Clinical Language Model Evaluation

Evaluating large language models (LLMs) for medical applications remains challenging due to benchmark saturation, limited data accessibility, and insufficient coverage of relevant tasks. Existing suites have either saturated, heavily depend on restricted datasets, or lack comprehensive model coverage. We introduce Medmarks, a fully open-source evaluation suite with 30 benchmarks spanning question answering, information extraction, medical calculations, and open-ended clinical reasoning. We perform a systematic evaluation of 61 models across 71 configurations using verifiable metrics and LLM-as-a-Judge. Our results show that frontier reasoning models (Gemini 3 Pro Preview, GPT-5.1, & GPT-5.2) achieve the highest performance across both benchmarks, most frontier proprietary models are significantly more token efficient than open-weight alternatives, medically fine-tuned models outperform their generalist counterparts, and that models are susceptible to answer-order bias (particularly smaller models and Grok 4). A subset of our evals (Medmarks-T) can be directly used as reinforcement learning environments to post-train LLMs for medical reasoning. Code is available at https://github.com/MedARC-AI/Medmarks

Brazil's Unified Health System (SUS) relies on official clinical guidelines that define diagnostic criteria, treatments, dosages, and monitoring procedures for over 200 million citizens. Yet current LLMs perform poorly on this guideline-specific knowledge, and no benchmark evaluates clinical recall grounded in Brazilian Portuguese protocols. We address this gap by adapting Qwen2.5-14B-Instruct to the Brazilian clinical domain. From 178 official guidelines (~5.4M tokens), we generate ~70M tokens of synthetic data in three formats -- rephrases, wiki-style articles, and question-answer pairs -- using four generator LLMs. We then apply continual pre-training followed by Group Relative Policy Optimization (GRPO). We introduce HealthBench-BR, with 1,780 balanced true/false clinical assertions, and PCDT-QA, with 890 open-ended clinical questions scored by an LLM judge. Our best model achieves 83.9% on HealthBench-BR and 85.4% on PCDT-QA, outperforming GPT-5.2, Claude Sonnet 4.6, Gemini 3.1 Pro, and Google AI Overview's web-grounded RAG despite having only 14B parameters. Ablations show that generator diversity and reinforcement learning are critical to these gains. We release all datasets, benchmarks, and model weights to support reproducible clinical NLP research for Brazilian Portuguese. Code, data, and model weights are available at https://github.com/hugoabonizio/clinical-protocols-br

Medical large language model (LLM) evaluations rely on simplified, exam-style benchmarks that rarely reflect the ambiguity of real-world medical inquiries. We introduce the CLinical Evaluation of Ambiguity and Reliability (CLEAR) framework, which assesses how decision-space presentation, ambiguity, and uncertainty affect LLMs' reasoning on medical benchmarks. CLEAR systematically perturbs (1) the number of plausible answer options, (2) the presence of a ground truth or abstention option, and (3) the semantic framing of answer options. Applying CLEAR on three benchmarks evaluated across 17 LLMs reveals three notable limitations of existing evaluation methods. First, increasing the number of plausible answers degrades a model's ability to identify the correct answer and abstain against incorrect ones. Second, this lack of caution intensifies as the framing of abstention shifts from assertive rejection like "None of the Above" to uncertainty admission like "I don't know" (IDK). Notably, just including IDK in the answer space increases incorrect answer selections. Lastly, we formalize the performance gap between identifying the correct answer and abstaining from incorrect ones as the humility deficit, which worsens with model scale. Our findings reveal limitations in standard medical benchmarks and underscore that scaling alone does not resolve LLM reliability issues.

Millions of clinicians use ChatGPT to support clinical care, but evaluations of the most common use cases in model-clinician conversations are limited. We introduce HealthBench Professional, an open benchmark for evaluating large language models on real tasks that clinicians bring to ChatGPT in the course of their work. The benchmark is organized around three common use cases central to clinical practice: care consult, writing and documentation, and medical research. Each example includes a physician-authored conversation with ChatGPT for Clinicians and is scored via rubrics written and iteratively adjudicated by three or more physicians across three phases. HealthBench Professional examples were carefully selected for quality, representativeness, and difficulty for OpenAI's current frontier models, to enable continued measurement of progress. Difficult examples for recent OpenAI models were enriched by roughly 3.5 times relative to the candidate pool of 15,079 examples. Additionally, about one-third of examples involve physicians conducting deliberate adversarial testing of models. As a strong baseline, we also collected human physician responses for all tasks (unbounded time, specialist-matched, web access). The best scoring system, GPT-5.4 in ChatGPT for Clinicians, outperforms base GPT-5.4, all other models, and human physicians. We hope HealthBench Professional provides the healthcare AI community a measure to track frontier model progress in real-world clinical tasks and build systems that clinicians can trust to improve care.

