A New Ai Tool Can Calculate the Risk of Someone Becoming Ill in the Coming Years. The Program Has Been trained with anonymized medical data and can predict the risk of individuals developing a specific disease.
The System Can Predict the Onset of 1,231 Different Diseases, Researchers Write in The Scientific Journal Nature . The Tool Can, For Example, Calculate the Chance of Type 2 Diabetes and Heart Attacks.
One of the Scientists Tells BBC News That the System Can Predict Diseases as if it was a weather forecast. “Just as we can predict a 70 percent Chance of Rain, we can do that in Healthcare,” Says professor Ewan Birney of the European Molecular Biology Laboratory. “We can do that not only for one disease, but for all those diseases at the same time. That was not Possible before.”
The AI Model, official called delphi-2M, is trained on anonymous data from 400,000 people from the uk biobank. Also, 1.9 Million Medical Records from Danes Were Used. The Data From Both Countries Provided Reliable Results Accordination to the Scientists. The Accuracy Score is on Average 0.7 for predictions about ten years. The best score would be 1.0.
If the model Registers Risks of Diseases in Time, Doctors May Be Able to Advise Patients Earlier to Change Their Lifestyle or Prescrobe Certain Mediations. Yet, Delphi-2M is not yet ready for use in the medical world. More tests are needed for that.