Note this is a not-yet-peer-reviewed preprint.
https://www.medrxiv.org/content/10.1101/2021.06.11.21258690v...
IMO, while I'm not inclined to take either study at face value immediately, I think a serious person should be equally reluctant to join the "it's just a flu" camp either.
https://i.ibb.co/5YcxJHH/EDB20654-900-A-427-C-8063-B5-FA667-...
"Nearly half the case participants had more gray matter after getting COVID" - as did the controls. It's almost like there's variability in the readings. If only someone would invent a subspeciality of mathematics about how to interpret noisy data and attempt to draw conclusions about the results.
I'm not totally sold on this study either, mind you. I'm just amazed at how many people are rushing to judgement in the other direction - the whole "just a flu" conclusion is wildly premature.
Yes, I mention the high variance. And indeed there are standard ways to interpret the effect size magnitude, given noisy data. The authors chose not to report Cohen's d. They do provide an r value (.16), which we can use to compute the the coefficient of determination. Since you are an advocate for reliance on traditional statistical interpretations, you should appreciate the meaning of r^2=.025
Furthermore the authors report a P=.01. They bootstrap this P, presumably because alpha significance level after multiple comparison adjustment renders cutoff well below .01. They also claim no a priori hypothesis wrt. gray matter increase or decrease. A 2-tailed alpha is typically shifted from p<.05 to p<.01 which their measured P value does not surpass. Nevertheless they claim statistical significance. Finally, after normalization, there is roughly an equal chance any given person will have more or less gray matter after getting covid - this doesn't require one to glean the importance of each ounce of gray matter to understand the effect size is small. I contend there is good reason this study has not yet passed peer review, if it ever will.
Figure 2 is the one you want https://els-jbs-prod-cdn.jbs.elsevierhealth.com/cms/attachme...
Which does show effect increasing with severity of the case. I've seen plots like that before in similar articles.
See also table 2.
The long-term effects of flu are not studied that well, by the way. On the basis we're used to it from before such studies were possible.
None of those symptoms we hear about, getting our breath, unable to smell, brain fog etc., none of them are new. And they imply internal damage that lasts.
Your second remark is also peculiar: the "obvious sampling problem" to which you allude is either (1) a subtler-than-expected-point about the UK Biobank's program (designed to be a mass participation program to observe longitudinal effects in previous healthy or as yet undiagnosed people) or (2) evidence you are just making stuff up.
There's also not a globally accepted definition of the syndrome yet, so it won't show up in stats.
If you go by "still experiencing at least one symptom after 3 months", it's closer to 10-20% than 1% : https://twitter.com/Dr2NisreenAlwan/status/13775492575703162...
Another study found that around a third of these 10-20% consider it debilitating (significantly affects their daily lives)
I don't know the appropriate way to extrapolate out but we're well past 200M globally.
0: https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/burd...
1: https://graphics.reuters.com/world-coronavirus-tracker-and-m...
Or are you making some other point I'm missing?
Getting estimates from the same people will provide a much lower variance estimaion of any potential effect.
Also, adding hospital records to this study would potentially provide a better estimation.
All studies have flaws, but this one does look pretty reasonable (the principal components method used should have been better described, for instance, and confidence bounds on the effect sizes would have been super useful).
Basically this is junk science. The only proper way to do such a study would be to compare the same subjects before and after infection. With so many cases such a study should be feasible.
I get the objection, but I'm getting annoyed at comments that state what's invalid with the study where we don't have the data available for that claim. What's wrong with "we don't know if the study is valid since we don't have details on X"?
Overall the majority of COVID-19 clinical research has been rushed and very low quality.
I spent years being dismissed as depressed Or having anxiety. When I finally was diagnosed and started treating condition with medications I could handle my cognitive problems and depression lifted.
Support groups for autoimmune conditions are full of people who routinely are dismissed by doctors.
I don't know how we teach general practitioners, but we are doing something horribly wrong. A lot of long covid people must be finding this out too, and it is just super sad.
One of my biggest lessons early in life was that not all doctors are the same (by a long shot). This is why it’s so important to see specialists related to your condition. And ultimately you’re going to be most responsible for getting proper care as often only you know what you’re experiencing.
Medicine and pathology are far to wide of subjects for generalists to operate efficiently across every field and niche. Which is why specialization is so important and connecting the patients to the right specialists.
There’s a lot of analogies to technology, where people tend to expect the experts to do and know everything and completely defer to them. I think we all have experienced this with our parents or people who didn’t grow up with computers. When ultimately the individual is always going to be a huge part of the puzzle and must accept that fact.
We send them to a school which emphasizes cramming vast amounts of established knowledge into their heads, at the expense of critical thinking skills, social skills, and for that matter sleep. In a world where doctors use Google and WebMD the same as everyone else (because of course they do, no one retains an encyclopaedic knowledge of every medical condition), it's basically a hazing ritual.
