Mammogram Math

John Allen Paulos in the New York Times Magazine:

ScreenHunter_01 Dec. 15 10.52 In his inaugural address, Barack Obama promised to restore science to its “rightful place.” This has partly occurred, as evidenced by this month’s release of 13 new human embryonic stem-cell lines. The recent brouhaha over the guidelines put forth by the government task force on breast-cancer screening, however, illustrates how tricky it can be to deliver on this promise. One big reason is that people may not like or even understand what scientists say, especially when what they say is complex, counterintuitive or ambiguous.

As we now know, the panel of scientists advised that routine screening for asymptomatic women in their 40s was not warranted and that mammograms for women 50 or over should be given biennially rather than annually. The response was furious. Fortunately, both the panel’s concerns and the public’s reaction to its recommendations may be better understood by delving into the murky area between mathematics and psychology.

Much of our discomfort with the panel’s findings stems from a basic intuition: since earlier and more frequent screening increases the likelihood of detecting a possibly fatal cancer, it is always desirable. But is this really so?

More here.