Article · Infectious Disease Modelling · October 13, 2026

Seasonal dynamic model for projecting tuberculosis epidemiological trends in Kyrgyzstan

А.К. Байтелиева, Pengwei Lou, Kalmambetova Gulmira, Daniil Shauer, Jiabo Xu, Kai Wang

Abstract

ABSTRACT Tuberculosis (TB) remains a major public health burden in Kyrgyzstan; however, the seasonal transmission dynamics of TB in the country have not been quantitatively characterized. This study aimed to develop a seasonal dynamic model to simulate TB transmission, estimate the basic reproduction number R 0 , assess transmission intensity, and inform evidence–based prevention and control strategies. Using monthly reported new TB cases from the national TB surveillance and registration system in Kyrgyzstan from 2022 to 2025, we constructed a seasonal SEIR model incorporating periodic transmission and reactivation rates. Model parameters were estimated using a Bayesian Markov chain Monte Carlo (MCMC) method. R 0 was calculated using two methods: the periodic time–varying method and the annual average transmission rate method. Local sensitivity analysis was performed to evaluate the impact of the baseline transmission parameter a s and the baseline reactivation parameter a e on TB transmission dynamics. The model adequately captured the seasonal fluctuations and overall downward trend in new TB cases, achieving a mean absolute percentage error (MAPE) of 8.66% and a root mean square percentage error (RMSPE) of 9.95%. The estimated periodic R 0 was 0.8842 (95% credible interval [CrI]: 0.8811–0.8873), which remained below the threshold of unity. Assuming no change in current prevention and control measures, projections suggest a gradual decline in monthly new TB cases, with the annual number of new cases declining to approximately 2,902 by 2126. These results indicate that TB transmission has not been fully interrupted and that elimination is unlikely to be achieved within the projection period. Sensitivity analysis revealed that both a s and a e substantially affected the number of new TB cases and R 0 , and that modest changes in these parameters could drive R 0 above the threshold of unity. The seasonal SEIR model provided a good fit to the monthly number of new TB cases in Kyrgyzstan. Although the estimated R 0 is below the threshold of unity, indicating a declining epidemic trend, the rate of decline remains relatively slow, highlighting the need to sustain and strengthen prevention and control measures.

Tuberculosis Research and EpidemiologyMathematical and Theoretical Epidemiology and Ecology ModelsHIV, TB, and STIs Epidemiologytuberculosisseasonal dynamicsepidemic projection

Seasonal dynamic model for projecting tuberculosis epidemiological trends in Kyrgyzstan

Abstract

More by Altynai Baitelieva