Association of Coffee Drinking with Total and Cause-Specific Mortality

December 25, 2012 4:08 AM

Association of Coffee Drinking with Total and Cause-Specific Mortality [FULL TEXT HTML]: "We used data from a very large study, the National Institutes of Health (NIH)–AARP Diet and Health Study (ClinicalTrials.gov number, NCT00340015), to determine whether coffee consumption is associated with total or cause-specific mortality. The current analysis, involving more than 400,000 participants and 52,000 deaths, had ample power to detect even modest associations and allowed for subgroup analyses according to important baseline factors, including the presence or absence of adiposity and diabetes, as well as cigarette-smoking status."

"Coffee is one of the most widely consumed beverages, both in the United States and worldwide. Since coffee contains caffeine, a stimulant, coffee drinking is not generally considered to be part of a healthy lifestyle. However, coffee is a rich source of antioxidants and other bioactive compounds, and studies have shown inverse associations between coffee consumption and serum biomarkers of inflammation and insulin resistance.

Considerable attention has been focused on the possibility that coffee may increase the risk of heart disease, particularly since drinking coffee has been associated with increased low-density lipoprotein cholesterol levels and short-term increases in blood pressure. Results from a number of studies have been inconsistent. The heterogeneous findings may be due to differences between case–control and prospective study designs and possibly also to inconsistent control for important confounders such as tobacco smoking. In addition, the numbers of deaths have been small in most studies. Cohort studies do not support a positive association between coffee drinking and mortality, however, and some even suggest a modest inverse association.

Previous studies have also investigated the association between coffee consumption and other major causes of death, and they have shown inverse associations with diabetes, inflammatory diseases, stroke, and injuries and accidents, although associations with cancer have generally been null. The results of studies of coffee consumption and total mortality have been mixed, with associations that have been consistent with either the null hypothesis or a modest inverse effect. Data are lacking to clarify the association between coffee drinking and mortality, to determine whether there is a dose–response relationship, and to assess whether associations are consistent across various subgroups."

posted by Blasdelb (76 comments total)

summary-
We noted inverse associations between coffee drinking and most major causes of death, with the exception of cancer.
...a modest borderline positive association was observed in men for coffee consumption and mortality from cancer, with a null association observed in women.
posted by bhnyc at 4:20 AM on December 25, 2012 [3 favorites]

I love the java jive and it loves me.
posted by seanmpuckett at 4:29 AM on December 25, 2012 [4 favorites]

Can someone tl dr this for me? I need to know if I should finish this cup.
posted by empath at 4:40 AM on December 25, 2012 [8 favorites]

Can someone tl dr this for me?

It's pretty clear that coffee does or does not have something to do with health.
posted by twoleftfeet at 4:42 AM on December 25, 2012 [69 favorites]

It's amazing how badly some people seem to want coffee to be proven harmful. After all, it is pleasurable!
posted by thelonius at 4:43 AM on December 25, 2012 [42 favorites]

Can someone tl dr this for me? I need to know if I should finish this cup.

People who drink coffee die of something eventually.
posted by dortmunder at 4:48 AM on December 25, 2012 [18 favorites]

However, coffee drinkers were also more likely to smoke, and, after adjustment for tobacco-smoking status and other potential confounders, there was a significant inverse association between coffee consumption and mortality.
...
Inverse associations were observed for deaths due to heart disease, respiratory disease, stroke, injuries and accidents, diabetes, and infections, but not for deaths due to cancer.

So...if I'm reading that right, coffee may be good for you. Or at least, it correlates inversely with harmful effects WHICH I WILL HAPPILY INTERPRET AS POSITIVE RESULTS THANKS WHY YES I'LL HAVE ANOTHER CUP PLEASE
posted by dubitable at 4:49 AM on December 25, 2012 [2 favorites]

The only way to safely drink coffee is to use beans that have been excreted from the anuses of the Asian Palm Civet. And only if you imbibe it via an enema.

It's actually quite popular at some coffee houses now. Ask for the "enema room" (or, in some cities, the "output to output room".)
posted by twoleftfeet at 4:53 AM on December 25, 2012 [1 favorite]

I read this as "morality" about 100 times and thought "No wonder I'm such a jerk!"
posted by Potomac Avenue at 4:53 AM on December 25, 2012 [14 favorites]

Tl dr = "is there an association between coffee and death? The data are lacking. Here s our overinterpretation of crap incidental data scraped from other studies, because we can"
posted by eustatic at 4:55 AM on December 25, 2012 [1 favorite]

Does this mean that I can continue to have my coffee standing outside with the cigarette?
posted by infini at 5:02 AM on December 25, 2012

Does this mean that I can continue to have my coffee standing outside with the cigarette?

