Correctness Coverage Evaluation for Medical Multiple-Choice Question Answering Based on the Enhanced Conformal Prediction Framework

Large Language Models (LLMs) hold the potential for clinical task-shifting by processing unstructured clinical text, enabling tasks such as clinical concept extraction and medical question answering from electronic health records. If implemented reliably, such approaches may benefit over-burdened healthcare systems, particularly in resource-limited settings and for traditionally overlooked populations, provided that local fine-tuning is supported by appropriate clinical and technical expertise. However, this powerful technology remains largely understudied in real-world contexts, particularly in the Global South. This study aims to assess whether openly available LLMs can be used reliably for processing medical notes in real-world settings in South Asia. We used publicly available LLMs to parse de-identified clinical notes from a large electronic health records (EHR) database in Pakistan, containing hospital records for 8.2 million patients. ChatGPT (GPT-3.5) as a general-purpose LLM, and GatorTron (base), BioMegatron, BioBert and ClinicalBERT as medical LLMs were evaluated when applied to these data, after fine-tuning them with (a) publicly available clinical datasets namely Informatics for Integrating Biology & the Bedside (I2B2) and National NLP Clinical Challenges (N2C2) for medical concept extraction (MCE) and emrQA for medical question answering (MQA), and (b) the local Pakistani de-identified EHR dataset, which includes inpatient Discharge Summaries (DS) and Subjective, Objective, Assessment, and Plan (SOAP) notes, as detailed in this paper. MCE models were applied to these clinical notes using both 3-label and 9-label formats, while MQA models were applied to medical questions. Internal and external validation performance was measured for (a) and (b) using F1 score, precision, recall, and accuracy for MCE and BLEU and ROUGE-L, which measure lexical and sequence similarity, for MQA. When clinical LLMs were not fine-tuned on the local EHR dataset, their performance during external validation on local data was notably poorer compared to internal validation on the dataset used for fine-tuning, with reductions of at least 15% in F1 scores for MCE and 35% in ROUGE-L and BLEU scores for MQA tasks. This suggests potential bias and highlights the inability of the medical LLMs to reliably handle the data distribution of the local population without further fine-tuning and adaptation. This trend persisted across two distinct natural language processing tasks: concept extraction and question answering, spanning a spectrum of task complexities. However, fine-tuning the LLMs with local EHR data significantly improved model performance across both tasks, yielding a 7.5% to 15% increase in the F1 score for MCE and a 27% to 53% increase in ROUGE-L and BLEU scores for MQA. Notably, ChatGPT, as a general-purpose LLM, stood out as an exception, demonstrating superior performance across all measured metrics on the local dataset compared to the publicly available dataset, with improvements ranging from 3% to 17% on the local EHR dataset, even without fine-tuning on the local data. Publicly available LLMs, predominantly trained on data from high-income regions, were found to be unreliable when applied in a real-world clinical setting in Pakistan. Fine-tuning them with local EHR data and regional clinical contexts improved their reliability, demonstrating a feasible adaptation strategy that is substantially less resource-intensive than training large language models from scratch. Close collaboration between local clinical and technical experts to curate and leverage more representative, inclusive, and unbiased medical datasets, can play a crucial role in further ensuring reliability of LLMs for resource-limited, overburdened settings, to be used in ways that are safe, fair, and beneficial for all.

