Layers of information: Interacting constraints on breast cancer risk-management by high-risk African American women

Abstract

Objectives:

To understand how various decision-making dynamics interact to shape the risk-management choices of African American women at high-risk of breast cancer, and to explore whether African American and White women have differential access to the information and interactions that promote proactive, confident risk-management behavior.

Design:

This paper draws on fifty original in-depth, semi-structured interviews with African American and White women at elevated risk of breast cancer. We used inductive grounded-theory methodology to explore the processes by which women make risk-management decisions and to compare those processes between racial groups. Moving backward from women’s decisions about whether or not to engage in specific risk-management behaviors, we explored the patterns that underlie behavioral decisions.

Results:

We find that decisions to engage in risk-management behavior rest on three accumulated layers of information. The layer most proximal to making risk-management decisions involves specific information about risk-management options; the middle layer involves general information about managing breast cancer risk; and the foundational layer involves personal perceptions of breast cancer risk and prevention. African American and White women experience distinct dynamics at each of these levels, and these differences may help explain racial differences in risk-management behavior. Compared to their White counterparts, African American women faced additional burdens at every step along the risk-management journey.

Conclusion:

These findings suggest that information gathering is more complex than has previously been addressed, that information access and provider access are closely related, and that African American women may be systematically disadvantaged with respect to information-generating experiences. Preventing cancer morbidity and mortality requires that all high-risk women have access to the layers of information necessary to engage in cancer screenings and preventive interventions. These results exemplify the ways that structural, social, and interpersonal inequalities combine to influence risk-management choices.

Introduction

Racial disparities have been well documented in a range of health conditions (Williams and Jackson 2005; Williams and Mohammed 2009), and are shaped by interacting levels of structural, social, psychological, and biological determinants (Warnecke et al. 2008). Breast cancer is no exception, with African American women diagnosed at similar rates as White women, but with more advanced and deadlier cancers (Iqbal et al. 2015; Richardson 2016). With 5–10% of breast cancers attributable to inherited genetic mutations (National Cancer Institute 2018), one promising front in the fight against breast cancer is prevention and early detection among women known to be at elevated risk due to diagnosed mutations or extensive family history of the disease. This paper addresses the question of how African American and White women at elevated risk of breast cancer engage with their risk-management options.

Women who know they are at higher-than-average risk of breast cancer may consider a range of choices to help manage this risk. Preventive surgeries (prophylactic mastectomy to remove one’s breasts, or prophylactic salpingo-oophorectomy to remove one’s fallopian tubes and ovaries) and chemoprevention (taking a daily dose of a hormonal therapy for several years) each reduce the risk of breast cancer for high-risk women by 50% or more (Daly et al. 2016; Domchek et al. 2010; Lehman et al. 2007; Lostumbo, Carbine, and Wallace 2010). Enhanced surveillance routines (earlier and/or more frequent mammograms plus breast MRIs) offer the promise of detecting cancer early so it can be effectively treated (Daly et al. 2016; Domchek et al. 2010; Lehman et al. 2007; Lostumbo, Carbine, and Wallace 2010). Choosing to engage - or not engage - in these behaviors requires healthy women to navigate complicated decision-making processes. These processes involve women’s perception of personal risk, access to information about risk and prevention options, understanding of and confidence in that information, interactions with healthcare providers, social support, financial capacity, and more (Padamsee et al. 2017).

Prior research by this group has demonstrated that many features of decision making are different for African American and White women (Blinded, Under Review). This insight stems from our inductive analysis of fifty in-depth, semi-structured interviews, in which women at elevated risk explained their feelings, thoughts, and decisions in their own words. The objective of the current study is to delve more deeply into women’s stories in order to understand how various decision-making dynamics interact to shape risk-management choices. In particular, we ask whether African American and White women have differential access to the information and interactions that promote proactive, confident risk-management behavior.

To answer this question, we move logically backward from the decision itself – to engage in risk-management behavior – digging through the layers of information that build to that end point. In order to choose to have a prophylactic surgery, take a preventive medication, or embark on a routine of intensified screening tests, a woman must know that the option exists, and that her own cancer risk is high enough to make it personally relevant. To make sense of these specific pieces of information, women must also possess a more general base of information about breast cancer risk and prevention, and this requires that they have access to trustworthy information sources. Underlying even that general information, women must perceive of cancer risk and cancer prevention as salient issues, such that there is sufficient motivation to learn about prevention and act on those insights. As we examine these layers of the decision-making context, we look for racial patterns that may distinguish the experiences of White and African American women and impact the potential of women in each group to make informed, health-protective decisions.

Background

Breast cancer risk estimation models identify high-risk women based on factors such as age, family history, and known genetic mutations (i.e. BRCA1, BRCA2, others) (Antoniou et al. 2004, 2008; Claus, Risch, and Thompson 1994; Constantino et al. 1999; M. Gail et al. 1989; M. H. Gail and Mai 2010; Tyrer, Duffy, and Cuzick 2004). High-risk women can face a lifetime risk for breast cancer of up to 84%, far greater than the average 12% lifetime risk for the general population of US women (Bevers et al. 2016). It is estimated that 10 million women in the USA meet the high-risk criteria for breast cancer (Freedman et al. 2003). The prevalence of BRCA mutations in particular is comparable among White, African American, Hispanic, and Asian American breast cancer patients, at 1–4% per gene (John et al. 2007; Kurian 2010).

Several risk-reduction methods are available to support both the physical health and psychological well-being of high-risk women; they can be used separately or in conjunction. Bilateral prophylactic mastectomy is the most effective method, and reduces breast cancer risk by about 90% and breast cancer-specific mortality by 80% (Domchek et al. 2010; Lostumbo, Carbine, and Wallace 2010). Bilateral prophylactic salpingo-oophorectomy is recommended for all BRCA mutation carriers between the ages of 35 and 40 years (or when childbearing is complete). It reduces the risk of ovarian, fallopian tube, or peritoneal cancer by up to 88% (Ludwig et al. 2016), and the risk of breast cancer by 50% (Eisen et al. 2005). Two chemoprevention medications are FDA-approved for use in high-risk women. A 5-year course of premenopausal Tamoxifen reduces risk of breast cancer by 30% to 50%. Raloxifene – approved for post-menopausal women – is about 76% as effective as Tamoxifen (Cuzick et al. 2013).