It's the fresh air that'll kill ya!
posted by fairmettle at 5:03 AM on December 25, 2012 [2 favorites]

Can someone tl dr this for me? I need to know if I should finish this cup.

People who drink coffee die of something eventually.

I had the exact same question. Now that I know the answer, I am going to drive to Starbucks, use the gift card that I got, order a quadruple espresso something and go for a drive.

Because studies have shown there is nothing safer than drinking coffee while driving.
posted by lampshade at 5:05 AM on December 25, 2012

If you drink espresso and steamed milk you will achieve immorta-latte.
posted by twoleftfeet at 5:07 AM on December 25, 2012 [30 favorites]

Now, here's a question - an inverse correlation with all causes except cancer. Does that mean a positive correlation with cancer, or just there is no net change with cancer?
posted by Kid Charlemagne at 5:12 AM on December 25, 2012

No correlation with cancer. " associations with cancer have generally been null."
{getting a second cup now . . .}
posted by Obscure Reference at 5:38 AM on December 25, 2012 [2 favorites]

It's amazing how badly some people seem to want coffee to be proven harmful. After all, it is pleasurable!

Please substantiate your assertion. Anything as widely-used, effective, and as a part of our culture as coffee needs to be studied. Your statement implies that, if they had found substantial harm, you would have discarded it as naysayers and killjoys trying to wreck things for everyone, much like how some people felt about investigation into the dangers of tobacco smoke.
posted by JHarris at 5:52 AM on December 25, 2012 [4 favorites]

As a long time asthma sufferer, I can testify to the inverse impact of coffee drinking on airway constriction. A really good triple ristretto, is particularly helpful.
posted by Abinadab at 5:54 AM on December 25, 2012 [1 favorite]

I say we ban all coffeemakers with a capacity larger than two cups. Why do you need to make more coffee than that?
posted by laconic skeuomorph at 6:17 AM on December 25, 2012 [1 favorite]

Does this mean that I can continue to have my coffee standing outside with the cigarette?
It's the fresh air that'll kill ya!

Although I'm not a smoker, it's my hypothesis that with all the toxic office buildings the folks that go outside regularly for 5-10 minutes will have slightly reduced mortality indicators.
posted by sammyo at 6:19 AM on December 25, 2012 [1 favorite]

laconic skeuomorph: "I say we ban all coffeemakers with a capacity larger than two cups. Why do you need to make more coffee than that?"

They already tried this, oh how they tried...
posted by goodnewsfortheinsane at 6:22 AM on December 25, 2012

By which I mean to say that in the Netherlands at least, Senseo machines (which do in fact serve up no more than two cups at a time) have become so ubiquitous that I am considered a (reverse?) snob in some circles for having the gall of preferring simple, honest drip brew.
posted by goodnewsfortheinsane at 6:25 AM on December 25, 2012 [1 favorite]

Please substantiate your assertion

What assertion? That's an impression I have formed over the years. You can disagree with it, that's fine. I don't do a peer-reviewed study before I write a comment on a message board.
posted by thelonius at 6:28 AM on December 25, 2012 [12 favorites]

"After multivariate adjustment, coffee appeared to be inversely associated with most major causes of death in both men and women, including heart disease, respiratory disease, stroke, injuries and accidents, diabetes, and infections."

tl;dr: Coffee prevents accidents - and injuries. This statement is equally true of any other health benefit you could ascribe to coffee.

Time for another cup.
posted by grajohnt at 6:32 AM on December 25, 2012

I say we ban all coffeemakers with a capacity larger than two cups. Why do you need to make more coffee than that?


You go to hell, sir. You go right to hell.
posted by TheWhiteSkull at 6:32 AM on December 25, 2012 [18 favorites]

*tsks* peace on earth joy to world remember? *slurp* *puff*
posted by infini at 6:35 AM on December 25, 2012 [1 favorite]

On a side note, since this is an FPP on coffee drinking habits...

