Surgical skill is predicted by the ability to detect errors

Objective analysis methods of surgical performance are now available so comparison between surgeons is available. One such method is by direct observation using the Objective Structured Assessment of Technical Skills (OSATS), but this is a time-consuming process; therefore, a simple screening tool for the ability to detect errors (previously validated) was analyzed and considered as a predictor of qualitative performance.

Thirty-eight volunteer surgeons were recruited to the skills laboratory to undertake 3 exercises. Two were bench-top surgical tasks that were scored using the global rating of the OSATS technique. The third task was the ability to detect simple errors in 22 synthetic models of common surgical procedures, some of which contained purposefully made errors. P<.05 was deemed to be statistically significant.

The scores (interquartile ranges in parentheses) for the 3 sections were excision of sebaceous cyst=21 (19,24), closure of small bowel enterotomy=23 (21,27), and identification of errors=31 (27,34). Three scorers blinded to the operative models exhibited an interobserver reliability of .9 and .91 for the video tasks, respectively. Spearman's rank correlations between the error examination and performance on the 2 tasks were both statistically significant at .69 (cystectomy) and .54 (enterotomy).

The ability to detect simple surgical errors is a predictor of technical skill and performance of bench tasks. What must be answered is whether the use of such models and principles can shorten the qualitative surgical learning curve.