Beyond binary acute kidney injury (AKI) classification: development and external validation of a distributional model predicting serum creatinine and urine output trajectories in intensive care units (ICU) patients
Authors' conclusion
This multi-step, multivariate distributional regression model is a reliable, more flexible, transparent, and clinically interpretable approach for AKI prediction compared to traditional classification methods.
Study context
- Population
- Adult ICU stays; training MIMIC-IV, external validation MIMIC-IV, UZGent, eICU and AmsterdamUMCtraining cohort
- Participants
- 4,118 participants
- Comparison
- Distributional regression model vs Tree-based classifiers
- Outcome
- Prediction of KDIGO stages 1-3 AKI and persistent stage 3 AKI
- Follow-up
- 24-hour ahead
Methodological detailsmulticentre, external validation in four independent cohorts, distributional regression
- Multicentre
- External validation in four independent cohorts
- Distributional regression
Performance
As reported in the abstract
AKI prediction, all stages, external cohorts
| Test or model | Value | AUC-PR |
|---|---|---|
| Distributional regression model | mean 0.774 | 0.774 |
Distributional regression model
Valuemean 0.774
AUC-PR0.774
Further findings
External validation cohorts (n)
- MIMIC-IV
- 3,838
- UZGent
- 4,442
- eICU
- 10,760
- AmsterdamUMC
- 6,129
Original publication
Beyond binary AKI classification: development and external validation of a distributional model predicting serum creatinine and urine output trajectories in ICU patients. Crit Care. Published 10 May 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →