When Less Is More: Balancing Harms, Benefits of Mammography

reprinted from Issue 14, Spring 2012 of Frontiers of Medicine (PDF)

Karla Kerlikowske, MD (left), and Louise Walter, MD, have led groundbreaking research into the best uses of screening mammography.



Early in her career, Louise Walter, MD, struggled with the decision about which of her elderly patients she should encourage to obtain a screening mammogram. “When I looked at the guidelines, the U.S. Preventive Services Task Force recommended that you stop at age 70, the American Geriatrics Society said to stop at 85, and the American Cancer Society said, you never stop screening,” recalls Walter, now a geriatrician at the San Francisco Veterans Administration Medical Center (SFVAMC). “I was like, ‘What do you do?’”

That question launched a research career devoted to developing more tailored guidelines for cancer screenings among older patients. Walter is one of a number of UCSF faculty members whose discoveries are helping to guide the use of screening tools in ways that have the best chance of helping patients while minimizing harm.

One of the challenges that Walter faced is that most randomized controlled clinical trials exclude patients above age 70. “With most of the things we do in medicine, we are extrapolating from things that have been shown to work in a younger population,” says Walter. “Also, the one-size-fits-all approach to medical care based on age does not work. I’d see a really healthy 80-year-old woman who was hiking up Mount Tam, and then I’d see a really sick 75-year-old with severe dementia. Their medical care should be tailored to their health, not just their age.”

With her colleagues, Walter developed a framework for individualized decision-making, helping physicians weigh the harms and benefits for a particular patient, as well as the patient’s life expectancy and values. Because mammography is not a very precise test, it yields many more false-positive test results than diagnoses of actual cancer. Consequently, patients may receive unnecessary biopsies, as well as surgery, chemotherapy and radiation to treat slow-growing cancers that would never have been life threatening.

Walter also found that frail elders are at increased risk of suffering negative outcomes from false-positive test results and unnecessary cancer treatment. “The person has a major surgery, gets wound infections, rips off all their bandages because they get delirious, ends up spending a year of their life getting wound care, and then dies of a stroke,” says Walter. “That is a huge harm from a screening mammogram, in someone that we knew had a very low chance of benefit.”

Walter has also done research on the unintended downstream harms suffered by frail elders who are screened for other types of cancer, including colon, cervical and prostate cancer. “The lag time for benefit from these screenings is at least five to ten years,” says Walter. “We should be doing things that are going to benefit these patients in the next couple of years, not in 10 years. We can do so many things to help frail elders have a better quality of life, such as helping with their depression or maintaining their independence at home, rather than sending them for a screening test that is trying to find an asymptomatic problem when they already have multiple problems.”

“When people come in wanting antibiotics for a cold, when you tell them all the reasons why you don’t recommend it, the vast majority say, ‘Okay, I don’t want it,’” says Walter. “The same thing with mammograms. When I explain to the caregivers of my frail, elderly patients with dementia what could happen, they say, ‘Oh, we don’t want that. Let’s focus on these other issues.’ The problem is, it takes time to have that conversation. I am always struck by how it is easier to do something – give people antibiotics, send them for a mammogram – rather than have a harder conversation about why these could be potentially harmful. I think that is why we sometimes do too much in medicine.”

Stratifying Screening by Breast Cancer Risk

Another UCSF researcher who has studied mammography is Karla Kerlikowske, MD, who also serves on the faculty at the SFVAMC. She recently co-authored a study published in the Annals of Internal Medicine that found that after 10 years of annual screening, more than half of women will receive at least one false-positive test result. She and her colleagues also found that receiving a mammogram every other year instead of annually reduced the likelihood of a obtaining a false-positive result from about 60 percent to about 40 percent within 10 years. There was no statistically significant increase in the number of late-stage cancer diagnoses by screening every two years rather than annually.

These findings are in line with guidelines issued in 2009 by the U.S. Preventive Services Task Force, a respected, independent panel of medical experts. The task force recommended biennial screening mammography for women aged 50 to 74 years, replacing its earlier guidelines recommending that starting at age 40 women should have mammograms every one to two years. The new guidelines were controversial, and were cited by opponents of the health care reform bill as an example of government “rationing” of health care.

“Countries all over the world use a two-year screening interval,” says Kerlikowske. “In the U.S., somehow we think that annual is better, but there is no data to support that recommen-dation. From a benefit-harm ratio, it doesn’t make sense. You get the same benefit from screening every two years, and reduced harms.”

Another challenge is that not all breast cancers are created equal. For example, ductal carcinoma in situ is a premalignant abnormality located inside the milk ducts, which has not invaded nearby tissues, and is more of a risk factor rather than cancer itself. There are also slow-growing invasive cancers that can be detected by clinical breast exam or mammography and have the same impact on life expectancy. Highly aggressive cancers are often deadly even if they are identified through mammography. In the remaining 20 to 30 percent of breast cancers, identi-fication of tumors with mammography can improve outcomes and potentially decrease women’s chance of dying of breast cancer. Unfortunately, mammography is unable to distinguish among these different types of breast cancer.

Kerlikowske is also developing better ways to determine women’s risk level for developing breast cancer, and tailoring screening protocols that maximize benefit while minimizing harm. For example, she found that a 40-year-old woman with dense breasts and a family history of breast cancer has a similar risk level as a typical 50-year-old, and should probably have screening mammography every two years. A 50-year-old with fatty breasts and no family history probably has reduced risk, and probably only needs to be screened every three years. Women with strong family history of breast cancer may be identified with the BRCA 1 or 2 gene mutations and are at higher risk and therefore have different screening protocols than the general population.

“The whole idea is to risk-stratify women for screening,” says Kerlikowske. “Just like for coronary artery disease, where you don’t give everybody simvastatin for their cholesterol, we’re trying to develop an analogous approach for breast cancer screening.”

“Right now, people think that if a little health care is good, then more health care is better, which is not always the case," said Rita Redberg, MD, MSc, FACC, director of the Women’s Cardiovascular Services

What’s Best for Patients?

Both Kerlikowske and Walter have produced groundbreaking work in discovering how to best use screening tools like mammography. “Their research has really been seminal,” says Rita Redberg, MD, MSc, a UCSF cardiologist who is the editor of the Archives of Internal Medicine.

Redberg has also been an active contributor to this wider discussion. The backlash against the U.S. Preventive Services Task Force guidelines on mammography inspired a long conversation between her and Deborah Grady, MD, MPH, associate dean for Clinical and Translational Research. “We talked about how the reaction to that very scientific and evidence-based announcement was negative, and how we needed a whole public re-education process,” says Redberg. “Right now, people think that if a little health care is good, then more health care is better, which is not always the case.”

The two developed the idea for “Less is More,” and launched the series in the Archives of Internal Medicine in 2010. Grady is section editor, and the articles highlight areas of health care with no known benefit, and definite harms. Several of the articles have prompted the FDA to issue warnings about faulty drugs and devices, which were described in the series. Redberg and Grady also welcome more UCSF contributions of research, cases and commentaries to this series.

“Part of our American culture is that we embrace technology,” says Redberg. “But not all technology is great for all people at all times…. People are quick to label as rationing, anytime the government is not recommending particular health tests. Perhaps if the focus of the communications was on the fact that these additional mammograms were hurting women, there would have been a more positive reaction.”

She notes that cost is not a criterion for the series. “We’re looking at it from the perspective of, what’s best for the patient,” says Redberg. “Certainly, that means not putting themselves at risk when there’s no benefit.”




Frontiers of Medicine Main Page