Large Language Models (LLMs) have the potential of facilitating the development of Artificial Intelligence technology to assist medical experts for interactive decision support. This potential has been illustrated by the state-of-the-art performance obtained by LLMs in Medical Question Answering, with striking results such as passing marks in licensing medical exams. However, while impressive, the required quality bar for medical applications remains far from being achieved. Currently, LLMs remain challenged by outdated knowledge and by their tendency to generate hallucinated content. Furthermore, most benchmarks to assess medical knowledge lack reference gold explanations which means that it is not possible to evaluate the reasoning of LLMs predictions. Finally, the situation is particularly grim if we consider benchmarking LLMs for languages other than English which remains, as far as we know, a totally neglected topic. In order to address these shortcomings, in this paper we present MedExpQA, the first multilingual benchmark based on medical exams to evaluate LLMs in Medical Question Answering. To the best of our knowledge, MedExpQA includes for the first time reference gold explanations, written by medical doctors, of the correct and incorrect options in the exams. Comprehensive multilingual experimentation using both the gold reference explanations and Retrieval Augmented Generation (RAG) approaches show that performance of LLMs, with best results around 75 accuracy for English, still has large room for improvement, especially for languages other than English, for which accuracy drops 10 points. Therefore, despite using state-of-the-art RAG methods, our results also demonstrate the difficulty of obtaining and integrating readily available medical knowledge that may positively impact results on downstream evaluations for Medical Question Answering. Data, code, and fine-tuned models will be made publicly available.11https://huggingface.co/datasets/HiTZ/MedExpQA.

PurposeCurrent benchmarks evaluating large language models (LLMs) in ophthalmology are narrow and disproportionately prioritize accuracy. We introduce BEnchmarking LLMs for Ophthalmology (BELO), a standardized evaluation benchmark developed through multiple rounds of expert checking by 13 ophthalmologists. BEnchmarking LLMs for Ophthalmology assesses ophthalmology-related knowledge and reasoning quality.SubjectsThis study did not involve human participation.DesignCross-sectional study.MethodsUsing keyword matching and a fine-tuned PubMed Bidirectional Encoder Representations from Transformers model, we curated ophthalmology-specific multiple-choice questions (MCQs) from diverse medical data sets (Basic and Clinical Science Course [BCSC], Multi-Subject Multi-Choice Dataset for Medical domain [MedMCQA], Medical Question Answering [MedQA], Biomedical Semantic Indexing and Question Answering [BioASQ], and PubMed Question Answering [PubMedQA]). The data set underwent multiple rounds of expert checking. Duplicate and substandard questions were systematically removed. Ten ophthalmologists refined the explanations of each MCQ's correct answer. This was further adjudicated by 3 senior ophthalmologists. To illustrate BELO's utility, we evaluated 8 LLMs (OpenAI o1, o3-mini, GPT-5, GPT-4o, DeepSeek-R1, MedGemma-4B, Llama-3-8B, and Gemini 1.5 Pro).Main Outcome MeasuresThe 8 LLMs were evaluated in terms using accuracy, macro-F1, and 5 text-generation metrics (Recall-Oriented Understudy for Gisting Evaluation, BERTScore, BARTScore, Metric for Evaluation of Translation with Explicit Ordering, and AlignScore). In a further evaluation involving human experts, 2 ophthalmologists qualitatively reviewed 50 randomly selected outputs for accuracy, comprehensiveness, and completeness.ResultsBEnchmarking LLMs for Ophthalmology consists of 900 high-quality, expert-reviewed questions aggregated from 5 sources: BCSC (260), BioASQ (10), MedMCQA (572), MedQA (40), and PubMedQA (18). To demonstrate BELO's utility, we conducted a series of benchmarking exercises. In the quantitative evaluation, GPT-5 achieved the highest accuracy (0.90, 95% confidence interval [CI]: 0.89–0.92) and macro-F1 score (0.91, 95% CI: 0.89–0.93). On the other hand, the models' performance on text-generation metrics varied and were generally suboptimal, with scores ranging from 20.4 to 72.0 (out of 100, excluding the BARTScore metric), indicating room for improvement in clinical reasoning. In expert evaluations, GPT-4o was rated highest for accuracy and readability, while Gemini 1.5 Pro scored highest for completeness. A public leaderboard has been established to promote transparent evaluation and reporting. Importantly, the BELO data set will remain a hold-out, evaluation-only benchmark to ensure fair and reproducible comparisons of future models.ConclusionsBEnchmarking LLMs for Ophthalmology provides a robust clinically relevant benchmark for evaluating both the accuracy and reasoning capabilities of current and emerging LLMs in ophthalmology. Future BELO benchmarking efforts will expand to include vision-based question answering and clinical scenario management tasks.Financial Disclosure(s)The authors have no proprietary or commercial interest in any materials discussed in this article.