Risk-reduction options are significantly underused by high-risk women who may benefit from reductions in both cancer risk and cancer-related worry. Furthermore, usage rates vary across populations. Uptake rates for prophylactic surgeries differ significantly by country, age group, parity status, family history, and Jewish ancestry (Friebel et al. 2007; K. A. Metcalfe et al. 2008; Kelly A. Metcalfe et al. 2008, 2012). Rates of prophylactic mastectomy among cancer-free BRCA mutation carriers vary between 11% and 50% for specific populations (Kelly A. Metcalfe et al. 2012; Skytte et al. 2010). Rates of prophylactic oophorectomy vary across groups of BRCA mutation carriers, but are generally at least 50% and can exceed 90% in Jewish and/or genetic counselling samples (Friebel et al. 2007; Kelly A. Metcalfe et al. 2008, 2012; Morgan et al. 2009; Skytte et al. 2010). Prevalence of chemoprevention use in the USA is estimated to be 1–8% (Smith et al. 2016; Waters et al. 2010), although uptake averages 25% in the clinical trial setting (Smith et al. 2016). Between 20–50% of BRCA mutation carriers engage in surveillance alone, without the use of a specific biomedical reduction intervention (Morgan et al. 2009; Schwartz et al. 2012). Guidelines recommend an annual mammogram and breast MRI (Bevers et al. 2016; Daly et al. 2016) for women with 20% or higher lifetime risk of breast cancer. However, fewer than 30% of women follow evidence-based screening recommendations, and the use of screening varies by race and ethnicity (Miles et al. 2018; Wernli et al. 2014).

If racial differences exist in the uptake of risk-reduction methods, they could account for some proportion of the known disparities in breast cancer morbidity and mortality (DeSantis et al. 2016). Research on usage of these methods has not systematically compared racial groups, but some patterns have been identified. African American women are less likely to participate in genetic risk assessment and counselling (Armstrong et al. 2005; Halbert 2006; Sherman et al. 2014; Simon and Petrucelli 2009), and less likely than White women to know about risk-reducing therapies (Kaplan et al. 2006). Compared to White women, African American women also have less knowledge about risk factors for breast cancer (Alexandraki and Mooradian 2010; MacNew et al. 2010; Manning et al. 2013), and perceive themselves to be at lower risk for developing it (Orom et al. 2013).

We also know that African American and White breast cancer patients make different treatment choices. African American women with early-stage breast cancer more frequently choose full mastectomy (as opposed to the less-invasive lumpectomy plus radiation option) for surgical treatment (Yeboa et al. 2016). They are only half as likely as Whites, however, to choose contralateral prophylactic mastectomy (having the healthy breast removed along with the cancer-affected breast) – a procedure not associated with improved clinical outcomes among average-risk women that has nonetheless become increasingly popular (Brown et al. 2016; Wong et al. 2017). Little is known about the factors that drive these decisions (Kim et al. 2017), or why they differ by race.

Based on the existing research on racial disparities related to breast cancer treatment, genetic testing, and knowledge about risk, it seems likely that risk-management decision making also differs between African American and White high-risk women. This study aims to use women’s own stories to explore the knowledge, understanding, resources, and choices that inform risk-management decisions within these two groups, tracing backward from the point of choosing whether or not to take a preventive action through the layers of information that would likely precede that choice.

Data & Methods

Data for this study consist of fifty original in-depth, semi-structured interviews with African American and White women at elevated risk of breast cancer, about their experiences with risk, risk-management behaviors, and decision-making processes. Adult participants at known elevated risk of breast cancer but without prior cancer diagnosis were recruited through clinics of the Ohio State University James Comprehensive Cancer Center and research volunteer databases. The study was conducted with approval of the Ohio State University’s Institutional Review Board (Protocol # 2014C0101).

Transcribed data were analyzed using grounded theory methods. Analysis proceeded in five stages, with at least two analysts involved at each stage: (1) exhaustive coding of initial interviews for inductive themes; (2) systematic and comprehensive coding into themes and unifying nodes; (3) in-depth exploration of the nodes and themes using comparative tables and exploratory memos; (4) theorizing based on analysis of emergent, core decision-making dynamics; and (5) comparison of African American vs. White women’s experiences with respect to each core decision-making dynamic. Additional information about the study’s sample, analytic methods, and the emergent theoretical concepts is available in the Methodological Appendix. In the text below, each informant is identified by pseudonym, race, and age.

Methodological Appendix.

Data for this study consist of fifty in-depth, semi-structured interviews with women at elevated risk of breast cancer about their risk-management behaviors and decision-making processes. Encouraging women to speak in their own words, these interviews explored women’s awareness of breast cancer and perceived risk status; the sources and content of risk information they have obtained; their understanding and consideration of prevention options; their expenditure of financial resources, time, and energy in coping with risk and prevention; decision-making processes; decision-making networks; and psychosocial well-being. Informants were recruited through multiple methods: in-person recruitment of patients at the High Risk Breast Program and Cancer Genetics clinics of the Stefanie Spielman Comprehensive Breast Center (James Cancer Hospital, Ohio State University Comprehensive Cancer Center); from ResearchMatch and StudySearch databases of research volunteers; and via snowball sampling from interviewed informants. Study participants included women who were at least 18 years old; had no prior cancer diagnosis; and had a breast cancer risk defined as ‘moderate’ (above average family history of breast or ovarian cancer), ‘high’ (family history that includes multiple, young, and/or bilateral cases of breast or ovarian cancer), or ‘severe’ (diagnosed BRCA1 or BRCA2 gene mutation) (Hampel et al. 2004). Because one of the main goals of the study was to enable comparison of decision-making dynamics between racial groups, we engaged in purposive sampling by race in order to recruit a final sample of 30 non-Hispanic White and 20 African American women. Interviews were conducted by the first author between May 2015 and March 2016. All but one informant agreed to have their interviews digitally recorded and professionally transcribed. Interviews ranged in length from 22 to 120 minutes, averaging 57 minutes. The study was conducted with approval of the Institutional Review Board of The James Cancer Hospital, The Ohio State University Comprehensive Cancer Center, and The Ohio State University (Protocol # 2014C0101).

