Modeling COVID-19 as a National Dynamics with a SARS-CoV-2 Prevalent Variant: Brazil - A Study Case

Celaschi, S.

COVID-19 global dynamics is modeled by an adaptation of the deterministic SEIR Model, which takes into account two dominant lineages of the SARS-CoV-2, and a time-varying reproduction number to estimate the disease transmission behavior. Such a methodology can be applied worldwide to predict forecasts of the outbreak in any infected country. The pandemic in Brazil was selected as a first study case. Brazilian official published data from February 25th to August 30th, 2020 was used to adjust a few epidemiologic parameters. The estimated time-dependence mean value to the infected individuals (confirmed cases) presents - in logarithmic scale - standard deviation SD = 0.08 for over six orders of magnitude. Data points for additional three weeks were added after the model was complete, granting confidence on the outcomes. By the end of 2020, the predicted numbers of confirmed cases in Brazil, within 95% credible intervals, may reach 6 Million (5 to 7), and fatalities would accounts for 180 (130 to 220) thousands. The total number of infected individuals is estimated to reach (13 +/- 1) Million, 6.2% of the Brazilian population. Regarding the original SARS-CoV-2 form and its variant, the only model assumption is their distinct incubation rates. The variant form reaches a maximum of 96% of exposed individuals as previously reported for South America.