Development and validation of the STeP score for predicting tracheostomy in patients with sepsis using a nationwide intensive care unit (ICU) database: a retrospective observational study
Authors' conclusion
We developed and validated a robust prediction model and simplified risk score (STeP score) for tracheostomy within 14 days in ICU patients with sepsis.
Study context
- Population
- Adults with sepsis (SOFA ≥ 2, excluding viral pneumonia) requiring mechanical ventilation in Japanese ICUs, 2018–20221,013 (13.8%) tracheostomy
- Participants
- 7,357 adults
- Exposure
- Training vs Validation
- Trainingn = 5,374
- Validationn = 1,983
- Outcome
- Tracheostomy
- Follow-up
- Within 14 days of ICU admission
Methodological detailsprediction model development and validation, nationwide, ICU database (JIPAD), least absolute shrinkage and selection operator (LASSO) regression with tenfold cross-validation, temporally independent validation cohort
- Prediction model development and validation
- Nationwide
- ICU database (JIPAD)
- LASSO regression with tenfold cross-validation
- Temporally independent validation cohort
Main finding
As reported in the abstract
Discrimination of the STeP model (8 LASSO-selected variables) and STeP score
| Group | AUC (STeP model) | AUC (STeP score, range 0-17) |
|---|---|---|
| Training | 0.76 | — not stated in the abstract |
| Validation | 0.74 | 0.73 |
AUC (STeP model)
Training0.76
Validation0.74
AUC (STeP score, range 0-17)
Training— not stated in the abstract
Validation0.73
— not stated in the abstract
Observational data: associations, not effects.
Further findings
Tracheostomy rate by STeP score risk group
Original publication
Development and validation of the STeP score for predicting tracheostomy in patients with sepsis using a nationwide ICU database: a retrospective observational study. J Intensive Care. Published 14 November 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →