Table 3.

Binomial logistic regression estimates of sexual and reproductive and HIV-related variables on the odds of consistent condom use among women living with HIV/AIDS of reproductive age in Brazil, 2003-2004^†.

	Consistent condom use (N = 834)
	Bivariate model	Adjusted model
	OR (CI)	OR (CI)
SEXUAL AND REPRODUCTIVE VARIABLES
Method of contraception
No method	ref.	ref.
Irreversible method	0.64	0.59
	(0.45, 0.93)	(0.39, 0.90)
	P = 0.018	P = 0.013
Reversible method	0.44	0.48
	(0.28, 0.68)	(0.30, 0.77)
	P < 0.000	P = 0.002
Desire to have children
No	ref.	ref.
Yes/don't know	0.72	0.63
	(0.52, 0.99)	(0.44, 0.92)
	P = 0.043	P = 0.018
Lifetime history sexual violence
No	ref.	ref.
Yes	0.67	0.64
	(0.46, 0.96)	(0.43, 0.96)
	P = 0.031	P = 0.030
HIV-RELATED VARIABLES
HIV status of partner
HIV positive	ref.	ref.
Does not know	1.12	1.30
	(0.71, 1.77)	(0.81, 2.11)
	P = 0.632	P = 0.282
HIV negative	2.91	3.23
	(2.01, 4.21)	(2.08, 5.02)
	P < 0.000	P < 0.000
Woman on ART
Yes	ref.	ref.
No	0.53	0.56
	(0.38, 0.75)	(0.39, 0.81)
	P < 0.000	P = 0.002

Logistic regression; OR: odds ratio; CI = 95% confidence interval.

^†Adjusted models are controlled for by number of children, stable partner, history of STD in last 6 months, age, and education.

ART: antiretroviral therapy.