“Colon cancer screening saves lives.” That’s how I began my latest article for Nature Medicine. The news story was about new non-invasive tests that aim to make screening a lot more pleasant. The lead wasn’t great. You could probably find a dozen other stories that begin the same way. But it at least seemed solid. Science is all about qualifiers and caveats, and I was delighted that, for once, I didn’t have to add an “appears to” or a “perhaps.”
When I fact checked the article, I didn’t really stop to question that line. Colon cancer screening saves lives. Of course it does. If it didn’t, we wouldn’t devote time and money to screening. Duh.
But then a few days ago I received an email from a reader. “I am puzzled by your opening statement in your recent Nature Medicine article, that colorectal screening saves lives,” she wrote. And then she quoted the National Cancer Institute: “Based on solid evidence, screening for colorectal cancer (CRC) reduces CRC mortality, but there is little evidence that it reduces all-cause mortality, possibly because of an observed increase in other causes of death.”
It took me a moment to process what that means. Screening reduces the number of deaths due to colon cancer, but when you compare overall mortality in people who participate in a screening program versus people who don’t, there’s no difference.
How can that be? I decided to do a little digging.
One possibility is that screening does have an effect, but we need much larger studies to detect it. Colon cancer accounts for only 2-3% of all deaths. So even a really awesome screening test would have only a small impact on overall mortality. And picking up that small impact “would require trials that are too large to be feasible,” writes Robert Steele, the head of cancer research at the University of Dundee who penned a 2011 British Medical Journal article arguing against the use of total mortality to assess the value of cancer screening programs.
Last year, researchers reported the latest results of the Minnesota Colon Cancer Control Study, a study that examined the effect of a stool test—called the fecal occult blood test—on mortality. The team assigned more than 46,000 participants to one of three groups: 1. annual screening, 2. biennial screening, 3. or a control group, and then they tracked their health for 30 years. The researchers found that annual screening reduced the risk of death from colorectal cancer by 32%. However, that corresponds to only a 0.9% reduction in overall mortality. The study wasn’t powered to pick up such a small effect, says Tim Church, a public health researcher at the University of Minnesota who led the study. “So we don’t know whether it reduces overall mortality.”
But how big of a study would you need? A 2006 meta-analysis, which analyzed three separate trials involving 245,000 people, still couldn’t find any difference in overall mortality. “This meta-analysis has almost 3 million patient-year follow-up and this sample size should have the power to detect very small differences in death rates between the two groups,” the authors write.
“Another possibility is that screening actually causes deaths, which offset the ones it saves,” Fletcher says, although he doesn’t think that’s likely. The studies I mentioned were looking at a non-invasive test that only asks participants for stool samples, a nearly risk-less procedure. But even easy-to-execute tests can cause stress, which has been linked to higher mortality. And patients who test positive need to get a more invasive test, which comes with its own risks. It’s also possible that screening causes physicians to catch and treat cancer cases that would not have lead to death, and that the treatment has a detrimental effect on mortality. Surgery can be deadly, and even chemotherapy can sometimes kill.
So does colon cancer screening save lives? According to the Centers for Disease Control and Prevention pamphlet “Colon Cancer Screening Saves Lives,” it does. But I think the truth is a little less clear than that pamphlet suggests. According to the NCI Bulletin, “the only reliable way to know if a screening test saves lives is through a randomized trial that shows a reduction in cancer deaths in people assigned to screening compared with people assigned to a control group.” And colon cancer screening has cleared that hurdle. But others argue that total mortality should be the standard because it’s less prone to bias.
If you’re an individual rather than a statistic, you have a choice to make. If you have a high risk of developing colon cancer, screening is a no brainer. But for the rest of us, science doesn’t have all the answers. The evidence is, well, mixed. So it’s frustrating to see pamphlets and ad campaigns that relentlessly push screening and provide no information about the risks as well as the benefits.
That’s not to say I’m anti-screening. After all, mortality isn’t the only consideration. We also need to think about quality of life. “Most people would value not dying of colorectal cancer— with a cancer diagnosis, major surgery, and possibly colostomy and chemotherapy along the way—regardless of whether they lived longer because of screening,” Fletcher writes.
*******
Image courtesy of the Providence of British Columbia on Flickr.
Thanks, Cassandra, for answering my question. Another possibility you should have considered, the one suggested by the quote from NCI, is that people who are “saved” from dying of colorectal cancer because of screening, are merely surviving to be killed by something else instead, e.g. heart attacks or other types of cancer, and that’s why screening doesn’t save lives. To see how this would work, imagine if you screened one million people aged 99 for colorectal cancer. You would probably save no lives at all, because they would almost all be dead of _something_ within the next year or two.
Are you familiar with the work of J Gilbert Welch and his colleagues Lisa Schwartz and Steve Woloshin? They do a great job of explaining this stuff e.g. in Welch’s book entitled , “Should I Be Screened for Cancer? Maybe Not and Here’s Why.”