Error Disclosure: What residents say and what patients find effective

Emily Grossniklaus^1*Angelo D'Addario²Ann King²Thomas Gallagher¹Kathleen Mazor³Andrew A White¹

• ¹University of Washington, Seattle, United States
• ²National Board of Medical Examiners, Philadelphia, Pennsylvania, United States
• ³University of Massachusetts Medical School, Worcester, Massachusetts, United States

Background: Effective medical error disclosure to patients is a critical skill that is often not taught effectively in medical training. The Video- based Communication Assessment (VCA) software enables trainees to receive feedback on their error disclosure communication skills. The VCA method also allows examination of the specific types of error disclosure responses that patients value most. Objective: The primary aim of this study was to describe the language medical residents use to disclose a hypothetical harmful medical error, and to determine the language associated with higher ratings by crowdsourced laypeople. A secondary aim was to examine the alignment between error disclosure content recommended by experts and communication behaviors that contribute to higher layperson ratings of disclosure. Methods: 102 resident physician responses to a case depicting a delayed diagnosis of breast cancer and their crowdsourced ratings were analyzed using thematic content analysis. We assessed the presence of specific themes in response to three sequential video prompts within a clinical case. Linear regressions were performed for each prompt’s response to examine the extent to which each theme predicted overall communication scores from layperson raters. Results: Nearly all (N = 92, 90.2%) residents provided responses which included either a general apology or a specific apology in at least one of the three prompt’s responses, and nearly all (N = 98, 96.1%) residents provided at least one response expressing a component of empathy. A few residents used rationalization (5.9%) or minimization (4.9%) behaviors; these were associated with negative beta-coefficients, although did not reach statistical significance. In a linear regression analysis, the strongest positive associations between resident responses and patient ratings were expressions of accountability (0.48), personal regret (0.47), apology (0.34), and intentions to prevent future mistakes (0.34). Conclusion: Resident physicians vary in which communication elements and themes they include during error disclosure, missing opportunities to meet patient expectations. While infrequent, some residents employed minimization or rationalization in their responses. Utilizing an assessment and feedback system that encourages responders to include themes layperson raters value most and to omit harmful expressions could be an important feature for future software for error disclosure communication training.