The Centers for Medicare & Medicaid Services (CMS) continues to evaluate our research data request and access policies, processes, and tools to ensure we are maximizing protections for controlled unclassified information (e.g., personally identifiable information and protected health information) while also enabling important research using CMS data. To further inform our next steps, CMS is seeking additional feedback from the research community. We are also releasing a summary of researcher feedback we received from CMS’ Request for Information released in February 2024.
Please review the questions below and submit responses to VRDCfeedback@cms.hhs.gov by April 15, 2025.
As noted in previous updates, CMS will not make significant changes to the research data request program until at least 2026. CMS will share guidance on these changes at a future date to allow for researchers to have adequate time for planning.
Background
In February 2024, CMS announced planned changes to our research data access policy to help protect sensitive beneficiary information. These changes included transitioning most research studies into the Chronic Conditions Warehouse Virtual Research Data Center (CCW VRDC) and changes to fees for physical data to cover growing costs associated with physical data provisioning. CMS released a Request for Information to gather feedback from researchers on these potential changes. Based on researcher feedback, CMS announced that policy changes related to the required use of the CCW VRDC and the structure of physical data fees would not go into effect until at least 2026.
CMS received close to 300 responses from a wide range of academic and non-academic researchers, as well as State and Federal agencies. The initial feedback we received expressed concerns about the impact of the potential policies on ongoing and future research. The three broad themes of the responses were CCW VRDC technical concerns, data access costs, and the timeline related to moving projects into the CCW VRDC. We summarize the feedback under each of these themes below.
Additional Feedback Opportunity
To help us determine next steps, we have identified additional questions for the research community. Your insights will help us ensure we continue safeguarding controlled unclassified information while meeting the needs of researchers. To ensure your feedback is specific and actionable, please respond to the questions listed below, addressing each of the questions and sub-questions separately. CMS does not expect you to provide feedback on all questions; you should focus on questions that are most relevant to you.
The comment period is open until April 15, 2025. Please provide feedback by sending your responses to VRDCfeedback@cms.hhs.gov. Your email should state ‘CMS Data Request Feedback’ in the subject line with responses provided as in-line text or in an attached Word document.
Feedback Summary and New Questions
CCW VRDC Technical Concerns
The technical feedback was the broadest category of responses. Respondents currently receiving physical data shipments expressed concern about understanding which CCW VRDC options best meet their needs, as well as the burden of learning a new system. Others shared concerns about the CCW VRDC environment such as: limitations on available programming languages, tools, libraries, etc.; suboptimal user experience; and potential challenges for advanced analytic techniques such as machine learning and microsimulation modeling. Many researchers highlighted that their projects utilize proprietary datasets that they do not currently have permission to upload into the CCW VRDC—such as electronic health records or Census records. Respondents felt that requiring a move to the CCW VRDC environment could curtail their research or add significant logistical barriers. Finally, respondents indicated that many institutions have already heavily invested funds, time, and effort in establishing their own specialized data environments, security infrastructure, and uniquely curated linked data assets. The respondents expressed concern that the new policies could undermine the value of that investment and require researchers to use an environment less tailored to their needs.
CCW VRDC Technical Follow-up Questions:
- Connecting to the CCW VRDC currently requires a supported version of Windows 10 or newer Windows operating system. Do you have a different operating system (e.g., macOS 15) on the computer you use to conduct research?
- Does your current research utilize any additional data source(s) other than CMS data? If yes:
- Are the data individually identifiable?
- Are the data proprietary and/or subject to other restrictions that impact moving the data into the CCW VRDC?
- What is the approximate volume (in GB or TB, as appropriate) of the additional data? Please separate totals by type (e.g., PII, proprietary) as appropriate.
- What file format(s) are the additional data files, and would you be compressing them prior to uploading to the CCW VRDC?
- Researchers have requested additional analytic tools in the CCW VRDC. Please help us refine the list of options by providing a ranked list of tools (up to 5) that you’d like to see available. See list of current tools available here https://www2.ccwdata.org/web/guest/about-vrdc.
- Researchers have requested additional programming languages in the CCW VRDC. Please help us refine the list of potential options by providing a ranked list of languages (up to 5) that you’d like to see available.
- Are there other additions or improvements to the CCW VRDC beyond new analytic tools and programming languages that you’d like CMS to explore?
- Are there capabilities—currently available or not—that you would like to have as a la carte purchasing options (e.g., paying for additional compute resources)? Currently researchers can purchase additional Databricks credits, additional storage space, additional output reviews, or access to an analytic container.
- If CMS were to explore moving away from SAS, how do you envision this would impact your work?
- What code conversion support or training would ease this transition for you?
- What CCW VRDC analytic tool(s) or development platforms would you use if SAS were not available?
