Many journalists fail to question new Cancer Society colorectal cancer screening guidelines

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

The American Cancer Society (ACS) has updated its colorectal cancer screening guidelines, lowering the recommended age to start screening from 50 to 45.

Many national outlets picked up the potentially practice-changing news with these headlines:

Why is the ACS recommending that people get screened at a younger age?

The rate of colon cancer in younger people is rising, according to the stories, all of which featured some version of this quote from NBC’s coverage:

“People born in the 80s and 90s are at double the risk for developing cancer of the colon and four times the risk for developing cancer of the rectum compared to people born decades earlier like the 40s and 50s,” said Dr. Richard Wender, chief cancer control officer for the American Cancer Society.

“We’re actually seeing colorectal cancer developing in people in their teens, which is something that was extraordinarily rare in the past, and it’s all too common today,” Wender told NBC News.

Missing from all of these stories, however, is any quantification of the actual rate of colon cancer in these groups, which might help readers understand what the ACS spokesman means by “all too common.”

As the New York Times noted last year, a person who was born in 1950 had a three in a million risk of developing colon cancer in their twenties, compared with a five per million risk for someone born in 1990. The risk of rectal cancer for someone in their twenties increased from 0.9 per million for those born in 1950, compared with 4 per million for someone born in 1990.

Those increases are in the same ballpark as the risk of being struck by lightning.

For comparison, the incidence of colon cancer in those over age 50 was 119 per 100,000 in 2013 — several orders of magnitude higher.

A modeling study is not a clinical trial

Perhaps more importantly, few of the stories I looked at raised any concerns about the evidence supporting the new guidance, even as they noted that other professional groups are sticking with their current recommendations to start screening at age 50.

Kenny Lin, MD, MPH, professor of clinical family medicine at Georgetown University, offered a counterpoint to predictions that the new guidance would save lives.

“The rationale for screening at an earlier age is that the incidence of colorectal cancer is higher [than in the past]… so there are more cancers to be found by screening,” he wrote in an email. But he pointed out that this rationale makes some assumptions — notably that the cancers being found in younger adults behave the same way, and can be treated as effectively, as those in older adults. It also assumes that outcomes are better when these cancers are found via screening, before they start causing symptoms (e.g. rectal bleeding).

“If either assumption doesn’t hold true, then earlier screening could fail to improve colorectal cancer mortality in real life, no matter what the models say,” Lin said.

Richard Hoffman, MD, MPH, director of the Division of General Internal Medicine for the University of Iowa Carver College of Medicine, pointed out another assumption baked into the guidelines — that patients will follow the ACS screening recommendations down to the letter and seek follow-up screens at the appropriate intervals.

“Given that we are being told that only 2/3 of adults 50 to 74 are adherent [with screening recommendations], this raises questions about the potential effectiveness of screening younger patients, especially with uncertain insurance coverage,” Hoffman said.

Internists Pieter Cohen, MD and Michael Hochman, MD raised other concerns in their Updates in Slow Medicine email newsletter, noting the data wasn’t from a clinical trial.

The change … is not based on the results of a new clinical trial but rather modeling studies based on increasing prevalence of colorectal cancer among US adults in their 40s.  The updated guideline encourages clinicians to discuss different screening modalities with patients, but, unfortunately, does not suggest counseling patients that there are a variety of expert opinions on when to start screening and that beginning screening before age 50 has known risks but only theoretical benefits.

Little attention to harms

Although there are several ways to screen for colon cancer, some professional groups, including the American Gastroenterological Association, endorse colonoscopies over other methods. But is this the safest approach? Cohen and Hochman point to recent data that shows “you are about 8 times more likely to end up in the hospital within a week of your screening colonoscopy than you are to be saved from colon cancer death over 10 years.”

USA Today was one of the few outlets that called attention to these potential harms, which include bowel perforation and complications from anesthesia. It also featured an extensive discussion of costs, and noted the ACS recommendation was “qualified” due to limitations in the evidence. According to an ACS guideline author, this means that “we hope that doctors will look at this and at least start discussions of colorectal cancer screening with their 45-year-old patients.”

However, Hoffman wondered whether clinicians have enough data to make these kinds of discussions meaningful. He noted that truly personalized decision making “requires informing patients about the benefits and harms of early screening and it’s not clear that we have sufficient clinical data to support these decisions.”

To their credit — and reflecting the emphasis in the ACS news release announcing the guidelines — most stories generally did a good job of discussing noninvasive colorectal cancer screening options like blood and stool tests.

But overall, few outlets challenged the narrative that more lives would be saved with an earlier start to screening and that the benefits would outweigh the harms. Those claims are speculative and may be plausible, but they are not yet proven. Some experts believe the attention focused on younger patients could be more effectively channeled elsewhere.

“Something like 1/3rd of the population age 50-75 isn’t up-to-date on any method of colorectal cancer screening,” Lin said. “The benefits of extending the start-screening age by 5 years, even if they exist, pale compared to the potential benefits of getting more people in the universally agreed-upon age range screened.”

You might also like

Comments (7)

Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

Pat Bowne

May 31, 2018 at 9:08 am

When I see US standards and recommendations change in ways that will increase the needed amount of treatment, I always wonder what the equivalent standards are in countries with different medical systems. As you pointed out in 2016, Canada does not view colonoscopies as any better than fecal tests. I decided after one colonoscopy that I would not do that again unless the less invasive test suggested I had a problem.

