Perceived Discrimination and Ethnic Identity Among Breast Cancer Survivors

Abstract

Purpose/Objectives

To examine ethnic identity and sociodemographic factors in minority patients' perceptions of healthcare discrimination in breast cancer care.

Design

Mixed methods.

Setting

Participants' homes in the metropolitan areas of Phoenix and Tucson, AZ.

Sample

39 women treated for breast cancer in the past six years: 15 monolingual Spanish-speaking Latinas, 15 English-speaking Latinas, and 9 African Americans.

Methods

Two questionnaires were administered. Individual interviews with participants were conducted by nurse researchers. Quantitative, qualitative, and matrix analytic methods were used.

Main Research Variables

Ethnic identity and perceptions of discrimination.

Findings

Eighteen women (46%) believed race and spoken language affected the quality of health care. Perceived disrespect from providers was attributed to participant's skin color, income level, citizenship status, and ability to speak English. Discrimination was more likely to be described in a primary care context, rather than cancer care. Ethnic identity and early-stage breast cancer diagnosis were the only study variables significantly associated with perceived healthcare discrimination.

Conclusions

This article describes the first investigation examining ethnic identity and perceived discrimination in cancer care delivery. Replication of this study with larger samples is needed to better understand the role of ethnic identity and cancer stage in perceptions of cancer care delivery.

Implications for Nursing

Identification of ethnic-specific factors that influence patient's perspectives and healthcare needs will facilitate development of more effective strategies for the delivery of cross-cultural patient-centered cancer care.

Current reviews have illustrated that research since 1970 has produced little progress toward the elimination of racial and ethnic disparities in cancer health outcomes (Aziz, 2007; Kagawa-Singer, Valdez Dadia, Yu, & Surbone, 2010). Complex social-ecologic mechanisms contribute to racial and ethnic cancer disparities, including sociodemographic and healthcare system characteristics, tumor biology, and cancer screening behaviors. However, studies have consistently demonstrated that racial and ethnic differences in cancer morbidity and mortality outcomes exist independently of those social, biologic, and clinical variables, suggesting that processes related to poorly understood cultural factors may be involved (Morris, Rhoads, Stain, & Birkmeyer, 2010; Virnig, Baxter, Haberman, Feldman, & Bradley, 2009). In a comprehensive review of cancer disparities research, Kagawa-Singer et al. (2010) stated, “The path of cancer care we have been traveling requires that we rechart our course, for we know what is wrong, but we are unclear why” (p. 35).

Attention has been increasingly focused on the exploration of institutional and interpersonal discrimination in healthcare delivery, with both overt and subtle forms of discrimination contributing to racial and ethnic health disparities (Smedley, Stith, & Nelson, 2003; van Ryn & Fu, 2003). Substantial evidence shows that perceived discrimination is associated with a broad range of poor mental and physical health outcomes in the general population (Facione & Facione, 2007; Williams & Mohammed, 2009). Although researchers are beginning to consider how perceived discrimination may contribute to cancer-related disparities, most studies in this area have focused on the effects of discrimination on cancer screening behaviors, with few examining perceptions of healthcare discrimination in the cancer treatment context (Campesino, 2009; Howard, Balneaves, & Bottorff, 2007; Mandelblatt et al., 2003).

Little is known about the determinants related to perceptions of discrimination in a healthcare context or conditions in which such perceptions may occur. Social psychology laboratory-based studies demonstrated that ethnic identity influences perceptions of racial and gender discrimination, although whether it has moderating or mediating effects is unclear (Eccleston & Major, 2006; Major, Quinton, & Schmader, 2003). Ethnic identity is defined as one's sense of belonging to his or her ethnic group and the centrality or importance of that identity to the individual's self-concept (Phinney, 1992). Because ethnic identity comprises a central component of one's self-concept, laboratory demonstrations wherein ethnic identity influences how individuals respond to situations characterized by intergroup disparities are not surprising. However, the role of ethnic identity as it affects perceptions of discrimination in actual healthcare contexts remains largely unexplored (Karlsen & Nazroo, 2002). Notably, healthcare contexts, unlike laboratory settings, usually involve a complex set of conditions. For example, patients in a cancer care context face highly emotional matters, sometimes including life or death considerations. In addition, uninsured patients may experience limited opportunities in healthcare delivery and, therefore, lack a sense of freedom in considering all available treatment options. In short, the healthcare setting can place individuals in an affectively charged, high-dependency situation in which their well-being or survival may be at stake.

