(1327) Pediatric ICU Risk-Adjusted Outcomes Based Upon U.S. Geographic Regions and Levels of Care

Introduction: Over the past three decades, there has been ongoing quality improvement (QI) and a decrease in mortality rates in pediatric intensive care units (PICUs) in the U.S. New recommendations for admission and levels of care in PICUs have been developed (Frankel, et al., 2019). Previous research has shown improved outcomes from tertiary care centers as compared to non-tertiary centers (Pollack et al., 1991). This study aimed to evaluate differences in outcomes based upon geographic regions and levels of care in the U.S., and identify QI opportunities.

Methods: 415,430 cases from 2021-23 comprising 125 PICUs in the VPS registry were analyzed to determine the relationship between levels of care and U.S. geographic regions simultaneously with standardized mortality ratio (SMR) and standardized length of stay ratio (SLOSR) using the PIM3 and PRISM III models. PICUs were classified as community, tertiary or quaternary, and U.S. Census region classification was used to define geographic regions. Two-way Analysis of Variance (ANOVA) was used to account for interactions between region and level of care. Tukey’s Honestly Significant Difference (HSD) test was used for the comparisons and adjusted pairwise Tukey HSD p-values were calculated with significance defined as p value < 0.05. To minimize the impact on outcomes based upon transfers across different levels of care, the analyses were performed both with and without transfers from PICUs.

Results: One way ANOVA showed improvement in SMR in community PICUs compared to higher levels of care. Two-way ANOVA including geographic regions indicated no differences in SMR and SLOSR between levels of care except PIM3 SMR South-quaternary PICUs compared to West-community PICUs (p=0.017) and PRISM III SMR South-quaternary PICUs compared to West-community (p=0.018). These differences remained significant after re-analyzing the data after excluding transfers from PICUs.

Conclusions: 1) Differences in SMR based upon levels of care may be explained by geographic regions in the U.S.
2) There is a significantly lower SMR in West-community PICUs compared to South-quaternary PICUs. This may be secondary to differences in race and access to critical care.
3) Future studies should evaluate the impact of social determinants on SMR in PICUs and identify targeted QI opportunities.