By Gil Vicente Reis de Figueiredo*
Scenarios and estimates about the spread of Covid-19 in Brazil
Introduction
The pandemic caused by the new coronavirus has been evolving at an alarming rate, surpassing one million confirmed cases, with more than fifty thousand deaths. The United States, Italy and Spain were hit hard, and other countries are following suit; many adopted severe measures, some belatedly, in an attempt to contain the disease, determining strong social confinement. This situation, unprecedented in many respects, has been raising a series of doubts, in particular regarding the intensity, duration and effectiveness of the imposition of measures to restrict social contacts, as a way of confronting Covid-19, with a view to preserving the capacity of the health systems, and, therefore, the defense of life.
It is not possible, today, to respond precisely to the above concerns. However, two recent studies, one produced by Dowd, Mills and collaborators, at the University of Oxford[I], and the other by Walker, Whittaker et al. Imperial College[ii], make it possible to assess a series of relevant aspects linked to these issues, and even suggest some alternatives for action with the aim of reducing the negative impact of the ongoing pandemic, albeit in a preliminary manner, subject to careful monitoring and subsequent verification, given that there are characteristics of the dynamics of the disease that are still not sufficiently clear at this time.
For these reasons, an initial warning is in order: the data that follows, although they contain numbers in several passages, should only be seen as qualitative indications, given the lack of reliable information on several factors, among them the number of infected people – than it follows that we do not even know for sure what the lethality of Covid-19 is or how it varies in different countries.
Scenarios
One of the main challenges that Covid-19 has been imposing refers to the dilemma of what to do to avoid accelerated contagion, since examples from several central countries have shown that very well-structured health systems can, in this case, be quickly taken to the collapse. Walker and Whittaker's article analyzes, in this regard, four scenarios: that of “maintenance”, defined by the pure and simple continuity of all the usual activities, as if nothing were happening; that of “reduction”, characterized by a 40% decrease in everyone's social contacts; the “extended reduction”, the same as the previous one, with an additional feature: the social contacts of the elderly are reduced by a higher percentage, 60%; and, finally, “suppression”, which consists of increasing these percentages to 75% for the entire population. This last scenario is subdivided into two: that of “early suppression”, in which the necessary measures are taken when the number of deaths is around 0,2 per week, per 100.000 inhabitants; and that of “delayed suppression”, in which case the number of deaths is already 1,6 per week per 100.000 inhabitants.
The authors, in their study, present a comprehensive approach, in which they analyze the pandemic globally, taking into account several factors that, in different regions, influence the action of Covid-19, among which they highlight: the level of income; the age pyramid; the average number of people in each dwelling; and the structure of the health system.
Our purpose here is to apply the model and simulations presented in the cited article to the Brazilian situation, with the necessary details so that the impact of decisions related to the pandemic can be evaluated, in particular with regard to – in the different scenarios – the projections referring to : the number of infections; the demand for hospital beds; the demand for ICU beds; the number of deaths; peak demand for ICU beds; and the expected duration of the crisis. The article prepared by Oxford researchers, in turn, will be used to focus on the impact of Covid-19 in Brazil, by age group.
Projections for Brazil
We present below the results of applying the Imperial College study to the Brazilian case, with data on the impact of COVID-19 in Brazil in the four scenarios described above. The projections shown in the graphs that follow were obtained considering simulations[iii]com Ro ('Basic Reproduction Number') equal to 3. As for the Early Suppression scenario, as proposed by the authors of the article in question, we used the value No = 0,2 deaths/week per 100.000 inhabitants.
Demand for hospital beds
In the scenario of maintaining social contacts at the usual levels, without any restriction, the projections for the total demand for hospital beds (i.e., extra beds, during the pandemic period) in Brazil for COVID-19 patients would be almost 5,9, 250 million; at the other end, in the early suppression scenario, around XNUMX beds would be needed. The table presents the projection of demand for hospital beds in Brazil, in the four scenarios.
Source: The Global Impact of COVID-19 and Strategies for Mitigation and Suppression.
Author's elaboration.
To evaluate the projected demand in regions, states or municipalities, the table below can be used, under the hypothesis that it is proportional to the national[iv].
Demand for ICU beds
With regard to the demand for ICU (Intensive Care Unit) beds, the projections are as follows:
Source: The Global Impact of COVID-19 and Strategies for Mitigation and Suppression.
Author's elaboration.
Here is the projected total demand for ICU beds per 100.000 inhabitants during the pandemic:
Peak demand for ICU beds
An important indicator is the forecast of peak ICU demand, because this is one of the factors that can lead to the destabilization and collapse of the health system. The graph below shows the projections for the ICU demand curves during the course of the pandemic, in the four screened scenarios. The horizontal axis is time, in days.
