A Cybernetics-based Dynamic Infection Model for Analyzing SARS-COV-2 Infection Stability and Predicting Uncontrollable Risks

Wenlei Xiao; Qiang Liu; J Huan; Pengpeng Sun; Liuquan Wang; Chenxin Zang; Sanying Zhu; Liansheng Gao

At the beginning, when the novel coronavirus from Wuhan causes concern in public, delayed and missed detection may exist in the reported number of cases. It brings troubles in estimating the epidemiological parameters and epidemic predictions [4, 5] . Most estimated R 0 ranges from 1.5 to 4.0 [1, 6, 7, 8] . As of Jan 29, 2020, the first investigated incubation period from patients was reported, which had a mean of 5.2 days(95% confidence interval [CI], 4.1 to 7.0) and followed a Poisson distribution, and the basic reproductive number was estimated to be 2.2 (95% CI, 1.4 to 3.9) [6] . However, the early sampled 425 patients had a median age of 59 years and 56% were male. Given those bias on the samples, the estimated epidemiological parameters may have deviations that might lead to great errors in the simulation. In the view of this, we used the data from Shanghai, a relatively well controlled city, to identify and calibrate the key parameters of the incubation period and the basic reproductive number. Subsequently, those parameters were used to evaluate the status of other cities (except for those cities in Hubei province). In Shanghai Model, there is no worry about the shortage of medical supplies, so a negative summation channel performs a direct control effect on the positive feedback infection loop, which is thus of paramount importance in reducing the number of total infectious cases. Two factors in the system can be regulated by the administration, C 0 and C 1 . The confirmation delay does not affect the infecting process, but brings a direct hinder in inspecting the real number of confirmed cases. The input of the Shanghai Model mainly comes from the inspected cases from Wuhan, so the number of daily imported cases should be given by R 1 x 1 , where R 1 is the infection rate of immigration, and x 1 is the daily immigrated population from Wuhan. On Jan 23, 2020, Wuhan went into lockdown to contain the outbreak of the epidemic. Since then, the input from immigration was switched off. Those sudden events changed the system dynamics dramatically, and is difficult to be estimated or approximated by other transmission models. Other similar cities can reuse the Shanghai Model with some modification on the parameters C 0 and C 1 . Nonetheless, the 4 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.