Study Sets Off Debate Over Mammograms' Value

New York Times, December 9, 2001

By GINA KOLATA

A new study in a British medical journal has stirred a passionate debate among doctors in Europe and the United States by asserting that mammograms do not prevent women from dying of breast cancer or help them avoid mastectomies.

The question is dividing experts and women's health advocates, many of whom acknowledge that they do not know what to think about the new report. For more than two decades, annual mammograms have been part of life for millions of women, with the American Cancer Society and the National Cancer Institute urging women to have them.

Experts are still digesting the new findings, which appeared in the Oct. 20 issue of the journal The Lancet, and few if any authorities in the United States are suggesting that women abandon routine mammography on the basis of this study.

Women have long been urged to have the test every year starting at age 50 or sooner, and promised that early detection will reduce their chances of dying from breast cancer by about 30 percent. And detecting cancer early, they were told, would allow women to avoid extensive surgery and harsh treatments that might be needed to control a larger tumor found later.

But the new analysis, of seven large studies of mammography conducted over the past few decades, says those promises are an illusion. It calls into question the assumption that early identification of breast tumors, before they can be felt in an external examination, improves the chances of a cure.

Those studies that found benefits from mammography were flawed, say the investigators, Dr. Peter Gotzsche, director of the Nordic Cochrane Center in Copenhagen, and Ole Olsen, the deputy director. Recent studies, more rigorously designed and conducted, found no such effects, they assert.

For example, they criticize a New York study from more than a quarter of a century ago finding that women who never had a mammogram died of breast cancer at a rate 30 percent higher than those who had the test. (Of the 30,565 who were never screened, 196 died over an 18- year period; of the 30,131 who had the test, 153 died.)

Dr. Gotzsche and Mr. Olsen say this study, and four others, do not meet agreed-upon standards for well-conducted and reliable research. They question whether the subjects who had mammograms might have been substantially healthier than those who did not, and whether deaths among women who had mammograms might be less likely to be ascribed to breast cancer than deaths among women who did not have them.

"The quality of the trials was very surprising because it is pretty low," Dr. Gotzsche said in a telephone interview. "Even if they are judged by yesterday's standards, the quality is low. In some cases, we know why that happened - these trials were conducted by people who were unfamiliar with clinical trial methodology. They were run by enthusiastic clinicians."

The researchers cite with greater approval a more recent study in Malmo, Sweden, that compared 21,088 women who had mammograms to 21,195 who served as controls. After nearly nine years, 63 women in the mammogram group had died of breast cancer, compared with 66 in the control group - an insignificant difference. The other study the researchers approved of, done in Canada, involved 44,925 women who had mammograms and 44,910 who did not. There were 120 deaths from breast cancer in the screened group and 111 among the women who served as controls.

Nor did mammography lead to fewer mastectomies, the investigators say. In the Malmo study, for example, 424 women in the mammography group and just 339 in the control group had mastectomies. One reason may be that doctors aggressively treated some tiny tumors found in mammograms - tumors that might never have developed into cancer or might never have been noticed in a woman's lifetime.

So far, just one country, Switzerland, has taken action as a result of the study, deciding not to offer a national mammography screening program. Dr. Gianfranco Domenighetti of the Swiss Network for Health Technology Assessment said the decision was heavily influenced by the Danish research.

But Switzerland did not have a national program; it was thinking of starting one. It is a different matter in a country like the United States, which has a longstanding policy of urging women to have mammograms. Once a program has been highly promoted and advanced as a way to save lives, said Dr. Barnett Kramer, the associate director for disease prevention at the National Institutes of Health, it can be difficult to suggest that guidelines be revised.

Nevertheless, some American experts, including researchers at the National Institutes of Health, say that the analysis deserves consideration, and that women should at least be aware of the debate.

But others, like experts at the American Cancer Society, say the study is unconvincing. And some advocacy groups say they are agonizing over how to advise women. They say some of their members, whose cancers were found by mammography, will always be convinced that the screening test saved their lives.

