RIO DE JANEIRO, BRAZIL – The study group, which includes researchers from Brazilian and foreign universities, has estimated that in its most pessimistic scenario, the number of deaths could be up to nine times the official figure.
The Covid-19 BR Observatory is a scientific front launched by researchers who seek to understand the development of the pandemic in Brazil, through mathematical models. The group has been analyzing the coronavirus situation for the São Paulo Municipal Health Secretariat.
The study was conducted using data provided by the Ministry of Health through last Wednesday, April 15th, when the country had recorded 1,736 deaths. Based on this number, the country probably has had somewhere between 3,800 and 15,600 deaths, according to the calculations made by UOL using the Observatory study.
To conduct the study, the observatory used data on the date of deaths submitted by the government in the reports for March 29th and 30th, and April 2nd, 4th, 5th, 6th, 7th, 8th, 10th, 11th, 13th, 14th and 15th.
Confirmations over ten days later
As an example, based on the reports released by the Ministry of Health, one can note that the number of deaths recorded on April 3rd varied from 21 to 77 during a two-week period, as the test results are slow to be released.
Later tests showed that more Covid-19 deaths occurred on that date than were originally reported.
In each of the Ministry of Health’s reports, there are updates on how many people died from the novel coronavirus by date. The data use the SIVEP-Gripe (Flu Surveillance Information System) records since the first death was reported.
The first confirmed death illustrates this lag. On March 17th, it was reported that the death had occurred the preceding day. Two weeks later, in April, it was found that another death had been the first in Brazil and that it had occurred on March 15th.
There was even a case in which the death took almost a month to enter the official body count statistics.
According to the Observatory’s survey, 53 percent of Covid-19 deaths in Brazil take ten days or longer to be reported and featured in reports. “That is, fewer than half of the deaths are confirmed within ten days from date of death,” reports the study.
Nowcasting, an estimation model of the present looking at the recent past, was used by the Observatory to make its calculations. The projection takes into account how many deaths should occur considering a lag in distribution.
“We look at how long, on average, it takes between the date of death and the date of the announcement of the death [by the ministry],” explains Rafael Lopes Paixão da Silva, PhD student at the IFT (Institute of Theoretical Physics) of Unesp (Universidade Estadual Paulista), and a member of the Covid-19 BR Observatory. “This gives us the distribution of the delays”.
The Observatory projected two scenarios on the actual number of deaths in the country. According to the pessimistic scenario, the country has had nine (9) times more deaths than those reported; under the more conservative scenario, there have been twice as many deaths.
Ten-day scenario
To analyze the data, the Observatory designed two scenarios. In the most conservative, in which deaths would take up to ten days to be reported, the country had around 122 percent more deaths than were officially reported.
Using the officially reported data from April 15th, this would lead to the calculation of about 3,869 deaths, which is more than twice the number of deaths officially announced at the time: 1,736.
In this scenario, Brazil would already be at the same mortality level as China, where the pandemic broke out in December of 2019. On Wednesday, the Asian giant counted over 3,300 deaths (the figure has since been revised to around 4,600). Covid-19 reached Brazil in late February.
20 days delay
In a scenario where deaths from coronavirus take up to 20 days to be accounted for, the study projects that the country has had mortality about nine times that of the official data. “This means that for each death reported, we would have eight more deaths not yet recorded.” In this hypothesis, the country has already had over 15,648 deaths.
This number is close to the number of deaths recorded in the most heavily impacted countries in Europe on April 15th:
France: 17,167
Spain: 18,579
Italy: 21,647
“As there are too few reports [with date of death], it is not yet feasible to decide which scenario is more likely between the two options,” comments the Observatory, which points out that “the results are very sensitive to the maximum time for notification that is assumed.”
“Normal” underreporting
“We know that underreporting is not uncommon, it’s normal,” says Dr. da Silva. He says some of the reasons for the outdated data are: delays in the tests; the time to report the deaths; and the updating procedure of the state health departments’ data system, which they forward to the Ministry.
“Until these deaths are confirmed by the Ministry, it takes some time,” he says.
Researchers from universities such as Unesp, USP (University of São Paulo), Unicamp (University of Campinas), Unb (University of Brasilia), UFABC (Federal University of ABC), University of Berkeley, USA, and University of Oldenburg, Germany, among others, are involved in the Observatory initiative.
Case monitoring
For infectologist Mateus Westin, a professor at UFMG (Federal University of Minas Gerais), knowing the number of cases is more significant than the number of deaths, when considering measures to prevent the disease from spreading.
By Friday, April 17th, Brazil had 160 cases for each million Brazilians. In the same proportion, there are ten deaths per million.
Westin – who is not part of the Observatory group – says that “the number of deaths by Covid-19 is a late marker” on the circulation of the virus. In his opinion, by observing only deaths, we would be “analyzing the situation with three to four weeks delay”, considering the beginning of the period in the asymptomatic stage of the disease.
In the document “Guidelines for Diagnosis and Treatment of Covid-19,” dated April 6th, the Ministry of Health said that “the early detection of new cases is paramount for preventing the infection from spreading”.
In order to have a real picture of the presence of the novel coronavirus in the country, Evaldo Stanislau, director of the São Paulo Infectology Society, stressed the value of testing the population. “Only then can we contain the cases more effectively.”
According to him, this “would have a huge impact,…more than looking at terminal and underestimated statistics, like cases and deaths.”
Source: UOL