Improving large language models (LLMs) for electronic health record (EHR) reasoning is essential for enabling accurate and generalizable clinical predictions. While LLMs excel at medical text understanding, they underperform on EHR-based prediction tasks due to challenges in modeling temporally structured, high-dimensional data. Existing approaches often rely on hybrid paradigms, where LLMs serve merely as frozen prior retrievers while downstream deep learning (DL) models handle prediction, failing to improve the LLM’s intrinsic reasoning capacity and inheriting the generalization limitations of DL models. To this end, we propose EAG-RL, a novel two-stage training framework designed to intrinsically enhance LLMs’ EHR reasoning ability through expert attention guidance, where expert EHR models refer to task-specific DL models trained on EHR data. Concretely, EAG-RL first constructs high-quality, stepwise reasoning trajectories using expert-guided Monte Carlo Tree Search to effectively initialize the LLM’s policy. Then, EAG-RL further optimizes the policy via reinforcement learning by aligning the LLM’s attention with clinically salient features identified by expert EHR models. Extensive experiments on two real-world EHR datasets show that EAG-RL improves the intrinsic EHR reasoning ability of LLMs by an average of 14.62%, while also enhancing robustness to feature perturbations and generalization to unseen clinical domains. These results demonstrate the practical potential of EAG-RL for real-world deployment in clinical prediction tasks.

Hongxin Ding, Baixiang Huang, Yue Fang, Weibin Liao, Xinke Jiang, Jinyang Zhang, Yinghao Zhu, Zheng Li, Liantao Ma, Junfeng Zhao, Yasha Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Clinician skepticism toward opaque AI hinders adoption in high-stakes healthcare. We present AICare, an interactive and interpretable AI copilot for collaborative clinical decision-making. By analyzing longitudinal electronic health records, AICare grounds dynamic risk predictions in scrutable visualizations and LLM-driven diagnostic recommendations. Through a within-subjects counterbalanced study with 16 clinicians across nephrology and obstetrics, we comprehensively evaluated AICare using objective measures (task completion time and error rate), subjective assessments (NASA-TLX, SUS, and confidence ratings), and semi-structured interviews. Our findings indicate AICare’s reduced cognitive workload. Beyond performance metrics, qualitative analysis reveals that trust is actively constructed through verification, with interaction strategies diverging by expertise: junior clinicians used the system as cognitive scaffolding to structure their analysis, while experts engaged in adversarial verification to challenge the AI’s logic. This work offers design implications for creating AI systems that function as transparent partners, accommodating diverse reasoning styles to augment rather than replace clinical judgment.

Machine learning for electronic health records (EHR) is advancing rapidly and already underpins risk stratification, readmission and mortality prediction, and decision support, yet reliable translation still stalls on fragmented data pipelines, inconsistent medical-code handling, and hard-to-reproduce eval-uation-barriers that especially hinder R-centric clinical teams. Despite impressive methodological gains in temporal modeling, attention mechanisms, and strong classical baselines, most turnkey toolchains live in Python; as a result, many healthcare researchers and clinical data scientists working in R lack a single, integrated path from raw multi-table EHR to calibrated, auditable models. We address this gap with RHealth, an open-source, R-native toolkit that plays the role of an end-to-end conductor: from data harmonization and medical-code normal-ization to task specification, model training, and standardized reporting. Concretely, RHealth provides adapters for widely used public datasets (e.g., MIMIC-III/IV, eICU), utilities to traverse and map ICD-9/10 and CCS codes, task templates for common outcomes (mortality, 30-day readmission, length of stay), and a modeling stack that-at this development stage-supports standard recurrent baselines (e.g., RNN) and offers an extensible interface for user-defined architectures under active development, all evaluated with reproducible splits, AUROC/AUPRC, and cali-bration diagnostics. By packaging the full pipeline-from data to evaluation granularity-into modular, composable components, RHealth lowers the entry barrier for R users, reduces “glue code,” and promotes transparent, people-centric experimentation that can also serve as a trustworthy upstream substrate for LLM-enabled applications. To our knowledge, it is among the first comprehensive, integrated deep-learning toolkits for EHR in the R ecosystem. The code and documentation will be released after the double-blind review process.

