2009—2022年景德镇市手足口病流行特征及SARIMA模型的应用  

Prevalence characteristics of hand,foot and mouth disease and application of SARIMA model in Jingdezhen City from 2009 to 2022

作　　者：陈肖 彭志斌 刘苗 潘凯伦 姚瑶 李思思[1] CHEN Xiao;PENG Zhibin;LIU Miao;PAN Kailun;YAO Yao;LI Sisi(School of Public Health and Health Management,Gannan Medical University,Ganzhou,Jiangxi 341000,China;Department of Information,Jingdezhen Center for Disease Control and Prevention,Jingdezhen,Jiangxi 333000,China)

机构地区：[1]赣南医学院公共卫生与健康管理学院,江西赣州341000 [2]景德镇市疾病预防控制中心信息科,江西景德镇333000

出　　处：《职业与健康》2024年第21期2964-2970,共7页Occupation and Health

摘　　要：目的分析2009—2022年景德镇市手足口病(hand,foot and mouth disease,HFMD)疫情监测数据,了解HFMD流行特征。利用季节性时间序列预测模型对景德镇市HFMD的年发病率进行预测和拟合推断,并评估模型的准确性。方法从景德镇市疾病预防控制中心信息数据中心收集2009—2022年景德镇市HFMD月发病率数据,使用Excel 2022软件建立样本数据库和整合预测数值表,利用SPSS 25.0进行数据处理及绘图,根据残差白噪声检验和贝叶斯信息规则(bayesian information criterions,BIC)小原则选出最优模型,利用2009—2022年景德镇市HFMD月发病率数据建立相关模型并进行拟合外推,以2022年1—12月实际数据检验预测效果,采用绝对误差和相对误差评价模型可信程度。结果2009—2022年景德镇市HFMD累计报告达31989人,每年均有发病,主要集中在每年的4—6和9—10月,5月为发病高峰。采用季节性差分自回归移动平均模型(seasonal autoregressive integrated moving average model,SARIMA)对HFMD的发病率进行预测,发现SARIMA(1,0,1)(0,1,1)12模型的拟合效果最好。选择2022年12个月HFMD发病率的拟合推断值,实际值在预测值的95%CI范围内波动,整体预测较理想。结论景德镇市HFMD存在季节性和周期性,SARIMA模型比较准确地拟合和预测HFMD的发病趋势,可以用来预测短期内HFMD的发病率,为HFMD的防控提供一定的参考依据。Objective To analyze the epidemic monitoring data of hand-foot-and-mouth disease(HFMD)in Jingdezhen City from 2009to 2022,and understand the epidemic characteristics of HFMD.The seasonal time-series prediction model was utilized to predict and fit the annual incidence rate of HFMD in Jingdezhen City,and the accuracy of the model was evaluated.Methods The monthly incidence rate data of HFMD in Jingdezhen from 2009 to 2022 were collected from the information data center in Jingdezhen Center for Disease Control and Prevention.The sample database and the integrated prediction value table were established using Excel 2022 software,the data processing and plotting were carried out using SPSS 25.0 software,the optimal model was selected according to the residual white noise test and the Bayesian information criterions(BIC).The monthly incidence rate data of HFMD in Jingdezhen City from 2009 to 2022 were used to establish the relevant model and perform fitting extrapolation.The actual data from January to December 2022 were used to test the prediction effect.The absolute error and relative error were used to evaluate the credibility of the model.Results From 2009 to 2022,a total of 31989 cases of HFMD were reported in Jingdezhen City,occurring every year,mainly concentrated in April to June and September to October,with May being the peak period of incidence.The seasonal autoregressive integrated moving average model(SARIMA)was used to predict the incidence rate of HFMD,and the SARIMA(1,0,1)(0,1,1)12was found to be the best fit.The fitted extrapolated values of the12 month HFMD incidence rate in 2022 were selected,and the actual values fluctuated within the 95%CI range of the predicted value,and the overall prediction was ideal.Conclusion Seasonality and periodicity exist in HFMD in Jingdezhen,and the SARIMA model fits and predicts the trend of HFMD more accurately,which can be used to predict the incidence of HFMD in the short term and provide certain scientific reference for the prevention and control of HFMD.

关 键 词：手足口病 发病率 流行特征 自回归移动平均模型 预测 

分 类 号：R181.81[医药卫生—流行病学]