key: cord-0869902-nqauuzw0 authors: Castro, Marcia C; Resende de Carvalho, Lucas; Chin, Taylor; Kahn, Rebecca; Franca, Giovanny V. A.; Macario, Eduardo Marques; de Oliveira, Wanderson Kleber title: Demand for hospitalization services for COVID-19 patients in Brazil date: 2020-04-01 journal: nan DOI: 10.1101/2020.03.30.20047662 sha: 954576043076819041bde838775dc7929be81f5b doc_id: 869902 cord_uid: nqauuzw0 COVID-19 is now a pandemic and many of the affected countries face severe shortages of hospital resources. In Brazil, the first case was reported on February 26. As the number of cases grows in the country, there is a concern that the health system may become overwhelmed, resulting in shortages of hospital beds, intensive care unit beds, and mechanical ventilators. The timing of shortage is likely to vary geographically depending on the observed onset and pace of transmission observed, on the availability of resources, and on the actions implemented. Here we consider the daily number of cases reported in municipalities in Brazil to simulate twelve alternative scenarios of the likely timing of shortage, based on parameters consistently reported for China and Italy, on rates of hospital occupancy for other health conditions observed in Brazil in 2019, and on assumptions of allocation of patients in public and private facilities. Results show that hospital services could start to experience shortages of hospital beds, ICU beds, and ventilators in early April, the most critical situation observed for ICU beds. Increasing the allocation of beds for COVID-19 (in lieu of other conditions) or temporarily placing all resources under the administration of the state delays the anticipated start of shortages by a week. This suggests that solutions adopted by the Brazilian government must aim at expanding the available capacity (e.g., makeshift hospitals), and not simply prioritizing the allocation of available resources to COVID-19. show that hospital services could start to experience shortages of hospital beds, ICU beds, and 27 ventilators in early April, the most critical situation observed for ICU beds. Increasing the 28 allocation of beds for COVID-19 (in lieu of other conditions) or temporarily placing all resources 29 under the administration of the state delays the anticipated start of shortages by a week. This 30 suggests that solutions adopted by the Brazilian government must aim at expanding the available 31 capacity (e.g., makeshift hospitals), and not simply prioritizing the allocation of available 32 We conducted forward simulations of the demand for hospital beds, ICU beds, and mechanical 77 ventilators by health macro-region in Brazil. In the initial phase of a new infectious disease 78 outbreak, the number of cases grows exponentially. 12,13 Using the number of cases reported to 79 All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03.30.20047662 doi: medRxiv preprint date and estimates of doubling times from the outbreak in China, 14-16 we estimated the number of 80 future cases each day. As the outbreak spreads, the exponential model becomes less realistic, 13 so 81 we stopped the simulations once 10% of the population was estimated to be infected. 82 We used the age structure of each macro-region combined with age-specific attack and severity 83 rates drawn from the literature. We used attack rates by age from the outbreak in China, 17 84 considering a range of ±3%, and considered that 86% of all infections were undocumented. 18 85 Demand for hospitalization was calculated using the severity of cases by age from China. 17 Time 86 from illness to hospitalization was 3-7 days, 19,20 and length of hospitalization ranged from 7-15 87 days; 21 for ICU hospitalizations these parameters were 8-15, 19,20 and 7-15, 21,22 respectively. We 88 considered that 5% of the cases needed ICU admission 21,23 and half of those in ICU needed 89 mechanical ventilation 20,21 for an average of 5 days. 24 We also simulated an alternative scenario 90 based on Italy and assumed that 12% of the cases needed ICU admission. 4,6 These parameters 91 were combined with assumptions of hospital occupancy and type of service to produce twelve 92 scenarios, summarized in Table 1 100 Figure 1 shows the fraction of public hospital beds and ICU beds in each health micro-region, as 101 well as the percentage of the population in those units that rely solely on the SUS. Inequities in 102 supply and demand exist, and mirror regional inequalities commonly found in social and health 103 indicators. 9 104 All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03.30.20047662 doi: medRxiv preprint As of March 27, 3.417 cases were confirmed in Brazil, 70% concentrated in nine cities: São 105 Paulo, Rio de Janeiro, Fortaleza, Brasília, Porto Alegre, Salvador, Belo Horizonte, Curitiba, and 106 Manaus. The growth pattern of the epidemic is not the same across the country. Therefore, we 107 chose to simulate scenarios for the nine health macro-regions associated with those cities; 108 combined these macro-regions represent 79·6% of reported cases. In the case of São Paulo and 109 Brasília, the macro-region includes only the municipality itself. hypothesis of access to public resources, scenarios 2, 5, 8, and 11 assumed that 80% of the 123 demand would come from individuals that solely rely on SUS. Since the offer of SUS services 124 presents regional inequities (Figure 1) , the timing of shortage depends both on the share public x 125 private in each area. 126 127 All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which was not peer-reviewed) is the . São Paulo, where the first case was reported, the first 100 cases were recorded in 19 days. 137 All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Brazil would start to face shortages of hospital beds, ICU beds, and ventilators, with ICU beds 147 being the most immediate problem. The timing of shortages across the country depends on the 148 onset and the intensity of transmission. Also, the population that relies solely on SUS may bear 149 the largest burden, further exacerbating the existing inequalities, which calls for a reflection 150 around equity and ethics in service allocation. Avoiding this scenario is the paramount task of the 151 It is unreasonable to expect that all hospital resources could be entirely dedicated to Although elective procedures can be postponed, other health emergencies compete for resources. 