传染病暴发疫情中传播动力学参数计算及疫情规模预测实现——基于R  

Calculation of transmission dynamics parameters and prediction of epidemic size in infectious disease outbreak—based on software R

作　　者：刘天[1] 侯清波[1] 姚梦雷[1] 黄继贵[1] 陈红缨[2] Liu Tian;Hou Qingbo;Yao Menglei;Huang Jigui;Chen Hongying(Department for Infectious Disease Control and Prevention,Jingzhou Prefectural Center for Disease Control and Prevention,Jingzhou 434000,Hubei,China;Hubei Provincial Center for Disease Control and Prevention,Wuhan 430079,Hubei,China)

机构地区：[1]荆州市疾病预防控制中心传染病防治所,湖北荆州434000 [2]湖北省疾病预防控制中心,湖北武汉430079

出　　处：《疾病监测》2022年第9期1211-1215,共5页Disease Surveillance

基　　金：荆州市2021年度医疗卫生科技计划(No.2021HC20)。

摘　　要：目的以美国新型冠状病毒肺炎(COVID-19)数据为例,利用R软件演示暴发疫情处置中代间隔(SI)、基本再生数(R_(0))、有效再生数(Re)、倍增时间和疫情规模的计算,为今后暴发疫情处置提供参考。方法收集美国2020年2月27日至2020年11月30日COVID-19逐日发病数据、国内2起疫情170对感染者与被感染者的代间隔(SI)数据。采用“fitdistrplus”包自助法对SI数据拟合“gamma”分布,得到SI均数和标准差。采用指数增长(EG)、极大似然法(ML)、序贯贝叶斯估计(SB)、时间相关再生数(TD)4种方法,利用获取的SI计算R_(0),选择拟合效果最好的作为R_(0)估计值,评价指标选择R^(2)。倍增时间基于EG的指数增长率进行换算。利用“EpiEstim”包计算Re;利用SEIARD模型估计不采取任何干预措施美国未来的病例数,用“deSolve”、“FME”包。结果170对感染者与被感染者数据拟合得到SI=4.78(95%CI:4.27~5.31),标准差sd=3.56。EG、ML、SB、TD拟合R_(0)依次为2.31(95%CI:2.30~2.31)、1.96(95%CI:1.95~1.97)、3.01(95%CI:2.53~3.55)、3.07(95%CI:2.20~4.22),R^(2)依次为0.95、0.73、0.99、0.96。倍增时间为3.75 d(3.74~3.77 d)。利用SEIARD估计理想状态下美国将于2020年5月30日全民感染,至2021年6月7日所有感染者清零,累计死亡5502466人。结论利用R软件能对R_(0)等传播动力学参数、疫情规模快速计算,在暴发处置中值得学习使用。动力学参数、疫情规模快速计算,在暴发处置中值得学习使用。Objective Taking the COVID-19 data of the United States as an example,using software R to calculate of the serial interval(SI),basic reproduction number(R_(0)),effective reproduction number(Re),doubling time and the number of COVID-19 using software R to provide a reference for the future epidemic response.Methods The daily incidence of COVID-19 in the United States from February 27,2020 to November 30,2020,and the SIs of 170 pairs of primary infection cases and and secondary infection cases in the two epidemics in the United States were collected.The“fitdistrplus”package bootstrap method was used to fit the“gamma”distribution of the SI to obtain the mean and standard deviation.Based on the obtained SI,four methods including exponential growth(EG),maximum likelihood(ML),sequential Bayesian estimation(SB),and time-dependent reproduction number(TD)were used to calculate R_(0).The R_(0) with the best fitting effect based on R^(2) was selected.The doubling time was converted based on the exponential growth rate of the EG method.The“EpiEstim”package was used to calculate Re.The SEIARD model was used to estimate the incidence of COVID-19 in the United States without any intervention,using the“deSolve”and“FME”packages.Results After the fitting of the gamma distribution of SIs of 170 pairs of primary infection cases and secondary infection cases,the SI was=4.78(95%CI:4.27−5.31)and the sd was=3.56.The fitted R_(0) of EG,ML,SB and TD methods were 2.31(95%CI:2.30−2.31),1.96(95%CI:1.95−1.97),3.01(95%CI:2.53−3.55),3.07(95%CI:2.20−4.22),the R^(2) were 0.95,0.73,0.99,0.96,respectively.The doubling time was 3.75 days(3.74−3.77 days).The SEIARD model estimated that all people in the United States would be infected by May 31,2020 without intervention measures,and no cases would be reported by June 7,2021,with a total of 5502466 deaths.Conclusion The use of software R can quickly calculate the transmission dynamics parameters such as R_(0) and predict the size of the epidemic,which is useful in outbr

关 键 词：代间隔 基本再生数 有效再生数 SEIARD 新型冠状病毒肺炎 

分 类 号：R211[医药卫生—中医学] R181