Moderate alcohol consumption inversely correlated with all-cause mortality
My roommate recently sent me a review article that LW might find interesting:
Conclusions: Low levels of alcohol intake (1-2 drinks per day for women and 2-4 drinks per day for men) are inversely associated with total mortality in both men and women. Our findings, while confirming the hazards of excess drinking, indicate potential windows of alcohol intake that may confer a net beneficial effect of moderate drinking, at least in terms of survival.
Personal observation says that LWers tend not to drink very much or often. Perhaps that should change, to the degree suggested by the article?
Full article here.
I started drinking around a year ago just because of studies like this.
Me too. But I stopped after a month of 1-2 drinks/day (for no particular reason). Now I drink at parties (so, essentially, never).
Different drinks have different negative side effects. I actually like the taste of most beer, wine, and scotch but don’t like the feeling I get after/during a night’s sleep.
Even if moderate drinking is good for people who like to do it, the good effects might not be there for people who don’t like it.
Yes. I thought so too—if it’s really helping me, mightn’t I see some incremental short-term benefit?
Huge (often genetic) variations in response to particular foodstuffs are already well known. (though my intoxication : dose response seems typical, really).
I can’t tell if this is supposed to be a joke or not.
It’s funny nonetheless because of the double meaning.
I don’t get it.
The second meaning in the double-entendre would be isomorphic to “Reading studies like this drove me to drink.”
It’s not a joke.
Just fermented drinks (wine or beer) or distilled drinks (whisky, vodka, etc.) as well? Just curious, my personal preference is for vodka, but with age advancing I’m thinking of making a move back to the more natural fermented drinks.
Mostly red wine, although I’ve recently started drinking piña coladas.
The loved one has started brewing mead. It’s fun having an alchemy lab in one’s kitchen, blooping away, and drinking the products of one’s hard work. (Home brewing is legal in the UK as long as you don’t sell the results or use distillation—in either case the revenooers come down on you like a ton of bricks, and in the latter case it’s too easy to make methanol if you don’t know what you’re doing.)
Me, too, but about two years ago. Unfortunately, I’ve had a hard time liking wine, so I’m hoping that moderate amounts of scotch and/or rum have a similar effect.
A post-doc at my lab told me that the non-drinker group include sober alcoholics, something that might cause the non-drinker group to have a higher overall mortality.
A mention of a study which excluded sober alcoholics, and hypothesizes that a little alcohol is good because it makes socializing more likely.
Good read, aside from reliving social anxieties, there is also the self-medicating aspect; since alcohol dulls anxieties in general some people claim that it actually has a net positive effect on public mental health, highly speculative, but worth considering.
Another possible confounding factor is that the amount people drink is correlated to some genetic differences, which might have other health consequences.
I don’t drink at all due, among other things, to the fact that alcohol tastes so awful to me that I cannot help involuntarily grimacing after taking a sip of wine, beer, or liquor.
I generally don’t like alcohol, though my tolerance has gradually increased (to way below average) over time.
Does distaste for alcohol correlate with anything else?
Supertasting. I strongly suspect I am one. There are only two things on the list of seven things supertasters often don’t like which I’ll consume (Brassica oleracea cultivars and soy) and I hate the other things.
I think of myself as a sub-supertaster. No to grapefruit and coffee (unless the coffee flavor is very mild as in ice cream). Yes to kale. No to alcohol, except in small quantities. Green tea is more tolerable than black. I seem to like olives less than a lot of people, but the problem is with bitterness more than saltiness. I don’t seem to like salt as much as a lot of people do. I like hot peppers a lot.
In general, my tolerance for bitter flavors has increased with time.
I wonder why it’s called “super” if it’s a bad thing...
Adverb: Especially; particularly.
Who says it’s bad? It means being more sensitive to taste; it just turns out that some things aren’t very nice if you can really taste them. Other stuff tastes great!
Thanks for the link! I look at how it describes my tastes: I hate beer, coffee, grapefruits, tonic. I dislike uncooked cabbage, green tea, soda. I’m OK with cooked cabbage, soy; and I like chilli and olives.
Unadjusted associations in observational studies should not guide decisions (“hospitals have a lot of sick people, therefore I should stay away from hospitals because they will make me sick!”). Either use a randomized trial, which is the gold standard for establishing causal association, or use an observational study and adjust for confounding and other biases appropriately.
I do avoid hospitals (except in emergencies) and precisely because they are dangerous.
“hospitals have a lot of sick people, therefore I should stay away from hospitals because they will make me sick!”
