Improving assessment practice through cross-institutional collaboration: an exercise on the use of OSCEs

BACKGROUND: This study was undertaken to improve assessment practice on OSCEs through collaboration across geographically dispersed medical schools in Australia. METHODS: A total of eleven OSCE stations were co-developed by four medical schools and used in summative 2011 and 2012 examinations for the assessment of clinical performance in the early clinical and exit OSCEs in each school's medical course. Partial Credit Rasch Model was used to evaluate the psychometric properties of the shared OSCE data. Evaluation of the quality assurance reports was used to determine the beneficial impact of the collaborative benchmarking exercise on learning and teaching outcomes. RESULTS: The data for each examination demonstrated sufficient fit to the Rasch model with infit mean square values ranging from 0.88 to 0.99. Person separation (1.25-1.63) indices indicated good reliability. Evaluation of perceived benefits showed that the benchmarking process was successful as it highlighted common curriculum areas requiring specific focus and provided comparable data on the quality of teaching at the participating medical schools. CONCLUSION: This research demonstrates the validity of the psychometric data and benefits of evaluating clinical competence across medical schools without the enforcement of a prescriptive national curriculum or assessment.