Association between Hospital Volume and Within-Hospital Intensive Care Unit Transfer for Sickle Cell Disease in Children's Hospitals

Abstract

To assess the relationship between hospital volume and intensive care unit (ICU) transfer among hospitalized children with sickle cell disease (SCD). We conducted a retrospective cohort study of 83,477 SCD-related hospitalizations at children's hospitals (2009-2012) using the Pediatric Health Information System database. Hospital-level all-cause and SCD-specific volumes were dichotomized (low vs high). Outcomes were within-hospital ICU transfer (primary) and length of stay (LOS) total (secondary). Multivariable logistic/linear regressions assessed the association of hospital volumes with ICU transfer and LOS. Of 83,477 eligible hospitalizations, 1741 (2.1%) involving 1432 unique children were complicated by ICU transfer. High SCD-specific volume (OR 0.77, 95% CI 0.64-0.91) was associated with lower odds of ICU transfer while high all-cause hospital volume was not (OR 0.87, 95% CI 0.73-1.04). A statistically significant interaction was found between all-cause and SCD-specific volumes. When results were stratified according to all-cause volume, high SCD-specific volume was associated with lower odds of ICU transfer at low all-cause volume (OR 0.46, 95% CI 0.38-0.55). High hospital volumes, both all-cause (OR 0.94, 95% CI 0.92-0.97) and SCD-specific (OR 0.86, 95% CI 0.84-0.88), were associated with shorter LOS. Children's hospitals vary substantially in their transfer of children with SCD to the ICU according to hospital volumes. Understanding the practices used by different institutions may help explain the variability in ICU transfer among hospitals caring for children with SCD.