Authors :
Presenting Author: Carson Gundlach, BS – Weill Cornell Medical College
Ashwin Mahesh, BA – Medical Student, Weill Cornell Medical College; Alexander Zhao, BS – Medical Student, Weill Cornell Medical College; Sarah Chowdhury, BA – Medical Student, Weill Cornell Medical College; Nuran Golbasi, BS – Medical Student, Weill Cornell Medical College; Lauren Sam, LCSW – Social Worker, Pediatric Social Work, Weill Cornell Medicine; Natasha Basma, MPH BS – Research Coordinator, Pediatric Neurology, Weill Cornell Medicine; Zachary Grinspan, MD MS – Associate Professor, Director of Pediatric Epilepsy, Healthcare Policy & Research, Pediatrics, Weill Cornell Medicine
Rationale:
Epilepsy is a neurological disorder that affects 470,000 children in the United States. Care patterns and clinical outcomes are influenced by social determinants of health (SDOH), such as environment, neighborhood, and sociocultural influences, and there are known disparities in care. Understanding the association between neighborhood-level social determinants of health (SDOH) and outcomes may provide insight into strategies to improve care for underserved populations.
Methods:
We used the DM_peds database at Weill Cornell, New York Presbyterian Hospital (2019 through 2022) to identify encounters with children with epilepsy based on the International Classification of Disease (ICD10) G40 codes. Emergency department (ED) visits were extracted, aggregated by patient identifier numbers, and merged with the encounter database. Those patients absent from the original ED visit extraction were assigned zero ED visits per year. We then linked the 2020 Agency for Healthcare Research and Quality (AHRQ) dataset with SDOH variables using each patient’s US Census Federal Information Processing Standard (FIPS) codes. All statistics were performed in R version 4.3.0. Logistic and Least Absolute Shrinkage and Selection Operator (LASSO) regressions identified key variables aligned with the National Institute on Minority Health and Health Disparities (NIMHD) framework. The NIMHD framework categorizes SDOH based on Domains of Influence, Levels of Influence, and Health Outcomes to better classify, analyze and address disparities in health outcomes.
Results:
Analysis of 8,186,934 pediatric encounters between 2019 through 2022 yielded 4,472 unique patients with epilepsy, including 479 with ≥ 2 average ED visits per year and 3993 with 0 or 1. Regression analysis narrowed the AHRQ dataset's 320 variables to 32 (logistic) or 22 (LASSO). After reviewing the highest LASSO coefficients and inspection of the NIMHD framework, we selected five variables. These variables showed that the odds of an ED visit decreased with neighborhoods having a higher percentage of renter-occupied housing units with children, units with incomplete kitchen facilities, and in-state county movers. The odds of an ED visit were increased in neighborhoods with a higher percentage of never-married women and widowed men. The final variables clustered within the NIMHD framework's Sociocultural and Physical/Built Environments at the Individual and Community Levels.
Conclusions:
The findings highlight the potential for census tract-specific interventions. Programs like family education, nutritional counseling, and support for single parents can be tailored to patients residing in high ED visit census tracts. Addressing these determinants can lead to more targeted care, potentially reducing ED visits and enhancing outcomes for pediatric epilepsy patients.
Funding: New York Academy of Medicine, David E. Rogers Fellowship Award