Anecdotal, non-quantitative, first-person overindulgent, with dramatic pauses for effect — what's the good of yet another fuzzy set of medical stories? Well, in the case of Jerome Groopman's How Doctors Think, quite a lot of good. It might be better titled How Doctors Should Think or perhaps How Good Doctors Sometimes Think. Groopman, occasionally featured in the New Yorker, writes about critical thinking, cognitive fallacies, and the need for proper (Bayesian) reasoning. As he notes in the Introduction, "... at each step along the way, we see how essential it is for even the most astute doctor to doubt his thinking, to repeatedly factor into his analysis the possibility that he is wrong. We also encounter the tension between his acknowledging uncertainty and the need to take a clinical leap and act."

So there are comments on anchoring and availability and confirmation biases, affective error and stereotyping, emotion, temperament, and irrational hope. There's citation of Kahneman and Tversky's work on Thinking, Fast and Slow. There's frank exposure of gift-giving by big pharmaceutical companies as they push new drugs and pressure doctors to prescribe them. There's philosophical musing about utility, how different patients can place different values on suffering versus longer life, on worst-case versus median expectation. There's thoughtful commentary on creativity and when it's most needed.

Groopman underscores the lack of simple answers in medicine, as in life. From Chapter 3, "Spinning Plates", for instance:

Most people believe that decisions in the ER must be made instantly, but Alter said that "is a misperception that we doctors in part foster." In order to think well, especially in hectic circumstances, you need to slow things down to avoid making cognitive errors. "We like the image that we can handle whatever comes our way without having to think too hard about it — it's kind of a cowboy thing." As if being swift and decisive saves lives. But as Alter put it, he works with "studied calm," consciously slowing his thinking and his actions with each patient in order not to be distracted or pressed by the hectic and sometimes chaotic atmosphere.

This is good stuff, generally well-balanced and well-written. A few more equations, and it would be even better (ok, at the cost of most of its readership!) ...

(cf. Statistics - A Bayesian Perspective (2010-08-13), Medicine and Statistics (2010-11-13), Introduction to Bayesian Statistics (2010-11-20), Statistical Hypothesis Inference Testing (2013-12-01), ...) - ^z - 2014-10-28