288 - Latino Community Validation of Race and Ethnicity Classification and Mapping in OMOP

Publication Number: 288.5679

Juan C. Espinoza, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States; Sadaf Javaid, Ann & Robert H. Lurie Children's Hospital of Chicago, des plaines, IL, United States; Jill Laeyendecker, Ann & Robert H. Lurie Children's Hospital of Chicago, Madison, WI, United States; Xaviera Vasser, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States

Background: The Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) is one of the most used CDM in healthcare research. The concepts of race and ethnicity are captured in the OMOP Person table and hold one value each, mapped to terms consistent with US federal definitions of race and ethnicity. In March 2024, OMB announced that race and ethnicity will be collapsed into a single race/ethnicity variable, and that two new options would be added: Middle Eastern or North African, and Hispanic or Latino. Mapping Latino ethnicity to Latino race will be challenging in older records that may not be updated via self-report (e.g., patients lost to follow up, no longer require care, deceased).

Objective: Examine historical EHR record at a single institution and explore potential mappings to convert Latino ethnicity to Latino race with community representatives.

Design/Methods: All unique EHR records were pulled, and the distribution and frequency of race and ethnicity were analyzed. The OMOP documentation was reviewed, and potential mappings were developed with considerations to the changes in race and ethnicity. To gather insights from the local Latino community served by our institution, we assembled focus groups of families seen in the past year who identified as Latino. Focus groups were help in Spanish and English.

Results: 1.7 million unique records had valid demographic data; 97.7% had a single race value. 78.9% of “other” race identified as Latino (Table 1). Based on our institutional data, we decided that all other, unknown, declined, and deferred race with Latino ethnicity will be mapped to Latino race. For other records, we developed 3 possible mappings: 1) prioritize Latino identity for all records; 2) prioritize Latino identify for some records; and 3) prioritize “more than one race” identity for all records (Figure 1). Focus group participants primarily favored option 1, stating the need to make sure that Latino records be “counted after sometimes being ignored.” Some participants leaned towards option 2, specifically to acknowledge groups who strongly identify with Latino and other identities, such as individuals of Afro-Caribbean descent.

Conclusion(s): Latino individuals are overrepresented among “other” race category and can be mapped to Latino race. Community members preferred prioritizing Latino identity for all mappings; a driving factor for participant’s preference is their perceived value on being distinctly counted and represented as Latino. Other organizations should consider a similar community based participatory approach to re-mapping historical records.

Table 1.
Counts of unique EHR records by race category showing the subset within each category that also identified as Hispanic or Latino Ethnicity.

Figure 1
Possible mappings for Hispanic ethnicity to Hispanic race