Abstract
BACKGROUND: A competency-based education approach calls for frequent workplace-based assessments (WBA) of Entrustable Professional Activities (EPAs). While mobile applications increase the efficiency, it is not known how many assessments are required for reliable ratings and whether the concept can be implemented in all sizes of residency programs. METHODS: Over 5 months, a mobile app was used to assess 10 different EPAs in daily clinical routine in Swiss anesthesia departments. The data from large residency programs was compared to those from smaller ones. We applied generalizability theory and decision studies to estimate the minimum number of assessments needed for reliable assessments. RESULTS: From 28 residency programs, we included 3936 assessments by 306 supervisors for 295 residents. The median number of assessments per trainee was 8, with a median of 4 different EPAs assessed by 3 different supervisors. We found no statistically significant differences between large and small programs in the number of assessments per trainee, per supervisor, per EPA, the agreement between supervisors and trainees, and the number of feedback processes stimulated. The average "level of supervision" (LoS, scale from 1 to 5) recorded in larger programs was 3.2 (SD 0.5) compared to 2.7 (SD 0.4) (p<0.05). To achieve a g-coefficient >0.7, at least a random set of 3 different EPAs needed to be assessed, with each EPA rated at least 4 times by 4 different supervisors, resulting in a total of 12 assessments. CONCLUSION: Frequent WBAs of EPAs were feasible in large and small residency programs. We found no significant differences in the number of assessments performed. The minimum number of assessments required for a g-coefficient >0.7 was attainable in large and small residency programs.