The author compares surgical event data from the first Minnesota state report on medical error with published research on similar incidents. He concludes that, because adverse events occur infrequently in individual institutions, large-scale databases are needed to aggregate data for study.