Fax: (410) 955-0241
Utilization of BRCA1/2 genetic testing in the clinical setting
Report from a single institution
Article first published online: 15 MAR 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 6, pages 1876–1885, 15 March 2002
How to Cite
Lee, S.-C., Bernhardt, B. A. and Helzlsouer, K. J. (2002), Utilization of BRCA1/2 genetic testing in the clinical setting. Cancer, 94: 1876–1885. doi: 10.1002/cncr.10420
- Issue published online: 15 MAR 2002
- Article first published online: 15 MAR 2002
- Manuscript Accepted: 26 NOV 2001
- Manuscript Revised: 20 NOV 2001
- Manuscript Received: 17 SEP 2001
- Ministry of Health, Singapore, through the Human Manpower Development Program (HMDP) Fellowship program
- genetic testing;
- hereditary breast carcinoma;
- hereditary breast and ovarian carcinoma
Clinical testing for BRCA1/2 has been available since 1996. Interest in testing in the research and hypothetical situations has been consistently high, but there have been limited reports on its clinical utilization.
This is a retrospective study of BRCA1/2 test utilization by high-risk patients who were seen at the Johns Hopkins Breast and Ovarian Surveillance Service.
Between February 1996 and December 1999, 258 families who had at least a 10% chance of carrying a BRCA1/2 mutation were offered genetic testing. Of these, 26 families seen between February 1996 and October 1996 had access to free testing. Overall, 68 of 258 (26%) underwent genetic testing. Educational level, number of children or daughters, breast carcinoma screening behavior, smoking and drinking behavior, perceived risk of breast carcinoma, and family history was not associated with test utilization. Eligibility for free testing, prior history of breast or ovarian carcinoma, Ashkenazi Jewish versus non-Ashkenazi Jewish heritage, genetic risk category, and age category were associated with test utilization, and in multivariate analysis, the first three remained statistically significant factors associated with genetic testing. Only 26% of the 50 patients who did not have access to free testing sought insurance reimbursement, of which greater than 50% (7 of 13) had a prior diagnosis of breast or ovarian carcinoma.
The actual utilization of BRCA1/2 genetic testing in a clinical setting is lower than in the research and hypothetical settings. Potential obstacles include cost, fear of insurance discrimination, and need to involve an affected family member in the testing process. Cancer 2002;94:1876–85. © 2002 American Cancer Society.
Breast carcinoma is the most common female malignancy in the United States, with an age-standardized incidence of 103 per 100,000.1 An estimated 5–15% of breast carcinoma is hereditary and caused by single gene mutations,2 such as the BRCA1/2 genes.3–5 The BRCA1/2 genes account for approximately 30% of breast carcinoma–only families6 and 50% of breast and ovarian carcinoma families.6 The American Society of Clinical Oncology recommends extensive discussion of the potential benefits, risks, and limitations before BRCA1/2 genetic testing, as identification of mutation carriers has a presumed but not established medical benefit.7 In practice, medical consideration of genetic testing is warranted when the patient's estimated risk of being a mutation carrier is at least 10%.7
Before the widespread availability of clinical BRCA1/2 testing, several studies reported a high level of interest in cancer susceptibility testing among physicians and patients in hypothetical8–16 and research situations.2, 17 The motivations governing genetic testing uptake in these study situations are likely to be different from that in clinical practice, and there is currently limited literature on the actual uptake of BRCA1/2 testing in the clinic.18 We report one of the first large clinical series on the rate and pattern of BRCA1/2 genetic test utilization and examined the utilization of genetic testing and associated factors among high-risk women.
This is a retrospective study of the utilization of BRCA1/2 clinical testing and the factors associated with genetic test utilization in a clinical population of women at increased risk for hereditary breast and ovarian carcinoma.
The study population comprised female patients who were seen in the Breast and Ovarian Surveillance Service at Johns Hopkins Hospital as a new consultation between February 1, 1996 and December 31, 1999 (see Fig. 1). The data were analyzed in August 2000, allowing a minimal follow-up period of 8 months, because patients may take weeks to months after the counseling session to make a genetic testing decision. Most patients were physician-referred and resided in the Baltimore–Washington area. Patients who had a personal and/or family history of breast and/or ovarian carcinoma and whose families were estimated to have at least 10% chance of carrying a BRCA1/2 mutation, to whom genetic counseling was provided and genetic testing discussed, were eligible for this study. Families that had more than one member attend the clinic on separate occasions were considered a single unit in our analysis, designating the person first seen as the proband. Patients whose families were estimated to have less than 10% chance of carrying a BRCA1/2 mutation were excluded, even if genetic counseling may have been provided and genetic testing discussed, because testing would not be considered strongly as an option on clinical grounds because of low yield. Patients who had been tested previously, or who had a family member who had been tested before the patient visiting our service, whether or not a BRCA1/2 mutation was identified, were excluded. These latter individuals formed a small minority of our study population and were excluded from analysis because the higher yield of genetic testing in this subgroup is potentially confounding. Only three male patients were seen during the study period, and these were excluded to maintain the homogeneity of our study population.
