Referring Hospital Characteristics Associated With Potentially Avoidable Emergency Department Transfers.

OBJECTIVE: Many emergency department (ED) transfers of children may be avoidable. Identifying hospital-level variables associated with avoidable transfers may guide system-level interventions to improve pediatric emergency care. We sought to examine hospital characteristics associated with ED transfers deemed "probably avoidable" in a large state Medicaid program.

METHODS: We performed a retrospective cohort study using 2009 to 2013 claims data for Pennsylvania Medicaid beneficiaries. We categorized all ED transfers of children < 17 years old as "probably avoidable," "possibly avoidable", or "unavoidable" based on ultimate disposition and procedures (including subspecialty consultations) at the receiving ED. Using descriptive statistics and multivariable regression, we examined hospital characteristics associated with probably avoidable transfers.

RESULTS: Among 2,839,379 pediatric visits to EDs across 158 Pennsylvania hospitals, 20,304 resulted in transfer. Among these, 3,764 (18.5%) were categorized as probably avoidable and 6,091 (30.0%) as possibly avoidable transfers. In adjusted analysis, compared to hospitals with no pediatric-specific capabilities, probably avoidable transfers were less likely from referring hospitals with pediatric-specific EDs and no other pediatric-specific capabilities (adjusted odds ratio [aOR] = 0.38, 95% confidence interval [CI] = 0.21-0.71) and from referring hospitals with pediatric-specific EDs and inpatient capabilities (aOR = 0.36, 95% CI = 0.20-0.64). Probably avoidable transfers were more likely from referring hospitals in large metropolitan areas (aOR = 2.64, 95% CI = 1.46-4.80) compared to those in rural areas.

CONCLUSIONS: Among pediatric emergency transfers in a large state Medicaid program with a nearly 20% probably avoidable transfer rate, there was significant hospital-level variation in the proportion of probably avoidable transfers. Transfers from hospitals in large metropolitan areas and transfers from hospitals without pediatric-specific capabilities had increased odds of being probably avoidable transfers, such that these hospitals may represent targets for interventions to reduce these transfers.