The integration of large language models (LLMs) into medical diagnostics has garnered substantial attention due to their potential to enhance diagnostic accuracy, streamline clinical workflows, and address health care disparities. However, the rapid evolution of LLM research necessitates a comprehensive synthesis of their applications, challenges, and future directions. This scoping review aimed to provide an overview of the current state of research regarding the use of LLMs in medical diagnostics. The study sought to answer four primary subquestions, as follows: (1) Which LLMs are commonly used? (2) How are LLMs assessed in diagnosis? (3) What is the current performance of LLMs in diagnosing diseases? (4) Which medical domains are investigating the application of LLMs? This scoping review was conducted according to the Joanna Briggs Institute Manual for Evidence Synthesis and adheres to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews). Relevant literature was searched from the Web of Science, PubMed, Embase, IEEE Xplore, and ACM Digital Library databases from 2022 to 2025. Articles were screened and selected based on predefined inclusion and exclusion criteria. Bibliometric analysis was performed using VOSviewer to identify major research clusters and trends. Data extraction included details on LLM types, application domains, and performance metrics. The field is rapidly expanding, with a surge in publications after 2023. GPT-4 and its variants dominated research (70/95, 74% of studies), followed by GPT-3.5 (34/95, 36%). Key applications included disease classification (text or image-based), medical question answering, and diagnostic content generation. LLMs demonstrated high accuracy in specialties like radiology, psychiatry, and neurology but exhibited biases in race, gender, and cost predictions. Ethical concerns, including privacy risks and model hallucination, alongside regulatory fragmentation, were critical barriers to clinical adoption. LLMs hold transformative potential for medical diagnostics but require rigorous validation, bias mitigation, and multimodal integration to address real-world complexities. Future research should prioritize explainable artificial intelligence frameworks, specialty-specific optimization, and international regulatory harmonization to ensure equitable and safe clinical deployment.

The rapid growth of biomedical literature poses challenges for manual knowledge curation and synthesis. Biomedical Natural Language Processing (BioNLP) automates the process. While Large Language Models (LLMs) have shown promise in general domains, their effectiveness in BioNLP tasks remains unclear due to limited benchmarks and practical guidelines. We perform a systematic evaluation of four LLMs—GPT and LLaMA representatives—on 12 BioNLP benchmarks across six applications. We compare their zero-shot, few-shot, and fine-tuning performance with the traditional fine-tuning of BERT or BART models. We examine inconsistencies, missing information, hallucinations, and perform cost analysis. Here, we show that traditional fine-tuning outperforms zero- or few-shot LLMs in most tasks. However, closed-source LLMs like GPT-4 excel in reasoning-related tasks such as medical question answering. Open-source LLMs still require fine-tuning to close performance gaps. We find issues like missing information and hallucinations in LLM outputs. These results offer practical insights for applying LLMs in BioNLP.

