Standardizing race, ethnicity, and preferred language data collection in hospital information systems: results and implications for healthcare delivery and policy

The Institute of Medicine has identified the need for healthcare organizations to collect standardized demographic data as a step toward reducing healthcare disparities. This observational study of patients discharged from a large academic medical center between 2005 and 2009, evaluates an organizational effort to standardize demographic data collection, characterizes limitations of the implementation, and assesses its utility in quality improvement and disparity reduction efforts. Primary measures include percentages of inpatient discharges with unknown race, ethnicity, and language data. Secondary measures include "ideal" cardiovascular care and readmission rates. From 2005 to 2009, the proportion of discharges with unknown race, ethnicity, or preferred language data decreased significantly. Among discharges with known ethnicity in 2009, Hispanic/Latino patients were significantly more likely to decline to specify their race or designate their race as a "Multiracial: Other Combination." No significant differences in ideal cardiovascular care were observed across demographic groups. Differences in readmission rates were observed among some groups. A provider organization can effectively standardize demographic data collection practices for use in quality improvement efforts. Current federal race categories are of limited utility for persons of Hispanic/Latino ethnicity, and performance measurement approaches that exclude demographic variables may fail to address healthcare disparities.