Acessibilidade / Reportar erro

Estimating Brazilian states’ demands for intensive care unit and clinical hospital beds during the COVID-19 pandemic: development of a predictive model

ABSTRACT

BACKGROUND:

The fragility of healthcare systems worldwide had not been exposed by any pandemic until now. The lack of integrated methods for bed capacity planning compromises the effectiveness of public and private hospitals’ services.

OBJECTIVES:

To estimate the impact of the COVID-19 pandemic on the provision of intensive care unit and clinical beds for Brazilian states, using an integrated model.

DESIGN AND SETTING:

Experimental study applying healthcare informatics to data on COVID-19 cases from the official electronic platform of the Brazilian Ministry of Health.

METHODS:

A predictive model based on the historical records of Brazilian states was developed to estimate the need for hospital beds during the COVID-19 pandemic.

RESULTS:

The proposed model projected in advance that there was a lack of 22,771 hospital beds for Brazilian states, of which 38.95% were ICU beds, and 61.05% were clinical beds.

CONCLUSIONS:

The proposed approach provides valuable information to help hospital managers anticipate actions for improving healthcare system capacity.

KEY WORDS (MeSH terms):
Public health administration; Bed occupancy; Coronavirus infections; Pandemics; Hospital bed capacity

AUTHORS’ KEY WORDS:
Public health planning; Population dynamics; COVID-19 pandemic; COVID-19 virus disease; Compartmental model

INTRODUCTION

The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) disease11. Huang C, Wang Y, Li X, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet. 2020;395(10223):497-506. PMID: 31986264; https://doi.org/10.1016/S0140-6736(20)30183-5.
https://doi.org/10.1016/S0140-6736(20)30...
33. World Health Organization. Coronavirus disease (COVID-2019) situation reports. 2020. Available from: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports. Accessed in 2020 (Sep 11).
https://www.who.int/emergencies/diseases...
was first reported in Wuhan, China, in December 2019. Since then, political leaders and healthcare managers have endeavored to estimate the demand for intensive care unit (ICU) and clinical beds to guard against the possibility of collapse of the healthcare system. Moreover, political leaders have adopted proactive non-pharmaceutical interventions (NPIs) to reduce fatalities arising from bed shortages. Although the quarantine and travel restrictions in Wuhan delayed progression of the epidemic on an international scale, thereby reducing case importations by nearly 80% until mid-February,44. Du Z, Wang L, Cauchemez S, et al. Risk for transportation of coronavirus disease from Wuhan to other cities in China. Emerg Infect Dis. 2020;26(5):1049-52. PMID: 32053479; https://doi.org/10.3201/eid2605.200146.
https://doi.org/10.3201/eid2605.200146...
66. Chinazzi M, Davis JT, Ajelli M, et al. The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak. Science. 2020;368(6489):395-400. PMID: 32144116; https://doi.org/10.1126/science.aba9757.
https://doi.org/10.1126/science.aba9757...
the virus has since then spread rapidly both domestically and internationally.55. Zhao S, Lin Q, Ran J, et al. Preliminary estimation of the basic reproduction number of novel coronavirus (2019-ncov) in China, from 2019 to 2020: A data-driven analysis in the early phase of the outbreak. Int J Infect Dis. 2020;92:214-7. PMID: 32007643; https://doi.org/10.1016/j.ijid.2020.01.050.
https://doi.org/10.1016/j.ijid.2020.01.0...
,77. Bogoch II, Watts A, Thomas-Bachli A, et al. Pneumonia of unknown etiology in Wuhan, China: Potential for international spread via commercial air travel. J Travel Med. 2020;27(2):taaa008. PMID: 31943059; doi: 10.1093/jtm/taaa008.
https://doi.org/10.1093/jtm/taaa008...
,88. Wu JT, Leung K, Leung GM. Nowcasting and forecasting the potential domestic and international spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study. Lancet. 2020;395(10225):689-97. PMID: 32014114; https://doi.org/10.1016/S0140-6736(20)30260-9.
https://doi.org/10.1016/S0140-6736(20)30...
In Brazil, confirmed cases started to be seen in state capital municipalities and then moved towards non-metropolitan cities, and the numbers of cases continue to grow in every state.99. Candido DS, Claro IM, de Jesus JG, et al. Evolution and epidemic spread of SARS-CoV-2 in Brazil. Science. 2020;369(6508):1255-60. PMID: 32703910; https://doi.org/10.1126/science.abd2161.
https://doi.org/10.1126/science.abd2161...
,1010. Cota W. Número de casos confirmados de COVID-19 no Brasil. Available from: https://covid19br.wcota.me/. Accessed in 2020 (Nov 25).
https://covid19br.wcota.me/...

