Table 3.

Comparisons between the predictive power of the new developed risk estimate models versus APACHE-II, SAPS-II, SOFA and MOSAIC scores in the complete case CRRT population.

ICU mortality				Hospital mortality
Prediction model	AUC	Contrast estimate (95% CL)	P-value	Prediction model	AUC	Contrast estimate (95% CL)	P-value
CRRT patients n = 462
ICU admission model	0.79			ICU admission model	0.76
SOFA admission	0.57	− 0.22 (− 0.29 to − 0.16)	< 0.0001	SOFA admission	0.57	− 0.18 (− 0.25 to − 0.12)	< 0.0001
APACHE-II admission	0.62	− 0.17 (− 0.23 to − 0.11)	< 0.0001	APACHE-II admission	0.60	− 0.15 (− 0.22 to − 0.09)	< 0.0001
SAPS-II admission	0.65	− 0.14 (− 0.20 to − 0.09)	< 0.0001	SAPS-II admission	0.65	− 0.11 (− 0.16 to − 0.06)	< 0.0001
CRRT patients n = 408
The MALEDICT RRT initiation model	0.82			The MALEDICT RRT initiation model	0.78
SOFA RRT initiation	0.69	− 0.13 (− 0.19 to − 0.06)	0.0001	SOFA RRT initiation	0.67	− 0.11 (− 0.17 to − 0.05)	0.0003
APACHE-II RRT initiation	0.68	− 0.14 (− 0.19 to − 0.08)	< 0.0001	APACHE-II RRT initiation	0.68	− 0.10 (− 0.16 to − 0.05)	0.0004
SAPS-II RRT initiation	0.63	− 0.19 (− 0.26 to − 0.12)	< 0.0001	SAPS-II RRT initiation	0.61	− 0.17 (− 0.24 to − 0.10)	< 0.0001
MOSAIC RRT initiation	0.69	− 0.13 (− 0.19 to − 0.07)	< 0.0001	MOSAIC RRT initiation	0.63	− 0.15 (− 0.21 to − 0.09)	< 0.0001
IHD patients n = 193
The best IHD ICU Admission model	0.82			0.81
The best IHD initiation model	0.80			0.81
The CRRT ICU Admission model	0.78			0.80
The MALEDICT RRT initiation model	0.74			0.77

Values for the ICU admission and RRT initiation models and comparisons between prediction models are mean values for 25 multiple imputations.

CL confidence limit, ICU intensive care unit, SOFA-score Sequential Organ Failure Assessment score, APACHE Acute Physiology and Chronic Health Evaluation II score, SAPS Simplified Acute Physiology II score, CRRT continuous renal replacement therapy, RRT renal replacement therapy.

Significant values are in bold.