Identification of high-risk residents allows remediation and support for administrative action when necessary. This study characterizes differences in documentation of marginally performing residents in a general surgery residency.High-risk residents were identified by the former program director. Twenty-four of one hundred fifteen residents over a 10-year period had one to four problematic areas: cognitive, synthetic, family/health, and interpersonal skills. Outcomes included finished (18), voluntary withdrawal (1), and involuntary withdrawal (5). A case-control study matching controls to cases by date of entry into the training program was used. Records were reviewed for demographics, preentry qualifications, American Board of Surgery In-Training Exam (ABSITE) scores, letters of complaint or praise, events of counseling, and monthly ratings. The records of 48 residents were reviewed. Ward evaluations were on eight categories with a 5-point Leikert scale (3-unacceptable to 7-outstanding). The evaluation score assigns points only to low ratings. High scores represent progressively poorer performance. A Wilcoxon signed ranks test was used to compare the cases and controls for continuous variables. The McNemar test was used in comparisons of categorical data with binary outcomes. Exact P values are reported.Objective data were similar for both groups. Study residents tended to score higher on monthly evaluations at Year 2 and by Year 3 this achieved significance (0.026). Study residents were more likely to have negative faculty letters (0.016) and events of counseling by a faculty member (0.017) and the program director (0.005).Identification of residents at risk should begin as early as possible during training. A combination of faculty evaluations and evidence of letters of counseling can detect high-risk residents. Programs may use such indicators to support decisions regarding remedial work or administrative action.