Brazil identified its first COVID-19 case on February 25, 2020, and its first death on March 17, 2020. A lack of consensus on public healthcare policy1111. Shalders A. Mandetta é demitido do Ministério da Saúde após um mês de conflito com Bolsonaro: relembre os principais choques. Available from: https://www.bbc.com/portuguese/internacional-52316728. Accessed in 2020 (Sep 8).
https://www.bbc.com/portuguese/internaci...
1212. BBC News Brasil. Ministro da saúde Nelson Teich pede demissão menos de um mês depois de assumir. Available from: https://www.bbc.com/portuguese/brasil-52683285. Accessed in 2020 (Sep 8).
https://www.bbc.com/portuguese/brasil-52...
compromised use of integrated and coordinated NPIs to reduce COVID-19 mortality and healthcare demand. Moreover, the lack of integrated methods for bed capacity planning compromised the effectiveness of public and private hospitals’ services.

Mathematical models are used to evaluate the efficacy of specific interventions that were implemented in the past, in order to identify future strategies and effectively inform public health policy.1313. Metcalf CJE, Morris DH, Park SW. Mathematical models to guide pandemic response. Science. 2020;369(6502):368-9. PMID: 32703861; https://doi.org/10.1126/science.abd1668.
https://doi.org/10.1126/science.abd1668...
,1414. IHME COVID-19 Forecasting Team. Modeling COVID-19 scenarios for the United States. Nat Med. 2020. PMID: 33097835; https://doi.org/10.1038/s41591-020-1132-9.
https://doi.org/10.1038/s41591-020-1132-...
The challenge in this work consists of forecasting the numbers of individuals requiring hospitalization and predicting the availability of hospital beds for such patients, taking into account the technical limitations of the data,1515. Souza WM, Buss LF, da Silva Candido D, et al. Epidemiological and clinical characteristics of the early phase of the COVID-19 epidemic in Brazil. medRxiv. 2020. https://doi.org/10.1101/2020.04.25.20077396.
https://doi.org/10.1101/2020.04.25.20077...
1717. Joe H, Esteban OO, Roser M. Coronavirus (COVID-19) testing. Our World in Data. Available from: https://ourworldindata.org/grapher/full-list-cumulative-total-tests-per-thousand. Accessed in 2020 (Sep 8).
https://ourworldindata.org/grapher/full-...
infection rates and under-reporting of cases1818. de Oliveira ACS, Morita LHM, da Silva EB, et al. Bayesian modeling of COVID-19 cases with a correction to account for under-reported cases. Infect Dis Model. 2020;5:699-713. PMID: 32995681; https://doi.org/10.1016/j.idm.2020.09.005.
https://doi.org/10.1016/j.idm.2020.09.00...
because of low numbers of tests. Therefore, we limited our analysis to the official data.1919. Ministério da Saúde do Brasil. Coronavírus Brasil. Available from: https://covid.saude.gov.br/. Accessed in 2020 (Nov 25).
https://covid.saude.gov.br/...
,2020. OpenDataSUS. Available from: https://opendatasus.saude.gov.br/. Accessed in 2020 (Sep 8).
https://opendatasus.saude.gov.br/...

Tertiary-care hospital beds are not equally available for all citizens. In Brazil, there are 2.4 beds/1,000 inhabitants and 34,318 ICU beds. 54% of the ICU beds are public, assigned to 78% of the population, while the remaining 22% of the population that has access to private care receive this medical care on 46% of the ICU beds.

