The lady in the white lab coat smiled.
This was her favourite time of the year. The new batch of freshmen, their enthusiasm and chatter... and the opportunity to run some of her favourite experiments again. The subjects today were seated in a circle, happily conversing in the manner that only adolescents could, and she liked to imagine that some lasting acquaintances would be made this day.
There was a little twist, of course. Other than one of the subjects - the "mark", if you would - the others were actually her assistants. They were all in on the jig.
The session would begin proper, with a fun introduction into some optical illusions, about which the mark appeared largely unaware of before, and now suitably impressed. Five minutes later, the questions started. A few items were worked through, and the mark would be able to provide the right answer, whenever his turn came. He had been informed, after all.
And then, the whole point of this experiment. The head investigator slowly held up a plain card, with just two lines on it, one red and one blue:
We're not doing it with pills today
(Original Source: flickr.com)
"Which of these lines is longer?" she asked, unable to avoid a faint grin. She lived for these moments.
It was her senior assistant's cue to start, and she suspected that Nelia loved her part in it. "The blue line."
The mark appeared slightly taken aback - but not too much - but then the other four conspirators would affirm with the same answer. "Blue."; "It's blue.", "Blue, of course.", "Blue."
It was now the mark's turn. He dithered.
"May I have the card?" he asked.
"Sure." she smiled again, as she handed it over. Most didn't even get to this stage. She watched as the poor guy brought the card up to his face, then lowered it, before a brief, haunted glance around at his peers. She savoured the angst, and wondered about exactly what was going through his mind.
"Blue." he finally said.
"Brilliant!" her smile widened.
But it was not over yet. The Prestige was yet to come. To fulfil their module requirements for the semester, the students had to take part in two such sessions, and the mark - who was actually named Mark - would dutifully return next week. The general setup was all but identical, but the questions had become a mix of basic factual and mathematical items. Oh, and the Line Test. And another teeny detail - all the subjects this second time around were the various "marks" from last week.
After everyone had agreed that 31 plus 27 was 58, she would hold the same card, with the same one red line and one blue line, up once more. "Which of these lines is longer?" she repeated. Mark was due to respond first, today.
"Blue, obviously!" - without any hint of hesitation.
My clever readers will, of course, recognize the above as a form of the Asch conformity experiment, indeed retaining the classic length-of-line task, with a dash of the Milgram obedience study (white lab coat). The follow-up is based on what seems like an apocryphal telling on training for persistent behaviour with monkeys and bananas - but I would not be overly surprised if a real-life trial turned out as described above. Note that the original Asch experiment had the lines being very clearly longer or shorter, which still reliably had over a third of the subjects swayed; one imagines the additional doubt were the lines rather closer in length - which may be instrumental, as we shall see.
Now, I might be a bit of a troll at times, but I am also something of a fanboy of science (okay, including psychology), technology and raw human endeavour at heart; I quite like immersing myself in what us hairless apes have managed to achieve. We have landed enough rovers on Mars to hold a Battlebots match, so it seems, but even over four millenia ago, a bunch of our desert dwellers managed to organize to haul 2.3 million two-tonne stone blocks, into a huge pyramid. Just last year, we have analyzed stardust particles billions of years old under an electron microscope, identified tyrannosaur embryos as being between some 71 and 75 million years old while possibly extracting DNA from similarly-dated fossils. We have closely examined the surface of Betelgeuse - a star a mere 600-plus light years distant. We have figured out how to assemble transistors a single nanometre apart. We are well on the path of editing the human genome at will, and achieving electrical superconductivity at room temparature. I often get the warm fuzzies while flipping through my Nature subscription - what can humanity not achieve, if we set our efforts to it? Consumer space travel? Invisibility cloaks? Universal translators? Teleportation? Sometimes, it all feels within grasp.
Well, very sadly, there appears one specific task that humans are entirely incapable of, for some reason: running proper early treatment trials using repurposed widely-available and commonly-used medicines, in the midst of a pandemic. It's just impossible! There is no possible way, it seems, to provide a few thousand volunteers with off-the-shelf candidate remedies at dosages long accepted as safe, maintain a control group, and record their outcomes on disease severity as measured by some quantitative test. Can't be done!