With the rapid development of large language models (LLMs) in natural language processing and generation tasks, researchers are continuously validating their application potential in the field of medical artificial intelligence (AI). Generative Pre-trained Transformer 5 (GPT-5), Pathways Language Model 2 (PaLM-2), Large Language Model Meta AI 3 (Llama-3), Gemini-2.5, and various medical vertical models have demonstrated excellent performance in many medical and health-related tasks. However, LLMs still face key bottlenecks in real-world medical scenarios, including “illusions” slow knowledge updates, and a lack of interpretability. These problems severely restrict the safe deployment of large language models in high-risk medical scenarios and have become a major obstacle to the transition of medical artificial intelligence from research to clinical application. Retrieval-augmented generation (RAG) technology has therefore become an important solution for improving the credibility of medical LLMs. RAG significantly reduces the risk of information errors by retrieving authoritative medical knowledge before generation while maintaining knowledge updates without retraining. A review of research progress on RAG in the medical field is of great theoretical value and practical significance for promoting the design, evaluation, and standardized application of trustworthy medical AI. This review aims to systematically summarize the research progress of RAG in medical scenarios, including the technical frameworks, typical applications, and development trends of three types of methods—naive RAG, advanced RAG, and modular RAG—and further discuss their significance in medical reliability, health equity, and personalized medicine. The naive RAG architecture, through a standard “index-retrieval-generation” process, transforms external knowledge bases into a structured vector space that can be queried by an LLM, achieving positive results in electronic medical record (EMR) summary, preoperative assessment, and medical question answering (QA). However, naive RAG still has limitations in retrieval accuracy and cross-modal processing. To address these shortcomings, advanced RAG methods have been devised that significantly improve the model’s decision-making capabilities by improving retrieval strategies, enhancing inference chains, and introducing self-reflection mechanisms. Furthermore, modular RAG design offers composable, multi-module systems to support multi-source knowledge integration and complex task decomposition. In terms of application value, RAG’s contributions to the healthcare field are primarily reflected in these aspects. On the one hand, it significantly improves the reliability of medical AI by introducing traceable knowledge to reduce illusions. On the other hand, it helps promote healthcare equity by driving a shift toward a patient-centered healthcare model through localized knowledge bases and multilingual support. However, the deployment of RAG in medical scenarios still faces many challenges, such as data privacy risks. Future research should focus on improving self-supervised reflection capabilities and developing cross-modal knowledge fusion technologies. As research develops, RAG is expected to become a core foundational capability of future intelligent healthcare systems, further promoting the development of safe and efficient intelligent healthcare.

Over the past five years, medical question answering (QA) has witnessed rapid advancements, driven by the evolution of large language models (LLMs), explainability techniques, and the growing emphasis on human-centered AI in healthcare. In clinical settings, where decisions directly impact patient outcomes, the need for transparency, interpretability, and trust in AI-generated answers is not merely beneficial but essential..Traditional QA systems, although reasonably accurate, often lack explainability, thus continuing to function as black boxes. This opacity raises fundamental concerns about accountability, safety in high-stakes medical domains. We introduce ExplainMed++, a novel system designed to generate human-centered medical QA summaries with intelligible explanations. Our approach enhances the interpretability and reliability of QA outputs by fine-tuning state-of-the-art LLMs using domain-specific datasets such as MedQA. This design is influenced by cognitive models of physician decision-making and recent progress in self-rationale LLMs. This literature increasingly highlights the limitations of accuracy-centric but also provides evidence-based and clinically aligned explanations.We incorporate an explanation module that leverages retrieval-augmented generation and reflective reasoning to align model outputs with clinical reasoning patterns. This literature review synthesizes findings from 70 influential studies published between 2020 and 2025,spanning foundational work in biomedical NLP, explainable AI(XAI), highlighting key technological advances, unresolved challenges, and promising future directions in explainable medical QA. These works collectively illuminate both the promise and limitations of current methodologies, including prompt-based reasoning and counterfactual generation. ExplainMed++ contributes to this landscape by bridging the gap between model performance and clinical usability and also, advancing both the performance and the trust-worthiness of QA systems in healthcare.

A survey for large language models in biomedicine.

Large language models (LLMs) are rapidly advancing natural language processing and driving increasing interest across biomedical research, clinical care, and healthcare operations. This systematic review synthesizes current empirical evidence on LLM applications in healthcare and biomedical informatics, focusing on their capabilities, evaluation practices, and reported outcomes. Existing studies highlight promising uses in research, where LLMs assist literature-informed genomic interpretation, functional annotation, and biological hypothesis generation, as well as support tasks related to protein and multiomics analysis. In clinical contexts, LLMs are primarily evaluated for natural language-driven tasks, including electronic health record summarization, clinical documentation support, medical question answering, and information extraction, rather than autonomous diagnostic or therapeutic decision-making. Early investigations also describe potential value in healthcare operations, patient communication, clinician education, and drug discovery workflows largely through knowledge retrieval, text generation, and semantic search. However, current evidence remains preliminary: most studies are retrospective, benchmark-based, simulation-driven, or limited to controlled research settings. Key challenges include hallucination and unreliable reasoning, bias and inequitable performance across populations, data privacy and security constraints, reproducibility limitations, and lack of prospective clinical validation and regulatory guidance. Emerging strategies, such as domain-specific pretraining, retrieval augmentation, multi-modal architectures, and privacy-preserving learning, aim to improve reliability, safety, and real-world applicability. This review concludes by outlining methodological, infrastructural, and governance requirements needed to responsibly integrate LLMs into biomedical workflows, emphasizing that clinical deployment remains exploratory and must be rigorously evaluated.

