Response to a request for information on improving data on Medicare Advantage

Medical stethoscope with transparent overlay of data and line graph.
Shutterstock / janews
Editor's note:

The authors submitted this comment letter to the Centers for Medicare & Medicaid Services and the Department of Health and Human Services on May 29, 2024.

Loren Adler and Matthew Fiedler responded to a request for information from federal officials on how the Centers for Medicare and Medicaid Services (CMS) can improve data related to Medicare Advantage (MA). The authors note that MA now accounts for more than half of Medicare enrollment, so robust data on MA plans’ interactions with beneficiaries, providers, CMS, and other entities is essential to understand how well the Medicare program is operating and how to make it work better.

Their letter makes several specific recommendations about how CMS can improve the data available on the MA program:

  • Improve MA encounter data: The most important priority for CMS is improving MA encounter data. These data are the only source of comprehensive, granular information on the health care received by MA enrollees and, as such, are the linchpin of efforts to understand MA. To make them more useful, CMS should:
    • allow researchers to access data on provider payments that CMS already collects as part of encounter records and, if necessary, improve these data to ensure that they reliably measure both enrollee cost-sharing and plan payments;
    • allow researchers to access data on claim denials that CMS already collects as part of encounter records and, if necessary, improve these data to ensure that they can be used to reliably identify encounters that result in a denied claim;
    • proceed with plans to collect detailed data on utilization of supplemental benefits and make these data available to researchers once they become available; and
    • consider collecting granular information on prior authorization requests.

CMS should also take steps improve data quality since there is evidence that the encounter data suffer from some data quality problems. More generally, CMS should recognize that while the encounter data began as a tool for administering risk adjustment, they now serve many other purposes, and its approach to these data should evolve accordingly.

  • Collect data on MA plans’ non-fee-for-service provider payments: MA plans often pay providers using non-fee-for-service methods, but, unlike fee-for-service payments, many such payments are not captured on encounter records. This gap impedes efforts to understand the overall level of MA provider payments, plans’ utilization management and diagnosis coding efforts, and compliance with medical loss ratio (MLR) rules. CMS should collect plan- and provider-level data on the size and characteristics of these payments.
  • Release ownership data for all provider types and improve ownership data quality: Data on provider ownership has many uses, but it can be particularly important to understanding MA since owning providers may help plans increase diagnosis coding intensity or circumvent MLR regulations, among other effects. CMS generally collects data on provider ownership as part of the Medicare enrollment process but has only released these data for certain provider types, notably excluding physician practices. CMS should release these data for all provider types. Additionally, since analyses of released data have identified data quality problems, CMS should take steps to improve data quality.
  • Make available certain other data that CMS already holds: CMS can also improve understanding of MA by releasing (or more promptly releasing) certain other data that it already holds. Specifically, CMS should: (1) release data on plan bids promptly, rather than waiting five years to do so; and (2) make final MA risk score data available to researchers.

Read the full comment letter here.

  • Acknowledgements and disclosures

    The Brookings Institution is financed through the support of a diverse array of foundations, corporations, governments, individuals, as well as an endowment. A list of donors can be found in our annual reports published online here. The findings, interpretations, and conclusions in this report are solely those of its author(s) and are not influenced by any donation.