RESUMO
In addition to current admissions criteria, the assessment of qualities related to humanistic medicine may help predict clinical and academic performance in dental education. The objective of this feasibility study was to develop and assess the reliability of a coding scheme to quantify and rate qualitative formative narratives describing individual student performance in a medical interview course that utilizes principles of humanistic medicine. A mixed method, using focus groups and individual interviews, was used to develop a coding strategy. Three coders retrospectively rated 209 formative narratives evaluating first-year students at the Harvard School of Dental Medicine to determine high and low performance in this course. Interrater reliability was tested. The focus groups yielded the use of superlatives to identify high performance, with their absence indicating low performance. This study found a high level of calibration among independent coders (Cronbach's alpha 0.75). Twenty-four (11.5 percent) of the narrative evaluations were coded as "high performers" and seven (3.3 percent) as "low performers." The results demonstrate the feasibility of quantifying narrative evaluations to determine high and low performance in a patient-centered course for dental students. It may be that humanistic qualities taught in a patient-centered medical interview course are more significant competencies than previously thought.