摘要
目的通过构建季节性差分整合移动平均自回归模型(ARIMA模型)预测江苏省镇江市肺结核流行趋势并验证模型的有效性,探讨新型冠状病毒感染疫情对肺结核流行情况的影响。方法收集江苏省镇江市2014-2022年肺结核月发病数资料,构建季节性ARIMA模型,以2022年1-12月肺结核发病数验证预测模型效果,并分析预测误差产生的原因。结果2014-2022年镇江市共报告肺结核病例11316例,除2017、2019年发病率有所回升外,总体发病率呈下降趋势,发病主要集中在3-8月。ARIMA(1,1,1)(1,1,0)_(12)的BIC值(5.913)最小,残差白噪声也通过检验。但短期自相关部分的AR系数不显著,因此建立ARIMA(0,1,1)(1,1,0)_(12)。2022年镇江市肺结核月发病数实际值与预测值存在一定的偏差(平均相对预测误差为19.20%),但均在拟合值的95%可信区间内,实际月发病数(平均78例/月)与预测值(平均78例/月)变化趋势基本一致,模型拟合度较好,可用于预测镇江市肺结核流行情况。结论利用该模型对短期内镇江市肺结核发病数进行预测,认为镇江市肺结核流行总体上仍将长期保持下行趋势。
Objective To predict the trend of pulmonary tuberculosis prevalence in Zhenjiang City of Jiangsu Province by constructing a seasonal autoregressive integrated moving average(ARIMA) model and verify the effectiveness of the model,and to explore the impact of the COVID-19 pandemic on the prevalence state of pulmonary tuberculosis.Methods The pulmonary tuberculosis monthly incidence data during 2014-2022 in Zhenjiang City of Jiangsu Province were collected to construct a seasonal ARIMA model.The model′s predictive performance was validated by using the onset number of pulmonary tuberculosis from January to December 2022,and the causes of prediction errors were analyzed.Results A total of 1 316 cases of pulmonary tuberculosis were reported in Zhenjiang City during 2014-2022.The overall incidence rate showed a downward trend,except for the slight increase in 2017,2019.The onset was mainly concentrated from March to August.The ARIMA model with parameters(1,1,1)(1,1,0)12 had the lowest BIC value(5.913),and the white noise residuals also passed the test.However,the AR coefficient in the short-term autocorrelation was not significant,so the ARIMA model with parameters(0,1,1)(1,1,0)12 was established.There was a certain deviation between the actual value and predictive value in monthly incidence number of pulmonary tuberculosis in Zhenjiang City during 2022(average relative prediction error of 19.20%).However,all were within the 95% confidence interval of the fitted values.The change trend of the actual monthly incidence number(average 78 cases/month) was basically consistent with the predicted value(average 78 cases/month).The model fitting degree was well and could be used to predict the epidemiological situation of pulmonary tuberculosis in Zhenjiang City.Conclusion This model is used to predict the incidence number of pulmonary tuberculosis in Zhenjiang City in the short term,and it is considered that the overall trend of pulmonary tuberculosis epidemic in Zhenjiang City will remain the downward trend in the long run.
作者
伍鸿远
夏媛媛
WU Hongyuan;XIA Yuanyuan(School of Medical Administration,Nanjing Medical University,Nanjing,Jiangsu 211166,China)
出处
《现代医药卫生》
2024年第1期20-25,30,共7页
Journal of Modern Medicine & Health
基金
国家社会科学基金重大项目(20&ZD224)
江苏高校哲学社会科学重点研究基地项目(2020RWPT0101)
国家级大学生创新创业训练计划(202210312047Z)。
关键词
ARIMA模型
肺结核
传染病预测
新型冠状病毒感染
镇江
Autoregressive Integrated Moving Average model
Tuberculosis
Prediction of infectious diseases
COVID-19
Zhenjiang
