This was an antibody test, indicating a cohort that already had the disease and survived. They should be compared to the cohorts that acquired the disease within the same time-frame and subsequently died. For those that die, from first symptoms to death is about 14 days.[1] Antibodies also do not appear till about 7 to 10 days after infection.[2] If you include deaths in the next 10 days, you must also include people who have been infected around the same period but have not yet developed antibodies (a presumably even larger number than 6000.)
You sir, are of course correct. You can directly compare currently reported deaths with currently reported anitbody tests because of the time for them to develop vs. time for death is approximately equivalent.
I've updated my post and redacted my initial comments, and drop by fatality rate estimate from 1% to 0.7%.