As one of the non-physicians here I would like to appeal for a guest blog we really need: Something from a doc who treated hospitalized Covid patients in 2020-21, and continues to do so now. Ideally, an infectious disease or pulmonary/critical care specialist.
What did the patients look like then, and now? How many are there? Are they admitted because of Covid, or with other conditions made worse by Covid, or is Covid incidental? How many develop a significant Long Covid syndrome—both at the beginning of the pandemic, and now?
I think the pressures on doctors are largely political—they’re made to feel that if they admit that things have changed, or that any precaution might be dropped, they are simply bad people who want to abandon the “most vulnerable among us”.
Hearing very concrete, day-to-day details from a doctor who is genuinely on the “front lines” would be so valuable! Most doctors who are asked for guidance by their friends or political allies do not have experience of this type. So they just succumb to the pressure to endorse whatever hysterical concept of “safety” feels good on political grounds to their non-medical friends.
This is good work by the student reviewers. The problem is that the review is not nearly critical enough. The foundational paper is flawed in so many ways (selection, definition of PASC [88 selection criteria?], testing ambiguities, etc.) that it does not deserve critique so much as withdrawal. It is work like this that has deprecated the NEJM and other once-good journals to yellow rag status.
Next time you do this, instead of parsing nuanced failings, it is OK to come out and note that the foundations of the paper are bereft and that none of it merited publication...Marcia Angell was so right when she resigned years ago. That would be an even more worthwhile critical contribution.
Congrats to these Med students for having the bandwidth to tackle critical appraisal. Esp in 2nd year, I was up to eyeballs in assimilating the basics and would not have been capable of this. It’s also a skill that was not taught back in my day in Med school, so good for them to start the process of honing this aspect early. It’s the key to maintaining competence over a decades long career.
That said, time management is also a skill that needs to be sharpened, and I’m not sure a study like this is the best way to spend their precious time.
I would note their summary of author conclusions uses the word “contributes” wrt vaccine effect on long covid incidence. This is NOT what the author wrote (at least in the abstract; this is not the type of paper I would spend time reading the entirety of). They imply the author made causal inferences where it appears he did not. He seems to simply describe incidences, which seems like a fair use of observational data (whereas any causal inference clearly would not be).
That said, long COVId at this point seems to be such a morass of diagnostic confusion as to bedevil any sort of serious study that would be ready for clinical prime time. Any “diagnosis” that proposes to amalgamate 80+ different conditions spanning multiple organ systems would seemingly be telling the TB’s and SLE’s of the world to “hold my beer” where it comes to being “the great mimicker”.
I don't see the operational definition of long covid, especially vis-a-vis duration. The study patients were followed for at least a year, suggesting that putative symptoms of long covid must persist for at least a year. Is this the case? The CDC first used a duration of persistent symptoms of one month duration, but I believe the WHO uses three months. Did the study use a one month duration, and count these as cases, even if fully resolved by one year? Was it three months even if resolved by one year? Or was it a requirement that symptoms must persist without resolution ever?
Another area of further research is to do a cross national comparison of long covid. Many of the symptoms are subjective and thus incentivized for secondary gain, such as eligibility for disability. It would be interesting to compare PASC rates in the USA to nations lacking disability insurance. Or to other nations with a different cultural expectation of disease experience.
"PASC remains extremely difficult to study" is almost the correct conclusion. Actually it is impossible. The quality of the data is the most important and essential part of any scientific study. With "covid" we have a fictional virus that was never physically isolated or cultured and "infections" defined by lab procedures (PCR) and tests ((antibodies) that are extremely unreliable. During the "pandemic" the diagnosis of covid was tossed around like confetti at a political rally. The students' time would have been better spent reviewing the original study out of Wuhan that "identified" a novel corona virus. Pay particular attention to the Methods section that describes how a few short segments of RNA were sequenced and run through a computer program that assembled them into a large number of presumed viral genomes---some say possibly several million---picked one, and slapped the name covid on it. Subsequently almost all of our usual seasonal respiratory disorders formerly known as colds and flu were labelled covid.
