Pediatricians excel at detecting serious bacterial illness

To gain insight into the best strategies for detecting serious bacterial illness, investigators examined data on more than 3,000 infants 3 months or younger who had a temperature of at least 38° C, and who were being treated by 573 practitioners nationwide from the Pediatric Research in Office Settings (PROS) network of the American Academy of Pediatrics. The purpose of the investigation was threefold:

• to characterize how practitioners managed these infants' disease and the spectrum of their disease

• to develop a clinical prediction model for identifying infants with bacteremia or bacterial meningitis

• to compare the accuracy of practitioners' diagnostic strategies with that of existing guidelines.

Here is what investigators found:

• Most infants (64%) were treated exclusively outside of the hospital. Practitioners performed laboratory testing (blood, urine, or cerebrospinal fluid testing) in 75% and initially treated 57% with antibiotics. PROS practitioners were significantly more likely to perform laboratory evaluations, give antibiotics, or hospitalize younger infants, those who appeared more ill, and infants with higher fever. Practitioners followed current guidelines for treating febrile infants in only 42% of episodes. Bacteremia was detected in 1.8% of infants (2.4% of those tested) and bacterial meningitis, in 0.5%. Frequency of other illnesses included urinary tract infection, 5.4%; otitis media, 12.2%; upper respiratory tract infection, 25.6%, bronchiolitis, 7.8%; and gastroenteritis, 7.2%.

• Age of less than 30 days and very ill appearance are the strongest predictors of bacteremia and bacterial meningitis. To develop a clinical prediction model, investigators analyzed collected data to determine which clinical features are, potentially, useful for initially identifying infants at high risk and the predictive value of laboratory testing. High temperature, an abnormal cry, and an abnormal white blood cell count are also strong risk factors for bacteremia and bacterial meningitis.

Using individualized clinical judgment to treat febrile infants, the pediatric clinicians in this study detected as many cases of bacteremia and bacterial meningitis as they would have by strictly adhering to practice parameters, or by following the model developed by investigators. At the same time, they performed fewer tests and hospitalized fewer infants (Pantell RH et al: JAMA 2004;291:1203).

Commentary: This study is a product of the PROS research network. It depicts what is really happening in pediatric office settings across the United States. The PROS network is a particularly important tool in the study of this topic because most data on febrile infants and management guidelines are obtained from the emergency departments of hospitals affiliated with urban universities. This study shows that office management of these children is a different beast. The article and the accompanying editorial, by Kenneth Roberts, MD, are worth a read.