O3218: Admissions for Diabetic Ketoacidosis Correlate With Neighborhood Disparities in Child Opportunity

Background: Diabetic ketoacidosis (DKA) is a life-threatening but preventable complication of diabetes, and is commonly the initial presentation of new-onset diabetes among pediatric patients. With rates of DKA rising, and further exacerbated by the COVID pandemic, there is an increased need to investigate risk factors to aid efforts at prevention. The Child Opportunity Index (COI) 2.0 provides composite scores based on 29 indicators of resources and conditions that promote childhood thriving, with neighborhood specificity at the level of census tracts. We hypothesized that DKA admissions at our hospital would be related to patients’ neighborhood COI in our catchment area.

Methods: We reviewed DKA admissions (n = 640) from January 2020 through July 2022 at our tertiary care children’s hospital. Patient addresses were converted to census tracts and aggregated in a deidentified manner for analysis. State-normed COI scores were collected for census tracts in our catchment area, ranging from 1 (least opportunity) to 100 (greatest opportunity), as well as stratified by quintiles. COI scores were available for 633 admissions (99%); ratings were lacking for 7 census tracts with 1 admission each. We examined for relationships between admission frequency and COI using chi-square test and linear regression. This study was approved by our institutional review board (#5230524).

Results: DKA admissions were unequally distributed among neighborhoods, showing an inverse relationship with opportunity levels. Of the 633 admissions during the study period with COI data available, 302 (48%) were from census tracts in the lowest quintile (very low opportunity), while only 21 (3%) were from census tracts in the highest quintile (very high opportunity) (Figure 1A, chi-square = 380, df = 4, p = 6 x 10^-81). Linear regression of individual census tracts showed a significant negative correlation (Figure 1B, slope = -0.012, 95% CI -0.020 to -0.003, p = 0.006).

Conclusion: We found disparities in frequency of DKA admissions across neighborhood opportunity levels, with more admissions being concentrated in areas of lower COI scores (Figure 2). We have begun a root cause analysis to further investigate the reasons behind this pattern, such as patients’ ability to follow up with a pediatric endocrinologist after hospital discharge, or possible lack of primary care in areas of physician shortage. Future work also includes testing whether food deserts in areas of low COI may be contributing to DKA risk due to lack of access to healthy foods. Geospatial analyses such as ours can be used to identify “hotspots,” geographic areas that are at greater risk for poor health outcomes. Such investigations can inform both physicians and policymakers by directing interventions toward the communities that would receive the most benefit.