Transcribed data were analyzed using grounded theory methods and the NVivo software package for qualitative data. The analysis reported in this paper proceeded in five stages. First, three separate coders generated exhaustive lists of the themes that emerged in the first eight interviews. These three lists were compared to generate one comprehensive list of themes, some of which were anticipated by original interview questions and others that emerged from women’s own accounts; related themes were then organized into nodes. Second, all interviews were coded systematically based on this list of inductively-derived nodes, with segments of women’s stories assigned to as many nodes and themes as relevant. Additional nodes and themes were created as they emerged from systematic, inductive exploration of the data. Two coders coded each transcript, catching one another’s omissions and discussing any areas of disagreement until a reliable coding principle could be agreed upon. Transcription, coding, and memoing stages all began while interviews were still being conducted; the conceptual content of later interviews thus incorporated insights generated by the analysis of early interviews to facilitate theoretical saturation before the end of data collection. The third stage of the inductive analysis involved in-depth exploration of the nodes and themes created in the first stage. Excel tables were created to chart node/theme attributes by subgroups or individual informants, and analytic memos were generated to explore patterns within nodes/themes and relationships between them.

Building on this previous work, the fourth stage of analysis for this study focused on three core decision-making dynamics: access to information, access to health care providers, and prevention-related beliefs. Within these core dynamics, several theoretical concepts were explored in-depth.

Use of, knowledge of, and disposition toward risk-reduction options:

Whether participants had heard of and/or used each risk-reduction option: prophylactic mastectomy, prophylactic oophorectomy, chemoprevention, genetic testing, and enhanced screening methods. Categories of disposition toward each option were inductively generated: chose the option; aware and leaning toward; aware of it but not considering; would consider under certain conditions; and unaware of the option.

Access to health care providers:

Primary care providers (PCPs) included family doctors, nurse practitioners, obstetricians, and gynecologists. Specialists included oncologists, breast surgeons, gynecologic oncologists, and genetic counsellors. An informant was categorized as having seen a specialist if she sought or received information about breast cancer risk from that provider.

Trust in physicians:

Categorized based on informants’ thoughts about and interactions with health care providers, as high (choosing to remain with a physician, praising his/her expertise, explicitly mentioning quality/trust) or low (choosing to switch doctors, seeking more proactive cancer prevention advice, explicitly mentioning lack of trust).

Sources of information:

Categorized as follows: personal medical knowledge (gained from medical/scientific education or work experience), family, friends, celebrities, co-workers and acquaintances, healthcare providers, medical or research journals, online research, magazines and newspapers, pamphlets and other printed information, individuals with similar experiences, support groups, social media, and other.

Main source of information:

The source of information the informant discussed most, described as having provided the most information relevant to risk and prevention, or turned to first when seeking information.

Information gaps:

Informant mentioned missing pieces of knowledge she desired or needed, or gaps in her ability to gather desired information (not knowing where to go for information or what questions to ask).

Orientation toward information:

Categorized as information seeking (e.g.: wanting to more fully understand one’s risk, collecting more information about one’s family history, seeking out health care providers with expertise in breast cancer risk) or information avoiding (for various reasons, including: not worrying about one’s risk for now, the belief that knowing more about personal risk will not affect one’s chances of getting cancer, the sense that learning specific information would require action).

Cancer prevention beliefs:

Beliefs about whether cancer can be prevented or one’s risk of cancer lowered. Included beliefs about actions that can reduce risk, statements of faith and the role of God.

Generalized vs. specific cancer risk:

Categorized as specific (informant perceives that she is specifically at high risk for breast/ovarian cancer) vs. general (informant perceives that she is at high risk for cancer in general) perceived personal cancer risk.

Current health conditions:

An informant was categorized as having a current health condition if she described experiencing ongoing illnesses, diseases, or health risks other than breast/ovarian cancer risk. Major health conditions further defined as diagnosed major illnesses that limit everyday activities or qualify one for disability benefits, such as lupus, congenital joint disease, or HIV.

Finally, the narratives of White women were compared to those of African American women with respect to each dynamic we analyzed, again involving a minimum of two coders in identifying and exploring patterns within each node, and additional memos describing similarities and differences between the two groups.

Informants are identified in the text by pseudonyms chosen by women themselves or by the interviewer to match the ethnic origin and first letter of the informant’s first name. Women’s words are quoted verbatim, except for trimming of stuttering and repetitive verbalizations such as ‘you know,’ or ‘um’; race and age are also indicated alongside each quote.

Results: Tracing the Precursors of Racial Differences in Risk-Management Decisions

Our informants’ narratives reveal both the risk-reduction decisions they have made and the wide range of psychological, interpersonal, and social dynamics that contribute to those decisions. We begin by describing usage rates for each risk-management method within this sample, and then work backward from those decisions through the three layers that underlie them. In each layer, differences between racial groups are prominent.

Risk-Reduction Decisions

Among our fifty participants, there were noticeable racial differences in the use of several risk-reduction interventions (see Table 1). For aggressive prevention methods (prophylactic surgeries and chemoprevention), we report both the proportion of women who actually chose to use that option, and the (more inclusive) proportion who either chose it or were leaning toward doing so. Seven women, all White, had already chosen to undergo prophylactic oophorectomy at the time of the interview, and another two White women were leaning toward having this surgery. Similarly, four women, all White, had already chosen to undergo prophylactic mastectomy; one additional White woman and one African American woman were also leaning toward having their breasts removed in the near future. Chemoprevention was a less popular choice among all women, chosen by only two Whites and one African American, with one additional White woman leaning toward pursuing it. Enhanced surveillance was the most popular risk-management option across groups, with 80% of African American and 77% of White women choosing more frequent screening, tests other than mammograms, and/or an early start to screening. Finally, genetic testing – often seen as a necessary precursor to aggressive preventive action – was also more widely used by Whites, with 67% of White interviewees (or their cancer affected relative) having been tested, vs. only 20% of African American interviewees (or affected relatives).

Table 1.

