For over 12 months I’ve been waiting for the outcome of the Coronavit trial, a large trial into the use of vitamin D supplements to help prevent Covid-19. Last month, to little fanfare, the trial results were published in a scientific paper (albeit a pre-print for now).
And the paper is very disappointing — their analysis found no benefit of vitamin D. But that’s not what was disappointing…
…the disappointing aspect was that their analysis had many small but important flaws, and one absolutely huge flaw that puts the whole of their conclusion into doubt. These are the main results they found, covid cases by supplement status:
On the face of it that seems conclusive enough — there’s no difference between the two vitamin D supplementation groups and the no vitamin D group.
The flaw is in the ‘out of’ part of the table above. The authors undertook a study that required their participants to fill in a web-survey once every week, but for the denominator they used the total number of participants enrolled in the trial. With studies such as this you don’t use the entirety of those that said they’d fill in the form, but the actual number of forms filled in. The authors don’t report on the response rate anywhere in their paper, let along use it for the analysis.
It is difficult to explain how fundamental a mistake this is — it isn’t a technical mistake that ‘anyone could make’, but rather the sort of mistake made by an undergraduate that didn’t listen properly during their lectures. As an example of how ludicrous this mistake is, imagine a survey where 100 random people were asked on the street if they were male or female, and the results came back that 25 of those asked responded ‘male’ and 25 responded ‘female’ — the conclusion should be that 50 didn’t respond to the question, not that 50% of the population are gender non-binary.
But perhaps the supplement and no-supplement groups had similar proportions filling in the forms? Here we come to an important aspect of the trial — those that took vitamin D supplements were ‘engaged’ in the trial; they were given something for free (the vitamin D capsules) and had to do something every day ‘for the good of science’, however, those in the no-supplement arm were given nothing for their trouble and were in what might be called the ‘you don’t get the special treatment’ arm of the trial. In such situations it is to be expected that there’ll be a lower participant engagement in the no-supplement arm. Indeed, this is one of the reasons for using placebo tablets in such situations — no so much for ‘the placebo effect’, but simply to keep engagement at similar levels in the ‘no supplement’ arm of the study.
We can gain an indication of the level of the engagement that participants had in the study, as right at the end they offered a free ‘vitamin D’ test to a subset of the participants. Approximately 93% of those in the groups that took the vitamin D supplement responded to this offer, but only 75% of those that didn’t take a supplement responded. Note that this is almost certainly an overestimate of the engagement of the no-supplement group in the trial, as this was the response to ‘an offer of a free thing’, whereas there was no benefit whatsoever to the no-supplement group in actually filling in the weekly web-survey.
Just about all of the results of the trial are affected by this flaw. The exception is the mortality data which was obtained via medical records (as the dead can’t fill in a web-form) — however this is a moot point as no-one died from (with) Covid during the trial.
Finally, another aspect of the trial that is suspicious is that according to official data from the UK government, during the study period approximately 5% of the population was infected with Covid-19 yet only approximately 3% of the study population were infected. Might it be the case that around 5% of the no-supplement group actually caught Covid during the trial? We just won’t know until the authors report on the experiment properly.
I very much hope that the paper’s peer-reviewers pick up on this mistake (frankly, I just can’t see how they couldn’t).
I think we know this study is bullshit. While it is not as simple as just taking a supplement and all will be fine, anyone that doubts that raised levels Vit D is beneficial for your immune system is a moron. As Uncel states, all this system is corrupt. Give the human body what it needs, what it needs is the conditions in which we naturally evolved in....ie.sunshine. We have progressed indoors quicker than we can evolve to cope with the withdrawal from sunlight and are deficient in Vid D because of this. It's just plainly fucking obvious.
I have these questions and comments:
Over what duration was the trail, as my take is the incidence of test positivity is remarkably low for ALL - I mean what are we all worried about?. Also ‘cases’ means very little - how many were ill with actual symptoms and the main question surely is : “ did vit D3 reduce disease SEVERITY “