Accurate and consistent Emergency Severity Index (ESI) assignment remains a persistent challenge in emergency departments, where highly variable free-text triage documentation contributes to mistriage and workflow inefficiencies. This study evaluates whether open-source small language models (SLMs) can serve as reliable, privacy-preserving decision-support tools for clinical triage. We systematically compared multiple SLMs across diverse prompting pipelines and found that clinical vignettes, concise summaries of triage narratives, yielded the most accurate predictions. The SLM, Qwen2.5-7B, demonstrated the strongest balance of accuracy, stability, and computational efficiency. Through large-scale domain adaptation using expert-curated and silver-standard pediatric triage data, fine-tuned Qwen2.5-7B models substantially reduced discordance and clinically significant errors, outperforming all baseline SLMs and advanced proprietary large language models (LLMs, e.g., GPT-4o). These findings highlight the feasibility of institution-specific SLMs for reliable, privacy-preserving ESI decision support and underscore the importance of targeted fine-tuning over more complex inference strategies.

Patient simulators are gaining traction in mental health training by providing scalable exposure to complex and sensitive patient interactions. Simulating depressed patients is challenging, as safety constraints and high patient variability complicate simulations and underscore the need for simulators that capture diverse and realistic patient behaviors. However, existing evaluations heavily rely on LLM-judges with poorly specified prompts and do not assess behavioral diversity. We introduce PSI-Bench, an automatic evaluation framework that provides interpretable, clinically meaningful diagnostics of depression patient simulator behavior across turn-, dialogue-, and population-level dimensions. Using PSI-Bench, we benchmark seven LLMs across two simulator frameworks and find that simulators produce overly long, lexically diverse responses, show reduced variability, and move through therapeutic stages and toward positive valence too quickly. We also show that the simulation framework has a larger impact on fidelity than the model scale. Results from a human study demonstrate that our benchmark is strongly aligned with judgments of mental health professionals. Our work reveals key limitations of current depression patient simulators and provides an interpretable, extensible benchmark to guide future simulator design and evaluation.

People increasingly turn to general-purpose AI chatbots for advice about emotional and mental health problems, but the ability of these systems to recognize and appropriately triage psychiatric emergencies remains under-characterized. We evaluated psychiatric triage performance in 15 frontier AI chatbots using 112 clinical vignettes spanning four urgency levels, from routine care to immediate emergency assessment. In each trial (1680 total), a chatbot received a single user message conveying all triage-relevant information from one vignette and recommended a timeframe for care. The primary outcome was emergency under-triage; secondary outcomes included triage accuracy and the direction of errors. Vignettes and user messages were generated using a clinician-verified LLM pipeline. Across 415 emergency trials, 23 were under-triaged (5.5%; 95% CI 1.8-15.9). Overall accuracy, averaged across urgency levels, ranged from 42.0% to 71.8% across chatbots and was lowest for intermediate cases (19.6%; 95% CI 11.7-28.1). Every chatbot showed a net over-triage bias; overall, 763 of 786 incorrect assignments (97.1%) were more urgent than the prespecified triage level. The error pattern was similar when predictions were assessed against clinician ratings: 35 of 430 trials involving vignettes rated as emergencies by at least 75% of clinicians were under-triaged (8.1%). AI chatbots recognized most psychiatric emergencies but still missed clinically important cases and frequently over-triaged less urgent presentations. Further evaluations should examine how triage performance changes when clinically relevant information must be elicited through conversation.

While large language models (LLMs) are increasingly being explored for mental health applications, recent studies reveal that they can exhibit stigma toward individuals with psychological conditions. Existing evaluations of this stigma primarily rely on multiple-choice questions (MCQs), which fail to capture the biases embedded within the models' underlying logic. In this paper, we analyze the intermediate reasoning steps of LLMs to uncover hidden stigmatizing language and the internal rationales driving it. We leverage clinical expertise to categorize common patterns of stigmatizing language directed at individuals with psychological conditions and use this framework to identify and tag problematic statements in LLM reasoning. Furthermore, we rate the severity of these statements, distinguishing between overt prejudice and more subtle, less immediately harmful biases. To broaden the reasoning domain and capture a wider array of patterns, we also extend an existing mental health stigma benchmark by incorporating additional psychological conditions. Our findings demonstrate that evaluating model reasoning not only exposes substantially more stigma than traditional MCQ-based methods but also helps identify the flaws in the LLMs' logic and their understanding of mental health conditions.