Source: The Global Impact of COVID-19 and Strategies for Mitigation and Suppression. Author's elaboration.
In the scenario of Early Suppression[v], according to projections, the peak demand for ICUs will be 7,3 per 100.000 inhabitants, while, in the other scenarios, it will vary between 70 and 220. In Brazil, this translates into demand peaks, which depend on the scenario . For the Maintenance scenario, the projection is around 467.000 ICU beds; for the Scale-down scenario, 170.000 would be required, for the Extended Scale-down scenario, 148.000, and for the Early Suppression scenario, approximately 15.000.
In order to have an idea of the meaning of these numbers, it is worth mentioning the data published on March 16, 2020 by the Ministry of Health (https://www.saude.gov.br/noticias/agencia-saude/46547-estados-terao-r-432-milhoes-para-enfrentar-covid-19): “The country has an offer of 55.101 intensive care beds. Of this total, 27.445 are from the SUS, with an average occupancy rate of 78%. The folder already plans actions to improve the use of idle units with management measures such as: bed rotation, optimization of resources and occupancy (adapting the scale of non-emergency procedures). Currently, the average length of stay in an intensive bed in the public network is 8 days. The Ministry of Health must also regulate unoccupied beds in the private network”.
Further on, the Ministry of Health reported that it intends to make available another 2 beds in mobile ICUs, with quick installation. As you can see, therefore, the maximum that can be expected today, in terms of ICU beds available for Covid-19 patients, is the approximately 22% of beds eventually not occupied by other patients – 12.000 beds – and more the 2.000 that the Ministry of Health hopes to obtain. These numbers are compatible, and even so already at the limit, with the Early Suppression scenario, in which confinement takes place with a 75% reduction in social contacts for the entire population, since, even in this scenario, it is projected to demand for 15.000 ICU beds dedicated to Covid-19 patients.
In the other scenarios, which would demand somewhere between 148.000 and 467.000 new ICU beds, the Brazilian health system would completely collapse and the number of deaths would be catastrophic. There is, therefore, not the slightest possibility of thinking about adopting any alternative other than this one, otherwise there will be a total strangulation of the Brazilian health system.
Projections of the number of deaths
The numbers of deaths, in scenarios other than Early Suppression, result from the fact that COVID-19, in severe cases, strongly attacks the lungs and airways and, as a result, it is a necessary condition to have ICUs with respirators . Therefore, at peak demand – in all scenarios except the one mentioned – there will be several hundred thousand patients in need of ventilators and ICU beds, but only a few tens of thousands will be available, i.e., an order of magnitude less . Hence the very high numbers of deaths that we will present below.
It is worth insisting, once again, on the fact that these values, like all others mentioned, have qualitative content, but cannot be thought of with numerical determinism, since there is uncertainty and even lack of knowledge in relation to information of various orders.
Early Suspension
No other decision is appropriate at this time, other than confinement, until the course of the pandemic is clearer. There are at least three important reasons for this.
The first, and obviously the most important, is the preservation of lives. From this point of view, the comparison of existing possibilities leaves no room for doubt: the model in question points, in the case of Early Suspension, to the loss of a much smaller number of lives than would happen in the other alternatives.
A second issue concerns the number of people who – according to projections – would get sick. Here, again, the contrast is stark: the number of contagions is significantly lower in the Early Suspension hypothesis; that is, a significant portion of the population, in this case, will be spared the physical strain imposed by a serious illness.
A third and no less relevant aspect is the collapse of the health system which, according to the projected values, would happen in all scenarios, except for Early Suspension.
Consequences: (a) workers in the health sector becoming sick en masse (as happened in other countries); (b) doctors having to choose who has access to a ventilator, and therefore who lives and who dies; (c) lack of basic inputs of all kinds, such as those for preventing contamination; and (d) lack of beds, with sick people without access to necessary health care.
For a comparison, just see that, in Early Suspension, the projected peak of additional demand for ICU beds (dedicated exclusively to Covid-19 patients) is around 15.000, indicative of a strong stress for the Brazilian health system, but still within the limits of the possible; whereas, in any of the other scenarios, this peak demand would be more than ten times greater, which would very possibly lead to a situation of complete chaos and untold loss of human life. There is no doubt, then, that the choice of confinement is the only sensible alternative available to us.
Impact by age group
Another aspect to consider, especially when moving back to normal life, is the fact that the risk of death is very different by age group, although this difference varies from country to country. The article published by researchers at the University of Oxford[vi] presents numerical projections, by age group, on mortality from COVID-19 in several countries, including Brazil. Comparing the results of this article with the IBGE numbers (inhabitants by age group, in 2020 – reference already mentioned above), we arrive at the following table.
