Table 3.

Logistic regression main effects and interactions of neighborhood poverty and primary health insurer on early diagnosis of node negative breast cancer for Mexican American and non-Hispanic white women

Predictor Variables		Mexican American				Non-Hispanic White
		Sample	OR	(95% CI)		Sample		OR	(95% CI)
Neighborhood poverty
< 5% poor		28	1.00			1,525		1.00
5-29% poor		77	1.07	(0.40, 2.89)		1,300		0.90	(0.76, 1.07)
> 30% poor		198	0.66	(0.26, 1.69)		786		0.84	(0.68, 1.04)
Neighborhood Mexican American prevalence
< 50%		133	1.00			3,338		1.00
> 50% (Barrio)		170	0.68	(0.36, 1.27)		273		0.99	(0.74, 1.33)
Primary health insurer
Uninsured or underinsureda		115	1.00			433		1.00
Adequately insuredb		188	0.87	(0.44, 1.73)		3,178		1.72	(1.31, 2.27)
		Poverty by barrio by insurer				Poverty by insurer
Significant interactions		303	2.33 *	(0.94, 5.82)		3,611		1.61	(1.07, 2.42)
Poverty by insurer interaction for non-Hispanic white women
		< 5% poor				> 5% poor
Predictor Variables		Sample	OR	(95% CI)		Sample		OR	(95% CI)
Primary health insurer
Uninsured or Medicaid		129	1.00			304		1.00
Private or Medicare		1,396	1.75	(1.20, 2.57)		1,782		1.29 *	(0.99, 1.67)
Poverty by barrio by insurer interaction for Mexican American women
				> 30% poor
	< 30% poor			< 50% Mexican American			> 50% Mexican American
	Sample	OR	(95% CI)	Sample	OR	(95% CI)	Sample	OR	(95% CI)
Primary health insurer
Uninsured/Medicare	30	1.00		32	1.00		53	1.00
Private/Medicaid	75	1.45	(0.58, 3.60)	33	0.69	(0.22, 2.11)	80	2.09 *	(0.96, 4.58)

Notes: All effects were adjusted for age, grade and all other main and interaction effects. Adjustment of the MA regression model for birthplace made no practical difference in findings.

^a Uninsured or Medicare for MA women and uninsured or Medicaid for NHW women.

^b Private insurance or Medicaid for MA women and private insurance or Medicare for NHW women.

^*p < .10.