Abstract P370: Implicit Bias Among Vascular Surgeons

Abstract

Objective: Implicit bias is an unconscious bias that may influence behavior, cognitive processes and decision-making. In clinical settings, implicit bias present in clinicians may influence the treatment that people receive and may be a root cause of health disparities. Our study measured implicit bias among vascular surgeons who participate in the management of individuals with peripheral artery disease (PAD). Methods: Our sampling frame included all vascular surgeons that contribute to the Vascular Quality Initiative (VQI) registry. We invited these individuals to take the race implicit association test (IAT) used to measure level of unconscious bias. The IAT asks participants to classify items presented on a computer by quickly pressing computer keys. The underpinning of the IAT is that people will match a representative group (images of African American or European American faces) to an attribute (good and bad words, such as excellent and angry) more quickly if there is already an implicit connection. Based on the differences in reaction time from several trials, participants are classified as having a no preference between African American and European Americans or a preference for one or the other. The level of preference can be slight, moderate, or strong. We stratified implicit findings by physician race/ethnicity. Results: Among 2765 vascular surgeons that participate in the VQI, 338 (12%) completed the IAT. Most participants (70%) showed an automatic preference for European Americans over African Americans, with more than half of participants (51%) showing a moderate or strong automatic preference. Of 195 participants who identified as Non-Hispanic White, 25% showed a slight automatic preference for European Americans, 41% showed a moderate automatic preference, and 34% showed a strong automatic preference. Of 16 participants who identified as Non-Hispanic Black, 38% showed an automatic preference for European Americans over African Americans and 31% showed an automatic preference for African Americans. We also found that participants were often unaware of their biases: More than two-thirds (n=64) of the 94 participants who disagreed with a statement asking if they are biased (“I am biased”) had an automatic preference upon testing. Conclusions: Implicit bias was found among vascular surgeons. The next step in our research will focus on the ability of implicit bias scores to predict a surgeon’s actual delivery of guideline care. Pending our findings, training and clinical support tools offer techniques to mitigate the impact of implicit biases on PAD-related health care.