An artificial intelligence model called Delphi-2M is being designed to predict people’s health problems up to a decade in advance, with an accuracy of around 70%, in much the same way weather forecasts work, according to the BBC.
The system is not yet ready for clinical use and will not be able to provide exact dates, but it can estimate the likelihood of 1,231 different diseases by analyzing medical records. This could help tailor healthcare to individuals and allow hospitals to anticipate demand much earlier.
Delphi-2M uses technology similar to well-known AI chatbots such as ChatGPT, detecting patterns in medical records.
Such predictions might translate into targeted medication or lifestyle guidance. For example, people at risk of developing a liver disorder could be advised to cut down on alcohol, benefiting them more than the general population.
The tool could also support disease-screening programs by analyzing entire sets of medical records from a given area, forecasting scenarios such as how many heart attacks are expected in a region by 2030. This, in turn, could guide resource planning for healthcare systems.
“This marks the beginning of a new way of understanding human health and the progression of disease,” said Professor Moritz Gerstung, head of the Artificial Intelligence in Oncology department at DKFZ, the German Cancer Research Center.
