Date of Award
Doctor of Philosophy (PhD)
Dr. Lee R. Brooks
By definition, the act of decision-making requires the consideration of multiple alternatives. Medical decision-making is no exception in that to decide whether or not a particular diagnosis is probably correct, a clinician must consider other diagnostic possibilities. Failure to do so is referred to as premature closure. If clinical instruction is provided regarding this diagnostic hazard, it tends to consist of advice to think of differential diagnoses before concluding one's diagnostic search. However, it is not yet known if such instruction is effective, let alone sufficient to eliminate diagnostic errors that arise from the tendency to under-weight non-focal hypotheses. I examined this issue by presenting case histories to experimental participants and manipulating the diagnoses that were explicitly presented with a request to assign probability ratings, a paradigm extensively used by Tversky and Koehler (1994). Fluctuations in the probability assessments provided an empirical measure of the degree to which diagnostic hypotheses were considered when they remained implicit in an "all other diagnoses" category. Results showed that diagnosticians tend to under-weight diagnostic alternatives that they are not explicitly asked to evaluate (Experiments 1 & 2). Furthermore, this effect does not solely arise from non-generation of alternative diagnoses. Diagnoses that medical students themselves generate are under-weighted relative to when the same diagnoses are presented explicitly (Experiment 3). By extension, this suggests that educational instruction to 'consider alternatives' is insufficient to help students eliminate premature closure. I hypothesized further that premature closure might result from a confirmation bias because the explicit presentation of a diagnostic alternative might cause diagnosticians to differentially process the evidence relevant to the diagnostic possibilities. That is, judges might focus their attention on the signs and symptoms of a case that are consistent with the diagnosis they are asked to evaluate at the expense of reducing consideration of the information consistent with other hypotheses. The results of Experiments 4 to 6 support this hypothesis. In these experiments it was found that participants who were prompted to re-evaluate the evidence were less susceptible to the biasing influence of the explicitly presented alternatives as long as this re-evaluation took place at the moment of decision-making. This was true even though no additional information was provided through prompting. Finally, additional factors that might lead to biased processing of the available evidence were examined. Formal medical terminology will cause differential processing, possibly because it is more mnemonically effective (Experiment 7). Diagnosticians are less susceptible to under-weighting non-focal alternatives when they themselves generate the focal hypotheses (Experiment 8). And finally, the diversion of attention created by the explicit mention of a specific diagnosis is not driven purely by an analytic shift of attention towards that particular diagnosis. Non-analytic influences are also influential (Experiment 9). The importance of these results for medical education and psychological theory of subjective probability is discussed.
Eva, Kevin Wayne, "The influence of differentially processing evidence on diagnostic decision-making" (2001). Open Access Dissertations and Theses. Paper 2474.