Operationalizing Sepsis Definitions With Artificial Intelligence: Toward Smarter Antibiotic Use

Abstract

Sepsis Definitions to BetterTarget and Tailor Sepsis Care." [1] We agree that a revision of sepsis definitions is needed to identify better which patients necessitate immediate antibiotic delivery.However, to prevent inadvertent harm, this strategy must be implemented carefully.Artificial intelligence (AI) may help provide an objective assessment of a patient's need for immediate vs. delayed antibiotics.The Surviving Sepsis Campaign's (SSC) Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) stratify patients into four risk groups, which provide antibiotic timing recommendations according to severity of illness (presence or absence of shock) and probability of sepsis ("definite or probable" vs "possible"). .[2] There is mounting evidence that early antibiotic administration (i.e.<1 hour) is necessary only in patients with septic shock, and antibiotics can be safely delayed in lower risk patients.[3]Nonetheless, the Centers for Medicare and Medicaid (CMS) Severe Sepsis and Septic Shock Management bundle (SEP-1) mandates antibiotics within three hours for all patients who meet criteria for severe sepsis.[4]However, determining infected from non-infected patients within this restrictive time frame is challenging, leading to

Ähnliche Arbeiten