Source: SSRN, By Michael Keane
ABSTRACT ( See the full paper as PDF here )
It is deliberately misleading to make general claims that hydroxychloroquine doesn’t work for CoViD-19. It is even more misleading to claim that “the evidence” proves that hydroxychloroquine doesn’t work.
To understand this, consider the use of parachutes. Consider someone who claimed that parachutes don’t work to stop death or serious injury when jumping out of a plane. What if they further claimed that “the evidence” proves that parachutes don’t work? Now consider if this same person advocated that because parachutes “don’t work”, we should ban their use and demanded that conscientious pilots who give a parachute to someone jumping out of a plane should lose their pilot’s license. Surely, such claims would be considered false and misleading.
Two classic parodies, published in the British Medical Journal, showed that the above statements about parachutes are technically true. In 2003, a straight-faced literature search found no high quality, randomized, placebo-controlled trials of parachutes. The fact that people usually get squashed to smithereens when they hit the ground without a parachute was just epidemiological data. There’s no proof the parachute makes any difference. It could be due to confounding factors or bias.
Subsequently, in 2018, a group of researchers actually did a randomized, placebo-controlled trial of the use of parachutes when jumping from a plane. There was no difference in death or serious injury in those who wore a parachute and those who didn’t. However, the participants jumped out of a plane 60cm off the ground while it was stationary.
So, it would, or course, be misleading to make the claim that the “evidence says that parachutes don’t work.” That is, the general claim that parachutes “don’t work” cannot be derived from studies where they are not used in the situation where they can provide benefit. Of course, everyone is going to survive whether you give them a parachute or not when jumping from 60 cm. It is egregiously misleading, and it would be lethal, to deprive someone of a parachute when jumping from 10,000 feet based on a study that tested a parachute at 60 cm.
Now let’s look at hydroxychloroquine. CoViD-19 is a disease with very different stages. And the vast majority of people, especially young people, are not going to die from CoViD-19.
The suggestion from a significant amount of epidemiological data is that, if given early in the course of the disease, HCQ might prevent progression to critical illness or death in at-risk population groups.
However, the randomized controlled trials (RCTs) performed to test the effect of hydroxychloroquine have been the equivalent of testing the effect of parachutes in the following circumstances: after a 60 cm jump; or pulling the rip-cord 2 feet above the ground after free-falling; or putting the parachute on someone after they hit the ground.
Different RCTs on hydroxychloroquine have had variable characteristics including the following: control groups with average ages in the 30s and 40s; a cohort with 99% of patients with mild to moderate disease; a cohort with death rates of 0.4%; a cohort with zero incidence of death or mechanical ventilation; patients already on ventilators or even ECMO; patients having symptoms up to 14 days before being commenced on hydroxychloroquine; patients having symptoms an average of 16 days before commencement of hydroxychloroquine.
Full Paper
Censoring medical communication has the potential to be dangerous. The process of censorship can be more detrimental than the alleged misleading information that is the focus of the censorship. The ongoing controversy regarding hydroxychloroquine (HCQ) and COVID-19 should prompt debate. Recently, Big-Tech has censored claims that hydroxychloroquine might benefit some coronavirus
patients.
Big-Tech justifies this under their policies of preventing false or misleading information related to COVID-19. Of course, deliberate and egregiously misleading information is harmful. But it cuts both ways. To be rational, it would also be necessary to label as misleading those obviously false or misleading claims on the other side of the HCQ ledger; for example, general statements that hydroxychloroquine doesn’t work for COVID-19.
To understand this, consider the use of parachutes. Consider someone who claimed that parachutes don’t work to stop death or serious injury when jumping out of a plane. What if they further claimed that “the evidence” proves that parachutes don’t work? Now consider if this same person advocated that because parachutes “don’t work”, we should ban their use and demanded that conscientious pilots who give a parachute to someone jumping out of a plane should lose their pilot’s license and have their life “cancelled”.
