I examine the nature medicine paper
As always I appreciate your analysis of papers that make headlines and appear to be “game changers” to people without the knowledge/training to put in proper context. I have been in medicine for 30 years. You are correct that skepticism in medicine is dead. I call todays physicians “bandwagon doctors”, always jumping to put all their patients on the latest, greatest (most expensive) medications and our medical schools are churning them out by the thousands. Bring back #conservativemedicine
Journalism is dead too in terms of providing balance. And it’s too bad that a tiny fraction of the population will get exposure to Vinay’s perspective vs. that exposed to the headlines. Goldmines proliferate, but more of them exist in the space to keep COVID top of everybody’s mind.
Vinay's points are spot on. This paper is far less than it is reported to be.
Interestingly, Igor Chudov does analyses that are often excellent and, using the supplemental data tables published in this same paper, posits that vaccines may have actually had a negative effect on cardiovascular complications. His analysis is worth a look. If Vinay ever read these comments (I do not believe he does) his thoughts about Igor's analysis would be contributory. But read it for yourself in any case.
Thanks for performing a brilliant and informative exegesis of the "steaming pile of doo-doo" that this article from worker bees in the DVA (Department of Veterans Affairs) has turned out to be. Despite your use of two run-on sentences and some rather glaring verb-subject disagreements as well, the post you have just made here in the last hour earns a score of 9.375 out of 10 possible points from me. Its last three paragraphs were a stunning wrap-up feature, by the way, and a real treat found there was the great but old punch line you inserted so deftly, and not once but technically twice (" . . . .show me a person with no diagnoses and I'll show you someone who's not had enough tests . . . ."). Finally, it's just pathetic that the authors who wrassled together their "control group" apparently have had no proper exposure, if any, to the key concept of "exchangeability" in the context of designing clinical outcome studies.
But . . .
In your erudite comments there was *one dog that did not bark in the night* (Reference: "Silver Blaze", Doyle A.C., 1892). You expressed no curiosity whatsoever about how in hell this article, in manuscript form, survived any peer-review worthy of the name, and we have here an article accepted by NATURE MEDICINE no less! I just checked on the status of this journal and here's what is trumpeted: "It is a monthly journal publishing original peer-reviewed research in all areas of medicine on the basis of its originality, timeliness, interdisciplinary interest and impact on improving human health". Hmmmm.
The DVA article you dissected in fact provides another exemplar of something I've been moaning about for the past 18 months -- Academic Meta-Issues Stimulated by COVID -- viz. a whole bunch of stuff, and especially in epidemiology's bailiwick, has been rushed into print without attention to critically checking details, without the application of careful scholarship, and without worthy expenditure of sufficient elbow grease. I've been around the "academic game" for a long time and it does seem to me that a whole bunch of folks are "building their C.V.s" and doing it almost willy-nilly. Of course, now we get to await the huge Blowback (that's a CIA term, actually) because the capped-teeth, hair-sprayed hosts on cable TV news shows have been handed very rich grist for their odious mills. Need I say that the major "click bait" aspect here -- so obviously -- is the presence of the magic words "cardiac disease". Were it not for the impending invasion of Ukraine, were it not for the Super Bowl coverage that's well underway, it's a safe wager that Wolf Blitzer and his ilk would have gone wild about the message in this faux "epidemiology work product" emanating from DVA. God help us.
I’ve subscribed. Thank you for your detailed analysis and opinions. There just has to be a way to get back to rela science. What you are doing here will hopefully push in in that direction. I’m curious though, you work at a prestigious medical university, UCSF, and there has to be others like you. Are they simply to afraid to speak?
On some level, your deep concern for the deterioration of science should depress me. But your explanations just give me a little peace. A glimmer of hope. Somebody isn’t just going along with the craziness and acting like it’s perfectly reasonable.
Bravo Dr Prasad!!!
Dear Dr. Prasad:
I recently saw your podcast with Zdogg! First I have to say: what great guys you are! Funny, nuanced, collegial, and decent. I've loved watching the podcast throughout this pandemic.
I want to discuss your discussion with Zdogg about: "what is the evidentiary basis for the vaccines being effective against hospitalization due to covid?" And you and Zdogg concede that, well, the clinical trials DO NOT prove this, though there is a hint of a signal of this in the clinical trials. You then mention all the real-world data on hospitalization rates based on vaccination status. Though this data is confounded and has lots of problems, the strong effect observed justifies, in your mind, the suggestion to take the vaccine.
I don't think there is any problem with your analysis per se, but there is a glaring flaw. It is the flaw of OMISSION. Specifically, you are omitting the alternative of early treatment!
Now, I know what you will say in response to this: ahh, but there's problems with those studies! doctors who treat early can claim they cured everyone, but they would have just been cured on their own! etc. etc. etc.
But do you see your inconsistency here? Why are you applying one standard of evidence to vaccine analysis and another standard of evidence to early treatment protocols?
Sure, the evidence for, say, the Fareed Tyson protocol is not GOLD STANDARD double blind randomized clinical trials, but NEITHER is the evidence for the vaccines! You admit that yourself!
And indeed, it is FAR EASIER to show that an early treatment protocol is effective because you only give it to those testing positive or exposed to covid. The "clinical effect" against hospitalization is much easier to observe and takes far, far fewer samples to detect. Indeed, THIS IS what is detected with early treatment protocols. And this is detected across multiple RANDOMIZED trials to a statistically significant degree. Whereas the randomized trials for covid vaccines DO NOT show a statistically significant effect.
My question is thus this: what DISTINGUISHES the evidence for the vaccines compared to the evidence for early treatment protocols? Why do you suggest the vaccine, but seem to outright dismiss early treatment protocols discussed on Rogan?
Thanks, and I greatly appreciate your thoughts on this. This is the biggest weakness I see in the pro-vaccine side, and I want to know the best argument for this. Thanks!
I would actually defend the study as being well done and pretty rigorous. I absolutely agree that we shouldn't read too much into these types of studies due to biases galore. And of course, the way people are talking about this study on Twitter is overwrought, but is that really the authors' fault?
It did give me pause that the study found an association with _every_ cardiovascular outcome they tested. That seems a little too pat. As someone who has worked with data like this a lot, I always get nervous when the results look this neat. (if everything is associated, believe none of them)
For this reason, I was please to see the authors included negative-outcome controls. Great sanity check! And indeed, zero of seven pre-selected negative controls were association with COVID+. This is a good sign, except... where the hell did they get these outcomes??
- Melanoma in situ
- Sickle cell trait
- Perforation of the tympanic membrane
- Malignant neoplasm of the tongue
- B-cell Lymphoma
- Hodgkin’s Lymphoma
I'm really scratching my head here. How the heck did they come up with this list? Was it truly random (like pick a list of seven ICD codes from a bag)? And is a diagnosis like hypertrichosis really an appropriate negative control, and how many cases do they have (paper doesn't say)? I know that there is evidence for covid causing just about anything, but still, I think you could come up with a better set of diagnoses, and maybe try to chose acute conditions (like the study outcomes) w/ similar incidence. Am i missing something?
That little detail does make me wonder if something else is going on... But otherwise, I think it was a nice study, IMHO. Do I trust the conclusions absolutely? No, but that's a really high bar. Like most studies, followup is needed to solidify or reject the hypothesis. Thanks again for the thoughtful commentary.
Patient Zero. Scary.