Danielle S. Bitterman

Marco Guevara, Shan Chen, Spencer Thomas et al. · harvard

Social determinants of health (SDoH) have an important impact on patient outcomes but are incompletely collected from the electronic health records (EHR). This study researched the ability of large language models to extract SDoH from free text in EHRs, where they are most commonly documented, and explored the role of synthetic clinical text for improving the extraction of these scarcely documented, yet extremely valuable, clinical data. 800 patient notes were annotated for SDoH categories, and several transformer-based models were evaluated. The study also experimented with synthetic data generation and assessed for algorithmic bias. Our best-performing models were fine-tuned Flan-T5 XL (macro-F1 0.71) for any SDoH, and Flan-T5 XXL (macro-F1 0.70). The benefit of augmenting fine-tuning with synthetic data varied across model architecture and size, with smaller Flan-T5 models (base and large) showing the greatest improvements in performance (delta F1 +0.12 to +0.23). Model performance was similar on the in-hospital system dataset but worse on the MIMIC-III dataset. Our best-performing fine-tuned models outperformed zero- and few-shot performance of ChatGPT-family models for both tasks. These fine-tuned models were less likely than ChatGPT to change their prediction when race/ethnicity and gender descriptors were added to the text, suggesting less algorithmic bias (p<0.05). At the patient-level, our models identified 93.8% of patients with adverse SDoH, while ICD-10 codes captured 2.0%. Our method can effectively extracted SDoH information from clinic notes, performing better compare to GPT zero- and few-shot settings. These models could enhance real-world evidence on SDoH and aid in identifying patients needing social support.

Shan Chen, Yingya Li, Sheng Lu et al. · harvard

Recent advances in large language models (LLMs) have shown impressive ability in biomedical question-answering, but have not been adequately investigated for more specific biomedical applications. This study investigates the performance of LLMs such as the ChatGPT family of models (GPT-3.5s, GPT-4) in biomedical tasks beyond question-answering. Because no patient data can be passed to the OpenAI API public interface, we evaluated model performance with over 10000 samples as proxies for two fundamental tasks in the clinical domain - classification and reasoning. The first task is classifying whether statements of clinical and policy recommendations in scientific literature constitute health advice. The second task is causal relation detection from the biomedical literature. We compared LLMs with simpler models, such as bag-of-words (BoW) with logistic regression, and fine-tuned BioBERT models. Despite the excitement around viral ChatGPT, we found that fine-tuning for two fundamental NLP tasks remained the best strategy. The simple BoW model performed on par with the most complex LLM prompting. Prompt engineering required significant investment.

Shan Chen, Marco Guevara, Nicolas Ramirez et al. · harvard

Radiotherapy (RT) toxicities can impair survival and quality-of-life, yet remain under-studied. Real-world evidence holds potential to improve our understanding of toxicities, but toxicity information is often only in clinical notes. We developed natural language processing (NLP) models to identify the presence and severity of esophagitis from notes of patients treated with thoracic RT. We fine-tuned statistical and pre-trained BERT-based models for three esophagitis classification tasks: Task 1) presence of esophagitis, Task 2) severe esophagitis or not, and Task 3) no esophagitis vs. grade 1 vs. grade 2-3. Transferability was tested on 345 notes from patients with esophageal cancer undergoing RT. Fine-tuning PubmedBERT yielded the best performance. The best macro-F1 was 0.92, 0.82, and 0.74 for Task 1, 2, and 3, respectively. Selecting the most informative note sections during fine-tuning improved macro-F1 by over 2% for all tasks. Silver-labeled data improved the macro-F1 by over 3% across all tasks. For the esophageal cancer notes, the best macro-F1 was 0.73, 0.74, and 0.65 for Task 1, 2, and 3, respectively, without additional fine-tuning. To our knowledge, this is the first effort to automatically extract esophagitis toxicity severity according to CTCAE guidelines from clinic notes. The promising performance provides proof-of-concept for NLP-based automated detailed toxicity monitoring in expanded domains.

Shan Chen, Marco Guevara, Shalini Moningi et al.