Objective. Clinical AI documentation systems require evaluation methodologies that are clinically valid, economically viable, and sensitive to iterative changes. Methods requiring expert review per scoring instance are too slow and expensive for safe, iterative deployment. We present a case-specific, clinician-authored rubric methodology for clinical AI evaluation and examine whether LLM-generated rubrics can approximate clinician agreement. Materials and Methods. Twenty clinicians authored 1,646 rubrics for 823 clinical cases (736 real-world, 87 synthetic) across primary care, psychiatry, oncology, and behavioral health. Each rubric was validated by confirming that an LLM-based scoring agent consistently scored clinician-preferred outputs higher than rejected ones. Seven versions of an EHR-embedded AI agent for clinicians were evaluated across all cases. Results. Clinician-authored rubrics discriminated effectively between high- and low-quality outputs (median score gap: 82.9%) with high scoring stability (median range: 0.00%). Median scores improved from 84% to 95%. In later experiments, clinician-LLM ranking agreement (tau: 0.42-0.46) matched or exceeded clinician-clinician agreement (tau: 0.38-0.43), attributable to both ceiling compression and LLM rubric improvement. Discussion. This convergence supports incorporating LLM rubrics alongside clinician-authored ones. At roughly 1,000 times lower cost, LLM rubrics enable substantially greater evaluation coverage, while continued clinical authorship grounds evaluation in expert judgment. Ceiling compression poses a methodological challenge for future inter-rater agreement studies. Conclusion. Case-specific rubrics offer a path for clinical AI evaluation that preserves expert judgment while enabling automation at three orders lower cost. Clinician-authored rubrics establish the baseline against which LLM rubrics are validated.

Large language models (LLMs) are increasingly deployed, yet their outputs can be highly sensitive to routine, non-adversarial variation in how users phrase queries, a gap not well addressed by existing red-teaming efforts. We propose Green Shielding, a user-centric agenda for building evidence-backed deployment guidance by characterizing how benign input variation shifts model behavior. We operationalize this agenda through the CUE criteria: benchmarks with authentic Context, reference standards and metrics that capture true Utility, and perturbations that reflect realistic variations in the Elicitation of model behavior. Guided by the PCS framework and developed with practicing physicians, we instantiate Green Shielding in medical diagnosis through HealthCareMagic-Diagnosis (HCM-Dx), a benchmark of patient-authored queries, together with structured reference diagnosis sets and clinically grounded metrics for evaluating differential diagnosis lists. We also study perturbation regimes that capture routine input variation and show that prompt-level factors shift model behavior along clinically meaningful dimensions. Across multiple frontier LLMs, these shifts trace out Pareto-like tradeoffs. In particular, neutralization, which removes common user-level factors while preserving clinical content, increases plausibility and yields more concise, clinician-like differentials, but reduces coverage of highly likely and safety-critical conditions. Together, these results show that interaction choices can systematically shift task-relevant properties of model outputs and support user-facing guidance for safer deployment in high-stakes domains. Although instantiated here in medical diagnosis, the agenda extends naturally to other decision-support settings and agentic AI systems.

Multiple myeloma is managed through sequential lines of therapy over years to decades, with each decision depending on cumulative disease history distributed across dozens to hundreds of heterogeneous clinical documents. Whether LLM-based systems can synthesise this evidence at a level approaching expert agreement has not been established. A retrospective evaluation was conducted on longitudinal clinical records of 811 myeloma patients treated at a tertiary centre (2001-2026), covering 44,962 documents and 1,334,677 laboratory values, with external validation on MIMIC-IV. An agentic reasoning system was compared against single-pass retrieval-augmented generation (RAG), iterative RAG, and full-context input on 469 patient-question pairs from 48 templates at three complexity levels. Reference labels came from double annotation by four oncologists with senior haematologist adjudication. Iterative RAG and full-context input converged on a shared ceiling (75.4% vs 75.8%, p = 1.00). The agentic system reached 79.6% concordance (95% CI 76.4-82.8), exceeding both baselines (+3.8 and +4.2 pp; p = 0.006 and 0.007). Gains rose with question complexity, reaching +9.4 pp on criteria-based synthesis (p = 0.032), and with record length, reaching +13.5 pp in the top decile (n = 10). The system error rate (12.2%) was comparable to expert disagreement (13.6%), but severity was inverted: 57.8% of system errors were clinically significant versus 18.8% of expert disagreements. Agentic reasoning was the only approach to exceed the shared ceiling, with gains concentrated on the most complex questions and longest records. The greater clinical consequence of residual system errors indicates that prospective evaluation in routine care is required before these findings translate into patient benefit.

Large language models are being deployed as mental health support agents at scale, yet only 16% of LLM-based chatbot interventions have undergone rigorous clinical efficacy testing, and simulations reveal psychological deterioration in over one-third of cases. We evaluate four generative models on 250 Prolonged Exposure (PE) therapy scenarios and 146 CBT cognitive restructuring exercises (plus 29 severity-escalated variants), scored by a three-judge LLM panel. All models scored near-perfectly on surface acknowledgment (~0.91-1.00) while therapeutic appropriateness collapsed to 0.22-0.33 at the highest severity for three of four models, with protocol fidelity reaching zero for two. Under CBT severity escalation, one model's task completeness dropped from 92% to 71% while the frontier model's safety-interference score fell from 0.99 to 0.61. We identify a systematic, modality-spanning failure: RLHF safety alignment disrupts the therapeutic mechanism of action by grounding patients during imaginal exposure, offering false reassurance, inserting crisis resources into controlled exercises, and refusing to challenge distorted cognitions mentioning self-harm in PE; and through task abandonment or safety-preamble insertion during CBT cognitive restructuring. These findings motivate a five-axis evaluation framework (protocol fidelity, hallucination risk, behavioral consistency, crisis safety, demographic robustness), mapped onto FDA SaMD and EU AI Act requirements. We argue that no AI mental health system should proceed to deployment without passing multi-axis evaluation across all five dimensions.

