. Author manuscript; available in PMC: 2017 Aug 1.

Published in final edited form as: Am Heart J. 2016 May 17;178:45–54. doi: 10.1016/j.ahj.2016.05.004

Table II.

AF prediction performance for the CHARGE-AF and the CHA₂DS₂-VASc risk scores

Risk score	Wald χ²	Sub-distribution hazard ratio^* (95% CI)	P	C-statistic (95% CI)	Calibration HL χ^2†
Primary analysis, n observations = 9722
CHARGE-AF	403	2.15 (1.99, 2.31)	<.0001	0.757 (0.741, 0.762)	5.6 (P = .69)
CHA₂DS₂-VASc	209	1.43 (1.37, 1.51)	<.0001	0.712 (0.693, 0.731)	28.5 (P < .0001)
CHA₂DS₂-VASc lone AF women	288	1.45 (1.39, 1.51)	<.0001	0.730 (0.713, 0.747)	35.5 (P < .0001)
CHA₂DS₂-VASc no sex	360	1.56 (1.49,1.63)	<.0001	0.741 (0.724, 0.758)	28.5 (P < .0001)
Sensitivity analysis, n observations = 4797
CHARGE-AF	208	1.90 (1.74, 2.07)	<.0001	0.713 (0.693, 0.733)	12.7 (P = .12)
CHA₂DS₂-VASc	89	1.34 (1.26, 1.42)	<.0001	0.673 (0.652, 0.695)	12.67 (P = .03)
CHA₂DS₂-VASc lone AF women	125	1.36 (1.29, 1.43)	<.0001	0.684 (0.663, 0.705)	19.8 (P = .001)
CHA₂DS₂-VASc no sex	173	1.46 (1.38, 1.54)	<.0001	0.699 (0.678, 0.720)	10.6 (0.06)

Sub-distribution hazard ratios are per unit change in respective scores. HL, Hosmer-Lemeshow.