Analyzing the health status of patients based on Electronic Health Records (EHR) is a fundamental research problem in medical informatics. The presence of extensive missing values in EHR makes it challenging for deep neural networks (DNNs) to directly model the patient's health status. Existing DNNs training protocols, including Impute-then-Regress Procedure and Jointly Optimizing of Impute-n-Regress Procedure, require the additional imputation models to reconstruction missing values. However, Impute-then-Regress Procedure introduces the risk of injecting imputed, non-real data into downstream clinical prediction tasks, resulting in power loss, biased estimation, and poorly performing models, while Jointly Optimizing of Impute-n-Regress Procedure is also difficult to generalize due to the complex optimization space and demanding data requirements. Inspired by the recent advanced literature of learnable prompt in the fields of NLP and CV, in this work, we rethought the necessity of the imputation model in downstream clinical tasks, and proposed Learnable Prompt as Pseudo-Imputation (PAI) as a new training protocol to assist EHR analysis. PAI no longer introduces any imputed data but constructs a learnable prompt to model the implicit preferences of the downstream model for missing values, resulting in a significant performance improvement for all state-of-the-arts EHR analysis models on four real-world datasets across two clinical prediction tasks. Further experimental analysis indicates that PAI exhibits higher robustness in situations of data insufficiency and high missing rates. More importantly, as a plug-and-play protocol, PAI can be easily integrated into any existing or even imperceptible future EHR analysis models. The code of this work is deployed publicly available at https://github.com/MrBlankness/PAI to help the research community reproduce the results and assist the EHR analysis tasks.

Medical Lay Language Generation (MLLG) plays a vital role in improving the accessibility of complex scientific content for broader audiences. Recent literature to MLLG commonly employ parameter-efficient fine-tuning methods such as Low-Rank Adaptation (LoRA) to fine-tuning large language models (LLMs) using paired expert-lay language datasets. However, LoRA struggles with the challenges posed by multi-source heterogeneous MLLG datasets. Specifically, through a series of exploratory experiments, we reveal that standard LoRA fail to meet the requirement for semantic fidelity and diverse lay-style generation in MLLG task. To address these limitations, we propose Magical, an asymmetric LoRA architecture tailored for MLLG under heterogeneous data scenarios. Magical employs a shared matrix A for abstractive summarization, along with multiple isolated matrices B for diverse lay-style generation. To preserve semantic fidelity during the lay language generation process, Magical introduces a Semantic Invariance Constraint to mitigate semantic subspace shifts on matrix A. Furthermore, to better adapt to diverse lay-style generation, Magical incorporates the Recommendation-guided Switch, an externally interface to prompt the LLM to switch between different matrices B. Experimental results on three real-world lay language generation datasets demonstrate that Magical consistently outperforms prompt-based methods, vanilla LoRA, and its recent variants, while also reducing trainable parameters by 31.66%. Our code is publicly available at https://github.com/tianlwang/Magical.git.