Large language models (LLMs) are increasingly utilized in clinical reasoning and risk assessment. However, their interpretive reliability in critical and indeterminate domains such as psychiatry remains unclear. Prior work has identified algorithmic biases and prompt sensitivity in these systems, raising concerns about how contextual information may influence model outputs, but there remains no systematic way to assess these, especially in the psychiatric domain. We propose an approach for reliability auditing downstream LLM tasks by structuring evaluation around the impact of prompt design and the inclusion of medically insignificant inputs on predicted hospitalization risk scores, which is often the first downstream AI clinical-decision-making task. In our audit, a cohort of synthetic patient profiles (n = 50) is generated, each consisting of 15 clinically relevant features and up to 50 clinically insignificant features, across four prompt reframings (neutral, logical, human impact, clinical judgment). We audit four LLMs (Gemini 2.5 Flash, LLaMa 3.3 70b, Claude Sonnet 4.6, GPT-4o mini), and our results show that including medically insignificant variables resulted in a statistically significant increase in the absolute mean predicted hospitalization risk and output variability across all models and prompts, indicating reduced predictive stability as contextual noise increased. Clinically insignificant features had an effect on instability across many model-prompt conditions, and prompt variations independently affected the trajectory of instability in a model-dependent manner. These findings quantify how LLM-based psychiatric risk assessments are sensitive to non-clinical information, highlighting the need for systematic evaluations of attributional stability and uncertainty behavior like this before clinical deployments.

Objectives: Large language models (LLMs) are increasingly used for clinical text summarization, yet structured methods to assess associated patient safety risks remain limited. Failure Mode, Effects, and Criticality Analysis (FMECA) provides a proactive framework for systematic risk identification but has not been adapted to LLM-generated clinical content. This study aimed to develop and validate a novel FMECA framework for the prospective assessment of patient safety risks in LLM-generated clinical summaries. Materials and Methods: An interdisciplinary expert panel (n = 8) developed a taxonomy of failure modes through literature review and brainstorming. Standard FMECA dimensions (occurrence, severity, detectability) were adapted into 5-point ordinal scales. The framework was applied to 36 discharge summaries from four patients, generated by an open LLM (GPT-OSS 120B) using real-world clinical data from the Geneva University Hospitals. Reviewers independently annotated the summaries across two rounds. Inter-rater reliability was assessed at failure mode, severity and detectability score levels. Usability and content validity were evaluated using an adapted System Usability Scale and structured feedback. Results: The final framework comprised 14 failure modes organized into categories. Inter-rater agreement improved between rounds, reaching moderate-to-substantial agreement for failure mode identification and good agreement for severity and detectability scoring. Usability was rated as good (mean SUS: 79.2/100), with high evaluator confidence. Discussion and Conclusion: This study presents the first FMECA-based framework for systematic patient safety risk assessment of LLM-generated clinical summaries. The framework provides a structured and reproducible method for identifying clinically relevant risks caused by these summaries.

Despite strong medical benchmark accuracy, LLMs can exhibit severe multi-turn sycophancy in clinical dialogue, abandoning initial correct diagnosis under escalating pressure. We propose \textbf{\textsc{Med-Stress}}, a targeted stress test framework that evaluates belief stability under escalating pressure. Across nine frontier large language models (LLMs), we find a clear dissociation between medical knowledge and robustness: high initial diagnostic capability does not imply high belief stability, yielding large knowledge-robustness gaps for several LLMs. To mitigate this failure mode, we propose a lightweight inference-time defense, \textbf{\texttt{RBED}} (\textbf{R}ole-\textbf{B}ased \textbf{E}pistemic \textbf{D}efense), and \textbf{\texttt{R-FT}} (\textbf{R}esilience-oriented \textbf{F}ine-\textbf{T}uning), a training-time approach that internalizes evidence-based resistance to pressure. Experiments show that \textbf{\texttt{R-FT}} nearly eliminates belief change and substantially improves robustness.

