Under Medicare, hospitals generally receive fixed payments for inpatient stays based on diagnosis-related groups (DRG), a system that classifies stays by patient diagnoses and procedures. The Centers for Medicare & Medicaid Services (CMS) annually uses its own data to reclassify DRGs. CMS also makes add-on payments for stays involving new technologies that meet three eligibility criteria. Stakeholders may submit data that are external to CMS as part of a DRG reclassification request or an add-on payment application. The Medicare Prescription Drug, Improvement, and Modernization Act of 2003 required GAO to examine whether CMS could improve its use of external data, including using data collected by other government agencies for DRG payments. As discussed with the committees of jurisdiction, GAO examined (1) to what extent CMS has used external data in determining payments for inpatient stays involving new technologies, and (2) to what extent can external data from other government agencies be used by CMS in determining DRG payments for inpatient stays involving new technologies. GAO interviewed officials from CMS and industry stakeholders. GAO interviewed officials from Bureau of Labor Statistics (BLS), Department of Veterans Affairs (VA), Department of Defense (DOD), and Agency for Healthcare Research and Quality (AHRQ) because these agencies may have data useful to CMS. GAO also reviewed regulations and other CMS materials.

CMS has used external data for two purposes: to inform DRG reclassification and to evaluate new technology add-on payment applications. To inform DRG reclassification, CMS accepts the submission of external data that are intended to demonstrate that inpatient stays involving a new technology are costlier on average than the other inpatient stays in the same DRG. CMS uses its data from the Medicare Provider Analysis and Review (MEDPAR) file to validate the external data submitted. Specifically, when external data are submitted for a proposed DRG reclassification for a procedure or new technology, CMS's policy is to find the same or similar evidence in the MEDPAR file. Generally, CMS will not make a reclassification decision for a DRG involving a new technology if the technology is so new that it does not appear in the MEDPAR file. To evaluate new technology add-on payment applications, CMS has generally used external data in conjunction with data from the MEDPAR file to evaluate whether a new technology meets one of the three eligibility criteria, specifically the criterion related to cost. Data from other government agencies have limitations for CMS's use in setting DRG payments for inpatient stays involving new technologies. This is because when setting DRG payments, CMS generally needs data that are representative of the Medicare population, timely, and complete in that the data include the total charge or other measure of costliness for all services provided during an inpatient stay, including new technologies. The data we identified from other government agencies were either not representative of the Medicare population, were not timelier than data from the MEDPAR file, or were not complete. Data from the MEDPAR file remain the primary data source for setting DRG payments because they include all charges from paid inpatient claims for inpatient services provided to all Medicare beneficiaries across all hospitals paid under the IPPS. In instances where data from the MEDPAR file have lacked charge information for certain stays involving new technologies, CMS has used external data to inform the DRG reclassification process and to evaluate new technology add-on payment applications. To set DRG payments, CMS needs data that meet criteria of being representative, timely, and complete. Although BLS, VA, DOD, and AHRQ collect data for their own purposes that could potentially be useful to CMS, these data are limited in their utility to set DRG payments because they do not always meet CMS's criteria. In commenting on a draft of this report, CMS stated that it agreed with GAO's findings.