Catering to the patient - development, validation and psychometric properties of an innovative assessment instrument

Introduction: It has been shown that communication skills acquired during undergraduate medical education are of great importance. Hence, many countries require teaching communication as part of their medical curricula. To assess students' learning progress, "Catering to the Patient", as an aspect of showing empathy, should be evaluated. Since there was no description of a validated instrument fitting for this purpose, one had to be developed. To describe its process of development and its psychometric properties were the aims of this study. Methods: Based on the Calgary-Cambridge Observation Guide (CCOG), items describing catering to the patient were selected and modified. Cognitive pretest interviews were conducted to check understandability. Therefore, 7 raters assessed 1 video each (R=7, V=1). In a following pilot study (R=3, V=10) first psychometric properties were evaluated and necessary corrections in the preliminary evaluation form were carried out before the final evaluation form was used to assess students' ability to cater to the patient and psychometric properties were described in detail (R=2, V=35). Results: The final assessment instrument, "catering to the patient - Marburg evaluation form", contains 11 checklist items and two global ratings (items 12 and 13). In the final evaluation the inter-rater reliability (IRR) ranged from 0 to 0.562, the median was r=0.305. Concerning item 13 (a global rating), 88.6% of the videos were scored with the maximum difference of one point. The internal consistency was very high (Cronbach's α: α=0.937 and α=0.962), and the correlation between the checklist items and the global rating was high (Pearson's r: r=0.856 and r=0.898). Discussion: The assessment instrument "catering to the patient" is suitable for giving feedback and for using it in formative examinations. Its use for summative examinations can be considered. Further examinations should evaluate if a three-point Likert scale could reach higher values and if item 13 can be used as a stand-alone item.