A Framework to Tackle Risk Identification and Presentation Challenges in Sepsis

Area of Science:

• Critical Care Medicine
• Health Informatics
• Epidemiology

Background:

• Sepsis onset and recovery are challenging to assess.
• Electronic health records (EHRs) offer a dynamic method for capturing sepsis episodes.
• Understanding sepsis trajectories is crucial for timely intervention.

Purpose of the Study:

• To utilize EHR data for identifying sepsis as a dynamic episode.
• To systematically detect risk factors associated with sepsis onset.
• To identify factors linked to in-hospital mortality in sepsis patients.

Main Methods:

• Retrospective analysis of 276,722 clinical observations from EHRs (2015).
• Cox proportional hazard model to assess risk factors for sepsis onset within 72 hours.
• Logistic regression using the PIRO score framework for in-hospital mortality.

Main Results:

• 47.54% of patients with infection experienced a sepsis episode within 72 hours.
• Sepsis episodes were associated with higher mortality (7.24%) compared to infection or organ dysfunction alone.
• Identified specific risk factors for sepsis onset and in-hospital mortality.

Conclusions:

• EHR data can accurately identify infection, organ dysfunction, and sepsis as dynamic episodes.
• A systematic approach using EHRs can detect critical risk factors for sepsis.
• Findings aid in developing targeted strategies to reduce sepsis-related mortality.