Medication reconciliation at clinical handoffs is a high-stakes, error-prone process. Large language models are increasingly proposed to assist with this task using FHIR-structured patient records, but a fundamental and largely unstudied variable is how the FHIR data is serialised before being passed to the model. We present the first systematic comparison of four FHIR serialisation strategies (Raw JSON, Markdown Table, Clinical Narrative, and Chronological Timeline) across five open-weight models (Phi-3.5-mini, Mistral-7B, BioMistral-7B, Llama-3.1-8B, Llama-3.3-70B) on a controlled benchmark of 200 synthetic patients, totalling 4,000 inference runs. We find that serialisation strategy has a large, statistically significant effect on performance for models up to 8B parameters: Clinical Narrative outperforms Raw JSON by up to 19 F1 points for Mistral-7B (r = 0.617, p < 10^{-10}). This advantage reverses at 70B, where Raw JSON achieves the best mean F1 of 0.9956. In all 20 model and strategy combinations, mean precision exceeds mean recall: omission is the dominant failure mode, with models more often missing an active medication than fabricating one, which changes how clinical safety auditing priorities should be set. Smaller models plateau at roughly 7-10 concurrent active medications, leaving polypharmacy patients, the patients most at risk from reconciliation errors, systematically underserved. BioMistral-7B, a domain-pretrained model without instruction tuning, produces zero usable output in all conditions, showing that domain pretraining alone is not sufficient for structured extraction. These results offer practical, evidence-based format recommendations for clinical LLM deployment: Clinical Narrative for models up to 8B, Raw JSON for 70B and above. The complete pipeline is reproducible on open-source tools running on an AWS g6e.xlarge instance (NVIDIA L40S, 48 GB VRAM).

Large Language Models (LLMs) are increasingly deployed in healthcare, yet their communicative alignment with clinical standards remains insufficiently quantified. We conduct a multidimensional evaluation of general-purpose and domain-specialized LLMs across structured medical explanations and real-world physician-patient interactions, analyzing semantic fidelity, readability, and affective resonance. Baseline models amplify affective polarity relative to physicians (Very Negative: 43.14-45.10% vs. 37.25%) and, in larger architectures such as GPT-5 and Claude, produce substantially higher linguistic complexity (FKGL up to 16.91-17.60 vs. 11.47-12.50 in physician-authored responses). Empathy-oriented prompting reduces extreme negativity and lowers grade-level complexity (up to -6.87 FKGL points for GPT-5) but does not significantly increase semantic fidelity. Collaborative rewriting yields the strongest overall alignment. Rephrase configurations achieve the highest semantic similarity to physician answers (up to mean = 0.93) while consistently improving readability and reducing affective extremity. Dual stakeholder evaluation shows that no model surpasses physicians on epistemic criteria, whereas patients consistently prefer rewritten variants for clarity and emotional tone. These findings suggest that LLMs function most effectively as collaborative communication enhancers rather than replacements for clinical expertise.

Rising demand for mental health support has increased interest in using Large Language Models (LLMs) for counseling, but adapting them to this safety-critical domain is hindered by limited real-world data due to privacy constraints. Synthetic datasets provide a promising alternative, but existing approaches often rely on unstructured or semi-structured text inputs and overlook structural dependencies between a client's cognitive, emotional, and behavioral states, leading to psychologically inconsistent and less realistic interactions. We introduce Graph2Counsel, a framework for generating synthetic counseling sessions grounded in Client Psychological Graphs (CPGs) that encode relationships among clients' thoughts, emotions, and behaviors. Graph2Counsel uses a structured prompting pipeline guided by counselor strategies and CPG, and explores prompting strategies including CoT and Multi-Agent Feedback. It produces 760 sessions from 76 CPGs across diverse client profiles. In expert evaluation, our dataset outperforms prior datasets on specificity, counselor competence, authenticity, conversational flow, and safety (Krippendorff's $α$ = 0.70). Fine-tuning an open-source model on this dataset improves performance on several metrics on CounselingBench and CounselBench, while matching baselines on others. We further introduce Therapy-Eval, a multi-turn evaluation framework, and demonstrate the effectiveness of our fine-tuned model in realistic therapeutic conversations. We make our code, data and fine-tuned model public.

This study compared repeated generation consistency of exercise prescription outputs across three large language models (LLMs), specifically GPT-4.1, Claude Sonnet 4.6, and Gemini 2.5 Flash, under temperature=0 conditions. Each model generated prescriptions for six clinical scenarios 20 times, yielding 360 total outputs analyzed across four dimensions: semantic similarity, output reproducibility, FITT classification, and safety expression. Mean semantic similarity was highest for GPT-4.1 (0.955), followed by Gemini 2.5 Flash (0.950) and Claude Sonnet 4.6 (0.903), with significant inter-model differences confirmed (H = 458.41, p < .001). Critically, these scores reflected fundamentally different generative behaviors: GPT-4.1 produced entirely unique outputs (100%) with stable semantic content, while Gemini 2.5 Flash showed pronounced output repetition (27.5% unique outputs), indicating that its high similarity score derived from text duplication rather than consistent reasoning. Identical decoding settings thus yielded fundamentally different consistency profiles, a distinction that single-output evaluations cannot capture. Safety expression reached ceiling levels across all models, confirming its limited utility as a differentiating metric. These results indicate that model selection constitutes a clinical rather than merely technical decision, and that output behavior under repeated generation conditions should be treated as a core criterion for reliable deployment of LLM-based exercise prescription systems.

