Economy or health? This false dichotomy used as an argument against pandemic control measures in Brazil has lost yet another nuance. If decision-makers were guided by ideal information that allowed them to adopt social distancing measures at the right moment and in the right dose, more lives could be saved, and the economic impact could be reduced during the pandemic. To make this possible, a predictive control system was developed within the CoVida Network, the result of a scientific collaboration between researchers from the Center for Data Integration and Health Knowledge (Cidacs/Fiocruz Bahia) and the control and automation engineering department of the Polytechnic School of the Federal University of Bahia (UFBA), which led the research.
The study was recently published in Scientific Reports, of the Nature group. In addition to Cidacs and UFBA, the paper is signed by researchers related to international institutions, such as the University of Almería, in Spain, the University of Oporto, in Portugal, and the Swansea University, in Wales, in addition to the Federal University of Santa Catarina and the University of São Paulo, both in Brazil.
In simulations made by the researchers, analyzing past measures and in an optimistic scenario, adopting the system could have avoided up to 64% of cases, hospitalizations, and deaths with better administration of the periods in which social distancing was enforced. “However, it is important to highlight that this is a hindsight analysis, and society’s compliance with government measures is crucial for the system to work”, emphasizes UFBA researcher Marcus Americano da Costa, one of the leaders of the study.
“Here in the state of Bahia, for instance, we have recently adopted stricter measures as the Saint John festivities arrived. With the method developed in the study, it is possible to make non-stop analysis anywhere, using objective metrics according to the level of transmission of the virus in the region under study and to dose restriction measures in a more suitable and effective manner”, says Pablo Ramos, Cidacs researcher.
The system developed by the group allows for the analysis, evaluation, and prediction, over time, of susceptible, exposed, symptomatic and asymptomatic infected individuals, occupation of hospital and ICU beds, recovered patients, and lethal cases. In addition, a social behavior variable was incorporated to the model, developed on a mapping of municipal and state decrees and population mobility.
To build the control system, researchers used computing techniques in areas such as dynamic systems, system identification, optimization, mathematical modelling, and control engineering. The control system was implemented on an epidemiological model previously proposed by researchers of the CoVida Network, which represents, in a realistic fashion, the non-linear, dynamic behavior of COVID-19 propagation, based on data such as hospitalizations in regular hospital beds, ICU beds, and mortality.
These combined techniques culminated in a system that generates predictions on the pandemic for the upcoming days, and therefore suggests the best decisions to be made to control disease propagation. To validate the results obtained through the system, researchers carried out simulations with real figures from different locations and made a detailed analysis of the evolution of the pandemic in Bahia, also assessing the effects of population compliance to restriction measures.
“From some of the innovations, such as a rigorous model that associates the epidemical response to social mobility indexes, our main result is that the study proposed here makes it possible to define the intensity of public policies of restriction (or flexibilization) that minimize infections and lethal cases, prevent the collapse of health systems, and do not needlessly jeopardize financial activities. To sum it up, these strategies allow for a better balancing of impacts on health and on the economy”, explains Americano da Costa.
The authors highlight that, even with the ongoing vaccination of the population, the adoption of the system developed here will be useful, not only for locations without sufficient quantities of vaccines, but also for other epidemics. “In the absence of a generalized vaccination program that can induce herd immunity, strategies to co-exist with the virus while simultaneously minimizing the risk of outbreaks are crucial, something that must work in parallel with the return of a ‘normal’ life”, states the article.
The system is already available to be adopted by any government manager, but in order to be used, it must be programmed specifically in a computer and requires that the administrator in question have basic knowledge in data analysis and interpretation of graphs. However, the researchers are now working to create an environment that is accessible for ordinary users and a model that includes the effects of vaccines, to adapt to the current moment.
The development of the system reinforces the need to bridge the gap between scientific knowledge and public management, especially in cases of large impact on social life, which is the case of the current pandemic, as mentioned by Mauricio Barreto, one of the leaders of the study and Cidacs coordinator. “Science has given great contributions to the control of the pandemic, but government managers are always lacking in instruments that allow for a more integrated management of the pandemic. The study shows the possibility of the existence of these tools by integrating knowledge originated in the modeling of complex systems, epidemiology, and control engineering”.