Hospital admissions of children from the emergency department: are decisions regarding children on public assistance different?

OBJECTIVE: To better understand the variables that influence the physician's decision to admit children from the emergency department (ED) for nonmedical problems. METHODS: A multicenter prospective questionnaire survey over a three-month study period. For all admitted children, the emergency physician completed a survey which recorded demographic data, insurance status, primary care provider (PCP), admitting diagnoses, and reason for admission. The reason for admission was noted as strictly medical or nonmedical (either an illness that could have been managed on an ambulatory care basis or a "psychosocial" admission). Group differences were analyzed by t test, chi 2, or logistic regression analysis where appropriate. RESULTS: There were 4318 ED admissions at five institutions of which 185 (4%) were judged to be nonmedical. No age or gender differences were found between the medical and nonmedical admission populations. Using logistic regression, adjusted odds ratios for nonmedical admissions were as follows: Medicaid insurance (2.34, 95% CI = 1.61-3.41), clinic-based primary care provider (1.54, 95% CI = 1.06-2.23), no or unknown primary care provider (2.40, 95% CI = 1.52-3.78), and after hours [eg, 5 PM to 8 AM] admissions (1.86, 95% CI = 1.31-2.63). CONCLUSIONS: These data suggest that children with lower socioeconomic status and limited primary care resources are more likely to be admitted from the ED for nonmedical reasons than children with commercial insurance resources or a private physician.

Clinical predictors of computed tomographic abnormalities following pediatric traumatic brain injury.

Children commonly seek attention in emergency departments following head injury. Head computed tomography (CT) is often used to decide subsequent disposition. Clinical criteria predicting CT abnormalities would allow effective and timely treatment and minimize unnecessary procedures depleting overburdened medical resources. We prospectively compared presenting clinical features with subsequent emergent head CT in 300 children less than 19 years old over a nine-month period. The disposition of patients following imaging was also recorded. Only suspected abuse was more than 50% positively predictive in children below age two and those above age two. Two signs were more than 67% positively predictive in both age groups: focal motor deficit and pupillary asymmetry. Patients with abnormal CTs were the only children to undergo emergent neurosurgery (30%) and were nearly five times as likely to be intensively monitored. Children with normal CTs were nearly five times as likely to be observed in a routine department or at home. We conclude that no single clinical feature can predict with certainty an abnormality on immediate head CT. However, children suspected of being abused, and those with focal motor deficits or pupillary asymmetry, should be imaged. Finally, emergent CT when judiciously ordered likely reduces unforeseen morbidity and minimizes costly intensive care observation.