154 author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which was not peer-reviewed) is the Second, there is a short window of opportunity to prepare. The response must be immediate, and 184 it will demand a concerted effort from society. Repurposing large spaces (e.g., arenas, 185 convention centers) to build makeshift hospitals for additional beds is critical. Calling on the 186 industry to produce the necessary equipment (e.g., ventilators, masks, gloves, protective gown, 187 etc) and to provide it to hospitals at minimum or no cost would also contribute to mitigating the 188 stress on the system, and to safeguarding the working conditions of health care professionals. 189 Mobilizing community leaders, artists, athletes, and local role models would help to convey a 190 unique message to the population, and to gather support from the wealthy that could provide 191 resources to expand the production of medical equipment. Brazil is, in theory, uniquely equipped to respond to the COVID-19 epidemic. It has a free and 202 universal health system, 7 it has one of largest community-based primary care delivery programs 203 that serves 74·8% of the population, 28 it can learn from the mistakes and success that other 204 countries hit by COVID-19 have made, and it has a history of responding to health threats by 205 implementing governmental action and by generating high-quality scientific evidence, such as 206 was done when Zika virus hit the country. 29 Yet, the current moment is unique. It requires a 207 unified message from the country's leadership at various levels: federal, state, and municipal. It 208 requires the industry to work in solidarity to produce needed inputs without aiming profit but the 209 collective wellbeing. It requires the population to realize the importance and the urgency to 210 comply. We hope our results will help to move forward this agenda. 211 212 All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03.30.20047662 doi: medRxiv preprint World Health Organization. Coronavirus disease (COVID-19) outbreak situation Fair Allocation of Scarce Medical Resources in 223 the Time of Covid-19 The Toughest Triage -Allocating Ventilators in a 225 Critical Care Utilization for the COVID-19 Outbreak 227 Early Experience and Forecast During an Emergency Response American Hospital Capacity And Projected Need for 229 COVID-19 Patient Care COVID-19 and Italy: what next? Brazil's unified health system: the first 30 234 years and prospects for the future The Brazilian health system: 236 history, advances, and challenges Education and Inequality in Brazil Paths of Inequality in Brazil: A Half-Century of Changes Avaliação dos parâmetros de oferta mínimos 241 Planejamento e Políticas Públicas -PPP/Ipea Idade e Sexo 2000-2015 -Brasil. Ministério da Saúde/SVS/CGIAE How generation intervals shape the relationship between growth 247 rates and reproductive numbers Infectious Diseases of Humans: Dynamics and Control Nowcasting and forecasting the potential domestic and 251 international spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study The effect of travel restrictions on the spread of the 254 2019 novel coronavirus (COVID-19) outbreak Estimating clinical severity of COVID-19 from the 256 transmission dynamics in Wuhan, China Epidemiology and Transmission of COVID-19 in Shenzhen 258 Analysis of 391 cases and 1,286 of their close contacts Substantial undocumented infection facilitates the rapid 261 dissemination of novel coronavirus (SARS-CoV2) Clinical features of patients infected with 2019 novel 263 coronavirus in Wuhan Clinical Characteristics of 138 Hospitalized Patients With 265 Novel Coronavirus-Infected Pneumonia in Wuhan, China Clinical Characteristics of Coronavirus Disease Clinical course and risk factors for mortality of adult 269 inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. The Lancet 2020 A model simulation study on effects of intervention measures in Wuhan 271 COVID-19 epidemic Retrospective Analysis of Clinical Features in 101 Death 273 Cases with Malaria 275 morbidity and mortality in Ebola-affected countries caused by decreased health-care capacity, 276 and the potential effect of mitigation strategies: a modelling analysis Spanish government puts private healthcare firms at the orders of 279 the regions. El Pais Coronavirus cases have dropped sharply in South Korea Brazil's Family Health Strategy -Delivering Community-Based 284 Primary Care in a Universal Health System All rights reserved. No reuse allowed without permission author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which was not peer-reviewed) is the MCC conceived the original, was responsible for data analysis, data interpretation, data 214 visualization, and wrote the manuscript. LRC was responsible for data curation, data analysis, 215 programming, and contributed to writing. RK and TC were responsible for programming and 216 contributed to writing. GVAF, EMM, and WKO were responsible for data curation and 217 interpretation. All authors approved the final version of the manuscript. 218 219