Spatially Explicit Modeling of 2019-nCoV Epidemic Trend based on Mobile Phone Data in Mainland China
Xiaolin Zhu, Aiyin Zhang, Shuai Xu, Pengfei Jia, Xiaoyue Tan, Jiaqi Tian, Tao Wei, Zhenxian Quan, Jiali Yu
doi: https://doi.org/10.1101/2020.02.09.20021360
Abstract
In December 2019, Wuhan, China reported an outbreak of atypical pneumonia caused by the 2019 novel coronavirus (2019-nCoV). As of February 7, 2020, the total number of the confirmed cases in mainland China reached to 34,546 of whom 722 have died and 2,050 recovered. While most Chinese cities have confirmed cases, the city-level epidemical dynamics is unknown. The aim of this study is to model the dynamics of 2019-nCoV at city level and predict the trend under different scenarios in mainland China. We used mobile phone data and modified the classic epidemiological Susceptible - Infectious - Recovered (SIR) model to consider several unique characteristics of the outbreak of 2019-nCoV in mainland China. The modified SIR model was trained using the confirmed cases from January 25 to February 1 and validated by the data collected on February 2, 2020. The prediction accuracy of new infected cases on February 2 (R2 = 0.94, RMSE = 18.24) is higher than using the classic SIR model (R2 = 0.69, RMSE = 40.18). We used the trained model to predict the trend in the next 30 days (up to March 2, 2020) under different scenarios: keeping the early-stage trend, controlling the disease as successfully as SARS in 2003, and increasing person-to-person contacts due to work/school resuming. Results show that the total infected population in mainland China will be 10.53, 0.15, and 0.41 million and 67%, 100%, 91% Chinese cities will control the virus spreading by March 2, 2020 under the above three scenarios. Our study also provides the city-level spatial pattern of the epidemic trend for decision makers to allocate resources for controlling virus spreading.
*注，本文为预印本论文手稿，是未经同行评审的初步报告，其观点仅供科研同行交流，并不是结论性内容，请使用者谨慎使用.
