Прототип цифрового помощника врача «Доктор Пирогов» создали в НГУ

The system covers 250 diseases and 20 medical specializations

Scientists at the Artificial Intelligence Center of Novosibirsk State University have developed a prototype of the digital assistant 'Doctor Pirogov', which supports doctors in providing primary medical care. The system is based on a hybrid architecture that combines neural networks and semantic knowledge graphs, and is already capable of working with data on 250 major diseases.

'Doctor Pirogov' conducts an interactive patient survey through a voice or visual interface, analyzes medical documentation, examination results, laboratory indicators, and genetic testing data. The system generates a list of probable diagnoses, recommends additional examinations, selects therapy taking into account drug interactions, and offers personalized preventive measures. The final decision remains with the doctor.

The prototype covers knowledge in 20 specialties, including therapy, cardiology, endocrinology, neurology, gastroenterology, pediatrics, oncology, and psychiatry. The system is particularly useful for sparsely populated and remote regions where patients are often seen by paramedics: the digital assistant preliminarily collects information, forms a list of possible diseases, and the doctor clarifies the diagnosis and receives recommendations for further treatment.

The hybrid approach allows combining the speed of neural network algorithms with the interpretability of decisions from the semantic network ANDSystem, created based on medical and scientific literature. The development of the semantic part took about 10 years, with participation from employees of the ICIG SB RAS, students, and scientific staff of NSU.

Clinical trials of 'Doctor Pirogov' will begin in 2026, and at the current stage, internal verification of scientific indicators is being conducted. The development opens up opportunities to reduce the routine workload on doctors, improve the quality of primary care, and expand access to medical support in remote regions.

Read more materials: