A Novel Survey Tool to Assess Pulmonary and Critical Care Fellows' Attitudes Regarding Acquiring Teaching Skills During Fellowship Training

Teaching is a core component of academic medical practice and is embodied in the etymology of the word doctor from the Latin doctorem, or “teacher.”1 The Accreditation Council for Graduate Medical Education's (ACGME's) practice-based learning and improvement competency requires trainees to “participate in the education of patients, families, students, residents, and other health care professionals.”2 

The literature indicates that residents consider teaching an important component of their training,3,4 and that attending physicians and residency programs encourage resident teaching to improve residents' knowledge, communication skills, time management, and clinical reasoning.5,6 

Fellows' attitudes toward teaching skills have rarely been examined,7–11 and there is only 1 prior study focused on pulmonary and critical care medicine (PCCM) fellows' perspectives.12 The importance that PCCM fellows place on acquiring and improving teaching skills during fellowship has not been previously measured. We developed and psychometrically characterized a survey tool to assess fellows' attitudes regarding the acquisition of teaching skills. Using this survey tool, we performed a prospective, observational cohort study to evaluate the survey tool's ability to assess the attitudes regarding and perceptions of the importance of acquiring teaching skills. We studied first-year through fourth-year fellows at the Harvard Combined PCCM Fellowship Program during 1 academic year.

As our literature review did not identify existing survey tools for assessing fellows' perspectives and attitudes regarding teaching skills, we developed questions to measure these domains. We designed questions to assess the domains of “interest in medical education” and “perception of support [in the acquisition of teaching skills].” Candidate questions were designed to minimize bias, redundancy, and confusion,13 and were edited by the authors in an iterative fashion to optimize content, length, and relevance to survey goals.14 We then asked 3 faculty members at our institution with medical education experience and familiarity with survey design to review the survey and edited the survey on the basis of this feedback. The final draft was piloted with 4 PCCM fellows to assess clarity, survey length, and identify redundant questions.15 The final version of the survey was administered to PCCM fellows at the 4-year Harvard Combined PCCM Fellowship Program (the survey instrument is provided as online supplemental material).

The anonymous survey was first administered online in August 2010, and the second administration occurred in June 2011. A self-generated identification was used to match both surveys for each participant.

The study protocol was reviewed by the Beth Israel Deaconess Medical Center Institutional Review Board and determined to be exempt.

Statistical analyses were performed with JMP Pro version 9.0 (SAS Institute Inc, Cary, NC). Cronbach α was calculated to assess for internal reliability of the survey items. We performed principal component analysis on correlations of responses from the beginning and end of the year (collapsed and decoupled from subject identification) to assess for variance per component and weighted correlations between questions. We also performed 2-tailed t tests to compare fellows' answers from the beginning of the year to those from the end of the year.

Twenty-seven of 34 fellows (78%) completed the survey at the beginning of the academic year, compared to 31 of 34 (91%) at the end of the year (table 1).

Principal component analysis demonstrated clustering of questions into components and graphically displayed between-question correlations (principal component analysis data are provided as online supplemental material). Cronbach α for the entire set was 0.648.

Responses on the 10-point Likert scales were collapsed into subgroups for ease of interpretation and generalizability of the results (table 2). Changes from the beginning to the end of the year were determined by comparing results from all fellows for a given question. Most felt that acquiring teaching skills was “important” or “very important” at the beginning of year (89%) compared to the end of year (84%, P  =  .15). Approximately half the fellows indicated they were “interested” or “very interested” in pursuing careers as medical educators (beginning of year  =  55%, end of year  =  46%; P  =  .30).

Despite positive attitudes toward teaching skills and interest in pursuing careers in medical education, most fellows were uncertain as to what skills they needed to become better teachers, with a minority indicating it was “very clear” what skills they needed (beginning of year  =  26%, end of year  =  16%; P  =  .92).

Our questions relating to fellows' attitudes about the feedback they received from faculty regarding their teaching skills revealed that the frequency with which fellows disagreed with faculty feedback showed a statistically significant increase during the academic year (beginning of year  =  11% disagreed with feedback, end of year  =  36% disagreed; P  =  .02). There were no across-group (by year) or within-group differences for a specific fellowship class when comparing beginning of year versus end of year.

We describe the development and preliminary psychometric properties of a survey tool to assess fellows' attitudes about teaching and present results from a single-center cohort study of PCCM fellows.

Despite a somewhat low Crohnbach α, our fellow survey tool meets criteria for content validity, as participants' answers indicate consistency with both pretesting expert validation and pilot testing.

The strongest indication of construct validity we offer is the clustering of similar questions on principal component analysis (provided as online supplemental material). Limited added evidence of construct validity is found in the consistency of the results of our study and prior studies of residents3,4 and fellows.7 Our survey found that most fellows were interested in improving their teaching skills and a sizeable percentage were interested in pursuing careers in medical education. During a year of fellowship training, fellows maintained positive attitudes toward teaching skills, yet most fellows reported they did not know what specific skills they needed, and this did not change after a year of fellowship training.

Fellows' increasing disagreement with faculty feedback as the year progressed may reflect fellows' increased clinical knowledge and increased experience as teachers from the beginning to the end of the year. However, the specific reasons for this attitudinal shift were not determined in this study. Our findings are consistent with prior studies of residents and their perceptions of obstacles to acquiring teaching skills.3,4 

Our study has limitations. This initial survey was conducted at a single fellowship program, which focuses on training academic physicians, and likely attracts and selects applicants with attitudes toward teaching skills that may differ from other programs.

In addition, the range of the Likert scales used in the survey tool may have been too wide, and future iterations of the tool will use scales ranging from 1 to 5 rather than 1 to 10 (provided as online supplementary material). This will obviate the need to cluster responses in future studies, increasing the utility and ease of interpretation of the survey results.

Our survey tool to assess PCCM fellows' attitudes toward learning teaching skills and interest in careers in medical education demonstrated moderate performance. Most fellows supported the importance of learning teaching skills, but most were unclear as to how to acquire these skills. Half of the fellows were interested in medical education careers and this interest persisted during fellowship. Our finding that most fellows did not know what would improve their teaching skills suggests that increased attention to explicitly developing fellows' teaching skills may fill an unfilled educational need.

Further studies of more heterogeneous populations of PCCM fellows and other trainees are necessary to draw definitive conclusions regarding fellows' attitudes toward acquiring teaching skills. Future research also is needed to assess construct validity in different populations of learners, generalizable to PCCM fellows in other settings and potentially other trainees.