Use of Breast Cancer Risk-Management Methods

	African American (n=20)	White (n=30)
Prophylactic Mastectomy
Participant chose	0% (0)	13% (4)
Participant chose or is leaning toward	5% (1)	17% (5)
Prophylactic Oophorectomy
Participant chose	0% (0)	23% (7)
Participant chose or is leaning toward	0% (0)	30% (9)
Chemoprevention
Participant chose	(5%) (1)	7% (2)
Participant chose or is leaning toward	(5%) (1)	10% (3)
Enhanced Surveillance
Participant has used	80% (16)	77% (23)
Genetic Testing
Participant or affected relative	20% (4)	67% (20)

These findings are consistent with those reported by other scholars. The use of in-depth interviews in this study, however, allows us to delve more deeply into the layers of decision-making dynamics that might explain these patterns. Because knowledge and uptake of breast cancer screening tests are relatively universal, the following exploration focuses on the other four key risk-management options: genetic testing, prophylactic mastectomy, prophylactic oophorectomy, and chemoprevention.

Specific Information about Risk-Management Options

The layer of information most proximal to risk-management decisions is specific information about one’s cancer risk, and about the relevant interventions one might consider. Genetic testing offers women individualized risk information that is difficult to replace; conversations with a genetic counsellor or other testing professional often also provide prevention information tailored to one’s own risk level. Undergoing genetic testing, however, requires both that a woman has heard of it and that she perceives of it as a personally relevant option. In our sample, all 30 White women had heard of genetic testing before the interview, while 3 of the African Americans (15%) had not. Even among those who had heard of the test, however, African Americans had a much lower rate (18% vs. 67% for Whites) of having been tested themselves, or having had their cancer-affected relative tested. Many White informants told stories about a relatively smooth path from a relative’s cancer, to a physician’s recommendation that she herself be tested, to her own swift decision to test. Others experienced a time gap between hearing about and undergoing genetic testing, but in the end were clearly motivated to do so. Chloe (White, 63) explained,

I wasn’t nervous. I just thought, ‘This’ll be great to know.’… I was doing it for myself but I was really doing it for [my daughter] and for my siblings because if I found out something then they should have [the test] too.

African American informants, in contrast, mostly reflected on why they had not (yet) had genetic testing. In many cases, women had considered testing – either on their own or after a doctor’s suggestion – but had not yet felt ready to move ahead. Financial constraints were also a common impediment to the genetic testing women might otherwise undertake.

In addition to deciding whether to undergo genetic testing, women at elevated risk may be faced with choices about undertaking aggressive prevention. Having prophylactic surgery or taking chemoprevention requires women to be aware these options exist, and to have sufficient information about them (see Table 2). Most participants in our study (84%) had heard of prophylactic mastectomy, but the proportions were unequal by race: 93% for Whites vs. 70% for African Americans. Among those who were aware of this option, about two thirds were positively disposed toward it, saying that they had chosen it, were leaning toward it, were considering it, or would consider it in some imaginable future circumstance. All but one (90%) of the positively disposed African Americans indicated that they would consider prophylactic mastectomy under some potential circumstance (such as if a genetic test came back positive, or a doctor specifically recommended it for them). For example, Ramona (African American, 38) explained:

Table 2.

Awareness of and Disposition toward Aggressive Prevention Options

	Prophylactic Mastectomy		Prophylactic Oophorectomy		Chemoprevention		≥ 1 Aggressive Prevention Option
	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)
Aware of Option1	70% (14)	93% (28)	35% (7)	60% (18)	20% (4)	57% (17)	75% (15)	93% (28)
Positively Disposed2	71% (10)	68% (19)	86% (6)	78% (14)	25% (1)	35% (6)	73% (11)	82% (23)
Chose3	-	11% (3)	-	39% (7)	25% (1)	12% (2)	7% (1)	29% (8)
Leaning toward3	7% (1)	7% (2)	-	11% (2)	-	6% (1)	7% (1)	18% (5)
Considering3	-	14% (4)	-	-	-	-	-	14% (4)
Might consider3	64% (9)	36% (10)	86% (6)	28% (5)	-	18% (3)	67% (10)	46% (13)
Not considering4	29% (4)	32% (9)	14% (1)	22% (4)	75% (3)	65% (11)	47% (7)	57% (16)

¹Percentages in the first 6 columns refer to percent of women aware of the option named in the top row; percentages in final two columns refer to percent of women aware of at least one aggressive prevention option (prophylactic mastectomy, prophylactic oophorectomy, and/or chemoprevention).

²Percentages in the first 6 columns refer to percent of women aware of the option who are positively disposed toward it (chose it, are leaning toward it, considering it, or would consider it under specific future conditions [labeled ‘might consider’]). Percentages in the final two columns refer to percent of women aware of and positively disposed toward at least one aggressive prevention option.

³Percentages in the first 6 columns refer to percent of women aware of the option who are in this particular category of positive disposition. Percentages in the final 2 columns refer to percent of women who are in this particular category of positive disposition with respect to at least one aggressive prevention option of which they are aware.

⁴Percentages in the first 6 columns refer to percent of women aware of the option but not considering it. Percentages in the final 2 columns refer to percent of women not considering at least one prevention option of which they are aware.

[If I tested positive for a BRCA mutation] I would think about [prophylactic mastectomy], think about the benefits, the pros, and the cons. Talk about it with my husband…If you have the gene it’s a huge possibility that you could get breast cancer… Because I see what [breast cancer] did to my grandmother. She died really young.

About half (53%) of positively disposed Whites also expressed a sense of openness to possible future surgery. For the other half, however, prophylactic mastectomy was a current consideration, something they were actively leaning toward, or a choice they had already made. Lainie (White, 59) expressed a matter of fact approach: ‘I decided before I even had the [genetic testing] results. If I [have] the mutation, I’m getting a mastectomy.’ Kathryn (White, 42) initially struggled, but soon made a confident decision.

[I watched] my Mom succumb to cancer and it devour her body. Literally, my mother died and the very next day I was on a plane…to Chicago for my [surgical] consult… I’m just a proactive person by nature. I make decisions and I move forward.

