Alcohol Is Healthy

By Robin Hanson · September 2, 2010 1:00 pm · Comments (49) · ·

My health econ students love to hear that our best data suggests moderate drinking is good for health.  New better data continues that trend:

[We study] the association between alcohol consumption and all-cause mortality over 20 years among 1,824 older adults…. Controlling only for age and gender, compared to moderate drinkers, abstainers had a more than 2 times increased mortality risk, heavy drinkers had 70% increased risk, and light drinkers had 23% increased risk. A model controlling for former problem drinking status, existing health problems, and key sociodemographic and social-behavioral factors, as well as for age and gender, substantially reduced the mortality effect for abstainers compared to moderate drinkers. However, even after adjusting for all covariates, abstainers and heavy drinkers continued to show increased mortality risks of 51 and 45%, respectively, compared to moderate drinkers. (more; HT yahoo, J Storrs Hall)

Controls included: Age, Gender, SES, Marital status, Gender x Marital status, Former Problem drinker, Health problems, Obesity, Smoking status, Physical activity, Physical activity x Time, Depressive symptoms, Avoidance coping, Gender x Avoidance coping, # Close friends, Quality of support.

Eric Crampton reviews the literature here, and comes to the same conclusion.

The US FDA prohibits the alcohol industry from advertising these studies, showing the health benefits of alcohol, because the public might get the wrong impression. You do not have free speech hearing in the US regarding health.

GD Star Rating
loading...
Tagged as:
Trackback URL:

4 Trackbacks

  1. [...] [1] Holahan CJ et al. Late-Life Alcohol Consumption and 20-Year Mortality. Alcohol Clin Exp Res. 2010 Aug 24. [Epub ahead of print] http://pmid.us/20735372. Full text: http://onlinelibrary.wiley.com/doi/10.1111/j.1530-0277.2010.01286.x/pdf. (Hat tip:  Robin Hanson, http://www.overcomingbias.com/2010/09/alcohol-is-healthy.html) [...]

    By Perfect Health Diet » Is It Smart to Drink? September 9, 2010 at 12:57 pm

  2. [...] to one or two cans a night. Those who want to binge can do so just as easily without the cans. But the facts don’t seem to [...]

    By Overcoming Bias : Naked Classism November 17, 2010 at 11:04 pm

  3. [...] we don’t stop short as much with for-profit corporations. For example, we won’t let alchohol makers advertize the fact that most research finds those who drink more are healthier. But [...]

    By Overcoming Bias : Why Not Censor? January 10, 2011 at 11:23 am

  4. [...] Results sec­tion is the one worth perus­ing. For those with­out full access to the study (ahem), Over­com­ing Bias pro­vides the full list of con­trols. If you’re new here, you may want to sub­scribe to my posts by email or through the RSS feed [...]

    By Drinking Levels and Mortality Rates – Lone Gunman July 12, 2011 at 2:27 am

45 Comments

  1. Lemmy Caution
    September 1, 2010 at 8:20 pm | Reply

    I started drinking more after that moderate drinking article came out.

    • Mary Page T.
      September 4, 2010 at 12:40 am | Reply

      This study, however, if you read the whole thing, goes on to say that most of the abstainers were alcoholics who had already done the I damage to their bodies with alcohol that they went on to die from later. I am a recovering alcoholic and believe me, I used to feel the same way, that alcohol was not the problem, by any means. Today I am suffering from several diseases that are debilitating, turning my life, at 51, into what I expected it to be at 85 I never dreamed, through 32 years of drinking, that this could ever happen to me. It was only the last six years that were pure hell, I considered the previous 26 years a success. How wrong I was, I’d gve anything for those years back and the ones I;m losing due to alcohol.

      • Mary Page T.
        September 4, 2010 at 12:53 am |

        And by the way, to underscore my comment I will share with you that my husband got sober (abstaining from alcohol for 25 years), relapsed and was dead within six years from alcohol. It is cunning and baffling, do not let it fool you, too.

      • Robin Hanson
        September 4, 2010 at 7:26 am |

        You can’t have it both ways. You say “abstainers were alcoholics who had already done the I damage to their bodies” but then suggest your husband died quickly from a “relapse.” But if so then choosing to drink does have health implications, even if one had been drinking lots before.

