Objective: Machine-learning (ML) algorithms allow for improved prediction of sepsis syndromes in the ED using data from electronic medical records. Transfer learning, a new subfield of ML, allows for generalizability of an algorithm across clinical sites. We aimed to validate the Artificial Intelligence Sepsis Expert (AISE) for the prediction of delayed septic shock in a cohort of patients treated in the ED and demonstrate the feasibility of transfer learning to improve external validity at a second site.
Methods: Observational cohort study utilizing data from over 180,000 patients from two academic medical centers between 2014 and 2019 using multiple definitions of sepsis. The AISE algorithm was trained using 40 input variables at the development site to predict delayed septic shock (occurring greater than 4 hours after ED triage) at varying prediction windows. We then validated the AISE algorithm at a second site using transfer learning to demonstrate generalizability of the algorithm.
Results: We identified 9354 patients with severe sepsis of which 723 developed septic shock at least 4 hours after triage. The AISE algorithm demonstrated excellent area under the receiver operating curve (>0.8) at 8 and 12 hours for the prediction of delayed septic shock. Transfer learning significantly improved the test characteristics of the AISE algorithm and yielded comparable performance at the validation site.
Conclusions: The AISE algorithm accurately predicted the development of delayed septic shock. The use of transfer learning allowed for significantly improved external validity and generalizability at a second site. Future prospective studies are indicated to evaluate the clinical utility of this model.