Introduction: We aimed to elucidate mechanisms driving disparate PARDS mortality by evaluating the influence of U.S. region and race/ethnicity utilizing joint modeling, hypothesizing that geographic region would compound race/ethnicity disparities in the South. Methods: We performed a retrospective cohort study of children ≤ 18 years with PARDS in the 2016, 2019, and 2022 HCUP Kids’ Inpatient Database. PARDS cases were identified by a published ICD-10 algorithm. Mixed effect logistic regression estimated adjusted odds ratios (aOR) for in-hospital mortality across US region (Northeast, Midwest, South, West) and separately across 16 joint exposure groups (region × race/ethnicity), controlling for age, sex, APR-DRG severity, complex chronic conditions, median ZIP-code income quartile, and hospital type. AI was used for grammatical and statistical code generation assistance. Results: In geographic analyses, children in the South (OR 1.18, 95%CI 1.05-1.33) and West (1.17, 1.03-1.33) faced increased PARDS mortality compared to those in the Northeast. In joint analyses, there were no differences in mortality for Black, Hispanic, and “Other” children in the Northeast, relative to White children in the Northeast (reference). In the Midwest, White, Black, Hispanic mortality did not differ from reference, but “Other” children (aOR 1.29, 95%CI 1.01- 1.65) had increased mortality risk. In the South, mortality risk was higher for White (aOR 1.27, 95 % CI 1.09–1.48), Black (aOR 1.39, 1.18-1.62), and Other (aOR 1.69, 1.36-2.09), but not Hispanic children (1.16, 0.98-1.38). In the West, all race/ethnicity groups experienced increased odds of mortality, with Black children (aOR 1.44, 1.16-1.79) facing the largest increase relative to reference. Similar patterns of disparities were observed when examining the years 2016, 2019, and 2022 separately. Conclusions: Regional disparities in PARDS mortality exist and are modified by race and ethnicity. These findings imply that structural and healthcare-system disparities operating at regional levels interact with racial/ethnic inequities. Interventions to improve outcomes should account for regional healthcare infrastructure differences, coupled with policies that address systemic drivers of inequity.
Keim et al. (Sun,) studied this question.