Lol, well Shouman had very large differences at baseline between groups, which is almost certainly because they report totally destroying their own randomisation schedule halfway through their trial. And that's just the *start*
-
-
Replying to @GidMK @GabinJean3 and
There are so many issues at this point I just assume people who trust these studies haven't actually read them
1 reply 0 retweets 0 likes -
Replying to @GidMK @jslocum1coxnet1 and
I trust the reproducibility despite the flaws. Unlike for HCQ I never saw the flaws consistently explaining the results. So I go to the most simple conclusion. I'm not seeing the baseline differences you are referring to. What I do see is the limited PCR testing.
1 reply 0 retweets 1 like -
Replying to @GabinJean3 @jslocum1coxnet1 and
But much of the reproducibility au this point is almost certainly down to fraud. And the baseline differences are obvious, just look at the index cases - the p-value for the chi-squared here is very much significantpic.twitter.com/lbIjAVAKir
1 reply 0 retweets 0 likes -
Replying to @GidMK @jslocum1coxnet1 and
I'm not sure how you infer from this that the ivermectin group was favoured by the difference in severity of the disease. If anything I would say it's the other way around, I would expect moderate disease to be most contagious. Seems far fetched this would explain results.
1 reply 0 retweets 1 like -
Replying to @GabinJean3 @jslocum1coxnet1 and
That's not true, but it's also not the point. Issues in randomisation are problematic, and this is just one of many problems
1 reply 0 retweets 0 likes -
Replying to @GidMK @GabinJean3 and
I mean, the blinding such as it was clearly failed. They appear to have almost exclusively tested people in the control group. They then included people who tested NEGATIVE in their primary outcome. They have numeric errors in the paper. It's not one problem, it's endless
1 reply 0 retweets 0 likes -
Replying to @GidMK @jslocum1coxnet1 and
The study was never blinded. I'm looking forward to your analysis on those points you raised. My general appreciation of your work is that you raise methodological doubts but are never really able to explain the results. Fraud is not a word to be lightly thrown around. Careful.
1 reply 0 retweets 2 likes -
Replying to @GabinJean3 @jslocum1coxnet1 and
The point about low-quality research is that the results such as they are become less trustworthy. Issues such as these could of course explain the results. And I know, I'm using it carefully and specifically although not about Shouman
1 reply 0 retweets 0 likes -
Replying to @GidMK @jslocum1coxnet1 and
The difference in results are usually much more easily explained by differences in dosing or delay in intervention. Reproducibility trumps methodological flaws. Some "higher quality" studies that show no benefit have very serious design flaws which I haven't seen you rant about.
2 replies 0 retweets 2 likes
That's just not true at all. Positive results can easily be explained by issues in study design, this is not some weird nit-picking but a very well-demonstrated fact
-
-
Replying to @GidMK @GabinJean3 and
We literally expect to see many poorly done, positive trials. It's precisely how bad quality research usually works, is exactly what we saw with HCQ, etc
0 replies 0 retweets 1 likeThanks. 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.