Data were obtained from review of the medical records, and a previsit questionnaire was completed by each patient. Basic demographic information (age, ethnic group, marital status, years of education, number of children and daughters), personal cancer history, screening practices (breast self-examination, mammography), lifestyle behavior (smoking and alcohol consumption), perceived lifetime risk of invasive breast carcinoma, and a three-generation family cancer history were obtained from the previsit questionnaire. The estimated familial risk of carrying a BRCA1/2 mutation as assigned by the provider and as communicated to the patient during the visit was obtained from the patient's written consultation note. Quantified risks were provided for some cases, whereas for others, descriptive terms were used to convey the information. For the purpose of this study, we considered the family to be “low-risk” when the estimated chance of carrying a BRCA1/2 mutation in the family was recorded to be between 10% and 20%. The family was considered “moderate-risk” and “high-risk” when the estimated risk was recorded to be between 20% and 50% and greater than 50%, respectively. When descriptive terms were used, we considered the family to be low-risk when the consult note reported that the family history is “mild,” “somewhat suggestive,” “has some suspicion,” “has a small risk,” or is “mildly suggestive” of a BRCA1/2 mutation. When the consultation note used terms such as “consistent,” “concerning,” “likely,” “maybe,” “possible,” “suggestive,” “reasonable,” or “suspicious,” we considered the family to be of moderate-risk. When the consult note used terms such as “highly suggestive,” “high probability,” “high-risk,” “significant,” “strongly suggestive,” or “strong possibility,” we considered the family to be high-risk.
Patients who were tested for mutations in the BRCA1/2 genes were identified from a master record kept in the genetic counselor's office (B.A.B.). Information regarding the timing of the test, person in the family who was initially tested, mode of payment, type of test, and test results were collected.
Description of Service Provided
The Breast and Ovarian Surveillance Service at Johns Hopkins Hospital was established in 1989 and is staffed by two attending physicians, two genetic counselors, and two nurse practitioners. An attending physician, who conducted the consultation with a genetic counselor and/or a nurse practitioner, saw each patient. The patient's family history was reviewed, and the patient given a risk estimate of the family carrying a BRCA1/2 mutation. Whenever possible, the patient's reported family history of cancer was verified through pathology reports. The Couch model,19 first published in 1997, was used to quantify the risk when deemed appropriate, because this was one of the most comprehensive models available during the study period. Because the Couch model provided risk estimates pertaining only to BRCA1, the number was doubled empirically to give patients an approximate estimate of their family's probability of carrying a BRCA1 or BRCA2 mutation. In some patients, the risk estimate was not quantified because the Couch model had not yet been published, because it was not appropriate to apply the Couch model, because the risk was difficult to quantify because of incomplete or uncertain family history, or because of provider preference. In these cases, clinical judgment was used to estimate the risk, and a descriptive term was used to convey the information to the patient and accordingly recorded in the consultation note.
During the genetic counseling session, information regarding the mode of inheritance of the BRCA1/2 genes and the associated cancer risks were discussed. Nondirective counseling was provided in all cases, because BRCA1/2 testing has not been established to result in definitive medical benefits. Patients were informed that the lifetime risk of breast and ovarian carcinoma in BRCA1/2 mutation carriers is 50–85% and 15–45%, respectively.20, 21 The benefits (improved cancer risk assessment, relief from uncertainty and anxiety about cancer risk, useful information for family members, lifestyle decision making, empowerment, etc.), risks (psychologic distress, loss of privacy, discrimination by employers and insurers, change in family dynamics, false sense of security, etc.), and limitations (not all mutations are detectable, uncertain significance of some mutations, negative result not informative unless a mutation has been identified in the family, results indicate probability and not certainty of developing cancer, unproven efficacy of most interventions, etc.) of genetic testing were discussed. Patients also were given screening recommendations based on their family history, as well as information on the surveillance options for breast and ovarian carcinoma for BRCA1/2 mutation carriers as recommended by the Task Force convened by the Cancer Genetics Studies Consortium.22 Preventive options (tamoxifen chemoprevention,23–25 prophylactic mastectomy,26, 27 and oophorectomy27, 28) were discussed as appropriate.