In the following sections, we briefly describe the model and the estimates for the required numbers of ICU and clinical beds for each Brazilian state that would be needed to avoid healthcare system collapse.

OBJECTIVES

The purpose of this paper was to present a combined approach to estimation of Brazilian states’ demands for ICU and clinical beds during a pandemic.

METHOD

ICU and clinical bed dynamics

Consider a compartmental model2121. McGee R. SEIRS Plus: SEIRS model and extended model framework. Available from: https://github.com/ryansmcgee/seirsplus. Accessed in 2020 (Sep 8).
https://github.com/ryansmcgee/seirsplus...
,2222. Dottori M, Fabricius G. Sir model on a dynamical network and the endemic state of an infectious disease. Physica A: Statistical Mechanics and its Applications. 2015;434:25-35. https://doi.org/10.1016/j.physa.2015.04.007.
https://doi.org/10.1016/j.physa.2015.04....
in which the population is divided into susceptible (S), exposed (E), infectious (I) and recovered (R) individuals (SEIR model). A rate of β of S individuals in contact with I (S-I contact) becomes E and progresses over the course of the incubation period at a rate σ, to state I. While a rate γ of I recovers from the disease, a rate μI of I evolves to death. A fraction ξ of R (recovered individuals) may become (or not, if ξ = 0) re-susceptible (S) and, therefore, a SEIRS model.

The effect of testing a segment of the population is modeled by introducing the rate of transmission for individuals with detected infection (βD), detected exposed state (DE) and detected infectious state (DI). Let ψE and ψI be the probabilities of positive tests for exposed and infected individuals, and Q, the rate of individuals with detected infection interacting with the population. Then DE and DI result from the rates θEψE and θIψI of testing exposed E and infected I individuals, respectively. The model describes the full spectrum of disease. Let N be the estimate of an affected population. Thus, N = S + E + I + DE + DI + R.

To represent the ICU and clinical bed dynamics, consider that a fraction of infected individuals is asymptomatic (IA). The rate α for symptomatic cases (IS = I – IA) requiring hospitalization (H) is αIS. Let T be the number of planning days during the pandemic with t corresponding to each admission day at the hospital. For clinical bed dynamics, consider LS and Lf as the average lengths of stay (LoS) of surviving patients and patients who died, respectively. For surviving patients in ICU beds, the average length of stay is Lb + Ld + La, where Lb represents the surviving patients in clinical beds who are re-directed to ICU beds, Ld is the average LoS of surviving patients in ICU beds, and La, the average LoS in clinical beds among surviving patients after being re-directed from ICU beds. For deceased victims of ICU bed dynamics, the average LoS is Lc + Li, representing the average periods for which a deceased patient will stay in clinical and ICU beds, respectively.

Given the average length of stay metrics for ICU and clinical dynamics, we calculate admission and leave days for each patient profile. For a surviving clinical patient, whose admission day is t, the expected day on which this patient leaves the clinical bed is To = t + LS – 1. A clinical patient who progressed to death is expected to be removed from the hospital in Tf = t + Lf – 1. A surviving ICU patient is admitted in Ti = t + Lb. In Tr = Ti + Ld, the patient returns to the clinical bed, and in Tc = Tr + La – 1, the patient leaves the clinical bed. For a deceased ICU patient, the admission day is Td = t + Lc, and Tu = Td + Li – 1 is the expected day on which the patient is removed from the hospital.

Let Ht be the daily admission of patients to hospitals, and τ the fraction of hospitalized cases that require critical care in the ICU. Also, consider ζ and η to be the rates of critical patients who evolve to death in ICU beds and clinical beds, respectively.

The integrated model is presented in Figure 1. The model equations and data are available in a Data Repository that is available from https://github.com/joaoflavioufmg/webcovid19.

Figure 1
Model for susceptible (S), exposed (E), infectious (I), recovered (R) and re-susceptible (S) individuals (SEIRS) in Brazilian states: Pandemic model integrated with dynamic intensive care unit and clinical hospital beds.

