Evaluating the Impact of Common Clinical Confounders on Performance of Deep-Learning based Sepsis Risk Assessment

Shikha Chaganti^1*Vivek Singh¹Alasdair Edward Gent²Rishikesan Kamaleswaran²Ali Kamen¹

• ¹Siemens Healthineers (United States), Princeton, United States
• ²Duke University, Durham, North Carolina, United States

Early identification of sepsis in the emergency department using machine learning remains a challenging problem, primarily due to the lack of a gold standard for sepsis diagnosis, the heterogeneity in clinical presentations, and the impact of confounding conditions. In this work, we present a deep learning-based predictive model designed to enable early detection of patients at risk of developing sepsis, using data from the first 24 hours of admission. The model is based on routine blood test results commonly performed on patients, including CBC (Complete Blood Count), CMP (Comprehensive Metabolic Panel), and lipid panels, as well as vital signs, age, and sex.To address the challenge of label uncertainty as a part of the training process, we explore two different definitions, namely, Sepsis-3 and Adult Sepsis Event. We analyze the advantages and limitations of each in the context of patient clinical parameters and comorbidities. We specifically examine how the quality of the ground truth label influences the performance of the deep learning system and evaluate the effect of a consensus-based approach that incorporates both definitions. Our results show that the consensus-based model identifies at-risk patients in the first 24 hours with 83.7% sensitivity, 80% specificity, and an AUC of 0.9.We also evaluated the model's performance across sub-cohorts, including patients with confounding comorbidities (such as chronic kidney and liver disease, and coagulation disorders) and those with confirmed infections. Our analysis revealed variable performance across different cohorts. This work highlights the limitations of retrospective sepsis definitions, and underscores the need for tailored approaches in automated sepsis detection, particularly when dealing with patients with confounding comorbidities.