoner: Reducing Language Disparities in Medical Reasoning via Language-Informed Co-Reasoning

TL;DR

This work tackles the persistent multilingual gap in medical reasoning by proposing Med-CoReasoner, a language-informed co-reasoning framework that runs parallel English and local-language reasoning, abstracts both into concept chains, and fuses them onto an English-centered scaffold augmented with local clinical knowledge. It ground’s reasoning with multilingual retrieval from MSD Manuals and AFRIDOC-MT, and leverages a threshold-based, concept-level fusion to preserve both logical structure and region-specific medical nuance. The authors introduce MultiMed-X, a seven-language benchmark with long-form QA and natural language inference to evaluate multilingual medical reasoning beyond MCQ tasks. Across heterogeneous backbones and tasks, Med-CoReasoner yields significant gains, especially for low-resource languages, and expert evaluations corroborate improved clinical soundness, localization, and safety, demonstrating a practical path toward equitable global medical AI deployment.

Abstract

While reasoning-enhanced large language models perform strongly on English medical tasks, a persistent multilingual gap remains, with substantially weaker reasoning in local languages, limiting equitable global medical deployment. To bridge this gap, we introduce Med-CoReasoner, a language-informed co-reasoning framework that elicits parallel English and local-language reasoning, abstracts them into structured concepts, and integrates local clinical knowledge into an English logical scaffold via concept-level alignment and retrieval. This design combines the structural robustness of English reasoning with the practice-grounded expertise encoded in local languages. To evaluate multilingual medical reasoning beyond multiple-choice settings, we construct MultiMed-X, a benchmark covering seven languages with expert-annotated long-form question answering and natural language inference tasks, comprising 350 instances per language. Experiments across three benchmarks show that Med-CoReasoner improves multilingual reasoning performance by an average of 5%, with particularly substantial gains in low-resource languages. Moreover, model distillation and expert evaluation analysis further confirm that Med-CoReasoner produces clinically sound and culturally grounded reasoning traces.

Figures (11)

• Figure 1: Performance gap between English-thinking and local-language-thinking settings under the same query: average scores of GPT-4o and DeepSeek-3.2 on MMLU-ProX-Health, with the largest degradation in Swahili.
• Figure 2: Illustration of the Med-CoReasoner framework. The system first translates user input into English, then conducts parallel reasoning in English and Italian via separate queries. Reasoning outputs are abstracted into concepts and fused into an English-anchored reasoning scaffold, where English provides a logical backbone and the local language supplies linguistically specific details. This concept-based scaffold is used to retrieve relevant knowledge and guide the generation of the final Italian reasoning output.
• Figure 3: Experimental results on MultiMed-X, where (#) denotes the ranking of our framework.
• Figure 4: Results on LFQA, judged by GPT-4o.
• Figure 5: Ablation results on selected languages across LFQA and MMLU-ProX.