A new generative AI model predicts the risk and timing of over 1,000 diseases using large-scale health record data from the UK and Denmark.
Key Details
- 1The model was trained on anonymised data from 400,000 UK Biobank participants and validated on 1.9 million Danish registry patients.
- 2It leverages generative transformer techniques similar to those in large language models to process health event sequences.
- 3Excels at forecasting diseases with clear progression patterns (e.g., heart attacks, cancer); less effective for variable conditions.
- 4Produces probabilistic, population-level risk estimates rather than individual certainties; shorter-term forecasts are more reliable.
- 5Underrepresents childhood/adolescent events and some ethnic groups due to training data limitations.
- 6Not yet ready for clinical use, but helps study disease progression and simulate outcomes for research.
Why It Matters
This research demonstrates the growing potential of AI to model disease trajectories at scale, paving the way for more preventive, personalized medicine. While not clinic-ready, such approaches could support risk stratification, resource planning, and ultimately improve patient outcomes in radiology and beyond.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.