Abstract
COVID-19 is an infection caused by the SARS-CoV-2 coronavirus, its first records were in the Chinese city of Wuhan in December 2019, and was considered by the World Health Organization (WHO) to be a worldwide pandemic in March 2020. In Brazil, COVID-19 spread to 27 states (UFs). As a result, decision-making to decrease the speed of transmission was based on WHO recommendations, where the main one is social isolation. However, due to the heterogeneity of the population in each of the UFs, the pandemic spread differently. Thus, it is interesting to group UFs by similarity due to some characteristics, and thus, observe the measures to combat COVID-19 carried out in each of these groups. The aim of this study was to group UFs using cluster analysis using the non-hierarchical k-means method considering the epidemiological coefficients such as incidence, prevalence, and lethality. The data were obtained from the website of the Ministry of Health of Brazil and consisted of the variables number of cases and new and accumulated deaths in UFs, in addition to the population at risk. For cluster analysis, the database was divided into three chronological periods for the three coefficients under study. With the cluster analysis, it was possible to verify the stratification of UFs according to their similarities in relation to COVID-19. Thus, the stratification of incidence, prevalence, and lethality by UFs can present itself as an additional resource to signal which places and which measures should be adopted and where these measures were effective.