Lifting The Veil
The following exposition will unavoidably retread some material already raised in previous blog posts, but which is probably necessary. The latest big development in America, from a couple of weeks back, has been that fully-vaccinated people can go unmasked in most places - curiously right on the heels of the WHO finally acknowledging that the coronavirus is likely airborne, ya know (year late, millions short)? The CDC is insistent that their newest about-turn on the mask mandate is based solely on science, though there has understandably been concern from many quarters - a Nature news feature, for example, unequivocally asserts that "the evidence is clear that masks cut down on COVID-19 deaths". On this, it might be helpful to review the stand being consistently adopted on this blog, against oft fast-changing official recommendations.
Should the vaccinated be required to mask up?
Back last February, the basic logic on masking had been laid out here as such: "All considered, it's tough to understand how wearing a mask would harm an individual, despite all the insistence that it's not necessary and that it could be defeated by dirty hands; of course, from society's point of view, not everybody needs a mask as much as others, and rushing to secure personal supplies could cause unwelcome panic, fair enough". I think this still holds. From the WHO's sad flip-flops, I gather it is fair to say that their esteemed experts were always mostly uncertain, about how the then-mystery virus was being transmitted. As such, my reasoning was quite simple: many similar contagious respiratory illnesses are effectively airborne; there is fairly little firm evidence either way; the novel coronavirus seems very bad news and possibly fatal; the negatives to wearing a face covering are relatively low; therefore, it seems reasonable to wear a mask, while the virus gets figured out; it shouldn't hurt, and it might help.
Please note that the above is not meant as an expression of i-am-very-smart; if it came off that way, I do apologize. But, honestly, it's... it's not exactly big brain stuff?
Well, for whatever reasons (many relating to the consequences of offshoring almost all of one's manufacturing capacity for critical medical supplies), scientific mask guidance took some time to turn around, with the likes of TIME equating mask wearing with superstition, and the director of the NIAID outright knowingly lying, while the world waited for high-quality evidence to come in. Along the way, an article was published in the Annals of Internal Medicine on how masks were ineffective - and probably referenced in support of that stand - before being retracted in July. Obviously, given that it involved only four subjects, it is fair to wonder why all that much weight was ever placed on it. Anyway, researchers in Denmark would finally complete a large, gold-standard randomized controlled trial (RCT) on surgical mask usage with some 6,000 subjects, just when I was beginning to despair that humanity had lost the ability to conduct proper trials. Finally also published in the Annals come November despite being completed by June, the RCT found that... wearing surgical masks didn't help that much.
Now, before the angry recriminations arrive, it might be clarified as to what exactly the DANMASK-19 RCT found. From Table 2, for the primary outcome involving "SARS-CoV-2 infection, defined as a positive result on an oropharyngeal/nasal swab test for SARS-CoV-2, development of a positive SARS-CoV-2 antibody test result (IgM or IgG) during the study period, or a hospital-based diagnosis of SARS-CoV-2 infection or COVID-19", the odds ratio for the face mask group as compared to the control group was 0.82 (0.54-1.23).
There are two parts to the presented result on the primary outcome, as illustrated. The red star shows the odds ratio (OR) for the face mask group as compared to the control group, and it being less than 1.00 suggests a positive effect, i.e. mask wearing reduces one's chance of testing/being diagnosed for the coronavirus. This might be approximated as a reduction of about 18% in relative risk, which is like, not bad? However, the standard 95% confidence interval (C.I.) for the OR ranges from 0.54 to 1.23, which indicates that the improvement is not that certain; there's ongoing debate on whether interpreting results with a 95% C.I. and its corresponding P-value is good practice, but by the "rules of the game", one would have to declare that face masks do not help to prevent coronavirus infection, i.e. the effect was not statistically significant at the 95% C.I.