That’s not necessarily as silly as your offhand treatment seems to suggest; it’s precisely one of the explanations proposed for some of the RAND health insurance experiment’s results. (One of the conclusions the RAND study suggested was that, for people who aren’t both poor and chronically ill, reduction in healthcare services (resulting from increased use fees) might not have a significant impact on health. It wasn’t designed to test this relationship, however, and I’m unaware of any subsequent studies which were.)
My go-to reductio is “Olympic sprinters have lots of gold medals; I should wear lots of gold medals to run faster!”
Clearly, it only fails because there’s too many Olympic sports.
You have no way of telling if it’s making you run faster, swim faster, shoot better, or do backflips better.
I dub this the Bling Fitness Theory.
This is the clearest example of what’s wrong with evidential decision theory I’ve ever seen!
Your reply applies (at least) equally well to the following argument: all plants have green leaves; roses have green leaves; therefore roses are plants. In both cases your reply would defend an obviously silly inference pattern by pointing to one instance in which it leads to a true conclusion. (Although in your case the inference actually predicts that hospital-provided medical care has a negative marginal effect on health, not a neutral one.)
Not really. Illness is a proximally transitive property; being in the same room as an ill person can, indeed, make you ill. It only leads to support of silly inference patterns if you disregard the actual meaning of the words used in the logical statement.
Consider the actual meaning of “therefore” used in IlyaShpitser’s post. It’s logically stronger than “and”.
In the least convenient possible world where health care is good and diseases aren’t transmissible, the naive causal inference would lead to the same conclusion.
+1 to writing up my tentative ‘How to use the LCPW principle safely and sanely’ post. Or at least sending a draft to Yvain for thoughts.
Until that post comes, I’m curious what kind of an example you intend to make of my use of the concept—safe/sane or otherwise?
Otherwise—the post was basically “remember that the least convenient possible world has to at least be possible”.
Except that it wasn’t formal logic; the conclusion didn’t follow from the premise, because there’s nothing indicating that sickness is something which should be avoided. Therefore, the reader is expected to use information contextual to the world the author is in; namely, that sickness is bad, and should be avoided. Because there’s also another possible world where -diseases- are good, and the construction wouldn’t even make sense.
Yes, I of course agree that hospitals are cauldrons of disease, and can make you sicker (especially if recovering from surgery). Medical errors by staff is a big issue as well.
even then you have placebo and file drawer effects.
from the article: “Whenever possible, adjusted relative risks were extracted; otherwise, crude relative risks and 95% confidence intervals (CIs) were calculated from the number of events.”
From the paper itself:
“Forty-eight curves (908 182 subjects and 86 941 deaths) were adjusted at least for age; among them, 28 were adjusted for social status too, and 10 for social status and dietary markers.”
This is adjusting for 1-4 confounders out of many possible confounders. Even doing so they lost half of the association. Note that the scientists themselves do not claim a causal effect, but only association. It’s reasonable to take their word for this.
I personally would not change dietary habits just based on studies like this.
They claim causation in several places, albeit sprinkled with perhapses and maybes. From the abstract:
From the Comment section:
This is all causal language.
No because saying alcohol is healthy could get them in lots of trouble.
Not even a little? Does it at least slightly increase your estimate of alcohol being healthy and so if at a social event deciding whether to have a drink shift your cost benefit analysis?
| No because saying alcohol is healthy could get them in lots of trouble.
You may be right—on the other hand, one can get a lot of publicity for a controversial finding. I don’t think these sorts of studies use conservative language because they fear getting in trouble due to subject matter. I think they fear getting in trouble for using the wrong statistical methodology or the wrong language to describe it.
| Not even a little? Does it at least slightly increase your estimate of alcohol being healthy and so if at a social event deciding whether to have a drink shift your cost benefit analysis?
I am not a rationalist, and I don’t use these kinds of cost benefit analyses when out drinking :). These sorts of studies are simply never a swing vote in my decision making. I agree that these studies are weak evidence.
Speak for yourself! :-)
Color me unconvinced. These “benefits” may come from any number of things, and taking alcohol as a general remedy may not be an advisable course of action because the problem is likely to be specific. Consider the following (I’ll be using “longevity” as shorthand for “improvement WRT total mortality”):
Alcohol → lowered social anxiety → more socialization → mental well-being → longevity
Alcohol → distraction from (seemingly) insurmountable problems → mental well-being → longevity
Alcohol → [insert chemical that triggers some elusive beneficial biological process that causes your cells to degenerate slower or whatever] → longevity
The last one seems least likely to me, and if you can get the social benefits through some other avenue, you may want to consider those first. I do recall reading up on some other classic studies that showed that red wine has some genuine antioxidant properties and such, but a significant impact of general longevity? I ’unno. You may still be better off using your beer bucks to buy supplements or exercise opportunities.