The debate has nothing to do with the effectiveness of breast cancer treatment. There is agreement that treatment, with surgery, hormones and chemotherapy, saves lives. Instead, the question - which has come up before with screening tests for other kinds of cancer - is whether earlier treatment is better.

At its heart, the analysis challenges the assumption that the period when a tumor can be seen on a mammogram but not felt in a breast examination is a critical period in which cure is possible. If that assumption is wrong - if cancers can be just as treatable, or just as deadly, whether they are found early or late - then mammography would offer no benefits.

A similar problem recently emerged with breast self-examination, another method of early detection that had been highly promoted. When studies in China and Russia indicated that it did not prevent breast cancer deaths, the cancer institute quietly dropped its emphasis on the method, saying in a database primarily for doctors that there was insufficient evidence of its value.

The institute will ask a panel of independent experts that advises it on that same database to look at the new mammography study, said Dr. Peter Greenwald, the institute's director for cancer prevention.

Dr. Greenwald added that there might be a simple explanation for the discrepancy among the studies analyzed by the researchers in Copenhagen. Newer treatments are saving women's lives, he said. That may mean that a woman's prognosis is nearly the same whether tumors are found early, with mammography, or later, when they can be felt. If this is the case, earlier mammography studies, conducted when treatments were less effective, would have found that screening prevents breast cancer deaths; more recent studies might not have found this effect.

But the most important question, Dr. Gotzsche said, is not whether women are saved from dying of breast cancer; it is whether mammograms prolong their lives. It is possible, he said, that a mammogram might find a cancer early, leading to treatment that might prevent a breast cancer death. But theoretically, the treatment might be so harsh that it precipitates another illness, so that the woman lives no longer, although her death is not attributed to breast cancer.

No one study was large enough to answer that question, so the investigators combined study data. They concluded that the overall death rate did not budge when women had mammograms compared with when they did not.

Dr. Greenwald said that while he had not formed an opinion about the validity of the analysis, he considered the study "important." But other experts, like Dr. David Freedman, a statistician at the University of California at Berkeley, dismiss it as deeply flawed. Dr. Freedman said some of the studies Dr. Gotzsche and Mr. Olsen cite as weak were actually the strongest, and vice versa. He called Dr. Gotzsche "incredibly alert to all difficulties in work he doesn't like" but added, "In papers he does like, he swallows anything."

Dr. Freedman said it would be ridiculous to demand evidence that mammography lowers the overall death rate. The reason, he said, is that breast cancer accounts for only 1 to 3 percent of deaths among women, making it impossible to see such an effect even if lives are saved.

"Mammography is not like aspirin and headache," he said. "The evidence is complicated. But the advice I would give someone is to get screened."

Dr. Robert A. Smith, the director of the division of cancer screening at the American Cancer Society, defended the studies criticized by the investigators in Denmark. "The trials have been gone over with a very, very critical eye by a lot of people," he said. "Although people differ in terms of which ones they prefer, most people would never have dismissed five out of seven as unworthy of consideration."

Dr. Smith said the cancer society would examine the new report as part of a previously scheduled review of publications on mammography. But, he said, "I do not see anything in the Olsen and Gotzsche analysis that would lead the review committee to question the evidence and wisdom of routine screening mammography for women 40 and older."

Dr. Stephen A. Feig, a radiologist who directs breast imaging at Mount Sinai School of Medicine in New York, said, "Screening has weathered controversies before and it will continue to do so."

Health insurance companies usually pay for mammograms. Joseph Luchok, a spokesman for the Health Insurance Association of America said, "If a procedure is covered it would stay covered until the medical community came to the opinion that it is not a useful procedure."

But some doctors who want to advise their patients confess that they are whipsawed by the dueling experts. "The debate has become so sophisticated from a methodology viewpoint that as a doctor my head is spinning," said Dr. Barron H. Lerner, an internist and historian at Columbia University's College of Physicians and Surgeons. Dr. Lerner wrote "Breast Cancer Wars" (Oxford University Press, 2001).