To this, an executive editor at The BMJ would pen "The curious case of the Danish mask study" as an editorial, in which he makes the case that the results should more properly be described as inconclusive rather than negative, and that there is a likely benefit of mask wearing to the wearer, even before considering potential reduced spread to others. Rebuttals notwithstanding, I regard the editor's viewpoint to be mostly accurate. The effect (red line) is indeed positive, while the uncertainty (blue line) is unfortunately large. The remedy, then, is to pursue more and larger trials, since that would reduce uncertainty (i.e. shrink the blue line). In the meantime, an estimated 18% improvement isn't to be sniffed at, and indeed well within my personal expectations that masks would help somewhat, but hardly solve the pandemic by its own. 18% here, 30% there, 60% inactivated vaccine, 15% concentrated power of will... it all adds up, man. Right?
As the BMJ editorial would also have to admit, the DANMASK-19 RCT would distressingly have trouble finding a major journal to publish it, thus the delay getting out, due to it running counter to the latest received wisdom. In my opinion, this is no way to conduct science - if the methodology is sound, the findings should be published and enter the literature, and the curation come later. In any case, perspectives in other renowned journals such as PNAS have (rightly) quoted Cochrane and the WHO in stating that "for population health measures, we should not generally expect to be able to find controlled trials, due to logistical and ethical reasons, and should therefore instead seek a wider evidence base", with them releasing a German synthetic control method trial claiming a 47% reduction in daily growth rate of infections from masking about the same period, and Science seeing fit to publish an entirely theoretical simulation on mask effectiveness just this month.
To the simulated study, some commentators would (not wrongly) point out that there wasn't even an attempt to validate the model in practice, and to be frank I'm slightly wary of pure simulations, from my reviewing duties. The mathematical framework is often impeccable, but then one often comes across details such as "wait, this simulation seems to assume that a patient in Los Angeles can freely employ available hospital resources in Chicago?". Suggestions to backtest the simulations on available data were somehow usually rejected, but this is not a knock on the researchers - accurate simulation is hard, and has destroyed the reputations of more than a few well-regarded groups. Many of these forecasts are probably on the level of predicting the stock market, and we all know how tough that is.
Double Standards, No Standards
Returning to what could have been. I have explained my support for face masks - even when they were discouraged via official "authority" - and believe my reasoning largely vindicated. I have undertaken vaccination due to its reported efficacy of 90+% (original coronavirus) to maybe 70+% (for variants), despite valid concerns on unknown long-term effects, because science fanboy, you see. These, note, happen to be prophylactic measures - masking hopefully aids the self and others from getting infected in the first place, while I suppose the vaccines aren't meant to be applied after infection is diagnosed (but if I am mistaken, please do educate me). This, however, leaves a glaring gap in the care pathway: what can and should medicos do, for the already-infected?
We've heard all about ventilators and Remdesivir and Dexamethasone for application to the desperately-sick, but standard of care for coronavirus-infected patients not yet manifesting severe symptoms appears to still be "sicken-in-place". The NIH's latest treatment guidelines for non-hospitalized (i.e. outpatient?) cases with mild or moderate COVID-19 seems to allow Bamlanivimab+Etesevimab or Casirivimab+Imdevimab (a.k.a REGEN-COV, as apparently deployed on POTUS) if a high risk of progression is expected, but the FDA appears to have revoked Emergency Use Authorization (EUA) for Bamlanivimab in April, so there's that, and the REGEN-COV cocktail appears to currently be marketed in India by Roche, at over US$800 a dose. Effective it may well turn out to be - accessible possibly not.
Perhaps a good way to explore this, might be to imagine ourselves in the shoes of a public health czar, at the beginning of the pandemic last February. Witnessing the coronavirus spreading over China on JHU's tracker, our dutiful czar might correctly read the situation as SNAFU. But it's not his job to mope about feeling sorry for everyone, and so the search for solutions begins: Face masks? Everyone gets some once supplies are assured, that's sensible. Vaccines? The record on that hasn't been great for coronaviruses, but fund the researchers generously, and hope for the best. Travel restrictions? Ban from countries where there's known serious outbreaks, quarantines for incoming visitors whenever possible. It's all logical?
