OBJECTIVE Timely provision of evidence-based recommendations through computerized physician order entry with clinical decision support may improve use of red blood cell transfusions (RBCTs). METHODS We performed a cohort study with historical controls including inpatients admitted between February 1, 2008, and January 31, 2010. A clinical decision-support alert for RBCTs was constructed by using current evidence. RBCT orders resulted in assessment of the patient's medical record with prescriber notification if parameters were not within recommended ranges. Primary end points included the average pretransfusion hemoglobin level and the rate of RBCTs per patient-day. RESULTS In total, 3293 control discharges and 3492 study discharges were evaluated. The mean (SD) control pretransfusion hemoglobin level in the PICU was 9.83 (2.63) g/dL (95% confidence interval [CI]: 9.65-10.01) compared with the study value of 8.75 (2.05) g/dL (95% CI: 8.59-8.90) (P < .0001). The wards' control value was 7.56 (0.93) g/dL (95% CI: 7.47-7.65), the study value was 7.14 (1.01) g/dL (95% CI: 6.99-7.28) (P < .0001). The control PICU rate of RBCTs per patient-day was 0.20 (0.11) (95% CI: 0.13-0.27), the study rate was 0.14 (0.04) (95% CI: 0.11-0.17) (P = .12). The PICU's control rate was 0.033 (0.01) (95% CI: 0.02-0.04), and the study rate was 0.017 (0.007) (95% CI: 0.01-0.02) (P < .0001). There was no difference in mortality rates across all cohorts. CONCLUSIONS Implementation of clinical decision-support alerts was associated with a decrease in RBCTs, which suggests improved adoption of evidence-based recommendations. This strategy might be widely applied to promote timely adoption of scientific evidence.