Sir.—The articles by DeAngelis et al,1 Joffe et al,2 and the accompanying editorial by Roberts3 in the December issue of AJDC ably presented the views of clinicians experienced in the continuing problem of the treatment of febrile infants. Yet, I was disappointed to note by its absence one predictive element, the presence or absence of fever in the siblings or family members of the affected infants.
During my third year of pediatric residency at the University of Miami, we became curious about that issue and conducted an informal review of our hospitalized febrile infants who were younger than 30 days of age. After reviewing 96 cases, we found that 18% were found to have a documented bacterial infection or neonatal pneumonia. The vast majority of those could be easily identified as being "sick" with the available data in the emergency room. Retrospectively, of those who were