Main

As medical science advances our understanding of the relationship between genetics and disease, tests for hereditary predisposition to disease will become more commonplace. Patients will increasingly face the question of whether to obtain genetic testing, and health care providers may be an important source of credible information. Understanding how patients obtain their genetic testing information and share it with family members is critical for assessing how well the informed decision-making process is working and where health care providers fit into this process. A current case in point is testing for mutations in the hereditary breast/ovarian cancer genes (BRCA1 and BRCA2).

The details of BRCA testing are complex and can be confusing. Only a small group of women are suitable candidates for BRCA testing (e.g., women with strong family history of premenopausal breast cancer), and even then, the majority of tests will be negative. BRCA1 and BRCA2 mutations are estimated to account for only 2–4% of all breast cancer cases for women between the ages of 15 and 69 at diagnosis, and for 2–9% of breast cancer cases for women diagnosed between the ages of 40 and 49.1 Estimates of lifetime breast cancer risk for BRCA1 and BRCA2 mutation carriers range from 40 to 95%, depending on the populations sampled and individual risk factors.1–6 Absence of a mutation does not imply negligible breast cancer risk. Instead, such individuals will still have, at a minimum, the baseline breast cancer risk of U.S. women with similar risk factors. BRCA1 mutations (and to a lesser degree, BRCA2 mutations) are also associated with increased risk of ovarian cancer.1,4 The estimated lifetime risk of ovarian cancer in BRCA1 carriers in a recent population-based study is 28%.1 If a BRCA mutation is identified, treatment options to reduce breast cancer risk include tamoxifen, prophylactic mastectomy, and prophylactic oophrorectomy.7,8 However, the efficacy of tamoxifen for mutation carriers has not yet been confirmed, and prophylactic mastectomy or oophorectomy may not be desirable for many mutation carriers.

Little is currently known regarding whether health care providers are discussing BRCA testing with their cancer patients. We found only one study, in which 16% of physicians of women currently under treatment for breast cancer reported talking with their patients specifically about BRCA testing.9 Long-term breast cancer survivors (5+ years) are typically reintegrated into mainstream practice because they no longer need specialized oncology care.

It is also not known if survivors of premenopausal breast cancer are discussing BRCA testing with their relatives. Such discussions may be important to allay fears, share information, and/or identify potential risks. We found only qualitative reports of communications within families regarding BRCA testing.10 Additionally, we found no population-based studies of BRCA knowledge, attitudes, and testing intentions in breast cancer patients or survivors. These individuals typically should be the first recipients of testing in a family. If the survivor does not have a BRCA mutation, it is unlikely that other blood relatives would benefit from BRCA testing.

The objectives of our study were (1) to determine whether a population-based sample of long-term breast cancer survivors speak with their health care providers and relatives about BRCA genetics; (2) to determine survivors'BRCA knowledge, attitudes, and intentions to obtain testing; and (3) to solicit opinions regarding preferred sources of information about BRCA genetics.

MATERIALS AND METHODS

Participants

Study participants were women enrolled in Group Health Cooperative of Puget Sound (GHC), a nonprofit health maintenance organization (HMO). Eligible participants had an initial diagnosis of invasive breast cancer or ductal carcinoma in situ (DCIS) 5 to 10 years prior to June 30, 1998, and were between the ages of 40 and 49 at the time of diagnosis. A total of 276 survivors were identified. This population was selected because women diagnosed under the age of 50 are more likely to have a heritable form of breast cancer6 and also to have children currently reaching child-bearing age where issues regarding BRCA mutations may be of heightened concern. Women younger than 40 were not selected because they were too few in number.

