Acute Kidney Injury
Conditions
Keywords
Acute kidney injury, Artificial intelligence, Explainable AI, Prediction
Brief summary
The goal of this clinical trial is to compare the occurrence of acute kidney injury (AKI) in inpatients when information from the 'PRIME solution' (AKI prediction program utilizing artificial intelligence) is provided. The main questions it aims to answer are: •\[When Artificial intelligence (AI) provides information regarding AKI occurrence prediction within 48 hours, what would change in the physician's behavior?\] •\[If provided with AI information, what would be the incidence of AKI, severe AKI (stage 2 or 3), kidney replacement therapy, and changes in mortality during hospitalization?\] In the case of the intervention group that receives AI information, autonomous treatment is conducted by referring to AI prediction information. Researchers will compare it with a usual-care group that does not receive AI prediction results.
Detailed description
The investigators designed a prospective, investigator-initiated, single-center, single-blinded, randomized controlled study with two experimental groups. A total of 1438 participants with hospitalized patients will be enrolled and randomized into two groups; intervention or usual-care groups. The investigators will apply an AKI prediction model based on the patient's demographic. The investigators will collect data on vital signs, laboratory test results, medication history, and surgical records. The investigators provide information on whether AKI develops within 48 hours and the top 10 explanatory factors for predicting AKI to the physician. The intervention group receives the prediction results daily until the patient is discharged or up to 7 days after admission. The usual-care group does not receive analysis results.
Interventions
OTHERPRIME solution
The investigators will apply an AKI prediction model (PRIME solution) based on the patient's demographic. The investigators will collect data on vital signs, laboratory test results, medication history, and surgical records. The investigators provide information on whether AKI develops within 48 hours and the top 10 explanatory factors for predicting AKI to the physician.
Sponsors
Seoul National University Hospital
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
PREVENTION
Masking
SINGLE (Subject)
Eligibility
Sex/Gender
ALL
Age
19 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Inpatients aged \> 18 years * Admitted to the Department of Nephrology, Geriatrics, Urology, Surgery, and Orthopedics * written consent
Exclusion criteria
* Patients with a scheduled hospital stay of 2 days or less * Patients who have been hospitalized for more than 8 days since the date of hospitalization * Patients scheduled to be discharged the next day * Patients with last measured serum creatinine greater than 4.0 mg/dL or less than Estimated Glomerular Filtration Fate Chronic Kidney Disease Epidemiology Collaboration (eGFR-CKD-EPI) 15 ml/min/1.73 m2 * Patients undergoing dialysis (Hemodialysis, peritoneal dialysis) due to end-stage renal disease * Declined to participate
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| physician's behavior change | From the date of radomization until the patient is discharged or date of death from any cause, whichever came first, assessed up to 8 days after admission | Physician's pattern of care changed after receiving AI prediction information The primary category of outcomes was the change in clinicians' behavior, including clinicians' medical treatment patterns, instances of overlooked AKI, and clinicians' consultation patterns. Medical treatment patterns of clinicians were classified into the following 5 categories: Patient assessment, Review of medication, Imaging workup, Monitoring hemodynamic stability, Evaluate furthermore. Overlooked AKI was defined as the absence of a follow-up Scr measurement within 2 weeks after AKI. Consultation to the nephrology division was classified into the following 3 outcomes: no consultation, early consultation, and late consultation. Early consultation was defined as consultation within 3 days from the AKI prediction. Consults issued more than 3 days after the AKI prediction were considered late consultations. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| incidence of AKI | From the date of randomization until the patient is discharged or date of death from any cause, whichever came first, assessed up to 8 days after admission | AKI is defined as any of the following (NotGraded): * Increase in SCr by ≥ 0.3mg/dl (≥26.5umol/l) within 48 hours;or * Increase in SCr to ≥1.5 times baseline, which is known or presumed to have occurred within the prior 7days; or * Urine volume \< 0.5ml/kg/h for 6hours. |
| severe AKI (stage 2 or 3) | From the date of randomization until the patient is discharged or date of death from any cause, whichever came first, assessed up to 8 days after admission | stage 2: Serum creatinine (SCr) 2.0-2.9 times baseline or Urine output \< 0.5ml/kg/h for ≥12hours stage 3: Serum creatinine 3.0 times baseline or increase in serum creatinine to ≥ 4.0 mg/dl or initiation of renal replacement therapy or Urine output \< 0.3ml/kg/h for ≥24hours or Anuria for ≥ 12 hours |
| renal replacement therapy | From the date of randomization until the patient is discharged or date of death from any cause, whichever came first, assessed up to 8 days after admission | Patients who underwent renal replacement therapy |
| death | From the date of randomization until the patient is discharged or date of death from any cause, whichever came first, assessed up to 8 days after admission | patient who died during hospitalization |
Countries
South Korea
Contacts
Primary ContactSejoong Kim, MD, PhD
Backup ContactGiae Yun, MD
Outcome results
None listed