Data sources and measurements

The model data included the social distancing index, which ranges from 30%, as observed at earlier times in the COVID-19 pandemic, to 100% in a possible lockdown situation.2323. Inloco. Índice de isolamento social. Available from: https://mapabrasileirodacovid.inloco.com.br/pt/. Accessed in 2020 (Sep 8).
https://mapabrasileirodacovid.inloco.com...
The latter would be expected to reduce the transmission rate of the model by 74%.2424. Jarvis CI, Van Zandvoort K, Gimma A, et al. Quantifying the impact of physical distance measures on the transmission of COVID-19 in the UK. BMC Med. 2020;18(1):124. PMID: 32375776; https://doi.org/10.1186/s12916-020-01597-8.
https://doi.org/10.1186/s12916-020-01597...
Furthermore, the basic reproduction number, which is the average number of secondary cases generated per case, was set to 2.9 in the case of the Brazilian outbreak.1515. Souza WM, Buss LF, da Silva Candido D, et al. Epidemiological and clinical characteristics of the early phase of the COVID-19 epidemic in Brazil. medRxiv. 2020. https://doi.org/10.1101/2020.04.25.20077396.
https://doi.org/10.1101/2020.04.25.20077...
After the first month, we set customized transmission rate values for each Brazilian state and also took the time-varying reproduction number into consideration.1616. Mellan TA, Hoeltgebaum HH, Mishra S, et al. Report 21: Estimating COVID-19 cases and reproduction number in Brazil. Imperial College of London. 2020. https://doi.org/10.25561/78872.
https://doi.org/10.25561/78872...
We used data from official reports on symptomatic infections.1919. Ministério da Saúde do Brasil. Coronavírus Brasil. Available from: https://covid.saude.gov.br/. Accessed in 2020 (Nov 25).
https://covid.saude.gov.br/...
,2020. OpenDataSUS. Available from: https://opendatasus.saude.gov.br/. Accessed in 2020 (Sep 8).
https://opendatasus.saude.gov.br/...
The asymptomatic fraction has been estimated variously in different reports,2525. Ferguson N, Laydon D, Nedjati Gilani G, et al. Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. medRxiv Imperial College of London. 2020. https://doi.org/10.25561/77482.
https://doi.org/10.25561/77482...
as follows: 18% on the Diamond Princess ship;2626. Mizumoto K, Kagaya K, Zarebski A, Chowell G. Estimating the asymptomatic proportion of coronavirus disease 2019 (COVID-19) cases on board the Diamond Princess cruise ship, Yokohama, Japan, 2020. Euro Surveill. 2020;25(10):2000180. PMID: 32183930; https://doi.org/10.2807/1560-7917.ES.2020.25.10.2000180.
https://doi.org/10.2807/1560-7917.ES.202...
31% in repatriation flight screening;2727. Nishiura H, Linton NM, Akhmetzhanov AR. Serial interval of novel coronavirus (COVID-19) infections. Int J Infect Dis. 2020;93:284-6. PMID: 32145466; https://doi.org/10.1016/j.ijid.2020.02.060.
https://doi.org/10.1016/j.ijid.2020.02.0...
and 50%-75% in the Italian village of Vo’Euganeo.2828. Qiu H, Wu J, Hong L, et al. Clinical and epidemiological features of 36 children with coronavirus disease 2019 (COVID-19) in Zhejiang, China: an observational cohort study. Lancet Infect Dis. 2020;20(6):689-696. PMID: 32220650; https://doi.org/10.1016/S1473-3099(20)30198-5.
https://doi.org/10.1016/S1473-3099(20)30...
The average infection lethality ratio and the percentage of symptomatic cases requiring hospitalization were obtained from recent studies2525. Ferguson N, Laydon D, Nedjati Gilani G, et al. Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. medRxiv Imperial College of London. 2020. https://doi.org/10.25561/77482.
https://doi.org/10.25561/77482...
,2929. Nascimento IJBD, Pinto LR, Fernandes VA, et al. Clinical characteristics and outcomes among Brazilian patients with severe acute respiratory syndrome coronavirus 2 infection: an observational retrospective study. Sao Paulo Med J. 2020:S1516-31802020005028202. PMID: 33263706; https://doi.org/10.1590/1516-3180.2020.00365.R1.08092020.
https://doi.org/10.1590/1516-3180.2020.0...
and were adjusted for each Brazilian state according to its demographic pyramid.3030. Instituto Brasileiro de Geografia e Estatística. Projeção da população do Brasil e das unidades da federação. Available from: https://www.ibge.gov.br/apps/populacao/projecao/. Accessed in 2020 (Sep 8).
https://www.ibge.gov.br/apps/populacao/p...