The rapid advancement of Large Language Models (LLMs) has stimulated interest in multi-agent collaboration for addressing complex medical tasks. However, the practical advantages of multi-agent collaboration approaches remain insufficiently understood. Existing evaluations often lack generalizability, failing to cover diverse tasks reflective of real-world clinical practice, and frequently omit rigorous comparisons against both single-LLM-based and established conventional methods. To address this critical gap, we introduce MedAgentBoard, a comprehensive benchmark for the systematic evaluation of multi-agent collaboration, single-LLM, and conventional approaches. MedAgentBoard encompasses four diverse medical task categories: (1) medical (visual) question answering, (2) lay summary generation, (3) structured Electronic Health Record (EHR) predictive modeling, and (4) clinical workflow automation, across text, medical images, and structured EHR data. Our extensive experiments reveal a nuanced landscape: while multi-agent collaboration demonstrates benefits in specific scenarios, such as enhancing task completeness in clinical workflow automation, it does not consistently outperform advanced single LLMs (e.g., in textual medical QA) or, critically, specialized conventional methods that generally maintain better performance in tasks like medical VQA and EHR-based prediction. MedAgentBoard offers a vital resource and actionable insights, emphasizing the necessity of a task-specific, evidence-based approach to selecting and developing AI solutions in medicine. It underscores that the inherent complexity and overhead of multi-agent collaboration must be carefully weighed against tangible performance gains. All code, datasets, detailed prompts, and experimental results are open-sourced at this link.

Recent studies have indicated that Large Language Models (LLMs) harbor an inherent understanding of truthfulness, yet often fail to consistently express it and generate false statements. This gap between ''knowing'' and ''telling'' poses a challenge for ensuring the truthfulness of generated content. Inspired by recent work on the practice of encoding human-interpretable concepts linearly within large language models, we treat truthfulness as a specially linearly encoded concept within LLMs, and introduce Adaptive Activation Steering (ACT), a tuning-free method that adaptively shifts LLM's activations in the ''truthful'' direction during inference. ACT addresses diverse categories of hallucinations by utilizing diverse truthfulness-related steering vectors and adjusting the steering intensity adaptively. Applied as an add-on across various models, ACT significantly improves truthfulness in LLaMA (↑142%), LLaMA2 (↑24%), Alpaca (↑36%), Vicuna (↑28%), LLaMA2-Chat (↑19%), and LLaMA3(↑34%). Furthermore, we verify ACT's scalability across larger models (13B, 33B, 65B), underscoring the adaptability of ACT to large-scale language models. Our code is available at https://github.com/tianlwang/ACT.

We introduce ColaCare, a framework that enhances Electronic Health Record (EHR) modeling through multi-agent collaboration driven by Large Language Models (LLMs). Our approach seamlessly integrates domain-specific expert models with LLMs to bridge the gap between structured EHR data and text-based reasoning. Inspired by the Multidisciplinary Team (MDT) approach used in clinical settings, ColaCare employs two types of agents: DoctorAgents and a MetaAgent, which collaboratively analyze patient data. Expert models process and generate predictions from numerical EHR data, while LLM agents produce reasoning references and decision-making reports within the MDT-driven collaborative consultation framework. The MetaAgent orchestrates the discussion, facilitating consultations and evidence-based debates among DoctorAgents, simulating diverse expertise in clinical decision-making. We additionally incorporate the Merck Manual of Diagnosis and Therapy (MSD) medical guideline within a retrieval-augmented generation (RAG) module for medical evidence support, addressing the challenge of knowledge currency. Extensive experiments conducted on three EHR datasets demonstrate ColaCare's superior performance in clinical mortality outcome and readmission prediction tasks, underscoring its potential to revolutionize clinical decision support systems and advance personalized precision medicine. All code, case studies and a questionnaire are available at the project website: https://colacare.netlify.app.

The integration of multimodal Electronic Health Records (EHR) data has significantly advanced clinical predictive capabilities. Existing models, which utilize clinical notes and multivariate time-series EHR data, often fall short of incorporating the necessary medical context for accurate clinical tasks, while previous approaches with knowledge graphs (KGs) primarily focus on structured knowledge extraction. In response, we propose EMERGE, a Retrieval-Augmented Generation (RAG) driven framework to enhance multimodal EHR predictive modeling. We extract entities from both time-series data and clinical notes by prompting Large Language Models (LLMs) and align them with professional PrimeKG, ensuring consistency. In addition to triplet relationships, we incorporate entities' definitions and descriptions for richer semantics. The extracted knowledge is then used to generate task-relevant summaries of patients' health statuses. Finally, we fuse the summary with other modalities using an adaptive multimodal fusion network with cross-attention. Extensive experiments on the MIMIC-III and MIMIC-IV datasets' in-hospital mortality and 30-day readmission tasks demonstrate the superior performance of the EMERGE framework over baseline models. Comprehensive ablation studies and analysis highlight the efficacy of each designed module and robustness to data sparsity. EMERGE contributes to refining the utilization of multimodal EHR data in healthcare, bridging the gap with nuanced medical contexts essential for informed clinical predictions. We have publicly released the code at https://github.com/yhzhu99/EMERGE.