I read in a PD James novel recently something to the effect of "He wondered if he should offer her some coffee, but it was too early in the morning"... it was 1030am... this is set in the UK. What would be the rationale here for this?
posted by infini at 6:37 AM on December 25, 2012

goodnewsfortheinsane, gawd, don't get me started. "Ah yes, American coffee." Which I understand to mean, "Ah yes, a beverage brewed from beaver urine and burnt acorns."

Trust me mijn vrienden, whatever advantages you think that thimble-sized cup of foul muck your Senseo machine spits out may have, flavor and caffeine content are not among them.
posted by 1adam12 at 6:44 AM on December 25, 2012 [2 favorites]

The only way to safely drink coffee is to use beans that have been excreted from the anuses of the Asian Palm Civet. And only if you imbibe it via an enema.

))<>((
posted by sourwookie at 6:47 AM on December 25, 2012 [5 favorites]

What assertion?

It is an assertion regardless of its originating impulse. I was trying to give you an opportunity to elaborate upon it.
posted by JHarris at 6:48 AM on December 25, 2012

WHERE IS THE STUDY ON FIVE HOUR ENERGY I HOPE IT IS GOOD FOR YOU AND NOT CAUSING BRAIN PROBLEMS
posted by orme at 6:57 AM on December 25, 2012 [15 favorites]

Yeah, I'll keep drinking coffee and tea, and wondering how in the world people are brave enough to drink a cocktail of inscrutability in a tiny, opaque plastic bottle.
posted by sonic meat machine at 7:06 AM on December 25, 2012 [1 favorite]

Taking coffee away from me would destroy my will to live another day. Did they factor that into the study?
posted by octothorpe at 7:13 AM on December 25, 2012 [6 favorites]

It is an assertion regardless of its originating impulse. I was trying to give you an opportunity to elaborate upon it.

Somebody hasn't had their cup of joe this morning.
posted by Celsius1414 at 7:16 AM on December 25, 2012 [1 favorite]

I really really dislike these kinds of studies because there is a monstrously huge self-selection issue. People who have liver problems will avoid too much coffee or drinking it all together because it can make them jittery as hell. So your coffee consumers are preselected for people with no issues metabolizing it and that messes up causal conclusions drawn by increasing the mortality among non-drinkers and decreasing it among the drinkers. I'm sure there are lots of other similar self-selection issues that are pretty much impossible to exclude the effects of.
posted by srboisvert at 7:17 AM on December 25, 2012 [2 favorites]

World Death Rate Holding Steady At 100 Percent
posted by The Card Cheat at 7:31 AM on December 25, 2012 [6 favorites]

By the way, I once DJ'd a radio program where I played -- no kidding -- 4 straight hours worth of country and western songs about coffee and/or coffee and cigarettes.

We require proof of this assertion. Still have that playlist anywhere? Because it sounds awesome.
posted by MonkeyToes at 7:31 AM on December 25, 2012 [5 favorites]

thanks spitbull, but I have some terrible news for you, I'll be 147 in March
posted by infini at 7:33 AM on December 25, 2012

I didn't read the study, did it consider the effects of sugar in coffee? Guess diabetes ahoy?
posted by elpapacito at 7:35 AM on December 25, 2012

modern age tldr; everything may prevent some diseases while causing others, so either consume nothing or ignore completely
posted by nathancaswell at 7:41 AM on December 25, 2012 [2 favorites]

The data are lacking. Here s our overinterpretation of crap incidental data scraped from other studies, because we can"

Meta-analyses like this one, when done properly (i.e. with proper statistical controls) are actually one of the most powerful research tools available. For those not familiar, the idea behind a meta-analysis is to gather information from the entire extant corpus of research on a subject, standardize it statistically (controlling for such things as variation in size, quality, and subject between different studies as well as confounders like smoking status, selection bias, genetic factors, and lifestyle choices -- all perfectly possible for a large meta-analysis like this one provided the appropriate data is present, itself a reasonable though by no means assured assumption) and then combine the aggregated data into a single dataset which has much greater statistical power than any of the component studies.

Done well (and I haven't read all of this study, but the idea seems perfectly reasonable and I trust Blasdelb not to post crap studies on the blue) this is an extremely effective technique. Meta-analysis is a large and growing field of research, especially in the medical research community where it has caused something of a revolution in the last several years.