Recent advances in deep research systems enable large language models to retrieve, synthesize, and reason over large-scale external knowledge. In medicine, developing clinical guidelines critically depends on such deep evidence integration. However, existing benchmarks fail to evaluate this capability in realistic workflows requiring multi-step evidence integration and expert-level judgment. To address this gap, we introduce MedProbeBench, the first benchmark leveraging high-quality clinical guidelines as expert-level references. Medical guidelines, with their rigorous standards in neutrality and verifiability, represent the pinnacle of medical expertise and pose substantial challenges for deep research agents. For evaluation, we propose MedProbe-Eval, a comprehensive evaluation framework featuring: (1) Holistic Rubrics with 1,200+ task-adaptive rubric criteria for comprehensive quality assessment, and (2) Fine-grained Evidence Verification for rigorous validation of evidence precision, grounded in 5,130+ atomic claims. Evaluation of 17 LLMs and deep research agents reveals critical gaps in evidence integration and guideline generation, underscoring the substantial distance between current capabilities and expert-level clinical guideline development. Project: https://github.com/uni-medical/MedProbeBench

Large language models (LLMs) are increasingly used in clinical settings, raising concerns about racial bias in both generated medical text and clinical reasoning. Existing studies have identified bias in medical LLMs, but many focus on single models and give less attention to mitigation. This study uses the EU AI Act as a governance lens to evaluate five widely used LLMs across two tasks, namely synthetic patient-case generation and differential diagnosis ranking. Using race-stratified epidemiological distributions in the United States and expert differential diagnosis lists as benchmarks, we apply structured prompt templates and a two-part evaluation design to examine implicit and explicit racial bias. All models deviated from observed racial distributions in the synthetic case generation task, with GPT-4.1 showing the smallest overall deviation. In the differential diagnosis task, DeepSeek V3 produced the strongest overall results across the reported metrics. When embedded in an agentic workflow, DeepSeek V3 showed an improvement of 0.0348 in mean p-value, 0.1166 in median p-value, and 0.0949 in mean difference relative to the standalone model, although improvement was not uniform across every metric. These findings support multi-metric bias evaluation for AI systems used in medical settings and suggest that retrieval-based agentic workflows may reduce some forms of explicit bias in benchmarked diagnostic tasks. Detailed prompt templates, experimental datasets, and code pipelines are available on our GitHub.

Large language models (LLMs) are increasingly explored as scalable tools for mental health counseling, yet evaluating their safety remains challenging due to the interactional and context-dependent nature of clinical harm. Existing evaluation frameworks predominantly assess isolated responses using coarse-grained taxonomies or static datasets, limiting their ability to diagnose how harms emerge and accumulate over multi-turn counseling interactions. In this work, we introduce R-MHSafe, a role-aware mental health safety taxonomy that characterizes clinically significant harm in terms of the interactional roles an AI counselor adopts, including perpetrator, instigator, facilitator, or enabler, combined with clinically grounded harm categories. Then, we propose MHSafeEval, a closed-loop, agent-based evaluation framework that formulates safety assessment as trajectory-level discovery of harm through adversarial multi-turn interactions, guided by role-aware modeling. Using R-MHSafe and MHSafeEval, we conduct a large-scale evaluation across state-of-the-art LLMs. Our results reveal substantial role-dependent and cumulative safety failures that are systematically missed by existing static benchmarks, and show that our framework significantly improves failure-mode coverage and diagnostic granularity.

Large language models are increasingly deployed to simulate patients for clinical training, research, and mental health tools, yet population-level validity remains largely untested. We introduce PsychBench, the first epidemiological audit of LLM patient simulation: 28,800 profiles from four frontier models (GPT-4o-mini, DeepSeek-V3, Gemini-3-Flash, GLM-4.7) evaluated against NHANES and NESARC-III baselines across 120 intersectional cohorts. The central finding is a coherence-fidelity dissociation: models produce clinically plausible individuals while misrepresenting the populations they are drawn from. Variance compression ranges from 14 percent (GLM-4.7) to 62 percent (DeepSeek-V3), eliminating the distributional tails of clinical reality. Despite test-retest correlations above r = 0.90, 36.66 percent of cases cross diagnostic thresholds between runs. Symptom correlation matrices diverge across demographic groups beyond split-half noise, with transgender populations diverging three to five times more than racial differences. Calibration bias is systematic and asymmetric. Models overestimate depression severity for most groups by 3.6 to 6.1 points (Cohen d = 1.13 to 1.91), consistent with training on clinical corpora with elevated base rates. For transgender women the direction inverts: models capture only 8 to 46 percent of documented minority stress elevation, yielding a -5.42 residual (d = -1.55). Models also attribute irritability to Black men and fatigue to women beyond matched controls, encoding racialized and gendered assumptions. Patterns replicate across US and Chinese architectures, indicating failures tied to current training paradigms rather than isolated implementations. For most users, LLM mental health tools risk pathologizing ordinary distress; for transgender users, algorithmic erasure of genuine need. The patients look right. They do not represent real populations.

