Correction in the projection of COVID-19 cases in Brazil based on data from the University of Maryland and Facebook.
On May 31, we published an article in Folha de São Paulo entitled "A Tsunami Approaches" about the arrival of a third wave of COVID-19 in Brazil. The analysis was based on data from Facebook users, with the University of Maryland, made available on platform Conass with Vital Strategies.
Trends pointed out by the survey had a direct correlation with the cases notified by the Ministry of Health, as observed on March 29th. However, we were informed by the Facebook team yesterday, June 7th, that there has been a change in the way the survey was conducted.
In short, the huge increase in Covid cases projected by the Facebook survey, while it did happen, did not occur in the magnitude previously envisioned. Part of the increase was caused by a methodological change that led to an increase in the survey response rate.
In our article, we point out that the country could reach 115,000 new cases a day in June, if the correlation between research and real data remained constant (chart below).
Now, knowing this change and correcting to use the survey version without methodological change (backup survey, which kept to 20% of users), this number is 85,000 new cases daily (simulation corrected below).
Although lower than the projection pointed out, this number is still the highest in the historical series, surpassing the peak of 03/29. Therefore, there is a clear reversal of the downward trend in cases and an increase in people with symptoms that surpass the peak of the second wave. The main message of the article, therefore, remains.
The magnitude is smaller, but expressive and worrisome. Unlike the previous wave, today we no longer have the structure to support an increase in hospitalized patients, as ICUs are already operating at full capacity in most of the country, given the change in the age profile of hospitalized patients. Importantly, the search is valid and the alert continues.