Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 20 of the 21 references it cites were found in a registry or an index; 13 of the 21 sources its claims cite were checked for support.

Machine learning model for early prediction of acute kidney injury (AKI) in pediatric critical care

Open access

Abstract

BACKGROUND: Acute kidney injury (AKI) in pediatric critical care patients is diagnosed using elevated serum creatinine, which occurs only after kidney impairment. There are no treatments other than supportive care for AKI once it has developed, so it is important to identify patients at risk to prevent injury. This study develops a machine learning model to learn pre-disease patterns of physiological measurements and predict pediatric AKI up to 48 h earlier than the currently established diagnostic guidelines. METHODS: EHR data from 16,863 pediatric critical care patients between 1 month to 21 years of age from three independent institutions were used to develop a single machine learning model for early prediction of creatinine-based AKI using intelligently engineered predictors, such as creatinine rate of change, to automatically assess real-time AKI risk. The primary outcome is prediction of moderate to severe AKI (Stage 2/3), and secondary outcomes are prediction of any AKI (Stage 1/2/3) and requirement of renal replacement therapy (RRT). Predictions generate alerts allowing fast assessment and reduction of AKI risk, such as: "patient has 90% risk of developing AKI in the next 48 h" along with contextual information and suggested response such as "patient on aminoglycosides, suggest check level and review dose and indication". RESULTS: The model was successful in predicting Stage 2/3 AKI prior to detection by conventional criteria with a median lead-time of 30 h at AUROC of 0.89. The model predicted 70% of subsequent RRT episodes, 58% of Stage 2/3 episodes, and 41% of any AKI episodes. The ratio of false to true alerts of any AKI episodes was approximately one-to-one (PPV 47%). Among patients predicted, 79% received potentially nephrotoxic medication after being identified by the model but before development of AKI. CONCLUSIONS: As the first multi-center validated AKI prediction model for all pediatric critical care patients, the machine learning model described in this study accurately predicts moderate to severe AKI up to 48 h in advance of AKI onset. The model may improve outcome of pediatric AKI by providing early alerting and actionable feedback, potentially preventing or reducing AKI by implementing early measures such as medication adjustment.

Overview

CoChat Summary

TL;DR

A single age-dependent ensemble machine learning model was developed on EHR data from 16,863 pediatric critical care patients across three institutions to predict creatinine-based AKI up to 48 h before conventional diagnosis. It predicted Stage 2/3 AKI with a median lead-time of 30 h and AUROC 0.89, while also forecasting any AKI and renal replacement therapy.

Key findings

  • The model predicted Stage 2/3 AKI before conventional criteria with a median lead-time of 30 h; AUROC reached 0.89 within the training prediction window (48–24 h before onset) and decreased slightly to 0.85 outside it.
  • It identified 70% of subsequent RRT episodes, 58% of Stage 2/3 AKI episodes, and 41% of any AKI episodes; PPV for any AKI was 47% (false-to-true alert ratio approximately 1:1), and PPV for Stage 2/3 AKI was 22% (1:4).
  • Among patients predicted by the model, 79% received potentially nephrotoxic medication after being identified but before AKI development.
  • The renal angina index comparator was less predictive, with AUROC around 0.57, sensitivity of 5% for any AKI and 3% for Stage 2/3 AKI, and PPV of 4% (1:21) and 1% (1:98), respectively.

Methods

  • Retrospective multi-center cohort using EHR data from 16,863 pediatric critical care patients aged 1 month to 21 years from three independent institutions (PICU and CTICU); each center’s data were split into derivation (70%), validation (15%), and holdout testing (15%) sets with no patient overlap.
  • A single age-dependent ensemble machine learning model used 15 final predictors plus age, selected from 250 candidate predictors; predictors were aggregated every 6 h, the model was trained on data 48–24 h before AKI onset, and tested on holdout data 48–6 h before onset. Primary outcome was Stage 2/3 AKI; secondary outcomes were any AKI and RRT requirement; the renal angina index served as baseline comparator.
  • AKI was labeled using KDIGO serum creatinine criteria; urine output criteria were not used due to unreliable records. Baseline creatinine was determined by mean normal creatinine for age and gender, and the model can make predictions when some predictors are missing without data imputation.

Limitations

  • Stated: the model uses creatinine alone to stage AKI due to data availability; up to 20% of AKI patients are diagnosed by urine output criteria alone, so the model may miss more urine-staged patients and reported PPV may increase if those patients were counted. Generalizability is not fully resolved, with a need for more validation on larger non-US datasets; the best way to present and integrate the model into workflow remains untested; and performance against admission reason and patient comorbidities is unknown.
  • Stated: the model intentionally avoided “black-box” neural network models to preserve transparency and actionability, which may have limited performance.
  • Evident from the text: validation was retrospective only, with no prospective clinical-impact study yet (the authors plan one); performance was worse on the UK hospital, likely due to its smaller data size and younger patients, and younger patients were predicted less accurately.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s13054-021-03724-0

