614 - Emergency Department Utilization Among Latent Classes of Social Risk in a Pediatric Primary Care Population

Publication Number: 3595.614

Michael Mattiucci, Golisano Children's Hospital at The University of Rochester Medical Center, Rochester, NY, United States; Yaxu Zhuang, University of Colorado Anschutz Medical Campus, Aurora, CO, United States; Guannan Shen, University of Colorado School of Medicine, Denver, CO, United States; Sana Yousuf, Children's Hospital Colorado, Denver, CO, United States; Elizabeth Bayliss, Kaiser Permanente Colorado, Aurora, CO, United States; Lisa Ross. DeCamp, University of Colorado| Children's Hospital Colorado, Aurora, CO, United States

Background: Health-Related Social Needs (HRSN) are associated with poor health outcomes and increased emergency department (ED) utilization, particularly for young children. There is a need for tailored interventions to address HRSN for high-risk children in primary care. In publicly insured adults, distinct social risk profiles have been developed using Latent Class Analysis (LCA), finding increased ED utilization in high social risk profiles. Development of social risk profiles can inform tailored HRSN interventions; however, this method has not been documented in a pediatric population.

Objective: To investigate ED utilization within social risk profiles of children age 0-3 in pediatric primary care using Latent Class Analysis.

Design/Methods: We created a dataset using electronic health record data at a large, academic primary care clinic in Aurora, CO. We included publicly insured patients < 3 who had a well child visit (WCV) between 10/1/21 and 3/31/22 with a social risk screening tool completed, and who had at least one additional WCV with social risk screening between 4/1/22 and 3/31/24. Indicator variables for latent class analysis (LCA) were: 1) food insecurity, 2) housing instability, 3) financial stress, and 4) difficulty keeping appointments. Bivariate analyses were used to examine differences in demographics and social risk factors between classes. Multivariable logistic regression was used to examine ED utilization between classes, adjusting for demographic and clinical factors.

Results: LCA identified three social risk profiles: High (2.0% of sample), Moderate (22%) and Low social risk (76%). When compared with the moderate-risk class, children in the high-risk class have higher likelihood of multiple social risks (Table 1, p< 0.001). Univariate logistic regression found that age, race/ethnicity, asthma, and Feudtner Complex Chronic Conditions were associated with increased ED utilization (all p< 0.01). No differences in ED utilization were found between the low social risk class and high (p=0.53) or moderate (p=0.23) social risk classes using multivariable logistic regression.

Conclusion(s): Distinct social risk profiles were identified using LCA in publicly insured children age 0-3 in an urban primary care setting, highlighting the potential for tailored interventions to address HRSN in higher risk groups. However, no differences in ED utilization were found between classes. Additional research is needed to determine effective methods to identify children who have high social risk and are more likely to use the ED.

Table 1
Characteristics of Latent Classes of Social Risk in Pediatric Primary Care