Decision aid helps inform older women of breast cancer screening benefits, harms

Robert Smith, PhD

When we look at the burden of disease, our focus is on the incidence of breast cancer after age 75, and the deaths that occur that would be attributable to that diagnosis. A significant fraction of deaths after age 75 is attributable to a diagnosis that occurred before age 75, which is not the issue in this analysis. Still, it is important to appreciate that a diagnosis of breast cancer after age 75 accounts for a significant proportion of the annual number of U.S. breast cancer deaths.

Since the average 75-year-old woman is expected to live 13 more years, this is an age group that needs regular assessment for the value of mammography screening. At the same time, as Schonberg and colleagues note, some women may be in poor health, have limited longevity due to life-limiting comorbid conditions and would be unlikely to benefit from the early detection of an asymptomatic cancer if that cancer would not have produced symptoms before they died from some other underlying condition. Further, the treatment of that cancer could further diminish their quality of life.

We have evidence that women with known, serious comorbid conditions still are referred to mammography, and died of a non-breast cancer-related cause soon after. Screening in these women should be avoided. In addition, mammography screening rates drop off in older women during a period when a majority of them are still likely to benefit from screening. This is due to their own sense that they are not at risk for breast cancer, their doctor’s declining attention to prevention and early detection as individuals get older and the likelihood that some non-life-threatening chronic conditions are the dominant feature in their health care, so screening gets neglected. It is also challenging and time consuming to assess longevity, especially out 10 years, and this projection can change from year to year. Thus, it is a big problem that women in this age group are under-screened and over-screened.

The American Cancer Society recommends that as long as a woman is expected to live 10 years or more, and she is in good health, she should continue getting mammograms. If a woman could be expected to live longer than 10 years, she should have an opportunity to avoid a premature death from breast cancer. There is value in decision aids to help women and doctors think about their current circumstances to support decisions about continuing or discontinuing screening. As you might expect, it is difficult to convey to a patient that they are facing the possibility of death and that preventive care is no longer an important priority in their health care.

A major problem in Schonberg and colleagues’ study comes from the standpoint of which women were included. A woman had to have undergone a recent mammogram to be eligible. In effect, this means that the outcome of interest was the proportion of women who were dissuaded from continuing to get mammograms. About 9% of women who were expected to live 10 or more years did not get a mammogram within 18 months. I would have preferred to see the intervention delivered to all women in the eligible age group to determine how well the tool performed in informing women in good health that it was too early to stop screening.

Building decision aids is complicated and most suffer from inherent limitations. This one is no exception.

First, the decision aid says we do not know if mammography benefits women 75 and older. This statement is overly orthodox with respect to evidence-based medicine. We do know, by any benchmark, breast cancer is a progressive disease in this age group, that it accounts for a significant fraction of annual breast cancer deaths, and that most breast cancers have better prognosis when you find them early. While we do not have any randomized controlled trials in this age group, that is a function of study design decisions made 30 to 40 years ago. There won’t be an opportunity to study screening in this age group with a randomized controlled trial, so we have to rely on observational data. So, this is a vague and misleading statement. Odd, then, that they present numbers based on a 20% mortality reduction. We do not know that. Could be higher, could be lower.

Second, it is not really relevant that breast cancers grow slower in older women. They do, but not that much slower compared with younger women.

Third, in the section on why women get mammograms, the decision aid says that experts (not some experts) think that small breast cancer found on an older woman’s mammogram may not cause problems for 5 to 10 years. Maybe. Maybe not.

Fourth, the concept of overdiagnosis is introduced. If a woman has been undergoing mammography, the odds that she would be diagnosed with a truly nonprogressive cancer at this age are theoretically nonexistent. Could she be diagnosed with a breast cancer that she would not have known about if she had not undergone screening because she dies from something else? That is more likely.

Fifth, there is no mention of statistics that the American Cancer Society states in its guidelines — that many breast cancer deaths occur in women diagnosed with breast cancer after they are 75.

Sixth, at this age, there is little value to the patient in providing her with statistics for women older than 75 years. If she is 75, the numbers that are relevant to her are the numbers for the next 5 to 10 years.

Schonberg and colleagues’ discussion of treatments would be dissuading to anyone. But there are studies underway challenging whether older women truly need aggressive treatment. Recent studies have shown that women with small, more favorable prognosis treatments also do not need very aggressive treatment (specifically, chemotherapy). Many of the harms associated with aggressive therapy, which in an older woman could measurably diminish quality of life, may not be needed as part of her treatment plan.

PCPs need to recognize that as women grow older, it is important to assess their preventive health plans and determine when to initiative discussions about cessation of screening. In general, the American Cancer Society advises that this should begin around the age of 75 years, but younger if a woman is in poor health. It also is important that everyone has a PCP who can take responsibility for preventive health, since it will be unlikely that the cardiologist or endocrinologist will feel an obligation to refer for mammography screening, or would be prepared to have these conversations. As the authors note, there are tools that help clinicians assess longevity and the likelihood of benefiting from screening, which will aid in these discussions.

References:

Oeffinger KC, et al. JAMA.2015;doi:10.1001/jama.2015.12783.

Schonberg MA, et al. J Am Geriatr Soc. 2013;doi:10.1111/jgs.12123.

Schonberg MA, et al. JAMA Intern Med. 2014;doi:10.1001/jamainternmed.2013.13639.