OBJECTIVE This study develops a methodology for identifying complications following carotid endarterectomy, using physician claims data. DATA SOURCES/STUDY SETTING We selected a random 20 percent sample of Medicare patients undergoing carotid endarterectomy in 1991 (n = 8,345) and extracted all of their claims. STUDY DESIGN Project neurologists identified the following services as indicative of complications following carotid endarterectomy if they were provided within 30 days of surgery: head CT, head MRI, and surgical exploration of the neck for hemorrhage, thrombosis, or infection. DATA COLLECTION/EXTRACTION METHODS Total costs were calculated from all claims associated with the hospitalization and the 30-day postoperative period. Outcomes included mortality (obtained from Medicare eligibility files), length of stay, discharge to an institution, and readmission to an acute care hospital (the latter obtained from claims data). PRINCIPAL FINDINGS Surgical complications were identified in one out of every ten endarterectomy patients (10.3 percent). Patients with complications were significantly more likely to die within 30 days of surgery (8.9 percent, compared with 1.1 percent of those not experiencing complications). They also were significantly more likely to be discharged to an institutional setting (24.9 percent versus 2.9 percent), and more likely to be readmitted to acute care hospitals (26.8 percent versus 8.2 percent). Patients with postoperative complications also were significantly more expensive: $22,187 versus $10,892. CONCLUSION Our findings suggest that physician claims could be used by PROs or similar entities as a screening tool to identify potential problem hospitals or problem surgeons. First, however, the methodology would need to be clinically validated.