A new review of studies on breast cancer and elderly women by researchers at UCSF and the Harvard School of Medicine suggests that doctors should focus on how long a patient is likely to live when deciding whether to order mammograms for their oldest female patients rather than order them routinely. The researchers say their study shows that the harms of screening likely outweigh the benefits unless women are expected to live at least another decade.

According to the study authors, national guidelines recommend that doctors make individualized screening decisions for women 75 and older. But the authors of the recent study review say that since women in that age group were not included in mammography trials, there is no evidence that screening helps them live longer, healthier lives.

According to the researchers, potential harms from routine screening in elderly women can include unnecessary treatment for slow-growing cancers or pre-cancerous lesions that pose no real threat to the women’s lives. The study authors recommend that women who are expected to live a decade or more should talk with their doctors and weigh the potential benefits of diagnosing treatable cancer through mammography against the possibility of being misdiagnosed or treated aggressively for a cancer that might not have posed any harm.

“People should be informed that everything we do in medicine can have good and bad effects, and that goes for mammography,” said Louise Walter, MD, UCSF professor of medicine and chief of the Division of Geriatrics.

The researchers reviewed all mammography studies conducted from 1990 to 2014 that identified risk factors for late-life breast cancer in women 65 and older, as well as the studies that assessed the value of mammography for women 75 and older. Since there were no randomized trials of the benefits of screening women over 74, they could not say whether mammography was beneficial for those women. Longitudinal studies found that healthy older women who were screened with mammography were less likely to die from breast cancer, but screening was not beneficial for women with serious medical problems.

Modeling studies indicate mammograms would prevent two cancer deaths for every 1,000 women in their 70s who were screened every two years for 10 years. However, these studies also predicted that about 200 of those women would receive test results indicating they had cancer when they did not, and about 13 would be treated for cancers that posed no harm.

The study was published in the JAMA.