关键词: 结核,肺, 发病率, 模型,统计学, 预测

Abstract:

Objective: To analyze the epidemic characteristics and changing rules of tuberculosis reported in Dongcheng District of Beijing from 2013 to 2022. The reported incidence from January to June 2023 was predicted through modeling the previously reported data using the seasonal autoregressive integrated moving average (SARIMA), so as to provide reference for tuberculosis prevention and control measures in the district. Methods: The reported incidence data of pulmonary tuberculosis in Dongcheng District, Beijing from January 2013 to June 2023 was obtained through the “Infectious Disease Monitoring System” subsystem of the “China Disease Prevention and Control Information System”. The epidemic characteristics of reported pulmonary tuberculosis from 2013 to 2022 were analyzed. The SARIMA model was established using monthly reported incidence data from 2013 to 2022, and the model was applied to predict and verify the reported incidence data from January to June 2023. Results: A total of 2505 pulmonary tuberculosis patients were reported from 2013-2022 in Dongcheng, with an average annual reported incidence rate of 28.81/100000. The highest reported incidence rate was in 2013 (38.68/100000, 379 cases), and the lowest rate was in 2020 (23.30/100000, 185 cases). The annually reported incidence rate showed an decreasing trend ($χ^{2}_{trend}=25.371$, P<0.001), and the average annual decline rate was 5.26%. The highest detection rate of pathogenic positive pulmonary tuberculosis was in 2022 (57.74%, 97/168), and the lowest rate was in 2017 (30.71%, 74/241), showing an overall upward trend year by year ($χ^{2}_{trend}=29.945$, P<0.001). The annual average reported incidence rate of male tuberculosis was 36.85/100000 (1559 cases), significantly higher than that of female tuberculosis (21.20/100000, 946 cases; χ²=184.738, P<0.001; n_male:n_female=1.73∶1). The proportion of patients in the 20-29 age group (19.88%, 498/2505) and retirees (37.72%, 945/2505) was relatively high and the patients were mainly concentrated in Yongdingmenwai Street (11.38%, 285/2505). Seasonal analysis showed that the seasonal index ranged from 0.83 to 1.09, with periodic fluctuations, and the epidemic period mainly concentrated from March to August, as well as in December. The SARIMA (0,1,2) (1,2,1)₁₂ model fit well with the reported incidence trend (AIC=657.67), with an average relative error of -17.72% and high prediction accuracy (root mean square error of 5.188 and average absolute percentage error of 22.01%). Conclusion: From 2013 to 2022, the reported incidence of tuberculosis in Dongcheng District of Beijing showed a steady downward trend, and the patients were mainly male and retired population. Attention should be paid to the prevention and control of tuberculosis and propaganda and education of the elderly and young adults in spring and summer. The SARIMA (0,1,2) (1,2,1)₁₂ model can well fit the trend of reported incidence of pulmonary tuberculosis in this region and has good predictive effects.

Key words: Tuberculosis, pulmonary, Incidence, Models, statistical, Forecasting