Health insurance mediation of the Mexican American non-Hispanic white disparity on early breast cancer diagnosis

During what has come to be known as the Great Recession, the census bureau estimated that the prevalence of poverty had increased nearly 25% in America over only four years (from 37.5 million in 2007 to 46.2 million in 2011). During the same era the prevalence of Americans without health insurance increased by more than 10% to 50 million (DeNavas-Walt et al. 2012). But if the underinsured were included, the estimate doubled to 100 million or approximately one of every three Americans (Kaiser Family Foundation 2012). When viewed through an ethnic lens, such inequities are even more startling. For example, the prevalence of being uninsured among Hispanic Americans (32.4%) was estimated to be nearly three times greater than among non-Hispanic white (NHW) Americans in 2009 (12.0% (Kaplan & Inguanzo 2011)). And among Hispanics, the prevalence was highest among Mexican Americans (MA), four of every ten of them lacking any form of health insurance, public or private (Miranda et al. 2011).

Even in the decade that preceded the great recession, the social risks that arose out of being uninsured and poor were stronger in some places and among some people. In California, the state with the largest MA population, the concentration of poor people into extremely poor neighborhoods rose substantially. And among Hispanic people, MA women in particular, the concentrations of the poorest poor and the least insured into so-called barrios were dramatic (Acosta 2010; Berube & Frey 2005; Jargowsky 2005). Our research group has been analyzing the effects of these social forces on cancer care in California over the past 15 years (Gorey et al. 2011; Gorey et al. 2009). This study aims to connect this work to the burgeoning social capital-based theoretical explanations for the Hispanic paradox. Diverse health benefits seem to be enjoyed by otherwise quite socioeconomically vulnerable people who live in Hispanic enclaves (Keegan et al. 2010; Mair et al. 2010; Cagney et al. 2007), especially in barrio neighborhoods that are predominantly populated by first generation immigrants from Mexico (Osypuk et al. 2010). When discussing healthcare disparities that Hispanics might face, it is important to analyze different Hispanic groups separately and avoid the risk of missing important clinical and social differences. Different groups may face different obstacles to accessing healthcare resources (Miranda et al. 2011).

Focusing on the cancer care and survival experiences of extremely poor people with cancers of great public health and human significance—breast and colon cancer—we have consistently observed that health insurance does not only matter, but indeed is critical. Adequate health insurance, be it private or public, seems to be strongly associated with access to the best available treatments and outcomes for all Americans. But for women, particularly women with the most treatable types of cancer such as localized, node negative (NN) breast cancer, health insurance seems to all but completely mediate or buffer the profoundly disadvantaging effects of poverty. We also observed two distinct, but theoretically related interactions of poverty and health insurance. First, the advantaging effects of health insurance were much stronger in low poverty neighborhoods, where less than five percent of the households were poor, than in high poverty neighborhoods, where thirty percent or more of the households were poor (Gorey et al. 2013; Gorey et al. 2012). Such high poverty neighborhoods have been described as places of prevalent demographic vulnerability that are particularly distressed for their lack of social and economic capital (Wilson 2012; Jargowsky 1997; Jargowsky & Bane 1991). It appears that women with breast cancer in more affluent neighborhoods, where more facilitative social and economic capital is available, are probably better able to absorb the indirect and direct, but uncovered, costs of care. Second, among MA women, the advantaging effects of health insurance were particularly strong in a certain type of high poverty neighborhood, that is, in barrios where the majority of the residents were MA (KMG, unpublished observations). Though seemingly paradoxical, such findings are consistent with the theory that MA barrios, even though they tend to be places of high poverty, may provide their residents with relatively more instrumental social and economic supports (Aranda et al. 2011; Markides & Eschbach 2005; Eschbach et al. 2004; Suarez 1994; Markides & Coreil 1986). This demonstrates that the effects of health insurance do not operate in a social vacuum. Health insurance surely matters, but so too does place and culture.

It seems that the interacting effects of being uninsured or underinsured, being poor and being an ethnic minority woman of color have been rather well studied during the post-diagnostic phase of breast cancer care. Much less is known about the diagnostic phase of care even though a number of studies have suggested that having adequate health insurance coverage at least partially mediates poverty and MA screening disadvantages (Miranda et al. 2011; Garcia et al. 2012; Gonzalez et al. 2012). It has already been established in this context that MA women are much less likely than NHW women to be diagnosed relatively early with NN breast cancer (KMG, unpublished observations), a type of breast cancer that has not yet spread to any regional lymph nodes and so typically has an excellent prognosis. Aiming to advance theoretical and practical understandings about this ethnic diagnostic gap, we advanced these hypotheses. First, the MA-NHW diagnostic gap is mediated by health insurance. Second, among NHW women the health insurance-early diagnosis relationship is moderated by poverty such that health insurance is less effective in high poverty neighborhoods. And third, among MA women in high poverty neighborhoods, the health insurance-early diagnosis relationship is moderated by barrio status such that health insurance is more effective in MA barrios.

