Table 3.

Net reclassification index (NRI) and integrated discrimination index (IDI) of Cox and Gradient Boosted Tree (GBT) modeling approaches for predicting 3-year risk of stroke.

Modeling approach and included data	Men			Women
	ΔAUC (p-value)	NRI [95%CI]	IDI [95%CI]	ΔAUC (p-value)	NRI [95%CI]	IDI [95%CI]
Single measurement of risk factor inputs
Cox: Single measurement at most recent visit	Referent	Referent	Referent	Referent	Referent	Referent
GBT: Single measurement at most recent visit	+ 0.032 (p = 0.08)	+ 0.14 [−0.17, 0.44]	−0.03 [−0.08, 0.01]	−0.013 (p = 0.59)	−0.25 [−0.51, 0.01]	−0.03 [−0.06, −0.01]
Sequential measurements of risk factor inputs
GBT: Sequential measurements at three visits	+ 0.016 (p = 0.57)	−0.20 [−0.52, 0.12]	−0.03 [−0.09, 0.02]	−0.015 (p = 0.42)	−0.13 [−0.39, 0.14]	−0.03 [−0.06, −0.01]
GBT: Longitudinal summary a of sequential measurements at three visits	+ 0.010 (p = 0.77)	−0.13 [−0.45, 0.18]	−0.05 [−0.11, 0.01]	−0.032 (p = 0.14)	−0.29 [−0.55, −0.03]	−0.04 [−0.06, −0.01]
GBT: Longitudinal summary of stroke risk estimates b at three visits + Single measurement at most recent visit	+ 0.007 (p = 0.38)	−0.02 [−0.34, 0.30]	0.00 [−0.02, 0.02]	−0.006 (p = 0.32)	+ 0.15 [−0.11, 0.42]	0.00 [−0.01, 0.02]

^a Longitudinal summary of risk factor inputs included (i) mean, standard deviation, minimum, and maximum values recorded for continuous risk factor inputs, and (ii) mean and standard deviation for binary risk factor inputs.

^b Longitudinal summary of risk estimates included individual-level random slopes and random intercepts from linear mixed effects models; variance of Cox-generated risk estimates across three visits; and absolute changes in Cox-generated risk estimates between consecutive visits.