Congratulations on the analysis. It is extremely complex to assert a causal inference—implicit by the stratification of analysis by vaccination—in a syndrome that is inadequately and chaotically defined.
Recent post-COVID data show that 9 out of 10 individuals lack metabolic health. COVID-19 and PASC are more prevalent in patients with metabolic disease, which suggests that this large subgroup, being more prone to COVID and its sequelae, likely has a higher vaccination rate.
Thus, it’s probable that the unvaccinated group, with a lower prevalence of metabolic disease or without evident clinical expression, may have a different susceptibility compared to vaccinated individuals, and not due to the vaccine itself, regarding the study’s outcome variables. This aspect could have been detected if a propensity score had been applied, particularly focusing on each of the 80 conditions of the syndrome in relation to the degree of metabolic alteration.
Another point highlighting the substantial confounding factor—metabolic syndrome—is that nearly 30% of severe COVID-19 cases result in a novo diagnosis of Type 2 Diabetes (requiring insulin), which is then classified as prolonged COVID-19.
From my perspective, these are likely metabolic or clinical conditions triggered by the effects of the infection. In other words, COVID infection and the development of disease have, as a significant causal substrate, metabolic disease, an aspect that should always be controlled in observational studies such as this one.
Thanks for your work—and please keep it up! Dr. Al-Aly’s poorly designed, irresponsible studies, based on creative misinterpretation of VA records, have been hugely influential unfortunately. Uncritical coverage in the mass media has stoked both irrational fears of Long Covid, and exaggerated ideas of Covid vaccine effectiveness.
In April 2024 I attended a big labor union conference which announced a universal mandatory masking policy for all participants. Here’s what they wrote to justify this:
“Even if you’ve had Covid before, we don’t want you catching it again here. Long Covid can hit anybody, and it’s likelier with each re-infection. Our movement needs all of us healthy and safe for the next fight!”
(The actual effectiveness of masks, either for self protection or to protect others, is another contested topic of course.). To their credit, the organizers backed off from actual enforcement of the mask mandate once the conference began. However, other politically Progressive gatherings in Chicago this year have doubled down on these mandates—making communication harder and walling off their followers from the rest of the population. We will continue to pay the price in increased division and isolation until more members of the medical profession gain the courage to challenge this evidence-free piety. THANK YOU!
I don’t understand why people think they can draw meaningful conclusions about the benefits of the covid vaccines by using observational data.
For example, a study done in Israel claimed a 90% reduction in fatalities by taking the covid booster. The raw data showed a 94.6 % reduction of covid deaths in the vaccinated arm. Looks impressive! The authors of the study left out an important detail.
Tracy B. Høeg, Ram Duriseti, and Vinay Prasad discovered that the data from that study showed the NON covid death rate in the vaccinated was also 94.8% lower. When you read the study, you think the vaccines are a miracle. But when you see the full picture, it looks like the vaccines are useless at best. This is far from the only example showing an extreme healthy vaccination bias.
Observational studies showing benefits for the covid vaccines are meaningless. The machine will never stop trying to convince the public how great these vaccines are. The people have heard so much BS they are growing more skeptical.
By the way, as the Wuhan virus does not cause viremia, Long-COVID is just a typical bacterial squellae post infection. As most physicians were following silly Chinese and WHO guidelines antibiotics were rarely used.
The authors do not understand that there were no epidemiology studies on the so-called variants. The variants were purely virtual and existed as sequences without patients' information.
To the authors of this paper: excellent work. Your careful critique of the study is very enlightening. I lack formal training so I appreciate your contribution, it's much better than anything I could do.
To look at the Long COVID phenomenon objectively I think that maybe the Bradford Hill criteria should be applied. A lot of the hard thinking has been done already in general, it just needs to be applied in this case. When the letter asked for stratification by severity, this corresponds to the dose dependent response criterium.