Documentation burden is a major contributor to clinician burnout, which is rising nationally and is an urgent threat to our ability to care for patients. Artificial intelligence (AI) chatbots, such as ChatGPT, could reduce clinician burden by assisting with documentation. Although many hospitals are actively integrating such systems into electronic medical record systems, AI chatbots utility and impact on clinical decision-making have not been studied for this intended use. We are the first to examine the utility of large language models in assisting clinicians draft responses to patient questions. In our two-stage cross-sectional study, 6 oncologists responded to 100 realistic synthetic cancer patient scenarios and portal messages developed to reflect common medical situations, first manually, then with AI assistance. We find AI-assisted responses were longer, less readable, but provided acceptable drafts without edits 58% of time. AI assistance improved efficiency 77% of time, with low harm risk (82% safe). However, 7.7% unedited AI responses could severely harm. In 31% cases, physicians thought AI drafts were human-written. AI assistance led to more patient education recommendations, fewer clinical actions than manual responses. Results show promise for AI to improve clinician efficiency and patient care through assisting documentation, if used judiciously. Monitoring model outputs and human-AI interaction remains crucial for safe implementation.

Xiaoye Wang, Nicole Xi Zhang, Hongyu He et al.

Recent advancements in artificial intelligence (AI), particularly in large language models (LLMs), have unlocked significant potential to enhance the quality and efficiency of medical care. By introducing a novel way to interact with AI and data through natural language, LLMs offer new opportunities for medical practitioners, patients, and researchers. However, as AI and LLMs become more powerful and especially achieve superhuman performance in some medical tasks, public concerns over their safety have intensified. These concerns about AI safety have emerged as the most significant obstacles to the adoption of AI in medicine. In response, this review examines emerging risks in AI utilization during the LLM era. First, we explore LLM-specific safety challenges from functional and communication perspectives, addressing issues across data collection, model training, and real-world application. We then consider inherent safety problems shared by all AI systems, along with additional complications introduced by LLMs. Last, we discussed how safety issues of using AI in clinical practice and healthcare system operation would undermine trust among patient, clinicians and the public, and how to build confidence in these systems. By emphasizing the development of safe AI, we believe these technologies can be more rapidly and reliably integrated into everyday medical practice to benefit both patients and clinicians.

Jack Gallifant, Shan Chen, Kuleen Sasse et al.

Sparse Autoencoders (SAEs) provide potentials for uncovering structured, human-interpretable representations in Large Language Models (LLMs), making them a crucial tool for transparent and controllable AI systems. We systematically analyze SAE for interpretable feature extraction from LLMs in safety-critical classification tasks. Our framework evaluates (1) model-layer selection and scaling properties, (2) SAE architectural configurations, including width and pooling strategies, and (3) the effect of binarizing continuous SAE activations. SAE-derived features achieve macro F1 > 0.8, outperforming hidden-state and BoW baselines while demonstrating cross-model transfer from Gemma 2 2B to 9B-IT models. These features generalize in a zero-shot manner to cross-lingual toxicity detection and visual classification tasks. Our analysis highlights the significant impact of pooling strategies and binarization thresholds, showing that binarization offers an efficient alternative to traditional feature selection while maintaining or improving performance. These findings establish new best practices for SAE-based interpretability and enable scalable, transparent deployment of LLMs in real-world applications. Full repo: https://github.com/shan23chen/MOSAIC.

Bingyang Ye, Shan Chen, Jingxuan Tu et al.

Large language models are increasingly being used to assess and forecast research ideas, yet we lack scalable ways to evaluate the quality of models' judgments about these scientific ideas. Towards this goal, we introduce PoT, a semi-verifiable benchmarking framework that links scientific idea judgments to downstream signals that become observable later (e.g., citations and shifts in researchers' agendas). PoT freezes a pre-cutoff snapshot of evidence in an offline sandbox and asks models to forecast post-cutoff outcomes, enabling verifiable evaluation when ground truth arrives, scalable benchmarking without exhaustive expert annotation, and analysis of human-model misalignment against signals such as peer-review awards. In addition, PoT provides a controlled testbed for agent-based research judgments that evaluate scientific ideas, comparing tool-using agents to non-agent baselines under prompt ablations and budget scaling. Across 30,000+ instances spanning four benchmark domains, we find that, compared with non-agent baselines, higher interaction budgets generally improve agent performance, while the benefit of tool use is strongly task-dependent. By combining time-partitioned, future-verifiable targets with an offline sandbox for tool use, PoT supports scalable evaluation of agents on future-facing scientific idea judgment tasks.

Katherine C. Kellogg, Bingyang Ye, Yifan Hu et al.

