Sciact
  • EN
  • RU

Bayesian Analysis of Tuberculosis Spread Scenarios in Regions of Russian Federation Full article

Journal Mathematics
, E-ISSN: 2227-7390
Output data Year: 2026, Volume: 14, Number: 10, Article number : 1600, Pages count : DOI: 10.3390/math14101600
Tags mathematical modeling; tuberculosis; multidrug-resistant forms; inverse problem; Sobol analysis; optimization; Bayesian approach; probabilistic forecasting
Authors Krivorotko Olga 1,2 , Neverov Andrei 1 , Schwartz Yakov 3 , Kaminskiy Grigoriy 4 , Zyatkov Nikolay 1 , Laushkina Zhanna 3
Affiliations
1 Research Center for Information Technologies and Artificial Intelligence, Sirius University of Science and Technology, 354340 Sirius, Russia
2 Sobolev Institute, Mathematics of SB RAS, 630090 Novosibirsk, Russia
3 Novosibirsk TB Research Institute, the Ministry of Health RF, 630040 Novosibirsk, Russia
4 Tula Regional Clinical Center for the Prevention and Control of AIDS and Infectious Diseases, 300002 Tula, Russia

Abstract: Understanding the heterogeneous spread of tuberculosis (TB), particularly multidrug-resistant (MDR) forms and the role of subclinical infection, is critical for achieving the WHO End TB strategy. This study develops a novel compartmental model that explicitly incorporates incipient and subclinical TB together with MDR forms, and links them to case detection and treatment pathways. The key innovation lies in integrating a sensitivity-based identifiability analysis with a Bayesian MCMC framework to quantify parameter uncertainty and correlations directly from regional surveillance data. Applied to five high-burden regions of the Russian Federation (2009–2020), the approach reveals strong heterogeneity in epidemic drivers: wide credible intervals for contagiousness, the rate of progression to bacterio-positive (BE+) states, and detection rates. The probabilistic forecasts up to 2025 are validated against 2021–2023 data. The region-specific differences in these correlated parameters dictate transmission dynamics, and improving detection of BE+ cases is the most effective lever for control.
Cite: Krivorotko O. , Neverov A. , Schwartz Y. , Kaminskiy G. , Zyatkov N. , Laushkina Z.
Bayesian Analysis of Tuberculosis Spread Scenarios in Regions of Russian Federation
Mathematics. 2026. V.14. N10. 1600 . DOI: 10.3390/math14101600 WOS Scopus OpenAlex
Dates:
Submitted: Mar 25, 2026
Accepted: May 6, 2026
Published online: May 8, 2026
Identifiers:
≡ Web of science: WOS:001774988600001
≡ Scopus: 2-s2.0-105040056694
≡ OpenAlex: W7160828698
Altmetrics: