Table 4.

Measures of discrimination for model 1 and model 2 in each data-set

	GAP (apparent performance)	GAP (optimism adjusted)	AESOP-10 (apparent performance)	AESOP-10 (optimism adjusted)
Model 1 (including all seven predictors)
C-statistic	0.718 (95% CI 0.711 to 0.724)	0.685	0.736 (95% CI 0.730 to 0.742)	0.691
Calibration slope	0.980 (95% CI 0.942 to 1.018)	0.804	1.017 (95% CI 0.981 to 1.052)	0.779
Calibration-in-the-large	−0.0432	−0.0452	0.040	0.030
Expected/observed	1.028	1.028	0.9745	0.981
Model 2 (excluding schizophrenia diagnosis)
C-statistic	0.698 (95% CI 0.691 to 0.705)		0.639 (95% CI 0.632 to 0.646)
Calibration slope
Calibration-in-the-large	−0.036039		0.033334
Expected/observed	1.025		0.9771

GAP, Genetics and Psychosis cohort; AESOP-10, Aetiology and Ethnicity in Schizophrenia and Other Psychoses.