Race, Ethnicity, and Language Data: Standardization for Health Care Quality Improvement
F. Granular Ethnicities with No Determinate OMB Race Classification
The Subcommittee recommends collecting an individual's Hispanic ethnicity, race, and granular ethnicity. Separate questions to collect these data ensure the ability to report Office of Management and Budget (OMB)-compliant data without needing to aggregate granular ethnicity categories back to one of the six OMB categories (e.g., an individual can self-identify as Asian, non-Hispanic, and of Laotian ancestry and all of these data should be retained in a data system). Nevertheless, in some instances, granular ethnicity data may need to be rolled up to one of the OMB categories for purposes of analysis or reporting. For example, an individual may not have responded to the question on race and only responded to the question on granular ethnicity. If the individual's health plan is required to report data to the state using only the OMB race and Hispanic ethnicity categories, the health plan may want to aggregate the individual's granular ethnicity to an OMB category, whenever possible.
Aggregating data to the OMB race categories through rollup schemes, though, may inevitably contradict or misrepresent an individual's self-identification as not all Americans of South African descent are Black, for example. An individual's granular ethnicity does not automatically determine his or her race; consequently, any rollup scheme may falsely classify some individuals. Certain granular ethnicity categories are more prone to misclassification than others, primarily because several individual races as well as multiracial persons are represented within a single ethnicity.
The subcommittee identified some of these ethnicities by cross-tabulating write-in responses to Census ancestry data by the OMB single-race and Hispanic ethnicity categories. Since many of the ethnicity groups had large proportions of individuals who reported more than one race, the subcommittee then cross-tabulated the ancestry responses with "alone or in combination with one or more other races" variable for each OMB group to see if 90 percent or more in the ethnicity group reported an OMB race either alone or in combination with another race.1 Many of the granular ethnicities that fell short of a 90 percent threshold based on single-race reporting exceeded that threshold when the identification was based on reporting the race group alone or in combination with other races.
Many of the granular ethnicity categories that still could not be assigned to an OMB race category using the 90 percent threshold for responses "alone or in combination" represented populations with long histories of intermarriage and multiracial identity (e.g., Native Hawaiian or Other Pacific Islander, American Indian or Alaska Native). These granular ethnicity categories could be assigned to the OMB race category of the same name on the basis of the OMB definitions for the Native Hawaiian or Other Pacific Islander and American Indian or Alaska Native categories. However, it is important to note that this assignment misclassifies many individuals based on the OMB race with which they would self-identify given the opportunity.
Additionally, Table F-1 includes granular ethnicities that are rolled up differently by different coding schemes. For example, the Centers for Disease Control and Prevention (CDC)/Health Level 7 (HL7) Race and Ethnicity Code Set 1.0 considers Madagascan in its Asian category while the Massachusetts Superset considers Madagascan under both Asian and African category. Thus, this ethnicity is included in Table F-1 and may be said to have "no determinate OMB race classification."
The subcommittee suggests that the Department of Health and Human Services (HHS) take into account that some ethnicities do not correspond with one specific OMB race category and that when rollup is necessary, these granular ethnicities be included in a category labeled "no determinate OMB race classification."
1 The 90 percent rule used in this analysis is not the only method for identifying granular ethnicity categories that cannot or should not be rolled up to one of the OMB categories. Census 2010, for example, is, when necessary, rolling up write-in responses based on the OMB definitions of each race and Hispanic ethnicity category. Then, all sub-Saharan African ethnicities will be coded as Black, where necessary for analysis.