So this isn't just an "overinterpretation of crap incidental data". Meta-analysis does have its pitfalls and must be undertaken with great care and skill, but it is a totally legitimate and respectable research technique when executed competently. And for those who actually want to know the results of this study, I'd recommend just clicking on the main link in this post (which goes to the study itself, a study which I'll add is commendably well-written and very accessible) and try reading the "discussion" section or at least the first paragraph of it.

For those who won't do even that, the takeaway is that there is a significant dosage-dependent effect (i.e. it gets stronger the more you drink) in which coffee drinking is correlated with lower overall mortality. This is when confounding factors (smoking is the biggest one) have been controlled for. So, bottom line: coffee is good for you.

This is probably by far the most powerful study on the subject to date, and the closest thing to a definitive answer we are likely to see assuming there are no major methodological flaws. I'm sure it will do nothing to resolve the media's ongoing ontological oncological project (the quest to figure out whether or not any given thing cures or causes cancer) but this study may still go a long ways toward actually providing an answer to whether coffee drinking is healthy or not.

That the media has done such a wonderful job of confusing the issue and undermining public confidence in science's ability to address this question does not invalidate the power of this analysis.
posted by Scientist at 8:05 AM on December 25, 2012 [29 favorites]

It's too early for me to overthink this cup of beans.
posted by rtha at 8:33 AM on December 25, 2012 [7 favorites]

who insists that coffee is an addiction and harmful for that reason alone

There's worse things from Columbia that you could get hooked on.
posted by radwolf76 at 8:58 AM on December 25, 2012 [1 favorite]

eustatic: Tl dr = "is there an association between coffee and death? The data are lacking. Here s our overinterpretation of crap incidental data scraped from other studies, because we can"

I think you are attacking this study for being a poorly done meta-analysis of multiple studies that were not designed to address this question using excessive statistical tests, but this is not a meta-analysis, it only considers data from a single study, that study was expressly designed for exactly this kind of analysis from the beginning, and Cox proportional-hazards regression models are both pretty damn conservative and much more powerful than Poisson model equivalents. The authors do all of the statistical tests necessary to show how lacking or otherwise the data are for the conclusions they draw for you and present them transparently. Indeed, all of the statements they make are backed by p-values less than 0.05, often much lower, and they even went through the trouble of nicely graphing out the 95% confidence intervals for everything they looked at to show that there wasn't any funny business going on.
posted by Blasdelb at 9:03 AM on December 25, 2012 [15 favorites]

There's worse things from Columbia that you could get hooked on.

Like patacones.

/homer drooling sounds
posted by Celsius1414 at 9:04 AM on December 25, 2012 [2 favorites]

That the media has done such a wonderful job of confusing the issue and undermining public confidence in science's ability to address this question

Now that's a topic that really begs for a really good deep meta-analysis of some kind.
posted by blucevalo at 9:06 AM on December 25, 2012 [1 favorite]

> I say we ban all coffeemakers with a capacity larger than two cups. Why do you need to make more coffee than that?

Sure! Limit blood transfusions to 4 ounces, too! Who needs more than that during open heart surgery?
posted by ardgedee at 9:10 AM on December 25, 2012

goodnewsfortheinsane: "By which I mean to say that in the Netherlands at least, Senseo machines (which do in fact serve up no more than two cups at a time) have become so ubiquitous that I am considered a (reverse?) snob in some circles for having the gall of preferring simple, honest drip brew."

Everyone looks at me funny when I ask for an Americano.
posted by Blasdelb at 9:12 AM on December 25, 2012

There's worse things from Columbia that you could get hooked on.

I know, right? Don't get me started on journalism degrees...
posted by dersins at 9:13 AM on December 25, 2012 [15 favorites]

> I say we ban all coffeemakers with a capacity larger than two cups. Why do you need to make more coffee than that?

I've found that most people can't make even two cups properly...
posted by mikelieman at 9:18 AM on December 25, 2012

The best ever is Lefty Frizzell's "Smoking Cigarettes and Drinking Coffee Blues," the original of which does not appear to be on YouTube.

