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

Area of Science:

• Neonatal Medicine
• Pediatric Infectious Diseases
• Clinical Risk Assessment

Background:

• The Kaiser Permanente Neonatal Early-Onset Sepsis (EOS) Calculator aids in risk stratification to reduce antibiotic exposure in newborns.
• The original calculator was based on data from 1993-2007, preceding changes in US obstetric practices like universal Group B Streptococcus screening and updated intrapartum antibiotic prophylaxis guidelines.

Purpose of the Study:

• To update the EOS calculator using a contemporary birth cohort (2010-2020).
• To evaluate the impact of updated data on EOS case ascertainment and antibiotic recommendations.

Main Methods:

• Re-estimated model coefficients and likelihood ratios for clinical status using data from 412,595 infants across 14 hospitals.
• Compared the sensitivity (correctly identified EOS cases) and the proportion of infants recommended for empirical antibiotics between the original and updated models.

Main Results:

• The updated model demonstrated a slightly higher sensitivity (0.80) compared to the original model (0.76), though not statistically significant (P = .15).
• Empirical antibiotic recommendations were similar: 3.5% for the original model and 3.7% for the updated model (P < .0001).
• The updated model identified more EOS cases, but treating each additional identified case would necessitate treating an extra 158 infants with antibiotics.

Conclusions:

• Both the original and updated EOS calculators are effective for quantifying early-onset sepsis risk in infants born at or after 35 weeks' gestation.
• The updated calculator offers improved sensitivity while maintaining low rates of empirical antibiotic use.