patient digital twin simulation

This study addresses the limited interpretability and decision-support capability of existing diabetes prediction models regarding intervention effects. The authors propose the first proof-of-concept digital twin framework tailored for clinical decision-making, integrating real-world clinical data, synthetically augmented longitudinal features, and continuous glucose monitoring analytics. By leveraging simulation-driven modeling, the framework generates interpretable glycemic trajectories and enables counterfactual intervention reasoning. This work pioneers the integration of counterfactual simulation with predictive modeling, demonstrating feasibility on both public datasets and controlled synthetic scenarios. The results highlight the potential of unified prediction and intervention simulation, laying the groundwork for future explainable, decision-oriented medical digital twin systems.

This work addresses the significant challenge of simulating clinical patient trajectories, which are shaped by complex biological and social factors, thereby hindering advances in personalized medicine and virtual clinical trials. To this end, we leverage over 200 million real-world electronic health records to develop the first large-scale, pre-trained generative simulator capable of modeling the probabilistic distribution of future clinical events, laboratory results, and their temporal dynamics based solely on a patient’s historical data. The generated trajectories exhibit high fidelity to real-world observations, with incidence rates, lab values, and temporal patterns closely matching empirical data. Notably, the observed-to-expected ratios for diverse clinical outcomes consistently approximate 1.0, demonstrating the model’s effectiveness and potential for high-fidelity patient trajectory simulation.

This study addresses the challenge of enabling clinical decision support systems to safely and dynamically adapt to evolving patient conditions in real time while optimizing treatment strategies. The authors propose the first online adaptive AI framework that integrates patient-specific digital twins with reinforcement learning, leveraging treatment effect estimation, a pre-trained outcome prediction model, and a rule-based engine to generate personalized sequential decisions. To ensure clinical safety, the framework incorporates an expert review mechanism that triggers human oversight only when necessary. Experimental results on both a synthetic simulator and the TCGA ovarian cancer dataset demonstrate that the proposed system significantly outperforms baseline approaches, achieves low decision latency, and requires manual intervention in only a small fraction of cases, thereby validating its efficacy and clinical feasibility.

This study addresses the challenge of balancing interpretability and scalability in the clinical deployment of cardiovascular digital twins. To this end, it proposes a hybrid modeling paradigm that integrates physics-based mechanistic models, data-driven approaches, and physics-informed graph neural networks, all unified through data assimilation to enable dynamic, patient-specific vascular network modeling. The work systematically traces the evolution of modeling paradigms in this domain, identifies key barriers to clinical translation, and establishes a technical roadmap alongside a validation framework for developing digital twins that simultaneously achieve physiological interpretability, computational efficiency, and clinical applicability.

This study addresses the challenge of dynamic individual health modeling for precision medicine by proposing a real-time updatable and privacy-preserving Human Digital Twin (HDT) framework. Methodologically, it integrates heterogeneous multimodal data—including clinical, physiological, behavioral, and environmental sources—using lightweight machine learning, multimodal fusion modeling, and edge-based anomaly detection, while embedding differential privacy and federated learning to ensure data security and compliance. Its key contributions are: (i) the first clinical deployment of a closed-loop, self-updating HDT enabling high-accuracy dynamic health trajectory prediction; and (ii) an integrated intelligent system supporting personalized diagnosis, treatment planning, and early warning. Experimental results demonstrate significant improvements: +12.3% in chronic disease progression prediction accuracy and 72-hour earlier detection of critical health anomalies. The framework provides a scalable, regulatory-compliant technical pathway toward operational intelligent healthcare systems.

This work addresses the lack of standardization in existing patient simulation methods, which exhibit incompatible data formats, prompt templates, and evaluation metrics, thereby severely hindering reproducibility and fair comparison. To overcome this limitation, we propose the first standardized and modular patient simulation framework that unifies the definition, composition, and deployment of simulated patients, enabling cross-method evaluation and seamless integration of custom metrics. Built upon a large language model–based role-playing architecture, the framework offers high flexibility and interoperability with diverse simulation strategies and assessment mechanisms. We demonstrate its effectiveness by successfully reproducing multiple representative approaches and rapidly developing two novel simulators, thereby validating the framework’s extensibility and capacity to accelerate research and development. The code is publicly released.

This work addresses the risk that unvalidated online decision algorithms may reduce user engagement in mobile health interventions, highlighting the urgent need for high-fidelity simulation environments for pre-deployment evaluation. The authors propose JITAI-Twins, a digital twin framework tailored to target subpopulations, which introduces conditional time-series diffusion models to mobile health for the first time. Through a three-stage pipeline—pre-training on observational data, fine-tuning with intervention data from related populations, and expert-guided inference-time calibration—the framework generates temporally consistent and individual-difference-sensitive dynamic simulations. JITAI-Twins supports multi-source information fusion and cross-population transfer, significantly outperforming baseline simulators in the HeartSteps v2–v4 studies by more accurately reproducing the temporal behavioral patterns of target populations, thereby providing a reliable testbed for just-in-time adaptive intervention algorithms.

This study addresses the challenge of reliably estimating treatment effects in single-arm clinical trials due to the absence of a concurrent control group. The authors propose a machine learning–based digital twin approach that constructs a synthetic control arm by generating personalized predictions of disease progression for untreated patients. Integrating doubly robust estimation with the U.S. FDA’s artificial intelligence/ML guidance principles, the method leverages historical data for model development and facilitates sample size calculations. Reanalysis of real-world trial data in amyotrophic lateral sclerosis and Huntington’s disease demonstrates that the proposed framework substantially improves the accuracy and robustness of treatment effect estimation, offering a flexible and scalable paradigm for causal inference in single-arm trials.

This work addresses a critical limitation of existing static clinical prediction models, which conflate disease biology with clinician behavior and struggle to account for treatment feedback, time-varying confounding, and non-random observation patterns. The authors propose the first unified framework for intervention-aware disease trajectory modeling that seamlessly integrates discrete and continuous time. This framework jointly models disease progression, treatment assignment, and observation processes, enabling both factual prediction and counterfactual inference as well as policy evaluation. It synthesizes multi-state/joint models, temporal point processes, deep sequential architectures, and longitudinal causal inference, while incorporating overlapping diagnoses, uncertainty quantification, and target trial emulation. The approach captures individualized treatment-sensitive trajectories and supports pre-deployment stress testing of clinical strategies, thereby generating reliable, decision-grade evidence for safe implementation in closed-loop learning health systems.

This study addresses critical limitations in existing medical world models—particularly their deficiencies in causal reasoning, uncertainty quantification, and prospective validation—and proposes a systematic framework for building trustworthy models in clinical settings. Drawing on a comprehensive review of 1,455 publications and a focused analysis of 98 core studies, the work establishes the first clear theoretical boundaries and empirical criteria for such models, centering on four key capabilities: patient state representation, temporal dynamics modeling, intervention simulation, and physician-in-the-loop planning. Introducing clinical digital twins as an integrative paradigm, the framework incorporates structured narrative synthesis, multimodal physiological modeling, uncertainty calibration, and safety-constrained planning. The authors identify 14 rigorously defined empirical studies demonstrating preliminary feasibility in trajectory prediction and intervention comparison, while also highlighting key bottlenecks and pathways toward credible clinical deployment.