. Author manuscript; available in PMC: 2020 Jan 1.

Published in final edited form as: Personal Disord. 2019 Jan;10(1):46–58. doi: 10.1037/per0000300

Table 2.

Performance of Three Risk Models of Four-Year MCI Incidence

	A	B	C
Overall Fit
Nagelkerke / Royston R²	.164 (.053, .297)	.247 (.113, .384)	.374 (.226, .502)
Scaled Brier Score	.278 (.239, .313)	.226 (.164, .280)	.375 (.326, .417)
Discrimination
AUC	.647 (.576, .718)	.680 (612,.748)	.741 (.677, .805)
Discrimination R²	.164 (.053, .297)	.247 (.113, .384)	.374 (.226, .502)
Calibration
Calibration in the Large	.101 (.090, .112)	.155 (.040, .171)	.002 (−.013, .01)
Hosmer-Lemeshow Statistic	18.90 (p = .0003)	41.89 (p < .0001)	3.11 (p = .742)

Notes: Measures of model performance with 95% Confidence Intervals in parentheses for Table 2 models. AUC = Area Under the Curve. Model A = demographics only, Model B = Model A + health factors, Model C = Model B + personality items/ facets.