The rapid adoption of large language models (LLMs) in healthcare has been accompanied by scrutiny of their oversight. Existing monitoring approaches, inherited from traditional machine learning (ML), are task-based and founded on assumed performance degradation arising from dataset drift. In contrast, with LLMs, inevitable model degradation due to changes in populations compared to the training dataset cannot be assumed, because LLMs were not trained for any specific task in any given population. We therefore propose a new organizing principle guiding generalist LLM monitoring that is scalable and grounded in how these models are developed and used in practice: capability-based monitoring. Capability-based monitoring is motivated by the fact that LLMs are generalist systems whose overlapping internal capabilities are reused across numerous downstream tasks. Instead of evaluating each downstream task independently, this approach organizes monitoring around shared model capabilities, such as summarization, reasoning, translation, or safety guardrails, in order to enable cross-task detection of systemic weaknesses, long-tail errors, and emergent behaviors that task-based monitoring may miss. We describe considerations for developers, organizational leaders, and professional societies for implementing a capability-based monitoring approach. Ultimately, capability-based monitoring will provide a scalable foundation for safe, adaptive, and collaborative monitoring of LLMs and future generalist artificial intelligence models in healthcare.

Mingxuan Liu, Yuhe Ke, Wentao Zhu et al.

The integration of large language models (LLMs) into healthcare holds promise to enhance clinical decision-making, yet their susceptibility to biases remains a critical concern. Gender has long influenced physician behaviors and patient outcomes, raising concerns that LLMs assuming human-like roles, such as clinicians or medical educators, may replicate or amplify gender-related biases. Using case studies from the New England Journal of Medicine Challenge (NEJM), we assigned genders (female, male, or unspecified) to multiple open-source and proprietary LLMs. We evaluated their response consistency across LLM-gender assignments regarding both LLM-based diagnosis and models' judgments on the clinical relevance or necessity of patient gender. In our findings, diagnoses were relatively consistent across LLM genders for most models. However, for patient gender's relevance and necessity in LLM-based diagnosis, all models demonstrated substantial inconsistency across LLM genders, particularly for relevance judgements. Some models even displayed a systematic female-male disparity in their interpretation of patient gender. These findings present an underexplored bias that could undermine the reliability of LLMs in clinical practice, underscoring the need for routine checks of identity-assignment consistency when interacting with LLMs to ensure reliable and equitable AI-supported clinical care.

Jirui Qi, Shan Chen, Zidi Xiong et al.

Recent Large Reasoning Models (LRMs) with thinking traces have shown strong performance on English reasoning tasks. However, their ability to think in other languages is less studied. This capability is as important as answer accuracy for real world applications because users may find the reasoning trace useful for oversight only when it is expressed in their own language. We comprehensively evaluate two leading families of LRMs on our XReasoning benchmark and find that even the most advanced models often revert to English or produce fragmented reasoning in other languages, revealing a substantial gap in multilingual reasoning. Prompt based interventions that force models to reason in the users language improve readability and oversight but reduce answer accuracy, exposing an important trade off. We further show that targeted post training on just 100 examples mitigates this mismatch, though some accuracy loss remains. Our results highlight the limited multilingual reasoning capabilities of current LRMs and outline directions for future work. Code and data are available at https://github.com/Betswish/mCoT-XReasoning.

Natalie Seah, Danielle S. Bitterman, Daphna Spiegel et al.

Accurately communicating the side effects of cancer treatments to cancer survivors is critical, particularly in settings such as informed consent, where clinicians must clearly and comprehensively convey potential treatment toxicities. However, this task remains challenging due to clinical knowledge deficits about adverse treatment effects and fragmentation across electronic health record (EHR) systems. Large language models (LLMs) have the potential to assist in this task, though their reliability in oncology survivorship contexts remains poorly understood. We present a deployment-oriented stress-testing framework for evaluating LLM-generated radiation side effect lists in breast cancer treatment and survivorship care. Using 21 breast cancer patient profiles, we construct paired patient clinical scenarios that differ only in radiotherapy regimens to evaluate seven instruction-tuned LLMs under multiple prompting regimes. We then compare LLM outputs to a clinician-curated reference derived from informed consent documents at two major academic medical centers and developed by a team including more than seven breast radiation oncologists. The reference maps radiation dose-fractionation, fields, and locations to associated toxicities, broken down by frequency and temporal onset. Across models, we reveal sensitivity to minor documentation changes, trade-offs between precision and recall, and systematic under-recall of rare and long-term side effects. When used alone, constraints on the number of side effects generated reduce precision, and grounding outputs in clinician-curated side effect lists substantially improves reliability and robustness. These findings highlight important limitations of LLM use in oncology and suggest practical design choices for safer and more informative survivorship-focused applications.

João Matos, Shan Chen, Siena Placino et al.