Surely, such claims would be considered false and misleading and, if anyone should be censored, this would be a prime case. Two classic parodies published in the British Medical Journal (BMJ) show that the above statements about there being no good evidence for parachutes and that parachutes don’t work are true. In 2003, Smith and Pell performed a straight-faced literature search and found no high quality, randomized, placebo-controlled trials of parachutes.[1] The fact that people usually get squashed to smithereens when they hit the ground without a parachute was just epidemiological data. There’s no proof it was the parachute that stopped death or serious injury. It could be due to confounding factors or bias.
These authors should win the Nobel Prize for their insight in publishing this satire. Subsequently, in 2018, a group of researchers actually did a randomized, placebo-controlled trial of the
use of parachutes when jumping from a plane.[2] And there was no difference in death or serious injury in those who wore a parachute and those who didn’t. The statistical analysis was pristine. However, the participants jumped out of a plane 60cm off the ground while it was stationary.
So, it would, or course, be misleading to make the claim that the “evidence says that parachutes don’t work.” That is, the general claim that parachutes “don’t work” cannot be derived from studies where
they are not used in the situation where they can provide benefit. Of course, everyone is going to survive whether you give them a parachute or not when jumping from 60 cm. It is egregiously misleading, and it would be lethal, to deprive someone of a parachute when jumping from 10,000 feet based on a study that tested a parachute at 60 cm.
Now let’s look at hydroxychloroquine. COVID-19 is a disease with very different stages. And the vast majority of people, especially young people, are not going to die from COVID-19. The suggestion from a significant amount of epidemiological data [3][4] is that, if given early in the course of the disease, HCQ might prevent progression to critical illness or death in at-risk population groups.
And this is very different to giving it late to people who are already very ill. And, furthermore, it is vital to appreciate the difference between mild disease and the early phase of disease. Even those who go on to die from COVID, start with lesser symptoms early in the disease before they get progressively sicker and sicker. But that is completely different to the vast majority who have mild disease throughout the illness and never get critically sick.
From Australia’s National COVID-19 Clinical Evidence Taskforce: “Evidence indicates that hydroxychloroquine is potentially harmful and no more effective than standard care in treating patients with COVID-19″.[5] They reference nine RCTs.[6-14] In particular, focus is given to five RCTs that have also been heavily cited in the media. The authors of these studies should be congratulated for conducting research during this hard-toimagine, overwhelming time.
But the studies didn’t necessarily address the question we need answered. In Annals of Internal Medicine,[12] a trial investigated patients early in the disease (which is important), but the patients were far too young. The average age of the control-patients was only 39. And “the incidence of hospitalization was only 3% and incidence of death only 0.4%”.
The authors concede, “Our population was relatively young with 77% of participants being aged 50 years or less, with few comorbid conditions; thus, our trial findings are most generalizable to such populations. It is possible that hydroxychloroquine is more effective in populations at higher risk for complications, such as older persons” This was the equivalent, therefore, of a 60cm parachute jump. So, this study cannot be cited as evidence that HCQ doesn’t work.
But despite that, there was still a 40% reduction in the composite end point of hospitalization and death in the HCQ group. And “at day 14 of the trial, 24% receiving hydroxychloroquine reported symptoms versus 30% receiving placebo (P = 0.21)”. Both these results were not statistically significant. But the study was simply not powered to investigate at-risk groups.
A study not powered to investigate the at-risk group can never provide evidence that an intervention “doesn’t work” in general. Furthermore, “Only 58% of participants received SARS-CoV-2 testing”. A similar study from Spain [11] showed the “risk of hospitalization was similar in the control arm (7.1%, 11/157) and the intervention arm (HCQ) (5.9%, 8/136; RR 0.75 [95% CI 0.32; 1.77])” However, “the mean age of patients was 41.6 years (SD 12.6)”. And importantly, “no patients required mechanical ventilation or died during the study period”.
A much-quoted trial in the New England Journal of Medicine,[14] is titled Hydroxychloroquine with or without Azithromycin in Mild-to-Moderate Covid-19; mild to moderate disease being the key. If, up to 14 days after symptoms had commenced, patients were not requiring any oxygen (58%) or a low amount, 4L/min or less (42%), that signifies a group that had selected itself out to have a less severe course regardless of treatment. And the average age was only 50.
