PROS Pearls:

* As part of the Febrile Infant Study, investigators examined the relationship between observed variability in the diagnosis and treatment of febrile infants and differences in clinical presentation, demographic features, practice/practitioner variables, and geographic region.

* In regards to patient's appearance, most febrile infants (73%) were described as mildly ill while 25% described as moderately ill and <2% as very ill. Temperatures ranged from a low of 38°C (minimal temperature for study inclusion) to 41°C. Results showed that practitioners hospitalized 61.3% of febrile infants younger than 28 days of age, although guidelines indicate that all infants in this age range should be hospitalized. The majority (86.7%) of the nonhospitalized infants were described as mildly ill compared to 56.1% of the hospitalized infants (P <.001).

* The overall model explained 46.5% of the observed variance in evaluation and treatment intensity, of which the clinical characteristics of the patient (eg, patient's appearance, temperature) alone explained 29.7%. Practice site fixed effects also contributed significantly to the model and explained 15% of the overall variance, but practice location measures and provider and practice characteristics explained only about 1%.

* These results imply that much of the observed variability in the evaluation and treatment of this common and potentially serious condition reflects providers tailoring their management to the patient's clinical presentation. These findings should help reassure health care consumers and payers that variability in treatment reflects a pattern that is rational and not a consequence of characteristics of the provider, his or her practice, or regional differences.