"You read the article in The Lancet and you nod your head yes. Then you read the studies by people on the other side and you nod your head yes," Dr. Lerner said. "We're witnessing this fight between the pro- and anti-mammography forces and they're both arguing that `my data is better and we're right and they're wrong.' "

What should be done, Dr. Lerner said, is "sitting back and trying to analyze how we as patients, as doctors, as a society, should deal with these conflicting data."

That is an issue that is plaguing some advocacy groups.

The report "is really causing a huge amount of concern among activists," said Cindy Pearson, the executive director of the National Women's Health Network. She said her group no longer advocates self- examination. Now, she said, if mammograms are not useful, "you're just left out there with nothing."

Fran Visco, president of the National Breast Cancer Coalition, said she welcomed the mammography dispute. "We know that mammography screening has serious limitations, yet it has been sold as the be-all and end-all for breast cancer," Ms. Visco said. "When someone says, `We have to question that assumption,' we're thrilled. We've been questioning it from the beginning."

"I'm not ready to tell women over 50 not to get screened," Ms. Visco said. "But what we are telling women is that we don't have a good screening test to detect breast cancer early and we're not sure what to do when we find it early." It is time, she said, to focus on these issues rather than assume that mammograms are the answer.

Evidence-Based Medicine

New York Times Magazine, December 9, 2001

By JACK HITT

When visiting our family doctor, most of us feel secure in the belief that modern science has purged medicine of such practices as cupping and bloodletting. But according to a recent article in the journal Patient Care, ''Some experts estimate that only 20 percent of medical practices are based on rigorous research evidence.'' The rest are based on what has been published in books repeatedly without independent testing or what doctors have always said should work. In other words, it's a kind of folklore.

A revolution is erupting in the wards of practical medicine these days, one defined recently by The British Medical Journal as ''the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients.'' The revolution is called evidence-based medicine, or E.B.M., and many traditional treatments are being run through the machinery of the scientific method and being found wanting.

One common E.B.M. approach is meta-analysis: collating data from far-flung studies to come up with a definitive answer to a medical question. Such studies are overthrowing some conventional wisdom. Mammogram screenings? They don't save lives. Remember the placebo effect? It doesn't exist. E.B.M. is also credited with validating some simple cures. Most people know that if you have a heart attack, you should immediately take an aspirin. Thank an E.B.M. study for proving that this works.

After colds, the second-most-common reason for a visit to the doctor is lower-back pain. The ''treatment'' has always been bed rest. Why? Because, as a recent article explained, ''The notion that rest is therapeutic and will relieve pain dates back to Hippocrates.'' But now that E.B.M. studies have used science instead of oral tradition to test this notion, they have found that bed rest ''may delay return to functional status.'' What works better? Light exercise and getting back on your feet. This past June, the Agency for Healthcare Research and Quality integrated the no-bed-rest approach into its guidelines. This new standard of care, which will probably save billions of dollars in unnecessary sick leave, marks the end of 2,400 years of misguided treatment.

E.B.M. is yet another idea that can be credited to the computer revolution. Doctors have long known that they learn very little after med school when their exhausting schedules and the baffling profusion of 4,000 monthly professional journals make it nearly impossible to keep up with innovations in treatment. The E.B.M. movement began when six doctors in Canada came up with the idea of skimming the most dependable studies and crunching the results into an accessible, reliable database.

Indeed, in the wake of E.B.M., journals are filling with terms that sound almost anthropological to describe longstanding treatments: ''local custom,'' ''witch-doctoring,'' ''myth.'' Or as one article this fall put it, ''This process of examining beliefs that have been based primarily on teaching and empirical experience rather than evidence has been compared to stripping the curtain away from the Wizard of Oz to reveal an ordinary man.