A calculated risk Risk stratification as a basis for medical management of early-onset sepsis risk

Invasive bacterial infection in the first days of life (‘early onset sepsis’, EOS) can have severe consequences. Early administration of empiric antibiotics is effective, but it is difficult to discern symptoms of EOS from normal newborn physiology or other neonatal problems. Consequently, large numbers of newborns are treated with antibiotics and exposed to the harmful effects of antibiotics, despite actually not having EOS. This results in negative effects for newborns and parents, and increases costs for society.
This thesis evaluates the EOS calculator, a new method to allocate antibiotics for suspected EOS. We found a significant reduction in unnecessary antibiotics associated with the EOS calculator in three different studies: a single center implementation, a systematic review and meta-analysis of 13 studies, and a sub-analysis of an observational multicenter prospective cohort study. We also found that the EOS calculator may reduce other health care utilization and related costs. Recommendations by the EOS calculator may be more in line with clinician intuition than current Dutch national guidelines; adherence to those is low.
Implementation and of the EOS calculator should happen thoughtful, and with caution. Interpretation of calculated EOS risk should be nuanced, and use of non-standard parameters should be preceded by recalibrations. However, use of different strategies to determine maternal presence of the dangerous Group B Streptococcus pathogen is unlikely to hinder implementation.
Further research should focus on cautious implementation and good monitoring of the EOS calculator in clinical practice, with comparisons and combinations with other strategies and biomarkers.