3 ms·

As someone who has closely followed the literature (I scraped and reviewed pretty much all of ACJN) - I find it fascinating that nobody has gone to the effort o

As someone who has closely followed the literature (I scraped and reviewed pretty much all of ACJN) - I find it fascinating that nobody has gone to the effort of doing a large-scale controlled study on this topic. Seriously - how difficult would it be to just get a large population in a controlled environment for period of time (12-18 months), separate them into groups, and feed them pre-determined diets, and observe the results. I'm sure there are lots of institutional opportunities where we have large populations of people who would be happy to volunteer for this type of study.

RobLach 13y ago
Doing some personal research a while back revealed to me that diet studies in general have ridiculously low standards. The vast, vast majority of them are volunteer, self-journaled data based that would be laughed out of any other scientific context. Controlled group studies are almost entirely non-existent and are typically limited to very specific conditions and goals, such as diets for epileptics. I hope at some point we'll sit down and do a proper broad study on different diets where all meals are controlled for x amount of time and get a range of data that you could further divide and conquer and figure such a crucial part of our existence out.
beagle3 13y ago
"Controlled environment" would require that these people would not see their friends/family (because almost every social gathering these days features calorie intake in some way), not be able to go into a convenience store to get the paper, and other things. It's much harder than you'd think. A friend of mine is an MD who does weight management research. Apparently, people misreport their intakes by significant factors - either because they don't consciously remember, or because they try to make themselves feel better - but generally, you can't trust what people do when you are not watching, or describe what they did.
DanBC 13y ago
There are ethical considerations that would mean it could never happen. People don't self report accurately. > I'm sure there are lots of institutional opportunities where we have large populations of people who would be happy to volunteer for this type of study. Foxcon?
gkelly 13y ago
I think the work from http://nusi.org/ http://nusi.org/ could be promising.

RobLach 13y ago

Doing some personal research a while back revealed to me that diet studies in general have ridiculously low standards. The vast, vast majority of them are volunteer, self-journaled data based that would be laughed out of any other scientific context. Controlled group studies are almost entirely non-existent and are typically limited to very specific conditions and goals, such as diets for epileptics. I hope at some point we'll sit down and do a proper broad study on different diets where all meals are controlled for x amount of time and get a range of data that you could further divide and conquer and figure such a crucial part of our existence out.

beagle3 13y ago

"Controlled environment" would require that these people would not see their friends/family (because almost every social gathering these days features calorie intake in some way), not be able to go into a convenience store to get the paper, and other things. It's much harder than you'd think. A friend of mine is an MD who does weight management research. Apparently, people misreport their intakes by significant factors - either because they don't consciously remember, or because they try to make themselves feel better - but generally, you can't trust what people do when you are not watching, or describe what they did.

DanBC 13y ago

There are ethical considerations that would mean it could never happen. People don't self report accurately. > I'm sure there are lots of institutional opportunities where we have large populations of people who would be happy to volunteer for this type of study. Foxcon?

gkelly 13y ago

I think the work from http://nusi.org/ http://nusi.org/ could be promising.