| Objective:According to the collected data of COVID-19 cases from 2020 to 2022 published on the websites of health committees of 13 provinces(municipalities)including Beijing,the epidemic characteristics of overseas imported cases were studied;at the same time,a combined SIR-multiple linear regression combined prediction model suitable for the prevention and control of overseas imported COVID-19 was established in combination with the Susceptible-Infected-Recovered(SIR)model and multiple linear regression model to provide a reference for the risk assessment of overseas imported COVID-19 at frontier ports.Methods:Firstly,the information of imported cases and countries of origin released by the health commissions of 13 provinces(municipalities)was collected,and secondly,the COVID-19 epidemic data of the countries of origin of imported cases(including cumulative confirmed cases,vaccination,variant infection,etc.)were collected on the WHO and Our World in Data data websites of the World Health Commission according to the information of imported cases collected by the health commission,and finally the epidemiological distribution of imported cases was described.The data from 2020 to2021 is the training set,and the data from 2022 is the test set,and after processing the data,the random coefficient method is used to construct the SIR model for predicting infectious diseases,and the epidemic situation of the country where the cases are exported is fitted,and finally the feature data set with direct significance with the infection status of each country is obtained.Then,the feature set data fitted by SIR model is divided into training set and test set(7:3)by random sampling method,and the multiple linear regression model is fitted and tested,and finally a SIR-multiple linear regression combination prediction model is established for overseas input.Results:1.Time distribution:From March 2020 to June 2022,the months in which imported cases exceeded 300 were as follows:In March 2020,527 confirmed cases(8.23%)and21 asymptomatic infections(0.70%);In April 2020,there were 357 confirmed cases(5.58%)and 34 asymptomatic infections(1.13%);In September 2020,there were 188confirmed cases(2.94%)and 140 asymptomatic infections(4.65%);In October 2020,there were 255 confirmed cases(3.98%)and 106 asymptomatic infections(3.52%);In May 2021,there were 185 confirmed cases(2.89%)and 187 asymptomatic infections(6.22%);In June 2021,there were 170 confirmed cases(2.66%)and 155 asymptomatic infections(5.15%);In July 2021,there were 354 confirmed cases(5.53%)and 235asymptomatic infections(7.81%);In August 2021,there were 397 confirmed cases(6.20%)and 243 asymptomatic infections(8.08%);In September 2021,there were 342confirmed cases(5.34%)and 161 asymptomatic infections(5.35%);In November 2021,there were 184 confirmed cases(2.87%)and 128 asymptomatic infections(4.26%);In December 2021,there were 409 confirmed cases(6.39%)and 39 asymptomatic infections(1.30%);In January 2022,there were 763 confirmed cases(11.92%)and 122asymptomatic infections(4.06%);In February 2022,there were 436 confirmed cases(6.81%)and 201 asymptomatic infections(6.68%);In March 2022,there were 245confirmed cases(3.83%)and 201 asymptomatic infections(6.68%).2.Spatial distribution:3347 cases(35.56%)were imported into Guangdong Province,2830 cases(30.07%)in Shanghai,558 cases(5.93%)in Yunnan Province,510cases(5.42%)in Sichuan Province,468 cases(4.97%)in Fujian Province,461 cases(4.90%)in Tianjin City,384 cases(4.08%)in Shaanxi Province,323 cases(3.43%)in Heilongjiang Province,190 cases(2.02%)in Beijing Province,124 cases(1.32%)in Shandong Province,103 cases(1.09%)in Hubei Province,84 cases(0.89%)in Gansu Province and 29 cases(0.31%)in Jilin Province.3.Recent country of residence distribution:Overseas imported cases mainly came from Asia,with the top ten countries with the most imported cases being the United States 992(10.5%),Russia 669(7.1%),Myanmar 563(6.0%),the United Kingdom 548(5.8%),Canada 367(3.9%),the Philippines 358(3.8%),the United Arab Emirates 343(3.6%),Japan 323(3.4%),France 279(3.0%),and Singapore 275(2.9%).4.Model Predictions:A combined SIR-multiple linear regression model was utilized to predict the number of COVID-19 cases resulting from foreign imports in China.The SIR model exhibited a satisfactory fit to the majority of the curves in the foreign epidemic data,with R~2values exceeding 0.75.This model proved effective in predicting the infection status of various countries and regions.After removing insignificant variables,the optimized model obtained through multiple linear regression demonstrated overall significance(P<0.05),with an adjusted R~2value of 0.7,indicating a good fit of the model.Conclusions:From March 2020 to June 2022,the months with higher numbers of imported cases were March,April,and October 2020,as well as July,August,September,and December2021,and January,February,and March 2022.Guangdong Province and Shanghai Municipality in China had a higher number of imported COVID-19 cases,primarily originating from the United States,Russia,and Myanmar.There was no apparent seasonality in the timing of the imported cases.By utilizing the data of imported cases,a combination predictive model for overseas-imported infections was constructed.The SIR model performed well in predicting the infection situation of overseas outbreaks,while the multiple linear regression model was somewhat deficient in predicting imported cases.However,the combined SIR-multiple linear regression model showed an overall good predictive performance. |
