Table 3. ACC Admission and Discharge Prediction Model’s Performance Measures.

Dataset	Performance Measures	Admission Model	Discharge Model
Derivation (n=94,859)	Discrimination
	C-statistic	0.76	0.78
	Calibration
	Hosmer-Lemeshow goodness-of-fit test (p-value)	36.0 (p<0.001)	31.1 (p<0.001)
	Overall Performance
	Brier Score (% improvement)	0.062 (7.6%)	0.060 (9.1%)
	Bootstrapping
	500 simulations average (min. to max.)	0.76 (0.75 to 0.76)	0.78 (0.77 to 0.78)
Internal Validation (n=31,619)	Discrimination
	C-statistic	0.75	0.77
	Calibration
	Hosmer-Lemeshow goodness-of-fit test (p-value)	23.5 (p=0.003)	19.9 (p=0.01)
	Overall Performance
	Brier Score (% improvement)	0.063 (6.6%)	0.061 (9.1%)
External Validation (Without Recalibration) (n=6,357)	Discrimination
	C-statistic	0.69	0.71
	Calibration
	Hosmer-Lemeshow goodness-of-fit test (p-value)	216.9 (p<0.001)	156.3 (p<0.001)
	Overall Performance
	Brier Score (% improvement)	0.065 (2.5%)	0.064 (4.0%)
External Validation (With Recalibration) (n=6,357)	Discrimination
	C-statistic	0.76	0.78
	Calibration
	Hosmer-Lemeshow goodness-of-fit test (p-value)	6.1 (p=0.641)	14.3 (p=0.074)
	Overall Performance
	Brier Score (% improvement)	0.061 (8.9%)	0.060 (9.1%)