IMPORTANCE Serious, preventable surgical events, termed never events, continue to occur despite considerable patient safety efforts. OBJECTIVE To examine the incidence and root causes of and interventions to prevent wrong-site surgery, retained surgical items, and surgical fires in the era after the implementation of the Universal Protocol in 2004. DATA SOURCES We searched 9 electronic databases for entries from 2004 through June 30, 2014, screened references, and consulted experts. STUDY SELECTION Two independent reviewers identified relevant publications in June 2014. DATA EXTRACTION AND SYNTHESIS One reviewer used a standardized form to extract data and a second reviewer checked the data. Strength of evidence was established by the review team. Data extraction was completed in January 2015. MAIN OUTCOMES AND MEASURES Incidence of wrong-site surgery, retained surgical items, and surgical fires. RESULTS We found 138 empirical studies that met our inclusion criteria. Incidence estimates for wrong-site surgery in US settings varied by data source and procedure (median estimate, 0.09 events per 10 000 surgical procedures). The median estimate for retained surgical items was 1.32 events per 10 000 procedures, but estimates varied by item and procedure. The per-procedure surgical fire incidence is unknown. A frequently reported root cause was inadequate communication. Methodologic challenges associated with investigating changes in rare events limit the conclusions of 78 intervention evaluations. Limited evidence supported the Universal Protocol (5 studies), education (4 studies), and team training (4 studies) interventions to prevent wrong-site surgery. Limited evidence exists to prevent retained surgical items by using data-matrix-coded sponge-counting systems (5 pertinent studies). Evidence for preventing surgical fires was insufficient, and intervention effects were not estimable. CONCLUSIONS AND RELEVANCE Current estimates for wrong-site surgery and retained surgical items are 1 event per 100 000 and 1 event per 10 000 procedures, respectively, but the precision is uncertain, and the per-procedure prevalence of surgical fires is not known. Root-cause analyses suggest the need for improved communication. Despite promising approaches and global Universal Protocol evaluations, empirical evidence for interventions is limited.