- Would you need assistance converting some or all of your SAS datasets to a different format? If possible, please provide a rough estimate of the file volume and average file size.
- How much time do you estimate you would need to prepare for this change?
- What other factors should we consider as we investigate this topic?
- Are there any existing resources (user guides, technical guides, data dictionaries, etc.) that you believe contain outdated, incorrect, or ambiguous information? Please include the specific name(s) of the resource and the section(s) you believe require updating, where possible.
- Are there any topics for which you cannot find an existing resource (i.e., topics for which we should create new documentation)?
- What were the most complicated things for you to learn when you first started using the CCW VRDC? If we were to create a new “Getting Started in the VRDC” resource, what information do you think would be essential to include?
- Are you aware of the VRDC Order Application and have you used it? If yes, did you have any issues or do you have suggestions for improving the tool?
- Researchers have commented on the data architecture within the CCW VRDC. Do you have concerns about the way the CCW VRDC stores data? If yes, please provide clear and detailed feedback about any specific issues as well as what changes, if any, you believe we should consider.
- Please review the available information about the current seat options: https://www2.ccwdata.org/web/guest/about-vrdc. Do you feel you have sufficient information to decide the seat type that would best meet your needs? (Yes / No) If no, what would make the decision easier? Do you need additional information about the features or capabilities? Is the language too technical? Please be as specific as possible.
- Please review the available information about the CCW VRDC: https://www2.ccwdata.org/web/guest/about-vrdc. Is there other information that would be valuable for you to prepare for a transition of your project to the CCW VRDC?
Data Access Costs Concerns
Respondents expressed concern about the proposed changes to the pricing structure of physical delivery of research files as well as the current pricing of the CCW VRDC. We received a significant amount of feedback indicating that the cost of CCW VRDC access would reduce the number of researchers who could afford seats, which in turn hinders a team-based approach to research, reduces code-checking, and more. Researchers particularly highlighted the impact of data access costs on students, trainees, junior staff, and researchers from marginalized communities.
Data Access Costs Follow-up Questions:
- CMS charges fees to researchers to recoup the cost of making data available for research purposes[1]. These fees are based on CMS’s costs and allow the agency to continue to make data available to researchers. Historically, CMS has infrequently reevaluated our pricing, leading to large pricing increases when changes are made. Would you rather: a) continue these larger, infrequent cost increases on a multiyear timeline, or b) switch to smaller annual increases?
- If CMS were to explore the creation of new CCW VRDC seat options, what would you need included in a lower cost, entry-level seat? Assume CMS would also offer alternative seat types which include all the existing tools and capabilities.
- What features would be necessary for you in this type of seat?
- Are there any features that would prevent your ability to use this seat type?
- Is there a price point at which this would not be a feasible option for your use?
- Would the development of a “viewer” seat in the CCW VRDC be beneficial for you or someone on your team? We envision this would be a very low-cost seat (non-transferrable) which could have access to CMS data under multiple data use agreements (DUAs) and primarily be used by those who supervise student work or oversee research teams.
- What features would be necessary for you in this type of seat?
- Are there any features that would prevent your ability to use this seat type?
- Is there a price point at which this would not be a feasible option for your use?
- While we understand that a limited data set (LDS) may not support all research needs, we have received support for expanding the range of lower-cost LDS files we make available for research. Please help us refine the list of options by providing a ranked list of Research Identifiable File datasets that you’d like to see available as LDS files.
Timeline Concerns
Respondents expressed concern about the implementation dates listed in the initial policy change announcement and felt that the amount of notice was insufficient for a change of this magnitude. For example, a number of researchers noted that they required more time to factor price changes into their budgets when applying for grant funding. Additionally, the feedback we received highlighted that if researchers are required to move ongoing projects into the CCW VRDC, many would have to duplicate work such as data cleaning which could negatively impact meeting grant timelines. Researchers also expressed concern that the technical changes required to use the CCW VRDC environment could not be implemented by CMS within the originally proposed timeline. Finally, researchers shared that the existing wait times and the high administrative burden associated with DUA approvals (initial requests, renewals, amendments) and CCW VRDC approvals (initial access, seat renewal, other actions) were already impacting project timelines, and expressed concern that the potential policy changes may further strain these processes.
Timeline Follow-up Questions:
- What is your average grant funding timeline, from initial proposal development through funding decision and award of funds? Please provide a timeline for each phase, if possible.
- When developing your proposal budget do you typically use the Estimate Study Size or the Data Pricing tools found on the CCW website (https://www2.ccwdata.org/web/guest/pricing)? If not, please describe the method you typically use when estimating CMS data costs for your grant proposal.
[1] See section 1106(B) of the Social Security Act