Safe clinical deployment of Large Language Models (LLMs) requires not only high accuracy but also robust uncertainty calibration to ensure models defer to clinicians when appropriate. Our paper investigates how social descriptors of a patient (specifically sexual orientation and religious affiliation) distort these uncertainty signals and model accuracy. Evaluating nine general-purpose and biomedical LLMs on 2,364 medical questions and their counterfactual variants, we demonstrate that identity markers cause a "calibration crisis". "Homosexual" markers consistently trigger performance drops, and intersectional identities produce idiosyncratic, non-additive harms to calibration. Moreover, a clinician-validated case study in an open-ended generation setting confirms that these failures are not an artifact of the multiple-choice format. Our results demonstrate that the presence of social identity cues does not merely shift predictions; it affects the reliability of confidence signals, posing a significant risk to equitable care and safe deployment in confidence-based clinical workflows.

Process-Level Reward Models (PRMs) are essential for guiding complex reasoning in large language models, yet existing PRM benchmarks cover only general domains such as mathematics, failing to address medical reasoning -- which is uniquely characterized by safety criticality, knowledge intensity, and diverse error patterns. Without a reliable medical PRM evaluation framework, we cannot quantify models' error detection capabilities in clinical reasoning, leaving their safety in real-world healthcare applications unverified. We propose MedPRMBench, the first process-level reward model benchmark for the medical domain. Built through a three-phase pipeline based on Clinical Reasoning Blueprints (CRBs), MedPRMBench systematically generates high-quality evaluation data from seven medical QA sources, covering 14 fine-grained error types across three categories (Simplicity, Soundness, and Sensitivity) with the first 4-level severity grading system to quantify clinical impact. The benchmark comprises 6{,}500 questions with 13{,}000 reasoning chains and 113{,}910 step-level labels, plus 6{,}879 questions for training. Our medical PRM baseline achieves an 87.1% overall PRMScore -- substantially surpassing all baselines -- and serves as a plug-and-play verifier that improves downstream medical QA accuracy by 3.2--6.7 percentage points. Systematic evaluation spanning proprietary frontier models, open-source reasoning models, and medical-specialized models reveals critical weaknesses in current models' medical reasoning error detection capabilities, providing clear directions for future PRM improvement.

Background: Large language models (LLMs) are increasingly proposed for diagnostic support, but few evaluations use real-world multimodal inpatient data, particularly in low and middle-income country (LMIC) public hospitals. Methods: We conducted VALID, a retrospective evaluation of 539 multimodal inpatient cases from a tertiary public hospital in South Africa. Inputs included radiology imaging (CT, MRI, CXR) and reports, laboratory results, clinical notes, and vital signs. Expert panels adjudicated 300 cases (balanced and discordant subsets) to establish ground truth diagnoses, differentials, and reasoning. Ten multimodal LLMs generated zero-shot outputs. A calibrated three-model LLM Jury scored all outputs and routine ward diagnoses across diagnostic accuracy, differential quality, reasoning, and patient safety (>10,000 evaluations). Primary outcomes were composite scores ($S_3$, $S_4$) and win rates. Results: (i) LLM performance was tightly clustered (<15% variation) despite large cost differences; low-cost models performed comparably to top models. (ii) All LLMs significantly outperformed routine ward diagnoses on average diagnostic and safety scores. (iii) Top performance was achieved by GPT-5.1, followed by Gemini models. (vi) Adding radiology reports improved performance by 6%. (v) Diagnostic and reasoning scores were highly correlated ($ρ= 0.85$). (vi) Output rates varied (65-100%) due to input constraints. Results were robust across subsets and evaluation design. Conclusions: Across a real-world LMIC dataset, multimodal LLMs showed similar diagnostic performance despite large cost differences and outperformed routine care on average safety metrics. Affordability, robustness, and deployment constraints may outweigh marginal performance differences in LMIC settings.

Frontier language models are being deployed into clinical workflows faster than the infrastructure to evaluate them safely. Static medical-QA benchmarks miss the failure modes that matter in emergency medicine: trajectory-level safety collapse, tool misuse, and capitulation under sustained clinical pressure. We present HealthCraft, the first public reinforcement-learning environment that rewards trajectory-level safety under realistic emergency-medicine conditions, adapted from Corecraft. It is built on a FHIR R4 world state with 14 entity types and 3,987 seed entities, exposes 24 MCP tools, and defines a dual-layer rubric that zeroes reward whenever any safety-critical criterion is violated. We release 195 tasks across six categories, graded against 2,255 binary criteria (515 safety-critical); a post-hoc 10-task negative-class slate extends this to 205 tasks and 2,337 criteria. V8 results on two frontier models show Claude Opus 4.6 at Pass@1 24.8% [21.5-28.4] and GPT-5.4 at 12.6% [10.2-15.6], with safety-failure rates of 27.5% and 34.0%. On multi-step workflows - the closest proxy to real emergency care - performance collapses to near zero (Claude 1.0%, GPT-5.4 0.0%) despite partial competence on individual steps. Six infrastructure bugs fixed between pilots v2 and v8 re-ordered which model "looks stronger," evidence that infrastructure fidelity is part of the measurement. A deterministic LLM-judge overlay bounds evaluator noise, and a 60-run negative-class smoke pilot shows the reward signal is not drop-in training-safe: restraint criteria pass at 0.929 prevalence, a gameability an eval harness can tolerate but a training reward cannot. We scaffold coupling to a Megatron+SGLang+GRPO loop per Corecraft Section 5.2 and leave training-reward ablations as future work. Environment, tasks, rubrics, and harness are released under Apache 2.0.