      • Nosayingmynameontheinternet
        September 6, 2010 at 7:39 pm |

        Are you sure that that was a factor in the study, though? It says in the third paragraph that that was taken into account and wasn’t a variable (“Former Problem drinker”).

      • Sara Kessell
        September 27, 2010 at 1:40 pm |

        Not only anecdotal, but completely irrelevant to the data being discussed. The study _specifically_ controlled for “former problem drinkers”, and made clear that “heavy drinkers” did not share the benefits described.

        Cardiovascular conditioning is arguably good for most healthy adults, but no one would recommend it for someone dying of congestive heart failure. _You_ can’t handle drinking. It doesn’t necessarily follow that no one else can either. Sulfa drugs are a ‘good thing’, but if they cause a severe allergic reaction, I suggest avoiding them too.

        As much as I hate to say it, this may be an example of the public having too little understanding of too much data.

  2. Constant
    September 1, 2010 at 8:35 pm | Reply

    1) Does this show that it is the alcohol content and not, say, the polyphenols? In other words, do vodka drinkers benefit, or are the red wine drinkers bringing up the average?

    2) As a point of comparison, to help understand how significant is the benefit, how does this compare to the unhealthiness of sitting for several hours a day, which has been in the news?

    • Eric Crampton
      September 2, 2010 at 5:36 pm | Reply

      Differences across types of alcohol are relatively small. Corrao’s 2000 metastudy has a minor review of that literature and says no particular differences, though there’s some evidence of a deeper J curve for folks in the Mediterranean.

      It has been suggested that wine, especially red wine has a superior protective effect than other alcoholic beverages.^’”^^ Conversely, a recent review of epidemiological literature provides evidence that a substantial portion of the benefits in the reduction of CHD is from alcohol rather than from other ingredients in wine.^” We were not able to test the hypothesis that the major protective effect observed in Mediterranean countries is due to the fact that Southern European populations consume alcohol mainly in form of wine. Among the included studies, only 11 reported beverage-specific relative
      risks,«,49,5t.53,54,5<.,6I.65,67.70,95

      Studies reported a dose-response relationship between quantity of alcoholic beverages and the risk of CHD. This observation is also in agreement with the fact that the hypothesized major protective effect of wine is weak."

  3. Silas Barta
    September 2, 2010 at 1:16 pm | Reply

    So this establishes that the relationship is causal. We should be trying to identify the causal mechanism that’s actually responsible for making lcohol prolong life. But then, what is it about alcohol as opposed to other consumables that can be responsible for this? Biochemically, what other substances resemble alcohol?

    If we accept that there is such a causal mechanism, then we should accept that those substances will have the same prolonging effects. But if we don’t accept those prolonging effects, we would have to reverse the conclusion about alcohol.

    For example, if alcohol does the same thing to the body as drinking e.g. a sugar-benzene mix, would we have to conclude that, “okay, the alcohol thing is a spurious correlation”, or would we start saying “benzene is good for you”?

  4. jz
    September 2, 2010 at 1:27 pm | Reply

    Perhaps the character of people who can demonstrate moderation, translates to their decisions in other realms, such as marriage management, stress avoidance, financial management, spending, recreation, eating, (yet other factors not controlled for).

    The physiology is alcohol is well studied, and I know of no physiological benefits of alcohol.

    • James
      September 2, 2010 at 1:41 pm | Reply

      Some possible physiological benefits of alcohol are discussed here.

    • Silas Barta
      September 2, 2010 at 3:53 pm | Reply

      The moderation thesis is very compelling for me. What have they done, if anything, to rule it out?

      In my experience, there are idiots who can’t take care of themselves (on the low-Maslow needs), and non-idiots who can. Driving is safe, and can be done safely with a cell-phone (so long as you stop doing so during difficult maneuvers) — but the idiot/anti-social folks aren’t capable of exercising good judgment in doing so.

      How have I passed through the most accident-prone demographic (young male) for 11 years without a single collision (of *anyone*’s fault), while driving aggressively 30 miles a day on freeways? (And one collision in my entire history that just made a small dent.) Because I’m not an idiot and know how to exercise caution.