Although the self-defining processes inherent in ethnic identity are believed to pervade an individual's thoughts, feelings, and behaviors, other processes (e.g., survival) may mitigate their impact on perceptions of discrimination. The authors believe that a deeper understanding of the conditions and contexts in which individuals may perceive discrimination is needed to develop interventions that will prevent or ameliorate negative psychological and physical health outcomes for populations typically marginalized by race, ethnicity, gender, and other social markers. Investigating complex social phenomena such as perceived discrimination in a cancer care context may be best served by innovative, transdisciplinary research that incorporates qualitative methodologies in the design. As Morris et al. (2010) stated, such studies may shed new light on “previously unexplained patient and provider influences on disparities in cancer outcomes” (p. 110). Therefore, the purpose of this mixed-methods field investigation was to explore perceived discrimination and its relationship to ethnic identity in a cancer treatment context.

Methods

A nurse-led, transdisciplinary research team used a mixed-methods design to explore ethnic identity and perceptions of discrimination in health care among a sample of women treated for breast cancer (stages I–IV). All participants (N = 39) had completed cancer treatment within the past six years in Arizona. The sample included three groups of breast cancer survivors: monolingual Spanish-speaking Latinas (n = 15, 39%), English-speaking Latinas (n = 15, 39%), and African Americans (n = 9, 23%). The research team represented nursing, surgical oncology, social psychology, and communication.

The conceptual framework guiding the current study was critical race theory (Delgado & Stefancic, 2001), which posits that race, ethnicity, and culture are socially constructed identities that operate within systems of power in the dominant society. Fundamental principles in critical race theory include recognition that (a) discrimination in U.S. society is based on a variety of social markers such as gender, class, and language, as well as race and ethnicity, and (b) discrimination may be best understood by listening to the experiential knowledge of people who are recipients of discrimination. Critical race theory was chosen as a useful framework for understanding power dynamics that may be present in healthcare delivery systems and potential barriers that impede intercultural healthcare relationships. In the qualitative portion of the current study, the authors elicited women's perceptions and experiences in healthcare delivery, specifically exploring any potential occurrences of discrimination. For this study, racial and ethnic discrimination is defined “not as an individual aberration but as a system of oppression that operates in multiple segments of society, often in implicit ways, and in conjunction with other systems of oppression based on categories of difference, such as class and gender” (Campesino, 2008, p. 300).

Procedures

All study procedures and materials were approved by the institutional review board of Arizona State University. Participants were recruited from community-based organizations in central and southern Arizona from 2008 to 2009. Data were gathered from individual interviews, conducted in the participant's choice of language (Spanish or English), by racially and linguistically matched nurse researchers. Using a recruitment script, researchers informed participants that the study purpose was to understand their experiences in cancer treatment, that the researchers were not employed by or associated with their healthcare provider or agency, and that their identity would remain confidential. The one-time interviews were conducted at the location of the participants' choice, most often in their home, lasting about one hour. Participants completed two questionnaires and then participated in a tape-recorded interview. All participants were compensated with a $20 gift card to a local retail outlet at completion of the interview.

Measures

Ethnic identity was assessed using a 10-item Multiethnic Identity Measure (MEIM) (Roberts et al., 1999), which has performed adequately in studies (α = 0.67–0.85) with Latino and African American populations (Green, Way, & Pahl, 2006; Roberts et al., 1999; Syed, Azmitia, & Phinney, 2007). Five items each from the MEIM affirmation and exploration subscales were used (total scale: α = 0.74 in the current study). Because some items of the original scale are worded for adolescent populations, the authors slightly modified wording for relevancy to the adult population. For example, “I have spent time trying to find out more about my own ethnic group” was revised to “I have spent time teaching others about my own ethnic group.” Likewise, “I often talk to other people about my ethnic group to learn more about my ethnic background” was shortened to “I often talk to other people about my ethnic group.”

