Clinical Reasoning
Clinical Reasoning in Differential Diagnosis Experts use 3 main methods or a combination: Analytic or Hypothetico-deductive Non-analytic or Pattern recognition Pathognomonic signs and symptoms
Analytic Process Presenting ClinicalDiagnosticPosterior FeaturesHypothesesProbability A Dx1 Pr (Dx1) B Dx2 Pr (Dx2) C Dx3 Pr (Dx3) Elstein, 1978
Non-analytic Process PresentingFilter Clinical through prior Diagnostic Featuresepisodes Hypotheses Pr (Dx1) Pr (Dx2) Pr (Dx3) A C D B A,B,D,F B,D,G,R C,F,G,H
Combined Model of Clinical Reasoning Both analytic and non-analytic processes combined Eva et al.,2002 Patient Presents Case Representation Hypotheses Tested Non-analytic Analytic Interactive
Implications for Clinical Teachers Teach around examples Few, complex examples - suboptimal Provide many examples Represent range of presentations of specific conditions
Implications for Clinical Teachers Practice with cases should mimic eventual use of knowledge Working through textbook cases is NOT enough Mixed practice with multiple categories mixed together
Implications for Clinical Teachers Do NOT rely on students to make comparisons across problems spontaneously Allow students to identify similarities in underlying concepts of distinct problems Relate principles in new examples with those in past examples Provide learners with an opportunity to reveal idiosyncratic mistakes
Implications for Clinical Teachers Encourage learners to use both analytical rule knowledge and experiential knowledge
Cognitive sciences- based training Research study 2 different methods for training 2nd year medical students Traditional classroom based lecture Cognitive sciences-based approach (KBIT) Papa et al. 2007
Cognitive sciences- based training Similarities Common problem Identified differentials for problem Introduced each case via use of prototype and case example
Cognitive sciences- based training Differences KBIT group - 4 example cases per disease FS group - 1 case example per disease KBIT group - actively required to apply knowledge base towards diagnosis of practice cases (35) FS group cases, with no control over students’ active engagement in the cases
Cognitive sciences- based training Differences KBIT - immediate online formative and contrastive feedback tailored to each student FS - not possible to deliver tailored feedback
Cognitive sciences- based training Results KBIT group diagnosed correctly more test cases than FS group 74.2% vs 59.9% (P < 0.001; effect size = 1.42)
Cognitive Biases Representativeness heuristic - overestimating similarity between people and events Availability heuristic - too much weight to easily available info Overconfidence Confirmatory bias - bias toward positive and confirming evidence Illusory correlation - perceiving two events as causally related when there is none Putting initial probability at too extreme a figure and not adjusting for subsequent info Klein, 2005.
Summary Expertise is not a matter of acquiring a general, all-inclusive reasoning strategy No one kind of knowledge counts more than any other Expertise in medicine derives from both formal and experiential knowledge Norman, 2007