Of 16 studies that met the criteria, only four (25%) randomized controlled trials were identified, and the remaining were observational studies. Seven observational studies (43.7%) were single-group pre-post tests. The mean number of trainees was 20.4 (SD, 14.1). Low-fidelity simulators were used in 13 studies (81.2%). Most of the studies (81.3%) were high quality based on a Medical Education Research Study Quality Instrument score of 12 or more. Evidence of assessment tool validation was absent among all studies. No study outcome measures were directed to skills transfer to the operating room or patient outcomes. Overall learning outcomes' effect sizes were consistently high (2.2; SD, 1.6), with junior residents benefitting most (effect size, 2.8; SD, 2.2) CONCLUSIONS: Simulation-based skill training is associated with improved learning outcomes for cardiac surgery trainees with large effect sizes, but more behavior-level outcomes are required to fully assess its value.