OUTCOME ORIENTATION a misconception of probability that harms medical research and practice

摘要

Human understanding of randomness and variation is shaped by a number of cognitive biases. In this article, we relate a lesser-known cognitive bias, the "outcome orientation," to medical questions and describe the harm it can do to medical research and practice. An outcome orientation means predicting outcomes one at a time, neglecting the fact that each event may be a member of a group of comparable events. People who reason according to an outcome orientation assign a subjective degree of belief to an outcome, but do so in a way that is incompatible with Bayesian reasoning or any other standard laws of probability. Instead of accepting that uncertainty is inevitable and generalizing from the frequency of similar events, the outcome orientation prefers one-off causal narratives. In medicine, the outcome orientation therefore erodes support for randomized controlled trials in favor of reductionist approaches. The rhetoric of personalized medicine resonates with, and can promote, the outcome orientation by emphasizing how the measurable attributes of individual patients, rather than chance or unknowable factors, causally produce each particular patient's outcome.