Are empiric antibiotic choices for orbital cellulitis in children presently unnecessarily broad (and unable to prevent surgery)?