Unfortunately it seems that the PASC is so amorphous as to be untestable by others. I think it would be helpful if the progenitor of the PASC concept would define a core set of symptoms/sequelae so that others could do independent analyses without being accused of selective inclusion of symptoms.
After working with "doctors" for several years, I haven't met with a single one who would even consider that option. Assessing radiation exposure is not rocket science:
Of course, even AIDS was an invented illness with no viable diagnosis (no virus has ever been properly isolated and the tiny partices, probably exosomes, proven pathogenic):
good start... but you guys seem far to generous... this paper seems like a whitwesh to try to justify vaccine use/mandates... I think you should focus on why the vax effects were likely way overstated and why the paper should have never been published as evidence of yet more corruption and heathcare malpractice during covid era
oh and one small nitpick: testing of asympt people was never a good idea. this bucks long standing resp virus best practice (see pandemic planner)
I agree that the vaccine effects were likely overstated and that it would have been good to address that. We could not come up with a specific reason other than chalking it up to poor methodology as to why the difference between vaccinated and unvaccinated was claimed to be 70% due to vaccines.
Thank you for the response. I know you wanted to be fair and credible and I appreciate that. I think the links cited above (the Hoeg/Prasad work for example and other links on health vax effect) could have been brought into the piece to show that was likely at play here as well. On top of that, asking the journal why they would publish a piece of shoddy work like this and to retract it would also be helpful (believe me, I emailed them myself but no luck obviously)... thanks again
Dear Bad Ass Motherfucker,
Be The First To Prove
That Gain Of Function Research
Has, In All Of Its Present Attempts,
Failed To Produce An Aerosolized Pathogen Capable Of Spreading To, And Among, The Entire Human Population.
And I Promise You The World.
Whoever You Are.
As one of the non-physicians here I would like to appeal for a guest blog we really need: Something from a doc who treated hospitalized Covid patients in 2020-21, and continues to do so now. Ideally, an infectious disease or pulmonary/critical care specialist.
What did the patients look like then, and now? How many are there? Are they admitted because of Covid, or with other conditions made worse by Covid, or is Covid incidental? How many develop a significant Long Covid syndrome—both at the beginning of the pandemic, and now?
I think the pressures on doctors are largely political—they’re made to feel that if they admit that things have changed, or that any precaution might be dropped, they are simply bad people who want to abandon the “most vulnerable among us”.
Hearing very concrete, day-to-day details from a doctor who is genuinely on the “front lines” would be so valuable! Most doctors who are asked for guidance by their friends or political allies do not have experience of this type. So they just succumb to the pressure to endorse whatever hysterical concept of “safety” feels good on political grounds to their non-medical friends.
This is good work by the student reviewers. The problem is that the review is not nearly critical enough. The foundational paper is flawed in so many ways (selection, definition of PASC [88 selection criteria?], testing ambiguities, etc.) that it does not deserve critique so much as withdrawal. It is work like this that has deprecated the NEJM and other once-good journals to yellow rag status.
Next time you do this, instead of parsing nuanced failings, it is OK to come out and note that the foundations of the paper are bereft and that none of it merited publication...Marcia Angell was so right when she resigned years ago. That would be an even more worthwhile critical contribution.
Congrats to these Med students for having the bandwidth to tackle critical appraisal. Esp in 2nd year, I was up to eyeballs in assimilating the basics and would not have been capable of this. It’s also a skill that was not taught back in my day in Med school, so good for them to start the process of honing this aspect early. It’s the key to maintaining competence over a decades long career.
That said, time management is also a skill that needs to be sharpened, and I’m not sure a study like this is the best way to spend their precious time.
I would note their summary of author conclusions uses the word “contributes” wrt vaccine effect on long covid incidence. This is NOT what the author wrote (at least in the abstract; this is not the type of paper I would spend time reading the entirety of). They imply the author made causal inferences where it appears he did not. He seems to simply describe incidences, which seems like a fair use of observational data (whereas any causal inference clearly would not be).