And that’s all assuming the researchers were conscientious enough to control for the other stuff in the first place. Apologies in advance if they actually did, but I’ve been generally unimpressed with the rigor of studies that claim to show correlations between Purportedly-But-Not-Really-Simple Thing X and Complicated Gestalt Such As Total Mortality, and so I deliberately skimped on the conscientiousness myself. Corrections are welcome in case you guys did read the whole article. But in the meantime, try these on for size:
Alcohol → indication that your income level is comfortable enough that you can afford to buy alcohol → selection bias → longevity
Alcohol → drink and drive → don’t die yourself, but WHOOPS, you just killed a pedestrian → the statistics give the cause of death as “car accident” rather than “alcohol” → longevity
Life sucks → alcohol → get wasted regularly rather than commit suicide → getting wasted gets in the way of fixing the actual problem → improved but still stunted longevity
Etc.
I was about to say, “Of course they controlled for income, that’s totally basic”, but I looked. Most of the studies didn’t control for class or income. :( Looking elsewhere, income is positively associated with drinking, so income could well be the hidden variable increasing drinking and decreasing mortality.
Also, some of the studies controlled for body mass index, meaning that if your beer gut increases your mortality, that doesn’t show up.
Well said.
It would actually be interesting to see some research on the biological side of alcohol consumption, say, some studies on the longetivity of rats consuming C2H5OH-containing drinks versus their non-alcoholic controls.
(At the very least, the rats might be saved from less pleasant experiments...)
This was already done.
Ah, thanks! The research was about alcohol and clofibrates:
That seems pretty significant! Cheers!
The only way that would contribute to the total mortality rate for drinkers being lower than for non-drinkers would be if I’m more likely to kill a pedestrian given that the pedestrian is sober than given that the pedestrian is drunk.
(OTOH, an effect such as “I’m (going to get) drunk, so I’m not driving tonight → I’m walking back home rather than driving to there → I’m less likely to die walking a mile than driving a mile” would be in the right direction, though—I guess—much smaller than other effects. My money’s on the biggest effect being the one about income.)
Apparently walking drunk is actually eight times more dangerous (per mile traveled) than driving drunk. So we would not expect moderate drinkers to have lower death rates due to walking home after drinking—if anything, this would increase their death rate.
I’m extremely surprised by this. I can’t see how walking can be that dangerous, provided you walk on sidewalks (or in pedestrian zones) and are careful when crossing the street. What population does the “eight times” statistic apply to?
The text is here. It’s based on a couple of assumptions—that drivers and pedestrians are equally likely to be drunk, and that drunken trips are representative of all pedestrian trips. I doubt there’s hard data out there on whether these things are true, since nobody does random checks of pedestrians.
I don’t assume these things of sober pedestrians, let alone drunk ones. I assume people with impaired reflexes and judgement are more likely to cross the street when it’s not safe to do so, fall down, etc.
Alcohol is also involved in a lot of hypothermia cases. Your circulation is worse, but you feel warmer and you’re more prone to falling down. I had a friend who nearly spent the night in a snowbank staggering home from a college party—if she’d been a bit drunker and passed out, she would have frozen.
Yeah, I’d bet that
underestimates the fraction of miles walked drunk by several times; if it’s underestimated by a factor of 8, then driving drunk is as dangerous as walking drunk. (But I might be overestimating it, due significant differences between my country and the US, e.g. here in Italy, whereas you don’t get arrested if you’re caught driving moderately drunk for the first time,¹ I’m pretty sure the probability of getting caught is much larger than 1 in 27,000 miles, given that I know at least four people to whom that happened. Therefore I guess people are much more reluctant to drive when drunk where I am than over there. Also, Italian towns are less car-friendly than I guess US towns are.)
Though they temporarily revoke your driver’s licence, and they make it a pain in the ass to get it back (lots of psychological tests, drug tests once a year for the first three years then once every five years—paid by you, etc.).
Yeah, I guess the equation was misapplied there. The point was that the statistics won’t (or might not) chalk the death up to alcohol like they should, which I’d say is a harmfully misleading omission; even if it’s not a longevity problem for the drunk driver, it is for the other person.
