The problem is in the cancer cargo-culting mindset that people without a scientic education or mindset tends to have.
A person I know doesn't freeze bread because she says it causes cancer, and she's not the only one (far from this...) with such ridiculous preconceptions.
Smoking is not as simple as many assume. Smoking actually changes the disease risk profile a lot, lowering risk for many diseases. It's probably an open question if light smoking is even bad for you. Super centenarians have disproportionately been light smokers.
The NHS makes me very proud here. A well written next-day response to widely published bad science journalism of the worst sort.
That headline was plastered across the front-pages of all the newspapers I read with no caveats. Both scare-mongering of unproven dangers and diminishing of more certain dangers like smoking & dietary excess. IMHO media outlets should be heavily fined for bad science journalism. Its a grave form of libel but poor injured science isn't able to sue them for it.
Don't forget that it is not easy to remove confounding with smoking, drinking, low physical activity etc. as, incidentally, those who eat most meat happen to also score worse with those factors. Of course, there are statistical methods, but there are reasons to believe that not all confounding may be properly removed.
For those interested in statistics behind studies like this, I think this (not very long) video is a good explanation why it is not so easy to remove all residual confounding:
The lector actually discusses a very similar or even the same study - mortality as a function of meat consumption. Not too complicated for a technical person even without a background in statistics, I believe.
> Don't forget that it is not easy to remove confounding with smoking, drinking, low physical activity etc. as, incidentally, those who eat most meat happen to also score worse with those factors.
And those who aim to be healthy will probably try to be at least moderate w.r.t their meat intake because they have heard that meat is unhealthy, which further "proves" that meat is unhealthy, while in reality it might be because of other habits that they might have, like exercising.
Things like paleo might make things a little less one-sided.
A cross-sectional study of this type of data will yield crappy, linkbait-friendly results.
For confounders, they didn't look at: income, geography, genetics (self-reported ethnicity is not a good proxy), etc. etc. etc.
The time and sample scale for powerful epidemiological research often seems higher than publish! publish! publish! researchers are given time and money to do.
Any morsel of truth to animal protein having a negative effect is lost in the structure of the study and data available.
Probably because the study was badly done. The difference in their results between 50-65 and 65+ is almost impossibly large and other studies done have never found those kinds of differences.
I once took a class in business school on medical pricing and the instructor said that if you look at a graph of life expectancy that a large percentage of the population dies in the 50-60 age range, but statistically people who live past the age of 60 or 65 (I can't remember the exact figure) will be likely to live to be over 75. In other words, once you make it past a certain age you are sort of screened out as somebody who is high-risk to die of certain ailments like heart disease or cancer.
So if we take that to be generally accurate, then these findings make sense because I would wager that if you studied any diet or habit thought to be unhealthy, more people would die off between 50-65 than after 65.
This also explains why averages don't make sense with regard to any conversation about life expectancy since the averages really need to be observed in bands.
If there's a confounding variable, it's likely that it is affected by retirement. For instance, a high protein intake in someone who's fifty might signify junk-food lunches, whereas the same thing in a retired person might be a sign of a more affluent lifestyle in general.
I suspect that there are confounding variables at work.
"Among those aged 50-65, those who ate a high protein diet were 74% more likely to die during follow up than those who ate a low protein diet (hazard ratio (HR) 1.74, 95% confidence interval (CI) 1.02 to 2.97). People in this age group who ate a high protein diet were more than four times as likely to die from cancer during follow up than those who ate a low protein diet (HR 4.33, 95% CI 1.96 to 9.56).
The results were similar once the researchers took into account the proportion of calories consumed from fat and carbohydrates. Further analyses suggested that animal protein was responsible for a considerable part of this relationship, particularly for death from any cause.
However, the opposite effect of high protein intake was seen among those aged over 65."
"Researchers doing observational studies have much the same problem. They try to think of all the differences between two large populations of subjects so that they can statistically negate them so that only the observation in question – the vitamin C level in the example above – is different between the groups. Problem is they can never possibly think of all the differences between the groups. As a consequence, they never have a perfect study with exactly the same number, sex, age, lifestyle, etc. on both sides with the only difference being the study parameter. And so they don’t really ever prove anything. In fact, we would all probably be a lot better off if all the researchers doing observational studies had followed my lead and fallen asleep mid study."
