It's not all White: Implicit Racial Bias in Imagery Used in Plastic Surgery Resident Education

Abstract

Objective: Plastic surgery education relies heavily on images featuring patient skin tone; while images can be useful representations, it is highly susceptible to implicit bias, a known contributor to healthcare disparities. Using skin tone as a proxy, this study evaluates patient representation in images used in the American Society of Plastic Surgery Resident Education Curriculum.

Design: Color photographs, graphics, and videos featured in the American Society of Plastic Surgery "Course Materials" for each module were categorized using the Fitzpatrick scale (I-II, III-IV, or V-VI) by six reviewers. Proportional data and average number ± standard deviation of photos and graphics for each category were reported. Significant difference between Fitzpatrick I to II and V to VI was investigated via a one-way analysis of variance with a Tukey's post-test to adjust for multiple comparisons.

Results: An average of 1861 photographs and 237 graphics were assessed with 82% (1518 ± 25.11) of photos and 97% (231 ± 24.45) of graphics categorized as Fitzpatrick I to II. A one-way analysis of variance with a Tukey's post-test demonstrates a statistical difference between images and graphics categorized as Fitzpatrick I to II and Fitzpatrick V to VI (p < 0.001).

Conclusions: Our data reveals an opportunity to improve racial representation in resident education. When 76% of patients in the United States are white and 13% are Black, our findings demonstrate both an unequal and unrepresentative distribution of photos and graphics of non-white patients. Residency is a formative time in a surgeon's career and therefore, exposure to accurate representation of a diverse patient population is of the utmost importance.

Keywords: imaging; implicit bias; plastic surgery education; racial disparities; skin tone.