Sber's AI-enabled development helped to almost halve pneumonia mortality
2021-05-26 01:06

Sber's AI-enabled development helped to almost halve pneumonia mortality

"Sber and Krasnoyarsk Krai Clinical Hospital are implementing a project using artificial intelligence (AI), which helps doctors to timely identify patients with high risk of severe viral pneumonia and, as a result, promptly adjust treatment tactics and reduce the number of fatal outcomes.

"In the first month of its use in the regional hospital, the proportion of high-risk patients who went to intensive care decreased by 14.3%, and the proportion of fatal outcomes decreased by 44.3%," Sber told reporters.

The project was launched in 2019. An impersonal database of patients with pneumonia treated at the hospital was used to train artificial intelligence. Thus, the system analyzes the information entered by the doctor in the electronic medical record of the patient, as well as the results of laboratory and instrumental tests obtained during the first day of hospitalization. Based on these data, the system assigns the patient one of four risk levels of severe pneumonia (ranging from 29% to 100%).

"Over the past year and a half, the team of Sberbank Artificial Intelligence Laboratory, Krasnoyarsk Krai Clinical Hospital and SberMedSI have gone a long way from the idea to debugging and implementing the model. As a result, doctors have received a tool that is good at identifying the risk of severe disease, and therefore makes it possible to start prevention in time and ultimately save lives," said Alexander Vedyakhin, First Deputy Chairman of Sberbank's Management Board.

Such a design is used in Russia for the first time. If the project shows high efficiency in treatment of pneumonia over the long term, its experience will be applied to other areas of medicine, where it is important to detect the risk of complications and death of patients in time. Krasnoyarsk Krai doctors have already positively evaluated the expansion of the system to other infectious diseases hospitals in the region. Pulmonologists and artificial intelligence researchers from other regions are also interested in the project.

Made in Russia // Made in Russia

Author: Karina Kamalova