Comparison of institutional guidelines with established early onset sepsis risk calculator in reducing antibiotic use in an inner-city NICU in US.

Objectives: Early onset sepsis (EOS) incidence has decreased since national guidelines and intrapartum prophylaxis were introduced. However, there has been a rising concern in antibiotic overtreatment for suspicion of EOS. A web-based EOS calculator has recently been used to evaluate the risk in newborns ≥34 weeks. Our purpose was to compare local strategies with the EOS calculator in our setting with an EOS incidence of 2/1000 live births.

Methods: A retrospective review of all newborns born ≥34 weeks from 1 January 2016 to 31 December 2017 was completed after receiving IRB approval. We applied the calculator to those eligible using an EOS incidence of 0.6/1000 and 2/1000 live births and divided the patients into four cohorts. The rate of antibiotic use was compared between local evidence-based guidelines and the EOS calculator.

Results: Of the 1367 newborns included in the study, 679 received antibiotics. Over the 2 years, antibiotic utilization decreased by 38%. The calculator would have recommended antibiotics for 468 patients (31% decrease) for an EOS incidence of 0.6/1000, but with a 2/1000 incidence rate the calculator recommended antibiotics for 673 patients (1% decrease).

Conclusions: The EOS calculator has been helpful in optimizing antibiotic use, but its use may result in suboptimal treatment without the knowledge of local EOS incidence. Our local antibiotic stewardship guidelines seemed to be comparable to the EOS calculator, especially by the last 6 months of the study period.