Underreporting of Covid cases masks up to 30% of deaths in Brazil, study shows

RIO DE JANEIRO, BRAZIL – The underreporting of severe cases and deaths of Covid-19, mainly due to the lack of diagnostic tests, is “masking” at least 30% of deaths that do not figure in the official statistics. Had they been considered, Brazil would count 530,000 deaths, considering that last Sunday, May 2, it totaled 407,000.

The conclusion is from a study by the Vital Strategies initiative, the main national Severe Acute Respiratory Syndrome (SARS) database based on Sivep-Gripe.

The study researched SARS cases and deaths since the beginning of the pandemic that are listed in the official statistics as “etiology unspecified or without final classification for Covid,” i.e., there was no identification of the agent causing the disease.

The data were compared to official records in the years prior to the pandemic (2018 and 2019) and subtracted the average numbers observed in these categories.

As of last April 19th, the results were as follows: an increase of 57.4% (from 1,098,254 to 1,728,955) for severe SARS cases and 29.9% (356,536 to 462,973) for deaths.

“There is no other explanation [if not Covid] for such a large increase in cases and deaths [from SARS]. There is no other acute respiratory virus acting concomitantly with Sars-CoV-2,” says researcher Ana Luiza Bierrenbach, senior technical advisor at Vital Strategies and the study’s author.

In confirmed Covid-19 cases, 96% had a diagnosis attesting to the infection. In the other SARS cases, 70% had some test, but the result was either negative or the confirmation data was missing, for example.

“Then you ask me if, despite this result, we are stating that it is Covid? I say yes,” she says.

She says that there are different diagnostic tests on the market, with a sensitivity that is not complete. There were also many SARS cases with negative PCR tests, but which had clinical and radiological findings compatible with Covid-19. In others, there was an epidemiological link (someone close, for example) with laboratory-confirmed cases.

Bierrenbach also explains that PCR/antigen test positivity decreases as the time between symptom onset and hospitalization or death increases.

“The disease has a few phases. In the replicative phase, you can find the virus in the nose or oropharynx. Then the virus goes into the tissues and you can no longer make a diagnosis only with the nasal swab [sample taken from the nose]. And sometimes the disease has progressed so far that the virus is no longer present in the tissues.”

The researcher recalls that in the Brazilian epidemiological surveillance guide and in the World Health Organization recommendations it is possible to assign a diagnosis of Covid even for people who don’t have a test confirmation. “It’s not our invention.”

According to Bierrenbach, the methodology used discounted the number of SARS cases and deaths that remain unspecified each year due to lack of investment in laboratory testing in emergency rooms.

“In Chile, a child will come in with a respiratory condition to the hospital and they will run a viral panel to identify [the agent]. In Brazil, many cases go without an etiological diagnosis.”

The study adjusted for another variable: the increase in hospitals that started reporting SARS hospitalizations and deaths. In 2019, there were 2,641 notifying units in 935 municipalities. In 2020, this number reached 6,621, representing 61.8% of the total of 5,570 Brazilian municipalities.

“By discounting these cases, we think that all the other cases are Covid. Underreporting is in the whole country. It is not localized in just one region or in just one age group, although in children the correction has been greater.”

According to the study, in the age group up to 1-year-old, there were 567 deaths attributed to Covid. With the adjustment, they rose to 1,397. Between ages 1 and 4, the difference is also large. There were 230 reported, and with the reclassification, they now total 498. Between 5 and 9 years of age, they rose from 143 to 321.

“At the start of the pandemic, there was no suspicion that the cases were Covid. The clinical manifestation of the disease in children is atypical, sometimes emerging after the viral symptoms and progressing rapidly to more severe disease and even death.”

According to her, because of low testing, the chance for early diagnosis is being lost, and people are often hospitalized late when testing is no longer as effective.

“Then it ends up like that. It looks like Covid, smells like Covid, has Covid symptoms, has an epidemiological link, but no diagnostic confirmation.”

For the researcher, the failure to classify cases and deaths without laboratory diagnosis is a major issue and should be addressed with better training or allowing these records to be reported not as confirmed for Covid-19, but as “probable.”

In England, for example, health services consider Covid deaths up to 28 days after confirmation of the disease. This would detect cardiac or neurological deaths that occur as a result of the disease’s sequelae, for example.

“Not here in Brazil. We are basically using laboratory confirmation. This in a country that doesn’t have testing or testing comes late.”

Bierrenbach says the correction of cases and deaths made in the study did not surprise the team. “The numbers are shocking, but we felt there was no other explanation for such a large number of unspecified cases. Many researchers had the feeling that there was something wrong. Our merit was to give voice to this.”

She says that the research is actually conservative because it does not take into account the mild cases diagnosed, severe cases that did not have access to diagnosis, eventual deaths that occurred at home, or even deaths as a result of Covid (a heart attack, for example) after a prolonged hospitalization.

“All of this is not accounted for in our figures. If we start adding it up, it will add up far. The numbers are frightening. I hope they will raise awareness among the population and our rulers.”

This article was produced by The Rio Times’ automated newsroom system. How we use AI · Report an error