There's been plenty of ridiculousness in the past few years, but to me -- perhaps influenced by my profession -- one of the greatest absurdities has been the touting of useless and dangerous medications as a partisan political cause. It makes no evident sense in terms of political ideology. There's nothing "conservative" about screaming that a useless medication actually works, or attacking doctors and hospitals for not prescribing it. There was however an insane idiot who was the leader of the Republican cult and he started it. Apparently that's all it takes nowadays.
So let's step back and consider how we determine that medications are safe and effective, and what the standard ought to be for making that decision. We can start with Carl Zimmer in the NYT today. If you have a subscription or a free read left this month you can take a look, but I'll tell you what you need to know in case you don't. First, ivermectin is completely useless against Covid-19. That is the conclusion of a well-conducted, fairly large clinical trial in which 679 people infected with the virus received the chemical and a comparably sized group did not. There was no difference in the risk of hospitalization.
As Zimmer explains:
Early in the pandemic, when researchers were trying thousands of old drugs against Covid-19, laboratory experiments on cells suggested that ivermectin might block the coronavirus. At the time, skeptics pointed out that the experiments worked thanks to high concentrations of the drug — far beyond safe levels for people. Nevertheless, some doctors began prescribing ivermectin for Covid-19, despite a warning from the Food and Drug Administration that it was not approved for such use.
Around the world, researchers carried out small clinical trials to see if the drug treated the disease. In December 2020, Andrew Hill, a virologist at the University of Liverpool in England, reviewed the results of 23 trials and concluded that ivermectin appeared to significantly lower the risk of death from Covid-19.
But critics pointed out that many of the studies had a high risk of bias and one was likely fraudulent. When Dr. Hill did a new review that included only higher quality trials, he found no evidence of benefit. Now the large scale trial has confirmed that. This is a consistent pattern in drug discovery. Small scale, weakly designed studies often find evidence of benefit, and it often gets hyped in the news media before large, well-conducted studies have concluded; and they often have negative findings.
Why does this happen? There are a few reasons. One is that small scale studies may seem to show a benefit just by chance -- if you compare, say, 20 people who get the drug with 20 who don't, that can easily happen even if there is no real benefit. But the studies that do seem to see a benefit will be promoted and get public notice, while those that don't will quietly disappear. This problem is reinforced by cognitive biases. The people who conduct the studies want them to show benefit, as do the patients generally, and there are many ways this bias can subconsciously influence the results. I won't go into all of them here but I think we can all agree that wishful thinking is powerful.
That's why, when they are feasible and ethical, large scale, placebo controlled, double blind randomized trials are considered the gold standard for deciding whether a medication is effective. Double blind means that neither the investigators nor the patients know who is getting the real drug. Large scale means there are enough participants that differences will be unlikely to be due to chance. Randomized means the investigators don't get to choose who will be in the intervention arm.
There are some problems and limitations with this, to be sure. There may be "heterogeneity of treatment effect," such that the medication benefits some people but harms others, and the effects cancel out. But obviously, until and unless we can sort that out, it doesn't make sense to prescribe the drug.
Actually the problems are usually in the other direction -- there isn't long enough follow-up or sufficient monitoring to detect all the possible adverse side effects. And even in large-scale trials, apparent effects can actually be due to chance. Unfortunately in these cases, historically, the trials showing some effect were more likely to be published. There is also the problem of retrospective analysis, in which investigators look at end points they didn't originally hypothesize. Statistical theory tell us this invalidates any inferences. The FDA and NIH are trying to address this problem by requiring registration of trials in advance.
In any case, for an unproven drug to be a political cause is ridiculous. I hope it never happens again.