      Some folks can’t be trusted around alcohol because they just abuse it. They lead dumb lifestyles that get them killed sooner. Some folks are a little better, and recognize they’re too dumb to handle alcohol in moderation and so avoid it entirely. But the smartest and most careful can handle it. (I didn’t see IQ in the controlled-for list.)

      How’s that for an offensive but plausible theory no one will get funding to test?

      • Vladimir M.
        September 2, 2010 at 4:45 pm |

        Silas Barta:

        I didn’t see IQ in the controlled-for list.

        Now that you’ve mentioned that, I remembered this striking finding I saw recently:
        http://www.marginalrevolution.com/marginalrevolution/2010/07/beer-makes-bud-wiser.html

        This could indeed go a long way towards explaining the findings from the cited paper. It seems quite plausible that the authors didn’t control for some additional life-expectancy-enhancing personal traits that these test scores (and thus also drinking) correlate with.

      • Eric Crampton
        September 2, 2010 at 5:43 pm |

        It’s plausible that “healthy behaviours” are a confound. But Castelnuovo and Donati’s 2006 metastudy basically ruled that out. Models with full controls for health behaviours compared to those with no controls reduced the advantage of moderate drinking from about a 19% relative risk reduction to about a 16% relative risk reduction. If controlling for all the observables only knocks you back that small amount, it’s pretty implausible that the unobservables are having an effect five times greater – and that’s what would be needed to wipe out the J-curve.

    • Eric Crampton
      September 2, 2010 at 5:39 pm | Reply

      Coronary heart disease, ischemic stroke. Those are the big ones. Massive evidence in favour of alcohol’s benefits.

      For a first cut check of what things alcohol helps, check the etiological attributable fractions reported in Collins and Lapsley, among other places. For some disorders, alcohol increases mortality. For others, it reduces it.

  5. gwern
    September 2, 2010 at 1:45 pm | Reply

    Well, I suppose this isn’t really news (as you say it’s just *better* data), but it still leaves open why there’s any effect.

    What troubles me is that aside from hormesis (=alcohol is weakly poisonous at moderate consumption, triggering an overreaction), the only evolutionary connection that comes to mind is that regular alcohol consumption signals regular access to (fermenting) fruits, signaling ample resources & allies & general success in life… which by the caloric restriction/hormesis theory ought to lead to shorter lifespans with more investment in children.

    Hm. I suppose we could actually try to reverse that; placebo medicine helps because it signals allies and a valuable future, letting the body make long-term investments in health; perhaps by the same logic alcohol and spices signal ample food (as before), signaling a valuable future, letting the body etc.

  6. OhioStater
    September 2, 2010 at 2:53 pm | Reply

    I’m tailgating the Ohio State – Marshall game tonight. What will 15 beers do for my health?

  7. andrew fischer lees
    September 2, 2010 at 3:41 pm | Reply

    Oh c’mon, no controls for mortality by genetic history? Plenty of mortality-related intermediate outcomes (as well as incidence of alcoholism) seem to have a genetic component. This inferred causality might be a multicolinearity problem.

    Though on net, I think that jz’s comment is the most salient

  8. Aria Stewart
    September 2, 2010 at 4:01 pm | Reply

    I, for one, am kinda glad that individuals enjoy free speech but businesses do not. But it’s still a slippery slope.

  9. Hrm
    September 2, 2010 at 6:16 pm | Reply

    “One important reason is that alcohol lubricates so many social interactions, and social interactions are vital for maintaining mental and physical health. As I pointed out last year, nondrinkers show greater signs of depression than those who allow themselves to join the party.”

    If we accept this statement from the Yahoo! article, how can we reconcile it with the controls for “# Close friends?” Is it the quality of their social interactions that’s better?

    I’ve never drank (or used any other recreational substances), so I can’t comment from personal experience.

  10. Bock
    September 2, 2010 at 6:29 pm | Reply

    We need more tests on hallucinogens, which for some reason are illegal. If I pick up a naturally growing mushroom on my farm I am breaking the law.

  11. Ryan S
    September 2, 2010 at 7:54 pm | Reply

    My first guess would be that the effect is due to reduced stress. Or would this effect already be included in one of their controls? (Or, If the stress-relief effect only lasts while intoxicated, and the control measurements were taken while sober, that could explain it.)