The primary data source for cancer diagnosis was the Cancer Surveillance System, one of nine standard U.S. Surveillance Epidemiology and End Results (SEER) cancer registries. Because some current GHC enrollees may have been diagnosed outside the SEER area prior to joining GHC, we also accessed self-reported information on breast cancer collected by GHC's screening program.11,12

Survey construction and dissemination

In January 1999, self-administered questionnaires were mailed to the study population in a four-wave approach, based on methods described in Dillman and Armstrong et al.13,14 We used the Health Belief Model and the Theory of Reasoned Action as the theoretical framework to guide survey development and analysis.15,16 Applied to a genetic testing scenario, the models suggest that sociodemographics, knowledge of genetics, attitudes toward testing, and opinions of others are important factors that influence intent to obtain genetic testing. In turn, intent to obtain testing is a key factor in the decision process to actually undergo testing. Our survey, therefore, included questions on each of these factors. For comparability, most knowledge and attitude questions were similar to those used in previous research.17–19 In addition, respondents were asked to assess the helpfulness of various types of information sources on BRCA. “Helpfulness” was defined as an answer of 3, 4, or 5 on a 5-point Likert scale from “not at all” to “very much.”

The survey was developed and pretested with assistance of a focus group of five long-term breast cancer survivors enrolled at GHC. These women represented a range of ages, lifestyles, and continuing problems associated with breast cancer. Our study was conducted under the supervision and guidelines of the GHC Center for Health Studies Human Subjects Review Committee.

Data analysis

Data were analyzed in SPSS Version 7.5 for Windows and SAS for Windows, Release 6.12. Analyses included descriptive statistics, ANOVAs, Pearson correlations, and bivariate and multivariate unconditional logistic regression. Regression analyses were used to describe the association between intent to obtain testing and other factors in the genetic testing decision process. For ease of interpretation and consistency with previous studies, the regression variables were dichotomized. However, to ensure that important information was not lost by dichotomizing, the regression analyses were also conducted using continuous and categorical variables.

Two separate dependent variables were used as measures for intention to obtain BRCA testing: (1) “I plan to have a genetic test for breast cancer only if my insurance covers the cost”; and (2) “I plan to have a genetic test for breast cancer, even if I have to pay for it myself.” Both measures used 4-point Likert scales ranging from “strongly disagree” to “strongly agree,” which we then dichotomized, comparing those who “somewhat or strongly agreed” to those who “somewhat or strongly disagreed.”

The independent variable “knowledge” was dichotomized based on whether more than 50% of seven true/false BRCA knowledge questions were answered correctly. Other independent variables included education, employment status, age, income, marital status, communication with health care providers regarding genetic testing, communication with family regarding genetic testing, family history of breast and ovarian cancer, other health behaviors such as smoking status, and attitude toward testing.

We used two dichotomous independent variables for attitude: “benefits” and “limitations” of BRCA testing. These variables were developed from 14 attitude questions with responses ranging from “strongly disagree” to “strongly agree” on 4-point Likert scales. Factor analysis (principal components analysis with Varimax rotation) indicated that the benefits and limitations comprised two distinct factors, with good reliability (Cronbach's alpha coefficient for standardized variables = 0.82 and 0.78, respectively). The variables representing the “benefits” and the “limitations” of BRCA testing were then dichotomized, comparing those whose overall “somewhat or strongly agreed” to those who “somewhat or strongly disagreed.”

RESULTS

The survey response rate was 79% (217 of 276 women). An additional 14 women were mailed the survey but excluded as eligible participants because two had recently died, four had undeliverable addresses, and eight did not believe they met the 5–10 year invasive or DCIS diagnostic criteria. An analysis of nonrespondents conducted using GHC's databases indicated no statistically significant differences in means between respondents and nonrespondents for age at diagnosis, estimated number of primary care visits/year, estimated total health care costs/year, years since initial diagnosis, and type of cancer (invasive or DCIS).

Sociodemographic and medical characteristics of respondents are shown in Table 1. The mean age at initial diagnosis was 44.4 years, and the average time since that diagnosis was 7.3 years. Fourteen percent experienced a recurrence and the average time since recurrence was 2.5 years.

