Pulse.

a daily field guide to health research that matters

◆ Console

‹ Sun · 26 Apr 2026
Near-term implementable finding

Association between platelet-albumin-bilirubin score and 30-day in-hospital mortality in patients with sepsis: evidence from a large database and machine learning modeling.

A simple blood test score from routine labs predicted sepsis survival risk, offering doctors a practical tool to identify high-risk patients earlier.

This MIMIC-IV-based study evaluated the platelet-albumin-bilirubin (PALBI) score as a sepsis mortality predictor, using machine learning to develop and validate a prognostic model for 30-day in-hospital mortality. The study contributes to the evidence base for CBC-accessible composite biomarkers in critical care risk stratification, with potential for near-term clinical application given the routine availability of the component laboratory tests.

What the study was

Study design
Retrospective database study + ML model development and validation
Population
Sepsis patients in MIMIC-IV database (pediatric ICU, Chongqing Medical University Children's Hospital)
Category
Diagnostics
Maturity
Exploratory
Journal
BMC Infectious Diseases

Why it surfaced

MIMIC-IV database + ML validation provides reasonable evidence level for a CBC-component composite score in sepsis ICU; sepsis mortality is a high-impact endpoint; scores derived from routine labs (platelets, albumin, bilirubin) are inherently implementable if validated.

A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.