I went through the source that you glowingly review as being such a wonderful trial and there are some problems I found that cast doubt on the value of the results.
https://dcricollab.dcri.duke.edu/sites/NIHKR/KR/GR-Slides-08-06-21.pdf
Several things about this slide deck stand out.
"Of the 2,908 trials captured in our registry, over
half (51%) intend to recruit 100 patients or less. The median sample size across all trials is
100"
The problem with this is you can't take a bunch of statistically insignificant results for a bunch of small sample sizes, and then bundle them together and say that shows statistical insignificance for a large population. 3000 trials with n= 100 are less valuable than 1 trial of n= 30 000. Especially given the criteria for selecting a sample may be different in each trial.
Trial design was quoted as being "Shared control patients" (across multiple medications being studied) and "No specific sample sizes".
If I'm understanding this correctly it means that the n for patients being given a placebo, was shared across the entire study which tested the effectiveness of 4 different drugs? So that can warp the results.
Inclusion criteria was listed as including several deadly conditions that could affect likelihood of hospitalisation or ER visits on their own, whether or not a placebo was given (such as stage 4 kidney failure and cancer) plus several conditions that are less likely to lead to ER hospitalisation on their own (being a smoker, being obese). And there's nothing showing that they ensured the proportion of people with more serious illness was equal between the control group and the group being given the medication. The randomisation was said to be stratified: "To account for other arms in the trial, Clinical site, Age (≥50 years vs <50 years)". So randomisation was NOT stratified according to severity or number of "high risk criteria" the patient had.
In the ivermectin trial specifically it said the people chosen were "Patients with COVID-19 and expected
hospital stays of ≤ 5 days". Implying they already were in hospital.
But it was trying to measure how much ivermectin affected "A composite of emergency room visits due to clinical worsening of COVID-19 (requiring
observation for > 6 hours) or hospitalization
due to the progression of COVID-19 within 28
days of randomization."
So a study on already hospitalised patients, was trying to measure the effect of ivermectin on rate of hospitalisation? The pre-screening occurred IN hospitals, primary care facilities and emergency rooms. Meaning that all of the people chosen for the study, were already hospitalised at some point.
The screening and randomisation occurred "up to seven days from symptoms start" - so if we assume that the people chosen were in hospital at the start of screening but were "expected to be there less than 5 days", and randomisation occurred within 7 days of them being selected (so some of them were still within their 5 day expected visit when randomisation started), and what's being measured is hospitalisation after randomisation, then the numbers counted as "hospitalisation" could also reflect patients who went home within their expected time frame of being in hospital.
Another shady aspect of the study is this:
"Hospitalization due to the progression of COVID-19 (defined as worsening of viral pneumonia) and/or complications within 28 days of randomization. "
Except these "complications" aren't stated to be complications specifically of the medication or complications of COVID, but could also mean "complications of cancer" or "complications of stage 4 kidney disease" etc. So that messes the data as well.
Finally, the people behind the study include:
MMS Holdings - a company whose mission is to help pharmaceutical companies get approval and to design scientific studies that help them get approval. The mission of this company is not to dispassionately test drugs to see if they work, but to design studies and produce results that get drugs approved. One of their clients is Pfizer. https://www.mmsholdings.com/pfizer-selects-mms-as-preferred-provider-for-plain-language-summary-writing-support/
Cytel Inc:
Another statistical modelling company designed to help pharmaceutical companies get approval - they work very closely with Pfizer:
https://www.cytel.com/blog/an-interview-with-pfizers-chris-conklinhttps://www.cytel.com/east-symposium-2019-slideshttps://www.cytel.com/blog/bayesian-methods-vaccines-research-covid-19
One of the principle investigators for the study as shown on the map is Dr Craig Rayner, President of Integrated Drug Development at Certara - another company with the same mission as MMS Holdings. They proudly proclaim on their website that: "Since 2014, our customers have received over 90% of new drug and biologic approvals by the FDA.
One of their clients is Pfizer:
While the presence of these companies helping to fund and direct the study, one must question how accurate and unbiased the study is, based on the potential conflicts of interest of those involved.