Table 1 Selected sociodemographic and medical characteristics of 5- to 10-year breast cancer survivors
Full size table

Respondents reported more discussions about breast cancer genetics/ BRCA testing with relatives than with health care providers (Table 2). Survivors were most often the conversation initiators in the family (70.5%), followed by daughters (26.8%) and sisters (21.4%).

Table 2 Communications about BRCA testing among breast cancer survivors' families and health care providers
Full size table

Table 3 shows BRCA awareness and knowledge scores and the percentage of women correctly answering seven true/false questions. The mean number of correct true/false answers was two. Of women who were conversation initiators, 62% answered less than half the true/false BRCA questions correctly. Of women who spoke with relatives specifically about seeking BRCA testing, 12% incorrectly thought they must have an altered gene because they already had breast cancer and 49% did not know the answer to this question. No women who indicated that a brother or son was the conversation initiator correctly answered the following true statement, “A father can pass down the altered BRCA genes to his daughter.”

Table 3 Knowledge of BRCA genetics among 5- to 10-year breast cancer survivors
Full size table

Most women had a positive attitude toward benefits of BRCA testing (Table 4). Over 60% agreed or strongly agreed with all seven BRCA positive attitude statements. Typically less than half agreed or strongly agreed with seven BRCA negative attitude statements. Women who answered more than 50% of the knowledge questions correctly were approximately 47% less likely to have a positive attitude toward testing compared with those who had lower knowledge levels (P = 0.095). No significant relationship between knowledge and negative attitude was observed.

Table 4 Attitudes and intentions regarding BRCA testing among 5- to 10-year breast cancer survivors
Full size table

Intent to obtain testing differed according to whether the individual must pay for the test or whether their insurance covers it (Table 4). Only 26% intended to obtain testing if they have to pay for it, and 93.1% of these women were willing to pay no more than $200, out of a range from $25 to $2,000. Approximately 67% planned to obtain testing only if their insurance covers it.

Two variables were significantly associated with intent to obtain testing “even if they had to pay for it,” based on unadjusted logistic regression analyses (Table 5). First, women with a more positive attitude toward testing were 5.8 times as likely to intend testing “even if they had to pay for it,” compared with women with a less positive attitude (P = 0.005). Second, women who talked with relatives about BRCA testing were over twice as likely to intend testing, compared with those who had not spoken with relatives (P = 0.020). Similar to the bivariate analyses, the multivariate logistic regression analysis showed an association between “intention to obtain testing even if the individual has to pay for it” and two variables: positive attitude toward testing (odds ratio [OR], 5.5; 95% confidence interval [CI], 1.6–18.9) and talking about testing with relatives (OR, 2.2; CI, 1.1–4.4).

Table 5 Factors associated with intent to obtain BRCA testing among 5- to 10-year breast cancer survivors
Full size table

Four factors were significantly associated with intent to obtain testing “only if insurance covers it” (Table 5), based on unadjusted logistic regression analyses. First, married women were 2.4 times as likely to intend testing, compared with unmarried women (P = 0.020). Second, women with a more positive attitude toward testing were 2.2 times as likely to intend testing, compared with women with a less positive attitude (P = 0.035). Third, women with any daughters were 2.2 times more likely to intend testing, compared to women with no daughters (P = 0.011). Lastly, women with any first-degree family history of breast cancer were 47% less likely to intend to obtain testing only if insurance covers it compared to women with no first-degree family history (P = 0.054).

Multivariate logistic regression demonstrated an association between “intent to obtain testing only if insurance covers it” and three factors: (1) marital status (OR, 3.91; CI, 1.56–9.81), (2) positive attitude toward test (OR, 3.72; CI, 1.60–8.69), and (3) having any daughters (OR, 2.27; CI, 1.07–4.82). Unlike the unadjusted regression results, the multivariate results did not show a significant relationship between having a first-degree family history of breast cancer and “intention to obtain testing only if insurance covers it.” To confirm that dichotomizing variables did not result in the loss of important information, we also performed the above regression analyses using continuous and categorical variables and obtained similar results.

