Revolutionary AI Platform Can Predict Personal Risk of Over 1,000 Health Conditions, Confirm Scientists

Experts have created a innovative machine learning system that can forecast a patient's likelihood of developing more than 1,000 diseases, and project health changes up to a long period in coming years.

The AI-powered tool was tailor-made by specialists using algorithmic concepts akin to those employed in advanced AI models.

It represents one of the most extensive showcases to date of how artificial intelligence can simulate health trajectories on a broad scope, having been refined on data from two entirely separate healthcare systems.

"Clinical occurrences often adhere to expected sequences," commented a lead scientist participating on the project. "Our AI model recognizes those sequences and can anticipate long-term wellbeing results."

This system operates by evaluating the probability of whether – and when – an person may experience conditions such as cancer, metabolic disorders, cardiovascular conditions, respiratory disease and many other health issues.

Called this advanced system, it analyzes "clinical occurrences" in a patient's record, such as when illnesses were detected, along with personal habits including weight issues, tobacco use, drinking habits, as well as years and gender.

The system also examines de-identified patient record data to forecast what may develop over the next decade and longer term.

This system was trained and tested using anonymised health records from a large number of participants in multiple extensive patient studies.

Medical probabilities are presented as likelihoods throughout durations, similar to predicting a likelihood of precipitation on a specific date.

An influential research director commented that patients might benefit from the platform within the next few years.

"You enter the clinic and the clinician is accustomed to using these tools, and they are capable of say: ‘These are four significant health concerns that are in your ahead and here’s two things you could do to significantly alter that.’"

The expert continued, "It's likely most individuals will be advised to improve diet, and if you use tobacco you will be encouraged to cease – and that will be reflected in your records so that advice is unlikely to shift remarkably – but for specific illnesses I think there will be personalized recommendations. This is the vision we aim to build."

This strength of the new AI tool over existing ones – such as single-disease risk calculators – is that "it can evaluate every condition together and over a prolonged span. This is the capability that condition-specific tools lack."

Researchers stated: "The tool predicts the likelihoods of over 1,000 diseases, conditional on each person’s health record, with precision comparable to that of current specialized calculators."

"Its AI-driven design also allows projection of synthetic future health trajectories, providing insightful projections of future health impacts for up to 20 years."

A respected professor in AI and health stated: "This represents the start of a new way to understand human health and disease progression."

"AI systems such as this one could eventually help tailor treatments and foresee patient services at a broad level."
Amy Jones
Amy Jones

Lena ist eine erfahrene Journalistin mit Schwerpunkt auf Politik und Gesellschaft, die regelmäßig über deutsche und europäische Themen berichtet.