Being on a roll now, our wise public health czar then figures that despite all precautions, it remains entirely likely that the coronavirus will penetrate the borders - and if so, it would be great to explore all means of reducing its impact. You'd want to treat the condition early if at all possible, that's just common sense, and it would be best if some cheap, old, widely-used drug could help. You'd know the proper safe dosages, you'd have long-term data, you could therefore test them with a clear conscience on first doing minimal harm. The assistants have been digging through the coronavirus literature for suitable candidates, and there was this promising paper from CDC authors on how "chloroquine is a potent inhibitor of SARS coronavirus infection and spread". Sure, COVID-19 is not quite the same as the original SARS, but they are similar in a lot of ways, and HCQ is a less toxic derivative of chloroquine with long-established dosages. Surely this might be worth a try? Can't hurt, might help... oh, and toss in the vitamins too.
And actually, this basic line of thought was indeed quite amenable to many scientists, because it really did make sense. The risks are bounded, because these are well-known and common medications; the potential benefits are attractive. At this point, sharp readers might realize - did you not complain earlier, about there being no proper early treatment trials? But there seem to be plenty of them displayed here?
Well, the twist here is on proper. It should be recognized that many of the trials represented in the aggregate above, have their faults. In particular, relatively very few of them are RCTs. But, as covered last June - and further supported by Cochrane and the WHO above in the context of face masks - RCTs are often impractical, and possibly even unethical. Consider an example of a disease with a baseline 20% fatality or permanent disability rate progressing from mild symptoms with standard of care (i.e. nil), which is not entirely outside the range of the coronavirus for elderly patients with comorbidities. To run an RCT with 1,000 subjects in each arm, one would expect 200 to perish from the control group. Then, if your treatment being tested does indeed have the 50% efficacy hoped for, one would have only 100 (or fewer) perish from the treatment group.
The trouble, of course, is that due to the required random assignment to the treatment and control arms, patients may be denied a chance at treatment when they might actually have preferred so, and suffer the consequences up to and including death. For the above example, it can hardly be denied that some 100 lives from the control group had been sacrificed in a sense, because they could otherwise have been given the treatment from the outset. Thus, the preponderance of cohort and case-control studies, where patients would be allowed informed choice of therapy following in-depth discussion with their physician. Obviously, this means that the experimental arms will likely be biased, but this can theoretically be mitigated to a large extent through statistical adjustments and matched analysis. Anyway, one hopes this sheds some insight into why RCTs can't exactly be run, in functional healthcare settings.
With this in mind, one might then consider the aggregate of nearly 300 studies on HCQ above, of which 220 have been peer-reviewed. Again, they are hardly perfect (though it would be more suspicious if they apparently were), and I suppose one could gripe on the evaluation of strength of evidence (e.g. as from PRISMA guidelines), but the most pessimistic reading would still seem to suggest that there's likely something in early treatment, from the above. One has to extend the same courtesy of basic trust to the (some four thousand) authors involved in the above HCQ papers, as given to the face mask and vaccine researchers, after all? These include, mind, some recent pretty large cohort studies from Iran and France, on over 28,000 and 10,000 subjects respectively. If most or some of the hundreds of groups involved in these papers turn out to be in cahoots, I would be very interested to know, and open to reconsidering my position on HCQ. But if not, it is hard to see why the judgment isn't at worst inconclusive, needs more study.
Anything interesting with this table extract from Figure 2 of the authoritative The BMJ living systematic review?
[N.B. Observations to be discussed a few paragraphs later]
One would be extremely hard-pressed to find that level of measured opinion on HCQ, etc. in the corrupt corporate media, however, and on this the justification has tended to be directed towards two meta-analyses: The BMJ's living systematic review (as cited solely by the WHO in recommending against HCQ), and the Cochrane Library's assessment. For both, one characteristic immediately stands out: only RCTs are to be considered. At a single fell stroke, the vast bulk of evidence on the matter has been unceremoniously denied as improper. Myself, I am unsure why peer-reviewed cohort studies should be essentially given an evidential weight of literally zero in the scientific endeavour of medicine, seeing as they (and worse) have been accepted by physicians and the mass media in advising the public on face masks, egg consumption and all sorts of health decisions. But the WHO, The BMJ and the Cochrane Library all say so, and they are all honourable organizations. So be it, then.
