Administrative data generated for billing purposes is often used as a means of detecting adverse events. This method is limited by the fact that the ICD-9 diagnosis coding system does not specifically define many common adverse events, and as a result AHRQ developed the Patient Safety Indicators (PSIs) to screen administrative data for this purpose. The updated ICD-10 coding system will be implemented nationwide in the United States in 2015. The Centers for Medicare and Medicaid Services has proposed translations of the PSIs for ICD-10, but this study found substantial inaccuracies when ICD-10-based administrative data was screened using these translations. For example, the PSI for pressure ulcers demonstrated a high false-negative rate, meaning that ICD-10 based data will likely under-report the true incidence of this never event. Other PSIs demonstrated the opposite problem—potential for over-reporting due to a high false-positive rate—and some PSIs have greater potential for variability in interpretation by coders due to the substantially larger number of diagnoses included in ICD-10 (a problem noted in a prior systematic review). The results of this study raise serious concerns about the validity of administrative data for measuring patient safety in the ICD-10 era.