The 24-item demographic form included socioeconomic and acculturation indicators and questions about health insurance, treatment methods and locations, and family involvement. The acculturation questions assessed participants' nativity, number of years living in the United States, parents' race or ethnicity, country of education, and how well they spoke English. Two discrimination items were derived from the national Commonwealth Fund's (2001) Health Quality Survey. The first assessed beliefs about healthcare discrimination: “Most people in the United States receive the same quality of health care regardless of their racial background or language spoken.” Participants responded on a four-point Likert-type scale (1 = strongly agree to 4 = strongly disagree). The second assessed discrimination more generally: “How much discrimination against minorities do you feel there is in the United States today?” Respondents again used a four-point Likert-type scale (1= none to 4 = a lot). Higher scores on both items indicate more perceived discrimination.

A 17-item semistructured interview guide, developed by the research team, was used to explore cancer treatment experiences. Interviewers began with open-ended questions about cancer care, such as “What has it been like for you to receive health care from your providers?” Participants were encouraged to fully describe their encounters with oncology providers (physicians, nurses, and ancillary staff) and experiences with the delivery systems in which they received care. Prompts were given to encourage more details, such as “Did you trust your doctor?” and “How did you decide what treatment to get for the cancer?”

Interviewers progressed to more specific questions related to potential experiences of discrimination related to race or ethnicity, language, income, or gender. Those questions were phrased to minimize the likelihood of bias or leading the participants. For example, “In thinking about all of the experiences you have had with cancer care visits, have there been times when you felt that you were treated differently by the doctors [and nurses or receptionists]?” was derived from previous studies in healthcare discrimination (Blanchard & Lurie, 2004). Participants were encouraged to elaborate and clarify meaning and perceptions. They then were queried about their perceptions or experiences regarding potential specific markers for discrimination, such as race and income. Those included, “Do you think African Americans [or Mexican Americans] receive the same quality of health care as Whites? What makes you think that?” and “Do you think lower income patients receive the same quality of health care as higher income patients? What makes you think that?” Interviewers did not use terms such as discrimination or racism in any of the interview questions. However, participants frequently referred to those descriptors in their responses.

Data Management

The authors examined quantitative data using descriptive statistics, t tests, chi square, analyses of variance, and correlation coefficients; coding and content analysis of the qualitative interview data were performed using Atlas.ti, version 5.0. Credibility and consistency in the interviewing process were established through weekly meetings among the investigators who were conducting the interviews and monthly meetings with the entire research team. After several interviews, researchers identified five categories that would facilitate the coding of participant responses to the interview guide and that related to the study aims of examining healthcare discrimination: (a) perceptions related to healthcare delivery, (b) economic and access issues, (c) race and gender perceptions in care delivery, (d) quality of care, and (e) healthcare system issues. Participants were asked about their cancer treatment experiences in each category to facilitate a full exploration of potential areas of perceived healthcare discrimination.

To ensure consistency and dependability in the coding, three nurse researchers coded portions of two randomly selected interviews from each ethnic or racial group. Discrepancies in coding resulted in a return to the data and subsequent iterations until 90% inter-rater reliability was obtained. A fourth qualitative research methodologist reviewed the development of categories and the Atlas.ti coding process. Based on data in those categories across all participants, three qualitative groupings were identified according to perceptions of discrimination: (a) participants who made clear statements of perceived discrimination in any of the coding categories; (b) those who were equivocal, such as making contradictory or ambivalent statements; and (c) those who clearly stated that they did not perceive healthcare discrimination.

In the final analysis stages, the research team triangulated quantitative data (demographic variables, MEIM scores, and discrimination item scores) with the qualitatively derived groups (e.g., perceived discrimination, equivocal, no discrimination) to assess patterns within and across groups. The triangulation process was accomplished using matrix analysis. Matrices are an analytic strategy to construct a visual “conceptually oriented display” of specific aspects of data from multiple sources (Miles & Huberman, 1994, p. 177). Similar in concept to a correlation matrix of variables, the triangulation matrices permitted investigators to examine the intersection of sources of data to assess patterns of similarities or differences in perceptions of care. Triangulation strengthened the verification process to achieve the overall goal of obtaining a deeper and broader understanding of contextual factors involved in perceptions of breast cancer care.