The percentage of symptomatic cases requiring hospitalization was estimated from the confirmed cases of SARS-CoV-2 infection in each state,3131. Portal da Transparência do Estado do Acre. Pacto Acre sem Covid. Portal de Informações sobre o combate à Covid-19. Available from: http://covid19.ac.gov.br/monitoramento/notificacoes/boletim. Accessed in 2020 (Dec 15).
http://covid19.ac.gov.br/monitoramento/n...
5656. Governo do Estado do Tocantins. Secretaria da Saúde. Covid-19. Casos confirmados. Available from: http://integra.saude.to.gov.br/covid19. Accessed in 2020 (Dec 15).
http://integra.saude.to.gov.br/covid19...
taking its demographic pyramid and the estimated incidence for each age group into account.2525. Ferguson N, Laydon D, Nedjati Gilani G, et al. Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. medRxiv Imperial College of London. 2020. https://doi.org/10.25561/77482.
https://doi.org/10.25561/77482...
Thus, we estimated that 6.6% to 7.7% (95% confidence interval, CI) of the cases of symptomatic infection in Brazil would require hospitalization.

The average proportion of the patients requiring critical care in an ICU was estimated to be 26.11% of the hospitalized cases, considering recent experience.2525. Ferguson N, Laydon D, Nedjati Gilani G, et al. Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. medRxiv Imperial College of London. 2020. https://doi.org/10.25561/77482.
https://doi.org/10.25561/77482...
We used the numbers of ICU and clinical beds for the month of April 2020 obtained from official Brazilian data sources.5757. Ministério da Saúde. Secretaria de Atenção à Saúde. CNESnet. Consulta leitos. Available from: http://cnes2.datasus.gov.br/Mod_Ind_Tipo_Leito.asp. Accessed in 2020 (Dec 15).
http://cnes2.datasus.gov.br/Mod_Ind_Tipo...
According to general bed utilization reports from before the pandemic,5858. Halpern NA, Pastores SM. Critical care medicine beds, use, occupancy and costs in the United States: a methodological review. Crit Care Med. 2015;43(11):2452-9. PMID: 26308432; https://doi.org/10.1097/CCM.0000000000001227.
https://doi.org/10.1097/CCM.000000000000...
34% of clinical beds and 21% of ICU beds were available for patients infected with COVID-19. The input data is presented in Table 1.

Table 1
Model parameters for each Brazilian state

RESULTS

The model considers two different periods of the pandemic: May and August 2020. It uses the historical records of COVID-19 cases5959. Labdec EMU. Laboratório de Apoio à Decisão, NESCON (Medicina-UFMG) e DEP (Engenharia de Produção-UFMG). Combate ao Coronavírus no Brasil. Available from: https://labdec.nescon.medicina.ufmg.br/webcovid19/. Accessed in 2020 (Dec 15).
https://labdec.nescon.medicina.ufmg.br/w...
to estimate future infections for each state and project the ICU and clinical bed use for a period of 365 days. Figure 2 shows an illustrative example of a forecast for August 2020, for the state of Minas Gerais. The model periodically adjusts to the real number of cases, thus providing a fair estimate of future cases of infection. Historical records of infection and future estimates feed the dynamic bed model, which forecasts a reduction in ICU and clinical bed capacity of the state healthcare system. In May 2020, the ICU bed capacity for Minas Gerais was different from the capacity in August 2020. Thus, we adopted a simplified assumption considering a single capacity increase in July 2020. The monitoring of bed procurement did not form part of the scope of the present study.

