Artificial intelligence can predict people's health problems over the next decade, 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 say it operates similar to a weather forecast that indicates a 70% chance of rain, but this time for human health.
With this AI model, the goal is to identify high-risk patients to prevent diseases and assist hospitals in understanding future healthcare demands.
The model, called Delphi-2M, uses technologies similar to AI chatbots. While chatbots learn language patterns, Delphi-2M analyzes anonymous medical records to predict future health outcomes.
It doesn't provide exact dates for conditions like heart attacks but gives estimates on the likelihood of various diseases.
So, just like in weather forecasting we might say there is a 70% chance of rain, we can do that for healthcare, stated Prof Ewan Birney, interim executive director of the European Molecular Biology Laboratory. What’s exciting is we can assess all diseases simultaneously, an unprecedented achievement.
The AI model was created using anonymous UK data, including hospital admissions and lifestyle factors from over 400,000 individuals participating in the UK Biobank project. It was later tested against data from Danish medical records, where it performed impressively.
The AI could eventually guide preventive measures like lifestyle changes or medications, similar to how cholesterol-lowering statins are currently prescribed based on heart attack risk.
Ultimately, researchers envision using this technology to enhance public health strategies, pinpoint high-risk areas, and improve the understanding of disease progression.
While the model requires refinement and robust testing before clinical application, it marks a significant advance towards scalable and ethically responsible health predictions in the medical field.