In general, patients were offered full sequencing of BRCA1/2 including all the coding exons and splice-site junctions. Beginning in 1997, patients of Ashkenazi Jewish heritage were offered the option of a screening test for the three common founder mutations, 185delAG and 5382insC in BRCA1 and 6174delT in BRCA2. Patients of Ashkenazi Jewish heritage were informed that these three mutations account for most identifiable BRCA1/2 mutations in their ethnic group.29 All patients who undertook genetic testing signed an informed consent form. Testing was performed at Myriad Genetics Laboratories (Salt Lake City, UT). The cost of the Ashkenazi screening panel was $300–350, whereas that of full gene sequencing was $2400. Patients could opt to pay for the test out-of-pocket or seek insurance reimbursement. Patients who were seen from February 1, 1996 to October 31, 1996 and who fulfilled the following criteria had access to free testing through the beta-testing phase offered by Myriad Genetics Laboratories: 1) an individual with breast cancer diagnosed before age 50, and who has a first-degree relative with breast carcinoma diagnosed before 50, or a first-degree relative with ovarian carcinoma diagnosed at any age; 2) an individual with ovarian carcinoma diagnosed at any age, and who has a first-degree relative with breast carcinoma diagnosed before 50, or a first-degree relative with ovarian carcinoma diagnosed at any age. At the patient's request, information regarding the genetic test and results could be kept in a shadow chart separate from the patient's medical records.
Patients were informed that the first person to be tested in the family should ideally be a member affected with breast or ovarian carcinoma. Patients who were unaffected and who did not have an available affected family member to participate in testing but who desired testing were not denied it, although they were informed that the information that was derived in such a situation could be limited.
The demographic characteristics, cancer history, screening and lifestyle behavior, perceived breast carcinoma risk, family history, family risk category, and access to free testing, were analyzed to determine if any of these factors were associated with genetic testing uptake. Two-sided Pearson chi-square test was computed for univariate analysis. Factors were included in multiple logistic regression analysis if they were associated with test uptake on univariate analysis with P value of 0.1 or less. All statistical analysis was conducted using the SPSS statistical software (SPSS Inc., Chicago, IL).
A total of 443 unrelated new patients were seen in consultation at the Breast and Ovarian Surveillance Service, Johns Hopkins Hospital during the study period. Of these, 258 patients (58%) were eligible and were included in our study. The remaining 185 patients were not included for reasons as detailed in Fig. 1.
The median time from the patient's first consultation with the service to the time of data collection and analysis for the 258 patients eligible for the study was 25 months (range, 8–55 months).
Demographics, Medical History, Screening and Lifestyle Behavior, and Perceived Breast Carcinoma Risk
Most of the patients were white, married, had completed at least 4 years of college and had children. Thirty percent of the patient population was of Ashkenazi Jewish heritage. Thirty percent had a previous diagnosis of breast and/or ovarian carcinoma, 57% practiced regular self-breast examination at least every month, and 86% reported having had at least 1 mammogram. Most did not smoke and had three or fewer alcoholic drinks per week. Of the patients who had no history of breast or ovarian carcinoma and who projected their risk of invasive breast carcinoma, 86% perceived their lifetime risk to be greater than 25% (Table 1).
|Characteristics (n = 258)a||No. of patients in each subcategory||Percentage tested in each subcategory (%)||P valueb|
|Years of education|
|Personal history of breast or ovarian carcinoma|
|Breast self-examination practice|
|Less than monthly||65||26|
|More than monthly||41||28|
|Ever had a mammogram|
|Alcohol consumption (drinks/wk)|
|Perceived life-time risk of invasive breast carcinoma (n = 118) (%)|
|Breast carcinoma–only family||187||25||0.69|
|Breast and ovarian carcinoma family||61||31|
|Ovarian carcinoma family||10||22|
|No. of relatives with breast carcinoma|
|No. of first-degree relatives with breast carcinoma|
|No. of first-degree relatives with ovarian carcinoma|
|Genetic risk category|
|Access to free testing|
Family History and Risk Category
Three-quarters of patients were from breast carcinoma–only families, approximately one-quarter were from breast and ovarian carcinoma families, and 4% were from ovarian carcinoma families. Slightly greater than half the patients had 3 or more family members with breast carcinoma, and 74% had at least one first-degree relative with breast carcinoma. Slightly greater than half the patients were from moderate- or high-risk families.
