OBJECTIVE: To assess the impact of an automated intraoperative alert to redose prophylactic antibiotics in prolonged cardiac operations. DESIGN: Randomized, controlled, evaluator-blinded trial. SETTING: University-affiliated hospital. PATIENTS: Patients undergoing cardiac surgery that lasted more than 4 hours after the preoperative administration of cefazolin, unless they were receiving therapeutic antibiotics at the time of surgery. INTERVENTION: Randomization to an audible and visual reminder on the operating room computer console at 225 minutes after the administration of preoperative antibiotics (reminder group, n = 137) or control (n = 136). After another 30 minutes, the circulating nurse was required to indicate whether a follow-up dose of antibiotics had been administered. RESULTS: Intraoperative redosing was significantly more frequent in the reminder group (93 of 137; 68%) than in the control group (55 of 136; 40%) (adjusted odds ratio, 3.31; 95% confidence interval, 1.97 to 5.56; P < .0001). The impact of the reminder was even greater when compared with the 6 months preceding the study period (129 of 480; 27%; P < .001), suggesting some spillover effect on the control group. Redosing was formally declined for 19 of the 44 patients in the reminder group without redosing. The rate of surgical-site infection in the reminder group (5 of 137; 4%) was similar to that in the control group (8 of 136; 6%; P = .42), but significantly lower than that in the pre-study period (48 of 480; 10%; P = .02). CONCLUSION: The use of an automatic reminder system in the operating room improved compliance with guidelines on perioperative antibiotic prophylaxis.