Data-Driven Study of the the COVID-19 Pandemic via Age-Structured Modelling and Prediction of the Health System Failure in Brazil amid Diverse Intervention Strategies

Askery Canabarro; Elayne Tenorio; Renato Martins; Lais Martins; Samurai Brito; Rafael Chaves

we follow a approach of Ref [9] . For the case of closure of school and universities (CSU), we use the census-based data of the number of students per age group to reduce the respective number of susceptible individuals. They are all supposed to be uninfected in the early stage of the epidemic and we assume that 100% of this target population will not disobey the policy as all schools and universities are closed in Brazil. The third column of Table II is the unit minus the values in the third column of Table I . For social distancing of people over 60 years old (SD60+) we assume that 75% will comply with this policy, meaning that 25% will leak the isolation, therefore the fraction 0.25 in the last three rows of Table II . For voluntary home quarantine (VHQ) and social distancing of the entire population (Quarantine) the compliance is 50% and 75%, meaning that the remaining quantity of susceptible individuals in each age group are given by factors of 0.5 and 0.25 of their population, respectively. See Table II to check g i for each NPI. As the measures are not capable of producing effects instantaneously, taking in general t β = 14 days to be effective [21] , we have to apply a modulation in the g i parameter in a similar fashion as done in Ref. [21] . Thus, we model g i as