April 20, 2023
Bob Sivinski, Chair
Interagency Technical Working Group on Race and Ethnicity Standards
1650 17th St. NW
Washington, DC 20500
Dear Mr. Sivinski,
The National Committee for Quality Assurance (NCQA) thanks you for the opportunity to provide feedback on the Federal Interagency Technical Working Group on Race and Ethnicity Standards’ (Working Group) Initial Proposals for Updating the Office of Management and Budget’s (OMB) Race and Ethnicity Technical Standards.
NCQA is a private, 501(c)(3) not-for-profit, independent organization dedicated to improving health care quality through our Accreditation and measurement programs. We are a national leader in quality oversight and a pioneer in quality measurement. Leveraging our strengths as a trusted third party, we are committed to helping organizations navigate the challenges associated with moving to an equitable health care system. Our mission to improve the quality of health for all Americans, with a focus on health equity and support for meaningful value-based payment models, propels our daily work.
Based on our breadth of experience and technical expertise incorporating race and ethnicity data to advance high-quality, equitable health care, we are pleased to provide our comments on the initial proposals outlined by the Working Group. Our general recommendations to OMB are:
- We support expansion of race and ethnicity categories beyond the current options. In expanding the categories, OMB should center community perspectives, avoid centering the concept of “White non-Hispanic” and allow flexibility for different use cases while providing standard mechanisms and guidance to maintain consistency.
- We recommend OMB require and enforce uniform and universal adoption of the updated OMB standards upon their release in 2024, for all government agencies and all private sector health care stakeholders, including payers and providers. In doing so, OMB should provide guidance to ensure patient and community protection and promulgate guidance for expected interoperability, to avoid undue burden through repeated fielding of the same question.
- We recommend OMB create an infrastructure to acknowledge and distinguish data source, including considering whether the proposed category expansion can accommodate uncertainty and setting a standard technical threshold for evaluating the quality of imputed values.
We acknowledge that our comments are from a health care perspective specifically, and we advise the OMB to consider the context of proposed updates across federal agencies and use cases.
NCQA experience with race and ethnicity standards
For close to two decades, NCQA has engaged in equity data-centered research and quality efforts. Among other undertakings, we have:
- Incorporated race and ethnicity stratification to 13 HEDIS measures to date and solicited public comment input on expanding the stratification to additional measures. We provide standardized guidance for mapping, managing and reporting health care quality data by race and ethnicity, as well as for integrating national health IT standards. We also launched the HEDIS Race/Ethnicity Stratification Learning Network, querying specifically on race and ethnicity data collection and management, use of standardized data categories and challenges and opportunities with reporting stratified data.
- Developed our first in class Health Equity Accreditation programs: Health Equity Accreditation and Health Equity Accreditation Plus. These programs include requirements for collecting data on race, ethnicity, sexual orientation and gender identity using standardized methods.
- Provided thought leadership in health equity through research to help the industry develop and implement best practices for race and ethnicity data. For example, we proposed recommendations for health plans and government agencies to improve collection and use of race and ethnicity data to advance health equity, building on earlier work in the same area, and created a resource guide for auditors to support health plans’ improved race and ethnicity data collection.
As we evolve to incorporate equity elements into new and existing programs and research, we continue to engage with users during their development—through pilot programs, public comment, ongoing customer support and engagement with stakeholders, who inform our recommendations. Building on our overall recommendations described above, we are pleased to provide our comments on the proposals outlined by the Working Group.
- Collect Race and ethnicity information using one combined question.
Because it is critical to capture the intersection between race and ethnicity, NCQA has historically recommended using separate questions to collect this information. For example, under the current schema, respondents can select both “Hispanic” (or “non-Hispanic”) and one or more category for race; this offers the ability to examine both race and ethnicity across data. We strongly recommend that if OMB uses a combined question to collect race and ethnicity information, the response include an option to select more than one race or ethnicity. We recommend operational considerations for implementing this change:
- Provide guidance for ensuring that data are meaningful and usable, particularly when multiple selections are possible. For example, we know that some health systems have encountered too many selections to be able to categorize individuals for the purposes of data and analytics.
- Provide guidance for managing data that come out of a combined-questions structure, including how data are rolled up when multiple race and ethnicity categories are selected. For example, during analytic processes, an individual who selects multiple categories (e.g., Asian Black Hispanic) could be included in each individual category, a single “higher level” category (e.g., Two or more…), or in a unique intersecting category (Asian-Black-Hispanic).
Each of these approaches may be needed, depending on the use case; for instance, some evaluations require total populations to sum to 100%. Guidance on how to execute such categorizations consistently will support common analytic approaches and interpretations.
- Add “Middle Eastern or North African” (MENA) as a new minimum category.
NCQA supports the addition of “Middle Eastern or North African” (MENA) as a new minimum category. We have routinely heard—through stakeholder feedback and public comment from plans, health systems and community members—a call for more granular categories for race and ethnicity than the current OMB standards. Many people of MENA descent do not consider themselves to be White, nor are they perceived as White by others. Including MENA as a new minimum category will align more closely with the diversity of experiences in this group and would represent an important step in validating community perspectives that have long been ignored.
- Require the collection of detailed race and ethnicity categories by default.