Multimodal/vision language models (VLMs) are increasingly being deployed in healthcare settings worldwide, necessitating robust benchmarks to ensure their safety, efficacy, and fairness. Multiple-choice question and answer (QA) datasets derived from national medical examinations have long served as valuable evaluation tools, but existing datasets are largely text-only and available in a limited subset of languages and countries. To address these challenges, we present WorldMedQA-V, an updated multilingual, multimodal benchmarking dataset designed to evaluate VLMs in healthcare. WorldMedQA-V includes 568 labeled multiple-choice QAs paired with 568 medical images from four countries (Brazil, Israel, Japan, and Spain), covering original languages and validated English translations by native clinicians, respectively. Baseline performance for common open- and closed-source models are provided in the local language and English translations, and with and without images provided to the model. The WorldMedQA-V benchmark aims to better match AI systems to the diverse healthcare environments in which they are deployed, fostering more equitable, effective, and representative applications.

Shan Chen, Pedro Moreira, Yuxin Xiao et al.

Large language models (LLMs) are increasingly envisioned as decision-support tools in clinical practice, yet safe clinical reasoning demands integrating heterogeneous knowledge bases -- trials, primary studies, regulatory documents, and cost data -- under strict accuracy constraints. Existing evaluations often rely on synthetic prompts, reduce the task to single-hop factoid queries, or conflate reasoning with open-ended generation, leaving their real-world utility unclear. To close this gap, we present MedBrowseComp, the first benchmark that systematically tests an agent's ability to reliably retrieve and synthesize multi-hop medical facts from live, domain-specific knowledge bases. MedBrowseComp contains more than 1,000 human-curated questions that mirror clinical scenarios where practitioners must reconcile fragmented or conflicting information to reach an up-to-date conclusion. Applying MedBrowseComp to frontier agentic systems reveals performance shortfalls as low as ten percent, exposing a critical gap between current LLM capabilities and the rigor demanded in clinical settings. MedBrowseComp therefore offers a clear testbed for reliable medical information seeking and sets concrete goals for future model and toolchain upgrades. You can visit our project page at: https://moreirap12.github.io/mbc-browse-app/

Mingye Gao, Aman Varshney, Shan Chen et al.

Cancer clinical trials often face challenges in recruitment and engagement due to a lack of participant-facing informational and educational resources. This study investigated the potential of Large Language Models (LLMs), specifically GPT4, in generating patient-friendly educational content from clinical trial informed consent forms. Using data from ClinicalTrials.gov, we employed zero-shot learning for creating trial summaries and one-shot learning for developing multiple-choice questions, evaluating their effectiveness through patient surveys and crowdsourced annotation. Results showed that GPT4-generated summaries were both readable and comprehensive, and may improve patients' understanding and interest in clinical trials. The multiple-choice questions demonstrated high accuracy and agreement with crowdsourced annotators. For both resource types, hallucinations were identified that require ongoing human oversight. The findings demonstrate the potential of LLMs "out-of-the-box" to support the generation of clinical trial education materials with minimal trial-specific engineering, but implementation with a human-in-the-loop is still needed to avoid misinformation risks.

Shan Chen, Jack Gallifant, Marco Guevara et al.

Generative models have been showing potential for producing data in mass. This study explores the enhancement of clinical natural language processing performance by utilizing synthetic data generated from advanced language models. Promising results show feasible applications in such a high-stakes domain.

Mingxuan Liu, Yilin Ning, Haoyuan Wang et al.

As machine learning models become increasingly integrated into healthcare, structural inequities and social biases embedded in clinical data can be perpetuated or even amplified by data-driven models. In survival analysis, censoring and time dynamics can further add complexity to fair model development. Additionally, algorithmic fairness approaches often overlook disparities in cross-group rankings, e.g., high-risk Black patients may be ranked below lower-risk White patients who do not experience the event of mortality. Such misranking can reinforce biological essentialism and undermine equitable care. We propose a Fairness-Aware Survival Modeling (FASM), designed to mitigate algorithmic bias regarding both intra-group and cross-group risk rankings over time. Using breast cancer prognosis as a representative case and applying FASM to SEER breast cancer data, we show that FASM substantially improves fairness while preserving discrimination performance comparable to fairness-unaware survival models. Time-stratified evaluations show that FASM maintains stable fairness over a 10-year horizon, with the greatest improvements observed during the mid-term of follow-up. Our approach enables the development of survival models that prioritize both accuracy and equity in clinical decision-making, advancing fairness as a core principle in clinical care.

Jack Gallifant, Katherine C. Kellogg, Matt Butler et al.

