Setting aside my bias about whether my large glass of wine last night was one drink or two, I find the claims in this post odd. None of the issues noted seem inherently problematic, and indeed several are as I would expect. Disclaimer: not an epidemiologist, although I do fit models to data and have seen more than a few epidemiology talks.
1: Asymmetric CIs aren't mysterious. Any parameter fit to anything beyond the simplest model and data will have asymmetric CIs. We often neglect this and assume a symmetric (normal) approximation, but in complex data you really should calculate the proper CIs.
2: The data in the age groups wouldn't be expected to be either fully dependent or independent. Looking at the data and assumptions that goes in, some affects all age groups, some only one, and there's model effects between groups (if you die due to alcohol at 30, you can't also die due to alcohol at 50). The CI size lying somewhere between independent and fully correlated is exactly what you'd expect there.
3. The last complaint is a more fair, but also common: most people who die are old, and most causes of death go up when you get old, so CIs get big when you have to subtract large numbers. More data is always welcome there, but also famously hard to get.
So, none of these seem like show stoppers to me, and the way some of them are presented in the post ("mysterious asymmetric CIs") seems to speak to a motivation to knock this study down.
I agree the stats methods in the paper are a bit thin, but I'm not sure how much of that is standard practice in the field. But I think that's a case for asking more detail, rather than denying the results out of hand.
<drug> consumers every time: it's actually not bad in moderate consumption/I can quit any time/medication is always a question of dosage
Looks to me like even taking meds for a prolonged amount of time shows you that to any problem you try to fix with externally administered chemistry always has a downside to deal with, too.
Please stop trying to frame your drug habit as a good thing, be it caffeine, weed or worse things like alcohol and beyond. Another one of those things where your mental health depends on you not understanding the problems you cause yourself.
If there's no explanation of why the 70+ bucket has so much larger CI band, having that dominate the totals seems like very fair critique. At the very least, they should have presented another totals with the outlier removed to show the impact.
Even the best people seem to go out of their way to rationalise (their own?) alcohol consumption. As a nowadays leading cause of cancer in many regions, and associated with dependency and other secondary negative outcomes, perhaps it is time to stop rationalising it?
And maybe also let off the rhetoric of "government sponsored" science, which is actually most science in many places. Written for views/clicks obviously. So tiering.
The following quote is especially problematic IMHO:
> For a paper claiming to overturn decades of alcohol drinking recommendations, the statistical foundation is underdocumented and, to be mild, statistically strange.
Given that "decades of alcohol drinking recommendations" lacked any statistical basis to begin with, the scientific-rigor bar to clear is quite low IMHO.
It's just writing for dramatic effect it seems. The claim that the paper is 'used as “definitive” argument in the medical discussions' seems quite absurd given that the paper itself was published just two months ago.
Your argument is misguided and tries to make the issue about the cultural background of the people making the argument rather than the merits of the argument itself. Most of the researchers I've worked with consume alcohol, they are taking an informed risk and are okay with that. There is not so much opposition as it is a reaction to having seen what alcohol does to the body and people.
I think nobody would argue drinking a lot of alcohol, or substantial alcohol, is detrimental for health.
The question being argued is whether drinking a small amount of alcohol is detrimental to your health. This is a different matter altogether.
We, humans, have been drinking alcohol for millenia. I'm skeptical to the idea that a small amount, say, a glass of wine for lunch every day or so, can harm us, especially for active individuals. There has been a natural selection process for us to adapt.
I believe what Taleb is doing here is of great public value, and I commend him for that.
I spent some time working with CBT for alcohol use disorder, mainly treating "active individuals". I've heard your line of reasoning many times. No level of alcohol consumption is considered risk free from a scientific point of view, at least the view of the senior researchers and professors I worked quite closely with.
So, I don't think this is providing "great public value", on the contrary I would suggest spending time with the research and treatment of the many individuals starting out thinking that "a glass of wine for lunch every day or so" can't "harm us".
From a certain point on we must start being adults again and become aware that some health risks are worth the price when it comes to drinking a glass or two of wine/beer every-once in a while. Because wine/beer have helped us become more of a social animal over the millennia, i.e. have helped us become humans, and suddenly reversing that for the sake of, I don't know, 6 months-a year of prolonged living in their late 70s - the 80s is not worth it.
