Introduction: High blood culture contaminant rates (BCCR) in the emergency department (ED) contribute to unnecessary tests and expenses. The national benchmark for BCCR is 2%. We had a baseline BCCR of >3% in our ED. Using the Model For Improvement framework, our global aim was to decrease the number of blood culture contaminants in our ED. Our specific aim was to decrease the BCCR by 50% within 12 months (Fig.1). We hypothesized that two key drivers would help achieve our specific aim: increasing venipuncture sterility and reducing the number of blood cultures. Methods: This prospective, single-center, quasi-experimental study was performed in an academic pediatric ED (90,000 patients/year volume). A multi-disciplinary team of front-line nurses and physicians formed a quality improvement team and created a Key Driver Diagram with multiple Plan-Do-Study-Act cycle interventions based on various change concepts (Fig.1). For our specific aim calculations, we included all peripheral blood cultures drawn in the ED except those from patients with cancer, central lines, ventriculo-peritoneal shunts, neutropenia, or transplant history. We classified positive cultures as pathogens or contaminants based on published guidelines. We used an interrupted time series design and applied validated tests to detect special cause variation with a statistical process control T chart. For secondary aims, we measured: (1) the differences in blood culture ordering rates (BCOR; = number of included blood cultures / number of patients with temperature ≤35.5 or ≥38.0) between the baseline and PDSA3 periods, and (2) the estimated annual savings in patient charges based on differences in BCCR and BCOR between the baseline and PDSA3 periods. We used an inflation-adjusted value from our previously published study ($1494 charges/contaminant). Results: Our one-year baseline period, PDSA1, PDSA2, and PDSA3 BCCRs were 3.02%, 2.30%, 1.58%, and 1.17%, respectively. We achieved our specific aim (61% BCCR decrease). Our T chart had special cause variation in PDSA1 and PDSA3, and demonstrated improvement from one contaminant culture every 1.87 days (baseline) to one every 7.32 days (PDSA3). The BCOR decreased from 26.2% (baseline) to 17.4% (PDSA3); p < 0.001. We estimate that we are annually preventing 19 contaminants, sending 1,615 fewer blood cultures, and saving patients with contaminant cultures a total of $28,332 in patient charges. Discussion: By standardizing our blood culture ordering and collection processes, we have decreased the number of blood culture contaminants in our ED. Our processes could be tested in other EDs to determine generalizability. We are now measuring for any differences in a balance measure: the rate of ED bounceback admissions for bacteremia who did not have a blood culture drawn on the first ED visit. Future PDSA cycles will aim to continue to improve our processes using other high reliability principles (e.g., clinical decision support) and change concepts.