The Study
I came across a New York Times article on my news feed suggesting that eating processed meat increased your chances of colon cancer by 4-18%. But, my analysis shows that in fact it could really only increase your risk by .14% – .64%.
https://www.nytimes.com/2018/12/14/well/eat/is-eating-deli-meats-really-that-bad-for-you.html
There are usually 3 types of reactions to articles like this:
- “Oh crap! Well, I’m never eating a deli sandwich again. And I’ll be swatting hot dogs out of little kids’ hands at every birthday party and cookout I attend. No way my little ones will munch on a serving of cancer on a bun.” Then they will immediately share this on social media.
- “Great, another thing I can’t eat. What the heck can I eat? No meat, no eggs, and even organic veggies have pesticides. Protein shakes have arsenic and smoothies are loaded with sugar which will give me diabetes. Fish are loaded with plastic, toxins, and mercury. Gluten is horrible, and even quinoa can cross contaminate with other grains so it’s not technically gluten free, so that’s out. No artificial sweeteners, because that can give rats cancer, so I guess its water (reverse osmosis of course) and organic hydroponic kale for the rest of my life. Screw that, pass the Doritos and beer! I’d rather die fat and happy. Enough with this diet nonsense.”
- “Hmm… let’s investigate further by reading the original source or pass it by someone who can better interpret it than I. Epidemiological nutritional studies tend to contain a lot of flaws which makes it really hard to make casual relationships. Also, the media poorly interprets them and shares sensationalized and erroneous conclusions. “
Which one do you have?
To validate my existence on this planet as a health professional/nutritionist while boosting my hopes for humanity, I sincerely hope you chose the third reaction whenever you encounter something like this.
Why the difference between what the article says, at 4-18% increased cancer risk versus my interpretation of the same research showing instead a .14% – .64%?
I think the most important thing to focus on in interpreting such claims about studies on nutrition and disease is the distinction between relative risk and absolute risk.
Failing to know the difference can be scary.
It can be the difference of hearing:
“You are 18% more likely to get cancer because you eat a couple slices of bacon each day!”
vs.
“Your risk of cancer might be raised by .64% because you eat bacon every day.”
That’s a big difference! And it’s unfortunately why reasons 1 and 2 are so common, which is not good.
My Analysis
So here’s what the research really found based on my analysis
The study quoted by the PhD interviewed, was an often sited systematic review of several studies, investigating dietary habits among hundreds of people in multiple countries.
Researchers grouped people based on how much processed meat they ate (which contains a wide variety of organ meats, deli meats, etc.)
They also tracked the incidence of colon cancer.
They noticed that the incidence of colon cancer correlated with the amount of processed meat they ate. Those that eat about 15 grams of processed meat a day (about .53 oz per day, or 3.70 oz a week) had a 4% increased risk compared to those who did not. Those who ate 50 grams of processed meat a day (about 1.76oz /day or 12.75 oz per week) had an 18% higher increased risk of colon cancer compared to those who didn’t.
A 4-18% risk sounds high. But that is a relative risk, not an absolute risk. Here’s the difference.
As human beings, our absolute risk of getting colon cancer is about 3.75%. So by increasing our relative risk by 18%, it increases our absolute risk by .64%.
You see, relative risk sounds scarier, that’s likely why it’s reported more often in the media. But absolute risk is more meaningful in thinking about the true impact decisions have on risk.
Here’s another example.
Consider these headlines:
“magic pill reduces risk of stroke by 50%”
“magic pill reduces strokes from 2 per 100 to 1 per 100 people”
The first headline reports relative risk, the second reports absolute risk.
Back to my analysis of the research showing the association between processed meat consumption and risk of getting cancer.
To focus on this important distinction regarding risk, let’s ignore the potential limitations of epidemiological studies (of which are many, but the primary one is that correlations found do not imply causation).
Let’s assume that the study is without limitation and could tell you that there is a clear connection between processed meat and cancer.
Using the relative risks found in the data, plus knowledge of the incidence of colorectal cancer in the US, we can see where we get the more meaningful absolute risk figure.
There are about 150,000 cases of colorectal cancer diagnosed each year in those older than 50 in the US (Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin 2019). There are approximately 42,000,000 people older than that in the US (https://www.kff.org/) . That means about 3.57% of the population will be diagnosed with colorectal cancer. So if you eat more than 50 gram a day of processed meat, then your risk goes up (18% increased relative risk) to 4.21%. Eating 15 grams per day increases your risk of colorectal cancer (4% relative risk) to 3.71%.
When you are deciding whether or not to have between 15g and 50 g of processed meat each day, you are essentially deciding to potentially increase your risk of colon cancer by .14% to .64% compared to if you decided not to consume processed meat.
Laying it all out
For many people, the convenience of having food they enjoy that gives them a good source of protein exceeds the very small increase in cancer risk. It may allow them to be less stressed and better handle bigger challenges, like reducing sugar intake, or keeping hunger at bay when lowering calories to lose weight.
For others, there is no amount of risk too small to warrant a diet change.
Both are valid reactions to the data, but rarely are decisions made when the data is correctly interpreted and the true risks laid out.
I think this concept is so important. Sensational claims based on statistical manipulation is at best a crucial mistake that misinforms, and at worst fear mongering that adds to the misinformation, confusion, and apathy that many have regarding the already challenging tasks of selecting and changing health behaviors.
Hopefully this helps for the next time you come across another study telling you that something is bad for you.
See if you can tease out the relative risk vs absolute risk. Of course, feel free to pass it over to me so myself or one of my colleagues can better evaluate it for you.
