Evaluation of clinical teaching by general internal medicine faculty in outpatient and inpatient settings.

PURPOSE: To determine the measurement characteristics of a form used to evaluate teaching in outpatient settings, and to compare ratings received by general internists in outpatient and inpatient settings.
METHOD: The physicians evaluated were 29 faculty who taught in both outpatient and inpatient settings affiliated with the Department of Medicine, University of Washington School of Medicine, over a five-year period (from 1985-86 through 1989-90). Residents completed 639 evaluations, using a six-point Likert-type scale (from 1, very poor, to 6, excellent) to rate instructors in eight categories and to provide an overall assessment of each instructor's teaching effectiveness. In addition, each evaluation contained an indication of the resident's perceived degree of involvement with the instructor. Statistical analysis involved two-tailed t-tests, analysis of variance and covariance, intraclass correlation coefficients, and the Spearman Brown prophecy formula.
RESULTS: When more than ten raters of teaching in outpatient settings were available, the reliabilities ranged from .58 to .81. The outpatient-setting ratings were significantly lower than the inpatient-setting ratings for seven categories, overall teaching effectiveness (5.01 versus 5.35, p < .05), and perceived degree of involvement of the instructor with the resident. When controlled for perceived degree of involvement, differences remained in only two categories: demonstration of clinical skills, rated higher for outpatient settings, and instructor accessibility, rated lower.
CONCLUSION: The evaluation form provides reliable ratings of teaching in outpatient settings when more than ten raters are available. The differences found in the ratings between the inpatient and outpatient settings may be explained by the factor of the resident's perceived degree of involvement with the instructor.