The starting dates of COVID-19 multiple waves

TL;DR

This paper develops a statistical methodology to identify the start dates of multiple COVID-19 waves, applied to Rio de Janeiro, revealing the virus was circulating earlier than initially detected, likely during carnival festivities.

Contribution

It introduces a parametric statistical approach combining surveillance data, nonlinear regression, and information criteria to pinpoint outbreak start dates.

Findings

Original virus strain circulating in late February 2020

Method successfully identifies multiple outbreak start dates

Early circulation linked to carnival festivities

Abstract

The severe acute respiratory syndrome of coronavirus 2 spread globally very quickly, causing great concern at the international level due to the severity of the associated respiratory disease, the so-called COVID-19. Considering Rio de Janeiro city (Brazil) as an example, the first diagnosis of this disease occurred in March 2020, but the exact moment when the local spread of the virus started is uncertain as the Brazilian epidemiological surveillance system was not widely prepared to detect suspected cases of COVID-19 at that time. Improvements in this surveillance system occurred over the pandemic, but due to the complex nature of the disease transmission process, specifying the exact moment of emergence of new community contagion outbreaks is a complicated task. This work aims to propose a general methodology to determine possible start dates for the multiple community outbreaks of COVID-19, using for this purpose a parametric statistical approach that combines surveillance data, nonlinear regression, and information criteria to obtain a statistical model capable of describing the multiple waves of contagion observed. The dynamics of COVID-19 in the city of Rio de Janeiro is taken as a case study, and the results suggest that the original strain of the virus was already circulating in Rio de Janeiro city as early as late February 2020, probably being massively disseminated in the population during the carnival festivities.