基于ARIMA的新冠肺炎疫情前后结核病流行趋势预测与分析

目的 构建ARIMA季节性模型，探讨新型冠状病毒肺炎疫情(简称新冠肺炎疫情)对结核病流行特征的影响，预测上海市宝山区结核病流行趋势。方法 收集上海市宝山区2009—2021年结核病月发病率资料，构建ARIMA季节性模型，验证预测模型效果，分析预测误差的原因。结果 上海市宝山区结核病月发病率模型为ARIMA(2,0,0)(0,1,1)₁₂,BIC值最小，Ljung-Box统计量Q=23.127,P=0.081,残差序列为白噪声。2019年实际月发病率与预测值变化趋势基本一致，且均在预测值95%可信区间内。受新型冠状病毒肺炎疫情影响，近两年观察值与预测值差异较大，2021年2月观察值在拟合值的95%置信区间外。结论 ARIMA(2,0,0)(0,1,1)₁₂模型能较为准确地预测宝山区新冠肺炎疫情前结核病发病趋势，受新冠肺炎疫情影响时，预测结果偏差较大，需要后疫情时代结核病发病数据来重新建模。

Prediction and analysis of epidemic trend of tuberculosis before and after the COVID-19 pandemic based on ARIMA

Objective To construct a seasonal ARIMA model, to explore the impact of coronavirus disease 2019 (COVID-19) on the epidemic characteristics of tuberculosis, and to predict the incidence trend of tuberculosis in Baoshan District, Shanghai. Methods Data about the monthly incidence rate of tuberculosis in Baoshan District of Shanghai from 2009 to 2021 were collected, and a seasonal ARIMA model was constructed to verify the effect of the prediction model, and the causes of the prediction error were analyzed. Results The monthly incidence model of tuberculosis in Baoshan District was ARIMA (2,0,0)(0,1,1)₁₂, with the lowest BIC, Ljung-Box statistic Q=23.127, P=0.081, and the residual order was white noise. The actual monthly incidence rates in 2019 were basically consistent with the predicted values, and all were within the 95% confidence interval of the predicted value. Due to the impact of the COVID-19 pandemic, there was a large difference between the observed values and the predicted values in the past two years, and the observed values in February 2021 were outside the 95% confidence interval of the fitted values. Conclusion ARIMA (2,0,0)(0,1,1)₁₂ model can accurately predict the incidence trend of tuberculosisbefore the COVID-19 pandemic in Baoshan District.However, when affected by the COVID-19 pandemic, the prediction results have a large deviation, which requires re-modeling of tuberculosis incidence data in the post-epidemic era.