Update to the Neonatal Early-Onset Sepsis Calculator Utilizing a Contemporary Cohort.

Abstract

Background and objectives: The Kaiser Permanente Neonatal Early-Onset Sepsis (EOS) Calculator has been an effective tool for risk stratification to safely reduce newborn antibiotic exposure. The calculator was derived from data on infants born between 1993 and 2007. Since that time, US obstetric practice has adopted universal antepartum screening for group B Streptococcus and intrapartum antibiotic prophylaxis guidance has changed. Our objective was to update the EOS calculator using a contemporary birth cohort and determine the effect of these changes on EOS case ascertainment and antibiotic recommendations.

Methods: The study included infants born at ≥35 weeks' gestation at 14 hospitals between January 2010 and December 2020 (n = 412 595 infants, EOS cases = 113). Model coefficients were re-estimated and the point estimates of the likelihood ratios for clinical status used to calculate the posterior probability of EOS. We compared the number of EOS cases correctly identified by each model (sensitivity) and the proportion of infants for whom empirical antibiotics are recommended.

Results: The original model had a sensitivity of 0.76 (95% confidence interval 0.63-0.85), while the updated model had a sensitivity of 0.80 (95% confidence interval 0.68-0.89), P = .15. The recommended empirical antibiotic use was 3.5% with the original model and 3.7% with the updated model, P < .0001. For each additional case identified by the updated model, an additional 158 infants would be treated with antibiotics.

Conclusions: Both the original and updated EOS calculators are effective tools for quantifying EOS risk among infants born at ≥35 weeks' gestation.