Harnessing AI for Improved Diagnosis and Management of Pediatric Sepsis: Current Advances, Challenges, and Future Directions

Abstract

Artificial intelligence (AI) has been applied to early recognition and management of rapidly progressive, community-acquired pediatric sepsis, a leading cause of childhood mortality. The broad adoption of electronic health records combined with rapid advances in digital technologies have enabled the federated training of both knowledge-driven AI, known as expert systems, trained by teams of collaborating clinicians, and data-driven AI, known as machine learning (ML), to derive predictive, clustering algorithms trained on "big data." An important subset of ML is "deep learning," which includes tools that understand, interpret, and manipulate human imagery and language, such as natural language processing and its subset large language models. We are in an era of rapid deployment of AI/ML-powered tools ranging from real-time electronic health records-embedded decision support tools to continuous wearable vital sign monitors and mobile/conversational virtual assistants/triage apps. These applications have the potential of transforming the timeliness of life-saving sepsis care delivery. This review explores the current and potential AI/ML applications in sepsis care, including tools for screening/early detection, risk stratification/outcome prediction, personalized treatment, and continuous patient monitoring. We highlight successful implementations and ongoing clinical trials, emphasizing the impact on patient outcomes. Finally, we address practical considerations for the future, such as bias mitigation and integration into clinical workflows.

Keywords: artificial intelligence; early diagnosis; health care technology; machine learning; pediatric sepsis.