With that said, I'm fully aware that excessive alcohol intake (i.e. alcoholism) has been destroying lives for millennia, too, but that's what life is, we can't turn it into this aseptic thing where every risk is accounted for.
But, alcohol is something that has been present in our ancestral environment (e.g. fermented fruits). If anyone's telling me, our bodies didn't learn to at least tolerate small amounts of it without any major side effects, I'm going to be asking a lot of follow-up questions.
I think it's much easier to argue this case than, say, tobacco or "drugs", which are arguably more recent inventions, depending on your exact definition of drug.
Same logic applies to heavy metal like lead and others. Also cyanide and arsenic. And so on. We still actively avoid them. Just because we can tolerate them via various process does not mean it wouldn't be better to avoid them altogether and minimise exposure.
If alcohol was invented today, people would be like "no way we allow a juice that makes you happy, stupid, aggressive and endangers yourself and others and has high addictive levels".
>And maybe also let off the rhetoric of "government sponsored" science, which is actually most science in many places.
Is it? Government science is usually very closely tied to politics, and that's so far from "science". I didn't see any real science in your post as well btw.
I am inclined to take Nassim's position as true, since it agrees with my desired beliefs: that moderate drinking is quite harmless, or if there are harms, they are outweighed by the benefits.
Modern medicine often gives us access to amazing drugs which have, in small print, a long list of associated, known, proven harms to our organs and health. The logic, as I understand it, is that the benefits of the drugs outweigh the harms.
The logic is not "that these drugs are harmless and possess only benefits".
This is important when discussing (eg) cancer treatments: a treatment may leave you alive for longer but in worse health — and therefore, it’s unclear which is preferable.
Arguably when left alive your health is by definition better than the alternative.
I suspect you mean the good health for a few months before death compared with a worse health during and after treatment. Still I had to make the quip.
QALY optimization doesn't necessarily optimize for quality of life like it claims on the tin, it optimizes for quantity of life. Would you like 1 year and unknown discomfort or 10 years and a colostomy bag?
Near end-of-life interventions (in the US) tend to be unnecessary and leave patients worse off, and it's done because hospitals and doctors are greedy.
You die in both scenarios, you live slightly longer (eg, 8 months versus 6 months) and during the overlap, you are less healthy than you would have been. Losing 6 months of seeing family, enjoying trips, etc due to side-effects from treatment just to live 2 months extra, but incredibly ill the whole 8 months.
In my perspective: you experienced less total QALY by taking the treatment and had more time, in worse health. Even counting those 2 months at zero, the non-treatment option experiences more vitality, in total — because 2 sickly months at the end doesn’t offset 6 months of lost vitality immediately before.
Alcohol is a carcinogen, famous for causing various upper gastrointestinal cancers. It also isn't great for brain cells. Why are these reasons not enough to quit it?
If you want something to drink in the evenings, how about a cup of mild bancha green tea andor rooibos tea andor hibiscus tea? If you do, please get the loose leaf product only, never the bagged product, as the latter infuses microplastics.
Agreed - it's all in the dose and the response. Not even mentioning the selection pressure for us to adapt to at least small amounts of alcohol in our diet, which has been present in basic things such as fermented fruits that we had to eat in our ancestral environment.
In small doses, alcohol may have other benefits that outweigh the risk of gastrointestinal cancer.
Whether it's as strong as the now Resveratrol scam or not is a matter to study honestly, though that quality seems beyond the reach of many humans motivated to prioritize something else.
Setting aside my bias about whether my large glass of wine last night was one drink or two, I find the claims in this post odd. None of the issues noted seem inherently problematic, and indeed several are as I would expect. Disclaimer: not an epidemiologist, although I do fit models to data and have seen more than a few epidemiology talks.
1: Asymmetric CIs aren't mysterious. Any parameter fit to anything beyond the simplest model and data will have asymmetric CIs. We often neglect this and assume a symmetric (normal) approximation, but in complex data you really should calculate the proper CIs.
2: The data in the age groups wouldn't be expected to be either fully dependent or independent. Looking at the data and assumptions that goes in, some affects all age groups, some only one, and there's model effects between groups (if you die due to alcohol at 30, you can't also die due to alcohol at 50). The CI size lying somewhere between independent and fully correlated is exactly what you'd expect there.
