Measuring family physician identity: the development of a new instrument

Our objective was to describe the development and psychometric assessment of an instrument designed to assess family medicine identity in residency training sites and compare responses from physician faculty and residents.

We conducted 28 focus groups between 2007--2008, 14 with faculty and 14 with residents who were part of the Preparing Personal Physicians for Practice (P4) Project. The first 22 focus groups were exploratory, and the second six were confirmatory where we shared working variable statements scored using a 5-point Likert scale. We then administered the survey to 223 faculty and 147 residents who were part of the P4 Project, followed by a principal component (factor) analysis, retaining items that reflected domains with eigenvalues higher than 1.0.

A total of 223 family physician faculty and 147 residents completed the identity survey. The item analysis extraction loadings ranged from 0.36 to 0.70. Based on item grouping patterns, five domains were reflected in the data: Patient/Family Relationships, Patient Advocacy, Career Flexibility, Balancing the Breadth and Depth in Practice, and Comprehensive Nature of Patient Care. Compared to residents, faculty conveyed stronger agreement about being comfortable balancing the breadth and depth of medical knowledge needed in practice and using a variety of approaches to supplement their medical knowledge about patient care compared to residents (90.6% versus 68.7% for breadth and depth, 95.9% versus 88.3 for using a variety of approaches). Compared to faculty, residents agreed more strongly that the ability to choose many options in how to build their practice appeals to them compared to faculty (89.1% versus 82.9%).

We successfully developed and tested a survey designed to measure family medicine identity in residencies, with five domains. Survey item responses were different between residents and faculty, which indicates the instrument may be sensitive to important changes over time.