"Information on food intake was only collected for one 24-hour period... ...this may have changed over the 18 years of follow up"
I wonder if the researchers' and USC's ploy to use sensationalist(though outright false) PR is going to end in them getting more funding. I hope not, but they do say no publicity is bad publicity.
I like how people think one study is enough to upend an entire body of knowledge. Never mind the fact that there is an intense amount of bias in nutritional research. Never mind the fact that the results of many studies cannot be accurately reproduced. Never mind the fact that observational studies generally suck and cannot 'prove' anything. Nah, this new study, which happens to conform to my preconceived notions, this new study changes EVERYTHING.
Read down toward the bottom: the study was based upon one self-reported day of each individual's diet extrapolated over 18 years.
This study is even less meaningful than most. I'd go so far as to say that it's utterly devoid of useful information and quite probably is misinformation.
I wonder how much was spent on performing the study?
> I like how people think one study is enough to upend an entire body of knowledge
almost nobody in this thread thinks this. in fact, quite the opposite. it seems everyone thinks, "well, this study doesn't conform to my preconceived notions, so it can't be true".
in fact, this is one study (although there was also a similar one done on mice) that shows a surprising result. i think an appropriate reaction is: maybe its true, and we should definitely do more research in this area.
If The Telegraph reported this as the NHS describes then this strikes me as profoundly irresponsible journalism. I'm no fan of The Telegraph, but I'm actually a little bit shocked
Looking at "animal protein" as a broad category is dumb. The interesting health questions are specifically about muscle meat. Eating a lot of muscle meat, to the exclusion of organ meat and collagen and dairy protein, probably provides too much iron and too much tryptophan in the amino acid profile.
Furthermore, you can't just talk about "meat". The distinction between fish, pork, chicken, and ruminants matters. The question of quality matters enormously. Are we talking about really cheap meat fed on moldy corn slop and handled improperly after slaughter (aged, dyed, pumped up with water and chemicals), or totally fresh grass fed beef?
There's also a significant different between unprocessed and processed meat, which these studies don't differentiate. Processed meat also varies widely in the kind of chemicals used in the processing.
[+] [-] kybernetyk|12 years ago|reply
[+] [-] pizza234|12 years ago|reply
A person I know doesn't freeze bread because she says it causes cancer, and she's not the only one (far from this...) with such ridiculous preconceptions.
[+] [-] Cless|12 years ago|reply
[+] [-] stdbrouw|12 years ago|reply
> “People who eat diets rich in animal protein carry similar cancer risk to those who smoke 20 cigarettes each day,” reports The Daily Telegraph.
[+] [-] anigbrowl|12 years ago|reply
[+] [-] a8da6b0c91d|12 years ago|reply
[+] [-] duncanawoods|12 years ago|reply
That headline was plastered across the front-pages of all the newspapers I read with no caveats. Both scare-mongering of unproven dangers and diminishing of more certain dangers like smoking & dietary excess. IMHO media outlets should be heavily fined for bad science journalism. Its a grave form of libel but poor injured science isn't able to sue them for it.
[+] [-] cjg|12 years ago|reply
[+] [-] Evgeny|12 years ago|reply
For those interested in statistics behind studies like this, I think this (not very long) video is a good explanation why it is not so easy to remove all residual confounding:
https://www.youtube.com/watch?v=3hnWaNThBFw
The lector actually discusses a very similar or even the same study - mortality as a function of meat consumption. Not too complicated for a technical person even without a background in statistics, I believe.
[+] [-] Dewie|12 years ago|reply
And those who aim to be healthy will probably try to be at least moderate w.r.t their meat intake because they have heard that meat is unhealthy, which further "proves" that meat is unhealthy, while in reality it might be because of other habits that they might have, like exercising.
Things like paleo might make things a little less one-sided.
[+] [-] jrockway|12 years ago|reply
Note: this post was sponsored by High Fructose Corn Syrup and Monosodium Glutamate Industries, Inc.
