Evaluating Surgical Residents' Patient-Centered Communication Skills: Practical Alternatives to the "Apprenticeship Model".

OBJECTIVES

The Accreditation Council for Graduate Medical Education (ACGME) requires residency programs to assess communication skills and provide feedback to residents. We aimed to develop a feasible data collection process that generates objective clinical performance information to guide training activities, inform ACGME milestone evaluations, and validate assessment instruments.

DESIGN

Residents care for patients in the surgical clinic and in the hospital, and participate in a communication curriculum providing practice with standardized patients (SPs). We measured perception of resident communication using the 14-item Communication Assessment Tool (CAT), collecting data from patients at the surgery clinic and surgical wards in the hospital, and from SP encounters during simulated training scenarios. We developed a handout of CAT example behaviors to guide patients completing the communication assessment.

SETTING

Independent academic medical center.

PARTICIPANTS

General surgery residents.

RESULTS

The primary outcome is the percentage of total items patients rated "excellent;" we collected data on 24 of 25 residents. Outpatient evaluations resulted in significantly higher scores (mean 84.5% vs. 68.6%, p < 0.001), and female patients provided nearly statistically significantly higher ratings (mean 85.2% vs. 76.7%, p = 0.084). In multivariate analysis, after controlling for patient gender, visit reason, and race, (1) residents' CAT scores from SPs in simulation were independently associated with communication assessments in their concurrent patient population (p = 0.017), and (2) receiving CAT example instructions was associated with a lower percentage of excellent ratings by 9.3% (p = 0.047).

CONCLUSIONS

Our data collection process provides a model for obtaining meaningful information about resident communication proficiency. CAT evaluations of surgical residents by the inpatient population had not previously been described in the literature; our results provide important insight into relationships between the evaluations provided by inpatients, clinic patients, and SPs in simulation. Our example behaviors guide shows promise for addressing a common concern, minimizing ceiling effects when measuring physician-patient communication.