Figure 2
Intensive care unit and clinical bed model for the state of Minas Gerais (MG).

Accordingly, the model forecasts the possibility of healthcare collapse, i.e. 100% utilization of ICU and clinical beds. Thus, the demand for ICU and clinical beds is established at the peak of infections, since the proportion of symptomatic cases requiring hospitalization produces the peak of capacity utilization of ICU and clinical beds. Although the maximum utilization of bed capacity occurred in May, June and July 2020, the estimate of SARS-CoV-2 infection levels continues to grow in the projection, which produces shortages of ICU and clinical beds for the states. Overall, the results suggest that 0.81% of the population had become infected by May 2020 and that this percentage was 2.87% by August 2020, excluding deaths and individuals who had recovered. Therefore, the majority of the population remains potentially vulnerable. Table 2 presents the results from the model projection of peak bed utilization and the demand for new hospital beds.

Table 2
Model projection of peak bed utilization and demand for new hospital beds

The outcomes show that the Brazilian states require 22,771 additional beds, of which 8,869 (38.95%) are ICU beds, and 13,902 (61.05%) are clinical beds. Populous states, like São Paulo, Rio de Janeiro and Minas Gerais account for nearly 40% of the projected demand.

DISCUSSION

Our method projected in advance the lack of ICU and clinical beds from May to July 2020. From March to April 2020, technical notes and print and digital media6060. Canzian F. Leitos de UTI do SUS devem acabar em maio na maioria dos estados. Folha de São Paulo. 03 mai 2020. Available from: https://www1.folha.uol.com.br/equilibrioesaude/2020/05/leitos-de-uti-do-sus-devem-acabar-em-maio-na-maioria-dos-estados.shtml. Accessed in 2020 (Dec 15).
https://www1.folha.uol.com.br/equilibrio...
warned about the estimated maximum utilization rate of public and private hospitals’ bed capacity. These warnings were borne out by reality in most states, as also shown through the model's projections.

These findings provide evidence that SARS-CoV-2 transmission in Brazil is not under control, and the number of active cases remains stable or is even growing in some states, despite the physical distancing policies adopted so far. This suggests that further action is required to prevent higher rates of mortality. Overall, these findings have filled an important gap in estimating the deficit of ICU and clinical beds within the Brazilian healthcare system, across the country. Furthermore, our study also provides a flexible tool that allows healthcare decision-makers to forecast the impact of the pandemic impact and to implement policies for reducing COVID-19 mortality.

The model was implemented in the Python software (Python Language Reference, version 3.8.2, available from http://www.python.org; Python Software Foundation, Amsterdam, 1995). The programming routine automatically captures historical and updated data on COVID-19 cases from a web-based repository that aggregates official data from all states and municipalities that present confirmed cases. The initial model6161. O’Brien J. See when states will face hospital bed capacity shortages during COVID-19 outbreak. HealthLeaders. Mar 16 2020. Available from: https://www.healthleadersmedia.com/clinical-care/see-when-states-will-face-hospital-bed-capacity-shortages-during-covid-19-outbreak. Accessed in 2020 (Dec 15).
https://www.healthleadersmedia.com/clini...
evolved from a simplified estimate on a spreadsheet to a sophisticated approach embedded in a web-based system that captures data, runs models and displays the results.5959. Labdec EMU. Laboratório de Apoio à Decisão, NESCON (Medicina-UFMG) e DEP (Engenharia de Produção-UFMG). Combate ao Coronavírus no Brasil. Available from: https://labdec.nescon.medicina.ufmg.br/webcovid19/. Accessed in 2020 (Dec 15).
https://labdec.nescon.medicina.ufmg.br/w...

The web-based system can be accessed online at https://labdec.nescon.medicina.ufmg.br/webcovid19/. It provides valuable up-to-date information for healthcare managers in advance, which is very useful because acquiring a large number of ICU and clinical beds in a short period is difficult.

