Table 4.

Regression model of perceived intervention support, by study arm at first and second follow-up using logistic regression analysis

	Unadjusted intervention (vs. control)		Adjusteda intervention (vs. control)
	OR (95% CI)	P-value	aOR (95% CI)	P-value
Talking with provider made positive differenceb
Follow-up 1
Agree a lot (vs. agree somewhat or disagree)	2.83 (1.78, 4.52)	< 0.001	2.76 (1.72, 4.42)	< 0.001
Follow-up 2
Agree a lot (vs. agree somewhat or disagree)	2.95(1.89, 4.60)	< 0.001	2.89 (1.84, 4.41)	< 0.001
Learned new things about a woman’s rights in her relationshipb
Follow-up 2c
Agree a lot (vs. agree somewhat or disagree)	3.79 (2.32, 6.19)	< 0.001	3.72 (2.27, 6.10)	< 0.001
Feel better able to take care of health than before visitd
Follow-up 1
Better (vs. same)	3.83 (2.03, 7.22)	< 0.001	4.00 (2.08, 7.71)	< 0.001
Follow-up 2
Better (vs. same)	1.23 (0.50, 3.03)	0.643	1.15 (0.47, 2.85)	0.754
Feel more confident in how deserve to be treatedd
Follow-up 1
Better (vs. same)	4.96 (3.30, 7.44)	< 0.001	5.03 (3.33, 7.86)	< 0.001
Follow-up 2
Better (vs. same)	2.81 (1.76, 4.49)	< 0.001	2.72 (1.70, 4.36)	< 0.001

^aAdjusted analyses also included sexual relationship power and experience of intimate partner violence within the past 12 months

^bResponse dichotomized as ‘Yes a lot’ vs. ‘A little’, ‘A fair amount’ or ‘Not at all’

^cVariable not comparable between rounds and only Follow-up 2 is displayed

^dResponse coded as ‘Better’, ‘Same’ or ‘Worse’. Note: No women responded ‘Worse’

Significance at p < 0.05