In a highly anticipated and scrutinized decision, the Centers for Medicare & Medicaid Services (CMS) released a new public data set on April 30, 2015, that includes detailed data regarding the prescribing patterns of physicians and other health care providers under the Medicare Part D Prescription Drug Program. Referred to as the Provider Utilization and Payment Data: Part D Prescriber Public Use File, the file comprises data relating to more than one million health care providers that prescribed $103 billion in prescription drugs across 36 million Medicare beneficiaries under the Part D program in 2013 (the Data Set). The decision reflects a growing effort by CMS to use data sharing to promote transparency and innovation, but users should be mindful of the context and limitations of the data, as cautioned by CMS.

The Data Set and Context of its Release

The Data Set is significant for its size and scope, covering approximately $103 billion in purchases of approximately 3,000 distinct drug products by Medicare Part D beneficiaries. The Data Set includes both brand name and generic prescriptions dispensed by individual prescribers (as identified by National Provider Identifier); it distinguishes between original prescriptions and refills; and it itemizes total drug costs paid by the Part D plan, Medicare beneficiary, other government subsidies, and any other third-party payers. The data is collated from the CMS Prescription Drug Event Standard Analytic File, which contains final-action claims submitted by Medicare Advantage prescription drug plans and stand-alone prescription drug plans. CMS has indicated its intention to publically release such data annually (“CMS releases prescriber-level Medicare data for first time” ).

CMS framed the historic data release in the context of a January 2015 announcement by the Secretary of the Department of Health and Human Services (HHS) that outlined goals and a timeline to shift Medicare payment toward value-based care. The agency underscored the promise of open sharing of data to promote new ideas and innovations around how health care is delivered. Emphasizing that the released data will enable stakeholders to conduct a wide variety of analyses on prescription drug utilization and spending, CMS stated its hope that the public availability and use of the Part D data will reveal new insights and instigate improvements and efficiencies in the private sector’s administration of the Part D program [“New Medicare prescription drug cost data available,” CTRS. FOR MEDICARE & MEDICAID SERVS.  (Apr. 30, 2015)].

In announcing the release of the Data Set, CMS importantly qualified that the total drug cost data in the Data Set does not reflect any manufacturer rebates paid to Part D plan sponsors, which result in lower consumer out-of-pocket costs. The agency also noted that the data does not reflect the quality of care offered by individual physicians and that it solely comprises beneficiaries enrolled in the Medicare Part D prescription drug program and thus is likely not representative of a physician’s broader practice or even his or her Medicare patient population. The caveats likely are intended to respond to concerns within the provider community that the data may be misconstrued, misunderstood or misused. In addition, in an effort to protect Medicare beneficiary privacy, the Data Set omits any information in instances in which 10 or fewer prescriptions were dispensed (“CMS releases prescriber-level Medicare data for first time,” supra).

Considerations and Implications

CMS’s decision to release public-use files that identify individual Medicare prescribers and prescribing costs exemplifies CMS’s initiative to make Medicare costs associated with individual physicians more readily available for research and public oversight. Physician-identifiable information collected by CMS is subject to the provisions of the Privacy Act of 1974. As such, prior to January 2014, the official position of HHS was that the privacy interests of physicians outweighed public interest in the disclosure of amounts paid to individual physicians under the Medicare Part B program [U.S. Dep’t of Health and Human Servs., Request for Public Comments on the Potential Release of Medicare Physician Data  (Aug. 6, 2013)]. CMS revisited this position after the U.S. District Court for the Middle District of Florida decided in May 2013 to lift a permanent injunction that had prohibited disclosures of such data. After soliciting public comment on whether and how to change its policy, CMS announced in January 2014 that it would make case-by-case determinations in response to requests for the Medicare payment data of individual physicians [Jonathan Blum, “CMS Modifies Policy on Disclosure of Physician Payment Information,”   CTRS. FOR MEDICARE & MEDICAID SERVS. (Jan. 14, 2014)], and, in April 2014, released the first public use files pursuant to this policy [“Medicare Provider Utilization and Payment Data: Physician and Other Supplier,”  CTRS. FOR MEDICARE & MEDICAID SERVS. (last updated Apr. 23, 2014)].

CMS has similarly modified its policy with respect to the release of Part D data. In 2012, CMS responded to a Freedom of Information Act request from ProPublica, a nonprofit investigative journalism organization, with Part D data that included aggregated prescribing information for individual physicians and other providers. As we discussed in a special report (“CMS Proposes to Expand Access to Part D Data for Research”), CMS finalized rulemaking last year that expands access by certain researchers to patient-level prescription drug event data with unencrypted prescriber, plan, and pharmacy identifiers [CTRS. FOR MEDICARE & MEDICAID SERVS., Medicare Program; Contract Year 2015 Policy and Technical Changes to Medicare Advantage and the Medicare Prescription Drug Benefit Programs Final Rule, 79 Fed. Reg. 29844, 29906 (May 23, 2014)]. During its rulemaking, CMS concluded that prescriber, plan, and pharmacy identifiers related to a prescription drug event (PDE) record are not commercially sensitive information in the view of the agency, and cited its public policy rationale for expanding access to PDE data for its decision to make these identifiers more readily available in unencrypted format as permitted under the new regulation (79 Fed. Reg. at 29908-9).

As CMS continues to shift from the role of a passive payer to a value-based purchaser, there are various ways that the agency might seek to leverage the public release of the Data Set and the potential insights from analysis of the Data Set and other Medicare data that CMS may release in the future. In particular, and as CMS emphasized at the time of release, publishing the Data Set is designed to contribute to an enhanced sense of transparency and accountability within the health care system. HHS may seek to use insights derived from the Data Set regarding prescribing patterns to support fraud- and abuse-enforcement activities as well as intensify collaborative efforts such as the public-private Healthcare Fraud Prevention Partnership. Findings from the Data Set also may enable CMS to refine how it measures the quality and cost-effectiveness of physicians and other health care providers. Analyses of the data may be used to support advocacy for integrating accountability for Medicare Part D in Accountable Care Organizations and Part D Plans, demonstrating both the role pharmacy services can play in generating cost savings and improving quality of care and opportunities for appropriate, cost-sensitive prescribing.

CMS’ decision to release this data, however, is not without controversy and concern.  Stakeholders within the pharmaceutical sector have raised concerns that the Data Set is misleading and misrepresents Medicare Part D spending because, as noted above, the data does not reflect manufacturer rebates and otherwise lacks necessary context. In that regard, industry stakeholders assert that the data incorrectly quantifies the cost of Part D covered drugs to the taxpayer and fails to demonstrate how these drugs contribute to lower Medicare spending by improving outcomes, helping patients avoid costly surgical interventions and reducing physician visits. Some providers have also expressed concern that the Data Set does not reflect nuances such as the effect of a physician’s patient population on drug costs or the necessity of prescribing branded drugs when there is no generic substitute.