https://github.com/mbevand/covid19-age-stratified-ifr … says that "The overall IFR estimates of COVID-19, with the exception of Levin et al., are relatively consistent with each other, usually within 30-40%. Levin et al. is often up to 2-fold higher than the others". What's the main reason for this discrepancy?
-
-
Replying to @hippopedoid @zorinaq
I don't think that's true actually. Our top range for the 80+ is a bit higher, but broadly all of the estimates that are aggregates rather than from a single sample are very similar
1 reply 0 retweets 0 likes -
They do look similar in the figure, but I think the quote refers to the age-integrated estimates for different countries as listed in the Table. E.g. for Italy: ENE-COVID / COVID: US CDC /COVID: Verity / COVID: Levin 1.065 /1.092 / 1.382 / 2.177 This does look different.
3 replies 0 retweets 1 like -
Atomsk's Sanakan Retweeted Atomsk's Sanakan
Re: "ENE-COVID" As noted, ENE-COVID leaves out nursing home deaths. https://twitter.com/AtomsksSanakan/status/1332799077122248706 … The CDC analysis leaves out those 80 and older. https://twitter.com/AtomsksSanakan/status/1310242459235090434 … Verity et al. worked from the Diamond Princess, and under-estimated deaths therehttps://twitter.com/AtomsksSanakan/status/1289061193471201281 …
Atomsk's Sanakan added,
Atomsk's Sanakan @AtomsksSanakanWhy didn't Ferguson et al. over-estimate IFR? 1) Verity et al. predict 12-13 deaths on the Diamond Princess, at a time when only 7 deaths had happened. 2) Verity et al use 12-13 for IFR. 3) Ferguson et al use Verity et al's IFR work. 4) 14 deaths happen https://twitter.com/AtomsksSanakan/status/1283302606065020928 …Show this thread2 replies 0 retweets 3 likes -
Hmm. Thanks. I am a bit confused about ENE-COVID because here in Lancet https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31483-5/fulltext … I don't see any mentions of the nursing home deaths excluded... but here in BMJ https://www.bmj.com/content/371/bmj.m4509 … they are indeed stated as excluded.
1 reply 0 retweets 1 like -
Levin et al. includes an IFR for Spain where nursing home deaths are put back in. That gave an IFR of ~1.9%, given the death lag used in Levin et al. for confirmed deaths. The BMJ ENE-COVID paper instead gave an IFR of ~0.8%. https://www.bmj.com/content/371/bmj.m4509 … https://link.springer.com/article/10.1007/s10654-020-00698-1 …pic.twitter.com/bex9gzjLe3
1 reply 0 retweets 2 likes -
I don't understand how Levin et al. got 1.9% for Spain. Seroprevalence was 5.0%, that's 2.3 mln people: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31483-5/fulltext …. Cumulative covid fatalities even 5 weeks later were only 28k (Table 2 in Levin et al.) That's IFR=1.2%. Even when I take 40k excess deaths, I get 1.7%.
2 replies 0 retweets 1 like -
Replying to @hippopedoid @AtomsksSanakan and
@GidMK I'd be very grateful if you could clarify the above question! Thanks.1 reply 0 retweets 1 like -
I'll have to check with my co-authors, I did not do the calculations for that graph
1 reply 0 retweets 2 likes -
Thanks! But I can also write an email to the corresponding authors if that's more convenient.
1 reply 0 retweets 1 like
I am one of the corresponding authors lol
-
-
Dmitry Kobak Retweeted Dmitry Kobak
The only one actually! :-) Unrelated: this might interest you guys. I'm trying to build an infrastructure to update this regularly.https://twitter.com/hippopedoid/status/1347655523861229568 …
Dmitry Kobak added,
Dmitry Kobak @hippopedoidExcess mortality mega figure. The latest data from@HMDatabase, CDC, FT (@jburnmurdoch), NYT (@atmccann), The Economist (@J_CD_T) + Iran, Ukraine, Armenia, Georgia. Shows excess mortality, % from annual, per 100k, and undercount. Code: https://github.com/dkobak/excess-mortality … pic.twitter.com/PJ7l63vZje1 reply 0 retweets 1 like -
Sorry, completely forgot to get back to you! The main divergence in our Spanish estimate, according to Prof Levin, is that we took pains to include nursing home deaths, which are opaquely reported in Spanish data
1 reply 0 retweets 2 likes - Show replies
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.