Ornish Claims High Protein Diets Are Deadly and Quotes a Study. His Claims Are Demonstrably False and the Study is Flawed. It’s Ornish’s diet, the vegan one, that is dangerous.

From the New York Times 3/23/2015 we have this op-ed title: “The Myth of High-Protein Diets.” Ornish states that a certain study attests to grave dangers if protein is eaten. His link to the study is pay-walled, however, fortunately for us, that same paper is available here, for one and all to see. So we read it, and so can you.

Now we may have our biases, but we do find the vegan movement, of which Ornish is certainly one of the high priests, is known for politics and cherry-picking. We suppose their justification is that all of this is in support of their noble cause.

It would seem any data can be used to prove animal product is bad for you. The China Study is often cited as proof, but it in fact proves the opposite. You can read about it in our What Does the China Study Really Prove? post, or see the graph we plotted from it at the end of this post..

Ornish claims that the paper proves there is a 400% increase in cancer among the higher percentage meat eaters aged 50 to 65. This indeed raises the bar for cherry picking far beyond anything we could have imagined. This is a rather spectacular and scary result. The heavy meat eaters were those getting more than 20% of their calories from protein.

Now 20% isn’t some off–the-charts number. The USDA recommends 10% to 35%. Would you believe- we have a bridge for sale!- that people getting 25% or more of their calories from protein are at such a risk and it was only now discovered to be so ?!? Well there is a reason for this odd untrue fact, and it has to do with how the paper chose its data. Let’s do a little de-construction.

Here is a direct quote from the paper: “Using Cox Proportional Hazard models, we found that high and moderate protein consumption were positively associated with diabetes-related mortality, but not associated with all-cause, CVD, or cancer mortality when subjects at all the ages above 50 were considered.”

Here are the key words: all the ages above 50. For the overall group, those eating greater than 25% of their calories as protein were NOT associated with an increase in all cause mortality, heart disease mortality, or cancer mortality. That wasn’t mentioned in the NYT op-ed.

But if the group 50 to 65 actually was dying at 4 times the rate, then it can only mean that once you turn 65, a high protein diet suddenly becomes beneficial. Now this is patent nonsense. No one suddenly starts thriving from meat that was dying from it a few years earlier, but Ornish just talks about the dying part. No where does he mention that the over 65 crowd were having less disease by eating more meat.

But is all this really true? Short answer: NO. Long answer: NO.

The real result, the one you can take to the bank, is that for people over 50, there was no association between meat and all cause mortality, cancer mortality, or cardiac mortality. So what is going on? Simple. When they looked at the overall numbers they could not see that meat was killing anyone. So they employed a technique called ‘slice and dice.’ You can read about this in our Understanding Medical Research – Part 1 – Spotting Doubtful Research post.

‘Slice and dice’ is simple. If you don’t have any compelling results, divide the cohort one way or the other until something publishable is found. In this case, their cohort size was 6,381 and they divided it by age, so 3,039 between 50 and 65 and 3,342 over 65. However, they divided the protein eaters such that only about 7% were in the low protein group. The combination of these causes a huge statistical flaw.

With things cut up this way, they found all sorts of wild ratios between the high and low protein eaters. And they would. Reason: People between 50 and 65 mostly don’t die, but people over 65 invariably do. Now in the paper, they DO NOT STATE how many deaths, they only report the scary ratios. Odd.

But, lets figure out how many would likely die of cancer in the younger group using standard tables. This would seem to be a number like maybe 25 for 3000 people. So in the low protein group, we would expect 7% of that, which works out to about two.

It could vary. It could be 4, it could be 0. This would prove nothing. The whole point of statistics is that you deduce something by looking at a large number of events. Now 1 or 2 is not a large number of events no matter how noble the cause it supports, and ratios constructed out of them are completely without meaning.

This may produce great headlines, but it’s lousy science. You cannot make conclusions based on such a small number of instances.

And, in the very same paper, when you look at the much larger number of cancer deaths in the older group, where there are enough deaths to have statistical meaning (1200 maybe died of cancer in this group) there is no such danger of meat found.

However, Ornish just reaches in the magic research-paper hat, pulls out 400% more cancer, and leaves it there. The same researchers got a 7300% greater prevalence of diabetes among the meat eaters. Ornish doesn’t mention this. It’s too preposterous even for him. The researchers, however, do sheepishly admit that “We underline that our hazard ratios and confidence intervals may be inflated due to our sample size and the extremely low incidence of diabetes mortality in the low protein group. Overall, there were only 21 diabetes deaths among persons without diabetes at baseline, only 1 of which was from the low protein group.”

It is too bad that they don’t state the exact number of deaths in each group. It is even worse that the peer reviewers didn’t point out that little numbers were getting compared to little numbers, resulting in big claims.

Red meat is good for you. If you don’t believe it, look at the China Study data itself. Here’s a plot of death rate versus red meat, with a trend line. More red meat, lower death rate. In fact in the China Study, all the animal product data was this way. More of any kind of meat, less death, more life.

At some level, we need to ask does even a noble cause justify a big lie?