Anyhow, the number of RCTs and participants considered in these systematic reviews on HCQ (and other candidate early treatments) are 10 trials with 4,935 participants for The BMJ, and 12 trials with 8,569 participants for Cochrane, which comes to about 500 to 700 participants per RCT. Here, it might be appropriate to recall how a totally fabricated paper published in The Lancet last May, completely tanked HCQ trials for some weeks, before being retracted in disgrace when its entirely unbelievable and baseless claims of HCQ doubling death rates contributed to its comprehensive debunking by outside parties. If you have forgotten all about that sordid affair, it is completely understandable too, because this episode has for some reason almost never been recounted in the corrupt corporate media, in the year since. As pointed out then, one has to imagine that intentional fraud against a plausible treatment for a virus ravaging the globe, must rightly be amongst the juiciest news opportunities of the year. Why? How? What for? If speculation was ever warranted, it would have to be around such, but, uh... might we interest you in fine celebrity boobs* instead?
But okay. Bygones can be bygones, the world's interest should rightly be centered around solutions to the coronavirus, so let's get back to finding them through holding trials... or not. And now, the discussion on the The BMJ table extract above. If one looks closely at what the table actually says, it suggests for instance that HCQ+AZ actually improves mortality from a standard care base rate of 130 per 1000 to 88 per 1000... just that the certainty is very low. Well, fine, one might say, before realizing that half the full table of treatment findings in Figure 2 are blank, and the remainder are nearly all of "very low" or "low" certainty, however with almost no harmful effects suggested. Given this, I think there may be a valid question here: why, after nearly one and a half years of the coronavirus, are we still at low certainty or worse for almost all early treatments?! Should not this sort of thing have been settled a lot earlier, from how many of these treatments have been floating around as plausible candidates since the beginning?
As another example, one would suppose that whether reasonable doses of HCQ, would help outpatients avoid eventual hospitalization. One figures this is of high practical interest, given all the concerns about hospital overcapacity in the event of an outbreak. Well, the Cochrane Library review presents the sum total of extant human knowledge on this important question, in Analysis 1.5 (Page 68):
On this, there is literally one single acceptable RCT examining the potential of HCQ to reduce hospitalization, with about 230 subjects in each arm, for which the HCQ arm recorded 4 hospitalizations and the control arm 10 hospitalizations, for a 0.41 risk ratio with HCQ. Certainly, if one were to argue that this hardly proves that early treatment with HCQ helps in reducing hospitalization, I guess I would have to agree. The point, I think, should be why the heck do we still not really know? Why does running low-risk trials towards evaluating potential early outpatient treatments, seem to be some minimum-priority vaguely-disreputable fringe position, when it should be by all rational reckoning, be amongst the lowest-hanging fruit towards pandemic management?
[*To clarify, I have nothing against these; I consider myself an equal-opportunity aesthete.]
Singapore To The Rescue?
For all that, the world continues to rotate. Millions are dead, as we are continually reminded by the mass media, but life goes on for the rest. Aircraft carriers, each containing thousands of crew, ply the seaways as ever, marvels of engineering all. Generals and terrorists, depending on which side one is standing on, keep exchanging missiles and firefights. Raising angry mobs, bombing residential buildings... you'd not want to admit it, but all these take minor wonders of organization. The soldiers have to be fed, the feints and counterstrikes planned, food and bullets and sundry supplies shuttled from truck to tunnel, to the minute sometimes. Thousands, tens of thousands, operating in clockwork precision, towards some set objective.
But gathering say about three thousand people, instructing them to take one pill a day for a month or so, and tracking whether they exhibit more severe symptoms? Now that's just crazy talk. It's beyond human ken. Perish the thought.
Well, I tend to be sparse in my praise for local organizations, perhaps overly so, but as hinted back in March, some of them have impressed with their instincts towards the ongoing scourge of muddled thinking at the WHO and some of the top academic journals. The Temasek Foundation has, for example, distributed a hundred thousand bottles of Vitamin D to the needy - which may or may not be related to preliminary indications that vitamin deficiency may be a significant predictor of resistance to acute respiratory infection. Again, can't see how it'd hurt, and it might help, as seemingly realized by some others.
