Daily Sepsis Research Analysis

IMPORTANCE: Hospital quality reporting remains a manual, costly process with critical limitations as a mechanism to improve care outcomes. OBJECTIVE: To assess whether near-real-time quality measurement, enabled by large language models (LLMs), can improve quality performance as measured by the Centers for Medicare & Medicaid Services (CMS) Severe Sepsis and Septic Shock Management Bundle (SEP-1) quality metric. DESIGN, SETTING, AND PARTICIPANTS: This single-blind, unstratified, cluster randomized trial was conducted between December 13, 2024, and July 8, 2025, at 2 academic emergency departments (EDs) within the University of California, San Diego (UCSD) health system. Participants included all 66 attending physicians who practiced in the UCSD EDs and worked more than 3 shifts per month prior to study initiation. INTERVENTION: Participants were randomized to receive targeted feedback from LLM-determined compliance with SEP-1 at the time of patient discharge or standard process. MAIN OUTCOMES AND MEASURES: The primary outcome was overall compliance with SEP-1. Secondary outcomes included expert agreement with the LLM SEP-1 determination, 30-day mortality, and intensive care unit admissions of patients with severe sepsis and/or septic shock in the ED. Effect sizes were estimated from a mixed-effects logistic regression model with the intervention group as a fixed effect and a random intercept for physician. RESULTS: The study population included 66 physicians who treated 301 patients (121 in the control group and 180 in the intervention group; median age, 64.3 [IQR, 51.1-75.7] years; 171 [56.8%] male; 52 [17.3%] with chronic kidney disease; 52 [17.3%] with chronic heart failure) who met CMS inclusion criteria for SEP-1. Physicians in the control group had a SEP-1 compliance rate of 70.1%, while those in the intervention group had a rate of 82.9%. Assignment to the intervention group resulted in a 13.0% absolute improvement in SEP-1 compliance (95% CI, 2.5%-23.4%; odds ratio, 2.10 [95% CI, 1.15-3.81]; P = .02) in the mixed-effects model. The largest difference between the intervention group and control group was in noncompletion of the 30-mL/kg fluid bolus component (3 of 180 [1.7%] vs 16 of 121 [13.2%]), a documentation-sensitive component of the quality measure. Agreement between LLM determination and expert review was 92%. No significant differences existed in intensive care unit admissions or 30-day mortality. CONCLUSIONS AND RELEVANCE: In this cluster randomized trial of artificial intelligence (AI)-enabled medical record abstraction for sepsis care, rapid assessment of SEP-1 performance and targeted feedback improved overall compliance with the measure. AI-driven quality clinical integration may address limitations in existing hospital quality reporting and better support a learning health system. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT07581340.

Microcirculatory dysfunction is a defining feature of septic shock and is strongly associated with mortality, yet its relationship to macrocirculatory haemodynamics remains poorly understood. In particular, the persistence of heterogeneous capillary perfusion despite restoration of blood pressure and cardiac output (termed haemodynamic incoherence) lacks a coherent mechanistic explanation. I developed a conceptual and computational model of the microcirculation in which network behaviour is constrained by three interacting variables: cardiac output, vasomotor state, and shear stress regulation. A network of one million parallel arterioles was simulated using physiologically plausible distributions of vessel radius. For each vessel, flow requirements were determined by an apparent shear target, reflecting endothelial sensing of shear rather than absolute physical values. Total cardiac output required to maintain network-wide shear was calculated as the sum of individual vessel demands. The model demonstrates that, for a given shear target, total flow requirements increase in proportion to the sum of vessel radii cubed, such that even modest global vasodilation produces a substantial increase in required cardiac output. Increasing the apparent shear target further amplifies this demand. When cardiac output is insufficient to meet these requirements, vessels experience low shear and undergo functional derecruitment, reducing total flow demand but resulting in marked heterogeneity and reduced functional capillary density. These behaviours reproduce key features of septic physiology, including the hyperdynamic circulation and microvascular shunting observed in severe sepsis. The model provides a unifying framework in which microcirculatory dysfunction emerges as an inevitable consequence of the interaction between vasodilation, flow limitation, and shear regulation, rather than as an independent pathological process. It further predicts that therapies which reduce global vasodilation or lower the apparent shear target may restore microvascular coherence without requiring supranormal cardiac output. This framework generates testable hypotheses and offers a physiologically grounded basis for reinterpreting haemodynamic management in septic shock.

BACKGROUND: Bronchoalveolar lavage fluid (BALF) metagenomic next-generation sequencing (mNGS) improves pathogen detection in severe pneumonia-related sepsis, but sampling is invasive and prone to false-positive results. Blood is easier to obtain, and broad-spectrum targeted NGS (tNGS) of microbial cell-free DNA may offer a practical alternative to BALF-based testing. We evaluated the diagnostic and prognostic value of blood-based bstNGS. METHODS: In this retrospective cohort, 122 adults with suspected severe pneumonia-related sepsis and paired BALF and blood samples underwent BALF-mNGS, blood-bstNGS and blood-mNGS. Pathogens were adjudicated using a composite clinical reference. We assessed blood-BALF concordance, compared diagnostic performance across methods, and examined whether blood-bstNGS could down-weight likely false-positive BALF-only detections and stratify prognosis. RESULTS: BALF-mNGS identified 414 microorganisms; 51% were adjudicated as causative or possibly causative, corresponding to 85.24% of patients. Among these pathogenic microorganisms, blood-bstNGS detected 45.02%, significantly more than blood-mNGS (22.27%), and nearly all pathogens detected by blood-mNGS were also detected by blood-bstNGS. Against the clinical reference, blood-bstNGS showed higher sensitivity (63.46%) than blood-mNGS (35.58%), conventional microbiological tests (CMTs) (49.04%), and blood culture (9.62%). Organisms detected only in BALF but not in blood were less likely to be classified as causative. Patients with concordant blood-bstNGS and BALF-mNGS profiles had significantly lower 30-day and 90-day mortality. CONCLUSIONS: In severe pneumonia-related sepsis, blood-bstNGS provides sensitive, non-invasive pathogen detection. It acts as a complementary tool rather than a replacement for BALF-mNGS, offering an important diagnostic alternative when BALF is unavailable and improving specificity and prognostic utility when used in combination.