2/n The study is here, and has an objectively large impact with nearly a dozen news stories and an Altmetric of 2750https://www.medrxiv.org/content/10.1101/2021.05.31.21258081v1 …
-
Show this thread
-
3/n The study is pretty simple - 116 people with Covid-19, randomized to either receive ivermectin or a placebo Reported results were viral load, and more specifically proportion of people with Ct>30 at various time points on PCRpic.twitter.com/ozXyxZwAOj
2 replies 6 retweets 75 likesShow this thread -
4/n The authors report a marginally significant benefit at various days using an uncorrected logistic regression, and from this the massive attention sprangpic.twitter.com/5xk5vLDkrH
2 replies 7 retweets 78 likesShow this thread -
5/n One of the first things I do when I read randomized trials like this is look at their pre-registration. This is basically what the authors SAID they would do registered BEFORE they did it Here's the pre-reg for this trial:https://clinicaltrials.gov/ct2/show/NCT04429711?term=NCT04429711&draw=2&rank=1 …
2 replies 10 retweets 105 likesShow this thread -
6/n Immediately on looking at the pre-registration, there is a HUGE discrepancy with the published results Here's what the authors pre-registered as their exclusion and the flow chart from the study. Notice the difference?pic.twitter.com/eMFPtwG9Y6
4 replies 10 retweets 93 likesShow this thread -
7/n It appears that the authors have added an extra exclusion criteria that is applied AFTER randomization and treatment. This is "tested Ct>35" in the first two testspic.twitter.com/V5ZfCZWhv7
5 replies 10 retweets 109 likesShow this thread -
8/n Based on the protocol, that implies the sequence of events went like this: 1. patient tests positive, randomized and started treatment 2. Day 2, patient tests Ct>35 3. Day 4, patient tests Ct>35
4 replies 6 retweets 72 likesShow this thread -
9/n This is an issue for two reasons - firstly, it's not what the authors said they'd do. Secondly, it means that they are excluding people WHO MEET THEIR PRIMARY OUTCOME (i.e. Ct>30) after randomization and treatment has started
2 replies 11 retweets 144 likesShow this thread -
Replying to @GidMK
While it might be reasonable to exclude those meeting the primary endpoint at baseline, this is not what is being done. They are including those w/ 30-35 PCR Ct, who then immediately meet the primary endpoint. When was this decision made? Is it specified in the protocol? SAP?
2 replies 1 retweet 3 likes -
Replying to @boulware_dr @GidMK
Clinicaltrials gov listing of endpoint is very vague. No Ct value appears there. When was the threshold determined? Before or after looking at unblinded data? (also listing is currently non-compliant w/ no updates in >1 yr) I would like to see the protocol & SAP (all versions).
2 replies 0 retweets 3 likes
Based on the preprint the original endpoint was Ct>40 but this was changed at some point. This also makes sense with the changed exclusion criteria for Ct values, but crucially no report of when all of this was changed
-
-
Replying to @GidMK @boulware_dr
You can check here his explanation.https://www.youtube.com/watch?v=htt1mlvBwBA&t=2701s …
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.