Table 3.

Area under the receiver operating characteristic curve (AUC) of different olfactory domains and their sensitivity, specificity, LR+, LR–, PPV, and NPV in single olfactory models and integrated models.

Models	Olfactory tests	Sensitivity	Specificity	PPV	NPV	LR+	LR–	AUC	p1-values	p2-values
(A) SINGLE MODELS*
	TDI score	0.802	0.722	0.891	0.259	2.885	0.274	0.828	<0.05
	TD score	0.692	0.748	0.884	0.235	2.746	0.412	0.772	<0.05	<0.05
	TI score	0.796	0.730	0.894	0.262	2.948	0.279	0.819	<0.05	0.306
	DI score	0.738	0.748	0.893	0.251	2.929	0.350	0.815	<0.05	0.239
	T score	0.612	0.783	0.888	0.229	2.820	0.496	0.731	<0.05	<0.05
	D score	0.544	0.791	0.875	0.215	2.603	0.576	0.723	<0.05	<0.05
	I score	0.712	0.757	0.894	0.247	2.930	0.380	0.818	<0.05	0.481
(B) INTEGRATED MODELS**
	TDI score	0.803	0.896	0.971	0.515	7.721	0.220	0.918	<0.05
	TD score	0.846	0.835	0.956	0.559	5.127	0.184	0.902	<0.05	<0.05
	TI score	0.838	0.870	0.965	0.556	6.446	0.186	0.913	<0.05	0.298
	DI score	0.783	0.913	0.975	0.495	9.000	0.238	0.913	<0.05	0.374
	T score	0.756	0.887	0.966	0.459	6.690	0.275	0.890	<0.05	<0.001
	D score	0.707	0.922	0.975	0.424	9.064	0.318	0.886	<0.05	<0.001
	I score	0.813	0.896	0.971	0.528	7.817	0.209	0.916	<0.05	0.797

p1-values show the significance of differentiating PD patients from controls calculated by SPSS. p2-values were calculated by MedCalc software. p2-values of (A) were calculated by comparing AUC between each score with TDI score in single models, p2-values of (B) were calculated by comparing AUC between each score with TDI score in integrated models. p < 0.05 were considered significant.

^*Single models were built upon each corresponding olfactory test, after adjustment of age and educational years.

^**Integrated models were built upon each corresponding olfactory test and MMSE, NMSS, after adjustment of age and educational years.

PPV, positive predictive values; NPV, negative predictive values; LR+, positive likelihood ratios; LR–, negative likelihood ratios; T, threshold; D, Discrimination; I, Identification.