Residential Economic and Racial Segregation Is Associated with Acuity and Severity of Pediatric Leukemia at Presentation

Introduction: Residential economic and racial segregation, based on geographic differences in social deprivation of resources and racial composition, has been linked to inferior outcomes in adult-onset cancers. However, the impact of residential economic and racial segregation on pediatric cancer outcomes remains understudied. This study aims to determine the association of such segregation with disease acuity and severity at initial presentation in a statewide population-based cohort of children with leukemia, the most common pediatric cancer.

Methods: We analyzed a cohort of children (ages 1-21 years) who presented at a large pediatric healthcare system in the southern U.S. between 2010-2018 with a primary diagnosis of leukemia (B-ALL, T-ALL, AML, or other types of leukemia including JMML, CML, and Burkitt’s). Outcomes included 1) high acuity (yes/no), defined as any intensive care unit (ICU) admission or ICU-level resource utilization (cardiovascular, respiratory, renal, hematologic, neurologic) within the first 72 hours of presentation, irrespective of physical location in the hospital; and 2) high severity (yes/no), defined as a white blood cell count ≥50,000 cells/microliter and/or central nervous system involvement at presentation.

To measure residential economic and racial segregation, we mapped the Index of Concentration at the Extremes (ICE) to the ZIP Code Tabulation Area (ZCTA) of patient residence at diagnosis. ICE was calculated as the difference between the number of non-Hispanic White people with at least 80th percentile household income and non-Hispanic Black people with less than 20th percentile household income, divided by the total number of people in the ZCTA. ICE values were then divided into three tertiles, with tertile 1 representing the most deprived/segregated areas and tertile 3 representing the most privileged areas. Sample characteristics were compared by ICE tertiles. Logistic regressions estimated outcome differences by ICE tertiles, with and without adjustment for demographic (age at diagnosis, sex) and clinical (leukemia subtype) factors.

Results: Of the 688 patients identified, the proportions with non-Hispanic Black patients (45% vs. 10%) and publicly insured patients (73% vs. 34%) were notably higher in ICE tertile 1 compared to tertile 3 (p-values<0.001). Patients in tertile 1 were more likely than those in tertile 3 to present with high acuity in both bivariate comparison (30.1% vs. 19.7%; unadjusted odds ratio [OR]=1.76, 95% CI=1.15-2.73, p=0.010) and adjusted model (adjusted OR=1.67, 95% CI=1.06-2.65, p=0.028). Similarly, patients in tertile 1 were more likely than those in tertile 3 to present with high severity in bivariate comparison (45.0% vs. 34.1%; unadjusted OR=1.58, 95% CI=1.09-2.31, p=0.017) and adjusted model (adjusted OR=1.44, 95% CI=0.97-2.15, p=0.072).

Conclusion: These findings suggest that living in racially segregated and economically deprived neighborhoods is associated with high acuity and high severity at initial presentation of leukemia among children. Further research is warranted to understand the underlying mechanisms through which residential segregation drives disparities in frontline presentation, which impact downstream outcomes in pediatric leukemia.