key: cord-0958229-obehsz2z authors: Silva, H. M.; Pereira, R. S.; Gritz, R. d. A.; Simoes, A.; Porto, F. A. title: Projection of hospitalization by COVID-19 in Brazilfollowing different social distances policies date: 2020-05-01 journal: nan DOI: 10.1101/2020.04.26.20080143 sha: ea5b82fa2641f2c7119d62b20912a66d16af3706 doc_id: 958229 cord_uid: obehsz2z Following the first infections in the Wuhan City, COVID-19 became a global pandemic as declared by the World Health Organization (WHO). Since it is an airborne disease transmitted between humans, many countries adopted a quarantine on their own population as well as closing borders measures. In Brazil many are discussing the best way to manage the opening of the quarantine under the constraints of hospitals infrastructure. In this work we implement a forecast of the demand for hospital beds for the next 30 days for every state in Brazil for different quarantine flexibilization scenarios and analyse how long it would take for the demand to exceed current available hospital beds. COVID-19, a disease caused by the SARS-CoV-2 virus, was first identified on December 31, 2019 in Wuhan City, and within weeks, it became a disease recognized as a global pandemic [14] by the World Health Organization. The disease began spreading in São Paulo, Brazil, on February 26, 2020, and later spread throughout the country. Subsequently its arrival in Brazil, the quarantine of the population began due to the high rate of transmission, lack of a vaccine which would result in high overcrowding of the health system.Today after almost a month of the adoption of the quarantine in Brazil, dialogues are occurring about its possible future. Evaluating this scenario, this article consists of the prediction in the next 30 days for the advancement of COVID-19 in Brazil for each state, analyzing when the health system would collapse, based on the 3 scenarios of quarantine adoption. Quarantine scenarios were based on the adoption of public policies on 3 countries [12] in the first 30 days of infection: South Korea (quarantine with more incisive policies, with the inexistence of prohibition of populational movement) [3] ; Germany (quarantine with more balanced public policies, recommending the use alcohol gel, masks and agglomerations, but without prohibiting the free movement of the population) [2] ; Italy (tenuous quarantine, without any countermeasure) [11] ,[9], [8] . The estimates will be compared against the number of available beds obtained from CNES[7] [1] , removing only the pediatric and neonatal beds, thus admitting that any other bed could receive a patient affected by the COVID-19, in such a way as to indicate up to when, in the best case or not, would supply the needed demand. The rest of the work is divided into session II methodology, III results, and IV conclusions. In this work we propose to analyze how COVID-19 will spread on each state in Brazil. To do so, we gathered data about COVID-19 spread over the world from John Hopkins resource center. Data about the spread of each Brazilian state was gathered from official sources [13] . The amount of available beds for COVID-19 patients was obtained from the CNES database on DATASUS. After obtaining the data the following procedure is applied: CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 1, 2020. . https://doi.org/10.1101/2020.04.26.20080143 doi: medRxiv preprint From the John Hopkins Coronavirus Resource Center data we subset only the information from our three scenarios which are South Korea, Germany, and Italy. The number of confirmed and dead cases is then normalized by dividing it by the population size of the respective country. This way we can compare different countries since we will be looking at the percentage of the infected population on then. Since COVID-19 has as its transmission vectors human themselves, it most basic model is an exponential function given by the solution of the equation dy dt = αy, where α denotes the infectivity of the disease. This model of course does not account for recovered or dead people from the disease, it also does not consider other measures that can be done to suppress infectivity. Since when taking the account these different scenarios our solution may change, and even these scenarios may happen at different points in time, we consider that the best curve to forecast may change across time. To do so for each scenario we calculate the infection rate increase as define by equation 1 where t+1 denotes the next day. For every scenario we calculate R during the first 31 days and from this we calculate the cumulative productory of R, this stores the percentage of the infected increased based on the number of infected already know. The resulting value does not take into account deaths or recovery, to take these into account the number of active people with the disease is given by formula 2 In which PL represents the projection of the number of beds in need, DI the number of days to occur hospitalization(10 days), CC the number of confirmed cases of COVID-19, TI the hospitalization rate of people affected with COVID-19 and TO the rate of death (3.2% estimate from World Health Organization (WHO) [4] , [5] , [10] . Since we are interested in analyzing how long would it take for the hospitals to be overcrowded we consider the number of people in hospitals, who need beds, is 5% [6] of the active infected as denoted in 2. So we can see for each state if or when the demand of hospital beds exceeds the supply. Observing the equation that predicts the increase in the number of people infected by COVID-19 needing hospitalization for each Brazilian state, the number of available hospital beds was considered. Therefore, it is possible to identify when each state may be overcrowded in its health facilities, without the ability to treat newly infected people, as well as not having minimal conditions to treat other types of pathologies. In this context, using the best-case scenario as a premise, we assume that if a patient diagnosed with COVID-19 needs hospitalization, all hospital beds will be available for use. Thus, for each Brazilian state, Intensive Care Unit (ICU) beds were considered in this study, as well as general hospital beds, consisting of beds for clinical and surgical hospitalization, with only pediatric and neonatal beds being removed. It is worth mentioning that, given the high demand for care and the number of beds available, it is assumed that all beds are considered for hospitalization and that they offer initial support for treatment in the event of a full ICU bed. Assuming April 22, 2020 as the milestone of the adoption of the possible scenarios mentioned, we created the projection of the need beds for hospital for the next 30 days. Figures 1,2,3,4 ,5 present the states with the most alarming scenarios, in which analyzing the upper part of the confidence interval, even if all the beds available today could accommodate those infected by COVID-19, the health system would collapse in all quarantine adoptions in this time window. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 1, 2020. . https://doi.org/10.1101/2020.04.26.20080143 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 1, 2020. . By estimating the need for beds by Brazilian states, even adopting the best possible scenario with the population hospitalized for COVID-19 being discharged in a period of 10 days, in addition to the premise that all hospital beds (except for neonatal and pediatric) would be willing to receive the infected population, 5 states would collapse in the model adopted by South Korea, 16 Brazilian states would collapse on Germany models and all Brazilian states would collapse their health system based on the Italian model, in a 30 day time window. Therefore, it is recommended the adoption of public policies implemented by South Korea, in addition to increasing the number of beds with public policies with field hospitals (emergency beds for COVID-19). It is noteworthy that in this scenario the Brazilian health system would make available for COVID-19 patients all available hospital beds, which would result in the said secondary deaths to COVID-19, people who need a bed to be hospitalized due to other diseases (heart attack, car accident, stroke, etc.), but because of the overcrowd of the public health system would return home, resulting in an end that could have been contoured if hospital treatment had been given to the pacient. Corona updates for internationals in germany Coronavirus cases have dropped sharply in south korea. what's the secret to its success? covid-19) mortality rate Coronavirus disease (covid-2019) situation reports Covid-19 dashboard by the center for systems science and engineering (csse) at johns hopkins university (jhu) Italy announces quarantine affecting quarter of population Who director-general's opening remarks at the media briefing on covid-19 -3 Two first coronavirus cases confirmed in italy: prime minister Covid-19 Coronavirus -brazil Covid-19: Epidemiology, evolution, and cross-disciplinary perspectives