ITMO: Digital Assistant for Treatment Recommendations for Patients with COVID-19

Main articles:

• Decision Support System (DSS)
• Artificial Intelligence in Medicine
• Drugs for coronavirus COVID-19

2021: Development of the Digital Doctor system

Scientists at ITMO University have developed an algorithm system that, based on analysis of patient data, COVID-19 offers the doctor the optimal treatment for each patient. This was announced by the University on February 2, 2021. The technology was created for people with cancer and cardiovascular disease, but given the pandemic, the program was "taught to help" and patients with COVID-19. Treatment algorithms are based on Russian and foreign official clinical protocols.

The "digital doctor" uses information from the electronic medical record and processes a lot of data from patients with COVID-19. The program includes an algorithm that takes into account a certain number of drugs and treatments, as well as possible conditions of the patient. The algorithms are based on drug interaction information taken from World Health Organization databases. By analyzing the data, the program tracks each stage of treatment. The system step by step prompts the doctor what to pay attention to, how to relate appointments to the list of drugs that a person usually takes. In this case, the program takes into account the material factor, so it can offer an alternative to expensive drugs or procedures if they are not available to the doctor or patient.

As of February 2021, the system passed a number of tests at the NMIC named after V.A. Almazov and several private clinics. Her utility was tested for diseases such as asthma, arrhythmia and multiple sclerosis. The creators of the project, together with thirteen doctors from various institutions, clinics and medical organizations, studied more than 300 real stories of diseases, on the basis of which they created several "template patients" of different severity. Doctors simulated different requests and separate examples of the management of "patients," and scientists compared the working time of doctors using the system and without it. The results showed that doctors spent two to four times less time on various actions.

It was important for us to achieve a minimum of false negative results, and to achieve a maximum of accuracy, even if there are false positive results. That is, the system will mark everything that it can, even if there are no problems, and the doctor will already make a decision. At the same time, the decision always remains with the doctor - the program only highlights the "suspicious" details and recommends a method of treatment, "said Alexandra Vatyan, associate professor at the Faculty of Information and Communication Technologies at ITMO University.

ITMO scientists are working to make the project a single system in which users - doctors - can exchange information and advise each other. As of February 2021, the verification process of all interfaces is underway. A fully completed working prototype is planned to be presented at the end of summer 2021.