Scientists have developed a new predictive model that can estimate a person’s risk of developing more than 1,000 medical conditions, including diabetes, cancer, and heart disease, years before symptoms emerge.
The research, published Wednesday in the journal Nature, describes the tool as one of the largest demonstrations to date of how advanced data models could be applied to healthcare. The system was trained on anonymised health records from 400,000 people in the United Kingdom and tested using data from 1.9 million patients in Denmark.
By analysing the sequence of past medical events—such as diagnoses, smoking histories, and the time gaps between them—the model can identify patterns that often precede serious illnesses. Researchers stressed that the results represent probabilities, not certainties, likening the output to a weather forecast.
“This is the beginning of a new way to understand human health and disease progression,” said Moritz Gerstung, head of the oncology modelling group at the German Cancer Research Centre (DKFZ), which helped lead the project.
The model proved most reliable in predicting conditions with well-documented and consistent progression pathways. These include certain cancers, diabetes, cardiovascular disease, and septicaemia, a severe blood infection. Its accuracy was also stronger in the near term, providing more dependable forecasts over a span of several years than over decades.
However, the system struggled with conditions that are less predictable, such as infectious diseases, mental health disorders, and pregnancy-related complications. Researchers said this limitation highlights the complexity of these conditions and the challenges in forecasting them based on past data alone.
Ewan Birney, interim director general of the European Molecular Biology Laboratory (EMBL), another partner in the project, described the results as “a big step towards more personalised and preventive approaches to healthcare.” He said the model demonstrated how long-term health data could be harnessed to generate meaningful predictions.
The collaborative effort also involved the University of Copenhagen. While the tool is not yet ready for clinical use, scientists believe it could eventually help doctors flag high-risk patients earlier and guide tailored preventive care. Gerstung said the model might “support earlier, more tailored interventions” once further validated.
Independent experts cautioned that the datasets used to build the system were limited in scope. Because they primarily reflect populations in the UK and Denmark, the findings may not be fully representative of other regions or diverse groups. More work will be needed to test the model across varied populations, particularly given differences in age, ethnicity, and healthcare access.
For now, researchers say the model’s value lies in deepening the understanding of how diseases evolve over time and how lifestyle factors interact with medical history to shape long-term health outcomes.
