As part of efforts to improve the quality of care, the Centers for Medicare and Medicaid Services (CMS) will not reimburse hospitals for additional costs incurred when a patient experiences certain complications of care (e.g., a decubitus ulcer), and some private insurers have followed suit. This study tested the validity of using administrative (billing) data to make these determinations, by evaluating the accuracy of the AHRQ Patient Safety Indicators at identifying in-hospital complications. For several patient safety indicators, the authors found that the majority of cases were actually pre-existing conditions, and thus did not represent a complication of care. Furthermore, cases where an in-hospital complication occurred did not receive significantly higher financial payments than cases without complications. A prior study also found that screening administrative data alone was an inadequate method for identifying post-surgical complications.