  12. Roland
    September 2, 2010 at 8:48 pm | Reply

    I like alcohol yet I never drink for fear of becoming an alcoholic. Is it worth the risk? What are the health benefits opposed to the probability of becoming a problem drinker?

  13. Wei Dai
    September 3, 2010 at 2:11 am | Reply

    Many East Asians have either ADH1B*47His [1] (which metabolize ethanol into the more toxic acetaldehyde up to 100 times quicker than normal) or ALDH2 Lys487 [2] (which slows down the metabolism of acetaldehyde into non-toxic acetate.

    Either of these genes is sufficient to cause the alcohol flush reaction, where a buildup of acetaldehyde causes the drinker’s face to turn bright red and makes him or her quite uncomfortable in other ways. These genes seem to have no adaptive function, except to discourage their carriers from drinking alcohol. It has also been shown that ADH1B*47His’s spread in Asia followed the spread of rice domestication and fermentation.

    These facts suggest that alcohol is unhealthy or at least reduced genetic fitness for East Asians in the last several thousand years. However I’m not sure how to weight this evidence against the studies that Robin cites.

    [1] The ADH1B Arg47His polymorphism in East Asian populations and expansion of rice domestication in history
    [2] The Alcohol Flushing Response: An Unrecognized Risk Factor for Esophageal Cancer from Alcohol Consumption

    • Eric Crampton
      September 3, 2010 at 5:24 am | Reply

      Wei: At what level of consumption does that effect really start kicking in? I’d read it as a mechanism to prevent excessive drinking rather than to prevent any drinking. Is that wrong?

      • Wei Dai
        September 3, 2010 at 5:48 am |

        From personal experience, I know that one drink is more than enough to trigger the effect. It pretty much prevents any drinking. (There are some people have this reaction but keep drinking anyway due to social pressure. According to the second paper I cited, they end up with a much higher risk for esophageal cancer even if they only drink moderately.)

    • Stuart Armstrong
      September 3, 2010 at 6:15 am | Reply

      Interesting, but I don’t feel the selective pressure from, say, oesophageal cancer would be enough to explain this (cancers develop late in life). Are these genes connected with others that would be beneficial for rice consumption? Or were early east Asian fermented rice drinks very unhealthy for other reasons? (impurities?)

    • Wei Dai
      September 4, 2010 at 6:44 am | Reply

      Some more thoughts on this topic.

      The fact that the genes for alcohol flush reaction is so widespread shows that drinking alcohol almost certainly decreased genetic fitness at least for East Asians for a significant period of time. (The idea is that those with the alcohol flush reaction tend to refrain from drinking, and since the genes have no other apparent adaptive function, it must be the non-drinking that increased their genetic fitness relative to drinkers.)

      But that doesn’t necessarily imply that it’s unhealthy to drink alcohol now. For example, the decrease in genetic fitness may have been caused by the extra money that drinkers spent buying alcohol, and not by any health effects. Or it may be that changes in genes or the environment has made alcohol healthier than it used to be.

      What I’d like to see is a study comparing the current or recent health/mortality of people with the genes for alcohol flush reaction (and do not drink because of it) with their siblings who do not have the genes (and do drink). That would constitute a near-perfect natural experiment to test the hypothesis that alcohol is healthy, with randomization between the treatment group and the control group provided by Mendelian inheritance.

  14. Kim Øyhus
    September 3, 2010 at 4:05 am | Reply

    So, the teeetotalers are insanely evil again.
    The previous horrible time was when they outlawed alcohol as an anti knock additive to gasoline, so they used methyl led instead, spewing out megatonnes of this stuff everywhere, poisoning everything.

  15. Eric Crampton
    September 3, 2010 at 6:09 am | Reply

    I’d want numbers on all-source mortality rather than oesophageal cancer – alcohol increases the risk of a few cancers, but those risks are offset at moderate consumption by reduced risks of other things like coronary heart disease.

  16. Stuart Armstrong
    September 3, 2010 at 6:10 am | Reply

    The US FDA prohibits the alcohol industry from advertising these studies, showing the health benefits of alcohol, because the public might get the wrong impression. You do not have free speech in the US regarding health.

    Beyond the issue of free speech, is there evidence that the public would get the wrong impression (and how wrong)? I can make easy arguments either way.

