Just over a year after the start of immunization against covid-19, Brazil is struggling to obtain data on the effectiveness of vaccines. The Ministry of Health, responsible for the data in the country, has not yet published any study that relates vaccination to hospitalizations and deaths caused by the disease.
According to experts in the area, in addition to not showing “interest” in measuring this effectiveness, the data management model used by the ministry makes it difficult for projects to carry out this survey.
THE UOL spoke with some of the main researchers of data projects on covid-19 in the country. Experts report that the difficulty occurs because there is a technological problem to arrive at percentages of people hospitalized or dead with complete, delayed or even non-immunized vaccination – and stratifying this in age groups, for example.
The biggest problem, they claim, is that the country has two separate data systems that are different and do not communicate with each other: the vaccination system (the SI-PNI, National Immunization Program Information System) and the hospitalization information system. and deaths (Sivep-Gripe, Influenza Epidemiological Surveillance System).
According to Isaac Schrarstzhaupt, coordinator at the Covid-19 Analysis Network, a nationwide effectiveness survey could only be done by crossing SI-PNI data with that of Sivep-Gripe. But this is not possible because the two public systems generate different (anonymous) identities in each database.
“When they anonymized the SI-PNI, they created a patient ID. When they anonymized Sivep, they created another one. So it is impossible to get the hospitalization or death information in Sivep and go to the SI-PNI to see when and how many doses the person took” , explains.
Without an ID (acronym that comes from the shortening of the word identity) to carry out this crossing, it is necessary to have access to so-called closed data, which gather personal information from patients and vaccinated persons. This, to ensure privacy, has highly restricted access.
In this case, states and municipalities have access to these closed databases and could make this local crossing using data such as name and date of birth. But it would be a lot of work, few end up doing it”.
Isaac Schrarstzhaupt, Covid-19 Analysis Network
Lack of integration requires more effort
Even for those who have access to complete data, such as federal research centers, cross-referencing information from databases becomes more complicated without a unique identifier.
“With a single number, we would be able to cross-reference the Sivep database with others, in particular the PNI. But this ended up being poorly done”, says Leonardo Bastos, a public health researcher at Fiocruz (Fundação Oswaldo Cruz) and a member of the Covid-19 BR Observatory.
According to him, without a common identifier, the crossing of information ends up demanding more effort from those who propose to perform the task.
“Doing this is difficult on several levels. These are huge databases.”
Leonardo Bastos, Fiocruz
Fiocruz has some research groups that have access to closed data. One of them is the Vigivac project, one of the few groups that published a study of the effectiveness of vaccines administered in the country until October 2021. According to the evaluation, they had an effectiveness between 83% and 99% for severe cases and deaths among age groups. 20 and 80 years old.
Bastos is preparing this cross-checking of data for a broad study on vaccine effectiveness in the country, but that will not be released in the next few days. “We are doing research at the pace of research, not at the pace that would be ideal for dissemination now to show the effectiveness”, he explains.
Recently, says Bastos, Sivep-Gripe underwent an update. Now, when you fill in an admission or death data, the vaccination status is automatically indexed. “But with the blackout, it didn’t work well. If you take the data from Sivep today, you’ll see that there is a hole. There are some data that you can still estimate, but not for most”, he adds.
No population, only estimation
Another problematic point he cites is the country’s population database — in the case of the IBGE (Brazilian Institute of Geography and Statistics). As the last Census was carried out in 2010, there are inaccuracies in the number of the population, as well as in the stratification by race, gender or age group, for example.
“Our population estimate is poor. First, because the last Census is 12 years old. In other words, the entire population from 2011 onwards is an estimate. We use the best data we still have, but it is not accurate. and a half we have percentages of vaccinated above 100%”, he says.
An example of this occurred with the state of São Paulo, which announced on the 22nd that it had vaccinated 102% of the eligible population with the first dose. The lack of knowledge about how the calculation is made even generated criticism from anti-vaccine bolsonaristas, who accused Governor João Dória (PSDB) of alleged fraud.
In the midst of this scenario, the coordinator of the Laboratory of Statistics and Data Science at Ufal (Federal University of Alagoas) and member of the Covid BR Observatory, Krerley Oliveira, says that the failure in the management of health data was clear in the blackout of information. (between December and January) —which generated a series of suspicions on the part of the researchers.
“I think people are very disappointed with the care that governments are taking with this information. And this is not due to a lack of tools or initiatives”, he says, citing that the crossing of data from the two systems is something that can be done in an internal environment of the Ministry of Health.
The ministry has all these bases, if they just want to. What we don’t have is the will to do things”.
Krerley Oliveira, Ufal
THE UOL asked the Ministry of Health about the reasons for not having a study crossing the SI-PNI data with the Sivep-Gripe to show vaccine effectiveness in the country, but received no response.
Local data show effectiveness
Due to the difficulty, few prefectures or state governments carry out vaccine effectiveness studies of patients hospitalized with covid-19.
One of the states that did this was Rio Grande do Sul. In the latest data presented, in December 2021, they point out that the complete vaccination schedule reduced the risk of death by 87% in people aged 20 years or older between August and November 2021. 2021. Among the elderly, booster vaccination reduced the incidence of death by 95% in the period.
Other states report the percentage of vaccinated and unvaccinated hospitalized patients. In Amazonas, on Thursday there were 581 hospitalized patients in Manaus divided into:
- 266 unvaccinated
- 154 have incomplete vaccination schedule
- 148 have a complete vaccination schedule
Pernambuco, which came to disclose the data, stopped. The State Health Department informed the UOL that “information on the vaccination status of some patients no longer appeared in the Ministry of Health’s information bank”. “The fact has already been reported to the federal agency for investigation”, says the folder.
In Rio de Janeiro, 88% of those hospitalized have not taken any or have incomplete doses of the vaccine.
These data are not unique to Brazil; we see it all over the world! In practice, when we see who has been hospitalized, we see a huge difference between vaccinated and unvaccinated or incompletely vaccinated. Anyone who has one dose or two — and has time for a third — is not adequately protected.”
Vice President of SBIm (Brazilian Society of Immunizations), Isabella Ballalai
Infectologist and professor at UFPE (Federal University of Pernambuco) Vera Magalhães says that all countries that have carried out vaccine effectiveness studies have reached the conclusion of a great reduction in hospitalizations and deaths.
“In the United States, 80% of hospitalized patients are not vaccinated. In England, the same thing is seen. So we can say that the vaccine, despite not preventing infection, is capable of preventing cases of hospitalization and death. was seen with the other waves, but now it is much more evident with the transmission of the absurd ômicron variant”, he concludes.