Hindsight Bias: An Impediment to Accurate Probability Estimation in Clinicopathologic Conferences

Abstract

Although clinicopathologic conferences (CPCs) have been valued for teaching differential diagnosis, their instructional value may be compromised by hindsight bias. This bias occurs when those who know the actual diagnosis overestimate the likelihood that they would have been able to predict the correct diagnosis had they been asked to do so beforehand. Evidence for the presence of the hindsight bias was sought among 160 physicians and trainees attending four CPCs. Before the correct diagnosis was announced, half of the conference audience estimated the probability that each of five possible diagnoses was correct (foresight subjects). After the correct diagnosis was announced the remaining (hindsight) subjects estimated the probability they would have assigned to each of the five possible diagnoses had they been making the initial differential diagnosis. Only 30% of the foresight subjects ranked the correct diagnosis as first, versus 50% of the hindsight subjects (p less than 0.02). Although less experienced physicians consistently demonstrated the hindsight bias, more experienced physicians succumbed only on easier cases.

The effects of mental contamination on the hindsight bias: Source confusion determines success in disregarding knowledge

“…Hindsight bias has been well-established as a robust effect whose consequences can be observed in physicians' overconfidence in the ability to diagnose difficult diseases (Dawson et al, 1988), monetary awards granted by biased juries (Casper et al, 1988), overconfidence in the correctness of answers to general knowledge questions (Davies, 1987;Hasher, Attig, & Alba, 1981;Wood, 1978), and, as the current study demonstrates, the inability to disregard new information in a subject with which one is already relatively familiar. Prior research in the area has focused primarily on hindsight procedures which either required participants to remember in hindsight their foresight responses (e.g., Davies, 1987;Hasher, Attig, & Alba, 1981;Wood, 1978) or asked one group of participants to make foresight judgments that were compared to the hindsight judgments of a different set of participants (e.g., Dawson et al, 1988). In the current study, participants were not asked to remember foresight judgments, nor did they give responses which were compared to those of foresight or hindsight counterparts.…”

Hindsight bias, outcome knowledge and adaptive learning

“…Similar susceptibility to hindsight bias occurs in clinicopathological conferences where less experienced physicians succumbed to the bias with both easier and more difficult cases, while more experienced physicians manifested the bias only with the easier cases. 12 …”

“…If insufficient attention is given to portraying the uncertainty and other possibilities that exist along the foresight path, then audience members, placed in an essentially passive observer role, can also fall victim to hindsight bias and wonder why the discussant couldn't see the obvious. 12 For maximum instructional value the discussant needs to withhold his or her final diagnosis from audience members who, in turn, need to become active participants in working out the diagnosis. Careful attention needs to be given to recreating the partial information and uncertain conditions which existed along the foresight path at the time critical decisions were being made.…”

Rationality in medical treatment decisions: is there a sunk-cost effect?

“…Medical decision making research has sought to uncover some of the parallels between cognitive biases which exist in the non-medical world and in the medical world (Detmer et al, 1978;Elstein et al, 1978;Dawson and Arkes, 1987;Elstein, 1988; for an inventory of medical decision-making biases, see Hershberger et al, 1994). Some of the biases that have been shown to infl uence medical judgment include the omission bias (Asch et al, 1994), availability bias (Poses and Anthony, 1991), hindsight bias (Arkes et al, 1981;Dawson et al, 1988), a bias to ignore negative evidence when attempting to synthesize information (Mazur and Hickam, 1990), framing effects (McNeil et al, 1984) and outcome bias (Gruppen et al, 1994).…”

“…For example, Dawson et al (1988) asked physicians attending clinicopathologic conferences to estimate the probability of fi ve possible diagnoses. They found that physicians who were informed of the correct diagnosis prior to making their probability estimates ranked it higher in their differential diagnosis than physicians who were not told the correct diagnosis beforehand.…”