3/n I should say at the outset here - the only personal comment I would like to make about Professor Ioannidis is that he is a very smart man who I respect tremendously I will, however, examine the paper, because I think that is what science is all about
-
Show this thread
-
4/n At first glance, and indeed on deeper reading, it is clear that very little has changed from my previous looks into the paper
1 reply 0 retweets 58 likesShow this thread -
5/n The methodology is still the same, and the eventual conclusion remains that the median IFR of COVID-19 is 0.27% (originally he estimated 0.26%)pic.twitter.com/kBhIWCEUAL
3 replies 2 retweets 64 likesShow this thread -
6/n The author then concludes that the IFR "tended to be much lower than estimates made earlier in the pandemic", which is odd because his own estimates made earlier in the pandemic (in May) were...lowerpic.twitter.com/7awIA4gEoW
4 replies 5 retweets 110 likesShow this thread -
7/n Indeed, as we can easily see, the resulting low IFR is simply a consequence of the low quality of the review itself and has very little to do with when the estimates were made
1 reply 5 retweets 92 likesShow this thread -
8/n For example, the review does not adhere to PRISMA guidelines (the most basic recommendations for reviews of this kind) which is very strange given that Prof Ioannidis himself is a co-author on the original PRISMA statementpic.twitter.com/fgwRh67bkL
2 replies 20 retweets 181 likesShow this thread -
9/n This has lead to a problematic situation, where there is no rating for study quality, publication bias, and indeed little consideration in the manuscript for how the quality of the published evidence might impact the review
1 reply 3 retweets 100 likesShow this thread -
10/n As we pointed out in our systematic review and meta-analysis of COVID-19 IFR, this is an issue because higher-quality studies tend to show a lower seroprevalence and thus a higher IFR https://www.sciencedirect.com/science/article/pii/S1201971220321809 …pic.twitter.com/Jf2TJH3GVP
3 replies 12 retweets 138 likesShow this thread -
11/n (Interestingly, Ioannidis cites our study but gets the numbers wrong, in what is distressingly something of a trend in the paper generally - we actually estimated 0.68% in the published paper which came out recently)pic.twitter.com/63z3Mw61uS
4 replies 11 retweets 117 likesShow this thread -
Replying to @GidMK
Independent double-checking of data extractions once more shows its importance. This is why we teach people not to do SRs on their own.
1 reply 1 retweet 7 likes
Yes! Never a good idea to take on such a mammoth effort by yourself, no matter how eminent a professor you may be
-
-
Replying to @GidMK
Exactly so. No one is above a simple transcription error
0 replies 0 retweets 4 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Replying to @GidMK @soniaahhh
Thanks ..My head hurts..
It boils down to ..looking at all the stats, the figure is probably around 5 but even that can’t be certain?0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
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.