Similar patterns were observed in women’s disposition toward prophylactic oophorectomy. White women had again more often heard of this method of reducing cancer risk, but both groups were less aware of it than of prophylactic mastectomy. Among the 60% of Whites and 35% of African Americans who had heard of it, 78% of Whites and 86% of African Americans were positively disposed toward the option. All 6 of these African American women, however, considered it a hypothetical option they were willing to consider under some conceivable circumstance, but had never thought seriously about before our interview. Among the 14 positively disposed Whites, in contrast, only 5 fell into this hypothetically positive category, while 7 had already chosen prophylactic oophorectomy and two others were currently leaning toward having this surgery. Kea (African American, 34) expressed a typical style of hypothetical thinking:

It’s something I think I might do. I guess I would have to have a conversation …what does [the genetic result] mean and does it need to be done right away or can I try and pop out one kid first? ...The alternatives [cancer] seem a lot worse and much scarier and you can always adopt kids…

Bronja (White, 68) also did not yet have her genetic test results, but represented a much more concrete style of thinking about the oophorectomy option.

• Q: Are you sure you want to [have the genetic test]?

• A: Yes, I want to know.

• Q: So, you’re not on the fence?

• A: No. I want to take everything out.

Patterns in women’s disposition toward chemoprevention were harder to discern, because fewer women had heard of this option. In our sample, just over half of White women (57%) and even fewer African Americans (20%) had heard of taking a medication to reduce one’s chances of breast cancer. Among the four African Americans who knew about chemoprevention, three were not considering it at all, while one – Teresa (48) – had chosen it, and spoke matter-of-factly about taking the medication simply because her doctor recommended it to reduce her cancer risk. Among the 17 White women who had heard of taking a medication to reduce risk, 65% also had no inclination to consider it. Of those remaining, two had already chosen it, one was leaning toward it, and three indicated that they would consider it under appropriate future circumstances. This pattern may be similar to that observed for preventive surgery, where most African American women were hypothetically open to intervention but had not considered it in a real-life context. Given that only four African Americans had heard of chemoprevention, however, this assertion cannot be made with confidence.

General Information about Managing Breast Cancer Risk

If specific, detailed knowledge about individual prevention options is the necessary precursor to using them, then more general information about breast cancer risk and risk-management is the underlying layer that makes that specific knowledge possible. We examine women’s access to general information from two angles: access to the specialist health care providers most likely to understand and share it, and general access to information from any source (see Table 3).

Table 3.

Access to General Information about Breast Cancer Prevention

	African American (n=20)	White (n=30)
Has seen at least 1 risk-relevant specialist	15% (3)	70% (21)
Received risk-related information or referral from a PCP	30% (6)	52% (16)
Aware of at least 3 of 4 key risk-reduction options	45% (9)	73% (22)
History of seeking out risk-related information	35% (7)	73% (22)
History of avoiding risk-related information	20% (4)	30% (9)

Specialists vs. Primary Care Providers

Within our sample, African American women had considerably less access to the specialists who provide this information. Most (70%) White participants had seen at least one genetics, breast, or cancer specialist about their breast cancer risk, while only 15% of African Americans had seen any of these relevant specialists. African American women were correspondingly more likely to rely on primary care providers (PCPs) for breast cancer information.

Our data reveal the impact of these racial differences on women’s access to risk-management information. Elise (White, 22) explains her experience gathering information from a genetic counsellor and breast oncologist after her mother was diagnosed with breast cancer and tested positive for a BRCA mutation.

[The genetic counsellor] went through my whole family history with me…and then she explained to me how the BRCA1 gene works and my chances of inheriting it… And then she told me that if I have inherited it the chances of me developing breast cancer and its link to ovarian cancer…about when I would get tested for the gene and if I should get tested and how I felt about it… [We discussed] my concerns about getting tested… about how knowing that information or not knowing could influence the rest of my life… The [breast oncologist then] talked to me a little bit more about the chances of developing [breast cancer].

Elise further reflected, ‘I have much more peace now that I’m on [the facility’s] radar as a patient... and they see me every year.’ Such experiences – involving specialist appointments, detailed discussions of risk and risk-management options, and a resulting sense of reassurance – were recounted by several White participants. We heard few such stories from African American participants, who most often saw PCPs for all their health issues.

I’ve had the same family doctor since I was 19, love him – and he said, ‘You know, looking at the family history, let’s just start screening everything here [breast, cervical, and ovarian cancer screening]’ And once a year, we just screen. [Because] my doctor knows, he knows I’m afraid of cancer. (Tanya, African American, 53).

Tanya explained that she had a positive relationship with this PCP and trusted him to help manage her cancer risk. Still, he did not offer her information, but provided it only when she made specific requests and focused his attention solely on routine screening.

The majority of women in both racial groups reported that PCPs did not offer personal risk or prevention information. Asked about discussing risk with her doctor, Wyleisha (African American, 45) reflected, ‘…We don’t go into that much detail about it. He knows I have a higher risk, but it’s not like we talk about it that much.’ Some women’s PCPs actually refused their patients access to genetic testing or preventive procedures for which they were clinically eligible, because the PCP assessed these interventions as too drastic. Jojo (African American, 25) asked her OB/GYN if she should get tested for BRCA mutations and was told, ‘The only people who ever come back positive for that test are Jewish ladies.’ Barberette (White, 58) intensely feared ovarian cancer and asked repeatedly to have her ovaries removed, even when she was mostly through menopause and about to go under anesthesia for her hysterectomy, but her OB/GYN steadfastly refused to take out ‘healthy ovaries.’

Some informants’ PCPs did offer referral to a specialist or information about genetic counselling, testing, or a specific risk-management option. This happened more often to our White participants (52%) than to their African American counterparts (30%). Furthermore, White women were more likely to be told of all the risk-management options, or referred to a cancer or breast specialist; the PCPs of African Americans usually recommended only an early start to mammograms.

Finally, we observed racial differences in women’s trust in their physicians. A third of White women sought out specialists for second opinions or switched doctors when they were unsatisfied. No African American women reported this; they were more likely to have a single long-term PCP (for more than 10 years, and sometimes even 20–30 years). White women talked about trusting their physician based on his/her knowledge, and were willing to switch doctors or add a doctor to get better information. Megan (White, 41) noted, for example, ‘[My OB/GYN] is an expert…a great baby doctor. She’s very good, and I went to her for years…but I felt more comfortable having an oncologist, a specialist, doing [my oophorectomy.]’ African American women, on the other hand, seemed to base trust on the length of the relationship, in some cases even leaning heavily on a PCP who had cared for multiple generations of a woman’s family. Planning how she would handle a cancer-related situation, Jojo (African American, 25) reflected:

My doctors from home I listen to more because… they see my family and they know my history a little bit more than the doctors here. I mean, here…when you go see a doctor, you give them your history and stuff like that. But [at home]… my gynecologist was my mom’s gynecologist when she was going through breast cancer.

