Kahneman versus Klein on medical decision-making
I just finished reading Daniel Kahneman’s bestseller Thinking, Fast and Slow. Regular readers of this blog will probably surmise that I don’t share the Nobel prize winner’s interest in cognitive biases (not that they don’t exist, of course).
Kahneman, like most social scientists, generally espouses an empiricist’s outlook on the human mind (the Humean mind…). For him, decisions are essentially a matter of calculation, and he documents the many ways we fool ourselves, and the many ways we can fall short. In that respect, the book is very interesting and often amusing.
Of course, one benefit of the research Kahneman has conducted with Amos Tversky is to poke holes in the flawed model of the world as made up of fully rational and selfish “Econs” as Richard Thaler put it. That model underlies much of mainstream, neoclassical economic theory.¹
What I found most interesting in Kahneman’s book is chapter 22, which deals with expert intuition, and whether it can be trusted. In that chapter, Kahneman candidly discusses his differences with Gary Klein, for whom he shows much respect. I wrote about Klein and how his insights can help us think about medical decision-making in my post “How experts really decide.”
According to Kahneman, he and Klein have come to agree that their differences may have to do with the kind of settings under which decisions are made, although he does not describe in formal terms what these different settings might be.
What is apparent, though, is that Klein, who studies his subjects “in the field,” trusts his own (and the average person’s) intuitive ability to identify who’s a good decision maker. Kahneman, ever the Humean skeptic, confines his study to the laboratory setting where repeatable and repeated measurements can sort out winners from losers.
In that regard, we should be careful not to draw premature conclusions about who best captures the essence of medical decision-making by simply looking at how doctors operate in the modern healthcare environment. It is quite possible—and, in fact, undoubtedly the case—that medical decision-making is very much conditioned by the theoretical framework under which the healthcare system operates.
If, as I have argued, we implicitly espouse the “machine model” of the organism, we will find plenty of examples that seem to vindicate Kahneman’s views (and Paul Meehl‘s), and contradict Gary Klein’s thesis. The machine model, however, may not be the best and most natural way to help us think about health and medical care.
You can find other articles that address different aspect of medical decision-making here.
1. Yet, even in his criticism of that model, and of “Bernoulli’s error,” Kahneman may be falling short, as David Howden has pointed out in his review of the book.