Findings

Quantitative Analyses

Demographics

Women in the Spanish-speaking Latina group (n = 15) all were born in Mexico and had lived in the United States for 2–34 years ($\stackrel{‒}{X}=13.5$ years). All interviews with that cohort were conducted in Spanish. Although the researcher did not inquire about participants' citizenship status, most revealed that they were not legal residents, which created significant restrictions in access to ongoing cancer treatment. Members of the English-speaking Latina group (n = 15) were predominantly U.S.-born Mexican Americans, although three were first-generation legal residents who had lived in the United States for most of their lives (range = 20–55 years, $\stackrel{‒}{X}=37$ years). Some (n = 6) reported being bilingual (spoke English and Spanish equally well), but all preferred the interviews to be conducted in English. Among the African American group (n = 9), all were U.S. born and spoke English. The smaller number of women in that group reflected recruitment challenges because of the small African American population (3%) in study sites.

Spanish-speaking Latinas had a mean family income level of $10,000–$20,000 per year versus $30,000–$45,000 per year for English-speaking Latinas and African Americans (not statistically significant). The Spanish-speaking group also had a lower mean education level (7.7 years) compared to African Americans (13.9 years) and English-speaking Latinas (14.1 years) (p < 0.001). Spanish-speaking Latinas were significantly less likely to have healthcare insurance, largely because they were not legal residents. Only five of the Spanish-speaking Latinas had healthcare insurance, whereas all participants in the other two groups were insured, except for one English-speaking Latina (p < 0.001).

Treatment and diagnoses

Among the total sample, 24 had been diagnosed with breast cancer in stages I or II, whereas 8 were diagnosed in stages III or IV. The remaining seven did not know the tumor stage; five were from the Spanish-speaking group and two were from the English-speaking Latina group. All women had received surgery (lumpectomy or mastectomy, with or without chemotherapy or radiation) as part of their breast cancer treatment, except for two women in the Spanish-speaking group who were diagnosed too late in the disease process.

Ethnic identity

The total mean scale score for the 10-item MEIM was 31.4 (SD = 4.82, α = 0.73); the item mean was 3.1. Those scores are comparable to the 12-item MEIM used in other studies with Latino and African American populations (Green et al., 2006; Roberts et al., 1999; Syed et al., 2007). No significant differences were observed in ethnic identity scores by race or ethnic group. The English-speaking Latinas had a mean ethnic identity score of 30.1 (SD = 6), African Americans' mean score was 31 (SD = 3.5), and Spanish-speaking Latinas had a mean score of 33.1 (SD = 3.9). The scores were unrelated to age, income, education, and language.

Perceived discrimination items

Almost half (n = 18, 46%) of the total sample disagreed or strongly disagreed (higher scores) with the item, “Most people in the United States receive the same quality of health care regardless of their racial background or language spoken.” Spanish-speaking Latinas had the lowest score ($\stackrel{‒}{X}=2$), with four (27%) disagreeing. Five African Americans (56%) disagreed with the statement ($\stackrel{‒}{X}=2.56$), and English-speaking Latinas had the highest disagreement at nine participants (60%) ($\stackrel{‒}{X}=2.6$).

The second item related to general discrimination asked, “How much discrimination against minorities do you feel there is in the United States today?” Most of the sample (77%) felt a bias exists against minorities: 18 (46%) responded with “a lot” and 12 (31%) responded with “some.” Spanish-speaking Latinas again had the lowest disagreement, with 10 (67%) reporting “a lot” or “some” ($\stackrel{‒}{X}=3.07$). For English-speaking Latinas, 12 (80%) reported “a lot” or “some” discrimination ($\stackrel{‒}{X}=3.13$). Finally, eight African Americans (89%) reported “a lot” or “some” discrimination ($\stackrel{‒}{X}=3.33$). Only three participants from each of the Latina groups felt “a little” discrimination existed against minorities, whereas one African American and one Spanish-speaking Latina reported “none.” No significant differences in mean scores were observed for the two quantitative discrimination items between the three racial or ethnic groups, nor did any differences emerge related to sociodemographic variables including income, age, and language spoken.

Qualitative Data

During the interviews, 14 participants (36%) clearly described experiences and perceptions of discrimination in healthcare delivery related to race or ethnicity, Spanish language, skin color, citizenship status, or having low income. Other participants described healthcare discrimination, but made ambivalent or equivocal statements (n = 16, 41%). A smaller group of participants did not perceive any healthcare discrimination (n = 9, 23%). Table 1 provides representative quotes for each of those categories of discrimination.

