The management of children with fever of indefinite source still remains controversial. This study aimed to compare different practice patterns between pediatric physicians (PPs) and emergency physicians (EPs) in the management of pediatric fever in the emergency department (ED) and correlate them to existing practice guidelines. Their impact on patient outcomes was also discussed.

Medical records of patients 3 to 36 months of age who presented to the ED with fever of indefinite source from June 1 to December 31, 2006, were retrospectively reviewed on day 5 after the patient's first visit. At the same time, telephone follow-up was carried out to determine whether the patient had been visiting or being admitted to another clinic or hospital after discharge. Variation in practice patterns were compared for the number of laboratory tests, ED length of stay (LOS), and the rate of immediate admission. Patient outcomes were measured as the rate of unscheduled revisit within 72 hours and the rate of subsequent admission. Compliance with existing practice guidelines between PPs and EPs were evaluated by dividing all eligible patients into 3 groups: (1) toxic appearing patients (group A), (2) nontoxic patients with body temperature (BT) ≧39°C (group B), and (3) nontoxic patients with BT below 39°C (group C).

A total of 345 patients who met the inclusion and exclusion criteria were enrolled into this study. Pediatric physicians and EPs treated 163 and 182 febrile children, respectively. In group A, PPs admitted more patients than EPs (41% vs 12 %), whereas more unscheduled revisits were seen in EP-treated patients (44% vs 10%). In group B, PPs ordered more laboratory tests than EPs (2.3 vs 0.7 tests per patient), and their patients also had a longer ED LOS (3.4 ± 3.2 vs 1.5 ± 1.1 hours). However, no difference was found in their rates of immediate admission and unscheduled revisit. In group C, PPs admitted more patients (15% vs 0%) and ordered more laboratory tests (2.0 vs 0.5 tests/patient) than EPs. Longer ED LOS (3.3 ± 3.9 vs 1.0 ± 1.4 hours) was also noted among PP-treated patients. However, no difference was noted in their rates of unscheduled revisit. In all groups, the rates of subsequent admission were similar.

Compliance with existing practice guidelines (admit the toxic cases and work up those with BT ≧39°C) was higher among PPs, which resulted in a lower rate of unscheduled revisit, but no significant difference was found in the rate of subsequent admission.