The government has implemented a hospital bed census and has obtained additional ICU and clinical beds. Up to November 2020, the Brazilian Ministry of Health had ordered 16,582 beds from suppliers. The cost of these additional beds will be US$ 437 million or R$ 2.34 billion.6262. Ministério da Saúde. COVID-19: Leitos. Available from: https://viz.saude.gov.br/extensions/DEMAS_C19Insumos_LEITOS/DEMAS_C19Insumos_LEITOS.html. Accessed in 2020 (Dec 15).
https://viz.saude.gov.br/extensions/DEMA...
Although the Ministry of Health has provided 73% of the overall projected demand, this provision of 16,582 hospital beds differs from (and is lower than) the projected demand for beds at the state level, as presented in Table 2. Meanwhile, during pandemics, we suggest that temporary hospitals should be provided, with the supplementary numbers of ICU and clinical beds for each state. This should incorporate use of NPIs, including physical distancing and personal protection policies6363. Andersen M. Early evidence on social distancing in response to COVID-19 in the United States. SSRN. 2020. https://doi.org/10.2139/ssrn.3569368.
https://doi.org/10.2139/ssrn.3569368...
such as use of masks, in order to avoid overloading the healthcare system.

In Brazil, physical distancing policies have reduced the intensity of transmission, thereby contributing towards saving many lives. However, in many states, the number of COVID-19 cases remains stable or is even rising, which indicates that distancing control policies should be intensified rather than loosened. Experiences have provided evidence that physical distancing policies have reduced SARS-CoV-2 infection rates,2424. Jarvis CI, Van Zandvoort K, Gimma A, et al. Quantifying the impact of physical distance measures on the transmission of COVID-19 in the UK. BMC Med. 2020;18(1):124. PMID: 32375776; https://doi.org/10.1186/s12916-020-01597-8.
https://doi.org/10.1186/s12916-020-01597...
and have also provided evidence that person-to-person activities have increased the SARS-CoV-2 infection rate.6464. Alfredo M. Casos na EPCAR: Barbacena passa a ter maior taxa de infectados de Minas por COVID-19. Available from: https://www.em.com.br/app/noticia/gerais/2020/05/25/interna_gerais,1150623/epcar-barbacena-passa-a-ter-maior-taxa-de-infectados-de-minas.shtml. Accessed in 2020 (Dec 15).
https://www.em.com.br/app/noticia/gerais...
These findings have demonstrated that economic, cultural, educational or social events involving proximity are dangerous during the COVID-19 pandemic. We identified certain limitations in our study that provide opportunities for future research. Our model is deterministic and assumes discrete values and a uniformly spread population; therefore, a geographical analysis on the spread of the virus is urgently needed in order to determine location-based transmission rates. Additionally, we used an average physical distancing index because physical distancing policies in several states were activated and deactivated at various times. One promising approach for future investigations may comprise formulation of a stochastic model with a multiperiod physical distancing index.

CONCLUSION

We proposed an integrated approach towards evaluation of the impact of the COVID-19 pandemic on the healthcare system capacity of Brazilian states, through estimating the demand for ICU and clinical beds for each state. The integrated model was applied to two periods of the pandemic for each state and it showed that healthcare would collapse at different times if bed demand, estimated as 22,771 beds, were not satisfied. The government has provided 16,582 hospital beds for the Brazilian states; however, the numbers of beds diverges from the estimated demand projections at the state level.

  • Universidade Federal de Minas Gerais (UFMG), Belo Horizonte (MG), Brazil
  • Sources of funding: None

REFERENCES

Publication Dates

  • Publication in this collection
    12 Mar 2021
  • Date of issue
    Mar-Apr 2021

History

  • Received
    30 Nov 2020
  • Reviewed
    11 Sept 2020
  • Accepted
    02 Dec 2020
Associação Paulista de Medicina - APM APM / Publicações Científicas, Av. Brigadeiro Luís Antonio, 278 - 7º and., 01318-901 São Paulo SP - Brazil, Tel.: +55 11 3188-4310 / 3188-4311, Fax: +55 11 3188-4255 - São Paulo - SP - Brazil
E-mail: revistas@apm.org.br