Table 1.

Representative Discrimination Quotes by Ethnic Group

Group	Perceived Discrimination	Equivocal	No Perceived Discrimination
English-speaking Latinas	Maybe because I'm Mexican…. Maybe if I was a White I'd be treated better.	[Do Mexican Americans have the same quality of care as Whites?] Well, no. And I'm not sure exactly why. I'm not sure if it's because they don't know the resources, or it's because they're shy about it, or the insurance.	I think there are avenues and ways to get care—I don't think anybody's ever turned away. This country is fantastic at providing for the immigrants.
African American	I do know that racism exists, but I don't have a way to prove it to you…. I don't see that we need to go to people who treat us poorly.	I bet most of the time it's money. I'm not saying that it's race oriented, or that it's gender oriented; there are more Black peoples that are affected by that.	The minorities that come in there, I would say, do get the same treatment.
Spanish-speaking Latinas	Everywhere in the news, everywhere you hear that because you're Hispanic, you have no right to medical services.	Friends tell me that at the other hospitals they don't have to wait as long as here, that it's faster, and that there's a lot of White people that go, and very few Hispanics. So I realize that, well, yes, there's a disparity with people, right? But I haven't had that treatment.	No, ethnicity and race don't matter to me, what matters is that they're a good doctor and that they're helping me.

Participants across the three racial or ethnic groups were more likely to describe discrimination in primary care settings, rather than in their oncology care. A predominant sentiment described across groups was gratitude toward oncology providers, even in cases when satisfaction with cancer care was lower, or in a few instances, where medical errors occurred. In those cases, participants attributed blame to healthcare systems rather than oncology providers.

Perceptions of Spanish-speaking Latinas

In response to interview questions (“Do you think Mexican Americans [or Mexicans] receive the same quality of health care as Whites?” and “Do you think lower income patients receive the same quality of health care as higher income patients?”), Spanish-speaking Latina participants reported perceived differences in the quality of health care based on spoken language, citizenship status (i.e., being an undocumented immigrant), and lack of healthcare insurance or restricted access to care. For example, a Mexican immigrant (Spanish speaking) participant stated, “The people who are from here [the United States], well, they have all the guarantees, they have everything, everything, everything.” Regarding the difficulty in relying on interpreters to communicate with providers, another Spanish-speaking participant remarked,

If they wanted me to make a decision about something, I couldn't understand it all … when there's an interpreter they don't repeat everything, they just summarize. So, of course I would rather have a Hispanic doctor because they could talk to me in Spanish.

Perceptions of English-speaking Latinas

English-speaking Latinas perceived healthcare discrimination based on spoken language, skin color, citizenship status, and lack of healthcare insurance or restricted access to care. However, those perceptions more often were in regard to challenges they believed Spanish-speaking immigrants experienced, rather than their own group. For example, many described the difficulty Latino non-citizens experience in accessing health care, as well as voicing awareness of anti-immigrant rhetoric prevalent in the state. One Mexican American (English speaking) participant stated,

In Arizona right now, the Mexican community is just being persecuted, you know? Which, yeah, they're not supposed to be here illegally, but for God's sake, they're human beings, you know?

Perceptions of African Americans

African American participants perceived healthcare discrimination related to race, gender, and lack of healthcare insurance or restricted access to care. As with the other two groups, perceptions related to discrimination more often were in relation to general health care rather than oncology care. For example, one participant from that group said,

I do know that racism exists, but I don't have a way to prove it to you… . But I'm coming as a patient, as a client, and so, my thing is to see what do I need to do to make this work… . It don't have to be the doctor's office. It can be me going to buy a ticket for whatever, but you get some kind of vibe, I can't explain it, but I know most African Americans understand it.

Data matrices

An integral step in the mixed-methods approach includes conducting multilevel matrix analysis of the qualitative and quantitative data. After the interview data were content analyzed and coded to identify responses pertaining to healthcare discrimination within each racial or ethnic group, as described in the previous section, data matrices were constructed for three qualitative groups (cross-cutting race and ethnicity) according to all participants who described perceptions of discrimination (n = 14), those who were ambivalent or equivocal (n = 16), and those who did not perceive discrimination (n = 9). As shown in Table 2, the resulting discrimination matrices contained even distributions of participants across the three racial and ethnic groups, except for the equivocal group, which included seven Mexican immigrants, seven Mexican Americans, but only two African American participants. No patterns existed in sociodemographic characteristics (age, income, or education) between the three qualitatively derived discrimination groups.

