Neonatal sepsis prediction through clinical decision support algorithms: A systematic review

Area of Science:

• Neonatal Medicine
• Clinical Informatics
• Biomedical Engineering

Background:

• Sepsis is a leading cause of neonatal mortality.
• Early and accurate sepsis prediction is crucial for timely intervention.
• Non-invasive parameters offer a safe and accessible approach to monitoring.

Purpose of the Study:

• To systematically review evidence on clinical decision support algorithms (CDSAs) for neonatal sepsis prediction.
• To evaluate the effectiveness of non-invasive parameters in these algorithms.
• To assess the current state of evidence and identify research gaps.

Main Methods:

• Comprehensive literature search across PubMed, CENTRAL, and EMBASE.
• Systematic screening, data extraction, and risk of bias assessment by two independent reviewers.
• Certainty of evidence evaluated using the GRADE framework.

Main Results:

• 36 studies including 18,096 neonates were analyzed.
• Heart rate (HR)-based parameters were the most frequently evaluated in CDSAs.
• One randomized controlled trial showed reduced 30-day mortality with HR-based CDSAs; other studies yielded inconclusive results.

Conclusions:

• HR-based parameters are valuable components of CDSAs for neonatal sepsis prediction.
• Combining HR with other vital signs and demographics may improve accuracy.
• Limited standardization and inconclusive evidence hinder widespread clinical implementation; further research is essential.