It is possible that women visiting their health care providers more often may have different opinions than other respondents. However, using ANOVAs and Pearson correlations, we did not observe significant associations between annual number of physician visits and the following variables: knowledge of BRCA genetics, attitude toward BRCA testing, and intention to obtain testing.

It is also possible that women at higher risk for a BRCA mutation might have different opinions than women at lower risk. Approximately 31% of our respondents may be at higher risk for a BRCA mutation, because they had breast cancer under the age of 50 and had a first-degree relative with breast and/or ovarian cancer. Women with a family history answered 39% of the true/false BRCA knowledge questions correctly, while women without the family history answered 24% of the questions correctly (P = 0.001). Also, 23% of women with a family history heard “a fair amount or a lot” about genetic testing for breast cancer, while only 15% of women without a family history heard “a fair amount or a lot.” Additionally, women with a family history were more likely to have discussed genetic testing with a health care provider (19% vs. 4%) (P = 0.001).

Respondents thought a number of sources and types of information on BRCA genetics would be helpful. Most women (80.6%) indicated that an informational booklet on BRCA genetics would be helpful, followed by discussions with doctors or women's health care specialists (79.8%), discussions with health educators (70.4%), informational videos (66.1%), summaries of Internet sites with genetic testing information (60.4%), informational group classes (48.9%), and discussions with a group of breast cancer survivors (47.9%).

DISCUSSION

Our results suggest that health care providers and long-term breast cancer survivors are not discussing BRCA genetics with each other. Only 8.3% of our sample spoke with health care providers about genetic testing, while 53% spoke with relatives. We also found that survivors played key roles as BRCA conversation initiators in their families and, therefore, may be an important information conduit for BRCA testing in the family. However, the survivors had a limited understanding of BRCA genetics, which could compromise an informed decision-making process for themselves and their family members. Our respondents had lower levels of BRCA knowledge than reported in existing studies involving female first-degree relatives of affected individuals, members of hereditary breast-ovarian cancer families, and general population studies. Only 17.3% our sample reported hearing “a fair amount” or “a lot” about genetic testing for breast cancer (Table 3). In contrast, other studies reported awareness ranging between 35% and 83%.17,20,21 Also, our sample correctly responded to an average of only 2 of 7 true/false questions, which is lower than reported in several other studies.17,18,22 Even among the higher risk group in our sample, knowledge of genetic testing for breast cancer was approximately 35% lower than prior studies, and discussions about genetic testing with health care providers were minimal.17,18,22

Several explanations regarding the lower level of knowledge in our study participants are possible. Knowledge may be limited because so few have spoken about breast cancer genetics with their health care provider. Additionally, unlike several previous studies, no participants in this study (to our knowledge) are involved in research protocols specifically targeting high-risk families of known mutation carriers. Further, our survey specifically targeted long-term breast cancer survivors in a population-based setting. Many previous studies were not population-based, and no previous studies looked at long-term survivors. Also, other studies may have focused on more recently diagnosed women, who may have been actively discussing genetics with their friends, family, and physicians; such individuals may have a better recollection of their discussions. Finally, studies several years ago were occurring when the discovery of BRCA1 and BRCA2 were covered in the news media. The level of knowledge in breast cancer survivors at that point in time may have been higher due to more active news coverage.

Consistent with several previous studies, the benefit of BRCA testing cited most often (88.3%) was learning about children's or relative's risk.18,19,23,24 The disadvantage of testing cited most often (50.9%) was concern about losing insurance if an altered gene is found. This is consistent with the results of Lerman et al.18

Consistent with prior research, respondents with more knowledge about BRCA genetics had a less positive attitude toward the benefits (P = 0.095) and a less positive attitude was associated with lower intention to obtain testing.18,23,25 This suggests that providing women with information on breast cancer genetics may not necessarily increase their intention to obtain testing. Increased understanding among survivors and their families- not necessarily increased intention to obtain testing among women who may not benefit from testing- is a desirable outcome from a medical perspective.