Table 2.

Ethnicity of Sample by Qualitatively Derived Discrimination Groups

	English-Speaking Latinas (N = 15)	African American (N = 9)	Spanish-Speaking Latinas (N = 15)
Discrimination	n	n	n
Perceived	5	4	5
Equivocal	7	2	7
Not perceived	3	3	3

Triangulation matrices

In the final step of analysis, a matrix was constructed to examine possible patterns between the three qualitatively derived discrimination groups and quantitative data (mean scores) from the discrimination survey items and ethnic identity measure (see Table 3). The matrix analysis showed congruency between qualitatively derived discrimination categories and scores on the healthcare discrimination survey item, such that the perceived discrimination group evinced the highest score on that item. However, the mean differences were not statistically significant.

Table 3.

Matrix of Discrimination Items and Ethnic Identity Scale by Qualitative Discrimination Groups

Variable	$\stackrel{‒}{X}$	Range	Perceived Discrimination	Equivocal	No Perceived Discrimination
Health	2.36	1–4	2.8	2	1.9
General	3.15	1–4	3.6	2.5	3.2
Ethnic identity*	31.4	20–40	33.4	31.1	29.1

^*p = 0.031

Note. Higher scores indicate stronger perceived discrimination and ethnic identity; lower scores indicate less perceived discrimination and weaker ethnic identity.

The authors then examined the relationship between the ethnic identity measure (MEIM) and qualitatively derived discrimination groups. Ethnic identity scores were significantly higher among participants who perceived discrimination ($\stackrel{‒}{X}=33.4$) than those who perceived no discrimination ($\stackrel{‒}{X}=29.1$), t(21) = 2.31, p = 0.031. Neither group differed from the equivocal group. In addition, no associations were found between ethnic identity and participants' scores for the two quantitative discrimination items.

Finally, unexpected significant differences were found among perceived discrimination categories in relation to stage of breast cancer diagnosis. Women who reported perceived healthcare discrimination in the interviews had earlier stage breast cancer diagnosis (stage I or II) compared to those who did not perceive discrimination (stage III or IV) (p < 0.001).

Discussion

Since 2000, racial and ethnic discrimination in healthcare delivery has received increased focus as a potential factor contributing to health disparities among target populations. The current study recruited breast cancer survivors from identifiable target groups (African American and Mexican-origin women) to understand their perceptions and experiences of discrimination in cancer care delivery. A key aspect in understanding healthcare discrimination lies in the identification of variables related to attributions of discrimination. Although laboratory-based studies have shown that perceptions of discrimination are influenced by ethnic identity (Eccleston & Major, 2006), the relationship has not been tested in a healthcare context. The current field study extends existing laboratory-based research by demonstrating that ethnic identity is associated with perceptions of discrimination among Latina and African American breast cancer survivors. To the best of the authors' knowledge, this article describes the first investigation examining ethnic identity and discrimination in a healthcare delivery context in general, and in cancer care specifically. Therefore, this study offers new knowledge that may be useful in understanding factors influencing perceived discrimination in healthcare delivery for future cancer and social behavioral science research. An additional significant and unexpected finding was that perceived healthcare discrimination was associated with early-stage cancer diagnosis.

Discrimination Against Minorities

Responses from a quantitative discrimination measure indicated that 30 participants (77%) felt “some” or “a lot” of discrimination exists in general against minorities in the United States. No significant differences existed in perceptions of discrimination between the racial and ethnic groups in the sample. Prior research (Campesino, Ruiz, Uriri Glover, & Koithan, 2009) has documented Latina cancer survivors' awareness of the anti-Mexican immigrant sentiment that predominates public discourse in Arizona and the United States (Fisher & McCombs, 2010; Pitzl, 2008). Despite the sometimes rancorous local immigration debate in Arizona, the prevalence of general perceived discrimination among the sample was similar to findings from nationally representative surveys. In the longitudinal Black Women's Health Study (N = 49,161) (Taylor et al., 2007), 66% reported experiencing discrimination in the workplace, housing, or by the police. In a national survey of U.S. Latinos (N = 2,003) (Pew Hispanic Center, 2007), 41% reported having had personal experiences with racial or ethnic discrimination. The current sample's attributions to discrimination are similar to the Pew national findings: skin color, income, education, immigration status, and language skills were identified as reasons for discrimination. Other studies have demonstrated increased perceptions of discrimination among foreign-born Latinos versus U.S.-born Latinos and among those who speak a language other than English at home, regardless of nativity (Lauderdale, Wen, Jacobs, & Kandula, 2006).

