Evaluation of the Effectiveness of an AI-Based Telemedicine System for Remote Screening of Chronic Disease Risks
Seliverstov Pavel VasilievichPeter the Great St. Petersburg Polytechnic University, St. Petersburg, Polytehnicheskaya St, Russia pavel.seliverstof@gmail.com0000-0001-5623-4226
Valentin ShapovalovPeter the Great St. Petersburg Polytechnic University, St. Petersburg, Polytehnicheskaya St, Russia. valshapovalov@mail.ru0000-0002-9764-4018
Andrey VasinPeter the Great St. Petersburg Polytechnic University, St. Petersburg, Polytehnicheskaya St, Russia. vasin_av@spbstu.ru0000-0002-1391-7139
Vladimir B. GrinevichPeter the Great St. Petersburg Polytechnic University, St. Petersburg, Polytehnicheskaya St, Russia. grinevich_vb@mail.ru0000-0002-1095-8787
Konstantin SemenovPeter the Great St. Petersburg Polytechnic University, St. Petersburg, Polytehnicheskaya St, Russia k_semenov@mail.ru0009-0002-4419-7776
Keywords: Artificial Intelligence, Medical Systems, Telemedicine, Telemedicine Technologies, Digital Medicine, Health, Remote Screening, Evaluation of Chronic Disease Risks.
Abstract
The goal of this paper is to evaluate efficiency of the developed diagnostic model for screening of students’ health. The basis for medical research was the city polyclinic No.76 in Saint Petersburg, which serves the young students of higher educational institutions in the Northern capital. The tool for remote examinations of students was a telemedicine system of remote screening which comprised web infrastructure (a virtualization platform in the form of a server, hardware, tools for accounting of calculations etc). The system also includes an original questionnaire on the basis of which a decision on presence of pathology risks and degree of their manifestation is issued automatically. Risks under the chosen 5 nosologies (cardiology, gastroenterology, pulmonology, endocrinology, oncology) are determined using a semantic network of questions (198 questions were formulated in total), answers to which help analyze and reveal symptoms of the nosological disease forms stipulated in the ICD. The system takes into account not only the quality of recognition of pathology risks, but is also capable of predicting the positive and negative, that is, absence of disease risks. Thereat, the risk factors were determined remotely taking into account three levels of threat probability: low, medium and high. According to the filled-in questionnaire, the system issues a formalized graphical summary conclusion on the state of risks and personified recommendations on examination, as well as healthy lifestyle tips. The work for determination of efficiency of telemedicine remote screening on the basis of the city polyclinic No.76 of Saint Petersburg involved 3155 persons – first-year students whose average age was 19.6±1.5 years. The research has showed good informativeness of the method, sufficient sensitivity and specificity. Based on the research results, it was concluded that the telemedicine system of remote screening allows for objective revelation of health risks under the specified nosologies, indication of early symptoms of adaptation disorders, and is an efficient tool of preventive medicine. The authors of the paper hope that this research will make a contribution to development of contemporary telemedicine information systems, so that doctors and health care professionals on the whole could improve the decision-making process while reducing time and financial expenses on diagnostics. Key words: Artificial Intelligence, Medical Systems, Telemedicine, Telemedicine Technologies, Digital Medicine, Health, Remote Screening, Evaluation of Chronic Disease Risks.