Construct validity and reliability of structured assessment of endoVascular expertise in a simulated setting

Objectives: To study the construct validity and reliability of a novel endovascular global rating scale, Structured Assessment of endoVascular Expertise (SAVE).

Design: A Clinical, experimental study.

Materials: Twenty physicians with endovascular experiences ranging from complete novices to highly experienced operators performed a video-recorded simulated contra-lateral iliac-artery-stenting procedure. The virtual-patient case was a novel technically challenging procedure presenting the distal arteries below the knee.

Methods: Three experts assessed the performances blinded to operator identity. Validity was analysed by correlating experience with performance results. Reliability was analysed according to generalisability theory.

Results: The mean score on the 29 items of the SAVE scale correlated well with clinical experience (R = 0.84, P < 0.01) and was found discriminative even among the more experienced participants having performed up to 500 endovascular procedures in total. Only the most experienced participants (>5000 procedures) obtained maximum scores. The inter-rater reliability was high (G = 0.94 and G = 0.95). The procedure time (median 69 min, range 32-86) correlated moderately with clinical experience (R = -0.53, P < 0.05), whereas the fluoroscopy time and amount of contrast fluid did not correlate.

Conclusions: The construct validity and reliability of assessment with the SAVE scale was high when applied to performances in a simulation setting with advanced realism. No ceiling effect was present in the assessment situation.