185 - Moving Beyond the Child Opportunity Index Quintiles; Exploring Neighborhood-Level Determinants of Pediatric Hospital Length of Stay

Publication Number: 1174.185

Temitope Osineye, Pediatric Physicians' Organization at Children's, Wellesley Hills, MA, United States; Patrice Melvin, Boston Children's Hospital, Boston, MA, United States; Melicia Y. Whitley, Boston Children's Hospital, Boston, MA, United States; Stephanie Tuttle, Boston Children's Hospital, Boston, MA, United States; Mia Umali, Boston Children's Hospital, Cambridge, MA, United States; Emily T. Correa, Boston Children's Hospital, Wellesley, MA, United States; Ravi R. Thiagarajan, Boston Children's Hospital, Chestnut Hill, MA, United States; Michelle L. Lock, Pediatric Physicians' Organization at Children's, Wellesley Hills, MA, United States; Valerie L. Ward, Boston Children's Hospital, Boston, MA, United States

Background: The Child Opportunity Index (COI) is a validated multidimensional composite of indicators that measure resources important for healthy child development based on their neighborhood. Little data exists on the neighborhood factors in COI subdomains of education, health/environment and socioeconomics with hospital length of stay (LOS), and these data are needed given that prolonged LOS can lead to increased risks of hospital acquired infections, safety events such as falls, and increased hospital costs.

Objective: To identify education, health and environment, and socioeconomic factors that may be associated with longer hospital LOS for children.

Design/Methods: A retrospective cohort study of hospital admissions (2021-2024) in a pediatric academic medical center for patients 0-18 years whose primary care provider was part of an affiliated statewide network of independent pediatric primary care practices. Primary outcome was hospital LOS, and primary exposures were 14 subdomains of COI. Clinical characteristics during admission were obtained via the Pediatric Health Information System (PHIS) database. Subdomain COI scores and levels were accessed by merging patient zip codes from electronic health records (Epic), with a publicly available COI v3.0 dataset. Neighborhoods were stratified into COI quintiles: 'Very Low' to 'Very High' opportunity, based on zip code. Strength of association between COI subdomain scores and hospital LOS was determined by Spearman correlation coefficient analysis. We assessed associations between COI subdomain levels and average LOS using one-way analysis of variance (ANOVA). Multivariable mixed effects linear regression models will be used to assess the association of COI subdomains with LOS, considering clinical factors and patient level random effects.

Results: In 16,706 patients (Asian 4%, Black 6%, Hispanic 10%, Multiracial 3%, Other 5%, White 60%, Unknown 12%), 37% had public insurance, 11% lived in the very low opportunity quintile, 44% had complex/chronic conditions while 12% were admitted in the intensive care unit. Mean age was 7.1 years and mean LOS was 3.1 days. Lower opportunity neighborhoods had longer average LOS. COI subdomains with the strongest correlation (all inversely correlated) with LOS included metrics related to employment, early childhood education, and social resources.

Conclusion(s): Variations exist in the association of COI subdomains with LOS in our study. Identifying the most impactful neighborhood-level factors can help prioritize interventions to reduce hospital LOS and improve health outcomes for all children.

Population Description
Population Description

Correlation of COI Subdomain Scores With LOS
Correlation of Child Opportunity Index Subdomain Scores With Hospital Length of Stay

Average LOS by COI Quintiles and COI Subdomains
Average Hospital Length of Stay by Child Opportunity Index Quintiles and Subdomains