What information have I ignored?
-
-
Replying to @GidMK @EdoajoEric
More: It's hard to believe that your are unaware of the ivermectin test in Mexico with 230k people and 76% reduction is hospitalization rate. But anyway, here it ishttps://osf.io/preprints/socarxiv/r93g4/ …
1 reply 0 retweets 2 likes -
Replying to @jjchamie @EdoajoEric
Oh yes I've read that. A very big issue with residual confounding makes it quite hard to interpret, particularly given that there are some numeric errors in the tables
2 replies 0 retweets 4 likes -
Replying to @GidMK @EdoajoEric
A very big issue with residual confounding? Numeric errors in the tables? Can you be more specific? I think @PPMERlNO would like to check that out.
1 reply 0 retweets 1 like -
Replying to @jjchamie @EdoajoEric
Sure thing - the numeric errors are pretty simple, it looks like some people were excluded that are not reported as far as I can see in the text. The text says 77,381 and 156,468 but the values in the tables add up to 77,327 and 156,011 respectivelypic.twitter.com/dmXP5gURtW
1 reply 0 retweets 2 likes -
This also matches the percentages given - 48,511 is 31.09% of 156,011, not 156,468. Not a major issue, but should probably be corrected
1 reply 0 retweets 2 likes -
As to residual confounding - there are numerous potential differences between people who got packs and those who didn't. For one thing, I don't see any discussion of SES in the paper, which is a pretty big potential confounder
2 replies 0 retweets 3 likes -
Replying to @GidMK @EdoajoEric
Again, you ground your mistrust in tiny impactless items. You have an enormous improvement in hospitalization (76%) and your arguments absolutely can't explain even 10% of the impact. Do you think your points would turn the results insignificant?
1 reply 0 retweets 4 likes -
Replying to @jjchamie @EdoajoEric
Of course - residual confounding can entirely reverse the effects of an analysis. That's a very well-demonstrated epidemiological fact https://sphweb.bumc.bu.edu/otlt/mph-modules/bs/bs704-ep713_confounding-em/BS704-EP713_Confounding-EM4.html …
1 reply 0 retweets 6 likes -
Replying to @GidMK @EdoajoEric
They controlled in the analysis by age and comorbidities, and looked at more than 230,000 people. The results was 76% less hospitalizations in the group with ivermectin. After the study they rolled our IVM in the country and every COVID metric improved. CFR dropped 80%
1 reply 0 retweets 3 likes
So firstly, same issue as in your analysis - it's almost certain people in Mexico were taking the drug well before these packs were handed out. Age and comorbidities were controlled for in a fairly crude way, and there's a large potential for residual confounding
-
-
Replying to @GidMK @EdoajoEric
Agreed, it's almost certain people in Mexico were taking the drug. This factor potentially reduces the study positive result because some in the control group are treated too Your recurrent card is "potential for residual confounding" This unknown can't explain big differences
1 reply 0 retweets 5 likes -
Replying to @jjchamie @EdoajoEric
Of course it can! This is a pretty fundamental part of epidemiology
2 replies 0 retweets 4 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.