The data above show that, in any scenario, it is prudent to adopt a progressively greater social distancing in the case of older people.
Financial support
The closure of trade, the pronounced reduction in the circulation of people and vehicles, and other characteristics of the imposition of social confinement, result in the near paralysis of economic activity. Informal workers cannot earn their daily living. Those who have a formal job run the risk of losing it, especially if they work in a micro-enterprise whose owner has not obtained financial support himself. Nor are small and medium-sized entrepreneurs able to support themselves, whose working capital reserves are scarce – which could generate a chain reaction, with mass layoffs. Not to mention the unemployed, who already depended on others who, in the current circumstances, will certainly have more difficulties to help them.
Therein lies the big problem: although, with confinement, there is the prospect of starting to get out of the health crisis in a shorter time than in other hypotheses, there is no way to sustain this situation over this period, even if it is shorter, without providing resources to those most in need. It is urgent, therefore, that a robust package of financial support be quickly allocated – as has been the case in several other countries – to those most vulnerable to the current crisis, as well as the agile and efficient distribution of these resources, otherwise the money will take too long to reach the tip .
It is time, therefore, to demand from the federal government the prompt solution of this impasse, even requesting the due contribution of those who hold big capital – from bankers to rentiers, from heavyweight entrepreneurs to owners of large fortunes. If this does not happen immediately, the pressure to return to economic activity will be immense, because the most diverse sectors, with good reason, will call for the end of confinement, without which their financial situation will be irreversibly made unfeasible – not to mention those that, today, they don't even have the resources to buy the basics; this could have extremely harmful consequences, since the early interruption of the restrictions on social contacts in force would leave the country in a scenario that has led to the disruption and collapse of some of the best health systems in the world.
The three structuring axes to face COVID-19
The first of these is the decision to implement tough measures to restrict social contacts – as, in a way, it has already been done, and needs to be sustained. The second is just as essential as the first: it is about, contrary to everything that federal governments have done in recent years, transferring resources to those who most need them, massively and in the very short term. The reason for this is simple. The vast majority of the Brazilian population will not be able to survive the journey of the next few weeks, weakened as it is economically by unemployment, by the reforms that took away income and wealth, by the real drop in their wages, by the change in labor relations in their disfavor, and by the accelerated reduction in the provision of essential public services, as a result of measures such as Constitutional Amendment 95. It is of immediate urgency to take this attitude, otherwise it will not be possible to sustain the social isolation that, from the point of view of public health and life, it is the only alternative to the catastrophic loss of life that can be drawn.
Finally, these two measures together, while crucial and necessary, are still largely insufficient. As it is possible to see in the projections presented, the success of social distancing will have, as a counterpart, a consequence that is positive and at the same time worrying: the contamination of a small portion of the population, which means that a good part of the rest will not have immunity to the disease.
Therefore, the relaxation of restrictive measures can only take place if we have secure information about who is not contaminated and who is – in which case you will need to maintain the proper quarantine. As, due to the very characteristic of the new coronavirus, many are asymptomatic, there is only one way to solve this problem: it is to obtain and apply a very expressive amount of rapid tests, capable of reaching a large portion of the population. Otherwise, as the study of Imperial College demonstrates, the suspension of restrictions will give rise to a new outbreak, although of much smaller dimensions than the original.
The central issue here is that the decision to lift confinement, even gradually, cannot be the result of 'guesswork'. It will be necessary to bring together researchers from the most diverse areas, so that, through tests to be applied to the population, with expertise in the field of health and statistics, they can serve as a source of data to feed models capable of making the most reliable projections possible on the dynamics of the pandemic when returning to usual activities and contacts. In this era of forced virtual meetings, it will be a great challenge to work collectively – because several areas will have to be involved – not only locally and regionally but also nationally. This is because the logistics of returning to 'normality' will have to take into account, in an integrated way, different situations, both from the point of view of health and economic activity and its priority sectors, and also from the logistics of the process itself.
The following will have to be evaluated: the percentage of people who, in a given community, have already acquired immunity and are no longer transmitting the disease; how many, in risk groups, are still susceptible; how many are still infected and in a favorable phase to spread the disease – among many other variables. With these data, it will most likely be possible, based on models built and adapted to the conditions of the pandemic in our country, to make projections on the magnitude of a “second wave” of contamination.
