Action vs. inaction

2 weeks ago, the U.S. Preven­tive Ser­vices Task Force came out with new recom­men­da­tions on breast can­cer screen­ing, in­clud­ing, “The USPSTF recom­mends against rou­tine screen­ing mam­mog­ra­phy in women aged 40 to 49 years.”

The re­port says that you need to screen 1904 women for breast can­cer to save one woman’s life. (It doesn’t say whether this means to screen 1904 women once, or once per year.) They de­cided that sav­ing that one woman’s life was out­weighted by the “anx­iety and breast can­cer worry, as well as re­peated vis­its and un­war­ranted imag­ing and biop­sies” to the other 1903. The re­port strangely does not state a false pos­i­tive rate for the test, but this page says that “It is es­ti­mated that a woman who has yearly mam­mo­grams be­tween ages 40 and 49 has about a 30 per­cent chance of hav­ing a false-pos­i­tive mam­mo­gram at some point in that decade and about a 7 per­cent to 8 per­cent chance of hav­ing a breast biopsy within the 10-year pe­riod.” The re­port also does not de­scribe the pain from a biopsy. This page on breast biop­sies says, “Ex­cept for a minor sting from the in­jected anes­the­sia, pa­tients usu­ally feel no pain be­fore or dur­ing a pro­ce­dure. After a pro­ce­dure, some pa­tients may ex­pe­rience some sore­ness and pain. Usu­ally, an over-the-counter drug is suffi­cient to alle­vi­ate the dis­com­fort.”

So, if we as­sume bi­an­nual mam­mo­grams, the con­clu­sion is that the worry and in­con­ve­nience to 286 women who have false pos­i­tives, and 71 women who re­ceive biop­sies, is worth more than one woman’s life. If we sup­pose that a false pos­i­tive causes one week of anx­iety, that’s a lit­tle over 5 years of anx­iety, plus less than one year of sore­ness.

(I heard on NPR that the USPSTF that made this recom­men­da­tion in­cluded rep­re­sen­ta­tives from in­surance com­pa­nies, but no ex­perts on breast can­cer. So per­haps I’m bark­ing up the wrong tree by look­ing for a cog­ni­tive bias more sub­tle than fi­nan­cial re­ward.)

I’m not shocked at the wrong­ness of the con­clu­sion; just at its di­rec­tion. The trade-off the USPSTF made be­tween anx­iety and death is only 2 or­ders of mag­ni­tude away from some­thing that could be defended as rea­son­able. Usu­ally, gov­ern­ment agen­cies mak­ing this trade­off are off by at least that many or­ders of mag­ni­tude, but in the op­po­site di­rec­tion. (F-18 ex­am­ple deleted.)

So, what cog­ni­tive bias let this gov­ern­ment agency move the dec­i­mal point in their head at least 4 points over from where they would nor­mally put it?

I think the key is that this re­port recom­mended in­ac­tion rather than ac­tion. In cer­tain con­texts, in­ac­tion seems safer than ac­tion.

Imag­ine what would hap­pen if the FDA were faced with an iden­ti­cal choice, but with ac­tion/​in­ac­tion flipped: Say you have an anti-anx­iety drug, which will elimi­nate anx­iety of the same level caused by a false-pos­i­tive on a mam­mo­gram, in 15% of the pa­tients who take it—and it will kill only 1 out of ev­ery 2000 pa­tients who take it. Per week.

Would the FDA ap­prove this drug? Ap­proval, af­ter all, does not mean recom­mend­ing it; it means that the de­ci­sion to use it can be left to the doc­tor and pa­tient. The USPSTF re­port stressed that such de­ci­sions must always be left up to the doc­tor and pa­tient; by the same stan­dards, the FDA should cer­tainly ap­prove the drug. Yet I think it would not.

A puz­zle is why we have the op­po­site bias in other con­texts. When Congress was de­bat­ing the bank bailouts and the stim­u­lus pack­age, a lot could have been said in fa­vor of do­ing noth­ing; but no one even sug­gested it. Em­piri­cally, we have a much higher suc­cess rate at in­ter­ven­ing in health than in eco­nomics. Yet in health, we reg­u­late ac­tions as if they were in­her­ently dan­ger­ous; while in eco­nomics, we see in­ac­tion as in­her­ently dan­ger­ous. Why?

ADDED: Per­haps we see reg­u­la­tion as in­her­ently safer than a lack of reg­u­la­tion. “Reg­u­lat­ing” (ban­ning) drugs is seen as “safe”. “Reg­u­lat­ing” the econ­omy, by bailing out banks, pass­ing large stim­u­lus bills, and pass­ing new laws reg­u­lat­ing banks, is seen as “safe”. Recom­mend­ing or not recom­mend­ing mam­mo­grams isn’t reg­u­la­tion ei­ther way; there­fore, we per­ceive it neu­trally.