NCQA supports the proposal to require collection of detailed race and ethnicity categories by default. We routinely hear—in our calls for public comment as well as in conversations with stakeholders (including health plans, health systems and community members)—that more detailed categories should be collected in order to accurately reflect patient identity. Our stakeholders have expressed, for example, the need for granularity in the “Asian” category because the category’s broad nature misrepresents both community identity and potential disparities in outcomes. With regard to the “Black or African American” category, we recommend expanding examples so a person with ancestry from the original Black racial groups of Africa can select this option; for example, by adding Fiji, Vanuatu, Papua New Guinea and the Solomon Islands. This expansion would include Black populations in Melanesia and the South Pacific whose heritage stems, at least in part, from Africa.
NCQA also recommends that OMB provide a standard mechanism for expanding and contracting race and ethnicity categories for different use cases across organizations. Organizations need the flexibility to support different levels of categorical collection to inform a variety of use cases. Stakeholders have expressed a need for processes that map or roll up from more detailed categories to those collected by the OMB, in order to support analyses across programs. We recommend processes that allow flexible collection, since different categorical approaches may be required even within the same organization. For example, health intervention planning may require a health plan to leverage more detailed values to ensure culturally competent and community-informed care in a geographic region, whereas population trending may require analysis of fewer categories to describe national trends.
We strongly recommend that significant cognitive testing—with a variety of represented groups and perspectives—be conducted for the question stem and the instructions for collecting more detailed race and ethnicity data. The instructions, number of choices, free text options and other aspects of the question could be confusing. Collecting a lot of information in a single question might lead to conflicting or unreliable data if instructions are not clearly understood.
- Update Terminology to SPD 15.
Terminologies Used Within Minimum Categories
NCQA supports the proposed updates to the terminologies used within the minimum categories, including any terminology updates that steer away from use of harmful, pejorative language, and promotes use of accurate, community-centered terms. For example:
NCQA supports discontinuing the use of the terms “majority” and “minority,” and believes that all category changes should center community perspectives and avoid focusing on “White, non-Hispanic” as a default reference group, so as not to imply that one group is superior to another. We also recommend that alternative descriptors such as “marginalized” or “minoritized” reflect the reality that these groups have been targeted and oppressed through inherently racist, biased systems. Updating the terminology will increase accurate identification while avoiding terms that evoke bias.
Question Stem and Instructions
NCQA recommends significant cognitive testing of the question stem and instructions. In particular, for the proposal to combine race and ethnicity into one question, the use of “or” may suggest that respondents cannot select both a race and an ethnicity. Testing the question stems and instructions with a variety of audiences and perspectives will inform language that most clearly conveys the question’s intent, and will result in accurate data.
- Guidance is necessary to implement SPD 15 revisions on Federal information collections.
Additional guidance needs not addressed in initial proposals
NCQA recommends that OMB create a standard expectation and implementation approach for collecting and stratifying race and ethnicity data across all applicable federal programs. Execution of the SPD 15 revisions requires effective cross-system research, program implementation and evaluation, which aligns with the White House Executive Order on Further Advancing Racial Equity and Support for Underserved Communities Through the Federal Government (Executive Order 14091). Given its influence, the OMB is uniquely positioned to be the catalyst, guide and accountability partner for ensuring that all federal programs align in race and ethnicity information requested from data providers, as well as expected attributes of the data.
Guidance on data collection methods (including self-identification)
Self-reporting is the gold standard, and national standards should advance efforts to collect and center race and ethnicity data. We have received many comments voicing concern about the use of imputed data and the implicit uncertainty, but we know self-identification is not always possible; we recommend that OMB consider whether the proposed changes to race and ethnicity categories can accommodate shifting values and imputation probabilities.
NCQA recommends the OMB consider the implications of revised categories for existing technical methods. For example, the ability to accurately impute missing values should be considered with proposed category approaches. There is a substantial technical infrastructure to support estimation of missing race and ethnicity data that will need to align with updated standards (including within-HHS use of imputed data in Medicare enrollment files, for example). Not accounting for this could lead to confusion, conflicting data or a spike in missing data.
We also recommend that the OMB set a standard technical threshold for evaluating the quality of imputed values. If statistical imputation for missing race and ethnicity data continues to be used for population-level analyses such as trending, public health and quality evaluation, then we need a standard technical threshold to ensure those methods meet the needs of their use cases, or we risk building bias into our systems and potentially misrepresenting the nature of health disparities or unmet needs for investment or action. These recommendations reflect feedback NCQA gathered in a 2022 roundtable of health care data experts that included plans, statisticians and policy research representatives.
- Comments On Any Additional Topics and Future Research
NCQA believes other sociodemographic data elements would benefit from a similar focused effort. We would recommend sexual orientation and gender identify as a natural next step, given the work begun by the National Academies of Sciences, Engineering, and Medicine and others on this topic.
hank you for the opportunity to comment. We remain committed to working with the federal government to build a more equitable, sustainable and responsible American health care system. We welcome a discussion on our experience and recommendations to continue to strengthen the OMB’s proposals for updating race and ethnicity categories. If you have any questions, please contact Eric Musser, Assistant Vice President of Federal Affairs, at (202) 955-3590 or at firstname.lastname@example.org.
Margaret E. O’Kane