Large language models (LLMs) integrated into agent-driven workflows hold immense promise for healthcare, yet a significant gap exists between their potential and practical implementation within clinical settings. To address this, we present a practitioner-oriented field manual for deploying generative agents that use electronic health record (EHR) data. This guide is informed by our experience deploying the "irAE-Agent", an automated system to detect immune-related adverse events from clinical notes at Mass General Brigham, and by structured interviews with 20 clinicians, engineers, and informatics leaders involved in the project. Our analysis reveals a critical misalignment in clinical AI development: less than 20% of our effort was dedicated to prompt engineering and model development, while over 80% was consumed by the sociotechnical work of implementation. We distill this effort into five "heavy lifts": data integration, model validation, ensuring economic value, managing system drift, and governance. By providing actionable solutions for each of these challenges, this field manual shifts the focus from algorithmic development to the essential infrastructure and implementation work required to bridge the "valley of death" and successfully translate generative AI from pilot projects into routine clinical care.

Fridolin Haugg, Grace Lee, John He et al.

Background: Facial appearance offers a noninvasive window into health. We built FAHR-Face, a foundation model trained on >40 million facial images and fine-tuned it for two distinct tasks: biological age estimation (FAHR-FaceAge) and survival risk prediction (FAHR-FaceSurvival). Methods: FAHR-FaceAge underwent a two-stage, age-balanced fine-tuning on 749,935 public images; FAHR-FaceSurvival was fine-tuned on 34,389 photos of cancer patients. Model robustness (cosmetic surgery, makeup, pose, lighting) and independence (saliency mapping) was tested extensively. Both models were clinically tested in two independent cancer patient datasets with survival analyzed by multivariable Cox models and adjusted for clinical prognostic factors. Findings: For age estimation, FAHR-FaceAge had the lowest mean absolute error of 5.1 years on public datasets, outperforming benchmark models and maintaining accuracy across the full human lifespan. In cancer patients, FAHR-FaceAge outperformed a prior facial age estimation model in survival prognostication. FAHR-FaceSurvival demonstrated robust prediction of mortality, and the highest-risk quartile had more than triple the mortality of the lowest (adjusted hazard ratio 3.22; P<0.001). These findings were validated in the independent cohort and both models showed generalizability across age, sex, race and cancer subgroups. The two algorithms provided distinct, complementary prognostic information; saliency mapping revealed each model relied on distinct facial regions. The combination of FAHR-FaceAge and FAHR-FaceSurvival improved prognostic accuracy. Interpretation: A single foundation model can generate inexpensive, scalable facial biomarkers that capture both biological ageing and disease-related mortality risk. The foundation model enabled effective training using relatively small clinical datasets.

David Restrepo, Chenwei Wu, Constanza Vásquez-Venegas et al.

The deployment of large language models (LLMs) in healthcare has demonstrated substantial potential for enhancing clinical decision-making, administrative efficiency, and patient outcomes. However, the underrepresentation of diverse groups in the development and application of these models can perpetuate biases, leading to inequitable healthcare delivery. This paper presents a comprehensive scientometric analysis of LLM research for healthcare, including data from January 1, 2021, to July 1, 2024. By analyzing metadata from PubMed and Dimensions, including author affiliations, countries, and funding sources, we assess the diversity of contributors to LLM research. Our findings highlight significant gender and geographic disparities, with a predominance of male authors and contributions primarily from high-income countries (HICs). We introduce a novel journal diversity index based on Gini diversity to measure the inclusiveness of scientific publications. Our results underscore the necessity for greater representation in order to ensure the equitable application of LLMs in healthcare. We propose actionable strategies to enhance diversity and inclusivity in artificial intelligence research, with the ultimate goal of fostering a more inclusive and equitable future in healthcare innovation.

Shan Chen, Jack Gallifant, Mingye Gao et al.

Large language models (LLMs) are increasingly essential in processing natural languages, yet their application is frequently compromised by biases and inaccuracies originating in their training data. In this study, we introduce Cross-Care, the first benchmark framework dedicated to assessing biases and real world knowledge in LLMs, specifically focusing on the representation of disease prevalence across diverse demographic groups. We systematically evaluate how demographic biases embedded in pre-training corpora like $ThePile$ influence the outputs of LLMs. We expose and quantify discrepancies by juxtaposing these biases against actual disease prevalences in various U.S. demographic groups. Our results highlight substantial misalignment between LLM representation of disease prevalence and real disease prevalence rates across demographic subgroups, indicating a pronounced risk of bias propagation and a lack of real-world grounding for medical applications of LLMs. Furthermore, we observe that various alignment methods minimally resolve inconsistencies in the models' representation of disease prevalence across different languages. For further exploration and analysis, we make all data and a data visualization tool available at: www.crosscare.net.