May 31, 2018 at 12:03 pm

Excellent review of a complex target. Reports of changing screening guidelines are aalso a great opportunity to educate media consumers about the basic reason why we do screening: to save lives. Unfortunately often cancer screening can reduce mortality that is related to the reason being screened (ie: fewer deaths from colon cancer in those screened for colon cancer) yet the overall death rates remain steady. Why? Because as Vinay Prasad and colleagues in the BMJ point out: “disease specific mortality reductions may be offset by deaths due to the downstream effects of screening.” They use the example of th 30 year follow up from the Minnesota Colon Cancer Control Study, which looked at annual fecal occult blood testing. (128 deaths from colon cancer per 10 000 participants in the screened group and 192 per 10 000 in the control arm—a statistically significant difference of 64 deaths per 10 000.) But was there an overall difference in the death rate? There wasn’t. What appears to happen is that even when they do meta-analyses of colon cancer screening, deaths from other causes increase in the populations being screened. The takeaway is that there are always unintended consequences of screening, one of which is death by other causes. (Full Disclosure: I wrote a book on screening called Seeking Sickness: Medical Screening and the Misguided Hunt for Disease and am an occasional reviewer/ blogger for HealthNewsReview.)
Prasad’s study cited above is here: Why cancer screening has never been shown to “save lives”—and what we can do about it. BMJ (Published 06 January 2016) 2016;352:h6080

Clara Chan

May 31, 2018 at 11:16 pm

The first argument about the increase in colon cancer rates of people in their twenties still being like being struck by lightning, while based on facts, doesn’t address the main topic of the articles which is the age of screening being recommended to begin at age 45 instead of 50. They are not recommending screening people in their twenties. Showing readers the rates of colon cancer in people ages 45-49 over time would be relevant.

    Kevin Lomangino

    June 1, 2018 at 7:03 am

    Hi Clara,

    Thanks for reading and commenting. The ACS spokesperson was the first to bring up the data on cancer incidence for people in their 20s, to make the case that earlier screening is justified. Don’t you think that’s misleading? If screening in people aged 45-50 is the main topic of the articles, why is the ACS talking about 20-year-olds? I think it’s necessary and appropriate to respond to that argument with facts.

    I can’t easily put my hands on data for people in the 45-50 age group but this study has some information:

    colon cancer incidence rates increased by 1.0% to 2.4% annually since the mid-1980s in adults age 20 to 39 years and by 0.5% to 1.3% since the mid-1990s in adults age 40 to 54 years;

    Although this is a worrisome trend, experts have noted that it could be a result of more colonoscopies being appropriately performed in people with a family history of colon cancer as recommended under current guidance. Also, colonoscopies are being performed more frequently to evaluate GI symptoms in younger people as Richard Hoffman, MD, MPH told me previously:

    One explanation could be that there is more recognition that a family history of colorectal cancer—and actually polyps—warrants screening with colonoscopy at an earlier age (age 40 or 10 years before the index diagnosis, whichever comes first),” he said. “This message could be accounting for the increase, particularly in the 40 year olds. Based on my experience I would agree that more diagnostic colonoscopies are being performed to evaluate GI symptoms. This could lead to more cancers—particularly early stage—being diagnosed.”

    Best regards,

    Kevin Lomangino
    Managing Editor

Clara Chan

June 1, 2018 at 10:09 pm

Hi Kevin,
Thanks for your reply. I read the comment in the NBC article about teenagers getting colon cancer as a side comment, not as a justification for screening people ages 45-49. In any case, it couldn’t be a justification because it’s not addressing the same ages. Yes, it could be misleading to some readers but it could also be misleading to present the data about people in their twenties to argue the other side about screening people ages 45-49.

It would be nice if the data about colon cancer rates in people ages 45-49 were easier to find and probably should have been presented in the American Cancer Society’s original announcement.

Although the data about people ages 40-54 is closer, it includes both younger people and people ages 50-54 who have been screening based on the existing guidelines. That leads to question such as: How many people ages 50-54 had cancers detected earlier which might increase the rates counted? How many years does it take for “earlier” stage to become “later” stage? If only a few years, then more people being detected at an earlier stage would not put most people in a different 5-year age category whether detected earlier or later. And how many people ages 50-54 had pre-cancerous polyps removed, how is that counted, and would that decrease the rates of cancer counted? Since the rates have been increasing for ages 40-54 since the mid-nineties, are the rates in 45-49 year olds now approaching the rates in 50-54 year olds when the screening guidelines for beginning at age 50 were established?

Richard Hoffman’s quote is still just speculation as it is just saying “it could be” this or that but doesn’t come with data to back it up.

Charles Carter

June 4, 2018 at 7:26 am

Excellent article and several well informed responses.
Many concerns persist about screening colonoscopies, not least the lack of overall mortality benefit. I don’t know if it’s included in the modeling or not but not only are there potential complications, but polyps and cancers are missed. The routine nature and low yield of these exams leads to faster exams (understandably but inexcusable) which are known to be less accurate.
Further at $3,000 or more per exam there is too much profit motive here. And the trend to more exams under anesthesia only adds to costs. CMS is attempting to institute value based payment using onerous intrusive quality measures. To me, the value and price of colonoscopy should be based on alternatives.
Lastly, redundant guidelines from different agencies and societies are a real problem. I quickly scanned the new recs but only brief mention of financial COI’s were documented. I doubt the guideline committee adhered to the IOM’s standards for panel members.

Mark Ebell

June 4, 2018 at 7:59 am

We wrote an article comparing cancer screening in high-resource economies (Pub Health Rev 2018: Vast majority of countries with CRC screening do FIT, only a few recommend colonoscopy as an option, and none recommend Cologuard or any other test. I was on USPSTF, we considered a 45 year start as one of the modeling options, but that was with a 15 year interval (i.e. 45/60/75). It is actually a reasonable option based on modeling (no RCT data for that age group or for colonoscopy at all) if you look at the efficiency frontier, but ultimately FIT-based strategies are by far the most cost-effective (again based on modeling).