Our study is the first to highlight how strongly the intent to obtain genetics testing is tied to insurance coverage for breast cancer survivors. Only 26.4% would seek BRCA testing if they had to pay for it themselves, whereas over 67% would seek testing if their insurance covered it. We found only one other study that distinguished between intent to obtain testing when the test is free versus when payment is required.26 In that study, conducted among women in a waiting room of an OB/GYN office and at a mammography center, approximately 60% would accept the test if they had to pay, whereas over 93% would accept the test if it were free. Other studies involving first-degree relatives of breast cancer patients, members of hereditary breast-ovarian cancer families, and the general population did not consider the influence of test cost and found moderate to high levels of intention (60–95%).21,26,27

In addition to attitude toward testing and cost issues, familial factors may influence the testing decision. Such factors include the survivor's marital status, communications with relatives, having daughters, and having a first-degree family history. These factors were significantly associated with intent, depending on insurance coverage (Table 5).

Health care providers and genetic counselors may have a variety of BRCA education options available for breast cancer survivors. Respondents indicated that written materials, discussions with doctors or health educators, videos, and the Internet would be helpful ways to obtain BRCA information. Brochures at mammography centers and doctors' offices are potential options, as are Internet sites. Respondents preferred written materials (80.6%) in contrast to a recent focus group session by Bernhardt et al., in which interactive group education with supplementary printed materials was preferred.25 Only 47.9% of our respondents indicated discussion groups would be helpful. It is possible that the results of Burnhardt et al. are more reflective of a self-selected sample already predisposed to participate in groups. Nonetheless, both results show a desire for written material.

A limitation of our research is the cross-sectional study design, from which causality cannot be established. For example, it is possible that women spoke with relatives about BRCA testing after having already decided to undergo testing, and not that speaking with relatives increased their intent to obtain testing. Nonetheless, these associations provide valuable insight into factors in the genetic testing decision process, and verify some of the work of previous researchers in other populations. Another limitation is that we do not have information regarding whether these women were actually tested. Testing is covered in this health plan after discussions with a geneticist, but we did not have access to data on whether this service was used. However, only five women in the study indicated that they had heard “a lot” about genetic testing for breast cancer. If a large number of women in our cohort had undergone testing, we would expect this number to be higher.

Additionally, we did not have information regarding our respondents' sources of information about BRCA genetics. Also, it would have been interesting to compare respondents' self-reported knowledge of BRCA to chart reviews or a survey of health care providers to identify what BRCA discussions or education, if any, may have transpired and when it occurred. Finally, our measure of willingness to pay for BRCA testing could have been more refined, such as “willingness to pay if the insurer covered testing but the individual had a 20% copayment of $516.”

In summary, our study is the first population-based study of long-term breast cancer survivors' communications, knowledge, attitudes, and intentions regarding BRCA testing. The demographic characteristics of women enrolled in our HMO are similar to the Puget Sound area of Washington State, though slightly higher in income level.28 Thus, our results can likely be generalized to survivors in this region who were diagnosed between 1983 and 1998 when they were 40 to 50 years old, and possibly to women of similar sociodemographic and survivorship status elsewhere in the country. Given that our sample includes women of higher incomes with access to medical care and these individuals have very limited knowledge of breast cancer genetics, we have no reason to suspect that conditions would be better in the general U.S. population.

Long-term breast cancer survivors have less BRCA knowledge than expected from previous studies. Our results suggest that increasing survivors' knowledge of BRCA is unlikely to result in a large influx of testing requests. Rather, it will provide information needed to make well-informed decisions. Though health care providers currently appear to be playing a limited role in the BRCA testing decision process, long-term survivors would find their input helpful. Providing genetic information to patients in a variety of ways will be an important new challenge for medical care providers.