Analysis of changes in tuberculosis incidence trends before and after COVID-19 based on time series models

Background This study aims to analyze the trends in tuberculosis (TB) cases from 2012 to 2023 in Yingjisha, providing valuable insights for improving TB prevention and control strategies. Methods The SARIMA, TBATS, TSLM, and STSM models were used for analysis. Results The number of reported TB incidences before COVID-19 was relatively stable and exhibited clearly seasonal patterns. The TBATS, TSLM, and SARIMA models fitted RMSE were 42.77, 55.57, and 64.02, the fitted MAPE were 1.46%, 1.92% and 3.20% respectively. The Box–Ljung test for the fitted sequences and the ADF test on the model fitted series yielded p < 0.05 for the above model. When using the models for prediction, the SARIMA model was lower than the other models. There is still a large deviation between the prediction and the actual number of reported TB incidences during the COVID-19 period. Conclusions Since the outbreak of COVID-19, there has been a decrease in the number of TB incidences, which has been largely influenced by epidemiological, social, and individual psychological factors.