Hospital-level factors associated with pediatric emergency department return visits

BACKGROUND: Return visits (RVs) and RVs with admission (RVAs) are commonly used emergency department quality measures. Visit- and patient-level factors, including several social determinants of health, have been associated with RV rates, but hospital-specific factors have not been studied. OBJECTIVE: To identify what hospital-level factors correspond with high RV and RVA rates. SETTING: Multicenter mixed-methods study of hospital characteristics associated with RV and RVA rates. DATA SOURCE: Pediatric Health Information System with survey of emergency department directors. MEASUREMENTS: Adjusted return rates were calculated with generalized linear mixed-effects models. Hospitals were categorized by adjusted RV and RVA rates for analysis. RESULTS: Twenty-four hospitals accounted for 1,456,377 patient visits with an overall adjusted RV rate of 3.7% and RVA rate of 0.7%. Hospitals with the highest RV rates served populations that were more likely to have government insurance and lower median household incomes and less likely to carry commercial insurance. Hospitals in the highest RV rate outlier group had lower pediatric emergency medicine specialist staffing, calculated as full-time equivalents per 10,000 patient visits: median (interquartile range) of 1.9 (1.5-2.1) versus 2.9 (2.2-3.6). There were no differences in hospital population characteristics or staffing by RVA groups. CONCLUSION: RV rates were associated with population social determinants of health and inversely related to staffing. Hospital-level variation may indicate population-level economic factors outside the control of the hospital and unrelated to quality of care.