Healthcare Discrimination

Among the sample, 18 women (46%) felt that the quality of health care in the United States differs according to a person's race or spoken language. Perceptions of healthcare discrimination were unrelated to sociodemographic characteristics including racial or ethnic group, income, education, age, availability of health insurance, or language spoken. By contrast, other studies have found that perceived bias in healthcare delivery was explained by demographic variables such as education and income (Hausmann, Jeong, Bost, & Ibrahim, 2008; Johnson, Saha, Arbelaez, Beach, & Cooper, 2004). Perez, Sribney, and Rodriguez (2009) found that uninsured Latinos were more likely to perceive lower quality of care. However, participants in the current study perceived lower quality of care, even when health insurance coverage and access to care were present.

The interview data revealed that perceived disrespect from providers and lower quality of health care was attributed to multiple factors, including skin color, low income level, immigration status (i.e., citizenship), and ability to speak English. Interestingly, those negative perceptions most often did not apply to participants' own cancer care delivery, in which they generally were quite satisfied, even in some cases where medical errors had occurred. Specifically, Spanish-speaking Mexican immigrant women described great difficulty in accessing cancer care and unsatisfactory communication with oncology providers, despite the use of healthcare translators. The lower mean score for the healthcare discrimination survey item in that group may be related to their interpreting the question with U.S. citizens as referents, rather than referring to themselves. Others in that group also felt that being an undocumented immigrant and Spanish speaker was related to disrespect from some providers, as well as restricting access to primary health care and cancer care. Therefore, triangulation of scores from the healthcare discrimination item with interview responses provided a contextualization of the data that yielded more accurate interpretations of findings than could have been achieved through reliance on the closed-ended discrimination items alone.

Reports of perceived healthcare discrimination across racial and ethnic groups vary widely in the literature. In a nationally representative survey of Latinos (N = 800), 43% believed race or ethnicity is very important in determining the quality of health care (Research! America, 2008). In contrast, the California Health Interview Survey (N = 42,044) found that only 13% of Latinos and African Americans reported they would have received better health care if they belonged to a different race or ethnic group (Lauderdale et al., 2006). Bird, Bogart, and Delahanty's (2004) study of people infected with HIV or AIDS (N = 110) found that 71% reported experiencing healthcare discrimination based on race, although the study authors noted that findings could have been complicated by providers' attitudes regarding sexual orientation. Note that the measures used to assess discrimination across those studies differ in their level of specificity, such that some allude to race or ethnicity in general, whereas others ask about participant's own attributes or experiences. Greater attention and consistency are warranted in the measurement of discrimination.

Studies of healthcare discrimination in a cancer context have focused largely on the inhibitory effects of perceived discrimination on cancer screening behaviors (Crawley, Ahn, & Winkleby, 2008; Facione & Facione, 2007; Shariff-Marco et al., 2009). A paucity of literature pertains to perceived discrimination in cancer treatment from which to compare the current findings. Campesino (2009) found that older Mexican American cancer survivors generally were satisfied with oncology and primary care providers, although the sample size was very small (N = 5). Breast cancer care research with African Americans and Latinas reported lower patient satisfaction related to perceived insensitivity from providers and denied access to care (Ashing-Giwa et al., 2004; Howard et al., 2007; Katz et al., 2005; Mandelblatt et al., 2003). However, a study of racially diverse patients with breast cancer (N = 789) found that most (62%) felt respected by their breast cancer surgeon, although satisfaction was lower among African American women (Thind, Diamant, Liu, & Maly, 2009). None of those studies included an examination of ethnic or cultural identity in the analyses.

