Proficiency-based residency training programs can be more efficient than the current duration-based formats. For their successful implementation, appropriate proficiency criteria must be developed. The objective of this study was to investigate the relationship between technical skill performances assessed using computer- and expert-based methods and training year. An assumption was that asymptotes in performance as a function of training year can be used to set the proficiency level for a technical skill, so the value at which the asymptote occurs can be labeled as the proficiency criteria.

Thirty-eight general surgery residents performed one-handed knot tying on bench-top simulators at two levels of difficulty: superficial and deep. Motion-efficiency measures and expert-based measures were used to evaluate performance. Total number of operations (ie, surgical volume) that each trainee participated in during residency was also acquired.

On the superficial model, asymptotes were observed at year 1 for motion-efficiency and year 3 for expert-based measures. On the deep model, asymptotes were observed at year 2 for motion-efficiency and year 4 for expert-based measures.

The data demonstrate the challenges associated with defining technical skills proficiency criteria. Different asymptotes were observed for the two assessment methods and neither covaried substantially with surgical volume. These data suggest that this asymptote approach in defining proficiency criteria can be suitable for development of proficiency-based residency training programs. The sensitivity of this approach to the type of assessment method and to the functional difficulty of the simulators used for assessment must be considered.