Examining the accuracy of residents’ self-assessments and faculty assessment behaviours in anesthesiology

Abstract

Background: Residents’ accurate self-assessment and clinical judgment are essential for optimizing their clinical skills development. Evidence from the medical literature suggests that residents generally do poorly at self-assessing their performance, often due to factors relating to learners’ personal backgrounds, cultures, the specific contexts of the learning environment and rater bias or inaccuracies. We evaluated the accuracy of anesthesiology residents’ self-assessed Global Entrustment scores and determined whether differences between faculty and resident scores varied by resident seniority, faculty leniency, and/or year of assessment.

Methods: We employed variance components modeling techniques and analyzed 329 pairs of faculty and self-assessed entrustment scores among 43 faculty assessors and 15 residents. Using faculty scores as the gold standard, we compared faculty scores with residents’ scores (x_i(faculty)–x_i(resident)), and determined residents’ accuracy, including over- and under-confidence.

Results: The results indicate that residents were respectively over- and under-confident in 10.9% and 54.4% of the assessments but more consistent in their individual self-assessments (rho = 0.70) than faculty assessors. Faculty scores were significantly higher (α = 0.396; z = 4.39; p < 0.001) than residents’ self-assessed scores. Being a lenient/dovish (β = 0.121, z = 3.16, p < 0.01) and a neutral (β = 0.137, z = 3.57, p < 0.001) faculty assessor predicted a higher likelihood of resident under-confidence. Senior residents were significantly less likely to be under-confident compared to junior residents (β = -0.182, z =-2.45, p < 0.05). The accuracy of self-assessments did not significantly vary during the two years of the study period.

Conclusions: The majority of residents’ self-assessments were inaccurate. Our findings may help identify the sources of such inaccuracies.