Many medical websites suggest so. If yes, then how? Alzheimer's is primarily incurable but then it can be delayed, how can coffee play a role?
References:
-
Can you link to somewhere that makes the claim?– fredleyFeb 9, 2015 at 12:40
-
Google Search Link: goo.gl/78hWjL– qedkFeb 9, 2015 at 13:27
-
2Sorry, I didn't mean in general, I meant the specific ones you had found. On Skeptics, where they get a lot of questions like this, all questions must be backed up with sources. Although we probably wouldn't need a strict policy here, it's still good to prompt for such information on questions like this to make sure it's a valid question (unlike 'my friend told me coffee gave her warts, is this true?').– fredleyFeb 9, 2015 at 13:30
-
It's there in the question description. Primary articles for verification and a search link if you want to obsess over it. ;)– qedkFeb 9, 2015 at 13:35
-
1@TomMedley, I'm a Wikipedia user and I'm used to verifying every claim I make when I add content. I do the same on Coffee.SE.– qedkFeb 9, 2015 at 13:36
Add a comment
|
1 Answer
tl;dr: Not sure yet, but it looks promising
So there are a whole bunch of studies that have been done that have investigated a potential link between coffee consumption and Alzheimers.
The problem, as always, is that these types of 'cohort studies' are really difficult. Basically the way it works is:
- Get a load of volunteers
- By doing questionnaires, work out everything about them
- Track them over a long period of time
- See if there's a statistically significant link between combinations of behaviours (such as drinking coffee) and a particular outcome (diagnosis of Alzheimers)
All the standard stuff applies here: you are not testing for causation, only for correlation. For example, ~80% of Americans drink coffee, but in other cultures this number varies a lot. If the coffee drinkers in your study are predominantly American, and the non-coffee drinkers are predominantly not, you have a whole bunch of confounding factors. Designing your study to avoid this kind of situation is really hard, especially when you're relying on volunteers.
There's also the issue of what exactly are you measuring? Someone might drink 3 cups a day now, but only has done for the past few years. Another person might drink 2 cups a day, but has done their whole life. Another person might drink 2 cups of decaff a day. Questionnaires are usually not sophisticated enough to capture this level of detail (partly because in your big expensive study you want to capture as much information about different behaviours as possible - not just coffee consumption).
Remember also that medical studies are chronically badly reported in the mainstream press. Even in 'high-brow' publications, by the time the headline-writers get their hands on a study, the message can have changed wildly. It's always worth trying to dig up the actual paper whenever new research is mentioned in the news. If no paper is linked, just ignore the headline.
One tool against all of this uncertainty is meta-analysis. This is where you look at every single study, weighted by how confident you are in the data (e.g. by weighting more to studies with more participants, and better questionnaires etc.). You then do some careful statistics work to get a measure of confidence across all of the studies. Think of it as a sophisticated way of 'averaging them out'.
One of these was done in 2002. Only 4 studies had been done by then, but the conclusion was:
Although our pooled estimates show that coffee consumption is inversely associated with the risk of AD, the four studies had heterogeneous methodologies and results. Further prospective studies evaluating the association between coffee consumption and AD are strongly needed.
(Heterogeneous is science for 'different')
i.e., "Not sure yet, not enough data, but it looks promising, give us more money for research."