MedSumGraph: enhancing GraphRAG for medical QA with summarization and optimized prompts.

Large language models (LLMs) often struggle with factual accuracy in knowledge-intensive domains like healthcare. We introduce LEAF (Learning and Evaluation Augmented by Fact-Checking), a framework for improving LLM factuality in medical question answering. LEAF comprises three components: (1) RAFE, a robust fact-checking system using open-source LLMs and domain-specific retrieval to evaluate response accuracy; (2) Fact-Check-then-RAG, which leverages fact-checking results to guide retrieval without parameter updates; and (3) Learning from Fact Check, enabling self-training through supervised fine-tuning or preference-based learning using fact-checking as pseudo-labels. Experimental results show that RAFE outperforms Factcheck-GPT in detecting inaccuracies, Fact-Check-then-RAG effectively corrects errors, and Learning from Fact Check improves performance without labeled data. In a real-world healthcare deployment with proprietary medical documents, LEAF achieved an 83% improvement in factuality scores, demonstrating practical applicability for adapting general-purpose LLMs to organization-specific knowledge. Our framework provides a scalable solution for industrial applications requiring high factual accuracy.

The assignment of current procedure terminology (CPT) codes to medical events is a highly cumbersome, logistic challenge for many healthcare organizations, as well as a significant contributor to medical expenses. Improvement in the allocation of medical resources and expenses dedicated to such tasks can be achieved through automation. However, because of the complex nature of medical records, automation of procedure terminologies is just now developing with the advent of machine learning methods. In this study, we develop a fine-tuned large language model (LLaMA-3B) as a high-reliability predictor for CPT codes. As input, we use 2018 pathology report text data, including gross report, microscopic description, brief medical history, and final diagnosis. We define our dataset with the top five most common technical component CPT codes, which account for 85% of all samples. As a (meta) predictor of the veracity of the large language model itself, we use the prediction’s softmax scalar value of the classification model, borrowing from similar recent approaches in conformal prediction. Specifically, the validation set, which is distinct from both the train and test set, is used to establish a prediction threshold below which the model withholds judgement. We show that, by fine tuning an off-the-shelf language model on pathology report text alone, we can achieve 95% prediction accuracy of the top 5 most common CPT codes on our dataset. We also then demonstrate the utility of combining large language models with conformal prediction. The two in combination raise our accuracy to 99.5% when we allow the model to abstain on making a prediction on 30% of the data, as is determined by a separate threshold on the scalar value of the predictor from the aforementioned validation set. We therefore present a highly flexible model for medical coding, which is simultaneously provably-reliable. Large language models can as such serve as a powerful tool for overcoming challenges in medical billing, thereby improving healthcare efficiency and reducing medical costs.

Medical Visual Question Answering (MedVQA) is a dual research field that intersects medical imaging and natural language processing, for better interpretability and accessibility of medical image data.Medical image quality is paramount for accurate diagnostics and subsequent medical visual question answering (MedVQA) tasks. This research focuses on applying and then analyzing results of different denoising methods on VQA-RAD, Medical VQA dataset to enhance quality of images. This study explores effectiveness of different traditional and deep learning based methods to reduce noise within medical images, thereby improving the accuracy and reliability of MedVQA task. We applied different traditional denoising filtering methods such as, gaussian filter, median filter, average filter, bilateral filter and convolutional autoencoder (CAE) based on deep learning on a VQA-RAD dataset to compare effectiveness of each denoising methods to improve image quality. Through comprehensive experiments and evaluations, this paper demonstrates that the convolutional autoencoder is potentially enhancing quality of medical images with an emphasis on preserving essential diagnostic information while suppressing unwanted noise with compare to other traditional denoising filters. The denoised images are then employed as input to improve accuracy for MedVQA tasks. The results of this research will help in optimizing medical imaging pipelines, ultimately benefiting clinical decision-making and healthcare outcomes.

