Table 5. Integrated discrimination improvement (IDI) and category-free net reclassification improvement (cfNRI) for [TIMP-2]*[IGFBP7].

Statistic	Value	p-value	95%-CI
IDI	0.207	0.0015	(0.079–0.336)
cfNRI	1.513	<0.001	(1.122–1.905)
AUC (reference model)	0.910	<0.001	(0.82–0.99)
AUC (new model)	0.967	<0.001	(0.92–1.00)
AUC (difference)	0.907	<0.001	(0.80–1.00)

IDI and cfNRI refer to the improved classification of patient using the reference model and the new model (reference model+[TIMP-2]*[IGFBP7]). AUC refers to the area under the receiver operator characteristic curve calculated from the model prediction against the outcome (AKI vs. no AKI). AUC difference refers to the ROC analysis using the difference between predicted probabilities for each patient. AUCs have been calculated in SPSS. IDI and cfNRI have been calculated in R using package PredictABEL. P-values and 95% confidence intervals (CI) for IDI and cfNRI are provided by the function reclassification and are determined by a z-score transformation.