335 - Associations between the Child Opportunity Index and pediatric mortality: A meta-analysis

Poster Number: 335
Publication Number: 335.2065

Background: Studies using diverse methodologies show that children experiencing adverse social determinants of health have disproportionately higher morbidity. With the increase of studies examining these relationships with the Child Opportunity Index ([COI], a composite measure of 29 place-based factors), there is an opportunity to synthesize associations between place and severe health outcomes among children.

Objective: To meta-analyze associations between the COI and pediatric mortality.

Design/Methods: Major databases, grey literature, and abstracts were searched from 2014 (COI inception) to Feb 2023. References of included full text articles (and studies citing them) were also searched. Observational studies in the United States evaluating 1) COI quintiles (very low, low, moderate, high, and very high) (exposure) and 2) pediatric mortality (outcome) were included. Excluded studies evaluated only adult populations or the continuous COI. Study data were collapsed into two groups: children with very low and low (VL/L) COI versus children with high and very high (H/VH) COI (reference). Using metafor (R), random-effects models created pooled odds ratios (OR) for pediatric mortality. To explore heterogeneity, analysis of the pooled OR was stratified by select variables to identify moderators: data source, diagnosis type, population scope (e.g. national v. regional), mortality time (during v. post-hospitalization), geographic unit of analysis, and restriction of population to specific diagnosis.

Results: The initial search yielded 183 unique studies: 9 articles were identified from full text articles. Seven studies met inclusion criteria with usable outcome data (Figure 1). All included studies were published in 2022-23 and examined in-hospital mortality. Two studies also evaluated post-hospitalization mortality. Five studies (71%) evaluated children with a surgical/trauma diagnosis. Children with VL/L COI had higher odds of mortality compared to children with H/VH COI, though heterogeneity was high (OR 1.59, 95% CI 1.31 1.94; I2 87%; Figure 2). No variables were significant moderators, however there were certain categories within the data source, diagnosis type, and population scope variables in which mortality was significantly higher among children with VL/L COI (Table 1). Heterogeneity was most reduced (I2 65%) by adding data source and diagnosis type to the model.

Conclusion(s): Experiencing lower COI is a significant risk factor for mortality among children. Collection and critical assessment of factors related to a child’s neighborhood should be routinely considered a key element of pediatric clinical care.