Artificial intelligence can predict people's health problems over a decade into the future, say scientists.
The technology has learned to spot patterns in people's medical records to calculate their risk of more than 1,000 diseases.
The researchers describe their vision as akin to a weather forecast that anticipates a 70% chance of rain – but for human health.
The AI model is intended to identify high-risk patients, enabling preventive measures, and assist hospitals in understanding future healthcare demands.
The model, called Delphi-2M, employs technology similar to popular AI chatbots like ChatGPT, which are trained to recognize patterns in language. Delphi-2M looks for patterns in anonymous medical records to predict health outcomes.
Prof Ewan Birney, interim executive director of the European Molecular Biology Laboratory, highlights the model's ability to provide disease likelihoods akin to predicting rain: “We can do that for healthcare – not just for one disease but for all diseases simultaneously. This is unprecedented.”
Delphi-2M was initially developed using anonymous UK data from over 400,000 individuals, encompassing hospital admissions, GP records, and lifestyle habits from the UK Biobank research project. Testing confirmed its accuracy with medical records of 1.9 million people in Denmark.
The model excels in predicting diseases with clear progression, such as type 2 diabetes and heart attacks, and it's expected to help anticipate healthcare needs and improve patient care.
Despite its promise, the AI model is not yet ready for clinical use. Future improvements aim to incorporate more diverse data such as genetics and imaging. However, researchers underscore the importance of thorough testing and regulation before widespread adoption.