I don’t think it would be right or proper to control for killing other people due to alcohol use even if you could. The social externalities of alcohol use are a separate question from the private benefits.
I agree that there’s some merit to treating alcohol’s effects on you and others separately, but if we do that, shouldn’t we then also work to exclude some of its benefits as “social externalities”? Like the whole “alcohol → socializing → mental well-being”-pattern?
You should exclude the mental well-being of the others you socialize with while drunk, yes. But that’s not going to show up on your personal longevity.
As has been pointed out., people who don’t drink are weird for reasons other than the fact they don’t drink (the most obvious one being that a large number of them are recovering alcoholics). Since an interventional study is pretty much impossible here, we’d need some natural experiment (something using drinking bans in Arab countries might work?) to have any real idea if there is causation. Until then, I suggest not-drinking is almost certainly less bad than drinking too much—and perhaps a more natural Schelling point than “I only drink 1-2 drinks per day”).
This looks like conjecture. I regularly have a pint of beer (about 1.5 UK standard measures) with dinner and don’t particularly feel I’m on a slippery slope either way.
I see this type of thing a lot. It’s kind of only interesting to me in an academic sense, because even if that amount of alchohol promotes longevity, I am not willing to drink due to being an utter control freak (and several relatives of mine had severe substance abuse problems, so if there’s any genetic component to that I want to avoid it). In any case, if it does, I wonder what the mechanism is?
You presume that I would rather live longer and drink than live for a shorter time and not drink. I deliberately take actions which do not maximize my expected lifespan, because a day of my life has finite value.
For example, drinking 2-4 drinks will impair me for a period of about 2-4 hours. That is a statistically significant portion of the day.
Nitpick: Why not just “significant”?
“Nitpick” understates it. That’s a pretty bad example of inflationary use of terms.
Because ‘statistically significant’ is falsifiable, while ‘significant’ is not.
‘Statistically significant’ refers to a specific criterion of falsifiability that doesn’t apply to your statement as phrased. You could invent a metric to make your claim more easily falsifiable: for example, “My happiness measured on a scale of 1-10 will decrease if I drink X amount of alcohol per day” is a testable claim, and having tested it, you could examine the results for statistical significance. But just adding the word ‘statistical’ to your sentence doesn’t make it a quantitatively falsifiable claim.
I wouldn’t call 2-4 drinks impairment… then again I don’t often drink in a time span that I would otherwise spend doing productive or difficult things.
Same here. Mild sleep deprivation usually impairs me much more (and the two things don’t seem to add linearly for me—if I’m already sleep-deprivated, drinking alcohol on top of that won’t make things much worse). Then again, I’m about 90 kg (198 lb) and this needn’t generalize to smaller people.
Correlation does not imply causation and such. For the latter to be tested for, a randomized study that makes people change their drinking habits would be required. The hard part is blinding it, of course.
Why do you have to blind it?
To make it most realistic, you would not only give the treatment group alcohol, you would tell them that controlled trials suggest that this treatment (self-referentially) will reduce all-cause mortality.
I presume that there is a joke in there somewhere, but I fail to see it.
I don’t think it’s a joke. I think paulfchristiano is saying that if we are interested in the effect treating studies like as reasons to change habits, then we should study the effects of changing behavior because of a study like this.
An unblinded study would not be able to account for placebo effects.
Why wouldn’t I experience placebo effects if I started drinking as the result of controlled trials which suggest that drinking will reduce all-cause mortality?
I want to know the effectiveness of the combination of chemical effects and placebo effects from the treatment I take, not the effectiveness of an outwardly similar treatment that has had significant changes made for the purpose of getting results different from the results I should expect.
Ah. I believe we have interpreted shminux’s top-level comment differently. I think shminux was stating that establishing the direct causal effects of alcohol consumption requires a blinded randomized controlled trial (RCT), which is true. However, blinding is indeed not required if one wishes to include the effects from the intervention method.
I am undecided on whether blinded RCTs are more cost-effective at the moment (obviously, it depends sensitively on what one wants to find out). In any case, I think we’ll agree that any interventional study – blinded or unblinded – will offer a larger degree of control than the current observational studies. However, I did a brief literature search, and found Klatsky’s (2010) comments on the difficulties in doing interventional studies on drinking:
I don’t think that the kind of people who are willing and able to take part in a randomized study of alcohol consumption are representative of the general population.
Gak, I misread that title initially as Moderate alcohol consumption inversely correlated with all-cause morality
Time for a break from lesswrong.