About one-third of Medicare beneficiaries now rely on Medicare Advantage (MA), the private
alternative to traditional Medicare, almost twice as many as did a decade ago. A mainstay of
how MA manages costs and quality is through restricting access to specific networks of covered
physicians and hospitals. However, we know little about how these limitations actually affect the
quality or beneficiary cost of care. This lack of information has two major implications:
consumers don’t know what they are buying in MA (in terms of the quality of care they will
receive) and policymakers can’t effectively regulate MA markets to optimize access, quality, or
costs. The lack of information is driven primarily by the fact that we have little, reliable data
about the networks MA provides its patients. Circumventing these challenges, our team has
pioneered an innovative approach to characterize the composition of MA physician networks
and access on a national scale and over time. This project would apply this approach to high-
cost specialists, as well as determine relationships between access and plan quality and costs
for beneficiaries.
Our broad goal is to understand MA plans’ care delivery though their networks and the
resulting impact on quality and costs for MA patients. We previously exploited Part D, the
Medicare prescription drug data, to infer networks from prescription events for primary care
physicians. The enabling principle is that physicians who prescribe more for enrollees in a plan
are more likely to be in-network for that plan than physicians who prescribe less. Though it
would be preferable to rely on plan directories for network information, our prior work shows that
plan-reported networks have substantial errors. Moreover, they do not capture which physicians
are accessible and accepting patients. While there are recently released MA claims data
available, they are limited to just three years and are of unknown accuracy. Thus, our approach
remains the only known, reliable way to assess MA networks.
The proposed project is crucial at this time because the extent to which MA, through
establishment of networks, influences the quality of care is unknown to policymakers. Moreover,
no specific guidance currently exists as to what beneficiaries gain or give up in exchange for
narrower or broader networks in MA. As there are likely tradeoffs across dimensions of quality,
premiums, and cost sharing, the findings of this study will be critical for beneficiaries (informing
how they choose plans), as well as policymakers, suggesting where greater regulatory scrutiny
around network adequacy and quality may be needed.