季节性自回归移动平均模型在秦皇岛市手足口病发病预测分析中的应用研究  

Applied research of seasonal ARIMA in prediction and analysis of hand-foot-mouth disease in Qinhuangdao

作　　者：陶旭[1] 史明坤[1] 董俊善[1] 郑立新[1] 杨健 韩子寅 TAO Xu;SHI Ming-kun;DONG Jun-shan;ZHENG Li-xin;YANG Jian;HAN Zi-yin(Qinhuangdao Center for Disease Control and Prevention,Hebei 066000,China)

机构地区：[1]秦皇岛市疾病预防控制中心,河北066000

出　　处：《医学动物防制》2022年第3期230-234,238,共6页Journal of Medical Pest Control

基　　金：秦皇岛市级科学技术研究与发展计划—重大突发公共卫生事件风险评估模型研究(202003B027)。

摘　　要：目的建立季节性自回归移动平均模型(Autoregressive Integrated Moving Average Model,ARIMA),并对秦皇岛市手足口病的发病率进行预测及效果评估,为今后手足口病防控措施的制定提供科学依据。方法通过"中国疾病预防控制信息系统"导出2009—2018年秦皇岛市手足口病确诊病例和临床诊断病例。利用SPSS 23.0软件对2009—2018年秦皇岛市手足口病发病率进行逐月统计,采用季节性ARIMA模型进行建模拟合及预测分析。结果秦皇岛市2009—2018年各月均有手足口病病例报告,且发病呈明显的季节性分布特征,病例主要集中在6—8月,7月为发病高峰,个别年份在10月或11月病例数略有上升,且具有每隔1~2年发病强度增强的周期性。选用季节性ARIMA模型对手足口病发病进行预测,以ARIMA(2,0,0)(0,1,1);模型最优,2019年各月手足口病发病率实际值均在预测值的95%CI范围内。结论季节性ARIMA模型能较好地拟合手足口病发病率变化趋势,可用于手足口病发病的预测预警,为手足口病防控工作提供依据。Objective To conduct the estimates and effectiveness evaluation on the incidence rate of hand-foot-mouth disease(HFMD)in Qinhuangdao by setting up the seasonal Autoregressive Integrated Moving Average Model(ARIMA),and to provide scientific basis for establishing prevention and control measures for HFMD in the future.Methods The confirmed cases and clinical cases of HFMD in Qinhuangdao from 2009 to 2018 were exported through the“China Disease Prevention and Control Information System”.The incidence rate of HFMD in Qinhuangdao was analyzed by SPSS 23.0 software on the month-by-month basis.Seasonal ARIMA was used for modeling,fitting and forecasting analysis.Results HFMD cases were reported in Qinhuangdao during years thorough from 2009 to 2018,and the incidence was featured with seasonal distribution,the cases were mainly concentrated during months from June to August,and the incidence peak was in July.The number of cases increased slightly in October or November in years,and the incidence intensity increased periodically every 1-2 years.Seasonal ARIMA was selected to predict the incidence of HFMD.ARIMA(2,0,0)(0,1,1);model was the best,and the actual incidence of HFMD incidence was within the 95%CI range of prediction in 2019.Conclusion Seasonal ARIMA is capable of fitting the trend of incidence rate of HFMD in better way,and is eligible to be used for prediction and early warning of HFMD and to provide basis for prevention and control of HFMD.

关 键 词：手足口病 发病率 自回归移动平均模型 预测 

分 类 号：R512.5[医药卫生—内科学]