3 ms·

> Another crappy diet study. Correlation does not equal causation, etc. etc. It’s well known that these sorts of studies have innate problems If you think scie

> Another crappy diet study. Correlation does not equal causation, etc. etc. It’s well known that these sorts of studies have innate problems If you think scientists don't know that correlation does not imply causation, I'm happy to report to you that they, in fact, do. Separating out incidental correlation from causation is actually kind of the point of science, and by extension, medicine. The primary population that gets it wrong is the people that read a single study, and immediately conclude something based on that one study, like "oh hey ketogenic diets reduces bad cholesterol, it must be super good." And it's simply intellectually dishonest (at best) to state that the motivation of the researcher is only to increase their paper count with presumable producing junk science, without presenting a shred of evidence to support such a claim(talk about bad science).

amanaplanacanal 5y ago
All this sort of study can do is to point to where there might be connections to follow up on by doing the real research: randomly assign people to diets and watch the results. But that step never seems to get done. Instead we have this endless series of “studies” which show us absolutely nothing. You said that all cause mortality increases on a ketogenic diet, and this study shows no such thing.
- deeviant 5y ago
  > You said that all cause mortality increases on a ketogenic diet, and this study shows no such thing. ... > The risks of all-cause and cause-specific death over an average 6.4-year follow-up rose with each fall in carbohydrate intake (see table), and remained significant after adjusting for all available factors that might have influenced the association (Model 2 in the table). Quoted from the study. It suggests exactly that thing. I suggest you read the study again.
- esyir 5y ago
  You know why these don't get done? Because that means that you're grabbing a large cohort of people, fixing their diets, and following them until a good number of them die. The timescales required here, not to mention the cost, means that such studies are virtually unfeasible.

amanaplanacanal 5y ago

All this sort of study can do is to point to where there might be connections to follow up on by doing the real research: randomly assign people to diets and watch the results. But that step never seems to get done. Instead we have this endless series of “studies” which show us absolutely nothing. You said that all cause mortality increases on a ketogenic diet, and this study shows no such thing.

deeviant 5y ago
> You said that all cause mortality increases on a ketogenic diet, and this study shows no such thing. ... > The risks of all-cause and cause-specific death over an average 6.4-year follow-up rose with each fall in carbohydrate intake (see table), and remained significant after adjusting for all available factors that might have influenced the association (Model 2 in the table). Quoted from the study. It suggests exactly that thing. I suggest you read the study again.
esyir 5y ago
You know why these don't get done? Because that means that you're grabbing a large cohort of people, fixing their diets, and following them until a good number of them die. The timescales required here, not to mention the cost, means that such studies are virtually unfeasible.