Electronic Health Records (EHRs) provide valuable patient data but often suffer from sparsity issue, posing significant challenges in predictive modeling. Conventional imputation methods inadequately distinguish between real and imputed data, leading to potential inaccuracies of patient representations. To address these issues, we introduce PRISM, a framework that indirectly imputes data through prototype representations of similar patients, thus ensuring denser and more accurate embeddings. PRISM also includes a feature confidence learner module, which evaluates the reliability of each feature considering missing statuses. Additionally, it incorporates a new patient similarity metric that accounts for feature confidence, avoiding overreliance on imprecise imputed values. Our extensive experiments on the MIMIC-III, MIMIC-IV, PhysioNet Challenge 2012, eICU datasets demonstrate PRISM's superior performance in predicting in-hospital mortality and 30-day readmission tasks, showcasing its effectiveness in handling EHR data sparsity. For the sake of reproducibility and further research, we have publicly released the code at https://github.com/yhzhu99/PRISM.

Predicting health risks from electronic health records (EHR) is a topic of recent interest. Deep learning models have achieved success by modeling temporal and feature interaction. However, these methods learn insufficient representations and lead to poor performance when it comes to patients with few visits or sparse records. Inspired by the fact that doctors may compare the patient with typical patients and make decisions from similar cases, we propose a Progressive Prototypical Network (PPN) to select typical patients as prototypes and utilize their information to enhance the representation of the given patient. In particular, a progressive prototype memory and two prototype separation losses are proposed to update prototypes. Besides, a novel integration is introduced for better fusing information from patients and prototypes. Experiments on three real-world datasets demonstrate that our model brings improvement on all metrics. To make our results better understood by physicians, we developed an application at http://ppn.ai-care.top. Our code is released at https://github.com/yzhHoward/PPN.

KDD AIDSH 2024 aims to foster discussions and developments that push the boundaries of Artificial Intelligence (AI) and Data Science (DS) in healthcare, enhance diagnostic accuracy and promote human-centric approaches to healthcare, thus stimulating future interdisciplinary collaborations. This year's symposium will focus on expanding the application of AI/DS in healthcare/medicine and bridging existing gaps. The workshop invites submissions of full papers as well as work-in-progress on the application of AI/DS in healthcare. The workshop will feature three invited talks from eminent speakers, spanning academia, industry, and clinical researchers. In addition, selected papers will be invited to publish in Health Data Science, a Science Partner Journal. This summary provides a brief description of the half-day workshop to be held on August 26th, 2024. The webpage for the workshop can be found at https://aimel.ai/kdd2024aidsh.

Graph data augmentation has shown superiority in enhancing generalizability and robustness of GNNs in graph-level classifications. However, existing methods primarily focus on the augmentation in the graph signal space and the graph structure space independently, neglecting the joint interaction between them. In this paper, we address this limitation by formulating the problem as an optimal transport problem that aims to find an optimal inter-graph node matching strategy considering the interactions between graph structures and signals. To solve this problem, we propose a novel graph mixup algorithm called FGWMixup, which seeks a "midpoint" of source graphs in the Fused Gromov-Wasserstein (FGW) metric space. To enhance the scalability of our method, we introduce a relaxed FGW solver that accelerates FGWMixup by improving the convergence rate from $\mathcal{O}(t^{-1})$ to $\mathcal{O}(t^{-2})$. Extensive experiments conducted on five datasets using both classic (MPNNs) and advanced (Graphormers) GNN backbones demonstrate that \mname\xspace effectively improves the generalizability and robustness of GNNs. Codes are available at https://github.com/ArthurLeoM/FGWMixup.