Despite large language models (LLMs) showing strong capabilities in general medical question answering, they still struggle with multi-domain knowledge coverage, consistent reasoning, and preventing hallucination in complex clinical conditions. This paper introduces a two-stage Classification—Debate Multi-Agent Framework that improves reasoning consistency, accuracy, and interpretability. In the first stage, the system performs MeSH-guided domain classification, routing medical multiple-choice questions into multiple LLM-based “medical specialist” agents, each of which has expertise in a specific medical subfield, such as anatomy, organism, diseases, etc. In the second stage, domain-specific agents independently generate answers and engage in structured debate rounds upon shared history, with the final decision determined by consensus or majority voting. Experiments on MedQA and MedMCQA demonstrate consistent accuracy gains across all model configurations. Results further show that fine-grained domain classification improves reasoning alignment, and that output format also significantly affects performance. Without requiring additional fine-tuning, retrieval augmentation, or complex pipelines, the proposed lightweight framework outperforms single-model baselines and several existing multi-agent systems, highlighting the effectiveness of integrating domain-aware classification with structured debate.

Deploying large language models (LLMs) in clinical settings is limited by security, reliability, latency, and accessibility concerns that favor smaller, on-device or on-premise models. However, these smaller models may struggle to meet accuracy requirements. While fine-tuning and retrieval-augmented generation (RAG) can improve domain-specific accuracy, these methods require additional labeled data, technical skill, and infrastructure. In contrast, test-time scaling —allocating extra token-budget during inference—offers a training-free alternative to increasing accuracy. However, the trade-offs between these strategies and their interaction with model size remain poorly understood for medical reasoning. To address this gap, we compare three approaches—test-time scaling, fine-tuning, and context grounding—using the Gemma and MedGemma family of LLMs (Gemma-3 1B, Gemma-3 4B, Gemma-3 27B, MedGemma-4B, and MedGemma 27B) and evaluate these systems across common biomedical question-answering (QA) datasets and a set of recently released medical exam questions with the performance of practicing clinicians available for comparison. We test baseline prompts (direct answer, Chain-of-Thought, and self-consistency) while introducing a new prompting method we call “prompt-chaining for continuous reflection” (PCCR) that forces inference time minimum token-generation budgets. We assess accuracy and tokens-generated, allowing us to investigate the accuracy–efficiency trade-offs across prompting, context-grounding, fine-tuning, and model scales. We discover equivalency points where smaller models perform comparably to larger ones with increased reasoning budgets, context-grounding, or fine-tuning. We also find inflection points where context-grounding and test-time scaling used together lead to degrading performance. Using these empirical results, we formulate a general framework with equations to balance cost-benefit trade-offs when engineering LLM-based systems for medical reasoning and QA. We recommend generalizable configurations, designs, and patterns to achieve accuracy and efficiency objectives for example use-cases relevant to healthcare organizations. Author summary When doctors and hospitals want to use artificial intelligence for medical tasks, they face difficult choices. The most capable AI systems are expensive to run and require sending sensitive patient and hospital data to external servers. Smaller systems that can run locally are more practical but may be less accurate. Our research asked: with what configurations can we make smaller AI systems perform as well as larger ones for medical reasoning? We tested five AI models of different sizes, including both generalist and medically-specialized models, on thousands of medical questions using various prompting strategies, including a new and very simple method we developed that encourages the AI to reason more extensively. We discovered several surprising findings which we describe in detail. Most importantly, we found that smaller, cheaper models can match larger ones when given the right combination of prompting strategy, specialized training, and supporting information. We translated these findings into practical guidelines that can help choose AI configurations that balance accuracy, speed, and cost. Our framework could help make medical AI more accessible to institutions with limited computational resources while achieving the highest possible accuracy.