Timothy Kelley
Senior Contributing Editor

Otis Brawley, MD, doesn’t often put his foot in his mouth. But when he responded to former Senator Bob Dole’s claim that prostate cancer screening had saved his life by questioning whether Dole’s life needed to be saved, the 1996 GOP presidential nominee had a quick retort.

“You’re obviously a Democrat,” he said.

Brawley didn’t mean that America could do without the droll Dole, only that the senator might have survived without being screened. Indeed, if a genie of perfect health care ever offered three wishes, one of them might be about cancer screening: If only it were as good as many of us—inadvertently, perhaps—have hoped it is!

Brawley, the American Cancer Society’s chief medical and scientific officer and executive vice president, gets frequent calls from women who complain that they followed ACS mammography guidelines religiously and still developed metastatic breast cancer. “How did this happen?” they demand to know.

Otis Brawley, MD

“Prostate cancer mortality is going down in 21 countries, only four of which have a lot of screening,” says Otis Brawley, MD, of the American Cancer Society.

“They assume a bad doctor read their mammogram or something,” says Brawley. “But all the studies we have on mammography indicate that at best it decreases risk of death by 30%. That means 70% of the people who were going to develop metastatic disease and die are still going to develop metastatic disease and die.” The problem, he says, isn’t just a public that is ill informed (some would say misled). “There are doctors who misunderstand this issue too.”

Meeting the devil

For a guy in a lofty post, Brawley, born 58 years ago on the Fourth of July and reared in a tough Detroit neighborhood, can be blunt when he needs to be. His provocative 2011 book, How We Do Harm: A Doctor Breaks Ranks About Being Sick in America, argued that the medical system wasn’t failing: “It’s functioning exactly as designed. It’s designed to run up health care costs.” One chapter recalled a chat with a cancer center marketer who—conceding that PSA screening hadn’t been proven to save lives—nevertheless boasted that he could project just how much business his center’s free prostate screenings would generate in radical prostatectomies, radiation therapy—even Viagra and incontinence procedures. Brawley realized, he wrote, that he’d been “granted an audience with Lucifer.”

Screening advice determines ups and downs of prostate cancer incidence

Prostate cancer incidence spiked in the early ’90s after the American Cancer Society and other groups recommended screening. Incidence fell sharply when the USPSTF came out against routine screening.

Source: National Cancer Institute, SEER Cancer Statistics Review 1975–2013

But Brawley is Dr. Moderate on screening when compared with two more ardent skeptics—one east, one west. One is H. Gilbert Welch, MD, professor of medicine at Dartmouth Institute and Dartmouth’s Geisel School of Medicine and author of 2011’s Overdiagnosed: Making People Sick in the Pursuit of Health and 2015’s Less Medicine, More Health: 7 Assumptions That Drive Too Much Medical Care. The other is Vinay Prasad, MD, assistant professor of medicine at Oregon Health and Sciences University, coauthor of 2015’s Ending Medical Reversal: Improving Outcomes, Saving Lives, and a much-followed tweeter (@vinayprasad82).

“Most people think of course you want to be screened—it’s the best thing since sliced bread,” says Welch. “But the truth is much more nuanced.”

Welch decries screening’s tendency “to turn people into patients unnecessarily.” The tests detect abnormalities, he says, but “we all harbor abnormalities.” And though we pretend otherwise, medical care has harms. For acutely ill or injured patients, those harms pale in comparison with the benefits. “But with a well population, everything changes,” says Welch. “It’s hard to make well people better, but it’s not hard to make them worse. Our society has promulgated the view that you can test yourself to health, and that’s not true.”

H. Gilbert Welch, MD

“There’s really only one screening test that we have pretty good evidence helps people live longer,” says H. Gilbert Welch, MD, of Dartmouth. “That’s lung cancer screening in heavy smokers.”

While false positives in mammography have received a lot of press, Welch says they’re just one of screening’s problems. More worrisome, in his opinion, is overdiagnosis: the way screening tests find genuine but nonaggressive cancers that would never have posed a problem but set in motion biopsies and treatment that can cause harm. And there’s the expense. Welch notes that medical costs are a leading cause of bankruptcy.

“There are real tradeoffs people need to understand,” says Welch. “And they’re particularly acute in breast- and prostate cancer screening.” But he sees difficulties even with screening for colorectal cancer, which he calls a “disappearing” disease. “Its death rate has been cut in half since my father died of it in 1978,” he says. This cancer’s sharp decline began before screening was routine, he says, contending that it can’t be credited fully to screening but may partly result from “changes in gut flora,” increased use of NSAIDs, and declining consumption of smoked and cured meats. He argues that America spends prodigiously on colonoscopy—a big money maker for hospitals—when no one has proved it superior to flexible sigmoidoscopy or fecal occult blood tests.

