Background: Accurate self-assessment of knowledge and technical skills is key to self-directed education required in surgical training. We aimed to investigate the presence and magnitude of cognitive bias in self-assessment among a cohort of surgical interns.
Methods: First-year general surgery residents self-assessed performance on a battery of technical skill tasks (knot tying, suturing, vascular anastomosis, Fundamentals of Laparoscopic Skills peg transfer and intracorporeal suturing) at the beginning of residency. Each self-assessment was compared to actual performance. Bias and deviation were defined as arithmetic and absolute difference between actual and estimated scores. Spearman correlation assessed covariation between actual and estimated scores. Improvement in participant performance was analyzed after an end-of-year assessment.
Results: Participants (N = 34) completed assessments from 2017 to 2019. Actual and self-assessment scores were positively correlated (0.55, P < .001). Residents generally underestimated performance (bias -4.7 + 8.1). Participants who performed above cohort average tended to assess themselves more negatively (bias -7.3 vs -2.3) and had a larger discrepancy between self and actual scores than below average performers (deviation index 9.7 + 8.2 vs 3.8 + 3.1, P < .05). End-of-year total scores improved in 31 (91.2%) participants by an average of 11 points (90 possible). Least accurate residents in initial self-assessments (deviation indices >75th percentile) improved less than more accurate residents (median 5 vs 16 points, P < .05). All residents with a deviation index >75 percentile underestimated their performance.
Conclusion: Cognitive bias in technical surgical skills is apparent in first-year surgical residents, particularly in those who are higher performers. Inaccuracy in self-assessment may influence improvement and should be addressed in surgical training.
Copyright © 2021 Elsevier Inc. All rights reserved.