Sepsis Prediction for the General Ward Setting.

Sean C Yu,Aditi Gupta,Patrick G Lyons,Philip R O Payne,Kevin D Betthauser,Marin H Kollef,Albert M Lai,Andrew P Michelson

doi:10.3389/fdgth.2022.848599

Abstract

ObjectiveTo develop and evaluate a sepsis prediction model for the general ward setting and extend the evaluation through a novel pseudo-prospective trial design.DesignRetrospective analysis of data extracted from electronic health records (EHR).SettingSingle, tertiary-care academic medical center in St. Louis, MO, USA.PatientsAdult, non-surgical inpatients admitted between January 1, 2012 and June 1, 2019.InterventionsNone.Measurements and Main ResultsOf the 70,034 included patient encounters, 3.1% were septic based on the Sepsis-3 criteria. Features were generated from the EHR data and were used to develop a machine learning model to predict sepsis 6-h ahead of onset. The best performing model had an Area Under the Receiver Operating Characteristic curve (AUROC or c-statistic) of 0.862 ± 0.011 and Area Under the Precision-Recall Curve (AUPRC) of 0.294 ± 0.021 compared to that of Logistic Regression (0.857 ± 0.008 and 0.256 ± 0.024) and NEWS 2 (0.699 ± 0.012 and 0.092 ± 0.009). In the pseudo-prospective trial, 388 (69.7%) septic patients were alerted on with a specificity of 81.4%. Within 24 h of crossing the alert threshold, 20.9% had a sepsis-related event occur.ConclusionsA machine learning model capable of predicting sepsis in the general ward setting was developed using the EHR data. The pseudo-prospective trial provided a more realistic estimation of implemented performance and demonstrated a 29.1% Positive Predictive Value (PPV) for sepsis-related intervention or outcome within 48 h.

Highlights

Sepsis is defined as a life-threatening organ dysfunction caused by a dysregulated host response to infection [1]
There is an abundance of Sepsis Prediction in General Ward literature focusing on the early detection and prediction of sepsis through traditional or newly developed scoring systems such as the Systemic Inflammatory Response Syndrome (SIRS) score, National Early Warning Score (NEWS), or quick Sequential Organ Failure Assessment score; or more recently through the use of machine learning models [6–8]
Most of these efforts focus on the Emergency Department (ED) or Intensive Care Unit (ICU) settings which are data-rich and have a higher prevalence of sepsis compared to the general ward setting [9– 11]

Summary

Introduction

Sepsis is defined as a life-threatening organ dysfunction caused by a dysregulated host response to infection [1]. There is an abundance of Sepsis Prediction in General Ward literature focusing on the early detection and prediction of sepsis through traditional or newly developed scoring systems such as the Systemic Inflammatory Response Syndrome (SIRS) score, National Early Warning Score (NEWS), or quick Sequential Organ Failure Assessment (qSOFA) score; or more recently through the use of machine learning models [6–8]. Most of these efforts focus on the Emergency Department (ED) or Intensive Care Unit (ICU) settings which are data-rich and have a higher prevalence of sepsis compared to the general ward setting [9– 11]. The objective of this study was to develop a machine learning model for predicting sepsis in the general ward setting, compare its performance to commonly used instruments for sepsis surveillance such as SIRS and NEWS, and extend the model evaluation using a novel simulated pseudo-prospective trial [13, 14]

Methods

Results

Conclusion