Cites

20 of 20 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 shows an example patient’s AKI disease trajectory and uses the prediction made 36 h before KDIGO Stage 2/3 AKI onset to demonstrate the inner-workings of the model. The top plot shows the patient’s measured serum creatinine values, with AKI onset time referenced as Time 0. The middle plot shows the predicted AKI risk up to prediction time—36 h before AKI onset. It also shows a mockup of the AKI alert that a user would see, including the patient context and suggested actions. The top three predictors contributing the highest risk to this specific prediction are displayed. The bottom portion demonstrates that the model is made up of age-dependent ‘weak classifiers’ of AKI risk based on single predictor values. The predicted AKI risk is the sum of weak classifier risks of all input predictors. Two example weak classifiers are shown. The first is the classifier for creatinine rate of change (CRoC). In the top plot, the example patient’s serum creatinine increases slowly under the AKI threshold prior to prediction time. The increase results in a positive CRoC value and elevated CRoC weak classifier risk of 0.60, as marked on the CRoC classifier plot. At the same time, the patient continuously received drugs with high nephrotoxic potential, shown by triangular ticks marking times of medication administration in the top plot. This results in the high-nephrotoxic drugs classifier risk being elevated to 0.14 (bottom plot). The ellipses (…) in the figure are placeholders for additional predictors not shown due to room constraints
Fig. 1 shows an example patient’s AKI disease trajectory and uses the prediction made 36 h before KDIGO Stage 2/3 AKI onset to demonstrate the inner-workings of the model. The top plot shows the patient’s measured serum creatinine values, with AKI onset time referenced as Time 0. The middle plot shows the predicted AKI risk up to prediction time—36 h before AKI onset. It also shows a mockup of the AKI alert that a user would see, including the patient context and suggested actions. The top three predictors contributing the highest risk to this specific prediction are displayed. The bottom portion demonstrates that the model is made up of age-dependent ‘weak classifiers’ of AKI risk based on single predictor values. The predicted AKI risk is the sum of weak classifier risks of all input predictors. Two example weak classifiers are shown. The first is the classifier for creatinine rate of change (CRoC). In the top plot, the example patient’s serum creatinine increases slowly under the AKI threshold prior to prediction time. The increase results in a positive CRoC value and elevated CRoC weak classifier risk of 0.60, as marked on the CRoC classifier plot. At the same time, the patient continuously received drugs with high nephrotoxic potential, shown by triangular ticks marking times of medication administration in the top plot. This results in the high-nephrotoxic drugs classifier risk being elevated to 0.14 (bottom plot). The ellipses (…) in the figure are placeholders for additional predictors not shown due to room constraintsCC BY · PubMed Central
Fig. 2 a AUROC of the model developed in this study and the renal angina index (RAI). Model AUROC increases closer to onset time, especially inside the training window of 48 to 24 h. b The receiver-operator curve at 30 h before onset time for the model and RAI. H1, H2 and H3 represent results from the holdout test data of the three hospitals. RAI results are shown for holdout test data from all three hospitals (H1-3)
Fig. 2 a AUROC of the model developed in this study and the renal angina index (RAI). Model AUROC increases closer to onset time, especially inside the training window of 48 to 24 h. b The receiver-operator curve at 30 h before onset time for the model and RAI. H1, H2 and H3 represent results from the holdout test data of the three hospitals. RAI results are shown for holdout test data from all three hospitals (H1-3)CC BY · PubMed Central

Verification

Claims the paper makes

9 Valid · 3 Not valid · 3 Not checked · 56 Own results and interpretations

39 more claims · 2 with their cited sources checked

Create a free account to see all 71 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 15
  • Probably not a substantial claim. 14
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 21 references

Research works printed but not linked: 2

  1. American College of Physicians. Annals of Internal Medicine. The TRIPOD Statement: Explanation and Elaboration. https://annals.org/aim/fullarticle/2088542/transparent-reporting-multivariableprediction-model-individual-prognosisdiagnosis-tripod-explanation.
  2. Kaddourah A, Basu RK, Bagshaw SM, Goldstein SL. Epidemiology of acute kidney injury in critically ill children and young adults. N Engl J Med. 2017;376:11–20. · DOI · Paper page

Cited by

See all 233

Metadata

Authors
Dong, J., Feng, T., Thapa-Chhetry, B., Cho, B. G., Shum, T., Inwald, D. P., Newth, C. J. L., & Vaidya, V. U.
Affiliations
  • Philips (United States) (US)
  • Phoenix Children's Hospital (US)
  • Addenbrooke's Hospital (GB)
  • University of Southern California (US)
Corresponding author
Dong, Junzi
Venue
Critical Care
Published
2021-08-10
Volume
25
Issue
1
Pages
288-288
ISSN
1364-8535 (ISSN-L)
OpenAlex
W3191280566
Cited by
233
References
21
Citations by year
2026: 40 · 2025: 75 · 2024: 65 · 2023: 33 · 2022: 20
FWCI
20.17
Type
journal article
Language
en
Open access
Licence
Locations
  • Critical Care (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Health Sciences › Medicine › Nephrology › Acute Kidney Injury Research
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
acute kidney injury, pediatric intensive care, early prediction, machine learning, serum creatinine, renal replacement therapy, clinical decision support alerts, nephrotoxic drugs, aminoglycosides, AKI risk stratification, electronic health records
MeSH terms
Machine Learning, Adolescent, Child, Child, Preschool, Computer Simulation, Critical Care, Female, Humans, Infant, Infant, Newborn, Male, Pediatrics, ROC Curve, Severity of Illness Index, Intensive Care Units, Pediatric, Cohort Studies, Area Under Curve, Young Adult, Acute Kidney Injury

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Machine learning model for early prediction of acute kidney injury (AKI) in pediatric critical care. https://app.cochat.ai/papers/10.1186/s13054-021-03724-0
The paper
Dong, J., Feng, T., Thapa-Chhetry, B., Cho, B. G., Shum, T., Inwald, D. P., Newth, C. J. L., & Vaidya, V. U. (2021). Machine learning model for early prediction of acute kidney injury (AKI) in pediatric critical care. Critical Care. https://doi.org/10.1186/s13054-021-03724-0

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.