Методи математичного прогнозування в медико-біологічних дослідженнях та їх теоретичні можливості в прогнозуванні перебігу СOVID-19

Abstract

SummaryToday, the predictors of the severe course of coronavirus disease (COVID-19) andthe fatal outcome of this disease, the criteria for classifying patients as at-risk for thecritical course are not well understood. The results of a retrospective multicenter studyof Chinese researchers were published in April this year in the journal ClinicalInfectious Diseases. In tis article, the first attempt was made to evaluate the possibilitiesof predicting the course of COVID-19, and the role of 7 clinical and laboratoryindicators (elderly patients, high serum levels of lactate dehydrogenase, C-reactiveprotein, coefficient of variation in the distribution width of red blood cells, ureanitrogen, direct bilirubin, and low albumin). Our review discusses the validity of theappropriateness of applying mathematical modeling methods in medicine and, inparticular, the multidimensional methods of regression, logistic regression, discriminantanalysis, Bayesian classification method, and the creation of neural networks. Thetechnique of mathematical modeling and quality assessment of models and theiroperational characteristics is considered.A review of factors that already have a theoretical justification for creating modelsfor predicting the course of COVID-19 had been conducted. It is shown that, accordingto the literature, predictors reflecting the development of a severe course of COVID-19may be indicators reflecting the development of a cytokine storm in a patient, asyndrome of systemic inflammatory reaction with an increase in the levels ofinflammatory and immunomodulation mediators as well as appearance of laboratorysigns of hypermetabolism syndrome and its extreme degree - multiple organ syndromeinsufficiency (abnormalities of protein, carbohydrate and lipid metabolism).Simple indicators such as an increase in blood glucose and insulin levels, a decrease inserum levels of protein (primarily albumin), and an increase in daily renal excretion of nitrogen(available laboratory indicators of hypermetabolism syndrome) can be red flags that shouldalert doctors regarding the risk of developing a severe / critical course of this disease, butrequire confirmation by the results of well-planned high-quality studies and calculated validatedmathematical methods of the COVID-19 course prediction.