Geographic Variation in Missing Race and Ethnicity Data in Minimum Data Set 3.0

Area of Science:

• Gerontology
• Health Services Research
• Data Science

Background:

• Administrative data on race and ethnicity in US nursing homes exhibit geographic variability.
• The extent of missing race and ethnicity data in nursing home administrative records is not well-characterized.

Purpose of the Study:

• To analyze geographic variations in missing race and ethnicity data within the Minimum Data Set (MDS) 3.0.
• To compare discrepancies in race and ethnicity data between MDS 3.0 and Medicare claims.

Main Methods:

• A cross-sectional study analyzed Medicare beneficiaries with MDS 3.0 records from 2014-2018.
• Missingness of race and ethnicity data in MDS was assessed by state.
• Misclassification rates of race and ethnicity variables from Medicare enrollment database (EDB) and Research Triangle Institute (RTI) were compared against MDS data.

Main Results:

• Significant geographic variation in missing MDS race and ethnicity data was observed, ranging from 1.2% to 14.7%.
• Misclassification rates varied substantially between MDS and Medicare data sources (EDB and RTI) for Hispanic, Asian American/Pacific Islander, and Black populations.
• RTI variables demonstrated superior sensitivity and specificity compared to EDB when contrasted with MDS data.

Conclusions:

• Geographic disparities in missing race and ethnicity data are evident in US nursing homes' MDS.
• Discrepancies between MDS and Medicare data sources (EDB, RTI) also vary geographically.
• Careful consideration of these data quality issues is crucial when utilizing MDS race and ethnicity information.