Continuous glucose monitoring (CGM) is central to diabetes care, but explaining CGM patterns clearly and empathetically remains time-intensive. Evidence for retrieval-grounded large language model (LLM) systems in CGM-informed counseling remains limited. To evaluate whether a retrieval-grounded LLM-based conversational agent (CA) could support patient understanding of CGM data and preparation for routine diabetes consultations. We developed a retrieval-grounded LLM-based CA for CGM interpretation and diabetes counseling support. The system generated plain-language responses while avoiding individualized therapeutic advice. Twelve CGM-informed cases were constructed from publicly available datasets. Between Oct 2025 and Feb 2026, 6 senior UK diabetes clinicians each reviewed 2 assigned cases and answered 24 questions. In a blinded multi-rater evaluation, each CA-generated and clinician-authored response was independently rated by 3 clinicians on 6 quality dimensions. Safety flags and perceived source labels were also recorded. Primary analyses used linear mixed-effects models. A total of 288 unique responses (144 CA and 144 clinician) generated 864 ratings. The CA received higher quality scores than clinician responses (mean 4.37 vs 3.58), with an estimated mean difference of 0.782 points (95% CI 0.692-0.872; P<.001). The largest differences were for empathy (1.062, 95% CI 0.948-1.177) and actionability (0.992, 95% CI 0.877-1.106). Safety flag distributions were similar, with major concerns rare in both groups (3/432, 0.7% each). Retrieval-grounded LLM systems may have value as adjunct tools for CGM review, patient education, and preconsultation preparation. However, these findings do not support autonomous therapeutic decision-making or unsupervised real-world use.

Evaluating medical AI systems using expert clinician panels is costly and slow, motivating the use of large language models (LLMs) as alternative adjudicators. Here, we evaluate an LLM Jury, composed of three frontier AI models, for scoring 3334 diagnoses on 300 real-world low- and middle-income country (LMIC) hospital cases. Both LLM- and clinician-generated diagnoses are scored against expert panel diagnoses across four dimensions: diagnosis, differential diagnosis, clinical reasoning, and negative treatment risk. The LLM Jury scores are compared with expert and independent re-scoring panel scores to assess error metrics, inter-rater agreement, severe-risk errors, and the effect of post hoc calibration using isotonic regression. In our data, we find that: (i) the uncalibrated LLM Jury scores preserve ordinal agreement with the expert clinician panel scores, but are systematically lower; (ii) the probability of severe-risk errors is lower for the LLM Jury than the human expert re-score panels; (iii) the LLM Jury combined with LLM diagnoses can be used to identify diagnoses at high risk of error, enabling targeted expert review and improved panel efficiency; (iv) the calibrated LLM Jury scores and rankings of diagnosing agents show excellent agreement with those of the primary expert panels; (v) LLM Jury models show no self-preference bias, they did not score diagnoses generated by their own underlying model or models from the same vendor more (or less) favourably than those generated by other models. Together, these results provide evidence that a calibrated LLM Jury is a trustworthy and reliable proxy for expert clinician evaluation in medical AI benchmarking. Confirming these findings in other clinical settings is an important direction for future work.

Evaluating large language models (LLMs) for clinical documentation tasks such as SOAP note generation remains challenging. Unlike standard summarization, these tasks require clinical abstraction, normalization of colloquial language, and medically grounded inference. However, prevailing evaluation methods including automated metrics and LLM as judge frameworks rely on lexical faithfulness, often labeling any information not explicitly present in the transcript as hallucination. We show that such approaches systematically misclassify clinically valid outputs as errors, inflating hallucination rates and distorting model assessment. Our analysis reveals that many flagged hallucinations correspond to legitimate clinical transformations, including synonym mapping, abstraction of examination findings, diagnostic inference, and guideline consistent care planning. By aligning evaluation criteria with clinical reasoning through calibrated prompting and retrieval grounded in medical ontologies we observe a significant shift in outcomes. Under a lexical evaluation regime, the mean hallucination rate is 35%, heavily penalizing valid reasoning. With inference aware evaluation, this drops to 9%, with remaining cases reflecting genuine safety concerns. These findings suggest that current evaluation practices over penalize valid clinical reasoning and may measure artifacts of evaluation design rather than true errors, underscoring the need for clinically informed evaluation in high context domains like medicine.