[+] [-] pistle|12 years ago|reply
A cross-sectional study of this type of data will yield crappy, linkbait-friendly results.
For confounders, they didn't look at: income, geography, genetics (self-reported ethnicity is not a good proxy), etc. etc. etc.
The time and sample scale for powerful epidemiological research often seems higher than publish! publish! publish! researchers are given time and money to do.
Any morsel of truth to animal protein having a negative effect is lost in the structure of the study and data available.
[+] [-] apierre|12 years ago|reply
[+] [-] cstavish|12 years ago|reply
[+] [-] mrfusion|12 years ago|reply
Also isn't it possible teenagers and young adults could need way more protein than the middle aged?
[+] [-] RyanZAG|12 years ago|reply
[+] [-] dkrich|12 years ago|reply
So if we take that to be generally accurate, then these findings make sense because I would wager that if you studied any diet or habit thought to be unhealthy, more people would die off between 50-65 than after 65.
This also explains why averages don't make sense with regard to any conversation about life expectancy since the averages really need to be observed in bands.
[+] [-] bencollier49|12 years ago|reply
I suspect that there are confounding variables at work.
[+] [-] JoeAltmaier|12 years ago|reply
The results were similar once the researchers took into account the proportion of calories consumed from fat and carbohydrates. Further analyses suggested that animal protein was responsible for a considerable part of this relationship, particularly for death from any cause.
However, the opposite effect of high protein intake was seen among those aged over 65."
[+] [-] RyanMcGreal|12 years ago|reply
[+] [-] himangshuj|12 years ago|reply
[+] [-] andyjohnson0|12 years ago|reply
[+] [-] tjaerv|12 years ago|reply
"Researchers doing observational studies have much the same problem. They try to think of all the differences between two large populations of subjects so that they can statistically negate them so that only the observation in question – the vitamin C level in the example above – is different between the groups. Problem is they can never possibly think of all the differences between the groups. As a consequence, they never have a perfect study with exactly the same number, sex, age, lifestyle, etc. on both sides with the only difference being the study parameter. And so they don’t really ever prove anything. In fact, we would all probably be a lot better off if all the researchers doing observational studies had followed my lead and fallen asleep mid study."
[+] [-] retrogradeorbit|12 years ago|reply
http://web.archive.org/web/20090223222003/http://www.nutriti...
And even though it was epidemiological (statistically based), it's findings were strong and it confirmed prior laboratory results and hypotheses.
[+] [-] raverbashing|12 years ago|reply
I'm wondering if there are better statistical tools to deal with it today.
Give it the individual profile of each person and let it draw the correlations.
[+] [-] shittyanalogy|12 years ago|reply
I wonder if the researchers' and USC's ploy to use sensationalist(though outright false) PR is going to end in them getting more funding. I hope not, but they do say no publicity is bad publicity.
[+] [-] Harj|12 years ago|reply
[+] [-] aaxe|12 years ago|reply
[+] [-] mberning|12 years ago|reply
[+] [-] crusso|12 years ago|reply
This study is even less meaningful than most. I'd go so far as to say that it's utterly devoid of useful information and quite probably is misinformation.
I wonder how much was spent on performing the study?
[+] [-] andylei|12 years ago|reply
almost nobody in this thread thinks this. in fact, quite the opposite. it seems everyone thinks, "well, this study doesn't conform to my preconceived notions, so it can't be true".
in fact, this is one study (although there was also a similar one done on mice) that shows a surprising result. i think an appropriate reaction is: maybe its true, and we should definitely do more research in this area.
[+] [-] circlefavshape|12 years ago|reply
[+] [-] jokoon|12 years ago|reply
I don't know why the bad health reference is smoking really.
[+] [-] tuan5|12 years ago|reply
[+] [-] antidaily|12 years ago|reply
[+] [-] a8da6b0c91d|12 years ago|reply
Furthermore, you can't just talk about "meat". The distinction between fish, pork, chicken, and ruminants matters. The question of quality matters enormously. Are we talking about really cheap meat fed on moldy corn slop and handled improperly after slaughter (aged, dyed, pumped up with water and chemicals), or totally fresh grass fed beef?
[+] [-] bryanlarsen|12 years ago|reply