Taking credit for a good trend

Colon cancer’s sharp decline actually started before colonoscopies became common in the late ’90s.

Age-adjusted U.S. colorectal cancer death rates by year

Source: National Cancer Institute, SEER Cancer Statistics Review 1975–2013

“There’s really only one screening test that we have pretty good evidence helps people live longer,” says Welch. “That’s lung cancer screening in heavy smokers.”

A proven lifesaver?

Prasad, in Oregon, doesn’t disagree—except to partly topple the one bowling pin Welch leaves standing. Yes, the National Lung Cancer Screening Trial (NLCST), which randomized 53,454 heavy smokers, did show reductions in all-cause mortality rather than just mortality from lung cancer, he says. But to Prasad, the study’s control arm wasn’t a proper control because its members underwent chest X-rays, and that’s not the standard of care. “Limited evidence shows that screening with chest radiography may even increase lung cancer mortality,” he writes. He says a control group with no screening would have been more appropriate. More important, the difference between the screening group and the control group in the NLCST was greater for overall deaths than it was for deaths from lung cancer.

“So if you believe that the test improves overall mortality, you have to believe that CT screening for lung cancer saves lives from conditions other than lung cancer too,” says Prasad. “To me, that’s not plausible.”

It exasperates Prasad that people ignore “the elephant in the room”: Screening is widely promoted as a lifesaver, even though (except possibly in the case of lung cancer) it hasn’t been proven in randomized controlled trials (RCTs) to reduce overall deaths. “That’s the definition of ‘saves lives,’” he says.

To prove that screenings save lives overall, says Prasad, would require high-powered trials 10 times the size of even the large extant studies of breast and lung cancer. To him, such an expenditure would be worthwhile in view of the billions now spent on screening. Whether that’s ever going to happen is, of course, another question, and the answer is probably no.

Prasad insists he’s no anti-screening crusader. “If an informed patient chooses to be screened, I support that,” says Prasad. “What I don’t support is any deceit or omission of truth in telling people about the process.”

And he believes such deceits and omissions currently rule the day. He cites a public-opinion research study in which 62% of women thought mammography at least halves the risk of getting breast cancer, and 75% believed a decade of screening can prevent 10 breast cancer deaths per 1,000 women. “Even the most optimistic estimates of screening do not approach these numbers,” he writes.

Cars that don’t go

If Prasad is exasperated, he’s got nothing on Daniel B. Kopans, MD, founder of Massachusetts General Hospital’s Breast Imaging Division and a professor of radiology at Harvard Medical School (who stresses that he no longer gets paid to read mammograms and thus has no personal financial stake in the debate). He’s fed up to the eyebrows with skeptics offering argument after argument against screening—all of which, he says, he’s refuted. “It’s like playing whack-a-mole,” he complains. “As soon as one argument against screening is refuted by science the skeptics dream up another.”

Kopans decries the distinction made by the U.S. Preventive Services Task Force (USPSTF) when it recommends biennial mammography for women 50 to 74 but says 40-somethings should discuss the screening with their doctor. “Everybody should talk it over with their doctor!” he says. “There’s no magic in age 50—and no data that support using that age as a starting point for screening.”

While agreeing that mammography is imperfect, Kopans contends that women should receive annual mammograms starting at 40. “The only harm is that more women will be recalled for a few extra pictures or an ultrasound,” he says.

Kopans charges Welch with deliberately confusing the issue by “throwing together ductal carcinoma in situ with small invasive cancers as if they’re the same.” While DCIS lesions cannot kill, they can turn into invasive ones, he says, and no one should leave an invasive cancer untreated. “We just don’t know which DCIS cases will go on and invade and potentially kill someone,” says Kopans. “But that’s a treatment issue, not a detection issue.”

Indeed, he exonerates mammography screenings themselves from the overdiagnosis and overtreatment that sometimes follow. “Folks like Welch” who suggest that women may wish to avoid these problems by not getting screened, he says, are like people who say, “If we just take the engines out of our cars, we can stop automobile accidents.”

Mammography “has been shown in RCTs to be able to reduce deaths from breast cancer,” says Kopans, who cites research going back to the first RCT of breast cancer screening, the Health Insurance Plan of New York study in 1971, which revealed a 23% decrease in breast cancer deaths for 32,000 women aged 40 to 64 who were invited to be screened. (Because it looked only at invitations—“no one can be coerced to participate in screening,” he writes—he says that study actually understated screening’s benefit.)

While critics of screening see conflicts of interest in the role of professionals—radiologists’ groups, for example—who advocate screening from within the screening business, Kopans turns the charge around. The National Cancer Institute, he says, is dominated by screening foes, and that affects what research it decides to support.