Ethnic Identity

The association between perceived healthcare discrimination and ethnic identity is a key study finding. Among the current sample, ethnic identity scores were significantly higher among those who perceived healthcare discrimination compared to those who perceived no discrimination, as reported in the qualitative data. The finding is congruent with laboratory-based studies that have found perceptions of discrimination are influenced by stronger ethnic identity. However, in the current study, the relationship emerged only when considering the qualitative measure of discrimination. The authors acknowledge that because the current study was correlational in nature, rather than experimental, speaking to the directional aspect of that relationship is not possible. However, individuals with higher ethnic identity scores also were more likely to report observed biases in the healthcare setting. Quantitative measures alone did not show the same pattern, as no association existed between ethnic identity and responses to the close-ended measures of discrimination. That finding may have occurred because the items used were general and overarching. In contrast, the qualitative methodology allowed participants to report their experiences using their own voice. That disparity across quantitative and qualitative measures suggests that capturing participants' experiences of discrimination through their own words may yield valuable information over and beyond the use of traditional quantitative measures. Future mixed-methods approaches may help to further delineate what may be a complex relationship between ethnic identity and discrimination. The authors believe that departures from close-ended items may be warranted, particularly when working with nontraditional populations, such as those in this sample. Likewise, use of larger samples and longitudinal approaches may help in elucidating the nature of the relationship between ethnic identity and perceived discrimination.

Tumor Stage

An unexpected finding was that perceived discrimination was significantly associated with earlier stage, rather than later stage, breast cancer diagnosis. No other studies were located that examined perceived healthcare discrimination with cancer staging as a variable. A possible explanation for the finding may be related to the vulnerability patients experience in facing a diagnosis of cancer, which may elicit heightened outcome dependency toward oncology providers. Patients with cancer are dependent on the physician for crucial information and support, that is, to identify, discuss, and recommend cancer treatment options that will best impact survival outcomes. A high dependency situation such as facing a cancer diagnosis may inhibit attributions to discrimination in the context of the patient-provider relationship. The fact that early-stage diagnosis was significantly related to greater perceived discrimination compared to later-stage diagnosis lends support to that possibility, presuming that a later-stage diagnosis elicits greater fear of cancer-related death and, therefore, greater outcome dependency. Laboratory-based experiments have determined that outcome dependency influences stereotyping and impression formation, particularly in interpersonal situations characterized by unequal social power (Depret & Fiske, 1999; Fiske, 2001). Those findings need to be further explored and replicated to better understand what role the context and meaning of a cancer diagnosis plays in a patient's perceptions of cancer care delivery. A potentially fruitful area for further research is examination of the effects of outcome dependency on perceived discrimination in healthcare contexts in which varying degrees of social power exist between the patient and provider, such as racial and gender concordant dyads (e.g., White male patient and White male physician) versus racial and gender discordant dyads (e.g., African American female patient and White male physician).

Limitations

The current study had several limitations. The sample size limits generalization to other populations of cancer survivors. In addition, all data were self-reported, which was the appropriate level of evidence given the mixed-methods approach, but verification of data such as cancer stage and other healthcare information was not possible without access to participants' medical records. However, studies with nationally representative samples of African American women found that self-reports of breast cancer had high degrees of accuracy (Lauderdale et al., 2006).

Conclusions

The current study expands the knowledge of factors that influence perceptions of healthcare discrimination among U.S.- and foreign-born Latinas and African Americans with breast cancer. The authors believe that this is the first study to examine the role of ethnic identity in perceived discrimination within a cancer care context. The mixed-methods approach provided a rich data set from which to analyze the complexities inherent in studying perceived discrimination.

Implications for Nursing

The development of culturally responsive clinical practice styles requires a more explicit understanding of how discrimination may operate at institutional levels and in everyday practice. Findings from this study may be used to inform the design of innovative, transdisciplinary research to better understand the role of ethnic identity in patients' perceptions of quality cancer care delivery. Identification of ethnic-specific factors that influence patient's perspectives and healthcare needs will facilitate development of culturally tailored strategies for the delivery of patient-centered care. Nursing, medical, and surgical oncology providers, like other healthcare providers in the United States, are experiencing increasingly diverse patient populations, which create complex and challenging healthcare environments. If those challenges are met as potential opportunities to improve cross-cultural oncology care, actual in-roads may be made in the journey toward eliminating cancer health disparities among racial and ethnic minorities.