Your YouTube Fu has failed you.
posted by PeterMcDermott at 9:34 AM on December 25, 2012 [3 favorites]

"I really really dislike these kinds of studies because there is a monstrously huge self-selection issue. People who have liver problems will avoid too much coffee or drinking it all together because it can make them jittery as hell. So your coffee consumers are preselected for people with no issues metabolizing it and that messes up causal conclusions drawn by increasing the mortality among non-drinkers and decreasing it among the drinkers. I'm sure there are lots of other similar self-selection issues that are pretty much impossible to exclude the effects of."

Reverse causality is a really big problem with observational studies like this one, but with a little bit of cleverness it is totally addressable, especially with the massive numbers of participants these authors are working with. While they did exclude participants who reported cancer or cardiovascular disease from the beginning, the survey did not ask about liver problems or self-reported ability to handle caffeine, which does indeed suck. However, the authors did find that the effects they saw were not only still present but even stronger among participants who reported themselves in excellent health at the beginning than for those who reported poor to fair health, which argues pretty strongly against the effect coming primarily from sick participants forgoing coffee.
posted by Blasdelb at 9:40 AM on December 25, 2012 [2 favorites]

Theres a few interesting things (to me) about this article, outside the research and the article content, specifically around the way it spread through social media:

-63% of the people who tweeted the article (and there were 1500+ of them), were the general public, rather than scientists
-This article spawned four separate reddit threads
-This article made Pinterest! How many NEJM authors can say that about their work?!

(data above from Altmetric)
posted by weaponsgradecarp at 9:41 AM on December 25, 2012 [2 favorites]

TL;DR = Tasty Liquid; Drink Repeatedly.
posted by benito.strauss at 9:44 AM on December 25, 2012 [8 favorites]

standardize it statistically (controlling for such things as variation in size, quality, and subject between different studies as well as confounders like smoking status, selection bias, genetic factors, and lifestyle choices -- all perfectly possible for a large meta-analysis like this one provided the appropriate data is present, itself a reasonable though by no means assured assumption) and then combine the aggregated data into a single dataset which has much greater statistical power than any of the component studies.

The thing is you simply cannot control for selection bias in nutritional studies. Simply having more data, all with similar selection bias problems, does not eradicate the problem. It exacerbates it by giving the conclusion an illusory weight.

The problem with metaanalytic studies is that to be effective they require considerable variability in methodologies with the hope that this variability cancels out the biases of the different methodologies. In this kind of nutritional research on mortality there just isn't sufficient methodological variability to cancel out the diet selection bias. There just isn't a method for dealing with diet selection factors that are correlated with mortality outcomes because you can't experimentally randomly manipulate the consumption of ubiquitous foods for entire lifetimes.

You might say they can adjust for it statistically but the adjustment, at least partially, if not entirely, presupposes the answer to the question.
posted by srboisvert at 9:46 AM on December 25, 2012 [1 favorite]

OK, so this is what I get for commenting without actually reading much of the linked study. (In my defense it was just before Christmas breakfast over here.) I was assuming from the sample size and the comments in this thread that we were looking at a meta-analysis here, but hadn't actually looked at the study's methods yet. As Blasdelb notes above this was not a meta-analysis but rather an extremely large observational study of a more traditional sort. Potentially that makes this even more powerful because it means that the study can be designed from the start to address the questions the researchers wanted to ask. (In meta-analysis one is limited by the suitability of the existing research.)

Part of the reason that I got caught out was because one so rarely sees traditional studies with such huge samples. They are very powerful but are generally expensive and difficult to run. Observational studies are easier do at large sample sizes than experimental ones of course, but it's still a serious undertaking. Confounding factors like self-selection are indeed a problem with such studies (as Blasdelb and others have said) but they can still be controlled for to a significant degree, as long as one is careful to get good data on the confounding factors so that one can use appropriate statistical controls.
posted by Scientist at 9:54 AM on December 25, 2012 [2 favorites]

I liked coffee but my wife didn't. It was grounds for divorce.
posted by twoleftfeet at 10:07 AM on December 25, 2012 [2 favorites]

Asian Palm Civet

2 civets 1 cup
posted by zippy at 10:10 AM on December 25, 2012 [1 favorite]

I get up in the morning and before I go to work I always drink a big cup of decaf coffee.

It's the only way I can muster the energy to go back to bed and blow off work.
posted by twoleftfeet at 10:29 AM on December 25, 2012 [1 favorite]

I'm surprised they're still studying this-- I considered this a settled issue, with most studies showing no effect, and studies showing a harmful effect, showing only a small effect. Of course people will continue to argue that coffee just has to be bad for you.

