Coverage with Evidence Development (CED) is a Medicare policy that provides coverage for promising but not yet fully proven medical treatments or services on the condition that patients participate in approved clinical studies or data registries. Although the policy is intended to balance faster patient access to emerging technologies with the need for continued evidence generation, in practice it creates a two-tiered healthcare system in which access to new medical technologies depends on where patients receive care. Our research shows that these requirements systematically disadvantage some of the most vulnerable Medicare beneficiaries, particularly rural and low-income patients.
The structure and implementation of CED may undermine the original goals of the program
CED was created by the Centers for Medicare & Medicaid Services (CMS) as a policy mechanism intended to balance early patient access to promising medical technologies with the need for stronger clinical evidence. Under CED, CMS may grant conditional Medicare coverage when a therapy appears promising but existing evidence is insufficient to satisfy the statutory standard that services be “reasonable and necessary” for Medicare beneficiaries. Rather than providing unrestricted national coverage or denying access outright, CMS requires that patients receiving these technologies participate in approved clinical trials, registries, or other evidence-generation programs designed to collect real-world data on safety, effectiveness, and clinical outcomes.
The structure of the CED program may create disparities in access. To qualify for Medicare reimbursement, studies conducted under CED must satisfy extensive methodological and oversight requirements established by CMS. Participating providers are often required to conduct randomized trials, prospective observational studies, or structured patient registries with rigorous data collection, longitudinal follow-up, and sophisticated statistical analysis. Additionally, CED frequently imposes minimum volume requirements on hospitals to participate. These requirements can create substantial administrative and financial burdens that are more easily absorbed by large academic medical centers than by smaller community hospitals or rural health systems. As a result, access to CED-covered technologies may become concentrated in well-resourced institutions, potentially limiting access among underserved populations, including rural beneficiaries, and patients with lower socioeconomic status.
In practice, the implementation of CED has raised additional concerns about equitable access to new technologies. Since 2005, CMS has issued approximately 27 CED decisions covering a range of technologies and services, yet only a small fraction of these decisions have undergone formal reconsideration of the evidence development requirements. This means that most of the CED decisions have resulted in prolonged periods of evidence gathering that restricts patients’ access to participating facilities. In fact, among the 27 CED decisions, only four have had their evidence-development requirements retired, typically after periods ranging from 4 to 12 years. Three of those reconsidered decisions ultimately resulted in broad national Medicare coverage, while the most recent example—Beta-Amyloid PET imaging for Alzheimer’s disease—had national coverage withdrawn entirely, leaving coverage decisions to local Medicare Administrative Contractors. More broadly, the duration of CED requirements has varied dramatically, with some remaining in place for nearly two decades and thus limiting access to innovative technologies for prolonged periods.
Examining CED decisions impact on patient access
To investigate the impact that CED decisions have on patients’ access to new treatments and services, we compared hospitals that have participated in CED registries and/or clinical trials to hospitals that have not. Specifically, we reviewed every CED decision and identified all hospitals that participated in CED trials using ClinicalTrials.gov and all registries, for instance the transcatheter aortic valve replacement (TAVR) and left atrial appendage occlusion (LAAO) registries. We then linked participating hospitals to Medicare’s “Hospital General Information” file that provides a unique facility ID for each hospital. That ID allowed us to merge discharge statistics for each hospital, facilitating our comparison of patient populations served by CED-participating and non-participating hospitals. We compared demographic characteristics along with measures of social disadvantage.
The geographic divide in healthcare innovation
The results highlight stark disparities in access to healthcare innovation:
- Rural patients are dramatically underserved. At hospitals participating in CED registries, rural patients accounted for 16.3% of discharges, compared with 25.4% at non-participating hospitals (Figure 1). A similar gap appears for hospitals participating in CED clinical trials (14.2% v 21.2%). In both cases, participating hospitals treated about one-third fewer rural patients than non-participating hospitals. This means rural Medicare beneficiaries face significant barriers to accessing innovative treatments like TAVR and LAAO, which require participation in a CED registry.
- Low-income beneficiaries face similar disparities. We also found disparities in socioeconomic status. Hospitals participating in CED registries and clinical trials treated fewer low-income subsidy (LIS)-enrolled beneficiaries and fewer patients from socioeconomically disadvantaged neighborhoods. For example, LIS-enrolled beneficiaries accounted for 27.5% of discharges at registry-participating hospitals compared with 31.0% at non-participating hospitals, with similar differences observed for CED clinical trial sites (27.7% vs. 29.1%) (Figure 1). We also found lower Area Deprivation Index (ADI) scores among patients treated at CED-participating hospitals, suggesting these hospitals served fewer patients from disadvantaged communities.
Figure 1. Comparison of CED vs Non-CED Hospital Discharge Statistics
How evidence requirements undermine their own goals
These access disparities create a fundamental tension within the CED program. One key goal of CED is to generate evidence on a nationally representative sample of clinically eligible patients. But when coverage is restricted to “centers of excellence” that are geographically concentrated and serve different patient populations, the resulting evidence base becomes skewed.
Instead of learning how new technologies perform across diverse patient populations and care settings, CED generates evidence primarily from patients with ready access to academic medical centers and other high-resource facilities. This limits our understanding of how these treatments might work for the broader Medicare population.
Some may argue that less-resourced hospitals may face barriers to adopting new technologies because they lack the patient volume, specialized equipment, or technical expertise to support implementation. However, CED has been applied to a wide range of interventions, ranging from continuous positive airway pressure (CPAP) for sleep apnea and home oxygen for chronic obstructive pulmonary disease (COPD) to pharmacogenomic testing for warfarin and for off-label oncology treatments. The primary requirements for these types of technologies are the ability to collect outcomes data and participate in evidence-generation activities. Consequently, concerns that rural or resource-constrained hospitals cannot participate in CED evidence-generation may be overstated. Even in cases where hospital resource constraints affect outcomes, the right approach is inclusion of and technical support for rural and other resource-constrained hospitals
The path forward: Rethinking evidence generation
Medicare needs to reconsider how it structures evidence requirements to avoid creating barriers to equitable access. Some potential approaches include:
- Expanding the network of participating sites to include safety-net hospitals and rural facilities
- Providing additional support to help resource-limited hospitals participate in registries
- Developing alternative evidence pathways that don’t require real-time registry participation for every patient
- Monitoring equity metrics when implementing CED decisions to ensure they don’t exacerbate disparities
Medical innovation can only improve population health if patients can access it. As Medicare continues to evaluate new and expensive technologies, it should be careful that evidence-generation policies do not unintentionally widen existing disparities in care. Otherwise, the patients most likely to benefit from innovation may continue to face the greatest barriers to receiving it.
