RPD205 - Identification of Behavioral Outcomes for Implicit Bias Mitigation Interventions

Key Message : Identifying behavioral outcomes for implicit bias mitigation interventions is imperative to measure their effectiveness. Verbal dominance and physical distance from the surrogate or patient during end-of-life communication may be appropriate outcomes to target.

Abstract:

Background: Given the substantial negative effects of racial implicit bias on healthcare communication (1), bias mitigation interventions have proliferated. However, rigorous evaluation of such interventions is lacking.

Objectives/

Purpose: We aimed to understand which communication behaviors correlate with implicit bias and perception of discrimination in end-of-life communication to identify behavioral outcomes for bias mitigation interventions.

Methods: Physicians caring for seriously ill acute care patients completed a simulated encounter with a Black standardized actor portraying a patient’s surrogate decision-maker (daughter). Verbal, nonverbal and paraverbal communication behaviors were coded from the videotaped simulations. Participants completed the Implicit Association Test (IAT) for racial bias and actors completed a Discrimination in Medical Settings scale.

Results: Of 43 respondents, 22 (51%) were female, 22 (51%) White, 17 (40%) Asian, and 4 (9%) Hispanic. Specialties included oncology (8, 19%), critical care (15, 35%), internal medicine (16, 37%), and Geriatrics (4, 9%). Fifteen (35%) were residents or fellows. Their mean IAT score was 0.32 indicating a moderate level of bias. Nonverbal and paraverbal behaviors were not correlated with IAT. Describing life-sustaining treatment positively was positively correlated with IAT (r=0.35, p=0.02), and orienting the surrogate to the conversation was negatively correlated with IAT (r=-0.38, p=0.02). Perceived discrimination was negatively correlated with nodding (r=-0.33, p=0.04) and mirroring the surrogate’s language (r=0.31, p=0.05). Perceived discrimination was positively correlated with time spent further away from the surrogate (r=0.37, p=0.02), describing life-sustaining treatment negatively (r=0.32, p=0.04) and verbal dominance (r=0.63, p< 0.01).

Conclusions: Few behaviors correlated with implicit bias. Targeting outcomes of bias mitigation interventions to behaviors that reduce perceived discrimination may be more valuable. How life-sustaining treatments are described has a complex relationship with bias and perceived discrimination. Reasonable outcomes to target in bias mitigation interventions include verbal dominance and physical distance from the surrogate, as well as mirroring, nodding and orienting.

References: 1. Johnson RL, Roter D, Powe NR, Cooper LA. Patient race/ethnicity and quality of patient-physician communication during medical visits. Am J Public Health 2004;94:2084-90.