Abstract P033: Identifying Racial/Ethnic Differences In Patterns Of Early Childhood Cardiovascular Health Using Electronic Health Record Data

Introduction: Racial/ethnic differences in CVH beginning at age 8 have been identified and linked with the development of cardiometabolic disease in adulthood; however, there is scarce research on CVH in very childhood. Our objective was to use a large, diverse pediatric EHR consortium to identify racial/ethnic patterns of clinical CVH from ages 2-12 years. Methods: We used ambulatory visit data spanning 2010-2018 from children aged 2-12 years within CAPriCORN - an EHR repository that combines medical records throughout the city of Chicago. The 4 clinical CVH metrics - BMI, blood pressure, cholesterol, and glucose - were categorized as ideal or non-ideal using available values of weight, height, blood pressure, laboratory readings, and ICD diagnosis codes. Multiple measurements within a given integer age were averaged by individual. Frequency of ideal and non-ideal status for each CVH metric was plotted by age in years and stratified by race/ethnicity (Figure). Results: There were 162,621 children included (47% female) with a median of 2 visits during follow-up. The race/ethnicity distribution was 4% Asian/Pacific Islander (API), 26% non-Hispanic Black (NHB), 44% non-Hispanic white (NHW), 18% Hispanic, and 8% other/unknown. Sustained decrease in ideal BMI occurred across race/ethnicity groups; however, proportion in ideal was consistently lower for NHB and Hispanic children. Ideal BP appeared to increase across childhood with few racial differences. Ideal cholesterol levels were constant across childhood, but the proportion of NHW children in ideal was lower than NHB and Hispanic children. Almost all individuals had ideal glucose levels throughout early childhood. Conclusions: Early childhood declines in CVH appeared to be driven by changes in ideal BMI. Racial/ethnic differences in ideal BMI and cholesterol were present by age 2 and were mostly sustained through age 12. Selection bias may account for some of these findings; consistent monitoring in early childhood is needed to better understand observed differences.