439 - Translating Caregiver-Identified Neighborhood Factors into Quantitative Variables for Statistical Analyses of Effects on Child Health

Félice Lê-Scherban, Drexel University Dornsife School of Public Health, Philadelphia, PA, United States; Maggie Beverly, Drexel University, Philadelphia, PA, United States; Kelly Courts, Drexel University College of Medicine, Collegeville, PA, United States; Natalie M. Rodrigues, Drexel Urban Health Collaborative, Philadelphia, PA, United States; Matthew Jannetti, Drexel Univsity School of Public Health, Philadelphia, PA, United States; Zachary H. Fusfeld, Drexel University, Dornsife School of Public Health, Philadelphia, PA, United States; Jayla Z. Norman, Drexel University, Dornsife School of Public Health, Philadelpihia, PA, United States; Jana A. Hirsch, Drexel University Dornsife School of Public Health, Philadelphia, PA, United States; Hans B. Kersten, St. Christopher's Hospital for Children, Philadelphia, PA, United States; Steven J. Melly, Drexel University, Philadelphia, PA, United States

Background: Food insecurity is a persistent problem in the US. Neighborhood conditions may be a promising target for interventions to mitigate the effects of food insecurity on child health but doing so requires identifying salient neighborhood features.

Objective: To identify neighborhood factors caregivers of young children find important for their families’ food choices, and translate these into quantitative variables to incorporate into statistical analyses of food insecurity and child growth.

Design/Methods: For this mixed-methods analysis, we conducted 9 focus groups in English and Spanish among caregivers (n=51) of young children aged < 5 years in Philadelphia, PA, to identify factors salient for their families’ food choices. Focus group transcripts were coded using a start-list method. We then used an iterative process with input from a scientific and community advisory board to translate caregiver-identified concepts into quantitative, area-level variables for linkage with pediatric patient addresses from medical records.

Results: Concepts identified by caregivers included neighborhood food retail options, transportation and logistics, financial considerations, safety concerns, social support, and structural determinants such as economic disinvestment. Advisory board input informed refinements such as defining relevant walking and driving distances, categorizing types of food stores, and abandoning area-level variables that did not adequately correspond to the concepts identified in the focus groups (e.g., multigenerational families as a proxy for proximity to family members). Sources of quantitative variables included the National Establishments Time Series database (e.g., corner stores within 400 m), American Community Survey (e.g., residential segregation, households with vehicle access), and prior population-based surveys in the study area (e.g., perceptions of safety). Approximately 450 variables were identified and are being linked with 78,512 addresses corresponding to 63,670 children living in the Philadelphia metro area 2007-2023 (Table 1).

Conclusion(s): Caregivers of young children in an urban metro area cite multiple features of their neighborhoods as salient when making food choices for their families. Incorporating caregiver-identified neighborhood factors into the identification of quantitative variables may increase the robustness and utility of statistical analyses of neighborhood impacts on child health.

Table 1. Example neighborhood features identified by caregivers, corresponding quantitative variables, and data sources
Q2Q PAS abstract 2024-11-01 TABLE.pdf