Psychometric Properties of the Chinese (Putonghua) Version of the Oxford Cognitive Screen (OCS-P) in Subacute Poststroke Patients without Neglect

Table 6

Results of binary logistic regression and the ROC analysis of four OCS-P subscales (with cutoff scores) for discriminating poststroke patients from the healthy older participants.

Variables/OCS-P Subscales (optimal cutoff)		SE	Wald	values	Sensitivity	Specificity	AUC	OR (95% CI)
Constant	48.65	10.87	20.04	<0.001	-	-	-	-
Picture Naming (⩽3)	−1.58	0.54	8.57	0.003	0.79	0.87	0.86	24.45 (10.08, 59.33)
Numerical Cognition (⩽6)	−3.45	1.23	7.92	0.005	0.59	0.98	0.79	84.90 (11.30, 637.67)
Praxis (⩽10)	−0.72	0.29	6.26	0.012	0.71	0.85	0.84	13.87 (6.05, 31.81)
Delayed Recall and Recognition (⩽7)	−1.51	0.56	7.36	0.007	0.71	0.92	0.83	26.93 (9.79, 74.11)

Note. Cut-off scores are based on optimal sensitivity and specificity for each subscale. SE, standard error; AUC, area under the curve; OR, odds ratio; and CI, confidence interval.