BACKGROUND: Accreditation Council on Graduate Medical Education work-hour restrictions are aimed at improving patient safety and resident well-being. Although surgical trainees will be dramatically affected by these changes, no comprehensive assessment of their well-being has been recently attempted. STUDY DESIGN: A multicenter study of psychological well-being of surgical residents (n = 108) across four US training programs before implementation of the 80-hour work week was performed using two validated surveys (Symptom Checklist-90-R [SCL-90-R] and Perceived Stress Scale [PSS]) during academic year 2002-03. Societal normative populations served as controls. Primary outcomes measures were psychologic distress (SCL-90-R) and perceived stress (PSS). Secondary outcomes measures (SCL-90-R) were somatization, depression, anxiety, interpersonal sensitivity, hostility, obsessive-compulsive behavior, phobic anxiety, paranoid ideation, and psychoticism. The impact of personal variables (age, gender, marital status) and programmatic variables (level of training, laboratory experience, institution) was assessed. RESULTS: Mean psychologic distress was significantly higher in general surgery residents than in the normative population (p < 0.0001), with 38% scoring above the 90th percentile and 72% above the 50th percentile. Mean perceived stress among surgery residents was higher than historic controls (p < 0.0001), with 21% scoring above the 90th percentile and 68% above the 50th percentile. Among secondary outcomes, eight of nine symptom dimensions were significantly higher in surgical residents than in societal controls. In subgroup analyses, male gender was associated with phobic anxiety (p < 0.001) and anxiety (p < 0.05), and junior level of training (PGY 1 to 3) with anxiety (p < 0.05), obsessive-compulsive behavior (p < 0.05), and interpersonal sensitivity (p < 0.05). CONCLUSIONS: More than one-third of general surgery residents meet criteria for clinical psychologic distress. Surgery residents perceive significantly more stress than societal controls. Both personal and programmatic variables likely affect resident well-being and should be considered in assessing the full impact of Accreditation Council on Graduate Medical Education directives and in guiding future restructuring efforts.