Identify Sepsis-3 patients using SOFA >= 2 and suspected infection criteria in ICU datasets.
npx skills add https://github.com/hannesill/m4 --skill sepsis-3-cohort
Please help me install this Agent Skill. Skill: sepsis-3-cohort Source: https://github.com/hannesill/m4/tree/main/src/m4/skills/clinical/sepsis-3-cohort Command: npx skills add https://github.com/hannesill/m4 --skill sepsis-3-cohort
Identify Sepsis-3 patients using the consensus definition (SOFA >= 2 + suspected infection) to enable cohort studies, outcome research, and quality metrics in ICU datasets.
Run the sepsis-3 extraction pipeline to identify ICU patients with suspected infection and SOFA >= 2 within the defined window around onset.
High-intent search queries and answers about installing and using this skill.
Sepsis-3 cohort selection combines suspected infection criteria with SOFA scoring. It flags ICU patients who have both antibiotic and culture events within a designated window and a SOFA score increase of 2 or more around the infection onset.
The SOFA criteria window requires a SOFA score of 2 or higher measured within 48 hours before to 24 hours after suspected infection onset. This timing window ensures accurate temporal alignment for sepsis-3 cohort selection in ICU datasets.
Yes, the Sepsis-3 extraction approach is cross-dataset applicable and works with eICU. It includes specific implementation guidance and notes on handling data differences between eICU and MIMIC-IV for accurate ICU cohort selection.
Limitations of using SOFA scores for sepsis cohort selection include ICU-specific data requirements, baseline SOFA assumptions, and strict timing windows. The implementation enforces these constraints and provides documentation on known limitations for outcome studies.
To define suspected infection for an ICU sepsis cohort, identify patients who received both antibiotics and cultures within a designated time window. You then evaluate their SOFA score changes to confirm they meet the Sepsis-3 consensus definition.
hannesillDetect suspected infection events by linking antibiotic administrations to culture timing.
hannesillCompute APS III scores from first-day ICU data to predict hospital mortality.
galeepFit Cox proportional hazards models to censored survival data with scikit-survival.
langcareAnalyzes FHIR patient data to compute qSOFA, SOFA, and SIRS sepsis risk scores.
