ORGANIZATION The Centers for Medicare and Medicaid Services (CMS) PROBLEM OR ISSUE ADDRESSED In 2005, CMS issued a national coverage determination for Positron Emission Tomography (PET) for oncologic indications not previously covered by Medicare, which provided payment for PET scans for these indications only in the context of an approved prospective clinical trial designed to assess clinical utility of PET. CMS instated this policy option, referred to as ‘Coverage with Evidence Development' (CED), due to the lack of evidence demonstrating the clinical effectiveness of PET scans for patient management, which is necessary to determine whether PET scans are reasonable and necessary for the diagnosis, staging, re-staging, and monitoring of various cancer types. CED allowed CMS to guide evidence development while providing access to this potentially beneficial technology. GOALS To determine whether PET imaging is clinically effective for previously uncovered oncologic indications. OUTCOMES ITEMS USED IN THE DECISION The main outcomes were change in intended management strategy and whether the PET scan allowed physicians to avoid other tests or procedures. Based on these, Medicare's Evidence Development and Coverage Advisory Committee (MEDCAC) members had to decide how confident they were that PET improves physician decision-making and clinical outcomes, and that the conclusions were generalizable to other cancers, to PET facilities in the general community, and to the Medicare population. IMPLEMENTATION STRATEGY Since CMS did not have the capacity to design and fund a study, the agency partnered with the American College of Radiology and the Academy of Molecular Imaging. The study itself had to be implemented within a short time do to federal regulations. The registry began patient and physician registration in May of 2006. This design was chosen in part because it balanced the desire for access to this service with the goal of generating evidence of reasonable quality that could be used by CMS to make a final coverage determination. RESULTS After the first year, most PET facilities in the United States had signed up to participate in the registry and a huge nationwide sample of data had been collected. The results from this data analysis demonstrate that physicians report a change in intended disease management strategy in about one out of every three cases and that when broken down by disease cancer type and indication, the figure remains fairly constant (Hillner et al. 2008). This evidence was reviewed in August 2008 at a meeting of MEDCAC. Based on these results and the results of a health technology assessment, MEDCAC members were asked to rate their confidence in the clinical utility of PET and the generalizability of the conclusions. The ratings demonstrated that MEDCAC members have limited confidence that PET improves clinical outcomes based on the evidence. Members pointed out that although the results show that physicians often change their intended management plan, there is no way to know if actual management changed and even if it did, it is also not known whether these changes actually lead to better patient outcomes. Still, on January 6, 2009, CMS issued a decision memo stating that there is now sufficient evidence showing that PET improves health outcomes when used for the diagnosis and staging of all previously uncovered cancer types, warranting coverage for these indications. However, CMS does not believe that there is enough evidence demonstrating the clinical utility of PET for monitoring response to treatment and re-staging. Therefore, these indications will still only be covered through the CED policy, likely necessitating the development of a new prospective study. LESSONS LEARNED The case presented above not only demonstrates that CED is operationally and technically possible while abiding to regulatory procedures, it also demonstrates CED can be used to help support research efforts designed to address questions of importance to health care decision-makers. In addition, as this is one of the first examples of Medicare taking advantage of CED, it also offers some lessons for the future. For instance, in order to streamline the process, it is necessary to identify a stable source of funding for these projects, to identify promising technologies which lack evidence of clinical effectiveness earlier in the development process, to reach a shared understanding among stakeholders of standards of evidence that are feasible and sufficiently robust for coverage decisions, and to reach a consensus as to the most efficient methods for conducting real world trials. In the future, it would be best if decisions of study design did not take place while Medicare was making a coverage determination and if private payers became involved in CED so a broader range of patients could participate in the studies.