Oh man, I remember this one. There was a study from a single clinic in one ward in Tokyo that did not correct for specificity. If you assumed that the uncorrected serology estimate of this one clinic was representative of the entire city of Tokyo, you got an IFR of 0.02%https://twitter.com/AtomsksSanakan/status/1423043538103803904 …
-
-
Apologies, I'm getting my serology studies confused. The 0.04% estimate is from the Tokyo study, the 0.02% estimate here is from a similar study conducted in Kobe City at a single hospital with no correction for specificitypic.twitter.com/FC1i1aMmgM
Show this thread -
Worth noting as always - this is not a criticism of the research, because these authors were very upfront about the limitations of their work. They even said that the confidence interval for these estimates was very wide because the sens/spec of the serological test was unknown
Show this thread -
Interestingly, a recent third-party evaluation of the serological test used indicates that it has a specificity of 98 (94-100)%, which means it's possible all positives in the sample were false positives and the true prevalence was 0% (increasing the IFR substantially)
Show this thread -
They used the Kurabo rapid antibody test, which was evaluated here:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7959261/ …
Show this thread
End of conversation
New conversation -
-
-
Wow, a guide of ‘what not to do in epidemiology.’ I reckon we could teach students just by using your Twitter feed as the core reading
-
Yeh, everyone learns about denominators in epi class and then published research just completely ignores them I guess
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.