Related Experiment Video
Updated: Jun 22, 2025
07:30
Design of Cecal Ligation and Puncture and Intranasal Infection Dual Model of Sepsis-Induced Immunosuppression
Published on: June 15, 2019
10.0K
Development and Validation of the Hospital Medicine Safety Sepsis Initiative Mortality Model
Hallie C Prescott1, Megan Heath2, Elizabeth S Munroe2
1Department of Internal Medicine, University of Michigan, Ann Arbor, MI; VA Center for Clinical Management Research, Ann Arbor, MI.
Chest
|July 4, 2024
Summary
A new sepsis mortality model accurately predicts risk and improves fair hospital comparisons. This risk-adjusted model reclassified one-third of hospitals, enhancing performance assessments.
Area of Science:
- Medical Informatics
- Clinical Epidemiology
- Health Services Research
Background:
- Fair comparison of hospital performance in sepsis care requires accounting for patient case mix.
- The Michigan Hospital Medicine Safety sepsis initiative (HMS-Sepsis) aimed to develop a risk-adjustment model for 30-day mortality.
Purpose of the Study:
- To assess if HMS-Sepsis registry data can adequately predict risk-adjusted 30-day mortality.
- To determine if performance assessments differ when using adjusted versus unadjusted data.
Main Methods:
- Retrospective cohort analysis of community-onset sepsis hospitalizations (April 2022-September 2023).
- Development and validation of a risk-adjustment model using physiologic, demographic, and baseline health data.
- Model performance evaluated using C statistics, Brier scores, and predicted vs. observed mortality comparisons.
Main Results:
- The final model included 13 physiologic and 16 demographic/chronic health variables; key predictors were age, metastatic solid tumor, temperature, altered mental status, and platelet count.
- The model demonstrated strong discrimination (C-statistic 0.81-0.82) and adequate calibration (0.0% overall error).
- Risk adjustment reclassified 33.9% of hospitals compared to unadjusted mortality assessments.
Conclusions:
- The HMS-Sepsis mortality model exhibits robust predictive performance for 30-day sepsis mortality.
- This validated model can facilitate equitable hospital benchmarking and performance evaluation.
- The model supports accurate assessment of temporal trends and observational causal inference in sepsis care.
