Thread by @otavio_ranzani

It's out 1st Brazilian national survey SarsCOV2 seroprevalence (preprint)

▶️113 sentinel cities
▶️Multistage sampling: probability proportionate to size in each city
▶️54.4% overall response rate (caveats👇)
▶️Lateral flow rapid test

Adj. prev. for 5 regions: from ~0.0 to 6.3%

Below the results for the 5 macro-regions.

We should remember that the epidemic is heterogenous in time within Brazil, so each region/area was sampled its own phase during the epidemic

Small variation by Age, but some variation for Color/ethnicity

Simple CORRELATION between prevalence and cases and deaths per capita

Brief of some limitations (read the manuscript):

1-Each region is on its own phase, so its a seroprevalence on an ongoing epidemic

2-The rapid test is not ideal, and some issues with sensitivity/specificity and all discussions around it (see validations in the text)

1/2

3-Response rate 54.4%: aimed 250 individuals in each city, which might be debatable (ie, proportionate to size within each city, not overall Brazil)

But interviewers were arrested, tests destroyed, by local police forces. The response rate varied among regions, cities, etc.

medrxiv.org/content/10.110…

cc @AdamJKucharski @nataliexdean @agbasteiro @parolin_ricardo @analise_covid19 @lsubissi @EricTopol @oatila @tarantulae @PauloLotufo @DrSamirBhatt
@CT_Bergstrom @TaschnerNatalia

cc @dadourado @leticiakawano @juliapescarini @MBarreto1954 @rede_covida @NOIS_PUCRio @legidoquigley @GidMK @carlos_chaccour

For those who don't know, the supplementary material can be found here

Supplementary material tab on @medrxivpreprint 👇