Exploiting the correlations between medical features is essential to the success of healthcare data analysis. However, most existing methods are either suffering large estimation variance for data insufficiency or inflexible in terms of demanding task-specific medical knowledge. In this paper, we propose a novel patient health representation learning framework dubbedSAFARI.SAFARIlearns a compact representation by imposing a clinical-fact-inspired task-agnostic correlational sparsity prior to the correlations of medical feature pairs. Specifically, we learn the compact representation by solving the bi-level optimization problem, which involves solving the high-level inter-group correlations and the nested lower-level intra-group correlations. We leverage the Laplacian kernel as a robust metric for feature grouping and graph neural networks for solving the bi-level optimization problem following the optimal value reformulation paradigm. Experiments on five datasets of various inputs and tasks demonstrate the efficacy ofSAFARI. The discovered findings are also consistent with our insights and medical literature, which can provide valuable clinical explanations.

Multimodal electronic health record (EHR) data are widely used in clinical applications. Conventional methods usually assume that each sample (patient) is associated with the unified observed modalities, and all modalities are available for each sample. However, missing modality caused by various clinical and social reasons is a common issue in real-world clinical scenarios. Existing methods mostly rely on solving a generative model that learns a mapping from the latent space to the original input space, which is an unstable ill-posed inverse problem. To relieve the underdetermined system, we propose a model solving a direct problem, dubbed learning with Missing Modalities in Multimodal healthcare data (M3Care). M3Care is an end-to-end model compensating the missing information of the patients with missing modalities to perform clinical analysis. Instead of generating raw missing data, M3Care imputes the task-related information of the missing modalities in the latent space by the auxiliary information from each patient's similar neighbors, measured by a task-guided modality-adaptive similarity metric, and thence conducts the clinical tasks. The task-guided modality-adaptive similarity metric utilizes the uncensored modalities of the patient and the other patients who also have the same uncensored modalities to find similar patients. Experiments on real-world datasets show that M3Care outperforms the state-of-the-art baselines. Moreover, the findings discovered by M3Care are consistent with experts and medical knowledge, demonstrating the capability and the potential of providing useful insights and explanations.

Deep learning models have been applied to many healthcare tasks based on electronic medical records (EMR) data and shown substantial performance. Existing methods commonly embed the records of a single patient into a representation for medical tasks. Such methods learn inadequate representations and lead to inferior performance, especially when the patient’s data is sparse or low-quality. Aiming at the above problem, we propose GRASP, a generic framework for healthcare models. For a given patient, GRASP first finds patients in the dataset who have similar conditions and similar results (i.e., the similar patients), and then enhances the representation learning and prognosis of the given patient by leveraging knowledge extracted from these similar patients. GRASP defines similarities with different meanings between patients for different clinical tasks, and finds similar patients with useful information accordingly, and then learns cohort representation to extract valuable knowledge contained in the similar patients. The cohort information is fused with the current patient’s representation to conduct final clinical tasks. Experimental evaluations on two real-world datasets show that GRASP can be seamlessly integrated into state-of-the-art models with consistent performance improvements. Besides, under the guidance of medical experts, we verified the findings extracted by GRASP, and the findings are consistent with the existing medical knowledge, indicating that GRASP can generate useful insights for relevant predictions.

