In Reply: Drs Utter and Romano have made important contributions to society's ability to measure patient safety using patient safety indicators based on administrative (billing) data. But their response to our article is puzzling. Our central claim was that billing data cannot be relied on to simultaneously measure quality, publicly report quality, and pay for performance. If they are, the ability to measure true changes in quality will be lost. Our secondary claim was that there is currently no substitute for billing data as a widely available basis for measuring outcomes. We concluded that there is an urgent need to develop alternate data sources not currently used for public reporting or reimbursement that will provide timeconsistent quality measures. 1 Rather than dispute our main claim, Utter and Romano state that a consultant report finds that the 2008 decreases in patient safety indicator 5 (leaving a foreign object in the body during surgery) and patient safety indicator 7 (CLABSI) reporting were only 15% and 23% (instead of the roughly 50% we found). This report 2 uses data we cannot verify, uses annual instead of quarterly data, and shows decreases of 26% and 30% from fiscal 2008 to 2009. Utter and Romano extract lower percentages from this report by including an increase in rates from 2009 to 2010, which is likely unrelated to the reimbursement change. Even if those percentages were accurate, our main message would remain unchanged. Decreases of this magnitude offer ample evidence that if payers stop paying for hospital-acquired conditions, many hospitals will stop billing for them. Utter and Romano also do not dispute our secondary claim. We agree that the determinants of gaming are the ease of doing so and the associated incentives. We disagree, however, that performance metrics based on administrative data are less susceptible to gaming than indirect measures of performance and measures that are purposely collected for quality measurement. Moreover, not billing for events that insurers will not pay for is not even gaming, in a sense. No rule requires billing for adverse events. If there was such a rule, we cannot imagine that it would be accompanied by a serious effort to audit the events for which hospitals are not billing. The intent of hospitals is to offer safe, high-quality care; however, we contend that financial imperatives are often paramount when generating administrative data. Significant discretion exists in the documentation and coding of patient episodes. This point was underscored by a recent Inspector General report 1 that found adoption of electronic medical records has led to higher billing levels without true changes in care. Gaming reflects manipulation of