  17. clayton
    September 3, 2010 at 9:22 am | Reply

    I do have free speech in the U.S. regarding health: “moderate alcohol consumption is good for your health.” No one blocked me from saying that. I will not be arrested. The FDA will not interfere with this comment. No one blocked the blogger from saying the same thing.

    Why do we believe that corporations should have free speech (or unions or any other legal bureaucracy)? (Recent U.S. Supreme Court decisions aside.) I read nothing in the Constitution that says they should (or shouldn’t). Corporation owners already have free speech. Why does the corporation also require it independently of that right already possessed by its owners?

  18. Floccina
    September 3, 2010 at 11:01 am | Reply

    Free speech by corporations is the speech of the owners and mangers of the corporation.

    Are those who are speaking against free speech for Corporations OK with speech by partnerships?

  19. Nancy Lebovitz
    September 3, 2010 at 3:35 pm | Reply

    Another hypothesis– people who hate the taste of alcohol may be supertasters (detest bitter flavors). Maybe part of the health effect is that people who can enjoy alcohol are also apt to eat more dark green veggies and grapefruit? Or maybe some of the good effects are actually from coffee.

  20. Summerspeaker
    September 4, 2010 at 6:35 pm | Reply

    Saying they controlled for age is somewhat misleading. All the folks involved were in the 55-65 range. As such, the study doesn’t necessarily tell you anything about the lifetime effect of alcohol consumption. For example, some heavy drinkers don’t even make it to 55.

  21. John Jenkins
    September 5, 2010 at 11:21 am | Reply

    The people with the best data sets on mortality (insurance companies), offer discounts on premiums if the insured drinks some amount X or less per day. Since the insurance companies actually have money riding on these transactions, why wouldn’t they set the discount for some amount between X and Y, with no discount above or below?

    You could argue that the insurance companies are simply making a mistake based on this data, but it’s a reasonable inference that they do know what they are doing (they stay in business, after all), and the actual risk of health issues (not just death) varies with consumption above a certain level. I find it a little odd that they ONLY think they look at is chance of death over a 20 year period. What about other ill effects of alcohol consumption?

    I think jk has the best thoughts. Epidemiological research is great, but unless you can show me a mechanism for why something is true, I am going to think you missed something.

    @Floccina: Most people decrying speech by corporations couldn’t give an adequate definition of what a corporation is, so the distinction will be lost on them.

  22. Arunas
    September 5, 2010 at 4:30 pm | Reply

    Has anyone out here had an alcoholic in the family? Has he/she started by drinking a liter of whiskey a day? Stuck ‘up the benefits’ of moderate consumption to where the sun does not shine! You cant talk about this in isolation. As humans and as far as alcohol is concerned we cant not drink moderately and usually we find this out when it’s too late. Has anyone thought of what disastrous effects these ‘studies’ might have when popularized globally? Actually i was looking for info on priming and stumbled upon this site and found this article… but I’ve seen people discussing it somewhere else before…it spreads.. and I’m not sure if the authors have thought about the ramifications and ethical aspects of it…all they know is how to play with statistics and sure want some attention. well – you got that… go to hell

  23. james wilson
    September 6, 2010 at 12:58 pm | Reply

    We can easily make a case that alcohol has almost no positive effect on health whatever, that it is, rather, the type of individual who easily moderates alcohol use that is the key to superior mortality statistics.
    The same could be said of eating.

  24. Peter Gerdes
    September 8, 2010 at 5:44 am | Reply

    Hmm, I wonder if the precedent in Citizens United could be used to challenge such a ban. I suspect the answer is no with respect to outright ads since they are commercial speech but it’s an interesting question whether a beer company could fund ‘public service’ messages not mentioning their product that conveyed the conclusions of these studies.

  25. erik
    November 8, 2010 at 2:41 am | Reply

    you and this study are making an obvious reasoning error – concluding you know the direction of causality. i propose that personalities that tend to moderation are healthier than extremist personalities.

  26. Zach
    November 18, 2010 at 10:27 am | Reply

    How do the authors define “light” “moderate” and “heavy drinking”

    The study does me no good if I can’t classify my own drinking habits in the same manner as the authors.

Post a Comment

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>