Narrative Approach to Goals of Care Discussions: Assessing the Use of the 3-Act Model in the Clinical Setting.

CONTEXT:The 3-Act Model is an innovative narrative approach to goals of care (GOC) discussions centered on patients' unique stories. Previous studies have demonstrated the effectiveness of the 3-Act Model training in enhancing trainees' skills objectively in role-plays with standardized patients. This study moves beyond the classroom to the clinical setting to assess whether learners preferred this approach, believed it to be clinically effective, and demonstrated proficiency objectively. OBJECTIVE:Our primary objective was to describe internal medicine interns' preferred approach to leading GOC discussions and their view of its efficacy, before vs. after training. Our secondary objective was to assess the proficiency of interns in leading GOC discussions with hospitalized patients. METHODS:We sent online surveys to a cohort of 22 interns both before (pretest) and mid-year after (posttest) training in the 3-Act Model. In addition, trainers objectively assessed GOC discussions led by a subset of trainees in the inpatient setting using a previously described rubric congruent with the 3-Act Model. RESULTS:In the posttest survey, many more interns reported using a narrative approach to GOC discussions most often (pretest, 1 of 22 [5%] vs. posttest, 16 of 18 [89%]). The percentage of interns reporting their preferred approach worked "very well" or "extremely well" increased from pretest to posttest (32%-89%, P = 0.002). Trainers assessed 13 completed GOC discussions led by 4 trainees (18% of original cohort) in the hospital setting and found all 13 to be proficient. CONCLUSION:The vast majority of interns reported preferential use and effectiveness of the 3-Act Model several months after training, and trainers found a subset to be proficient in GOC discussions with hospitalized patients. This study contributes evidence that training in the 3-Act Model positively impacted intern behavior in the clinical setting.