Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

“Health benefits” of moderate drinking disappear when controlling for wealth

theconversation.com

Post reply on HN
Print view View original

121–130 of 134 posts

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#121

Post by Retric »

mortenjorck wrote: post #2

The headline looks like another entry in the endless "x is good for you / x is bad for you" genre of science journalism, but the article is actually a valuable object lesson in the complexities of trying to control for variables in any study. It's also a timely reminder of how socioeconomic factors remain one of the biggest predictors of health outcomes.

It's really not based on a lot of data though. 2,908 people for 10 years while removing several groups and apparent health benefit is much reduced in women So, it still links drinking with health benefits even after splitting those people into a lot of small groups.

If anything it demonstrates how a bias may influence study design and reporting.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#123

Post by mark_l_watson »

I have no scientific basis for this, this is just an idea: all alcohol consumption might be bad for our health, but moderate alcohol might relax people enough to make up for the little harm done from consumption.

I would be interested to see a study that compared the health of moderate drinkers with similar people who didn't drink but did some meditation, yoga, or anything else that makes people more relaxed.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#124

Post by RationPhantoms »

ramblenode wrote: post #115

I just stopped reading after this sentence: > We split the adults aged 50 or older into groups reflecting their average level of drinking, from non-drinkers through to those drinking three or more drinks a day. The decision to split a naturally continuous or interval variable into discrete groups and then compare those groups has no valid statistical purpose. It serves no advantage over the regression alternative but…

And here I am, not believing their statistical prowess because their pages don't use HTTPS

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#125

Post by justbaker »

RationPhantoms wrote: post #124
ramblenode wrote: post #115

I just stopped reading after this sentence: > We split the adults aged 50 or older into groups reflecting their average level of drinking, from non-drinkers through to those drinking three or more drinks a day. The decision to split a naturally continuous or interval variable into discrete groups and then compare those groups has no valid statistical purpose. It serves no advantage over the regression alternative but…

And here I am, not believing their statistical prowess because their pages don't use HTTPS

This comment is so securely crafted sir.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#126

Post by ibejoeb »

lkesteloot wrote: post #119

I had also read elsewhere that the worse health of non-drinkers (relative to moderate drinkers) could be explained by the fact that many non-drinkers are recovering alcoholics, and had already trashed their bodies pretty badly. A study (can't find link) controlled for this and found that non-drinkers are healthier than moderate drinkers, even without controlling for wealth.

This study has accounted for what they call "sick quitters" and those lifelong abstainers. This is entertaining, to me, because it seems that they have arrived at exactly the opposite conclusion: when controlling for overall intake, and then controlling for wealth, moderate drinkers fare better than non-drinkers.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#127

Post by marchenko »

mark_l_watson wrote: post #123

I have no scientific basis for this, this is just an idea: all alcohol consumption might be bad for our health, but moderate alcohol might relax people enough to make up for the little harm done from consumption. I would be interested to see a study that compared the health of moderate drinkers with similar people who didn't drink but did some meditation, yoga, or anything else that makes people more relaxed.

Social correlates probably play a significant role. Moderate drinking may be correlated with sociality and preparing meals for one's family/ friends. Strong social ties and attention to diet are both correlated with good health outcomes. I think we see a hint of these lifestyle correlates in the findings that red wine is the best vehicle for alcohol consumption. The alcohol itself might be mildly harmful, but for people who are able to moderate their intake, the pleasure of imbibing and its associated rituals more than makes up for it.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#128

Post by DanBC »

Retric wrote: post #121
mortenjorck wrote: post #2

The headline looks like another entry in the endless "x is good for you / x is bad for you" genre of science journalism, but the article is actually a valuable object lesson in the complexities of trying to control for variables in any study. It's also a timely reminder of how socioeconomic factors remain one of the biggest predictors of health outcomes.

It's really not based on a lot of data though. 2,908 people for 10 years while removing several groups and apparent health benefit is much reduced in women So, it still links drinking with health benefits even after splitting those people into a lot of small groups. If anything it demonstrates how a bias may influence study design and reporting.

No!

For some time now we've thought that moderate drinking does not provide any health advantages[1] except to older women, and that it provides a small health advantage to those women.

This research just tells us that moderate drinking isn't useful to older women either.

[1] there are some advantages but these are outweighed by the disadvantages.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#129

Post by kochthesecond »

jimrandomh wrote: post #29

A lot of research has shown supposed health benefits of alcohol, but it all has something in common: it's based entirely on observational data, with no intervention performed. Observational studies have an extremely poor track record in health and especially in nutrition, and are not experiments, but they get reported as though they were science. So we get reversals like this one. (Which was entirely expected.)

Feynman actually has a good piece on "science" like this.

Re: “Health benefits” of moderate drinking disappear when controlling for wealth

#130

Post by Retric »

DanBC wrote: post #128
Retric wrote: post #121

Earlier quoted context omitted.

It's really not based on a lot of data though. 2,908 people for 10 years while removing several groups and apparent health benefit is much reduced in women So, it still links drinking with health benefits even after splitting those people into a lot of small groups. If anything it demonstrates how a bias may influence study design and reporting.

No! For some time now we've thought that moderate drinking does not provide any health advantages[1] except to older women, and that it provides a small health advantage to those women. This research just tells us that moderate drinking isn't useful to older women either. [1] there are some advantages but these are outweighed by the disadvantages.

This research really does not provide significant evidence for that idea. I realize your preconceived ideas align with this idea, but there is no evidence of harm from this study!

It also does show evidence of heath benefit for women. It's likely that men also benifit though to a lesser degree, which is to small to show in this data. Remember westerners had daily alcohol intake for a very long time. Much like Milk it's negative effects on that population are likely reduced.

PS: This is also analyzing 3rd party data for heath effects which is a very low order of evidence. https://xkcd.com/882/

Post reply on HN