Seeking and Obtaining Risk-Related Information

Health care providers were only the one of the sources of risk-management information participants described. The number and type of information sources participants used was associated with their knowledge of the four key risk-reducing options (see Table 4). Women who knew of all four options reported 2–12 sources of information; women who knew of one, two, or three options reported 2–9 sources; the three women who knew about none of the options reported only 2–3 sources. These ranges were similar across race, as were the mean (5.58) and median (5) number of information sources women described. Among women who knew of 2 or fewer options most were African American, while among those who knew of 3 or 4 most were White. Women who knew about three or four risk-management behaviors most frequently reported health care providers, or friends and family (who were often themselves health care providers or cancer patients who had learned from their own providers), as their main source of information. For women who knew of two or fewer of these behaviors, online research was the most frequently reported main source, followed by PCPs, family and friends, and magazines. Strikingly, the only three participants aware of none of the key prevention options were African American women who said a PCP was their main source of information.

Table 4.

Main Source of Information by Number of Key Risk-Management Options Known¹

	4 options known		3 options known		2 options known		1 option known		0 options known		Total
Main Source of Information	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)	African American (n=20)	White (n=30)
Specialist	1	6	-	3	1	1	-	-	-	-	12
PCP	-	2	3	2	1	-	-	-	-	3	11
Online	-	1	1	1	1	1	2	1	-	-	8
Family/friends	-	1	2	1	2	1	-	-	-	-	7
Medical journals	-	2	1	-	-	2	-	-	-	-	5
Medical knowledge	1	1	-	1	-	-	-	-	-	-	3
Magazines/newspapers	-	-	-	-	1	1	-	1	-	-	3
Similar experiences	-	-	-	1	-	-	-	-	-	-	1
Total	2	13	7	9	6	6	2	2	0	3
	15		16		12		4		3

¹Out of four key risk-management options: prophylactic mastectomy, prophylactic oophorectomy, chemoprevention, and genetic testing

Another factor affecting women’s potential to obtain general information about breast cancer risk and prevention was their ability to perceive gaps in their knowledge and to fill those gaps. Our data suggest important racial differences in these dynamics. Both White and African American women mentioned general information gaps that indicated lower levels of risk-related knowledge. These included not knowing one’s full family history, not understanding the importance of family history to individual risk, being unsure when to start mammograms, and a general lack of education about cancer. More specific knowledge gaps, such as about the hormonal after-effects of oophorectomy, statistical chances of inheriting a BRCA mutation, or long term effects of chemoprevention, were articulated mostly by White women. In addition, only White participants recalled moments of not knowing where to go for the information they needed. In 5 of the 6 cases where this came up, participants were faced with particularly weighty information-gathering and decision-making needs as a result of testing positive for a genetic mutation.

Perceived knowledge gaps also raised the related issue of women’s orientations toward seeking out missing information. These orientations were not constant across time, with many women describing both information-seeking and information-avoiding moments. It may be that White participants were more consciously aware of these information orientations, as more Whites than African Americans told both information-seeking and information-avoiding stories. Information-seeking behavior included searching online to verify information, reading medical journals to stay aware of the most recent information, and asking doctors for advice about what to do. Information seeking was often motivated by a desire to fill in gaps related to family history and genetic risk. Tamara (African American, 38) started researching her family history on Facebook after realizing the multiple cases of cancer among relatives could not be coincidental, and Sharon (White, 26) sought out genetic testing to compensate for family history information that was not available because her Ashkenazi Jewish grandmother had fled Austria as a child during the Holocaust.

Other participants described moments when they did not want more information. For White women, such information avoidance most often stemmed from a desire to avoid facing a dilemma about what steps to take next. Several avoided genetic testing, for example, because they presumed that finding out about a genetic mutation would require taking action they were not ready for - such as having their ovaries removed. Lucy (White, 45) feared that understanding her risk more thoroughly would only cause her anxiety:

I knew I wasn’t going to [have a prophylactic oophorectomy] because I knew I wanted to have kids, and so my whole thought was like, ‘Why possess this knowledge if I’m just going to end up living with it and it’s going to drive me crazy?’

For African American participants, information avoidance had different underlying motivations. Some had more pressing priorities in the form of other current health problems, while others assumed additional information would be futile because there was nothing to be done about risk.

For women who both perceived gaps in their risk-related knowledge and felt motivated to remedy them, engaging with a health care provider was often the next step. Again here, we observed racial differences. Most White participants described asking physicians questions as an unremarkable event, something they undertook as a matter of course and felt generally at ease doing. As Tara (White, 31) reflected:

I tend to have a pretty open dialogue with my doctors, so if I have any specific concerns, I feel comfortable bringing them to them ... [I say, for instance,] ‘I read this, can you help me understand what’s going on or how much of a concern is this really?’

For African American women, this ability to purposefully request medical information was less taken for granted; several recounted consciously preparing to be able to ask the right questions of their doctors. Tiffany (African American, 56) explained that she looks online for information first ‘so that when I do go to a doctor I’m a little more informed on my end instead of letting him just, you know, tell me things, that I know a little bit so I can say yes or no.’

Perceptions of Breast Cancer Risk and Prevention

Beneath the most proximal (specific knowledge) and mid-range (general information) layers shaping risk-reduction behavior lies the third, most basic layer of women’s perceptions relevant to breast cancer risk and prevention. These have to do with how women understand the broad concept of cancer prevention, how they perceive their own risk for cancer, and how breast cancer risk fits into the larger complex of personal health problems and risks with which they are coping (see Table 5).

Table 5.

Perceptions of Breast Cancer Risk and Prevention

	African American (n=20)	White (n=30)
Identifies ≥1 risk-reduction method	45% (9)	80% (24)
Belief that God is in control of health	25% (5)	10% (3)
Generalized notion of cancer risk	35% (7)	7% (2)
Coping with other health issues	50% (10)	33% (10)

When asked whether it is possible to alter one’s cancer risk, 80% of White women identified at least one risk-reduction method, compared to only 45% of African American women. Methods identified from this open-ended question included healthy diet, exercise, preventive surgeries, or praying to God. In keeping with differences in knowledge of specific risk-reduction options (described above), White women more often discussed clinical interventions (50% vs. 15% of African Americans), while African Americans somewhat more frequently discussed the use of lifestyle methods to reduce cancer risk (65% vs. 43% of Whites). A subset of women expressed the belief that it is impossible to control whether or not one gets cancer, using phrases such as, ‘if it’s going to happen, it’ll happen’ or ‘you can’t control it.’ A quarter of African American participants (and 10% of Whites) volunteered the idea that ‘God is in control’ of their health.