Details of Genetic Testing
A total of 68 patients (26%) elected to undergo testing, whereas the remaining 190 patients declined testing or expressed interest at the time of the consult but had not been tested at the time of the analysis. The proportion of patients undergoing genetic testing during each year of the 4-year study period was 20 of 48 (42%) in 1996, 11 of 56 (20%) in 1997, 13 of 54 (24%) in 1998, and 24 of 100 (24%) in 1999 (P = 0.057). Of note, 18 of 20 patients who were tested in 1996 had access to free testing (Table 2).
|Type of test|
|Full gene sequencing||59|
|Cancer-affected family member (cancer free proband)||18a|
|Patient seeking consult||82 (n = 56)|
|Prior breast or ovarian carcinoma||59 (33 of 56)|
|No prior breast or ovarian carcinoma||41 (23 of 56)|
|Time interval between consultation and genetic testing|
|> 12 mos||4|
|Mode of payment (screening panel, n = 28)|
|Mode of payment (full sequencing, n = 40)|
Thirty-three of 68 individuals who were tested were patients with cancer, 12 of 68 were affected relatives who were tested at the request of our cancer free patients, and 23 of 68 were unaffected patients who proceeded with testing without first testing an affected family member. Slightly greater than half the patients underwent full gene sequencing. Almost half the patients were tested on the same day of the consult and 96% within 1 year of the consult. Eighteen of 68 patients had free testing. Of the remaining 50 patients who had to bear the cost of the test, 0 of 28 who had the Ashkenazi Jewish screening panel and 13 of 22 (59%) who had full sequencing sought insurance reimbursement, of which 7 had a prior diagnosis of breast or ovarian carcinoma.
Results of Genetic Testing
Approximately one-third of the patients tested had a BRCA1/2 mutation identified. Among the patients who underwent full gene sequencing, 20% had a variant of uncertain significance. No deleterious mutations were identified in any of the 11 patients from low-risk families, whereas 32% and 67% patients from moderate- and high-risk families, respectively, were found to carry a deleterious mutation (P = 0.016). Fifty-eight percent of patients from breast and ovarian carcinoma families carried a BRCA1/2 mutation, compared with only 21% of patients from breast carcinoma–only families (P = 0.014; Table 3).
|Genetic test results||No. of patients (%)|
|BRCA1 mutation||15 (22)|
|BRCA2 mutation||7 (10)|
|Variant of uncertain significance||8 (12)|
|No mutation||38 (56)|
|Genetic test results by test type||No. of patients (%)|
|Screening panel (n = 28)|
|BRCA1 mutation||5 (18)|
|BRCA2 mutation||2 (7)|
|No mutation||21 (75)|
|Full gene sequencing (n = 40)|
|BRCA1 mutation||10 (25)|
|BRCA2 mutation||5 (13)|
|Variant of uncertain significance||8 (20)|
|No mutation||17 (43)|
|Genetic test results by family history||Mutation||Variant||No mutation|
|Breast carcinoma–only families (n = 47)||10 (21)||6 (13)||31 (66)|
|Breast and ovarian carcinoma families (n = 19)||11 (58)||2 (11)||6 (32)|
|Ovarian carcinoma–only families (n = 2)||1 (50)||0 (0)||1 (50)|
|Genetic test results by genetic risk category||Mutation||Variant||No mutation|
|Low-risk families (n = 11)||0 (0)||1 (9)||10 (91)|
|Moderate-risk families (n = 38)||12 (32)||6 (16)||20 (53)|
|High-risk families (n = 9)||6 (67)||1 (11)||2 (22)|
|Not categorized (n = 10)||4 (40)||0 (0)||6 (60)|
Factors Associated with Genetic Testing Uptake
Access to free testing, prior diagnosis of breast or ovarian carcinoma, and Ashkenazi Jewish heritage were the only factors associated with genetic testing on both univariate and multivariate analysis. Age and genetic risk category also were evaluated in multivariate analysis but were not significant.
