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.
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.
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.
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.
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"):
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").
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.
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.
Gak, I misread that title initially as Moderate alcohol consumption inversely correlated with all-cause morality
Time for a break from lesswrong.
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:
The RCT with pre-specified end points is considered the best path to scientific truth in medical matters. Randomization is the best method of controlling for known and unknown confounders. Blinding, often part of the process, minimizes bias. The logistics of RCTs are more difficult for study of lifestyle changes than for pharmacologic or procedural interventions. We have no RCTs of moderate alcohol drinking with CAD or other fatal event end points. For ethical reasons the effects of heavy drinking are not amenable to an RCT, but the wish for such studies of chronic disease effects of moderate drinking is often expressed. Generally there has been little discussion of practical considerations. The hypothesized fractional benefits would require large numbers in a costly multicenter trial of long duration. Even assuming such an effort, could a representative study group, after exclusions, be recruited? If an appropriate population was acquired, could compliance with randomization to daily/almost daily moderate drinking or none be maintained for years? Is blinding possible? What alcohol dose(s) should be used? How many arms are needed (e.g., beer, liquor, white wine, red wine, alcohol-water mixture, placebo)? These are formidable problems. We do have studies relevant to intermediate “surrogate” markers, such as high-density lipoprotein cholesterol, antithrombotic effects, and endothelial function. There is also promise in “natural’ experimental randomization by metabolic genetic polymorphisms related to alcohol metabolism. Unfortunately, it is likely that we will be left much-dependent upon observational epidemiology.
My roommate recently sent me a review article that LW might find interesting:
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.