Table 3.

Multivariable logistic regression, model-1, model-2 and model-3. Model performance measurements.

Variables	Odds-ratio and CI	P value	C-index	Likelihood ratio	BIC	p interaction
Model-1			0.830	51.73	188.2	0.002
PI-WSUR total (from 20 to 49)	3.45 (1.62–7.34)	0.04
Admission-month (from first month to third month)	0.33 (0.11–0.97)	0.01
Model-2			0.855	58.31	181.6	<0.001
PI-WSUR-C-OCD (from 8 to 27.5)	5.36 (2.11–13.61)	0.01
Admission-month (from first month to third month)	0.26 (0.09–0.87)	<0.001
Model-3			0.856	59.30	190.6	<0.001
PI-WSUR C-OCD (from 8 to 27.5)	4.87 (1.74–13.61)	0.02
Beck-depression inventory (from 2 to 9)	1.13 (0.46–2.79)	0.77
Beck-anxiety inventory (from 2 to 6)	1.28 (0.74–2.20)	0.36
Admission-month (from first month to third month)	0.32 (0.10–0.98)	<0.001

Data are given in Odds-Ratio (95%CI). Multivariable Logistic regression analyses was used All 3 models adjusted Age, Gender, Diabetes Mellitus, Hypertension, Smoking hystory, Pain time (daytime (from 8 a.m. to 6 p.m.) or at night (from 6 p.m. to 8 a.m.) and Coronary artery disease history. p value for odds-ratio; p interaction for model1- PI-WUR-total scores and admission month interaction p-value, for model 2–3 interaction between month and PI-WUR-contamination scores and admission-month interaction p-value. Abrevations: BIC: Bayesian information criteria (lower value depict better model), Harrel c-index statistics measures the discriminative ability of the model, and values closer to 1.0 are better. Likelihood Ratio X2 higher value represent better model performance. PI-WSUR:Padua inventory-Washington university revision, C-OCD:Contamination obsessive compulsive disorder.