Abstract. According to the COVID-19 pandemic, the Ukrainian regions significantly differ in the population’s vulnerability to this infection. Specific patterns (combinations) of factors identify the reasons for regional differentiation of morbidity and mortality from COVID-19. They were accumulated over a long period and formed the so-called «retrospective portraits of the region’s vulnerability to COVID-19» for each region. The main purpose of the study is to define such combinations of financial, economic, environmental and social factors causing many deaths and morbidity from COVID-19 among the population of different Ukrainian regions. The study is based on a constructed spatial nonlinear model. According to the step-by-step algorithm, individual factor variables are gradually added / removed from the model specifications by the Aitken method depending on their correlation with morbidity and mortality from COVID-19 in the region until the model’s specification with the highest adequacy by p-value and t-statistics is formed. The nonlinear multifactorial regression equations regarding the dependence of the resulting indicator (the level of morbidity and mortality of the region from COVID-19) on variables — 23 indicators of social, economic, environmental and financial development of each Ukrainian region and Kyiv are built for the creation of the «retrospective portraits of the region’s vulnerability to COVID-19». Besides, the correlation matrices and correlation pleiades are formed. Based on a correlation matrix, the multicollinearity test is performed using the Farrar — Glauber algorithm. The Durbin — Watson method checks residuals for autocorrelation. The heteroskedasticity test is performed using the Spearman rank correlation test. The empirical analysis results show that migration, population size, the environmental situation in the region, a significant index of medical institutions readiness for qualitative patient care during the pandemic and citizens’ income dynamics mostly affect the incidence of COVID-19 and the number of deaths. The retrospective research results can help create road maps of individual regions to overcome the future epidemiological influence effects. Keywords: COVID-19, epidemiological threats, retrospective portraits of regional vulnerability to COVID-19, step-by-step nonlinear regression, morbidity, regional morbidity differentiation, pandemic, multicollinearity, heteroskedasticity. JEL Classіfіcatіon С21, С51, C 31, C12, I15, I18, R58, R11 Formulas: 17; fig.: 3; tabl.: 2; bibl.: 36.