In addition to understanding that it is possible to exert some level of control over one’s cancer risk, women at objectively high risk must also perceive this risk before interventions specifically recommended for high-risk women are likely to appear attractive. Most participants articulated a sense of personal risk for breast/ovarian cancer specifically, but a smaller group seemed to view all cancers as a single set, and to articulate instead a more generalized sense of risk for any kind of cancer (Author 2017). A third of our African American participants voiced this generalized sense of cancer risk, in terms like those used by Monica (African American, 32):

Cancer in general has always been a worry, a concern, of mine…so many people are being diagnosed with every type of cancer at this point…cancer in general is just common.

In contrast, most (93%) White participants perceived of their risk in more specific terms focused on breast and ovarian cancer. Kathryn (White, 42), whose mother had breast cancer first and then ovarian cancer 25 years later, expressed her worry about these specific illnesses:

I was told that I had an 87% [greater] chance of getting breast cancer in my lifetime than the average woman, and a 60% [greater] chance of ovarian cancer than the average woman. And you know, you watch it happen to your mom and you see many relatives die of cancer…I was really scared.

Lastly, our participants’ narratives reflect their need to cope with breast cancer risk within the broader context of whatever other health conditions they face. Diabetes, heart disease, and gynecologic issues were common health problems at the forefront of women’s minds; individual participants were also dealing with lupus, HIV, and other illnesses. Trish (African American, 57) explained how her main health worry gets prioritized over her cancer risk: ‘For me, it’s [concern] for my heart… I have had two heart attacks…I don’t think of the other diseases... I’m more concerned with my heart than I am with the cancer.’ The higher level of worry current health conditions engender is not the only factor causing women to de-prioritize cancer risk; coping with active illnesses also demands considerable time. Tessa (White, 55) describes this competition for her health-management resources: ‘I have rheumatoid arthritis. I’m always going to the doctor’s for that. So I tend to put the other stuff on “I’ll-do-that-later”.’ African American participants were more likely than Whites to be focusing their attention on current health conditions. Half reported chronic health conditions that necessarily took higher priority in their lives than dealing with the risk of a future breast cancer, whereas only one third of the White women reported health problems other than cancer risk.

Discussion

Risk-management usage rates within our sample echoed the findings of previous studies. A higher proportion of White than African American women pursued genetic testing and both prophylactic surgery options, while more than three quarters of participants in both groups were strongly oriented toward enhanced surveillance and few women in either group chose chemoprevention. The narratives of our participants, however, reveal new information about the complex personal and social dynamics that drive these decisions. In particular, women’s stories suggest that (a) the decision to engage in risk-management behavior rests on several accumulated layers of information, and (b) that the content of these layers differs substantially for African American and White women. We theorize that the most basic layer consists of a sense that one is at unusual personal risk for breast cancer and a perception that it may be possible to mitigate that risk to some degree. With these foundations in place, women can access and understand general information about breast cancer risk and prevention through medical specialists or other sources. That layer of information then makes it possible to consider specific potential interventions clinically relevant to one’s own case, and – for some women – to decide to pursue specific risk-management behaviors. This metaphor of a build-up of layers on which an ultimate behavioral choice may rest offers a new way to make sense of the contextual elements women describe as relevant to their risk-management decisions. Our analyses further suggest that African American and White women experience distinct dynamics at each of these levels, and that these differences might help explain racial differences in prevention behavior. While the sections above moved backward from the information layer most proximal to risk-management decisions (specific information) to the most distal layer (perceptions of risk and prevention), the paragraphs below discuss our findings in the (reverse) order more likely to reflect lived experience.

At the foundation of a progression that may eventually culminate in risk-management behavior are two important, although not necessarily conscious, insights: that one is at personally high risk for a future health problem, and that it might be possible to mitigate this risk in some way. Within this sample of women at elevated risk for breast cancer, some participants perceived of themselves as at specific high risk for breast and/or ovarian cancer while others perceived of themselves as at generalized high risk for any cancer. Furthermore, the perception of generalized risk was more common among African Americans. Having a generalized vs. a specific perception of personal cancer risk may produce differences in how women understand and absorb risk-management information: medical interventions specific to breast cancer risk might seem less relevant to women whose risk perception is more generalized. The concept of generalized vs. specific risk perception is a new finding of this project (Author 2017), but is consistent with prior findings that African American women have less knowledge about breast cancer risk factors (Katapodi et al. 2004; MacNew et al. 2010; Manning et al. 2013). White participants also more often volunteered examples of risk-management options, while African American women more frequently volunteered statements reflecting a belief that cancer risk cannot be managed.

African American participants also more commonly described dealing with health problems other than cancer risk, and more often had major ongoing health problems. Coping with one or more present health conditions reasonably pushes thinking about cancer risk down a woman’s personal priority list. Time, attention, and money are finite, and the need to invest these resources in current or chronic health problems reduces their availability for addressing possible future threats to health. The fact that our African American participants faced more current health problems and devoted fewer resources to risk-management decision making is consistent with racial disparities found in all major illness categories in the USA.

If a woman perceives of breast cancer risk as a relevant personal issue, believes it can potentially be mitigated, and devotes resources to addressing it, she may seek and/or absorb additional information. Our data support the presumption that health care providers are the most common source of this risk and prevention information, and that specialists are particularly important. As documented by prior researchers (Eismann et al. 2016; Forman and Hall 2009; Nelson et al. 2014; Sabatino et al. 2007; Tan et al. 2013), the providers offering detailed information about risk and prevention in this study were most often specialists. PCPs, by contrast, provided less detail about risk and risk-reduction, avoided making referrals, and rarely recommended risk-reduction methods other than surveillance. Previous studies have similarly found that PCPs are relatively unlikely to educate patients about individual risk severity or specific prevention options, usually because they lack the knowledge, time, or confidence necessary to do so (Sabatino et al. 2007; Tan et al. 2013). At the extreme, some PCPs even prevented participants from accessing interventions that clinical guidelines would indicate are appropriate for them. Because African American women were considerably less likely than Whites to have seen a specialist (a pattern also documented by Kolb (2006)), the information disadvantage experienced by women who rely solely on PCPs is not race-neutral. Furthermore, these patterns in access to specialists and the information they provide may be exacerbated by patterns related to trust: our African American participants indicated greater trust of physicians they had known for a long time (most often PCPs) while White participants indicated greater trust in doctors who knew the most about their current cancer-related concern (most often specialists).

