The debate about using hydroxychloroquine to treat or prevent COVID has become the talk of the town. It can become so emotional that it pits neighbor against neighbor and brother against brother. Everyone seems to have an opinion.
Many of these opinions are based on poor studies and/or on a he said/she said approach. Why are many so gullible along these lines? In part, because folk healing has been going on for thousands of years. The other day some friends told me that their son has COVID. With conviction, his mother said that she told him to drink plenty of hot tea with lemon. That is good for COVID, she said. Then she realized who she was talking to and added, “At least, that is what I think is good.”
I didn’t correct her. I just said, “Could be.” I try not to correct along these lines when the interventions do no harm.
But hydroxychloroquine is a different matter. It can do harm. And, in my opinion, its greatest harm is to bolster distorted thinking and pseudoscience. I wish we could discuss the flaws in all the studies which claim that hydroxychloroquine is beneficial, but that would be too much information. So instead, I will use one study as an example. I would also like to explain some aspects of medical knowledge that may help folks understand this issue better.
I will start with an explanation of how we decide whether a medication works. One way is by observing what happens when a person is treated with a specific medication. Let’s consider hydroxychloroquine. Suppose every time we gave it patients said they felt much better. Suppose every time we gave it no one died. If such dramatic results happened with hydroxychloroquine then there would be no debate.
However, that is not what happened. Instead, quite a few who got hydroxychloroquine died. When people took it, they did not suddenly get better. When a medicine does not obviously work, then we turn to a statistical analysis to determine whether it is beneficial. This is where it gets tricky because statistics involve a lot of complicated mathematics. Also, for statistical analysis to be valid, variables need to be controlled and groups need to be comparable. Being comparable is quite important.
For example, if doctors divided patients into two groups and one group was over 65 and the other under 65, and they gave hydroxychloroquine to the under-65 group, then that would be a biased, invalid study. As a corollary, one way to manipulate that data is to not have comparable groups.
So one place to start in evaluating a medical study is to ask whether the study groups are comparable. The next place to go is the endpoints. One endpoint that has been used for some COVID trials is length of illness. That endpoint, in my opinion, is not of much value. A more solid endpoint is death. So what we really want to know about hydroxychloroquine is does it prevent death.
So suppose we set up a study of hydroxychloroquine with death as an endpoint. Once we get numbers of deaths we have to decide whether the differences between those who got HCQ and those who didn’t were significant. We decide this by calculating what is called a P value. If the P value is less than 0.05, then the difference is said to be significant. The mathematics behind the P value is too much to go into here. Suffice it to say that though those mathematics are solid, there is a misleading component to the P value. I will try to make this clearer below.
For example, there have been many studies of statins. Most of those studies have about 2,000 patients in each group – the treated group and the placebo group. Most of those studies show about a 1% difference between the two groups – i.e., in the group that got the statin medication, 2% had heart attacks; and in the group that got placebo, 3% had heart attacks. This 1% difference yields a P value of less than .05.
In my opinion, this 1% difference is not much. But what happened is once that P value was obtained, advertising and propaganda took over such that many believe that statins lower their individual risk of heart attacks and death by 30% or more. That belief is patently false. In other words, the P value of less than 0.05 allowed for exaggerated claims about the benefits of statins. In that sense, the P value is a misleading component to it.
This game of finding P values of less than 0.05 and then making exaggerated claims became part of the culture of academic medicine (i.e., medical schools and biomedical research organizations). Therefore, when hydroxychloroquine came along, it was easy to find those in academic medicine who would play the same game with hydroxychloroquine. They do this for attention, promotion, and glory.
For example, one study published in the Journal of Medical Virology is entitled “Chronic Treatment with Hydroxychloroquine and SARS-COV-2 Infection.” This study found 26,815 patients with COVID; 77 (0.29%) of those 26,815 were treated with hydroxychloroquine. They found 333,489 COVID-negative patients. Of those 333,489, 1,215 (0.36%) were receiving hydroxychloroquine. The researchers then looked for a P value with these numbers. They said the difference of 0.29% vs 0.36% gave a P value of less than 0.05. This led them to conclude that hydroxychloroquine protects from COVID.
