Loading
Life sciences · Preprint
arXiv · September 8, 2026
Raises a question worth testing. It does not answer one.
NOAH is a generative transformer model trained on retrospective multimodal clinical data from MIMIC to represent and forecast patient trajectories. The work demonstrates technical capability to process diverse data types and perform zero-shot classification and time-to-event prediction, but contains no prospective validation, clinical outcomes data, or comparison to established standards. This is a machine learning systems paper with no evidence that the model improves clinical decision-making or patient outcomes.
Retrospective machine learning model development and evaluation study. Historical patient records from MIMIC dataset family (299,000 patients, 431,000 hospital visits); retrospective database only, no prospective recruitment.. Intervention: NOAH generative transformer model with bidirectional time integration and variational latent space for multimodal patient data representation and forecasting.
Model trained on 559 million clinical events from 431,000 hospital visits of 299,000 patients NOAH processes medical images, time-series signals, numeric data, categorical events, and unstructured clinical records Model demonstrates capability in probing for 15 ICD chapters and 29 comorbidities, and time-to-event prediction
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
Clinicians and researchers should recognize this as a technical proof-of-concept that does not yet establish clinical utility, safety, or superiority over existing methods. Prospective validation in independent clinical settings with hard clinical endpoints would be necessary before considering implementation in clinical care.
This is a machine learning model development study with no clinical validation, comparative effectiveness data, or prospective clinical outcomes—it demonstrates technical capability on retrospective data but does not establish clinical utility or safety.
As stated by the source record.
Quoted from the source exactly as published.
Clinicians and researchers should recognize this as a technical proof-of-concept that does not yet establish clinical utility, safety, or superiority over existing methods. Prospective validation in independent clinical settings with hard clinical endpoints would be necessary before considering implementation in clinical care.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
The digitization of healthcare has generated vast, longitudinal, and multimodal patient records over a lifetime, yet fully exploiting these data to represent and predict patient state trajectories remains a critical challenge. Current AI models often struggle to capture the complex, irregular temporal dynamics and inherent stochasticity of real-world multimodal patient data. Existing AI approaches for modeling longitudinal patient records are predominantly discriminative, limited to a few modalities, constrained by closed categorical vocabularies, treating time as a monotonic inductive bias, or they are limited in forecasting future patient states. We introduce NOAH, a time-aware, task-agnostic, generative transformer model representing and forecasting the full multimodal patient journey. NOAH features a novel bidirectional time integration and a variational latent space to capture the continuous evolution of patient states and the stochasticity of clinical trajectories. Built from over 559 million clinical events from 431,000 hospital visits of 299,000 patients across the MIMIC dataset family, NOAH natively processes medical images, time-series and numeric signals, categorical events, as well as structured and unstructured clinical records. NOAH is the first truly holistic generative model in its field, enabling autoregressive forecasting with optional time control, zero-shot classification, and counterfactual intervention simulation. It generates highly informative and predictive patient state representations that demonstrate strong performance in probing for clinical outcomes, 15 ICD chapters, and 29 comorbidities, as well as in time-to-event prediction. Seamlessly handling diverse modalities and complex temporal dynamics, NOAH provides a versatile, task-agnostic, scalable foundation for intelligent predictive systems in personalized clinical care and digital medicine.
Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.