Study shows advantage of alternating quarantines in cities across the state of São Paulo

Strategy would maintain economic activity and protect healthcare services, according to system developed by a research center supported by FAPESP.

By Elton Alisson  |  Agência FAPESP – Researchers affiliated with the Center for Mathematical Sciences Applied to Industry (CeMEAI) have developed a system that indicates the best quarantine strategies for cities to control the spread of COVID-19.

According to the system’s simulations, cities in the state of São Paulo, for example, could begin isolating at different times in order to maintain state-wide economy activities while at the same time protecting the health service.

The system is described in an article published on the medRxiv platform, in a pre-print (not yet peer-reviewed) version. The source codes have been uploaded to GitHub, where they can be accessed by policymakers in the cities interested in using the system.

CeMEAI is one of the Research, Innovation and Dissemination Centers (RIDCs) funded by FAPESP and is hosted by the University of São Paulo’s Institute of Mathematics and Computer Sciences (ICMC-USP) in São Carlos.

“We intend to offer guidance to municipal governments that want to use the system, which can simulate scenarios in which cities don’t need to quarantine at the same time,” Tiago Pereira, a professor at ICMC-USP and a co-author of the article, told Agência FAPESP. 

According to Pereira, who studies mathematical modeling of infectious disease dynamics, if an entire region such as the state of São Paulo follows a unified quarantine protocol to control COVID-19 and implements social isolation in a synchronized manner, the result will be a longer period of restricted mobility.

Infectious diseases are not transmitted at the same rate in every location, he explained. Moreover, cities have very different population dynamics, economic needs and healthcare infrastructures.

“The quarantine protocol for Ribeirão Preto [the third most populous city in the interior of the state of São Paulo] can’t be the same as for São Pedro [with under 50,000 inhabitants],” Pereira said.

The system contributes to the decentralization of quarantine protocols by taking into account the population of each municipality, commuting and other forms of travel to nearby cities (estimated from anonymized smartphone usage data), local healthcare infrastructure in terms of the number of intensive care beds available, and the progress of the disease over time in the region.

Based on these parameters, the system suggests the ideal level and duration of the restrictive measures required to ensure that the number of infected people does not reach the point where it may cause a breakdown of local healthcare services. 

The system can also be used to coordinate quarantine strategies across several cities so that they do not have restrictions with the same level and duration.

“The model can alternate periods of lockdown or strict shelter-in-place with more relaxed measures and more normal activity,” Pereira said.

Simulations for São Paulo 

To validate the model, the researchers used data for the main cities in the state of São Paulo to estimate travel between them, healthcare service capacity and the number of infected people, simulating quarantine scenarios with either the official number of cases registered by the Brazilian health ministry or the probable number assuming that for every confirmed case of COVID-19 there are 12 more cases that have not been notified.

The results of the simulations showed that if the quarantine policy in place in the state of São Paulo is maintained until the end of June and the rate of contagion also remains unchanged, the number of people infected will peak early in July in the state capital and in August in the rest of the metropolitan region. 

Under this scenario more than 2% of the population of São Paulo City would be infected in two months, between the start of June and the start of August.

“The disease would peak a little faster in São Paulo, but the city’s health system would be overwhelmed,” Pereira said.

The best-case scenario showed that the city’s health system (including field hospitals) has the capacity to treat 1.5% of the population.

Further simulations were run using this estimate, assuming peak infection rates of 1.5% for the state capital and 0.7% for the rest of the state, and an extension of quarantine restrictions in the capital until September to avert a collapse of the city’s health system.

Under this scenario, lockdown would be necessary in São Paulo City now to bring down the contagion curve at the end of August, while cities in the rest of the state would need to implement intermittent quarantine protocols until the end of the year.

Cities in the north of the state such as Araçatuba, Ribeirão Preto and São José do Rio Preto would be able to start quarantine two weeks later than cities in the metropolitan area, according to the simulations. They would need to remain under quarantine until the end of September, alternating two weeks of isolation with two weeks of unrestricted mobility.

Most cities in the São Paulo metropolitan area could begin fortnightly intermittent social distancing in mid-August. Economic activities would not therefore need to shut down completely, especially where agricultural activities are concentrated, according to the projections.

“We found that even if the number of cases is multiplied by 12 to take undernotification into account it’s possible to protect city healthcare services by means of optimal quarantine control, according to the simulations performed by our modeling system,” Pereira said. 

The system can also simulate changes in situations, such as the impact of rising availability of intensive care beds in the state capital on control of the disease in the interior. 

“Currently São Paulo City is sending COVID-19 patients to the interior because it doesn’t have enough intensive care beds, but the flow will be reversed when the disease is controlled in the capital and still spreading in the interior. Our system can measure the impact of this change,” Pereira said.

The article “Robot dance: a city-wise automatic control of COVID-19 mitigation levels” (https://doi.org/10.1101/2020.05.11.2009854) by Paulo J. S. Silva, Tiago Pereira and Luis Gustavo Nonato can be read at: www.medrxiv.org/content/10.1101/2020.05.11.20098541v1.