3. The last complaint is a more fair, but also common: most people who die are old, and most causes of death go up when you get old, so CIs get big when you have to subtract large numbers. More data is always welcome there, but also famously hard to get.
So, none of these seem like show stoppers to me, and the way some of them are presented in the post ("mysterious asymmetric CIs") seems to speak to a motivation to knock this study down.
I agree the stats methods in the paper are a bit thin, but I'm not sure how much of that is standard practice in the field. But I think that's a case for asking more detail, rather than denying the results out of hand.
<drug> consumers every time: it's actually not bad in moderate consumption/I can quit any time/medication is always a question of dosage
Looks to me like even taking meds for a prolonged amount of time shows you that to any problem you try to fix with externally administered chemistry always has a downside to deal with, too.
Please stop trying to frame your drug habit as a good thing, be it caffeine, weed or worse things like alcohol and beyond. Another one of those things where your mental health depends on you not understanding the problems you cause yourself.
It's not bad in moderation. Sugar is much more addictive and harmful than alcohol, and it's also fine in moderation.
Please provide any scientific evidence of your claim.
[delayed]
Heroin is fine in moderation! /s
The thing is that it's impossible to moderate one's intake as, as you said, it's addictive.
[delayed]
If there's no explanation of why the 70+ bucket has so much larger CI band, having that dominate the totals seems like very fair critique. At the very least, they should have presented another totals with the outlier removed to show the impact.
I'll drink to that! Cheers!
Even the best people seem to go out of their way to rationalise (their own?) alcohol consumption. As a nowadays leading cause of cancer in many regions, and associated with dependency and other secondary negative outcomes, perhaps it is time to stop rationalising it?
And maybe also let off the rhetoric of "government sponsored" science, which is actually most science in many places. Written for views/clicks obviously. So tiering.
// Master of Science in Public Health
The following quote is especially problematic IMHO:
> For a paper claiming to overturn decades of alcohol drinking recommendations, the statistical foundation is underdocumented and, to be mild, statistically strange.
Given that "decades of alcohol drinking recommendations" lacked any statistical basis to begin with, the scientific-rigor bar to clear is quite low IMHO.
It's just writing for dramatic effect it seems. The claim that the paper is 'used as “definitive” argument in the medical discussions' seems quite absurd given that the paper itself was published just two months ago.
People from puritanical cultures rationalise their opposition to other people consuming alcohol.
Ad hominem arguments cut both ways.
Your argument is misguided and tries to make the issue about the cultural background of the people making the argument rather than the merits of the argument itself. Most of the researchers I've worked with consume alcohol, they are taking an informed risk and are okay with that. There is not so much opposition as it is a reaction to having seen what alcohol does to the body and people.
I think nobody would argue drinking a lot of alcohol, or substantial alcohol, is detrimental for health.
The question being argued is whether drinking a small amount of alcohol is detrimental to your health. This is a different matter altogether.
We, humans, have been drinking alcohol for millenia. I'm skeptical to the idea that a small amount, say, a glass of wine for lunch every day or so, can harm us, especially for active individuals. There has been a natural selection process for us to adapt.
I believe what Taleb is doing here is of great public value, and I commend him for that.
I spent some time working with CBT for alcohol use disorder, mainly treating "active individuals". I've heard your line of reasoning many times. No level of alcohol consumption is considered risk free from a scientific point of view, at least the view of the senior researchers and professors I worked quite closely with.
So, I don't think this is providing "great public value", on the contrary I would suggest spending time with the research and treatment of the many individuals starting out thinking that "a glass of wine for lunch every day or so" can't "harm us".
> No level of alcohol consumption is considered risk free from a scientific point of view
From scientific point of view life leads to death in 100% of cases.
I love the sock puppet throwaway account trolling here.
There are risks getting out of bed in the morning, but there are greater risks NOT getting out of bed in the morning.
Risk is apparently a very difficult subject.
From a certain point on we must start being adults again and become aware that some health risks are worth the price when it comes to drinking a glass or two of wine/beer every-once in a while. Because wine/beer have helped us become more of a social animal over the millennia, i.e. have helped us become humans, and suddenly reversing that for the sake of, I don't know, 6 months-a year of prolonged living in their late 70s - the 80s is not worth it.
