Predicting Multiple Illnesses from Symptoms Using Smart Computer Maps

Many people live with several health conditions at the same time. Doctors call this multimorbidity. It creates serious challenges for hospitals and clinics worldwide. Medical teams need better ways to spot these patterns before they get worse.

A new computer model offers one possible solution. It builds a web-like map that connects symptoms with diagnoses. The model also adds scientific knowledge about diseases and tiny molecules. It blends broad medical information with small biological details. Then it calculates the chances for many disease categories at once.

The researchers tested the model using two separate tasks. In the first task, they gathered all symptoms from a patient across different visits. They used that complete list to predict the diseases recorded for that same patient. In the second task, they focused on just one visit. They used the symptoms and some known diagnoses from that visit to guess the other diseases present.

The results showed that diseases tend to cluster in non-random ways. The model mapped connections between symptoms, illnesses, and biomedical facts. This method works for both individual patients and single hospital visits.