The Quick Screens
The next window is a capture of a website that promotes ivermectin for COVID. I have highlighted the author column yellow where I have done a more in depth review of the paper for either making it through the quick screening I’ve described below, or if I’ve found something important worth mentioning. I am only doing that highlight on the first instance of a citation from that website.
One of the easiest things to assess is the 95% confidence interval, which is a very useful quick glance, because if it includes the value of “1” in the range, then there is no statistical difference between the arms of the study (ie, ivermectin isn’t shown to be of any benefit). I’ve highlighted those in red. For consistency, I’ve done all of the highlighting based on quick screening in that risk ratio and confidence interval column.
Next, I looked for studies with relatively small sample sizes. I arbitrarily chose those with 100 or less participants. Given the number of providers using ivermectin and the number of people seeking to use it, this seems like a more than fair cutoff given how large the studies could have been. These are highlighted yellow and don’t warrant further analysis since they lack statistical power, which is a very complicated calculation and really not worth the time for something that clearly doesn’t work.
Another major red flag is if the author needed to pay to publish. Printed copies of pay to publish articles are really only useful as kindling or toilet paper. Those are marked green.
One surprise that came up was that some studies never were published that were used on this site. Those are also worthless as evidence, especially if it has been a long time since they were posted as preprints. Those are marked pink.
The ones that I have reviewed more carefully are marked in light blue. I would have left them without highlighting, but since many of these are duplicated throughout the website, it was an easier way for me to keep track if I have reviewed them or not as I go through them all. There are a couple of studies though that deserve special attention.
This is likely one of the studies that caused the noise about remdesivir. The authors make a very clear statement in their retraction.
Internal Flaws
The remaining studies are the ones that weren’t easy to quickly exclude. Some of the more legitimate authors point out their flaws, and I will use their words directly in quotes and made the text bold for important language. They are in alphabetical order by author.
This is another observational study and I am willing to assume that corticosteroid use is similar to that in India, as described in Behera below.
One thing particularly interesting about how this is listed on the website is that different data is provided in different sections of the site, particularly the confidence intervals. This suggests that the author of the website (not this study) hasn’t done due diligence in making these claims. Copy and paste isn’t that hard.
“The study analyzed a strategy implemented by IMSS to treat COVID-19, and it was not designed as a clinical trial. Therefore, it has the biases inherent to observational studies that decrease its validity. In addition, the two study groups were determined based on the delivery of the treatment kit, and delivery was based on the kit’s availability in the medical units and not on the patients’ clinical conditions. Because the intervention was not randomly assigned, we observed an imbalance in the distribution of age, previous medical conditions, and type of test performed between the comparison groups. To assure that the results were not modified by these variables, a multivariate model was carried out in order to adjust the RRs for hospitalization and death. Unfortunately, the intervention program did not include the registration of medications consumed by patients. Therefore, information regarding the specific use of each kit component was not available, and the data considering only the use of ivermectin could not be retrieved.”
One of the problems with studies on ivermectin in India is that corticosteroid use is rampant there, which does have benefit against COVID. “In India, knowledge and awareness regarding prescription drugs is lacking. Of these, corticosteroids are one of the most widely used drugs and also one with an array of side effects.“
If an ivermectin study out of India doesn’t control for this confounding variable, it does not provide evidence of efficacy. This study did not control for it.
Even though this is a pay to publish journal, it is regarded as reputable. There are two major flaws I quickly identified.
First, this is an observational study that collected data from surveys of clinical staff. Those types of studies do not have much weight of evidence, in that they don’t often account for confounding variables.
One of the questions I initially had was whether it was controlled for if staff worked on a COVID unit or not, so I looked through the survey to see if that question was asked. It was not. More concerning though was the lack of an analysis of a question that they did ask, “Any history of contact with positive case?” This could have at least provided a surrogate measure for my question. It’s telling though that no data was provided related to this question. I suspect it’s because it would have shown the real significant variable to becoming a COVID case. If that is true, this paper is dishonest, and the authors may have published it to try to get on the ivermectin efficacy bandwagon.
This is another observational study, if it can even be called that. It never even made it to a preprint server. It was a poster presentation at a conference, which has no peer review whatsoever. It’s also another example of where the website has inconsistent data summaries with the same study.
This was nothing more than a letter to the editor and not peer reviewed.
This compared ivermectin to HCQ. Not worth anything to assess the effectiveness of either.
The authors don’t even have a limitation section, which is a red flag. One obvious one is that this study compares countries that use ivermectin prophylaxis to those that don’t in Africa. It uses case counts as the outcome measure, which is flawed because of the wide range of ascertainment bias in these countries.
It is also worth noting that in these countries, ivermectin is used AS INTENDED for parasitic infections. As such, people who may have started taking ivermectin as prophylaxis may have felt better because it was taking care of a parasitic infection, so they may have been less likely to seek medical care (and hence, COVID testing) because they began to feel better than their baseline.
This one is outright garbage. I’ll let the published correction speak for itself from a screen capture. It’s outright fraudulent and unethical, to the degree of Andrew Wakefield and autism.
This was another observational retrospective cohort study out of the Dominican Republic among healthcare workers that self selected to be in the ivermectin group, hence, it doesn’t account for confounding variables, as the authors admit. “Since this is not a randomized study and the selection of groups was based on adherence, it does not allow us to clear certain confounding factors, such as the fact that the ivermectin group could be made up of healthcare personnel more concerned with prevention in general, including greater personal protection measures and more careful use of the PPE, reducing the risks of contracting SARS-CoV-2.” In addition, “There was no RT-PCR test at the start or exit of this study, neither in the ivermectin group nor in the control group. There is the possibility that asymptomatic cases were not detected.”
This is another observational study out of India that doesn’t account for confounding variables. In addition, the surveys of the healthcare workers were not anonymous, which makes them very prone to social desirability/conformity bias.
It’s interesting that they listed this one on the website as proof of efficacy. It specifically states “Among outpatients with mild to moderate COVID-19, treatment with ivermectin, compared with placebo, did not significantly improve time to recovery. These findings do not support the use of ivermectin in patients with mild to moderate COVID-19.”
Another one that was retracted.
This was a retrospective observational study where corticosteroid use was a major confounding variable. At the time, corticosteroids were reserved for the sickest patients since there had not been data published as to the benefit yet. They admitted “Although more of the control group was enrolled in the first weeks of the study, suggesting the possibility of timing bias, this may be offset by preferential treatment of more severe patients with ivermectin early in the study because of low initial availability.” This suggests that corticosteroids were skewing their results.
This “study” is amazingly bad. It is nothing more than a survey of physicians who treated patients with ivermectin or hydroxychloroquine and what they thought of their results.
Conclusion
One of the most telling things about this and mirror websites is that there is no information related to who is behind them. That is not how science works. I’d be willing to bet that they are being driven by those who are trying to profit off of the use of ivermectin.
To their credit, or more likely perhaps because of how bad they knew it would make them look, they took down a study by Elgazzar, which Jack Lawrence discovered was clearly fraudulent. He makes a nice summary of these sites. It is well worth the read.