Like jumping out of a stationary, grounded airplane, this group was mostly going to survive whether they had HCQ or whatever they were treated with. And giving HCQ up to 14 days post the onset of symptoms is like someone jumping from 10,000 feet and then pulling the ripcord and opening the parachute just 2 feet off the ground. That’s not when HCQ is thought to work, it’s too late. There were also some other issues affecting the reproducibility of this study. The RECOVERY Group study [10] from the UK, recruited large numbers of older, at-risk patients, which is important.
But RECOVERY studied hospitalized patients, and the average duration of symptoms at randomization, prior to receiving HCQ, was 9 days. An element that the study might be examining the wrong stage for HCQ is suggested by the fact that “written informed consent”, could be obtained, “from a legal representative if they (the patients) were too unwell or unable to provide consent”.
Indeed, before the study even started, many patients were already on a ventilator or even ECMO (artificial, out-of-body oxygenation of the blood) and for those who weren’t, the vast majority already needed additional oxygen before commencement of HCQ. Regarding the disease stages of COVID-19, HCQ is not considered to exert its benefit at these later stages. This is the equivalent of someone jumping out of a plane from 10,000 feet, and after they hit the ground, putting a parachute on their back and claiming that the evidence says parachutes don’t save lives.
In addition, RECOVERY used a uniquely high dose, which included 2400mg in the first 24 hours. Typically, studies have examined approximately 800mg in the first 24 hours. If the RECOVERY dosing was correct, that would make irrelevant all the other studies, and preclude them from being cited as evidence. A further study conducted earlier in the pandemic, demonstrated issues regarding disease severity, age groups and when, in the course of the disease, HCQ is given. The study title was “Hydroxychloroquine in
patients with mainly mild to moderate coronavirus disease 2019: open label, randomised controlled trial”. [8] From the study: “The mean age of the patients was 46 years.
The mean duration from onset of symptoms to randomisation was 16.6 (SD 10.5; range 3-41) days,” and “almost all (148; 99%) patients had mild to moderate covid-19″. In addition, there are more existential issues with the above studies with a number of provisos regarding interpretation. And this introduces a final element that must be discussed when discussing “the evidence”. Especially in rapidly changing, dynamic, prohibitively combinatorial systems, randomized controlled trials have consistently failed to have the fidelity to give reproducible results. See here [15] for an abridged discussion and here [16] for a more detailed discussion.
Experience pre-COVID has demonstrated that large RCTs have repeatedly failed to find beneficial treatment effects in perioperative medicine [17]; The NIH has now deprioritized large RCTs for sepsis research [18]; and there appears to be an over representation of probable false-negative results from large RCTs.[19] There are now a multitude of studies with accumulating epidemiological data [3][4] that HCQ might be beneficial, if given early to at-risk patients.
These studies are increasing in number and cannot be summarily dismissed. Like most academics, I genuinely want to know whether HCQ might have a role in reducing critical illness from COVID-19. I want to know the equivalent of whether, if you jump out of plane at 10,000 feet, it helps to have a parachute; not if you jump from 60cm, or if you pull the rip-cord 2 feet above the ground after free-falling, or if you put the parachute on someone after they hit the ground.
And like other innovations in human history, the effect of HCQ might depend on the system in which it is given and what other healthcare and treatment modalities are available. Trying to understand such a
complex system is somewhat reminiscent of the knowledge problem described by economists last century. [20] Based on the totality of current knowledge, the general claim that HCQ doesn’t work for COVID-19, or quoting the above studies without appropriate nuance, is misleading; as deliberately misleading as someone claiming that parachutes don’t work.
Related:
Belgium Study 8,075 patients: Low HCQ doses resulted in lower mortality in Covid patients
Dr James Todaro: An Effective Treatment for Coronavirus (COVID-19) – Previously Censored.