| shikamaru317 said: https://reason.com/2020/04/20/l-a-county-antibody-tests-suggest-the-fatality-rate-for-covid-19-is-much-lower-than-people-feared/?amp&__twitter_impression=true |
While it would be great to have a lower death rate, it also mean the virus is a lot more infectious and harder to stop. From that link:
Since the number of infections in Los Angeles County is much higher than the official numbers indicate, Ferrer said at the press briefing, the risk of infection is correspondingly higher, which reinforces the case for social distancing measures. At the same time, she said, the fact that 95 percent or so of the county's adult population remains uninfected shows those measures, including the statewide lockdown, are working.
Even with 40x higher than rest results, that still only gets Italy to 12% infected as a country. If it is that infectious as the study suggests herd immunity would also not be completely effective until most of the population has anti bodies. So it could still get 6 to 8 times as bad in Italy before they're done with it without stopping the spread early. That would mean another 150K deaths are possible in Italy, likely more. Taking the current 24K plus another 1K from the current serious cases, multiplied by 6. However there are likely still uncounted deaths.
So all this speculation does is reinforce the need for extending the measures.
Anyway a test of 863 adults, where about 35 were found to have anti bodies, is far too small imo. The margin of error they assume is too small imo, and it alread leads to 22 to 58 with 40 in the middle.
Plus you're always comparing apples with oranges. Testing methods and criteria are different everywhere as well as amount of available tests and testing capacity. In Germany it wouldn't be that much higher. What would be useful to do mas anti body tests in South Korea where the virus is now under control to see how many people it actually reached during the early stages. Testing all (or random 10%) of Wuhan would be better to get an accurate idea but China numbers are wonky anyway. 1.1 million tests is a lot as well, testing 1%, 110K tests is do-able but more and more bias can creep in. Small tests are only representative when results are big, not when it's a couple dozen positives.
