关键词: 结核, 肺, 流行病学研究特征(主题), 预测, 小地区分析

Abstract:

Objective: To analyze the epidemiological characteristics of pulmonary tuberculosis (PTB) in Tai’an City, Shandong Province, from 2014 to 2024, a seasonal autoregressive integrated moving average (SARIMA) model and a Prophet model were established to predict the PTB notification trend, thereby providing a scientific basis for the formulation of future prevention and control strategies. Methods: Data on PTB cases in Tai’an City from 2014 to 2024 were collected through the TB Information Management System, a subsystem of the China Disease Prevention and Control Information System. Descriptive epidemiological analysis was performed, and both SARIMA and Prophet models were established to analyze and predict the PTB notification trend. Results: The average annual notification rate of PTB in Tai’an City from 2014 to 2024 was 25.66 per 100000 population (15820 cases), showing an overall downward trend (${\chi }_{trend}^{2}$=2043.193, P<0.001). The notification rate was 45.60 per 100000 (2564 cases) in 2014 and 18.85 per 100000 (1008 cases) in 2024, with an average annual decline rate of 8.45% ((1-$\sqrt[10]{\frac{18.85/100000}{45.60/100000}}$)×100%). The notified cases were concentrated in December, accounting for 9.26% (1465/15820) of the total, while the fewest cases were reported in February, accounting for 6.45% (1020/15820). High-incidence areas included Dongping County (32.90/100000, 2772 cases) and Ningyang County (29.65/100000, 2538 cases). Most notified patients were male, with a male-to-female ratio of 2.94∶1 (11808∶4012), and the notification rate in males (37.61/100000) was significantly higher than that in females (13.03/100000), with a statistically significant difference (χ²=336.950, P<0.001). The age group of 55-64 years accounted for the largest proportion of cases (20.34%, 3217/15820), while the 0-14 years group accounted for the smallest proportion (0.56%, 89/15820). The SARIMA (0,1,1)(1,1,0)₁₂ model and the Prophet model were established using monthly notified PTB case data, demonstrated a good fit with the historical data. The SARIMA model predicted an increase in the number of notified PTB cases in 2025 (1249 cases) compared with 2024 (1008 cases), with a predicted peak in May (125 cases) and a trough in November (88 cases). In contrast, the Prophet model predicted a declining trend in 2025 (849 cases) compared with 2024 (1008 cases), with a predicted peak in December (95 cases) and a trough in February (65 cases). Error analysis indicated that the predictive accuracy of the Prophet model (15.26%) was better than that of the SARIMA model (19.29%). Thus, the Prophet model performed better in capturing temporal distribution characteristics and achieved higher prediction precision. Conclusion: The notification rate of PTB in Tai’an City showed an overall downward trend from 2014 to 2024. It is recommended to strengthen TB screening and health education for high-risk groups such as those aged 55-64 years, as well as in high-incidence areas including Dongping County and Ningyang County. The Prophet model performed better than the SARIMA model in predicting the PTB notification trend in Tai’an City. Based on the predictions of the Prophet model, it is recommended that Tai’an City prioritize enhanced PTB surveillance and health education resources deployment in advance in late autumn, so as to shift the prevention and control mode from passive response to active intervention and improve overall effectiveness.

Key words: Tuberculosis, pulmonary, Epidemiologic study characteristics as topic, Forecasting, Small-area analysis