This study, and studies of similar quality, have been cited by doctors who claim that hydroxychloroquine is effective for treating and preventing COVID. Just look on medical blogs and you will see it cited. (And please, those who attack me, don’t waste your time writing with links to such garbage.)
But anyone with a lick of sense should be able to see that this study in the Journal of Medical Virology in no way proves that hydroxychloroquine prevents COVID. The “statistically significant” difference between 0.29% and 0.36% is incredibly trivial; it is 0.07%. Also, “statistical significance” is easily manipulated by sample size. It is well known that the larger the sample size, the easier it is to make any difference show a P value of less than 0.05.
One reason such a trivial difference (0.07%) showed a P value of less than 0.05 is because their study consisted of roughly 360,000 people. Years ago some doctors proved this point by reanalyzing the results of a study based on astrological signs. They could get a P value of less than 0.05 for Aries versus Libra, because the sample size was so large.
Games, games, and more games. To a certain extent, this whole HCQ debate has little to do with medicine and science and a whole lot to do with human nature.
These games and statistical manipulations are not uncommon in the medical literature. This is a serious ethical problem, and over the years I discussed this problem with medical ethicists. None of those I spoke with wanted to touch it. They said it was out of their area of expertise. This is a problem. When the leadership in academic medicine not only allows this stuff to happen but promotes it, then quagmires, such as this HCQ quagmire, will arise.
In a sense, the chickens are coming home to roost. The sins of academic medicine are coming back to bite them. I think Dr. Fauci has felt the teeth.
These sins of academic medicine also give ammunition to those who want to claim they are right about HCQ, even when they lack evidence. That is, they say the New England Journal of Medicine and The Lancet are guilty of this, and Dr. Fauci is guilty of that, and therefore they are right. They don’t seem to understand that two wrongs don’t make a right. Using the mistakes of others to claim that your mistakes are right is a mistake.
It is clear from the data and from experience that hydroxychloroquine is not a miracle medication for COVID. If it has any benefit, it will be of the trivial variety. What makes people believe that a trivial benefit is a miracle medication is propaganda. This medical propaganda has become an integral part of American culture.
Just look at the supplement industry. It relies heavily on such propaganda. Recently, brain supplements are popping up left and right. At the end of each commercial about these things they say, “this product has never been shown to treat or prevent any illness.” But that is how stupid they think the public is – i.e., they can tell them that the stuff has never been shown to work but that they should take it because an actor tells them that it does. Or, they hire an “expert” to come up with some biochemical scheme that will sell the product.
This approach has also been seen in the HCQ debate. The zinc/hydroxychloroquine interaction is one such example. People are told that hydroxychloroquine is a zinc ionophore and that zinc inhibits COVID, and therefore it works. However, it would require a good clinical study to prove that hypothesis true. To date, there is no such study.
It should be kept in mind that many fancy schemes when tested with good clinical studies turn out not to be true. Also, it should be kept in mind that the history of medicine is filled with examples of doctors coming up with schemes that they claimed were true but were false. Many times those schemes hurt patients. Bloodletting in one such example. Doctors swore up and down that to cut a person and let the blood drain into a bucket was therapeutic.
The hydroxychloroquine debate has demonstrated that many are vulnerable to medical hustles. I would like to think that if educated correctly, humans have enough sense to avoid or overcome those hustles. When we are sick, we all want to get better. We all want medications that make us better. But we should not let those desires allow us to be exploited or led astray.
W. Robert Graham, MD, completed medical school and residency at UTHSC-Dallas (Parkland Hospital) and served as chief resident. Graham received a National Institutes of Health fellowship at the Salk Institute for oncogene research in 1985. He was a professor of medicine at Baylor College of Medicine from 1998 through 2016. In retirement, he enjoys writing and ranching.
Last Updated August 10, 2020