The fundamental thing will be to try to guarantee that this new episode, perhaps inevitable, does not produce a number of new cases of the disease that, once again, put the existing health systems at risk, and, therefore, do not generate new losses of life. If we are not able to produce knowledge that can support the lifting of confinement, we will be at the mercy of a debate that, unfortunately, can become sterile and irrational – and this will have to be avoided at all costs.
Conclusion
Serious studies and models that we have today show that the temporary implementation of strong restrictions on social contacts is the best way to face Covid-19. Sustaining this situation will require financial support for the most affected sectors and people. The massive application of tests and the construction of models that allow deciding when and how to lift confinement will also be essential tasks.
We have to face these three issues simultaneously. However, it is urgent and urgent to demand from the federal government and other powers of the Republic that they promptly take the measures that are their responsibility in the economic area, without which the consequences could be inexorable pressure to return to activities, which would almost certainly result in the collapse of the Brazilian health system and the inevitable and unacceptable loss of hundreds of thousands of lives.
*Gil Vicente Reis de Figueiredo is a retired professor at the Department of Mathematics at UFSCar.
Notes
[I] Demographic science aids in understanding the spread and fatality rates of COVID-19. March 15th, 2020. Jennifer Beam Dowd*, Melinda C. Mills*, Valentina Rotondi, Liliana Andriano, David M. Brazel, Per Block, Xuejie Ding, Yan Liu. Leverhulme Center for Demographic Science, University of Oxford & Nuffield College, UK. Available here: Accessed March 27, 2020.
[ii] The Global Impact of COVID-19 and Strategies for Mitigation and Suppression. March 26, 2020. Patrick GT Walker*, Charles Whittaker*, Oliver Watson, Marc Baguelin, Kylie EC Ainslie, Sangeeta Bhatia, Samir Bhatt, Adhiratha Boonyasiri, Olivia Boyd, Lorenzo Cattarino, Zulma Cucunubá, Gina Cuomo-Dannenburg, Amy Dighe , Christl A Donnelly, Ilaria Dorigatti, Sabine van Elsland, Rich FitzJohn, Seth Flaxman, Han Fu, Katy Gaythorpe, Lily Geidelberg, Nicholas Grassly, Will Green, Arran Hamlet, Katharina Hauck, David Haw, Sarah Hayes, Wes Hinsley, Natsuko Imai , David Jorgensen, Edward Knock, Daniel Laydon, Swapnil Mishra, Gemma Nedjati-Gilani, Lucy C Okell, Steven Riley, Hayley Thompson, Juliette Unwin, Robert Verity, Michaela Vollmer, Caroline Walters, Hao Wei Wang, Yuanrong Wang, Peter Winskill, Xiaoyue Xi, Neil M Ferguson, Azra C Ghani, Imperial College COVIE-19 Response Team. Available here: Accessed March 27, 2020.
[iii] The article by Imperial College researchers presents several simulations, for different parameters of Ro, (which is the 'Basic Playback Number'). The following values are considered: 2,4; 2,7; 3,0; and 3,3. In this work, we used the simulations referring to the value 3,0, because, according to some studies in progress (not yet conclusive, therefore), the most widespread virus in Brazil has genetic similarity with the European version, whose Ro is close to 3,0, unlike the virus variant most present in the United States of America, whose RO , also evaluated on a preliminary basis, is 2,4. It is also necessary to consider that any choice like this involves a certain degree of uncertainty, not least because the virus can present mutations that may change the adopted values.
[iv] According to the publication of the Brazilian Institute of Geography and Statistics, IBGE (2013), entitled 'Brazil, Population projection by sex and simple age on July 1st – 2000/2060), the Brazilian population on July 1st, 2020 will be of 212.077.375.
[v] Early Suppression, as defined in Article Imperial College, is the scenario resulting from the drastic reduction of social contacts, by at least 75%, for the entire population, provided that this occurs right at the beginning of the pandemic, with a number of deaths per week of less than 0,2 per 100.000 inhabitants. In the graph, this is the green curve, which has a peak demand of 7 ICUs per 100.000 inhabitants. The blue, yellow and pink curves correspond to the Suppression scenario, but starting at a time when there are already 0,6, 1,6 and 3,2 deaths per week per 100.000 inhabitants. The red curve corresponds to the scenario in which no action is taken to reduce social contacts (the 'Maintenance' scenario). And the curves on the right correspond to the Reduce and Extended Reduce scenarios.
[vi] Demographic science aids in understanding the spread and fatality rates of COVID-19. March 15, 2020. Jennifer Beam Dowd*, Melinda C. Mills*, Valentina Rotondi, Liliana Andriano, David M. Brazel, Per Block, Xuejie Ding, Yan Liu. Leverhulme Center for Demographic Science, University of Oxford & Nuffield College, UK.
