Spanish COVID-19 IFR

Well, that's terrifying (as usual, emboldening mine):

Torcuato writes:
Dear Twivers,

The Spanish Government has just announced the initial, preliminary results of its ENECOVID 19 serologic survey for SARS-cov-2. It is based on random, representative sample of over 60000 people selected by the Instituto Nacional de Estadística (akin to the US Census Bureau).  Rigorous stuff.

Everyone had a “quick” antibody test and most also consented to blood draws for an Elisa test (only a quarter of the samples have been analyzed so far though, so the Elisa results do not feature in the preliminary analysis). These will be repeated two more times, in 3 and 6 weeks’ time.

Formal publication is expected soon, but here are some highlights (alas, sans confidence intervals):

-Overall IgG prevalence of 5% nationally.

-Significant variation across provinces (from 1.6% to 14.2%).

-Among people who reported experiencing anosmia at some point in the previous two months, IgG prevalence is 42% (!).

The infection mortality rate was not specified, but here is my back-of-the-envelope calculation:

-Given that the population is 47 million, about 2,35 million are estimated IgG positive.

-PCR-confirmed deaths stand at 27100 (as of 5/13), but there is unaccounted for excess mortality (6000 as of 4/28), so actual number could be as high as 33000.

-If we are not too fuzzy, this gives a rate between 1,15% and 1,4%.

As a note, they mentioned in the press conference that in those individuals with Elisa already performed, the two kind of tests agreed 97% of the time, if I recall correctly (given the low prevalence, presumably they agree on a lot of true negatives). Official press note in Spanish:
https://www.isciii.es/Noticias/Noticias/Paginas/Noticias/PrimerosDatosEstudioENECOVID19.aspx

This is probably the highest IFR (infection fatality rate) I have seen published anywhere.