Due to the characteristics of COVID-19, the epidemic develops rapidly and overwhelms health service systems worldwide. Many patients suffer from life-threatening systemic problems and need to be carefully monitored in ICUs. An intelligent prognosis can help physicians take an early intervention, prevent adverse outcomes, and optimize the medical resource allocation, which is urgently needed, especially in this ongoing global pandemic crisis. However, in the early stage of the epidemic outbreak, the data available for analysis is limited due to the lack of effective diagnostic mechanisms, the rarity of the cases, and privacy concerns. In this paper, we propose a distilled transfer learning framework, which leverages the existing publicly available online Electronic Medical Records to enhance the prognosis for inpatients with emerging infectious diseases. It learns to embed the COVID-19-related medical features based on massive existing EMR data. The transferred parameters are further trained to imitate the teacher model’s representation based on distillation, which embeds the health status more comprehensively on the source dataset. We conduct Length-of-Stay prediction experiments for patients in ICUs on real-world COVID-19 datasets. The experiment results indicate that our proposed model consistently outperforms competitive baseline methods. In order to further verify the scalability of o deal with different clinical tasks on different EMR datasets, we conduct an additional mortality prediction experiment on End-Stage Renal Disease datasets. The extensive experiments demonstrate that an benefit the prognosis for emerging pandemics and other diseases with limited EMR.

Deep learning-based health status representation learning and clinical prediction have raised much research interest in recent years. Existing models have shown superior performance, but there are still several major issues that have not been fully taken into consideration. First, the historical variation pattern of the biomarker in diverse time scales plays a vital role in indicating the health status, but it has not been explicitly extracted by existing works. Second, key factors that strongly indicate the health risk are different among patients. It is still challenging to adaptively make use of the features for patients in diverse conditions. Third, using prediction models as the black box will limit the reliability in clinical practice. However, none of the existing works can provide satisfying interpretability and meanwhile achieve high prediction performance. In this work, we develop a general health status representation learning model, named AdaCare. It can capture the long and short-term variations of biomarkers as clinical features to depict the health status in multiple time scales. It also models the correlation between clinical features to enhance the ones which strongly indicate the health status and thus can maintain a state-of-the-art performance in terms of prediction accuracy while providing qualitative interpretability. We conduct a health risk prediction experiment on two real-world datasets. Experiment results indicate that AdaCare outperforms state-of-the-art approaches and provides effective interpretability, which is verifiable by clinical experts.

Predicting the patient's clinical outcome from the historical electronic medical records (EMR) is a fundamental research problem in medical informatics. Most deep learning-based solutions for EMR analysis concentrate on learning the clinical visit embedding and exploring the relations between visits. Although those works have shown superior performances in healthcare prediction, they fail to explore the personal characteristics during the clinical visits thoroughly. Moreover, existing works usually assume that the more recent record weights more in the prediction, but this assumption is not suitable for all conditions. In this paper, we propose ConCare to handle the irregular EMR data and extract feature interrelationship to perform individualized healthcare prediction. Our solution can embed the feature sequences separately by modeling the time-aware distribution. ConCare further improves the multi-head self-attention via the cross-head decorrelation, so that the inter-dependencies among dynamic features and static baseline information can be effectively captured to form the personal health context. Experimental results on two real-world EMR datasets demonstrate the effectiveness of ConCare. The medical findings extracted by ConCare are also empirically confirmed by human experts and medical literature.

This paper proposes a novel multi-task allocation framework, named PSAllocator, for participatory sensing (PS). Different from previous single-task oriented approaches, which select an optimal set of users for each single task independently, PSAllocator attempts to coordinate the allocation of multiple tasks to maximize the overall system utility on a multi-task PS platform. Furthermore, PSAllocator takes the maximum number of sensing tasks allowed for each participant and the sensor availability of each mobile device into consideration. PSAllocator utilizes a two-phase offline multi-task allocation approach to achieve the near-optimal goal. First, it predicts the participants' connections to cell towers and locations based on historical data from the telecom operator; Then, it converts the multi-task allocation problem into the representation of a bipartite graph, and employs an iterative greedy process to optimize the task allocation. Extensive evaluations based on real-world mobility traces show that PSAllocator outperforms the baseline methods under various settings.