Because of the risk of SBI, infants less than 2 yrs who are febrile and ill-appearing are frequently admitted to the hospital and treated with antimicrobials pending the result of culture of blood, urine and cerebrospinal fluid. Most infants treated in this manner have self-limited infections that do not require therapy. Molecular techniques that can accurately and rapidly detect the presence of bacteria in clinical specimens (Pediatrics 1995; 95:165) might impact on this practice by permitting identification of infants that could be discharged soon after admission. The potential impact of using these techniques was assessed in a retrospective analysis of all infants <2 yrs who were hospitalized for suspected SBI at The New York Hospital from 1/1/95 to 12/31/95. During this period 144 infants (4.9% of all pediatric admissions; mean age 2.4±3.8 months) were admitted and treated for SBI. SBI was ruled-out and infants were discharged within 100 hours of admission(mean 79±16hrs) in 77 (53.4%) episodes. These infants had remained in the hospital awaiting final culture reports. Based on average costs of hospitalization for days 2-4 ($1363/day) of infants suspected of having SBI, the savings in hospitalization costs if SBI was ruled out in these 77 infants at 24 or 48 hrs of hospitalization would have been $209,902 or $104,951 respectively. Molecular techniques aimed at detecting bacteria in clinical specimens within 24 hrs would need to be highly sensitive and have nearly 100% negative predictive value to be acceptable. If this testing was offered selectively to infants suspected of having SBI, testing would have to cost less than $1,458/patient to be cost effective in this tertiary care, pediatric service that is part of a teaching hospital.