The COPCOV HCQ Trial organizers have for one (entirely justifiably) denounced the WHO's recommendation against use of HCQ for COVID-19 due to "high certainty evidence that has emerged regarding the lack of effect of hydroxychloroquine prophylaxis", because as they rightly point out, there is no new information and certainly no high certainty evidence that it doesn't work, as self-admitted by The BMJ's systematic review previously discussed, that the WHO had referred to in their recommendation. But it's not like what one says has to correspond with actual proof nowadays. Still, the COPCOV group appears no closer to managing to come up with results, and it was left to us Singaporeans to actually complete a proper prophylaxis RCT, with HCQ and several other treatments. America, China, you lot can land bots on Mars? Well, we can run a balanced experiment involving application of throat spray towards reducing coronavirus spread, which neither of you seem at all capable of doing. No need to thank us.
And the results are... *drum roll* application of oral HCQ yields a 0.70 (95% C.I.: 0.44-0.97) relative risk as compared to Vitamin C dosing for the control group, i.e. a 30% improvement for the primary outcome of lab-confirmed coronavirus infection. Similar results were observed for povidone-iodine spray, and Zinc+Vitamin C (though the latter was unfortunately marginally non-statistically significant). Many thanks too to the participating dorm workers, happily none of whom succumbed to the disease, through the trial.
That... that wasn't that hard, was it?
The lead author of a previous famous paper on HCQ for postexposure prophylaxis (which actually doesn't even discount its possible efficacy) would put in some critique, which I think the trial's organizers will be happy to address. The authorship of this Singaporean trial on HCQ might be noted to consist of members from no less than nine different departments/divisions from the Yong Loo Lin School of Medicine and the National University Hospital, which I suppose could be interpreted as a strong consensus and united stand behind the findings... as it should be. Seriously, if one carefully runs a trial on accessible treatments for a pandemic, and some are actually shown to work, in what world would one refuse to publicize the results?!
Well, this breakthrough has made it to the front page of The State's Times (kudos for this), but I suppose it fair to say that there has been an... oppressive silence on this, in the Anglo media. This is hardly the last word, of course, but it remains an excellent proof-of-concept, that it is physically possible to run moderately-large RCTs with HCQ and Ivermectin and Zinc and Vitamin D and these kinds of inexpensive treatments. One small step for Singapore, one giant step for mankind, or something. Interestingly, this comes alongside Pfizer testing pill-based remedies, and as a Finnish firm patents - wait for it - a nasal spray containing HCQ, Ivermectin & Aprotinin. I really hope it works excellently, though being commercial it'd not unlikely be like US$10 a puff after insurance, but it'd be worth it to watch a certain set snort it (under some Pokémon-sounding brand name) while grouching about dumb HCQ believers.
The media materializes before me, in the figure of an androgyne in a lab coat. A card is held up. There is a red line and a blue line. The blue line is clearly shorter than the red line.
"Which is longer?" it drones, in a voice long shorn of residual humanity, but with the hidden force of nameless corporations unmistakably resonant in its timbre.
"Blue!" the answer comes, as a pallid grey hand claws, breaks the dead earth. The faces appear, bleak, soulless, alike, interchangeable. They simper at their False God, a mockery of the Truth, these piteous NPCs - so certain in their misguidance, so content in their ignorance. They swarm helplessly towards the wan reflected light, clambering over each other, towards the dribbles of sustenance that the figure emits. Facebook likes. Twitter retweets. Instagram tags. "Blue! Blue! Blue!"
Occasionally, a brief spark reappears in the eyes of one of the Greys. Always it is followed by horror, as it comprehends what it has become. It jerks, it turns, it moves against the press of its peers. They do not let it go. They never do. Baleful milky eyes fixated on their Idol, the same syllables falling from drooling mouths, they tear their erstwhile companion apart limb from limb, to the maddening flute-whines of CNN FAKE NEWS. I shiver and press closer to the cavern wall, I watch the remains twitch, now but bloody ribbons. It might yet be my fate.
I am Mark, and the red line is longer.
Next: Glory! Glory!
Copyright © 2006-2022 GLYS. All Rights Reserved.