We’ve gotten a bit into the weeds in the other comments, and in this one I’d like to zoom out a bit to see what argument you’re actually making. I’ll make an attempt to (re)construct your argument, and you can tell me where I’m misrepresenting it.
The health burden claimed in this post is extraordinarily high, so we should see the signal even in noisy data like news reports about athletes and celebrities
We’re not seeing those reports as much as we would expect
Conclusion: something in the report is wrong
I’ve already argued against 2 with the Airtable containing >100 athlete sudden deaths/collapses + a few news articles of arguably below-top-level.
[Jan 2023 EDIT : I don’t think this Airtable is strong data anymore, but very weak]
Re: 1.
I do think that it should be taken into account that the information ecosystem around COVID and Long Covid is really, really bad. Patients typically get misdiagnosed a lot before biological abnormalities are found, public health authorities spread a lot of misinformation, and most media outlets have pretty bad coverage. In this ecosystem, I don’t think it’s easy to spot athlete retirements due to confirmed Long Covid, or any other signals.
Re: 3
More importantly, I’m of the opinion that the evidence I offer in the post is of sufficiently higher quality than a google search for news reports: i.e. cohorts with controls, population samples, disability data, and biological data (e.g. seems like at least 50% of LC patients have COVID-specific markers). In my opinion, if you want to assert that the main claim in the report is wrong, you have to additionally argue that at least one of the following is wrong (If not, I think you’re only justified to claim ‘something here doesn’t make sense, but it’s not clear what’).
A. The controlled cohort studies & disability data is wrong
B. The controlled cohort studies & disability data are do not justify the high amounts in the UK population sample
C. The population sample is right, but professional athletes have significantly lower rates of Long Covid
D. The numbers are right, but a signifcant fraction are not attributable to COVID, but something else
E. Something else, or a combination of weaker versions of the above claims
I agree with you that C is unlikely.
Also, the rate would need to be substantially lower for my claim that ‘this is a major problem’ to be invalid (although you’re not explicitly claiming it’s invalid). E.g. at 60million/year, it’s still enormous. At 30million, arguably still big.
I had a quick look at the Airtable. Many of the people included do not seem to be professional athletes at the time that they died/collapsed. For example, this includes an ice hockey player at a university, someone who plays basketball in the fourth tier of the Spanish league and a former pro runner. This expands the sample so much as to make inference from the data impossible. There is a reason to focus on professional footballers in England because we know the sample size, there are a lot of them (5,000) and we should expect news about long covid-induced retirement to make it into the news.
By your estimates, 0.25% of people in the whole UK population are impaired a lot by long covid. We should therefore expect 13 of the 5,000 English pro footballers to have retired or gone on the news saying they can’t play because of long covid. I have looked into this and know of no cases of this. I know other people have looked into this after I offered them a bet and also haven’t found any.
I think the studies of long covid are wrong and that the controls are not good. The symptoms of long covid are vague, highly variable in severity, and already widely prevalent in the population (in the integers of percent).
It’s actually higher than 0.25%. More like 1 in 5 out of ~1.8% (avg. prevalence among 17-34, with shootings >3 months), so 0.36%.
Some of those will be recent though, so those we shouldn’t expect to be reported in the news even if the news was taking everyone. 30%?
Some will retire not knowing it’s actually Long Covid and state other reasons. 50%?
That leaves like 6 people, which to me is sufficiently small that it can be missed by chance (eg. no top level players have gotten severe Long Covid).
I’m also wondering if heart failure is another outcome rather than Long Covid and disability. ME/CFS is a really strange disease where people can push through a lot, and only get the bill later. It’s not that people literally can’t run.
Regarding the studies: I agree that there’s a lot to be desired regarding symptom measurement (I think we’ll see better measurement in the future). But even the vague symptom descriptions are significantly higher in PCR-confirmed covid cases, so I don’t understand your worry.
I went through the Airtable more systematically and found 7 English football players that had heart issues/collapsed on the field in 2021/2022. None were explicitly linked to covid, but only 1 had rumours of an underlying condition. 2 out of 7 players were in League 7 though. I think it’s still pro, not sure.
