Accurate and interpretable prediction of ICU-acquired AKI
Authors' conclusion
The two AKI prediction models have good discriminative performance using common features, which can aid in accurately and informatively monitoring AKI risk in ICU patients.
Study context
- Population
- Adult ICU admissions
- Participants
- 50,342 vs 39,087
- Exposure
- any-AKI model vs moderate-to-severe AKI model
- any-AKI modeln = 50,342
- Moderate-to-severe AKI modeln = 39,087
- Outcome
- Prediction of time of diagnosis of any-AKI or moderate-to-severe AKI (AUROC)
Methodological detailssingle-centre, prospective validation
- Single-centre
- Prospective validation
Main finding
As reported in the abstract
AUROC before diagnosis
any-AKI model
0.756
AUROC
Moderate-to-severe AKI model
0.721
AUROC
Observational data: associations, not effects.
Further findings
| Outcome | any-AKI model | Moderate-to-severe AKI model |
|---|---|---|
| PPV, prospective validation | ||
| PPV | 0.796 | 0.546 |
- Triggers per 8 h, AKI vs no AKI, any-AKI model (prospective): AKI median 1.82 (IQR 0–4.71) triggers per 8 h
- Triggers per 8 h, AKI vs no AKI, any-AKI model (prospective): No AKI median 0 (IQR 0–0.73) triggers per 8 h
- Triggers per 8 h, AKI vs no AKI, moderate-to-severe AKI model (prospective): AKI median 2.35 (IQR 0.14–4.96) triggers per 8 h
- Triggers per 8 h, AKI vs no AKI, moderate-to-severe AKI model (prospective): No AKI median 0 (IQR 0–0.8) triggers per 8 h
Original publication
Accurate and interpretable prediction of ICU-acquired AKI. J Crit Care. Published 10 February 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →