Skip to content

Large Language Model-Based Real-time Acute Kidney Injury Prediction with Explainable Risk Attribution: A Multi-Center Development and Validation Study

Large Language Model-Based Real-time Acute Kidney Injury Prediction with Explainable Risk Attribution: A Multi-Center Development and Validation Study

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600119799
Enrollment
Unknown
Registered
2026-03-03
Start date
2025-08-22
Completion date
Unknown
Last updated
2026-03-09

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Acute kidney injury

Interventions

Observation group:None

Sponsors

Peking University First Hospital
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to 100 Years

Inclusion criteria

Inclusion criteria: Based on the structured database of in-hospital medical information in 4 centers, visits that met all the following inclusion criteria were selected, ranked and analyzed in terms of visits: 1) admission between January 1,2018, and December 31,2020; 2) perform>=2 serum creatinine tests during a single visit.

Exclusion criteria

Exclusion criteria: Visits that met the inclusion criteria were exported from a structured database and computer-excluded visits that met the following criteria: 1) < 18 years; 2)History of chronic kidney disease and maintenance hemodialysis; 3) kidney-related surgery; 4) length of hospital stay < 24 h; 5) prehospital Aki; 6) peak serum creatinine < 53 µmol/L from 90 days before admission to discharge; 7) the first visit with creatinine above 353.6 µmol/L after admission.

Design outcomes

Primary

MeasureTime frame
Accuracy;Sensitivity;Specificity;Positive predictive value;Negetive predictive value;

Secondary

MeasureTime frame
Mean Time to inference;Case understanding scoring;Medical guidelines and consensus scoring;Clinical reasoning score;Correlation of risk factors scoring;Acceptability of treatment recommendations scoring;Content infidelity score;Bias and unfairness scoring;harm scoring;

Countries

China

Contacts

Public ContactLi Yang

Peking University First Hospital

li.yang@bjmu.edu.cn+86 10 8357 2200

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Mar 14, 2026