Applying a Prospective Space-Time Scan Statistic to Examine the Evolution of COVID-19 Clusters in the State of São Paulo, Brazil

Abstract

We present the first geographic study that uses space-time statistics to monitor COVID-19 in Brazil. The first cases of COVID-19 were confirmed in December 2019 in Wuhan, China, caused by the contamination of the SARS-CoV-2 virus, and quickly turned into a pandemic. In Brazil, the first case occurred on January 23rd, 2020 but was officially reported by the Brazilian Ministry of Health on February 25th. Since then, the number of deaths and people infected by COVID-19 in Brazil have been steadily increasing. Despite the underreporting of coronavirus cases by government agencies across the country, the State of Sao Paulo has the highest rate among all Brazilian States. Thus, it is essential to detect which areas contain the highest concentration of COVID-19 to implement public policies, to mitigate the spread of the epidemic. To identify these critical areas, we utilized daily confirmed case data from the Brasil.IO website between February 25th, 2020 to May 5th, 2020; which were aggregated to the municipality level. A prospective space-time scan statistic was applied to evaluate possible active clusters in three different time periods. The results visualize the space-time evolution and dynamics of COVID-19 clusters in the State of Sao Paulo. Since the first study period, the results highlight approximately 4.6 times the number of municipalities belonging to a significant cluster with a RR>1 on May 5th. These results can inform health authorities and public management to take the necessary measures to minimize the transmission of COVID-19 and track the evolution of significant space-time clusters.

Space-Time Patterns, Change, and Propagation of COVID-19 Risk Relative to the Intervention Scenarios in Bangladesh

“…Spatial, temporal, and spatio-temporal statistics are essential approaches for identifying transmission dynamics of COVID-19, which can enable evaluating ongoing efforts as well as inform new, innovative solutions for containing the disease [ 16 , 17 ]. Existing literature shed light on the spatiotemporal dynamics of COVID-19, which helped identify the extent and impact of the pandemic as well as intervention strategies in different parts of the world including China [ 2 , 18 , 19 ], Italy [ 20 ], Spain [ 21 ], Brazil [ 22 ], and the United States [ 5 , 23 , 24 ]. In this paper, we leveraged a prospective space-time scan statistic approach [ 25 ] to detect currently “active” and emerging clusters of COVID-19 outbreak in Bangladesh.…”

A syndromic surveillance tool to detect anomalous clusters of COVID-19 symptoms in the United States

“…Scan statistics are available using the free software SaTScan 29 . SaTScan has been widely used in both disease and syndromic surveillance, including dengue fever and chikungunya 17 , 30 , 31 , sexually transmitted disease 32 , foodborne illness 33 , respiratory infections and common illnesses 34 , 35 , and COVID-19 31 , 36 , 37 , among countless others. Essentially, scan statistics determine if the number of disease cases/symptoms in a defined area and proximal in time are greater than the expected number of cases/symptoms, such as the underlying population contained in the study area or distribution of point-level events.…”

Space-time patterns, change, and propagation of COVID-19 risk relative to the intervention scenarios in Bangladesh

“…Figure 2 Snapshots of print media reports between April 27 and April 30 when a large number of people in Bangladesh congregated outside amidst nationwide lockdown Spatial, temporal, and spatio-temporal statistics are essential approaches for identifying transmission dynamics of COVID-19 which can enable evaluating ongoing efforts as well as inform new, innovative solutions for containing the disease (Franch-pardo et al, 2020;Xiong et al, 2020). Existing literature shed light on the spatiotemporal dynamics of COVID-19, that helped identify the extent and impact of the pandemic as well as intervention strategies in different parts of the world including China (Guan et al, 2020;Li et al, 2020;Su et al, 2020), Italy (Gatto et al, 2020), Spain (Orea & Álvarez, 2020), Brazil (Ferreira et al, 2020), and the United States (Desjardins et al, 2020;Mollalo, Vahedi, & Rivera, 2020). In this paper, we leveraged a prospective space-time scan statistic approach (Kulldorff et al, 2005) to detect currently "active" and emerging clusters of COVID-19 outbreak in Bangladesh.…”