Let them eat their horse paste.
1 Like
Does anyone remember laetrile, the purported cancer treatment drug from the 1970âs? Same arguments (on both sides) were given back then too.
You can be more explicit about the criteria of a âwell-designedâ study for this purposeâitâs just statistical power. If you have a high-power study that fails to reject the null, then itâs strong evidence that the null hypothesis is true. If a low-power study fails to reject the null, then itâs not strong evidence of anything.
As an example, this study found that the patients who were treated with ivermectin got better two days earlier, on average, then patients who got a placebo. Now, thatâs not strong evidence that ivermectin makes people get better sooner, but itâs also obviously canât be strong evidence that it doesnât. (If the drug shortened recovery times by two days, this is exactly what youâd expect to see.). The reason is that the trial has low power for small effect sizes.
1 Like
This is wholly irrelevant. If you wish to make a point, feel free to make it directly.
No, it didnât:
â the duration of symptoms was not significantly different for patients who received a 5-day course of ivermectin compared with placebo â
1 Like
Just answer the questions man. If theyâre irrelevant, it shouldnât matter what you say, right?
It doesnât matter, which is why Iâm willing to do zero work on whatever nonsense youâre trying to do here. Iâm not willing to look into the data behind novavax whatsoever to argue with someone about a fundamental tenet of EBM. You can make any argument you want directly. Your approach here is not honest.
1 Like
Thank God we have you here to explain these things.
There is no data behind Novavax. Trials are ongoing. Does that help?
1 Like
I know you think this is inconsistent, but itâs not. The actual observation was that the patients who got ivermectin recovered sooner. The power of the study was such that that observation was not reliable evidence of a real effect.
I really think the play here was just trolling, but oh well.
Umm we do apparently? Bob made a wholly false claim about the topic. You now are saying that thereâs no data behind novavax when thereâs a stage 3 rct trial in JAMA.
Simply incorrect, which is why the paper doesnât actually say that at all. You can not use statistically insignificant data to make a claim like this. Youâll notice that the paper doesnât actually say this whatsoever.
1 Like
The fact that there was a difference between treatment and control which didnât rise to statistical significance was actually integral to the point he was making, but as usual your first port of call is just to assume everyone else is thick.
My bad. There are phase 3 trials still in progress, I didnât realise one was done already. I feel like you know very well what my point is and are pretending you donât though.
Jesus Christ man. I retract my post above about feeling like you know what my point is.
1 Like
Yâall have zero experience with evidence based medicine. Yâall are making super basic and obvious mistakes that are explicitly taught as things not to do. You guys havenât found this one quick trick with EBM, you just donât understand it.
The null hypothesis wins. You cannot say that statistical insignificant differences show differences between a control and experimental group because maybe it wasnât just powered enough.
4 Likes
Dicing a little deeper into this paper:
âThe primary outcome was originally defined as the time from randomization until worsening by 2 points on the 8-category ordinal scale. According to the literature, approximately 18% of patients were expected to have such an outcome.23 However, before the interim analysis, it became apparent that the pooled event rate of worsening by 2 points was substantially lower than the initial 18% expectation, requiring an unattainable sample sizeâ
Woof that is not good. They changed their primary end point after they started getting their data because it was going to be a negative study, and it was still negative.
1 Like
Youâre semantikally correct here actually. The median patient who took ivermectin recovered two days earlier than the median patient on a placebo. They didnât report the average (at least in the part of the paper I read).
As for the rest of it, weâre not talking about making claims, weâre talking about actual observations. They gave the drug and the placebo to real actual people, and the median person who got the drug got better sooner. This is a thing that happened and was observed and written down and reported in this paper. As you say, you canât make reliable predictions about whether other people would get better sooner by taking this drug. Itâs very very possible that the observed effect arose by chance.
I donât know man, I think we all have to assume you probably just donât have a high school level understanding of what statistical significance is.
1 Like
You can keep repeating the same false assertion all you want, itâs not going to make it true.
I donât see why weâre spending so much time arguing about ivermectin. I think we could all agree that (1) thereâs no robust evidence that it prevents or treats COVID-19 (at least no more than there ever was for hydroxychloroquine) and (2) that itâs almost exclusively being used by dumb-dumbs who refuse to take the steps that have been demonstrated to prevent COVID-19 (social distance, wear a mask, get vaccinated).
As far as the public narrative goes, itâs helpful to discourage people from taking ivermectin for a number of reasons, one of which is that there is some population of people who are persuadable.
7 Likes