There are a number of problems with the study that you can jump on if you want, but it's certainly well-written, and the problems aren't glossed over. Survey data is always questionable, especially with poor follow-up. Non-interventional, so not of a lot of use toward definitively answering what people want to know ("Should I start /stop drinking coffee?"). AARP members are not necessarily a representative slice of America. Plenty of unknown confounders. Risk of reverse causality. Lots of questions asked (doesn't sound like any correction for that), meaning some spurious associations may be present. (Again, really impressed with how the authors acknowledged all of those problems; that's not SOP.)

I'm surprised at the modest benefit coffee seems to show, mostly because I would expect the unknown confounders to work the other way: since coffee is seen most frequently as it just has to be bad for you, I would expect that people who were most concerned about their health, people who did everything right, would avoid it (you'd probably see this as a harmful effect within the "takes vitamin supplements, eats lots of vegetables" subpop).

If you want to continue to believe that caffeine is harmful to you, you can do so without contradicting this study. Coffee drinkers only live longer when you control for confounders-- particularly cigarette smoking. Smoking and coffee consumption are not necessarily independent of each other. There may be something about coffee that makes people more likely to smoke, or makes it harder to quit (I think a lot of smokers will argue that coffee definitely does make it harder to quit :) )

There's a good reason that coffee-drinkers are less likely to report having diabetes, btw: people with diabetes are frequently advised to refrain from coffee, as it can raise blood sugar. I checked this out once upon a time, but can't remember if there's any validity to that advice.
posted by nathan v at 10:35 AM on December 25, 2012 [2 favorites]

thelonius: "It's amazing how badly some people seem to want coffee to be proven harmful. After all, it is pleasurable!"

I think that's pretty subjective. For me, even a small amount of coffee throws me into a thirty-hour nightmare where I can't even sit down. So - not really pleasurable. I seem to be an outlier, though.

nathan v: "I'm surprised they're still studying this-- I considered this a settled issue, with most studies showing no effect, and studies showing a harmful effect, showing only a small effect. Of course people will continue to argue that coffee just has to be bad for you."

Thanks for your thoughtful comment regarding the few possible problems with the data. To give an utterly unscientific perspective - personally I think I can understand why at least some people have a hard time believing that coffee is good for you. I am personally very sensitive to stimulants. But even among most people who aren't, if one isn't accustomed to having coffee and then has some all of a sudden, it feels like it has a very large psychological and physical impact. I think it can be difficult for us to believe that something that impactful and (let's be honest) that addictive is really good for us.

But studies tend to show that it's actually a benign addiction, and so I figure it's strange but true: a coffee addiction is not bad for most people. It's bad for me - I have tried having mild green tea regularly, and had to quit because I found I only slept two or three hours per night - but I don't think most people are so sensitive.
posted by koeselitz at 10:52 AM on December 25, 2012 [1 favorite]

We need now a study to answer a more pressing question. Is it healthier to drink coffee instead of black teas?
posted by TwelveTwo at 11:07 AM on December 25, 2012 [2 favorites]

Scientist: "Done well (and I haven't read all of this study, but the idea seems perfectly reasonable and I trust Blasdelb not to post crap studies on the blue) this is an extremely effective technique. Meta-analysis is a large and growing field of research, especially in the medical research community where it has caused something of a revolution in the last several years."

I am not a scientist, so I'll be honest and say I don't know for sure. However - I have known several doctors who are adherents to evidence-based medicine who have expressed a lot of skepticism about the current trend toward meta-analysis. I didn't get the impression that this was because meta-analyses are inherently bad; it's rather because it is very difficult to do a proper meta-analysis. And, of course, actual double-blind studies are in almost every case preferable to meta-analyses. Meta-analyses are generally intended to get around difficulties in doing such studies (hard to get a large enough sample size, hard to associate a large number of variables, etc.)

