Artificial Intelligence Triage for Acute Kidney Injury Under Variable Creatinine Assay Standardization: A Calibration-Transfer Evaluation of Dialysis Referral Thresholds

Clinical decision support for acute kidney injury (AKI) must operate across laboratories whose creatinine assays are calibrated differently, yet the consequences of that heterogeneity for dialysis referral thresholds are rarely quantified. The framework treats assay-standardization offset as a shift in the latent severity distribution and derives the induced bias in threshold-referral rates. Using a before-after-control-impact design with an intraclass correlation of 0.64, between-cluster SD of 6.0 and residual SD of 4.5, power was estimated across 2000 replications per design cell for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The analysis indicates that 54 clusters are required to detect a 5-percentage-point effect at 80% power, and that calibration transfer degrades referral-threshold accuracy in a manner that is asymmetric across the severity distribution. The contribution is a transferable specification of when assay-standardization variation can be ignored and when it must be modelled explicitly before an AKI triage model is deployed for dialysis referral.