Do Emergency Medicine Residents Prefer Resident-initiated or Attending-initiated Feedback?


  • The authors have no relevant financial information or potential conflicts to disclose.



Real-time feedback is crucial to improving physician performance. Emerging theory suggests that learner-initiated feedback may be more effective in changing performance than attending-initiated feedback, but little is known about how residents perceive resident- versus attending-initiated feedback.


The primary aim was to determine whether residents' satisfaction varied by learner-versus attending-initiated feedback encounters. We hypothesized that residents would be more satisfied with resident-initiated feedback.


This was a multicenter study of five emergency medicine residency programs. We developed a milestones-based, real-time feedback intervention that provided behavioral anchors for ED subcompetencies and prompted a feedback discussion. The intervention was implemented at all sites for a 3-month period from March to November 2014. Residents were asked to initiate one card per shift; attendings were also invited to initiate encounters and, in either instance, asked to provide one specific suggestion for improvement. Residents confidentially rated their satisfaction with feedback on a 10-point scale. Reported satisfaction was categorized as “very satisfied” (score of 10) versus “less than very satisfied” (score < 10). Logistic regression was used to assess the difference in satisfaction between resident- versus attending-initiated feedback, and random effects were used to account for the clustering of repeated ratings within resident and by site.


A total of 785 cards was collected from five sites. Participation varied by site (range = 21–487 cards per site). Of the 587 cards with both feedback initiator and satisfaction data, 67% (396/587) were resident-initiated, and median satisfaction score was 10 (range = 4–10). There was no difference in the odds of being “very satisfied” by resident- vesus attending-initiated encounters (odds ratio = 1.08, 95% confidence interval = 0.41 to 2.83).


Our results suggest that residents are likely to be as satisfied with self-initiated feedback as attending-initiated feedback. Further research is needed to determine whether resident-initiated feedback is more likely to be incorporated into practice and result in objective performance improvements.