I didn't think of it as a blog post, I came across it on my Twitter search described in the methods. Not unfair to classify it that way, I'll have to have a think. Technically it's not a meta analysis because the IFR estimate doesn't involve statistical aggregation
-
-
Replying to @GidMK
It doesn't matter how you come across something, that's not the problem: if it doesn't fit your inclusion criteria, then it has to be excluded. (Or you revise your inclusion criteria & re-do all the methods needed to accommodate & start searching all over again.)
1 reply 0 retweets 2 likes -
Replying to @hildabast
Definitely, but the main issue was about how to classify it. I'm not sure it's fair to call the estimate a blog post, it's an expert review published on a university website - in this case, my initial thinking was that it is grey literature
1 reply 0 retweets 0 likes -
Replying to @GidMK
Fair enough - doesn't solve the original problem I was raising it for: double(triple &c) counting.
1 reply 0 retweets 0 likes -
Replying to @hildabast
Ah but I don't think it is double-counting, because the actual IFR estimate that they produce is not a statistical aggregate - it's just their expert opinion
1 reply 0 retweets 0 likes -
Replying to @GidMK
Wow. (a) If expert opinions in non-systematic review informal literature are eligible it has major implications for your inclusion criteria. (b) That's not how you describe it in the preprint table. (I'll repeat my advice: you really need a co-author with specialist SR expertise)pic.twitter.com/NBBWAq6NIO
2 replies 0 retweets 0 likes -
Replying to @hildabast
Hmmm true. I'll talk to my co-author but I think it's a good point. Probably will exclude in the next update, thanks!
1 reply 0 retweets 0 likes -
Replying to @GidMK
You need to talk to someone with more specialist methodological expertise in systematic reviews & meta-analysis, incl about whether meta-analysis is indicated at all. You need to address the point, for which CEBM was just 1 example: how often are Wuhan data counted, for example?
1 reply 0 retweets 0 likes -
Replying to @hildabast
Well there were two studies that used Wuhan data, but they were very different datasets (deaths/cases and serology two months apart). Another study inferred IFR from Beijing, and another used exported cases from China as a whole
2 replies 0 retweets 0 likes -
Replying to @GidMK
Quick glance at that table seems to be 3 with Wuhan & 2 with Diamond Princess (other than CEBM).
1 reply 0 retweets 0 likes
I suppose it does depend on what you mean by double-counting. The French study used DP and French data to build a model estimating IFR, rather than the DP modelling study from Imperial. And as I said, 2 using different data from Wuhan, one using evacuees to Japan from Wuhan
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.