What disease was it? What kind of medications have helped you?
I'm suffering from some kind of autoimmune-like disease with similar symptoms and trying to figure out what's wrong with me.
Plaqunial helped a lot, but I can’t tolerate it. Shame, it works well. Steroids are good for ending a flare but it’s hard to get prescribed, also side effects.
What really worked for me me was * blood thinners for cognitive defects * Avoid triggers, direct sun, sugar, caffeine, spice * autoimmune protocol diet * magnesium, vit d, e, b complex, fish oil
Get blood work done for each kind. ANA test as a starting point.
For me Early Sjo test confined diagnosis.
Make a list of ALL symptoms and take it with you to all doctors. A Neuro-ophthalmologist was the first to say Sjogrens based only off my sheet.
An elimination diet is probably the best place to start. I ate nothing but sweet potatoes for a month, my symptoms all cleared up. Stated coming back when I introduced random foods.
That’s how I started following AIP diet
People complaining of fatigue or having difficulty concentrating or having brain fog are in some studies already counted as suffering from long covid. No causual relationship required.
A friend of mine is a doctor and we were chatting about long covid 2 days ago, his description of it seemed pretty cut and dried to me - e.g. a swimmer here in Glasgow with hopes of getting to the 2024 olympics is currently unable to train due to breathing issues post-covid. They’re supposed to be fully immersed in training right now but can’t since covid in December 2020.
It's just not consistent with the "It's all in their minds theory," or "it's the same as stress from isolation" theory.
There's clearly something neurological going on, and we don't understand it yet.
Was a nightmare getting those under control but my symptoms eased up. Then I got covid and it all started up again. Took months to get back to a baseline, but symptoms were identical, aside from loss of smell.
Ofcourse long covid is real. I am just a bit skeptical about the diagnosis criteria.
I really don't understand what point you tried to make.
The whole point of scientific papers is to gather and document findings in a way to be subjected to critical analysis and serve as food for thought.
From your own example, which unfortunately is completely unsubstantiated, documenting symptoms reported by patients is a good starting point to form hypothesis to be verified or rejected.
"A common challenge in studies of COVID-19 is that differences between people who have vs. have not been ill could relate to premorbid differences. To address this issue, a linear model was trained on the broader independent GBIT dataset (N = 269,264) to predict general cognitive performance based on age (to the third order), sex, handedness, ethnicity, first language, country of residence, occupational status and earnings. "
"We controlled for various things in the original regression, but hey, there still could be unmeasured confounders that make people who got COVID score less on IQ tests, even before they got COVID! (Like coming from a poorer background, other health vulnerabilities etc. etc. etc.) Unfortunately, we don't have a measure of IQ from before they got COVID. So, we estimated one using a set of variables that we do have measures for!"
Erm... if those extra variables predict IQ, then why not just add them as controls? And of course, if you do so, then you'll still hit the problem that there are plenty of potential unmeasured confounders out there. This is just a silly way to pretend you've controlled for something, when in fact you can't.
The quality of statistics in medicine is so bad. Disgracefully bad. In particular, the Lancet seems to be a serial offender.
However, they did determine those individuals who contracted COVID were not different in their premorbid test performance from those who did not later contract COVID. This is a big part of what a "complete" longitudinal design would get you.
Let's say you did get that test data on follow-up. If post-COVID people who were infected were different cognitively, and not different premorbidly, that would suggest COVID was involved. If they weren't different cognitively, you'd have to explain why the first wave of individuals were different based on COVID history but not the later wave. Still important to show but maybe a different set of explanatory challenges.
That seems to add credibility here to a causal mechanism as opposed to a background correlation of risk of infection they are missing.
Is this not a randomised control trial? As I understand it, RCT is usually referred to as the “gold standard” - why would the gold standard be a flaw in this case?
What you describe sounds to me, a layman, as an improvement over and above but if RCT is good enough normally, why discount it entirely in this case because there’s some other improvement could be made?
No. I am not quite sure how one might even think the words apply here.
If it were, participants would be randomly assigned to the the "control" group where they do not get Covid, but would be told they had Covid (to control for the negative placebo of being told they are now afflicted with a horrible long term malady) and the "treatment" group which would actually be given a Covid infection. Assuming such a study were doable, one can then use a "within participant" comparison which is much more powerful than comparing statistics across groups.
>> We sought to confirm whether there was an association between cross-sectional cognitive performance data from 81,337 participants who between January and December 2020 undertook a clinically validated web-optimized assessment as part of the Great British Intelligence Test, and questionnaire items capturing self-report of suspected and confirmed COVID-19 infection and respiratory symptoms.
There is no randomized assignment here. Susceptibility to report a Covid infection without positive proof is a confounding variable. I suspect that susceptibility is correlated with other cognitive issues.
A "within participant" comparison as the GP brought up would avoid these kinds of issues.
That’s what’s require for the gold standard.