AI Model Predicts Disease Risk Decade Ahead Like Weather Forecast

European researchers have developed an artificial intelligence model that can predict people’s health problems over a decade into the future, using technology similar to ChatGPT to analyse medical records and forecast disease risk across more than 1,000 conditions.

Context and Background

The AI model, called Delphi-2M, was developed by researchers at the European Molecular Biology Laboratory, the German Cancer Research Centre (DKFZ), and the University of Copenhagen. Professor Ewan Birney, interim executive director of the European Molecular Biology Laboratory, explains the breakthrough as creating healthcare predictions “just like weather, where we could have a 70% chance of rain.”

The model was initially trained using anonymous UK data from over 400,000 participants in the UK Biobank research project, including hospital admissions, GP records, and lifestyle factors such as smoking habits. Validation testing with 1.9 million medical records from Denmark confirmed the model’s accuracy, with Prof Birney noting: “If our model says it’s a one-in-10 risk for the next year, it really does seem like it turns out to be one in 10.”

The AI performs best at predicting diseases with clear progression patterns, including type 2 diabetes, heart attacks, and sepsis, whilst proving less effective for random events like infections.

Looking Forward

The technology promises to revolutionise preventive healthcare by identifying high-risk patients whilst early intervention remains possible. Applications could include targeted medication prescriptions, personalised lifestyle advice, and enhanced disease-screening programmes tailored to individual risk profiles.

Healthcare systems could utilise the model for strategic planning, anticipating regional disease demand years in advance to optimise resource allocation. Professor Moritz Gerstung from DKFZ describes this as “the beginning of a new way to understand human health and disease progression,” with potential for personalised care delivery at scale.

Source Attribution:

Share this article