Children account for over 25% of emergency department (ED) visits in the United States. Almost ninety percent of pediatric ED visits occur in general community EDs, many of which have low pediatric volumes and sporadic exposure to critically ill children. The goal of this study was to assess the association between ED pediatric volume and mortality with the hypothesis that EDs with lower annual pediatric volumes will have higher mortality for similar life-threatening conditions.
We performed a secondary analysis of the 2012 Nationwide Emergency Department Sample (NEDS) dataset, the largest all-payer ED database in the United States. We included all patients ages 0-17 years. Using ICD-9 codes, the dataset was restricted to all patients with respiratory failure or cardiac arrest to identify critically ill children at high risk for mortality. Annual pediatric ED volume was divided into quartiles: low (<6373), medium (6374-12117), medium-high (12118-25376), and high (>25376). Descriptive statistics were performed and hierarchical multivariable logistic regression used to estimate associations between pediatric ED volume and mortality, adjusting for patient-level (comorbidity count, age, insurance, household income, sex) and hospital-level factors (urban-rural classification, teaching status, trauma designation, inpatient capacity, children’s hospital designation) that may offer competing explanations for this association.
Of the 6.4 million pediatric patient visits in the 2012 NEDS dataset, 2,629 had a primary ICD-9 diagnosis of respiratory failure or cardiac arrest and were included in the cohort. The mean age was 4.1 years (SD 5.4), 40% were female, and the mean comorbidity count was 1.6 (SD 1.77), with 83% having at least one comorbidity. Thirty-nine percent of the cohort died in the ED. The majority of patients had Medicaid insurance (60%), 26% were privately insured, 10% were self-pay, and 49% lived in an urban area. Eleven percent were seen at children’s hospitals, 42% at trauma centers, and 56% at teaching hospitals. Almost 19% of patients were treated at a low-volume (1st quartile) ED, 24% at a low-medium volume ED, 27% at a medium-high volume ED, and 31% at a high-volume ED. High-volume EDs were associated with decreased mortality (OR 0.40 95% CI 0.17-0.94) when compared to low-volume EDs. There was a trend toward incrementally lower odds of mortality in low-medium, medium-high, and high-volume EDs when compared to low-volume EDs, but only high-volume ED mortality reached significance. Using predicted probabilities to convert odds ratios to risk difference, children who present in cardiac arrest or with respiratory failure are 17% less likely to die in a high-volume ED than a low-volume ED, adjusting for patient- and hospital-level covariates available in the NEDS dataset.
In a population-based pediatric ED sample of respiratory failure or cardiac arrest, higher volume EDs were significantly associated with decreased mortality after adjusting for patient- and hospital-level confounders.
© 2017 Published by Elsevier Inc.