With that said, I'm fully aware that excessive alcohol intake (i.e. alcoholism) has been destroying lives for millennia, too, but that's what life is, we can't turn it into this aseptic thing where every risk is accounted for.
There is no evidence either small quantities of tobacco or most drugs harm you either.
People that want to drink will find excuses to do so.
Maybe — I haven't looked at the evidence.
But, alcohol is something that has been present in our ancestral environment (e.g. fermented fruits). If anyone's telling me, our bodies didn't learn to at least tolerate small amounts of it without any major side effects, I'm going to be asking a lot of follow-up questions.
I think it's much easier to argue this case than, say, tobacco or "drugs", which are arguably more recent inventions, depending on your exact definition of drug.
Same logic applies to heavy metal like lead and others. Also cyanide and arsenic. And so on. We still actively avoid them. Just because we can tolerate them via various process does not mean it wouldn't be better to avoid them altogether and minimise exposure.
Smoked meat has been a staple of human diet for millenia.
It's cancerous.
Which "drugs"?
The problem with highly addictive things is people do not stop at small quantities.
If alcohol was invented today, people would be like "no way we allow a juice that makes you happy, stupid, aggressive and endangers yourself and others and has high addictive levels".
>And maybe also let off the rhetoric of "government sponsored" science, which is actually most science in many places.
Is it? Government science is usually very closely tied to politics, and that's so far from "science". I didn't see any real science in your post as well btw.
I am inclined to take Nassim's position as true, since it agrees with my desired beliefs: that moderate drinking is quite harmless, or if there are harms, they are outweighed by the benefits.
Modern medicine often gives us access to amazing drugs which have, in small print, a long list of associated, known, proven harms to our organs and health. The logic, as I understand it, is that the benefits of the drugs outweigh the harms.
The logic is not "that these drugs are harmless and possess only benefits".
This is important when discussing (eg) cancer treatments: a treatment may leave you alive for longer but in worse health — and therefore, it’s unclear which is preferable.
Arguably when left alive your health is by definition better than the alternative.
I suspect you mean the good health for a few months before death compared with a worse health during and after treatment. Still I had to make the quip.
Everyone will die. The question is whether, say, two extra months spent in a miserable state is worth it.
QALY optimization doesn't necessarily optimize for quality of life like it claims on the tin, it optimizes for quantity of life. Would you like 1 year and unknown discomfort or 10 years and a colostomy bag?
Near end-of-life interventions (in the US) tend to be unnecessary and leave patients worse off, and it's done because hospitals and doctors are greedy.
https://www.scientificamerican.com/article/surgery-near-end-...
No — I meant worse health.
You die in both scenarios, you live slightly longer (eg, 8 months versus 6 months) and during the overlap, you are less healthy than you would have been. Losing 6 months of seeing family, enjoying trips, etc due to side-effects from treatment just to live 2 months extra, but incredibly ill the whole 8 months.
In my perspective: you experienced less total QALY by taking the treatment and had more time, in worse health. Even counting those 2 months at zero, the non-treatment option experiences more vitality, in total — because 2 sickly months at the end doesn’t offset 6 months of lost vitality immediately before.
Alcohol is a carcinogen, famous for causing various upper gastrointestinal cancers. It also isn't great for brain cells. Why are these reasons not enough to quit it?
If you want something to drink in the evenings, how about a cup of mild bancha green tea andor rooibos tea andor hibiscus tea? If you do, please get the loose leaf product only, never the bagged product, as the latter infuses microplastics.
Because risk/benefit assessment is different from person to person.
Dose makes the poison. And some substances are beneficial at low doses and harmful at high doses.
Do you foreswear aspirin?
Agreed - it's all in the dose and the response. Not even mentioning the selection pressure for us to adapt to at least small amounts of alcohol in our diet, which has been present in basic things such as fermented fruits that we had to eat in our ancestral environment.
In small doses, alcohol may have other benefits that outweigh the risk of gastrointestinal cancer.
Whether it's as strong as the now Resveratrol scam or not is a matter to study honestly, though that quality seems beyond the reach of many humans motivated to prioritize something else.
> never the bagged product, as the latter infuses microplastics.
Seriously? How can a paper tea bag infuse microplastics?