Errors in Racial Data Likely in Electronic Medical Records

THURSDAY, Sept. 5, 2024 (HealthDay News) — There are significant errors in recording racial data in children’s electronic medical records (EMRs) across health systems, according to a study published online Sept. 3 in JAMA Network Open.

Gary L. Freed, M.D., from the University of Michigan in Ann Arbor, and colleagues examined the error rate of racial and ethnic attribution in EMRs across the three largest pediatric health systems in Michigan. The analysis included comparisons between parent or guardian report of race and ethnicity for 4,333 children seen at outpatient clinics, emergency departments, and inpatient units (Sept. 1, 2023, to Jan. 31, 2024) and EMRs.

The researchers found that the greatest error rate across the health systems occurred with the exact match of parental report of racial designation with the EMR (range, 41 to 78 percent). The matching rate improved with consolidation of race options, with differences between the health systems narrowing to 79 to 88 percent matching. Across health systems, matching of ethnicity between the EMR and the parental report ranged from 65 to 95 percent. Missing race or ethnicity data in the EMR, which was counted as a nonmatch, varied across the health systems (range, 2 to 10 percent). The highest error rates were seen at the health system with the greatest number of options for race and ethnicity.

“Error rates of these magnitudes raise doubts regarding data, suggesting either the presence or absence of inequities and disparities in specific areas of clinical care, and may undermine strategies to improve care,” the authors write.

Abstract/Full Text