Assessment of performance measures and learning curves for use of a virtual-reality ultrasound simulator in transvaginal ultrasound examination

作者：Madsen M E^*; Konge L; Norgaard L N; Tabor A; Ringsted C; Klemmensen A K; Ottesen B; Tolsgaard M G

来源：Ultrasound in Obstetrics and Gynecology, 2014, 44(6): 693-699.

DOI：10.1002/uog.13400

• criterion-based training
• learning curves
• simulation-based training
• standard setting
• ultrasound simulation

Objective To assess the validity and reliability of performance measures, develop credible performance standards and explore learning curves for a virtual-reality simulator designed for transvaginal gynecological ultrasound examination.
Methods A group of 16 ultrasound novices, along with a group of 12 obstetrics/gynecology (Ob/Gyn) consultants, were included in this experimental study. The first two performances of the two groups on seven selected modules on a high-fidelity ultrasound simulator were used to identify valid and reliable metrics. Performance standards were determined and novices were instructed to continue practicing until they attained the performance level of an expert subgroup (n=4).
Results All 28 participants completed the selected modules twice and all novices reached the expert performance level. Of 153 metrics, 48 were able to be used to discriminate between the two groups' performance. The ultrasound novices scored a median of 43.8% (range, 17.9-68.9%) and the Ob/Gyn consultants scored a median of 82.8% (range, 60.4-91.7%) of the maximum sum score (P < 0.001). The ultrasound novices reached the expert level (88.4%) within a median of five iterations (range, 5-6), corresponding to an average of 219 min (range, 150-251 min) of training. The test/retest reliability was high, with an intraclass correlation coefficient of 0.93.
Conclusions Competence in the performance of gynecological ultrasound examination can be assessed in a valid and reliable way using virtual-reality simulation. The novices' performance improved with practice and their learning curves plateaued at the level of expert performance, following between 3 and 4 h of simulator training.

• 出版日期2014-12