That said, long COVId at this point seems to be such a morass of diagnostic confusion as to bedevil any sort of serious study that would be ready for clinical prime time. Any “diagnosis” that proposes to amalgamate 80+ different conditions spanning multiple organ systems would seemingly be telling the TB’s and SLE’s of the world to “hold my beer” where it comes to being “the great mimicker”.
I don't see the operational definition of long covid, especially vis-a-vis duration. The study patients were followed for at least a year, suggesting that putative symptoms of long covid must persist for at least a year. Is this the case? The CDC first used a duration of persistent symptoms of one month duration, but I believe the WHO uses three months. Did the study use a one month duration, and count these as cases, even if fully resolved by one year? Was it three months even if resolved by one year? Or was it a requirement that symptoms must persist without resolution ever?
Another area of further research is to do a cross national comparison of long covid. Many of the symptoms are subjective and thus incentivized for secondary gain, such as eligibility for disability. It would be interesting to compare PASC rates in the USA to nations lacking disability insurance. Or to other nations with a different cultural expectation of disease experience.
"PASC remains extremely difficult to study" is almost the correct conclusion. Actually it is impossible. The quality of the data is the most important and essential part of any scientific study. With "covid" we have a fictional virus that was never physically isolated or cultured and "infections" defined by lab procedures (PCR) and tests ((antibodies) that are extremely unreliable. During the "pandemic" the diagnosis of covid was tossed around like confetti at a political rally. The students' time would have been better spent reviewing the original study out of Wuhan that "identified" a novel corona virus. Pay particular attention to the Methods section that describes how a few short segments of RNA were sequenced and run through a computer program that assembled them into a large number of presumed viral genomes---some say possibly several million---picked one, and slapped the name covid on it. Subsequently almost all of our usual seasonal respiratory disorders formerly known as colds and flu were labelled covid.
Dear colleagues,
Congratulations on the analysis. It is extremely complex to assert a causal inference—implicit by the stratification of analysis by vaccination—in a syndrome that is inadequately and chaotically defined.
Recent post-COVID data show that 9 out of 10 individuals lack metabolic health. COVID-19 and PASC are more prevalent in patients with metabolic disease, which suggests that this large subgroup, being more prone to COVID and its sequelae, likely has a higher vaccination rate.
Thus, it’s probable that the unvaccinated group, with a lower prevalence of metabolic disease or without evident clinical expression, may have a different susceptibility compared to vaccinated individuals, and not due to the vaccine itself, regarding the study’s outcome variables. This aspect could have been detected if a propensity score had been applied, particularly focusing on each of the 80 conditions of the syndrome in relation to the degree of metabolic alteration.
Another point highlighting the substantial confounding factor—metabolic syndrome—is that nearly 30% of severe COVID-19 cases result in a novo diagnosis of Type 2 Diabetes (requiring insulin), which is then classified as prolonged COVID-19.
From my perspective, these are likely metabolic or clinical conditions triggered by the effects of the infection. In other words, COVID infection and the development of disease have, as a significant causal substrate, metabolic disease, an aspect that should always be controlled in observational studies such as this one.
Jairo Echeverry-Raad. MD
Thanks for your work—and please keep it up! Dr. Al-Aly’s poorly designed, irresponsible studies, based on creative misinterpretation of VA records, have been hugely influential unfortunately. Uncritical coverage in the mass media has stoked both irrational fears of Long Covid, and exaggerated ideas of Covid vaccine effectiveness.
In April 2024 I attended a big labor union conference which announced a universal mandatory masking policy for all participants. Here’s what they wrote to justify this:
“Even if you’ve had Covid before, we don’t want you catching it again here. Long Covid can hit anybody, and it’s likelier with each re-infection. Our movement needs all of us healthy and safe for the next fight!”
