[Fitting and predicting trend of COVID-19 by SVEPIUHDR dynamic model]

Abstract

Objective: To fit and predict the trend of COVID-19 epidemics in the United States (USA) and the United Kingdom (UK), and analyze the effect of vaccination. Methods: Based on the SEIR dynamic model, considering the presymptomatic infections, isolation measures, vaccine vaccination coverage, etc., we developed a SEIR with vaccine inoculation, Presymptomatic infectious, unconfirmed infectious, hospital isolation and domiciliary isolation dynamics model. The publicly released incidence data of COVID-19 from November 6, 2020 to January 31, 2021 in USA and from November 23, 2020 to January 31, 2021 in UK were used to fit the model and the publicly released incidence data of COVID-19 from February 1, 2021 to April 1 were used to evaluate the predicting power of the model by software R 4.0.3 and predict changes in the daily new cases in the context of different vaccination coverage. Results: According to the cumulative confirmed cases, the fitting bias and the predicting bias of the SVEPIUHDR model for USA and UK were less than 5%, respectively. From the model prediction results, the cumulative cases after COVID-19 vaccination in USA in early April reached 31 864 970. If there had not had such vaccination, the cumulative cases of COVID-19 would have reached to 35 317 082, with a gap of more than 3.4 million cases. In UK, the cumulative cases of COVID-19 after the vaccination was estimated to be 4 195 538 in early April, compared with 4 268 786 cases if no COVID-19 vaccination had been provided, there would have heen a gap of more than 70 000 cases. Conclusion: SVEPIUHDR model shows a good prediction effect on the epidemic of COVID-19 in both USA and UK.