[Analysis of the epidemic characteristics of reported pulmonary tuberculosis incidence in Kashgar Prefecture, Xinjiang Uygur Autonomous Region from 2015 to 2022 and establishment of SARIMA prediction]

Abstract

Objective: To analyze the epidemic characteristics of reported tuberculosis incidence in Kashgar from 2015 to 2022, and use the seasonal autoregressive integrated moving average (SARIMA) model to predict the incidence, providing references for the local control of pulmonary tuberculosis. Methods: The reported incidence data of tuberculosis in the Kashgar area of Xinjiang from January 2015 to August 2023 were collected through the"Infectious Disease Monitoring System", a subsystem of the "Chinese Disease Prevention and Control Information System". The epidemic characteristics of reported incidence in this area from 2015 to 2022 were analyzed. Two SARIMA models of monthly reported incidence number and rate were established. The prediction performance of the two models was evaluated using the reported incidence data of tuberculosis from January 2023 to August 2023. The χ² test was used to analyze population characteristics, and the Cochran-Armitage trend test was used to analyze annual incidence. Results: From 2015 to 2022, 133 972 cases of pulmonary tuberculosis were reported in Kashgar, with a yearly reported incidence rate of 383.64/100 000, showing a rising trend (T_CA=77.03, P<0.001) and then a declining trend (T_CA=176.16, P<0.001). The proportion of pathogenic positive pulmonary tuberculosis had increased yearly (T_CA=132.66, P<0.001). The reported onset time was concentrated from January to June each year, with a peak in April. Yengisar County, Zepu County and Yopurga County had the highest reported incidence rate in Kashgar. The sex ratio of men to women was 1.03∶1, and the reported incidence rate of men was higher than that of women (χ²=27.04, P<0.001). The reported incidence rate of the group aged 60 years and older was the highest. The patient's occupation was mainly farmers (84.99%). The average relative errors of the SARIMA (1, 1, 2) (0, 1, 1)₁₂ model and SARIMA (0, 1, 1)(0, 1, 1)₁₂ model in predicting the reported monthly incidence number and rate were 11.67% and -9.81%, respectively. Both models had good prediction accuracy (MAPE=33.55%, MAPE=38.22%). Conclusion: The average reported incidence rate of pulmonary tuberculosis in the Kashgar area shows a rising trend first and then a declining trend. The patients are mainly men and farmers, and attention should be paid to the prevention and control of tuberculosis among the elderly in winter and spring. The SARIMA (1, 1, 2) (0, 1, 1)₁₂ model and SARIMA (0, 1, 1)(0, 1, 1)₁₂ model can fit the trend of reported tuberculosis incidence in the Kashgar area well and have good predictive performance.