Racial differences in possession of information were also reflected in our analyses of information gaps. While women in both racial groups expressed some gaps in their store of basic prevention-related information, many White women also enumerated gaps in the more specific, detailed information they wished to have, or felt they should have. In addition, White participants more commonly mentioned not knowing where to go to get specific information. In combination with their objectively more complete knowledge of prevention options, these patterns suggest that White participants more often built enough of a knowledge base to identify higher-level gaps in their knowledge and information-seeking abilities. It is possible, in other words, that African American women more frequently experience fundamental knowledge gaps that impede their ability to know what else they need to know. What African Americans seemed more aware of, in contrast, was the need to consciously prepare to ask their doctors clarifying or detail-eliciting questions, while White women seemed to take these information-oriented conversations for granted.

When women possess foundational information about cancer risk and cancer prevention, they can build a more specific understanding of key risk-management options. While the percentage of participants who had heard of these varied by option, the general trend was consistent across race: all participants had heard of breast cancer screening and almost all knew about genetic testing, most knew about prophylactic mastectomy, fewer were aware of the prophylactic oophorectomy option (despite the fact that this is the surgery recommended for all BRCA mutation carriers), and fewer than half had heard of chemoprevention. A higher proportion of White than African American participants knew about each risk-management option.

Usage of genetic testing – a common but not ubiquitous risk-management step for this sample – varied by race. Much lower rates among African Americans who knew of this option indicate that knowing one is at risk and that a genetic test exists is not sufficient; some other dynamic must mediate the path from awareness to action. That mediator could be a dynamic described above, such as the features of women’s relationships with specialists vs. primary care providers, or one’s sense that risk-management is possible. Alternatively, African American women might face more financial, interpersonal or other impediments, such that their path from knowing one carries a high familial risk of breast cancer to genetic testing is less smooth than that of their White counterparts.

Given the unique nature of the analyses we could conduct with our interview data, we were also able to move beyond usage rates to provide a novel, more nuanced exploration of women’s disposition toward their risk-management options. Like so many other layers that contribute to risk-management behavior, women’s disposition toward specific options seems to vary by race. Our analyses indicate that while women in both racial groups were open to the possibility of each prophylactic surgery under some future condition, only White women (half of those positively disposed toward either surgery) were currently and actively considering those options. As with the use of genetic testing, it seems that knowing about surgical options and even thinking they might be personally relevant does not result equally in movement toward undertaking those surgeries. The reasons for this racial difference in disposition require further exploration; it may stem from differences in the ways health care providers approach the issue with patients, differences in when women think prevention is appropriate, differential financial impediments, or other distinctions between racial groups.

The central limitations of this research have to do with its highly specific sample. The constrained realistic sample size for an in-depth qualitative interview study required that we focus only on the two dominant racial groups in the USA, but future research should investigate how the information-related dynamics we have explored here play out for other racial and ethnic groups. Our sample did not contain sufficient SES variation to allow in-depth exploration of how insurance-, income-, and poverty-related issues may affect women’s decision making, or how these class dynamics may be intertwined with racial ones. In addition, while drawing on women’s narratives about their own risk-related experiences allows us to generate a range of new insights, it also omits the perspectives of the health care providers (and others) critical to women’s information accumulation and risk-management decisions.

Conclusions and Implications

Racial disparities have been documented across a wide range of health-related determinants, processes, and outcomes. In our study of women at elevated risk of breast cancer, participants of all races, ages, and risk levels articulated strong motivation to educate themselves about cancer risk and to proactively protect their health. However, our results reveal that compared to their White counterparts, African American women faced additional burdens at every step along the risk-management journey. Taken together, these findings suggest that information gathering is quite a bit more complex than has previously been addressed, that information access and provider access are closely related, and that African American women may be systematically disadvantaged with respect to information-generating experiences.

Preventing cancer morbidity and mortality is the central goal of risk-management efforts for groups known to be at higher than average risk. But achieving this goal is only possible if all high-risk women have access to the layers of information and understanding necessary to engage in cancer screenings and preventive interventions. The results of this study exemplify the nuances of racial inequalities in health – the ways that structural, social, and interpersonal inequalities combine to influence risk-management choices. Furthermore, these results come from a study of women personally motivated to volunteer themselves for cancer-prevention research; the disparities among the broader population of high-risk women might well be even more substantial. Societal-level changes that improve health equity and remove systematic racial disadvantage could clearly improve the ability of women at elevated risk of breast cancer to make informed, empowered risk-management decisions. But it may also be possible to work upward, from insights about this particular population of women to changes that could have broader systemic impact.

One particularly fruitful place to begin might be with the initial and continuing medical education of healthcare providers. Primary care providers could be trained to follow up more systematically on indications of familial risk, to provide some degree of risk-management information, and to refer high-risk women for genetic testing and specialist care. Specialists could more routinely encourage high-risk women to pass risk-management information along to their biological relatives, or to suggest that those relatives consult with their healthcare providers about breast cancer risk and prevention. All healthcare providers could be educated about the relevance of risk information and risk-management options for African American women, and the current disparities in provision of this information across race. Considerable research remains to be done to confirm the patterns found in this study, and to more deeply understand the origins of gaps and challenges that systematically disadvantage African American women. At this point, however, the evidence suggests that researchers and clinicians aiming to prevent cancers and deaths among women at elevated risk must attend to the layers of information that underlie women’s risk-management decisions, and to racial disparities in a wide range of decision-making dynamics.

Data Availability Statement

The data that support the findings of this study are available on reasonable request from the corresponding author, TJP. The data are not publicly available due to the inclusion of information that could compromise research participant privacy.