OBJECTIVE: To examine the impacts of a large-scale simulation -based extracorporeal cardiopulmonary resuscitation (ECPR) training program in an academic children's hospital. METHODS: The study followed a quasi-experimental, mixed method, time series design. Two-hour high-fidelity ECPR simulations were held monthly in the pediatric, cardiac, and neonatal intensive care units. Intensive care unit specific cases were used in each unit. The learning objectives for all cases were the same. Each simulation included an average of 11 health care professionals, including nurses, physicians, respiratory therapist, and perfusionists. Impacts of training were examined using Kirkpatrick' s 4-level model: reactions, learning, behaviors, and results. Participant surveys, semistructured interviews, facilitator observations, applied cognitive task analysis, and hospital code data were used to examine the impacts of training. RESULTS: From February 2014 to October 2016, a total of 332 health care professionals participated in 29 ECPR simulations. Participants enjoyed the simulations and reported learning gains. Applied cognitive task analysis revealed 2 specific behaviors, coordination of compressions with surgical cannulation and performing sterile compressions, that were targeted for further training. The rate of adherence to the ECPR activation protocol improved from 83% (48/58) before simulations started to 95% (92/97) after simulations (P=.02). ECPR activation time decreased from 7 minutes (interquartile range, 4-9 minutes) before simulations started to 2 minutes (inter quartile range, 1-4 minutes) after simulations (P<.01). CONCLUSIONS: Large-scale simulation-based ECPR training was associated with positive reactions, learning gains, behavioral change, improved adherence to the ECPR activation protocols, and faster activation times. Other children's hospital that perform ECPR should consider simulation-based training.