Multivariate risk and clinical signs evaluations for early-onset sepsis on late preterm and term newborns and their economic impact.

There is an increasing evidence that strict evaluation of clinical signs is effective in detecting newborns at risk of early-onset sepsis (EOS) that require antibiotic therapy. In a retrospective case control design, we compared EOS antibiotic indication by clinical signs surveillance with multivariate risk analysis (EOSCalc), and estimate their costs. Newborns ≥ 34 weeks who received EOS antibiotics from June 2014 to December 2016 were studied. Were considered symptomatic those with three clinical signs within first 24 h or two signs and one risk factor present. Cost estimative was done using bottom-up hospital's perspective. Eight thousand three hundred twenty-one were born, 384 were included. Two hundred nineteen (57%) would receive antibiotics by EOSCalc and 64 (16.7%) by clinical signs (p < 0.001). All patients with blood cultures were detected and false-negatives were absent. Total cost was US$ 574,121, estimate US$ 415,576 by EOSCalc, and US$ 314,353 by clinical signs (p < 0.001).Conclusions: The use of EOSCalc and clinical signs surveillance seem to be safe and accurate methods in EOS management. Additionally, the two approaches have shown an economic advantage when compared with the hospital's current practice. What is Known: • EOSCalc is a useful method for screening of EOS in late preterm and term infants. • Presence of clinical signs and/or maternal risk factors are present newborns with EOS. What is New: • Rigorous observation of clinical signs is a more accurate method than EOSCalc to screen for EOS in late preterm and term newborns. • Rigorous observation of clinical signs is more economic than EOSCalc in managing EOS in late preterm and term neonates.