Using Brunswik’s Probabilistic Functionalism to Test How Clinicians Make Judgments in Simulated Neonatal Resuscitation Scenarios

Abstract

Accurate clinical assessments are essential for providing appropriate patient care. However, clinicians do not necessarily make accurate assessments, either because they are overloaded with other tasks, or because the computation of assessments is ambiguous. We used Brunswik’s probabilistic functionalism to study how clinicians assessed the clinical state of a mannequin when they viewed recordings of simulated neonatal resuscitation scenarios and when they performed simulated resuscitations. Seventeen clinicians individually assigned an Apgar (neonate illness) score to 30 pre-recorded scenarios and also to 9 scenarios in which they played a hands-on role. We computed a judgment policy for each clinician showing the relative importance of five clinical signs that constitute the Apgar score. The accuracy of clinicians’ judgment policies was significantly correlated with the accuracy of their Apgar assessments for the pre-recorded scenarios (p<0.01) but not for the hands-on scenarios. The weighting for the clinical signs in the judgment policies was different from the unit weighting in the Apgar score itself. Brunswik’s approach provided a useful framework for testing clinicians’ assessments in simulated neonatal resuscitations. Future studies should determine the factors that affect accuracy in hands-on scenarios and test the applicability of the methods presented for other healthcare practice areas.