Table 2.

ROC analysis for identifying MCI using MMSE, CDT, and AQT in the training and validation samples

Total number of tests	Training sample (n = 326)				Validation sample (n = 140)
	Cut-off*	Sensitivity (%)	Specificity (%)	Youden index	Sensitivity (%)	Specificity (%)	Youden index
1	MMSE <  28	59.1	79.3	0.38	72.7	60.3	0.33
1	AQT > 83 s	54.7	73.8	0.28	68.8	49.2	0.18
1	CDT < 4 p	33.7	85.5	0.19	35.1	85.7	0.21
2	MMSE < 28 AND AQT > 52	58.0	80.7	0.39	40.3	87.3	0.28
2	MMSE < 27 OR AQT > 91	65.8	79.3	0.45	55.8	77.8	0.34
2	MMSE < 30 AND CDT < 5	48.6	75.2	0.24	42.9	82.5	0.25
2	MMSE < 28 OR CDT < 4	69.6	70.3	0.40	63.6	68.2	0.32
3	MMSE < 27 OR AQT > 91 OR CDT <  2	66.3	79.3	0.46	57.1	77.8	0.35

^*Cut-offs were established in the training sample according to the highest Youden index and tested in the validation sample. MMSE, Mini-Mental State Examination; AQT, A Quick Test of Cognitive Speed; CDT, Clock Drawing Test.