Impact of neonatal sepsis calculators on diagnostic accuracy and antimicrobial stewardship: A systematic review and meta-analysis

Background and Aim: Early-onset neonatal sepsis (EOS) is a life-threatening infection occurring within the first 72 hours of life, requiring timely treatment to optimize outcomes. Although group B streptococcus screening and intrapartum antibiotic prophylaxis have reduced EOS incidence, concerns regarding antibiotic overuse persist. The neonatal sepsis calculator, developed by Kaiser Permanente Northern California, is a risk-based prediction tool designed to improve antibiotic stewardship. This systematic review and meta-analysis evaluate its diagnostic accuracy and impact on antibiotic use. Methods: A systematic literature search was conducted across PubMed, Scopus, Web of Science, ScienceDirect, and ProQuest (up to July 1, 2024), following the PRISMA guidelines (PROSPERO ID 567269). Studies assessing the neonatal sepsis calculator against blood culture results were included. Data on antibiotic usage with and without calculator implementation were extracted. Meta-analysis of diagnostic accuracy was performed using STATA 16.0 to determine pooled sensitivity, specificity, diagnostic odds ratio (DOR), and area under the curve (AUC). Proportional and binary meta-analysis using RStudio 4.4.1 summarized antibiotic usage changes. Results: From 1,523 records, 21 observational studies were included. The neonatal sepsis calculator showed a pooled sensitivity of 68% (95% CI 49–82%) and specificity of 78% (95% CI 57–90%), with an AUC of 0.79 (95% CI 0.75–0.82) and a DOR of 7.25 (95% CI 2.49–21.08). Its implementation significantly reduced antibiotic use (OR 0.045, 95% CI 0.009–0.236, p=0.001). Conclusion: The neonatal sepsis calculator demonstrates good diagnostic accuracy and supports antibiotic stewardship. Future refinements incorporating neonatal factors may enhance its predictive performance. (www.actabiomedica.it).