Nature Paper: AI Model Forecasts Personal Health Trajectories Up to 20 Years

EricTopol · x · 2026-08-06

Eric Topol highlights a landmark Nature paper by Dr. Sarah Urbut from Harvard Medical School. The team developed an AI model named ALADYNOULLI that integrates electronic medical records and polygenic risk scores. It was trained on a dataset of 683,000 individuals, 348 diseases, and up to 52 years of follow-up.

The model forecasts a person's health trajectory and disease risk up to 20 years out. It dynamically adjusts predictions based on lifestyle changes or medications, functioning like a GPS for health. The research also identified temporal disease signatures, revealing distinct genomic mechanisms behind identical disease phenotypes.

Related event: Nature Study: AI Predicts 20-Year Health Trajectories(4 posts)→

Original post →

More from Research

Research channel →