Table 4.

Multivariate analyses of correlates of caregiver burden scores: stepwise regression analyses: final regression model

Dependent variables and significantly associated variables	%Variance(R2)	Standardized beta	T	P level	Tolerance*	VIF*
Dependent variable: total IEQ objective burden
Education of caregiver	6.5	-0.17	1.7	0.09	0.91	1.1
Age of patient	6.6	-0.27	2.9	0.005	0.98	1.02
Patient's general well-being (subjective QOL: LQLP)**	7.7	-0.29	3.0	0.004	0.92	1.09
Patient "been out shopping" (objective QOL: LQLP)	14.2	0.38	4.0	0.001	0.98	1.00
	total = 34.9
Dependent variable: global impression subjective burden
Education of caregiver	9.3	0.25	2.48	0.15	0.93	1.07
Family income	5.0	0.18	1.87	0.066	0.97	1.03
Patient's negative affect	7.3	0.25	2.58	0.01	0.98	1.03
Patient's general well-being (LQLP)	4.9	-0.22	2.26	0.027	0.94	1.07
Patient "been out for a ride"	3.9:	0.19	2.07	0.04	0.99	1.01
	total = 30.4
Dependent variable: IEQ supervision subscale score
Patient's basic unmet need (CAN***)	5.0	-0.16	1.56	0.12	0.83	1.20
Patient's health domain QOL	5.5	-0.23	2.26	0.027	0.89	1.13
Patient's positive affect	4.9	0.17	1.60	0.114	0.80	1.25
Patient "been out shopping"	7.6	0.32	3.18	0.002	0.92	1.09
	total = 25.2

* Tolerance (</= 0.2) and its reciprocal, VIF (>/= 4), are tests of multicollinearity between dependent and independent variables.

** LQLP: Lancashire Quality of Life Profile

*** CAN: Camberwell Assessment of Need

The independent variables entered were: (i) Step 1: caregiver and patient socio-demographic characteristics; (ii) Step 2: Patient's CAN and VSSS scores; (iii) Step 3: Patient's LQLP subjective scores, affect and self-esteem scores; (iv) Step 4: LQLP objective indices; (v) Step 5: patient's psychopathology (ICD-10 positive and negative syndrome) and GAF scores