MP22: Using galvanic skin response to identify resuscitation expertise in a pulmonary embolism simulation exercise

Abstract

Introduction: As physicians gain expertise in clinical settings, they are able to handle progressively more information, in both complexity and magnitude, as an organized schema. Expert physicians then, will be more likely to function with less cognitive load —the amount of mental effort someone exerts within their short-term working memory. Expert physicians will also retain more working memory capacity to process information during medical emergencies than novice physicians. While a physician's ability to process medical information may have implications for handling medical emergencies, there is a paucity of empirical research examining the link between physicians’ expertise and biometric measures of cognitive load. Using galvanic skin response (GSR) as a surrogate measure of cognitive load, we assess whether average cognitive load differs significantly between expert and novice physicians in a pulmonary embolism simulation exercise. Methods: We analyzed GSR data (n = 39) from a 10-minute simulated pulmonary embolism exercise among 18 faculty physicians and 21 residents. Cluster and factor analyses were used to identify novice, intermediate, advanced, and expert physicians with based on participants’ GRS scores. One-way ANOVA was used to analyze group differences. Descriptive statistical techniques were also used to describe the distribution of GRS expertise by participants’ level of training. Results: Contrary to expectation, we found more than two groups of resuscitation expertise in the simulation exercise. Respectively, we identified 7.7% and 20.5% of participants as novice and expert physicians. About 36% of participants were classified as intermediate (35.9%) physicians while another 36% were classified as having an advanced (35.9%) expertise in resuscitation. All the novice physicians identified were found to be PGY1 and PGY2 resident residents. A third (33.3%) of faculty physicians and 9.5% of residents were identified as experts. As expected, average GSR score for experts (x¯ = 0.60μS, SD = 0.26) was significantly (F = 137.6, p < 0.001) lower than the average GSR for novices (x¯ = 5.55μS, SD = 0.99), intermediate (x¯ = 2.84 μS, SD = 0.40), and advanced (x¯ = 1.57 μS, SD = 0.28) physicians. Conclusion: GSR measures of cognitive load may be used to identify resuscitation expertise in managing pulmonary embolism and related medical conditions through simulation exercises.