47,000 Women Could Die As a Result of the New Mammogram Guidelines

Cost-benefit analysis can kill. The failure to distinguish statistics from arithmetic can kill. In the current debate over mammograms, the number of women projected to be at risk of death due to cost-benefit analysis is about 47,000.

That is the approximate number projected to die by the United States Preventive Services Task Force (USPSTF), if its recommendations on scaling back mammograms had been accepted. It is the task force's number, if you do the arithmetic, which it apparently did not.

USPSTF statistics say that the life of “only” one woman in 1,900 will be saved if mammograms start at age 40 instead of age 50. In other words, a 40-year-old woman’s “risk” of dying from breast cancer in the next 10 years is only 1 in 1,900. That seems like no risk at all. 1 divided by 1,900 equals .000526. About half a woman per 1,000. Minuscule, right?

Now, how many women in America would be affected?

The most recent (July 2008) census figures say there are about 304,000,000 Americans, of which 50.7 percent are female. That’s about 154,000,000 females. Roughly 80,000,000 of them are under 40 and about another 20,000,000 between 40 and 50. Of the 80,000,000 under 40, each one, under the proposed guidelines, would not get a mammogram until age 50. If “only” 1 in 1,900 die as a result, that would be .000526 times 80,000,000, which equals about 42,000.

In short, moving the mammogram age from 40 to 50 would result in the deaths of 42,000 women now 40 or under, according to the statistics of the Preventive Services Task Force. Of the 20,000,000 between 40 and 50, it could mean the deaths of as many as 10,500 women, though the figure may be somewhat lower because half are more than halfway through the critical period. There might be as few as half, say, 5,000 deaths. Adding 42,000 and 5,000, we get a ballpark figure of 47,000 of currently living American females who would die needlessly under the proposed task force restriction on mammograms. Of course, as more are born, the absolute numbers would go up.

What is at issue is called “framing.” The Preventive Services Task Force chose the probability of risk frame: only 1 in 1,900. But the arithmetic frame reveals the more important truth.

Framing, in this case as in so many others, is a matter of life and death. Take the framing in the New York Times (Nov. 18, 2009) in the front-page news analysis by Kevin Sack and in the op-ed by Robert Aronowitz. Sack frames the mammogram debate as the “science of medicine” versus “medical consumerism.” Aronowitz calls it “wishful thinking” that early mammograms could help, and speaks of “the very small numbers of lives potentially saved.”

You can see why cost-benefit analysis can kill. Its use isn’t science. Real scientists do arithmetic as well as statistics. Medical science is about real people, not percentages or statistics, especially when large numbers of real people are involved and small differences in risk can produce large numbers of deaths.

The Preventive Services Task Force also uses the “harm” frame. The task force observes that more mammograms mean more false positives and claims that false positives do “harm.” But no science is presented showing that the harm done is greater than the deaths of 47,000 women.

What is the "harm"? Anxiety and unnecessary biopsies from false positives are listed as the harms. My wife had such a false positive. The anxiety came for economic reasons: she had to wait for a biopsy because no one who could perform one was present when the mammogram was done, due to economic restrictions. The biopsy when it came was simple: a needle inserted to withdraw fluid, like taking a blood sample. No harm. If the biopsy had been done immediately, there would have been no need for anxiety. But the task force does not recommend immediate biopsies as a way to eliminate such “harm.”

Aronowitz also claims the figures show that mammograms haven’t helped prevent breast cancer. He observes that the rate of 28 breast cancer deaths per 100,000 people has not changed substantially since the 1950s, despite more mammography and better treatments. But that could mean, and probably does mean, that there has been an increase in breast cancer offset by earlier detection and better treatment, saving tens of thousands of lives, but not affecting the overall rate. But he did not consider the possibility that the occurrence of breast cancer might have increased, while the rate of deaths did not change because of earlier detection due to mammograms.

I suspect that the real harm intended is economic harm – the costs of the “unnecessary” mammograms and biopsies. But the task force gives no figures weighing the economic costs versus the human cost of the deaths of 47,000 women. Now, in cost-benefit analysis, a commonly cited figure for the value of an American life is $6.5 million. Forty-seven thousand times 6.5 million is $305,500,000,000. That is, 305 billion 500 million dollars. Of course, that would be spread over the next 40 years, but it’s not clear that such a cost-benefit analysis would make this less than the cost of mammograms and biopsies, all moral issues and human costs aside. Unfortunately, the Preventive Services Task Force doesn’t do the calculation, so my figures may be off. The exact figures are not the point. The point is to go beyond rates to numbers.