Simulating clinical interventions with a generative multimodal model of human physiology

2026-04-30

AI Summary

This study addresses the substantial inter-individual variability in response to clinical interventions and the challenge of modeling multidimensional physiological dynamics within a unified framework capable of simulating personalized interventions. Leveraging deep phenotyping data from over 15,000 individuals across seven health domains—including 667 longitudinal biomarkers—the authors propose HealthFormer, a decoder-only Transformer architecture that unifies heterogeneous, multimodal health measurements through a shared tokenization scheme. This enables autoregressive prediction of physiological trajectories and in silico intervention simulation without task-specific fine-tuning. Evaluated across four independent cohorts, HealthFormer significantly outperforms conventional clinical risk scores on 27 of 30 endpoints, achieves 100% accuracy in predicting intervention directionality, and generates mean predictions within the 95% confidence interval in 30 of 41 trials, marking the first successful application of generative multimodal physiological modeling for cross-cohort disease prediction and clinical digital twins.

Abstract

Understanding how human health changes over time, and why responses to interventions vary between individuals, remains a central challenge in medicine. Here we present HealthFormer, a decoder-only transformer that models the human physiological trajectory generatively, by training on data from the Human Phenotype Project, a multi-visit cohort of over 15,000 deeply phenotyped individuals. We tokenise each participant's health trajectory across 667 measurements spanning seven domains: blood biomarkers, body composition, sleep physiology, continuous glucose monitoring, gut microbiome, wearable-derived physiology, and behaviour and medication exposure. We train HealthFormer to forecast individual physiological trajectories across these domains, and from this single generative objective a range of clinically relevant tasks can be expressed as queries on the model. We show that, without task-specific training, HealthFormer transfers to four independent cohorts and improves prediction for 27 of 30 incident-disease and mortality endpoints, exceeding established clinical risk scores in every comparison. We further show that the model can simulate interventions in silico: in a held-out personalised-nutrition trial, intervention-conditioned predictions recover individual six-month biomarker changes (e.g., Pearson r = 0.78 for diastolic blood pressure). Across 41 randomised intervention-outcome comparisons drawn from published trials, our results show that the predicted direction of effect agrees in every case, and the predicted mean falls within the reported 95% confidence interval in 30 cases. We position HealthFormer as an initial health world model, from which forecasting, risk stratification, and intervention-conditioned simulation arise as queries, providing a basis for clinical digital twins.

Problem

• clinical interventions
• individual variability
• physiological trajectories
• health modeling
• intervention response

Innovation

• generative multimodal model
• physiological trajectory forecasting
• in silico intervention simulation
• health world model
• clinical digital twin