NHANES Data, Obesity & Diabetes

The image below comes from the blog of Dr. Richard Feinman.   Free drink and an appetizer at the Asylum Pub & Grill tonight for the first person to point out what's wrong with this picture!

original image direct link

Comments

LeonRover said…
Surely the diabetic data might have been split male/female?

Is it not available within the NHANES sub-populations?

When Feinmann discourses on biochem, he is an expert.

When ANYONE uses epidemiologic surveys, I PERSONALLY place the same relevance on it as I would on an individual's ANECDOTAL conclusions WRT the individual's data.

Both are "interesting", but how is it relevant to ME.

The only epidemiolic study which passes my criteria for relevance was Sir Austin Bradford Hall's Study on Smoking and Lung Cancer - few since then had the required rigour. Why? Because very few have been conducted to the rigourous standards he employed.
MM said…
Hmm. Well one thing I see is that in his little diabetes graph the diabetes incidence had a huge jump between 1970 and 1975, but the diet remained stable during those years. Actually the diet remained stable up to about 1980 while diabetes continued to climb. However, if there's something else wrong with the picture I'm not seeing it.
CarbSane said…
MM wins! Indeed the "blips" in the climb occur during the steep part of the intake curves!
CarbSane said…
When Feinmann discourses on biochem, he is an expert.

True, which is why it's so sad the man is (by his own admission) so lacking in understanding of chemical thermodynamics.
ProudDaddy said…
Booze gives me migraines, so I don't regret being late to the contest (although I'm certain I would enjoy meeting Evelyn). I would have said the same thing as MM and went on to add that diabetes probably takes years to develop, so the graphs needed to start a lot earlier. I would also have noted that the only way the data would indicate ANYTHING would be if it was 100% known that maconutrient consumption is the sole cause of diabetes. I would have then used the word confounder. 'Nuff said...
Mario Iwakura said…
Actually, if you take in account that diabetes is a disease that take a long way to manifest and be diagnosed (10y+), this gap error is even more evident!
Geoff 99 said…
Yes - if you want to nail down a single cause for the increasing rate of diabetes, you essentially need to find something with a distinctly linear rate of growth probably starting 10 to 20 years before this chart's start date of around 1969? Im assuming when we factor out the simple absolute growth in the population, the linear growth will be as steady, but perhaps not quite so steep.

One answer may simply be the rising 'average' age of the population in general - and with it the associated risk of health issues. Add to that a health system increasingly based on diagnostic 'numbers' (coupled with drug based relief) and you have at least a partial explanation for the slope on the diabetes graph.

If you want to complicate things along the way you could throw in stuff like cultural and lifestyle changes (TV, cars, supermarkets, food technology/availability, refrigeration, environmental pollution, etc, etc).

Its a lot to factor out if you want to study the relationship between nutrient intake and diabetes.
KD said…
Eyeballing the trend line and taking into account that there could be measurement errors in any one of those figures, it looks like a pretty linear trend upwards- ie there may not be much point in trying to spot microtrends within the overall upward trend. In any case, it sure doesn't look like it's the carb per se that caused the rise.
MM said…
Yay! Is this the menu?

Cherry Pickin Martinis (Taubes), Spiked Pink Leptinade (Rosedale), Frozen Leptin Drops (Kruse), F'en Wheat Belly Buttons (Davis)

Gee, I can't decide. :)
Lesley Scott said…
looking at the line charting the growth of people with diabetes, I wonder what caused it to taper off in the mid-80s and even decline sharply in the early 90s? I remember aerobics & going to the gym started becoming a lot more mainstream around then (I was one of the legions furiously bouncing around in my shiny new Reeboks & legwarmers to songs from the Flashdance & the like). Plus bodybuilding became much more mainstream...remember the wonderful Rachel McLish? I wonder if all that enthusiastic cardio & working out more had the effect of cutting a lot of people's overall caloric intake? Plus doesn't intense exercise increase insulin sensitivity?

But then, the line spikes straight up again around '92. huh.
Galina L. said…
According to the graph it is logical to guess that population with diabetes on the rise changes the diet, not over-vise. Also sounds interesting.
Wearing shoulder pads heals diabetes. Notice how the diabetes slides back up after we gave up wearing the things after 1992 or so. ; ) Back to the DYNASTY look. We must test this theory!
bentleyj74 said…
Hey! I saw a commercial for some sort of Dynasty sequel...diabetes is goin' DOWN.
Unknown said…
Could these numbers simply reflect better record keeping, and a population that is better informed as to the ill effects of diabetes, thus, making there way to the doctors during this time period. Also, could the doctors be better informed, along with better testing methods, thus, more diagnoses? What I'm saying is that diabetes was not on the rise, it had always been high, the only rise came in the reporting and diagnosis.
Nigel Kinbrum said…
Where are the graphs of energy expenditure vs year?
Jason Sandeman said…
The trend of 92 can be explained as the WHO tightening the criteria for a dx of diabetes. IIRC, it was around then that it changed.
julianne said…
There is a much better correlation with the introduction of margarine and shortening in the 40's - accounts for the 20 year lag.
I like this graph - scroll down in this link
http://www.krispin.com/omega3.html

Sunscreens were also invented in the 40's, maybe it correlated with a drop in vitamin D? Given low vit D and insulin resistance are linked.

(Why do we pick the 'facts' that suit our theories?)
CarbSane said…
Good catch Mario! I agree with KD, that trend looks pretty consistently upward with no inflection mirroring either the carb or intake curve.
CarbSane said…
Each cocktail now comes in four versions: 2 virgin, 2 alcoholic, one each high carb and low carb :D
CarbSane said…
Welcome Terry! I suspect this is indeed part of it. When I was in my 20's I only went to the doctor for the annual gyn exam. No bloods drawn, just weight, BP, pap & pelvic exam. The only other times I had to see a doc back then was the seemingly annual bouts of strep throat (haven't had that since I stopped working around so many adults with kids in day care ...) for which I visited the walk-in. They just did a swab and culture and that was that. Ten years later you couldn't go to a walk-in for a bad case of poison ivy w/o a blood pressure and weight check. I assume incidence of hypertension has gone up as a result of more people becoming aware of it from such incidental visits. Might explain at least part of the rise. I'll take Jason's word below about the fluctuation in the early 90's.

I'm very wary or the fear mongering with diabetes. It's darned serious, don't get me wrong, but the conversion rate from prediabetes to frank diabetic is very low so I think it's a bit of a misnomer. I also see statistics lumping prediabetics in with diabetics and including estimates of *undiagnosed* people in the mix. So on top of having more people diagnosed with screening, "they" still need to add on more that would be diagnosed? The assumption being that folks who don't go to the doctor routinely are less healthy?
Nigel Kinbrum said…
Feinmann never did give a satisfactory response to your question HERE.

Which part of "Absolute carbohydrate intake was kept constant throughout the study." did he not understand?