Development and Validation of a Real-time Prediction Model for Acute Kidney Injury in Hospitalized Patients

Development and Validation of a Real-time Prediction Model for Acute Kidney Injury in Hospitalized Patients

Early prediction of acute kidney injury (AKI) may provide a crucial opportunity for AKI prevention. To date, no prediction model targeting AKI among general hospitalized patients in developing countries has been published. We developed a simple, real-time, interpretable AKI prediction model for general hospitalized patients from a large tertiary hospital in China, and validated it across five independent, geographically distinct, different tiered hospitals.

Early prediction of acute kidney injury (AKI) may provide a crucial opportunity for AKI prevention. To date, no prediction model targeting AKI among general hospitalized patients in developing countries has been published. We developed a simple, real-time, interpretable AKI prediction model for general hospitalized patients from a large tertiary hospital in China using the machine learning technique, and then validated the performance of the prediction model across five independent, geographically distinct, different tiered hospitals in China.

Inclusion Criteria:

• Adult patients (18 years and older) admitted to five hospitals during the study period

Exclusion Criteria:

• Have less than 2 documented serum creatinine (Scr) measurements during hospitalization
• Being diagnosed with end-stage renal disease (ESRD)
• Maintained on dialysis or had an initial Scr greater than or equal to 4.0 mg/dL at admission
• Developed AKI prior to admission or within 24 hours after admission
• Length of stay shorter than 24 hours
• Underwent kidney transplantation or nephrectomy during hospitalization
• With all Scr measurements lower than or equal to 0.6 mg/dL from 90 days prior to admission until discharge.