OBJECTIVES: Medical errors cause approximately 44,000 to 98,000 injuries annually in the United States (US). This study estimated the occurrence of medical errors and associated costs using actuarial methodology developed by Milliman, Inc. applied to Premier hospital data. METHODS: The Milliman study utilized an expert panel to estimate how often medical injuries were likely to be associated with a medical error rather than the underlying disease. Injuries were classified into five groups based on the likelihood that they were associated with a medical error. The midpoint of each range of likelihood of medical error was applied to the frequency of each medical injury to establish the rate of medical error. This study utilized Milliman’s ranges and midpoints applied to Premier data. Injury rates, error estimates and cost-per-error estimates were developed from Premier 2008-2009 data. A visit qualified as an injury visit if >1 of 97 injury groupings identified by Milliman occurred. Non-injury control groups were established using propensity score matching. Population estimates were projected from Premier to all US acute care hospitals. RESULTS: While the focus of the earlier report was health plan claims data, this study utilized hospital billing records to demonstrate direct cost of medical errors to hospitals. The rates of medical errors were similar to the Milliman study and both studies found pressure ulcers and postoperative infection were the most common types of medical errors. Milliman estimated the total cost of medical errors to be approximately $17 billion in 2008. This analysis found that hospitals bear approximately 1 billion in costs related to medical errors. CONCLUSIONS: Utilizing methods developed by Milliman, this study estimated the rates and cost of medical errors to hospitals utilizing hospital billing data, revealing similar patterns of errors.