Could finding more cancer lead us to understand risk factors less?

Left – Otis Brawley; Right – Gil Welch

An opinion piece in last week’s Annals of Internal Medicine argues that just how aggressively we screen for some cancers can actually distort our understanding of the risk factors for a particular cancer, as well as how common we perceive it to be.

The authors describe ‘scrutiny-dependent’ cancers  — those subtypes of cancers often picked up with screening that are commonly referred to as slow-growing, indolent, subclinical, or even as pre-cancerous — and that often don’t progress to cause health problems or shorten lives. Prostate cancer and thyroid cancer are two such examples.

They propose two common ways in which aggressive screening can distort our understanding of these cancers:

• The push for ‘early detection’ leads to more scrutiny-dependent cancers being found which, in turn, gives the false impression of an increased incidence of some cancers.
• Aggressive screening of the family members of someone with cancer means more cancer will be found. This could give the impression of family history being more of a risk factor than it may actually be.

Co-author Gil Welch, MD — a professor of medicine at the Dartmouth Institute for Health Policy and Clinical Research who specializes in overdiagnosis and overtreatment — explains it this way in a recent article in STAT:

“If we biopsied men without a family history of prostate cancer, at the same rate as we biopsy men with a family history, we’d find more prostate cancer in them as well. Family history influences how hard we look for prostate cancer and therefore how much we find. The risk factor becomes a self-fulfilling prophecy.”

Other examples from the article include:

• Breast cancer found by mammography, ultrasound, or MRI is more likely to be indolent than that found by self-examination. As we screen with more advanced technologies capable of finding smaller and smaller tumors, we will find more cases of cancer but also more cases that don’t develop into cancer. This not only represents overdiagnosis, but can also give the false impression of breast cancer being more common than it really is.
• Low grade thyroid cancer is common, and women are considered to be at a three times greater risk of developing it than men. But the mortality rates for thyroid cancer are nearly identical for men and women. So it’s possible women are simply getting scrutinized much more, having more nodules detected (both malignant and indolent), and the ‘risk factor’ of gender may be ‘more apparent than real’.

In other words, how many people we screen and how often we do so, can potentially mislead us regarding not only who is at risk for a particular cancer, but also how prevalent that cancer is.

And that raises an important question: What role does the media play in all of this?

The prevailing message

Here in the United States we are bombarded with pro-screening messages. They come from our celebrities, doctors, heroes, employers, family, and — above all — the media. The prevailing message is clear: “better safe than sorry” and “catch it early.” Although it’s usually well-intentioned and has strong emotional appeal, it’s not always based on strong evidence and often fails to adequately address the downsides of screening.

We’ve written about dozens of screening messages that cheerlead more than they inform about the benefits and risks. Some recent examples:

• A prostate ‘Pep Talk’ is the wrong way to get men thinking about cancer screening
• Imbalanced ‘Saved by the Scan’ campaign neglects big concerns over lung cancer screening
• What you’re not being told about ‘free’ public head and neck cancer screening events
• Sheryl Crow hawks 3D mammograms with fear and false hope

I asked co-author, Otis Brawley, MD —  the chief medical and scientific officer for the American Cancer Society — what he thinks about how news organizations handle stories about cancer screening.

“In the 80’s and 90’s the prevailing message from the media was screen, screen, screen,” said Brawley. “Only in the past 10-15 years have some reporters begun to question this. And this goes for advocacy groups too, who have an understandable emotional conflict of interest because they care about a particular disease. I know, I work for one. But, both reporters and advocates need to be truthful and accurate about screening. Because people can get hurt.

“We’ve seen plenty of examples — like with prostate disease in men and cervical dysplasia in women — of how over-enthusiasm for early detection caused us to jump ahead and do harmful things before we actually had the scientific evidence. And I want people to understand that the decision to get screened in not a simple, knee-jerk ‘why not?’ ”

Are Brawley & Welch anti-screening?

Brawley says he’s been accused of being ‘anti-screening’ many times — we’ve reported on these sometimes-vicious attacks.  I asked him how he felt about these accusations.

“I’m not anti-screening,” said Brawley. “What I’m against is the over-emphasis on screening, and the over-reliance on screening in instances with little evidence to support it.

“There’s good screening and bad screening. Good screening is tests where we have solid randomized, prospective trials which show as an end result that you have more people alive because they got screened, than if they had not gotten screened. I can quote you 11 studies showing that breast cancer screening for women over 50 reduces the risk of death. But for prostate cancer that’s a harder argument to make. And in the case of thyroid we have no randomized, prospective trials to support screening. That’s the other end of the spectrum.”

In short, Welch and Brawley firmly believe that screening should be based on rigorously tested evidence. In some cancers that evidence is clear, while for other cancers (‘scrutiny-dependent’ ones) the evidence is lacking. Media messages about cancer screening need to do a better job communicating that uncertainty. 

Looking forward it’s worth pointing out that advances in technology will likely push ‘detection’ earlier and earlier. We will be left wondering what we are actually detecting and what its clinical relevance really is.

That will make this little article loom very large indeed.