March is Colorectal Cancer Awareness Month: Here are 6 prevention recommendations that confuse association with causation

Kevin Lomangino is the managing editor of HealthNewsReview.org. He tweets as @KLomangino.

March is Colorectal Cancer Awareness Month, which the American Institute of Cancer Research is marking with a news release headlined “6 Steps to Prevent Nearly Half of Colorectal Cancer Cases.”

The recommendations are said to be “research-based” but that term appears to be used loosely. Here’s why.

association is not the same as causation

What’s wrong with issuing sensible lifestyle recommendations?

The AICR recommendations – with the exception of a puzzling endorsement of garlic — seem reasonable enough on their face. After all, who can argue with advice like “Stay a healthy weight,” “Fit activity into your day,” and “Go moderate on the alcohol?”

I don’t argue with that advice at all. In fact, I think it’s very good advice for general overall health.

What I argue with is the enthusiastic promise that these steps will reduce your colorectal cancer risk by half – which is highly uncertain and almost surely an overstatement.

“In every one of these cases the recommendation is based on observational, but not randomized data,” says Vinay Prasad, MD, MPH, an oncologist and Assistant Professor of Medicine at the Oregon Health and Sciences University. “The specific claims made here are the types of observational associations where drawing causal conclusions is notoriously unreliable.  For this reason, to recommend these practices seems to be careless extrapolation.”

In other words, the studies that support these recommendations are based on observations of people who followed them. But people who, for example, get lots of exercise probably have other habits or characteristics that make them less likely to develop colorectal cancer. (This is a problem known as “confounding.”) Is it the exercise or the other factors that are responsible for the benefits seen in studies like this? One can almost never be certain, which is why most scientists agree that a trial, which randomly assigns people to get more exercise or not, would be necessary to prove whether exercise has a protective effect against colorectal cancer.

The best evidence available, but still frequently wrong

A phone call with Alice Bender, MS, RDN, an AICR dietitian quoted in the release, confirmed that the evidence supporting the AIRC recommendations is by and large from observational studies. 

She noted that randomized controlled trials are very difficult to perform in cancer prevention, especially when it comes to diet. And so the best available evidence in this area is epidemiological in nature, often backed by other types of research, such as lab experiments and animal studies. Then, an expert panel reviews those studies to come up with recommendations like those issued in the AICR news release.

“The expert panel only issues recommendations when the evidence is strong enough and consistent enough that it’s considered unlikely to change,” Bender says.

The evidence in this area indeed may be unlikely to change, as Bender states. But in my opinion, that’s mainly because no randomized controlled studies will ever be done to test these hypotheses. The studies would be too big and run too long to be feasible. And there are no deep-pocketed sponsors lining up to provide the funding needed to conduct them.

In other areas where similar hypotheses can be and have been tested, we find that the conclusions of observational studies are frequently overturned. There’s even a case in point from this week: statin drugs. Although reams of observational data show that statin use is associated with better cancer survival rates, randomized controlled trials have so far failed to confirm any anti-cancer benefit.

Uncertainty in the science should be reflected in consumer health messages

Dr. Martin Wiseman, the Medical and Scientific Adviser for World Cancer Research Fund International and AICR, acknowledges that the evidence to date “cannot offer proof” that his organization’s recommendations will reduce the risk of colorectal cancer. However, given that randomized trials are not feasible and may also suffer from certain limitations, he says that the AICR expert review process is “the best way to approach whether the evidence is strong enough to take action as a health professional.”  

Wiseman’s argument is reasonable, and experts can disagree as to the standard of evidence required to show cause and effect. But since when does “taking action as a health professional” have to involve issuing an imbalanced, incomplete news release that makes exaggerated claims about the evidence?

If the evidence “cannot offer proof” that these actions will reduce the risk of colorectal cancer, is it wise to promise the public that those steps “will prevent nearly half of all colorectal cancer cases”?

And, if such promises must be made, shouldn’t there be an acknowledgment, somewhere, that the supporting evidence isn’t conclusive?

If such cautions aren’t introduced, don’t we risk sending readers for another spin on the health messaging merry go round?

HealthNewsReview.org contributor Richard Hoffman, MD, MPH called the AIRC recommendations “pretty straightforward” and said that  “Exercising, controlling weight, and eating healthy diets are good advice for warding off a host of illnesses.

“But it’s hard to quantitate the benefit because data are from observational studies that may have biased findings,” he said. (Similar to “confounding” which was mentioned above, “bias” is a type of systematic error that plagues observational studies and can skew their results.) “It’s also hard to know exactly what to recommend—how much and what type of exercise is associated with what reduction of risk? Is there any ‘safe’ level of alcohol or red meat?”

We’re not asking anyone to move mountains here

Let me spin a common thread of concern about observational studies from the experts that I reached out to.

  • these types of studies “cannot offer proof”
  • “drawing causal conclusions is notoriously unreliable”
  • “may have biased findings”

We think that descriptions reflecting these concerns need to find their way into the messages that ultimately reach the public.

It’s not hard to do this. We offer a primer: Observational studies: Does the language fit the evidence? Association vs. causation.

And we’ve praised organizations that make the choice to include such caveats in their news releases.

I hope the AIRC will soon join the ranks of organizations that we laud for their balanced messaging.

You might also like

Comments

We Welcome Comments. But please note: We will delete comments left by anyone who doesn’t leave an actual first and last name and an actual email address.

We will delete comments that include personal attacks, unfounded allegations, unverified facts, product pitches, or profanity. We will also end any thread of repetitive comments. Comments should primarily discuss the quality (or lack thereof) in journalism or other media messages about health and medicine. This is not intended to be a forum for definitive discussions about medicine or science. Nor is it a forum to share your personal story about a disease or treatment -- your comment must relate to media messages about health care. If your comment doesn't adhere to these policies, we won't post it. Questions? Please see more on our comments policy.

Comments are closed.