基于ARIMA乘积季节模型预测耐碳青霉烯类鲍氏不动杆菌流行趋势  

Multiplicative seasonal model based on ARIMA for prediction of prevalence trend of carbapenem-resistant Acinetobacter baumannii

作　　者：王玉沐 尹伟琴 杨乐[2] WANG Yu-mu;YIN Wei-qin;YANG Le(Changzhou Jintan First People's Hospital,Changzhou,Jiangsu 213000,China;不详)

机构地区：[1]常州市金坛第一人民医院感染管理科,江苏常州213000 [2]常州市第二人民医院感染管理科,江苏常州213000

出　　处：《中华医院感染学杂志》2024年第17期2575-2579,共5页Chinese Journal of Nosocomiology

基　　金：江苏省医院协会医院管理创新研究课题(JSYGY-3-2023-400)。

摘　　要：目的利用自回归求和移动平均(ARIMA)乘积季节模型,预测耐碳青霉烯类鲍氏不动杆菌(CRAB)的流行趋势,为医院制定针对性的防控策略提供理论依据。方法使用2018-2022年常州市金坛第一人民医院每月临床分离的CRAB菌株数量作为数据集,建立ARIMA乘积季节模型,并使用2023年的数据作为验证集,与模型预测值进行对比,从而评估模型的预测性能。结果2018-2022年医院CRAB检出率整体呈上升趋势(P<0.05);痰标本为主要检出来源,每年的3-4月为发病高峰,具有周期性和季节性;最优模型ARIMA(0,1,2)(0,1,1)12的正态化的贝叶斯信息准则(BIC)为3.867,残差序列Box-Ljung检验显示差异无统计学意义(Q=11.109,P=0.745),模型拟合良好;模型预测2023年CRAB的检出数量与实际值的平均相对误差为21.35%,实际值均在预测值的95%CI之内。结论ARIMA乘积季节模型能较好的预测CRAB的流行趋势,可为CRAB感染的短期预测、动态分析并及时采取针对性的防控措施提供理论依据。OBJECTIVE To predict the prevalence trend of carbapenem-resistant Acinetobacter baumannii(CRAB)by using autoregressive integrated moving average(ARIMA)multiplicative seasonal model so as to provide theoretical basis for formulating targeted prevention and control strategies.METHODS The numbers of the CRAB strains that were monthly isolated from Changzhou Jintan First People′s Hospital from 2018 to 2022 were chosen as the dataset.ARIMA multiplicative seasonal model was established,the data of 2023 were set as the validation set and was compared with the predictive value of the model,and the predictive performance of the model was assessed.RESULTS The isolation rate of CRAB showed an overall upward trend from 2018 to 2022(P<0.05).The strains were mainly isolated from sputum specimens,and the prevalence of the strains was at the peak in March-April every year,showing the periodical and seasonal characteristics.The normalized Bayesian information criterion(BIC)of the optimal ARIMA(0,1,2)(0,1,1)12 was 3.867,the residual sequence Box-Ljung test showed that there was no significant difference(Q=11.109,P=0.745),and the model was well fitted.The mean relative error of the model was 21.35%in prediction of the number of isolated CRAB strains and the actual values,and the actual values all fell within the 95%confidence interval(95%CI)of the predicted values.CONCLUSION ARIMA multiplicative seasonal model can predict the prevalence trend of the CRAB strains and provide theoretical bases for short-term prediction,dynamic analysis of the CRAB infection so as to take targeted prevention and control measures.

关 键 词：自回归求和移动平均模型 乘积季节模型 耐碳青霉烯类鲍氏不动杆菌 流行趋势 

分 类 号：R18[医药卫生—流行病学]