“That’s a huge conflict of interest that never gets mentioned,” he says.

This study, that study

Kopans blames the Canadian National Breast Screening Study for the NCI’s decision back in 1993 to drop support for screening women in their 40s. But he says that study was underpowered and afflicted with poor-quality mammography. (Brawley counters that it was upheld in a quality audit by the Canadian Justice Department, and Welch says the mammograms did “exactly what they were supposed to do.”)

William Catalona, MD

“We’re now seeing many more men come in with advanced, incurable disease—we’re undoing the good that was done over the past 25 years,” says William Catalona, MD, of Northwestern University.

This window on the decades-old mammogram controversy reminds us that the clamor over cancer screenings is more than just noisy and vituperative. It’s a seemingly endless thicket of methodological detail. Thus urologist William Catalona, MD, of Northwestern University, lauds the seven-nation, 162,387-man European Randomized Study of Screening for Prostate Cancer, which indicated that PSA testing lowered prostate cancer-specific mortality, and trashes the 76,693-man U.S. Prostate, Lung, Colorectal and Ovarian Cancer (PLCO) Screening Trial, which suggested it did not.

Otis Brawley does the opposite.

Even in the European trial, says Brawley, data from five nations out of seven showed no benefit, and one disproportionate contributor to the affirmative result—Sweden—has been shy about letting others review its data. Catalona points out that in the U.S. study, some 90% of the men in the theoretically unscreened control group underwent PSA testing outside the study, making the advantages of screening less apparent. “What he doesn’t tell you,” Brawley pipes up, “is that he worked really hard to get those biases in that study—to sabotage it.”

Catalona denies it was sabotage; he says many men who had been enrolled in a PSA study of his in St. Louis “subsequently enrolled in the PLCO trial because it offered free chest X-rays and colonoscopies,” and half of them were randomized to the control arm. He’s also critical of the “innuendo” in Brawley’s book about screening advocates’ motives and finds it “remarkable” that Brawley, who is black, would be “biased against prostate cancer screening” because blacks are especially vulnerable to the disease. And so it goes.

“I think people like Gilbert Welch are throwing away the baby with the bath water,” Catalona says. He concedes that two decades ago U.S. medicine was overdoing prostate cancer treatment. (As Prasad puts it, “the country was punch-drunk on PSA.”) But he regrets the USPSTF’s 2012 decision to cease recommending routine PSA screening. “It’s creating a generation of family practitioners and internists who’ve been taught that screening for prostate cancer is a bad thing,” he says, with the result that “we’re now seeing many more men come in with advanced, incurable disease—we’re undoing the good that was done over the past 25 years.”

Catalona believes all reasonably healthy men should have a baseline PSA blood test for risk stratification, so that if their level later rises dramatically they can be checked out to make sure prostate cancer isn’t the cause. He also thinks screening critics are aiming their guns partly at the overtreatment of the past rather than the reality of screening today. “I’m actually going to agree with Bill Catalona on that one,” says Brawley, noting that as overtreatment goes down, screening’s benefit/harm ratio changes.

Does he also agree that a new wave of advanced prostate cancers is upon us?

“I don’t know,” Brawley says. “I have an open mind about that. But I do know that prostate cancer mortality is going down in 21 countries, only four of which have a lot of screening. So it seems to be declining at least partly for some reason other than screening and aggressive treatment of localized disease.”

Putting screening to the test of what we know—or believe

It was in 1975 in the 11th grade English classroom of Father Richard Powalski, at the University of Detroit Jesuit High School, that Otis Brawley, MD, learned the critical-thinking maxim he applies to medicine today: “Say what you know, what you don’t know, and what you believe—and label it accordingly.”

Brawley, chief medical and scientific officer of the American Cancer Society, has made this mantra so well known that it pops up at the end of a 2016 BMJ analysis of cancer screenings co-written by Vinay Prasad, MD, of Oregon Health and Sciences University. But Prasad renders the key phrase as “what we simply believe.” His point is that, even though backers say screening “saves lives”—and even though some studies have shown that screening does reduce deaths from a particular cancer—almost no studies have demonstrated that screening reduces mortality from all causes.

Brawley himself, however, has a slightly different take. Carefully labeling what we know doesn’t mean we can’t act in the absence of certain knowledge, he believes. “I’m willing to settle for a cancer-specific mortality reduction,” says Brawley, noting that most of the studies now cited weren’t designed to show all-cause mortality. “It’s unfair, in my mind, to say we shouldn’t do a screening test because we have no study proving it increases overall survival when, indeed, we’ve never tried to run a study to show that.”