The sampling frame was the California cancer registry. Study participants were originally randomly selected from three geographic and three socioeconomic place strata. Geographic strata were very large metropolitan areas (San Diego, San Francisco and Los Angeles), smaller metropolitan areas (Salinas, Modesto, Stockton, Bakersfield and Fresno) and rural places. Socioeconomic strata were based on the prevalence of poor households in census tract-defined neighborhoods: high poverty (30% or more), middle poverty (5% to 29%) and low poverty (less than 5% poor (Census Bureau 2002)). Data was obtained for 303 MA and 3,611 NHW women with breast cancer, diagnosed and staged between 1996 and 2000, who were originally followed until 2011. Most of the MA women were identified directly through medical records (77.8%), the remainder through a validated algorithm using Hispanic surnames and maiden names, birthplace, race and other record linkages (sensitivity = 84.4% and specificity = 99.1% (NAACCR Race and Ethnicity Work Group 2009)). MA barrios were defined as neighborhoods where 50% or more of the residents were MA in 2000. We explored other barrio criteria from 33% to 75% MA. The 50% criterion had the most predictive validity for these analyses. Health insurance and breast cancer care variables were extracted from hospital and physician office charts and clinic reports.

We used logistic regression models to test hypotheses about the mediating and moderating effects of poverty and health insurers in predicting binary (node negative or positive) breast cancer stage at diagnosis. Odds ratios (OR) and 95% confidence intervals (CI) were estimated. We also provided practical assessments more germane to clinical or policy significance. All rates were directly adjusted for age and tumor grade using this study’s sample as the standard. Then we used standardized rate ratios (RR) for all between-group comparisons with pooled 95% CIs. Statistical model tests are presented in tables. Practical significance indices are presented in the text. Other methodological details have been reported (Gorey et al. 2013; Gorey et al. 2012).

MA women with breast cancer were much less likely to have been diagnosed relatively early, before their disease had spread to regional lymph nodes, but this MA-NHW disparity was completely mediated by the main and interacting effects of health insurance. Advantages of health insurance were largest in low poverty neighborhoods for NHW women, while among MA women they were, paradoxically, largest in high poverty MA barrios. In fact, the highest rate of early NN breast cancer diagnosis was among such MA women with adequate health insurance who lived in MA barrios, with three-quarters having NN disease at the time of their diagnosis. Consistent advantages of being privately insured were also observed for all study participants, MA and NHW, with Medicare coverage seemingly more instrumental for the older cohort of NHW women and Medicaid coverage more so for the poorer cohort of MA women.

It seems that the effectiveness of public and private health insurance programs is significantly impacted by the availability of other key resources. In more well-to-do neighborhoods where social and economic capital abound most NHW women with breast cancer seem quite able to absorb the indirect and additional uncovered, direct costs of care. High poverty neighborhoods on the other hand, with their relative lack of such capital reserves, seem to remain as described more than a generation ago by William Julius Wilson, places of “true disadvantage” (Wilson 2012), especially for the women who live there. Not only are they much more likely to be uninsured or underinsured (Gorey et al. 2013; Gorey et al. 2012), but even when publicly or privately insured, these programs seem much less effective there than they are in places of lower poverty. Seemingly paradoxically, even within high poverty neighborhoods, a very strong advantaging effect of having health insurance was observed among MA women who lived in barrios where the majority of their neighbors were MA. These findings in support of the “barrio advantage” theory, suggest that adequate health insurance, in concert with other social and economic resources that may be more available in largely MA neighborhoods likely potentiate each other. It stands to reason that having the additional capital of either private or public health insurance could operate to potentiate the strengths and resiliencies that already seem to exist in barrios.

These findings reaffirm the preventive impact of health insurance especially among those at greatest risk of not having adequate coverage. They are also consistent with the theory that more facilitative social and economic capital is available to MA women in barrios and to NHW women in more affluent neighborhoods. It is there that each respective group of women with breast cancer is probably best able to absorb the indirect and direct, but uncovered, costs of care. Policy makers need to understand that even covered health care presently comes with myriad of costs. And while many seem able to absorb them, many others do not.