Basically, from what I can tell (repeating the disclaimer that I am not a doctor or a scientist) the rule of thumb is that meta-analyses can be incredibly useful, but bear much more scrutiny, of only because they draw from larger pools of sources and therefore tend to multiply possible sources of error. As far as I know, that is not a reason to dismiss meta-analyses as bullshit out of hand. It's just a reason to submit them to closer scrutiny and study.
posted by koeselitz at 11:07 AM on December 25, 2012

(And that's what I get for not even reading the whole thread - ha. Sorry, please disregard my last comment and carry on.)
posted by koeselitz at 11:11 AM on December 25, 2012

2 civets 1 cup

Civet coffee (Kopi Lowak) is so 2010. Elephants are the new hotness.
posted by radwolf76 at 11:17 AM on December 25, 2012

Too Much Coffee Man wants to do ANYTHING NOW!
posted by Smedleyman at 11:49 AM on December 25, 2012 [2 favorites]

There is actually a good reason why many people suspected that coffee was bad for you - because historically the sort of people who drink a lot of coffee also do a lot of other things that are bad for them, like smoking cigarettes, drinking too much alcohol, drugs, and excessive consumption of rich foods.

My earliest memories of New York City involve quite a few people with cigarettes constantly dangling from their mouths and a seemingly bottomless cup of coffee.

This is why studies like this, which work hard to tease out a single independent factor, are so very important...
posted by lupus_yonderboy at 1:26 PM on December 25, 2012 [2 favorites]

Does this mean that I can continue to have my coffee standing outside with the cigarette?

Please ask your coffee to come inside because it's cold out, but tell your cigarette to remain standing outside.
posted by Napoleonic Terrier at 1:28 PM on December 25, 2012 [1 favorite]

I didn't get the impression that this was because meta-analyses are inherently bad; it's rather because it is very difficult to do a proper meta-analysis.

Although a bit off-topic, your comment deserves consideration. Meta-analyses were originally proposed for combining multiple randomized-control trials of the same treatment and endpoint. They have been extended to combining multiple studies of other designs, especially observational studies. The idea is to find every study done, evaluate them by a strict set of criteria and then use quantitative methods to combine the results to assess the likely value of the effect and its uncertainty. While many criticisms can be leveled at these studies in practice, it must be remembered that they are intended to replace the informal subjective combining of information from multiple studies, and whatever weaknesses you see in meta-analyses are there in such assessments ("literature reviews") with the added biases of subjectivity, selection, and publication.
posted by Mental Wimp at 1:37 PM on December 25, 2012 [2 favorites]

lupus_yonderboy: “There is actually a good reason why many people suspected that coffee was bad for you - because historically the sort of people who drink a lot of coffee also do a lot of other things that are bad for them, like smoking cigarettes, drinking too much alcohol, drugs, and excessive consumption of rich foods.”

I agree. In fact, even on a more basic level – the idea of a benign addiction is a bit counter-intuitive. Probably that's in large part because addiction itself has been pathologized in popular psychology and defined down as an inherently destructive disorder. But there's no rule that a physical dependence on a substance must be harmful.
posted by koeselitz at 2:54 PM on December 25, 2012 [1 favorite]

I wonder where the data stands on coffee verses tea. (Especially for those of us who drink tea like people drink coffee - big mugs with plenty of milk repeatedly throughout the workday.)
posted by maryr at 5:29 PM on December 25, 2012

So maybe, if I just took to popping coffee beans directly into my veins with a massive hypodermic needle, we could bolster the data for these guys.
posted by PuppyCat at 5:54 PM on December 25, 2012

To me, the interest in this study shows that people understand the social meanings of food better than the nutritional realities. We all learn to recognise foods according to social categories, and these categories become very hard to question, even if they're wrong, and particularly when these assumptions say something about our social identities, which we tend to defend beyond rationality.
posted by chrisgregory at 7:17 PM on December 25, 2012 [4 favorites]

Coffee is a good source of potassium.
posted by pointless_incessant_barking at 11:02 AM on December 26, 2012

koeselitz: However - I have known several doctors who are adherents to evidence-based medicine who have expressed a lot of skepticism about the current trend toward meta-analysis.

Assuming by "doctors" you mean practicing MDs and DOs, I don't know many doctors who are scientists (people who describe, test, and report hypotheses according to the generally accepted principles of science). I know many who are technicians with extensive training, both from school and in a sort of apprenticeship as an "intern".

I'd trust medical researchers before doctors, when it comes to performing and evaluating medical research.
posted by IAmBroom at 8:01 PM on December 26, 2012 [1 favorite]

As a coffee lover, this can be considered as a good news. As maryr mentioned above, I wonder what tea vs. coffee study would reveal.
posted by immunophen at 10:13 PM on December 27, 2012