Modeling Technical Skills Learning Curves of Incoming Surgical Trainees

OBJECTIVE: Not all trainees reach technical competency even after completing surgical training. While assessment of technical skill is not part of the residency interview process, identifying under-performers early on may help identify opportunities for individualized, targeted training. The objectives of this study were to (1) create predictive learning curve (LC) models for each of 3 basic laparo-scopic tasks to identify performers versus underperform-ers and (2) evaluate the use of LCs to identify underperformers during selection into surgical training. DESIGN: Predictive LC models were created for laparo-scopic pattern cutting (PC), peg transfer (PT) and intra-corporeal knots (IC) over 40 repetitions by 65 novice trainees in 2014. Trainees were categorized as performers and underperformers. Receiver operator characteristic analysis determined the minimum number of repetitions required to predict individual LCs, which were then used to determine the proportion of underperformers. SETTING: Technical performance was assessed onsite at the Canadian Residence Matching Service (CaRMS) inter-views, after interview completion (January 2015). PARTICIPANTS: Applicants to general surgery (GS) and gynecology (OBGYN) participated in a skills assessment during.RESULTS: The PC, PT and IC tasks required a minimum of 8, 10, and 5 repetitions respectively, to predict overall performance. Predictive values for each task had excel-lent sensitivity and specificity: 1.00, 1.00 (PC); 1.00, 1.00 (PT); and 0.94, 1.00 (IC). Eighty applicants completed 8 PC repetitions; 16% were identified as underperformers. CONCLUSIONS: Individual LCs for three different lapa-roscopic tasks can be predicted with excellent sensitiv -ity and specificity based on 10 repetitions or less. This information can be used to identify trainees who may have difficulty with laparoscopic technical skills early on. ( J Surg Ed 80:51-61. (c) 2022 Association of Program Directors in Surgery. Published by Elsevier Inc. All rights reserved.)