Ivermectin and statistics

Having been an economist, strategist, risk manager and investment manager, I have come across a lot of statements and propositions that have been 'supported' by statistical evidence. From experience, I know that you can almost prove anything by being selective enough in what statistics you use. After seeing quite a bit of statistics, one obtains some intuition when statistics are being used to send a message. My intuition for the use of statistics started to rattle when I read CNA's article on the use of Ivermectin in a Malaysian study. Following that intuitive hunch, the research effort that followed revealed some insights that I'd like to share.

COVID is a high stakes event and it may be expected that we will see statistics that on the one hand support certain conclusions and on the other hand debunk the same conclusions.

Ivermectin and who's interested

Within our struggles to deal with COVID, the use of Ivermectin is a case in point with certain studies being pro-use and others claiming that the use of Ivermectin doesn't provide any benefits. We have to dive deeper into the matter to get a clearer image on how statistics are used.

First, it is important to recognise the stakeholder interests around the use of Ivermectin. Ivermectin has been around for decades and is a generic medicine that is low-cost to produce and obtain, with low profit margins. In general, its application hasn't resulted in any significant amount of serious adverse drug reactions.

Should Ivermectine be an effective medicine in the battle against COVID, it would jeopardise the earnings models of new experimental medicines (like the jabs against COVID) that have recently been on offer and which provide big pharma with hundreds of billions in profits.

Ivermectin in Uttar Pradesh

Perhaps the most advocating case for the use of Ivermectin is the experience in Uttar Pradesh, India's most populous state, with over 200 million inhabitants and home to the Taj Mahal. Today, the state recorded only 17 new cases, in the context of a cumulative total of over 1.7 million cases since registration began (link to cases in Uttar Pradesh).

Uttar Pradesh's success story in fighting COVID is frequently attributed to the widespread use of Ivermectin. In August 2020, the Uttar Pradesh government gave the go-ahead to include Ivermectin as prophylaxis for close contacts of COVID-19 patients, health care workers, and the treatment of the patients.

The Merriam Webster dictionary defines prophylaxis as “measures designed to preserve health (as of an individual or of society) and prevent the spread of disease”. The stress words here are the words “preserve” and “prevent”.

Ivermectin studies in Malaysia and Argentina

The quoted Malaysian research studied the administration of Ivermectin in the first week on patients experiencing symptoms of COVID-19 and concluded that "Ivermectin does not reduce severe illness from COVID-19".

Of course, the administration of Ivermectin was NOT preventive in the Malaysian study, whereas it was in the case of Uttar Pradesh. We don't know what happened, but the quite positive results in Uttar Pradesh could well be attributed to wearing a seat belt before the crash whereas in Malaysia the seat belt was worn after the crash. We don't know for sure, but we do know for sure that the Malaysian study cannot be conclusively compared with the experience in Uttar Pradesh and therefore cannot be conclusive on the use of Ivermectin.

In addition, CNA's article refers to the Argentine study of IVERCOR being " in line" with the Malaysian study. So when I looked into the IVERCOR study, I was surprised to learn that conclusions were drawn that the use of Ivermectin prevented hospitalisation!

One passage in a related study, particularly caught my eye: “In an in vitro study, ivermectin was found to be an inhibitor of the SARS-CoV-2, with a single addition to Vero-hSLAM cells 2 h post infection with SARS-CoV-2 able to effect ~5000-fold reduction in viral RNA at 48 h.”

A 5,000 fold reduction in viral RNA, 48 hours after infection appears quite substantial. Of course, with COVID's long incubation time, when symptoms occur, chances are that one is most likely very well past those 48 hours after infection. It could well be that the administration of Ivermectin may be far less impactful when applied (much) late(r).

Concluding remarks and a call for action

Let's be clear on this. I am not a doctor and I don't have a medical background. The above certainly doesn't constitute medical advice and most certainly shouldn't be interpreted as an advice to buy medicines online with parties you don't know or don't trust.

What I do know is that statistics can be used to tell almost any story.

The above shouldn't be interpreted as critics against CNA's reporting or Singaporean or Malaysian policies from a foreigner that would wish to create a stir (which would be far besides the truth).

I would be the first to admit that the situation in my home country, The Netherlands, probably requires more attention. After decades of successfully administering Ivermectin, recently, GPs in the Netherlands can now be fined 150 thousand euros for prescribing Ivermectin to patients. The rationale for this sudden policy move still hasn't reached me and I guess MPs are still to dive deeper into this matter.

What I do want to say is that statistics can be used and interpreting them is a tricky business. This small post isn't even close from being scientifically conclusive.

The time has come to investigate the large scale use of Ivermectin as prophylaxis in the prevention of hospitalisatioon and the spreading of COVID, with truly independent funding, truly independent researchers and truly independent statisticians, with an ability to speak out when external pressure is placed on possible outcomes.

The stakes have become too high for any rigging of results. Let's keep each other sharp. In that context, I very much welcome critical comments on this post as well.