PREDICTION OF LONG-TERM TREATMENT IN PATIENTS WITH COVID-19 BASED ON ANALYSIS OF THEIR PRIMARY EXAMINATION DATA

Introduction. The COVID-19 pandemic has caused enormous social and economic damage. A significant part of people who have contracted this infection have serious health problems. To prevent the long-term consequences of COVID-19, models for predicting the duration of treatment based on prognostic factors obtained at the beginning of the disease are needed. Materials and Methods. In 2020-2021, a survey of 832 patients with COVID-19 was conducted. With the help of computer programs Microsoft Excel 2021 and Statistica 12.0 Trial Version for Windows, the collected database was processed. The relationship between numerical predictors was investigated using Spearman's correlation, and between categorical indicators – gamma correlation. To predict the duration of the patient's treatment based on the initial clinical symptoms and signs, the tools of the Statistica 12.0 program "classification trees" were used. The model was built using the tool Data Mining – Trees/Partitioning – C&RT (Classification and Regression Tree). Results and Discussion. During the study period, it was established that the long-term treatment of patients with COVID-19 depends on age, concomitant diseases, shortness of breath, body temperature, chest pain, frequency of respiratory movements in 1 minute, pain in the heart area, average blood pressure and heart rate according to 1 minute. This is confirmed by a significant difference between the groups of patients who were treated for up to 30 days and 31 days and more, as well as established correlations. On the basis of the above-mentioned predictors, a decision-making algorithm was developed to determine the duration of patient treatment. Conclusions. Established interrelationships between clinical symptoms in patients with COVID-19 will allow timely detection of possible complications in patients. Implementation of a built decision-making model based on leading predictors will help doctors predict the duration of the infection and develop effective measures to prevent serious post-infection consequences.