AUTHORS
Callie Fry, Morgan A. Paul, MB2, John A. Kairalla, PhD3, Leyna Xoi, MS4, Leanne Duhaney, MD, MPH5, Kristine Karvonen, MD, MS6, Colleen A. Kelly, MD7, Haley Newman, MD, MSCE8, Puja J. Umaretiya, MD, MS9, Daniel J. Zheng, MD, MHS, MSHP10, Peter D. Cole, MD11, Iris Paltin, PhD12, Elizabeth A. Raetz, MD13, Mignon L. Loh, MD14, Sarah Alexander, MD15, David T. Teachey, MD16, Sumit Gupta, MD, PhD, FRCPC17, Rachel E. Rau, MD18, Kira Bona, MD, MPH19, Rahela Aziz-Bose, MD, MPH20
AFFILIATIONS
1 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
2 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
3 University of Florida, Gainesville, Florida, United States
4 University of Florida, Gainesville, Florida, United States
5 Boston Children's Hospital, Boston, Massachusetts, United States
6 Fred Hutchinson Cancer Center, Seattle, Washington, United States
7 Boston Children's Hospital, Boston, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
8 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
9 University of Texas Southwestern Medical Center, Dallas, Texas, United States
10 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
11 Rutgers Cancer Institute, New Brunswick, New Jersey, United States
12 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
13 New York University Langone Health, New York, New York, United States
14 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
15 British Columbia Children's Hospital, Vancouver, British Columbia, Canada
16 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
17 Hospital for Sick Children, Toronto, Ontario, Canada
18 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
19 Boston Children's Hospital, Boston, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
20 Boston Children's Hospital, Boston, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
ABSTRACT
Introduction: Children’s Oncology Group (COG) trials have utilized household- and neighborhood-level poverty proxies to identify stark disparities in relapse and survival across pediatric cancers. The frontline trial AALL1731 (NCT03914625) prospectively collected a direct parent-reported poverty measure—household material hardship (HMH: food, housing, utility, or transportation insecurity). We aimed to evaluate whether HMH identifies the same cohort of children as historic poverty proxies (insurance and ZIP code), and to explore whether poverty measure selection influences associations with salient outcomes, utilizing parent anxiety as an exemplar parent-reported outcome (PRO).
Methods: This secondary analysis used prospectively collected data from the randomized, phase III COG trial AALL1731, which enrolled children 1-9 years with newly diagnosed B-cell acute lymphoblastic leukemia (ALL) across 217 US and Canadian sites (2019-2024). English-, Spanish-, or French-speaking families opted-in to the embedded HMH study and completed baseline surveys during the first cycle of chemotherapy. The analytic cohort was restricted to US participants with complete baseline data on HMH, ZIP code, insurance, and Generalized Anxiety Disorder-7 (GAD-7).
Poverty was defined as parent-reported HMH (≥1 resource insecurity) and trial-collected proxies at the household- (public insurance: sole Medicaid or Children’s Health Insurance Program insurance) and neighborhood-levels (high-poverty ZIP code: ≥20% of persons at or below the federal poverty level [FPL]). The primary outcome was the proportion of HMH-exposed children who would not have been identified as poverty-exposed by proxy measures. We subsequently explored the association of each measure with parent moderate-to-severe anxiety (GAD-7 ≥10) at diagnosis. Covariates included child sex, age, race/ethnicity, language, parent education, household size, household structure, and income (<200% vs ≥200% FPL).
Descriptive statistics characterized the frequency of each poverty exposure. Univariate logistic regression models evaluated associations between each poverty measure and parent anxiety. A multivariable logistic regression model examined the association between HMH and parent anxiety, adjusting for covariates significantly associated with parent anxiety on univariate analysis. Statistical significance was set at α=0.05 (two-sided).
Results: The analytic cohort included 945 children. Thirty-three percent were poverty-exposed by HMH (n=308), 39.6% by public insurance (n=374), and 13.9% by high-poverty ZIP code (n=131). Among HMH-exposed children, 35.4% (n=109) were classified as non-poverty-exposed by insurance, and 81.8% (n=252) by ZIP code.
Moderate-to-severe anxiety was reported by 35.0% (n=331) of parents. In univariate analyses, HMH was associated with moderate-to-severe parent anxiety (odds ratio [OR] 1.55, 95% CI: 1.17-2.06, p=0.002), while public insurance (OR: 1.00, 95% CI: 0.76-1.31, p=1.0) and high-poverty ZIP code (OR: 0.67, 95% CI: 0.44-1.00, p=0.052) were not. In multivariable analysis adjusting for child race/ethnicity and household structure, HMH remained independently associated with moderate-to-severe parent anxiety (OR 1.62, 95% CI: 1.18-2.24, p=0.003).
Conclusion: Relying on historic poverty proxy measures alone would miss a large proportion of children with parent-reported HMH in a national pediatric ALL trial—35% of HMH-exposed children would have been classified as non-poverty-exposed by insurance, and 82% by ZIP code. Only HMH was associated with parental anxiety—an exemplar PRO—while proxy measures did not demonstrate an association with this outcome. Poverty is a multidimensional construct without a single “gold standard” definition; relying on proxy measures alone will miss a large proportion of children exposed to HMH and its salient sequelae. Future analyses will evaluate associations of each poverty measure with disease and late-effect outcomes. These data underscore the value of embedding direct parent-reported poverty measures in future trial design to comprehensively identify populations at risk of disparate outcomes and enable targeted intervention.
Acknowledgements: COG AALL1731 was supported by funding from the NCTN Operations Center Grant (U10CA180886), NCTN Statistics & Data Center Grant (U10CA180899), St. Baldrick’s Foundation, and Amgen.
DISCLOSURES
Callie Fry: No relevant conflicts of interest to declare.