ASSESSMENT OF EMPATHY IN A STANDARDIZED-PATIENT EXAMINATION

Background: Empathy has long been thought to be an important character istic of a good physician, and a measure of empathy is needed to provi de feedback to medical students, residents, and physicians on this imp ortant aspect of their clinical performance. The standardized-patient- resting format provides a simple but intuitvely convincing approach to the assessment of empathy. Purpose: To determine the extent to which 4th-year medical students were checked ''empathic'' by standardized pa tients (SPs) on a performance-based examination, to evaluate the psych ometric properties of this simple empathy measure, and to see whether empathy was related to clinical performance on history taking and phys ical examination. Method: Analyses were performed on examination data for 1,048 senior medical students in the 8 member schools of the New Y ork City Consortium tested at The Morchand Center of Mount Sinai Schoo l of Medicine. Results: The percentage of students who were checked '' empathic'' ranged from 59% to 98% (M = 79%) across the 7 SP cases. Of the 1,048 students, 268 (26%) were checked ''empathic'' on all 7 cases ; however, 221 (21%) were checked on less than or equal to 4 cases, 90 (9%) on less than or equal to 3 cases, and 26 (3%) on less than or eq ual to 2 cases. The generalizability coefficient of the overall empath y scores was .43; the dependability index with cutoff was .81 for dete cting students checked ''empathic'' on fewer than half of the 7 cases. The correlations of the empathy item with the other checklist items s uggest that the empathy construct refers to behaviors that make the pa tient feel comfortable and important. Also, students who were checked ''empathic'' on fewer than half of the 7 cases performed lower on hist ory taking and physical examination. Conclusions: Empathy appears to b e reasonably acceptable in this sample of students, although it is of concern that on average, more than 200 students per case were not seen as empathic, and more than 200 were checked ''empathic'' on less than or equal to 4 of the 7 cases. These results show the potential useful ness of this simple measure of empathy and illustrate the need for fee dback to address any problems.