(The actual effectiveness of masks, either for self protection or to protect others, is another contested topic of course.). To their credit, the organizers backed off from actual enforcement of the mask mandate once the conference began. However, other politically Progressive gatherings in Chicago this year have doubled down on these mandates—making communication harder and walling off their followers from the rest of the population. We will continue to pay the price in increased division and isolation until more members of the medical profession gain the courage to challenge this evidence-free piety. THANK YOU!
In 2024?!? OMG!
I don’t understand why people think they can draw meaningful conclusions about the benefits of the covid vaccines by using observational data.
For example, a study done in Israel claimed a 90% reduction in fatalities by taking the covid booster. The raw data showed a 94.6 % reduction of covid deaths in the vaccinated arm. Looks impressive! The authors of the study left out an important detail.
Tracy B. Høeg, Ram Duriseti, and Vinay Prasad discovered that the data from that study showed the NON covid death rate in the vaccinated was also 94.8% lower. When you read the study, you think the vaccines are a miracle. But when you see the full picture, it looks like the vaccines are useless at best. This is far from the only example showing an extreme healthy vaccination bias.
https://www.nejm.org/doi/full/10.1056/NEJMc2306683
Observational studies showing benefits for the covid vaccines are meaningless. The machine will never stop trying to convince the public how great these vaccines are. The people have heard so much BS they are growing more skeptical.
By the way, as the Wuhan virus does not cause viremia, Long-COVID is just a typical bacterial squellae post infection. As most physicians were following silly Chinese and WHO guidelines antibiotics were rarely used.
The authors do not understand that there were no epidemiology studies on the so-called variants. The variants were purely virtual and existed as sequences without patients' information.
And what was the counting window for being “vaccinated” - 14 days after second dose ?
To the authors of this paper: excellent work. Your careful critique of the study is very enlightening. I lack formal training so I appreciate your contribution, it's much better than anything I could do.
To look at the Long COVID phenomenon objectively I think that maybe the Bradford Hill criteria should be applied. A lot of the hard thinking has been done already in general, it just needs to be applied in this case. When the letter asked for stratification by severity, this corresponds to the dose dependent response criterium.
Unfortunately it seems that the PASC is so amorphous as to be untestable by others. I think it would be helpful if the progenitor of the PASC concept would define a core set of symptoms/sequelae so that others could do independent analyses without being accused of selective inclusion of symptoms.
"Long covid," just like the convid symptoms, must have been due to radiation poisoning.
https://rayhorvaththesource.substack.com/p/the-invasion-of-the-infernal-towers-13b
After working with "doctors" for several years, I haven't met with a single one who would even consider that option. Assessing radiation exposure is not rocket science:
https://rayhorvaththesource.substack.com/p/do-you-want-to-know-your-radiation
Of course, even AIDS was an invented illness with no viable diagnosis (no virus has ever been properly isolated and the tiny partices, probably exosomes, proven pathogenic):
https://rayhorvaththesource.substack.com/p/catching-the-idea-of-a-virus
good start... but you guys seem far to generous... this paper seems like a whitwesh to try to justify vaccine use/mandates... I think you should focus on why the vax effects were likely way overstated and why the paper should have never been published as evidence of yet more corruption and heathcare malpractice during covid era
oh and one small nitpick: testing of asympt people was never a good idea. this bucks long standing resp virus best practice (see pandemic planner)
I agree that the vaccine effects were likely overstated and that it would have been good to address that. We could not come up with a specific reason other than chalking it up to poor methodology as to why the difference between vaccinated and unvaccinated was claimed to be 70% due to vaccines.
Thank you for the response. I know you wanted to be fair and credible and I appreciate that. I think the links cited above (the Hoeg/Prasad work for example and other links on health vax effect) could have been brought into the piece to show that was likely at play here as well. On top of that, asking the journal why they would publish a piece of shoddy work like this and to retract it would also be helpful (believe me, I emailed them myself but no luck obviously)... thanks again