acute respiratory viral infections
A.A. Uskov, M.V. Shipilov
In view of the widespread and nonspecific clinical symptoms of acute respiratory viral infections (ARVI), with a long time required for the specific diagnosis, as well as the frequent handling of these patients to doctors of various specialties, no doubt about the urgency of measures to improve the diagnosis and treatment ARVI. For this purpose, the authors suggested the following expert systems, «Express-diagnosis and prognosis of ARVI», «Violations of cytokine network functioning in patients with ARVI,» «Treatment guidelines for patients with ARVI».
Considered expert systems are based on algorithms of deductive fuzzy inference using fuzzy base of production rules obtained by the expert. This model allows to represent both declarative and procedural knowledge. A distinctive feature of the developed expert system was given the concentration of cytokines in serum of patients, reflecting an immune response to the introduction of the virus. The authors first introduced the concept of «level of the cytokine storm» that indirectly reflects the activity of effector cells of blood, measured in points with 10-point scale (zero level corresponds to the healthy body). Experimental research has shown high efficiency of expert systems described among physicians of various specialties, regardless of seniority and medical experience. The developed expert system can be used individually, solving individual problems, and together, allowing the complex to support decision making in the diagnosis, treatment, and determine the favorable or unfavorable prognosis of ARVI. When combined expert system it is possible to exchange data between them, which significantly reduces the time for data entry. Currently, the knowledge base of expert systems are modified (to improve), and created their online versions as to work in a local network of major medical institutions, and through the Internet.