In this study, data were collected from a retrospective set of de-identified pediatric inpatient and emergency encounters for patients between 2-17 years of age, drawn from the University of California San Francisco (UCSF) Medical Center. The performance of the machine learning algorithm was compared to the pediatric severe sepsis definition of Goldstein et al., 2005., and was validated by measuring Area Under the Receiver Operating Characteristic (AUROC). The algorithm AUROC values significantly outperformed the Pediatric Logistic Organ Dysfunction score (PELOD-2) and pediatric Systemic Inflammatory Response Syndrome (SIRS) in the prediction of severe sepsis four hours before onset using cross-validation and pairwise t-tests. These results suggest that a machine learning-based pediatric sepsis prediction tool may enable earlier sepsis recognition and treatment initiation.
Read the paper here.