Using the Modified Delphi Method to Develop a Tool to Assess Pediatric Residents Supervising on Inpatient Rounds

Consensus about behaviors that define effective supervision by residents of more junior trainees on pediatric inpatient rounds is lacking.

Use modified Delphi method to develop a checklist of essential supervisory behaviors pediatric residents demonstrate while leading inpatient, non-ICU, nonspecialty teaching rounds and pilot the checklist.

One hundred and forty-one initial candidate behaviors were identified through literature review and narrowed by local stakeholders. Forty-one behaviors were submitted to national experts identified through purposive sampling to be refined using the modified Delphi method. Participants indicated agreement with behaviors' inclusion in the checklist and whether any were nonobservable. Measures of central tendency and dispersion were used to identify consensus with a behavior's inclusion in the next cycle. A criterion was eliminated if >25% of experts felt it was not observable. Cycles continued until consensus was reached on ≤20 behaviors. The resulting checklist was piloted at 2 hospitals.

After 2 modified Delphi cycles, consensus was obtained on 18 behaviors for inclusion in the final checklist. These were spread across 3 subcategories: teaching, leadership, and patient safety. In the pilot, the checklist of behaviors discriminated between residents with differing levels of competency in supervising trainees on inpatient rounds. For residents who had the checklist completed by 2 faculties, there was 75% agreement in behaviors observed.

Based on literature review, local stakeholder input, and consensus of national experts through the modified Delphi method, we created and piloted a checklist of observable behaviors characteristic of effective clinical supervision by pediatric residents leading trainee teams on inpatient, non-ICU, nonspecialty rounds.