Comprehensive approach to the tuberculosis indicators assessment of the children population in the Republic of Crimea

Abstract

Introduction. Childhood tuberculosis (TB) control is relevant due to the peculiarities of its course in this age group, and the TB incidence in children is an important prognostic epidemiological indicator.Aim. Use of multivariate statistical analysis to estimate and predict childhood TB indicators in the Republic of Crimea (RC).Materials and methods. The official TB statistics in the Republic of Crimea for 2014-2018 are used. The calculated means of the indicators are checked for normality using the Kolmogorov-Smirnov and Shapiro-Wilk tests. Pearson correlation analysis is applied to determine pair correlation relationships. Stepwise multiple regression analysis is carried out to determine group conditionality of the indicators, where coefficients, with which significant pairwise correlations are found, are selected as independent variables. Based on the results, multiple regression equations are made to predict the values of dependent variables. The data is processed using Statistica 10.0 software.Results. For childhood TB incidence, strong direct correlations are established with the incidence and prevalence of pulmonary TB among children. The paired correlation coefficient between the incidence of childhood TB and childhood lung TB and the detection of active TB patients in preventive examinations of children varied in the range of 0.63-0.72. For the prevalence of TB among children, strong direct correlations were found with the incidence of TB and pulmonary TB in children. Multiple correlation coefficients for the incidence and prevalence of childhood TB exceeded the values of paired correlation coefficients and were in the range of 0.93 to 0.98 (p<0.001), indicating greater significance of group conditionality of the indicators. Determination coefficients R2 were between 0.87 and 0.96. Multiple regression models were built for the childhood TB incidence, childhood lung TB incidence, childhood TB prevalence, childhood lung TB prevalence.Conclusion. The found strong direct pairwise correlations for childhood TB incidence and prevalence and childhood pulmonary TB incidence and prevalence can serve as prognostic criteria and reflect the quality of antituberculosis interventions. High values of paired correlation coefficient between childhood TB incidence and childhood pulmonary TB and detection of patients with active TB in preventive examinations of children are a criterion of quality of both TB services and primary care, which can prevent the spread of TB and improve the epidemic situation of TB in Crimea. The calculated multiple regression models for the studied indicators can serve the needs of practical forecasting in Healthcare.