We’ve gotten a bit into the weeds in the other comments, and in this one I’d like to zoom out a bit to see what argument you’re actually making. I’ll make an attempt to (re)construct your argument, and you can tell me where I’m misrepresenting it.
The health burden claimed in this post is extraordinarily high, so we should see the signal even in noisy data like news reports about athletes and celebrities
We’re not seeing those reports as much as we would expect
Conclusion: something in the report is wrong
I’ve already argued against 2 with the Airtable containing >100 athlete sudden deaths/collapses + a few news articles of arguably below-top-level.
[Jan 2023 EDIT : I don’t think this Airtable is strong data anymore, but very weak]
Re: 1.
I do think that it should be taken into account that the information ecosystem around COVID and Long Covid is really, really bad. Patients typically get misdiagnosed a lot before biological abnormalities are found, public health authorities spread a lot of misinformation, and most media outlets have pretty bad coverage. In this ecosystem, I don’t think it’s easy to spot athlete retirements due to confirmed Long Covid, or any other signals.
Re: 3
More importantly, I’m of the opinion that the evidence I offer in the post is of sufficiently higher quality than a google search for news reports: i.e. cohorts with controls, population samples, disability data, and biological data (e.g. seems like at least 50% of LC patients have COVID-specific markers). In my opinion, if you want to assert that the main claim in the report is wrong, you have to additionally argue that at least one of the following is wrong (If not, I think you’re only justified to claim ‘something here doesn’t make sense, but it’s not clear what’).
A. The controlled cohort studies & disability data is wrong
B. The controlled cohort studies & disability data are do not justify the high amounts in the UK population sample
C. The population sample is right, but professional athletes have significantly lower rates of Long Covid
D. The numbers are right, but a signifcant fraction are not attributable to COVID, but something else
E. Something else, or a combination of weaker versions of the above claims
I agree with you that C is unlikely.
Also, the rate would need to be substantially lower for my claim that ‘this is a major problem’ to be invalid (although you’re not explicitly claiming it’s invalid). E.g. at 60million/year, it’s still enormous. At 30million, arguably still big.
I had a quick look at the Airtable. Many of the people included do not seem to be professional athletes at the time that they died/collapsed. For example, this includes an ice hockey player at a university, someone who plays basketball in the fourth tier of the Spanish league and a former pro runner. This expands the sample so much as to make inference from the data impossible. There is a reason to focus on professional footballers in England because we know the sample size, there are a lot of them (5,000) and we should expect news about long covid-induced retirement to make it into the news.
By your estimates, 0.25% of people in the whole UK population are impaired a lot by long covid. We should therefore expect 13 of the 5,000 English pro footballers to have retired or gone on the news saying they can’t play because of long covid. I have looked into this and know of no cases of this. I know other people have looked into this after I offered them a bet and also haven’t found any.
I think the studies of long covid are wrong and that the controls are not good. The symptoms of long covid are vague, highly variable in severity, and already widely prevalent in the population (in the integers of percent).
Thanks for the clarity John!
It’s actually higher than 0.25%. More like 1 in 5 out of ~1.8% (avg. prevalence among 17-34, with shootings >3 months), so 0.36%.
Some of those will be recent though, so those we shouldn’t expect to be reported in the news even if the news was taking everyone. 30%?
Some will retire not knowing it’s actually Long Covid and state other reasons. 50%?
That leaves like 6 people, which to me is sufficiently small that it can be missed by chance (eg. no top level players have gotten severe Long Covid).
I’m also wondering if heart failure is another outcome rather than Long Covid and disability. ME/CFS is a really strange disease where people can push through a lot, and only get the bill later. It’s not that people literally can’t run.
Regarding the studies: I agree that there’s a lot to be desired regarding symptom measurement (I think we’ll see better measurement in the future). But even the vague symptom descriptions are significantly higher in PCR-confirmed covid cases, so I don’t understand your worry.
I went through the Airtable more systematically and found 7 English football players that had heart issues/collapsed on the field in 2021/2022. None were explicitly linked to covid, but only 1 had rumours of an underlying condition. 2 out of 7 players were in League 7 though. I think it’s still pro, not sure.
Analysis here.