The tempest that greeted the United States Preventive Services Task Force guidelines on mammography screening for women in their 40s prompted the Senate to insert a mandate in its health care reform bill that every insurer cover every mammography screening test at no cost to beneficiaries. If it passes, it will spark an upsurge in mammography screening, especially among women under 50, and raise the nation's health care tab.

The Journal of the American Medical Association this morning provides a timely article (subscription required) reminding physicians and women about the serious health costs of adopting that policy.

The authors, Dartmouth's Steven Woloshin and Lisa Schwartz, use their trademark "number needed to treat" analysis to point out that:

Without screening, 3.5 of 1000 women in their 40s will die of breast cancer over the next 10 years (ie, 996.5 of 1000 will not die of the disease).

Screening reduces the chance of breast cancer death from 3.5 to about 3 of 1000. In other words, 2000 women between 40 and 49 must be screened annually for the following ten years to save one life.

For most women with cancer, screening generally does not change the ultimate outcome; the cancer usually is just as treatable or just as deadly regardless of screening.

Finding cancers that were never destined to cause symptoms or result in death is the biggest problem with mammography, especially among younger women. Since it is impossible to know which cancers caught early are benign, all are treated with surgery, chemotherapy, radiation, or some combination. Overdiagnosed women undergo treatment that can only cause harm, and must live with the ongoing fear of cancer recurrence.

While only 7% of women believe there could be breast cancers that grow so slowly that leaving them alone would not affect their health, randomized clinical trials have consistently shown that the groups undergoing mammography have more breast cancer, even after 15 years of follow-up. This persistent difference represents overdiagnosis.

Estimates of the rate of overdiagnosis range from 2 women overdiagnosed for every breast cancer death avoided in one trial, to 10 to 1 in another.

A companion piece by Steven Woolf of Virginia Commonwealth University points out that the guidelines recommend physicians discuss these relative risks and benefits with women before proceeding with annual mammograms. It wasn't a blanket call for a ban.

Advocates of mammography and cancer survivors often belittle these harms, but a moral duty exists when subjecting millions of asymptomatic women to a procedure that benefits relatively few. Whether hundreds of women should endure the consequences of inaccurate mammograms to save 1 woman's life is a legitimate ethical question.

"The politicalization of medical care is wrong," the Woloshin and Schwartz conclude. "Promoting screening irrespective of the evidence may garner votes but will not create healthier voters. It may do the opposite."