Sepsis Identification Tools: A Narrative Review

Background Although sepsis remains a medical emergency, there is no standard test for diagnosing it. Current sepsis management guidelines strongly recommend screening for sepsis but do not identify a specific tool to use. Objective To summarize the evidence for sepsis screening tools and triggers, identify the current tools used, and describe their effectiveness. Methods A review of the literature from January 2019 through June 2024 was performed. Studies were included if they described sepsis screening tools used for adults in the emergency department or adult inpatients, including intensive care unit patients. Studies were excluded if they described tools specific to machine learning with artificial intelligence or biomarkers and biologics. Results A total of 300 articles were screened. The final set of 26 studies included articles on computerized clinical decision support systems (8 studies), existing early warning systems (14 studies), and new or novel tools (4 studies). Sepsis definitions were heterogeneous and generally based on disease classification codes, criteria from the Sepsis-2 or Sepsis-3 definitions, or combinations thereof. The most commonly used early warning system tools used that had superior evidence were the National Early Warning Score versions 1 and 2. Little evidence supported the use of the quick Sequential [Sepsis-related] Organ Failure Assessment alone for sepsis identification. The use of computerized clinical decision support systems is varied; both proprietary and individual system–developed tools are available, with little consensus on standards for reporting accuracy. Conclusion It is clear that all currently available tools function only as adjuncts to clinical acumen.