The BRCA1/2 genes were cloned in 1994 and 1995, respectively, and testing has been available commercially since the end of 1996. Before the availability of clinical testing, various studies in breast carcinoma patients,14 healthy family members of breast or ovarian carcinoma patients,8–12 and the general population13–16 have reported high level of interest in BRCA1/2 genetic testing, ranging from 69% to 96%.8, 9, 11–14, 16, 30, 31 Since genetic testing for BRCA1/2 became available, there have been few published studies on the actual utilization of testing in clinical practice.18
Only one-quarter of patients who were eligible to consider testing in our clinical population eventually were tested or arranged to have an affected family member tested. Although we did not systematically recontact our patients, we did see most in follow-up or maintained telephone contact with them after their consultation. It therefore is unlikely that many were tested elsewhere without our knowledge. The uptake for BRCA1/2 testing in the clinical situation thus is lower than has been reported in previous hypothetical and research situations. The only factors that were associated with genetic test utilization in our study were access to free testing, personal history of breast or ovarian carcinoma, and Ashkenazi Jewish heritage. Of note, many factors that have been reported previously to be predictive of interest in genetic testing in hypothetical and research situations were not associated with test utilization in our study. These include high educational level,10, 12 health-seeking behavior,16 having children or daughters,2, 8, 32 healthy lifestyle,16 high perceived risk of developing cancer,9 and high predicted genetic risks,18, 32 all factors that were highly represented in our study cohort. It is possible that the significance of these factors in the clinical situation is overshadowed by practical considerations such as cost, fear of insurance discrimination, or the recommendation to involve an affected relative in the testing process.
Approximately one-third of the entire patient population tested was found to carry a BRCA1/2 mutation. In keeping with what has been reported by others,19, 33 breast-ovarian cancer families were more likely to carry deleterious BRCA1/2 mutations than breast carcinoma–only families in our series. We also found good correlation between the likelihood of finding a deleterious mutation and genetic risk categories.
Potential Obstacles to BRCA1/2Testing
Cost is likely to be one of the most important obstacles to genetic testing. The current cost of BRCA1/2 testing ranges from $300 to $2680 and is prohibitive for many individuals. Previous studies of patients' response to prospective offers of genetic testing have found significant reduction in interest when cost became a factor.9, 30 This would explain why nearly 70% of patients who were eligible for free testing chose to be tested, as opposed to only 22% when patients had to bear the cost of the test. Although the greater estimated genetic risk in the subgroup that had access to free testing may have played a role in the higher test utilization, we do not believe this to be the main reason. We also reported significantly higher genetic test uptake among Ashkenazi Jewish patients, who could be screened with the less costly three-mutation panel, suggesting again that cost is an important consideration.
Fear of insurance discrimination often has been cited as a potential barrier to genetic testing.2, 8, 34 Most of our patients were middle- to upper-class based on educational attainment, and greater than 99% had health insurance. Insurance plans increasingly provided reimbursement for genetic testing over the course of our study period. In 1996, Maryland enacted a law prohibiting health insurers from using genetic test results to make coverage decisions or to modify rates or terms of health insurance policies. Despite this, many remain concerned about the potential negative impact of genetic information on insurability,16, 35 and many women in our series chose to pay out-of-pocket for confidentiality reasons. Of the patients who sought insurance reimbursement for the more costly full gene sequencing, greater than half (7 of 13) had a prior diagnosis of breast or ovarian carcinoma and would have been less concerned about insurance discrimination based on the results of a genetic test.
Discrepancy between Hypothetical or Research Interest and Actual Utilization
What accounts for the discrepancy between actual clinical test utilization and reported interest in the hypothetical situation and research setting? Patients seen in clinical settings often are given detailed information on hereditary breast and ovarian carcinoma and genetic testing. As such, these patients are better informed than those who were offered testing hypothetically. Although there may be genuine interest in genetic testing, actual utilization of the test may be low after the individual weighs the potential advantages against the disadvantages and limitations. Conversely, uptake is high in the research setting despite participants being fully informed of its disadvantages and limitations. However, research participants do not bear the cost of the test, are relatively protected against insurance discrimination, and are generally higher risk individuals, as opposed to a clinical population, which usually comprises of a mix of low-, moderate-, and high-risk families. Although genetic risk category did not have an impact on test utilization in our study, this nevertheless could have been a confounding factor. Finally, individuals who have chosen to participate in research are self-selected, highly motivated, and often have altruistic desires to help science17 and thus are not representative of a clinical population.
Limitations of Our Study
The genetic risk estimates provided to patients were not based on a single model, nor were they delivered to patients in a standardized manner. Of note, a significant proportion of patients were not given a quantitative risk estimate for various reasons, and the assignment of genetic risk categories for these patients became somewhat arbitrary and subjective. The inability to categorically give a quantitative risk estimate to certain patients reflects the actual clinical situation. Furthermore, there are significant limitations to any single risk assessment model. The Couch model, for example, does not take into account male breast carcinoma, bilateral breast carcinoma, or cancers apart from breast and ovarian carcinoma that may be increased in BRCA1/2 mutation families. In addition, we have empirically doubled the risk estimates derived from the Couch model to provide an approximate estimate of each family's probability of carrying a BRCA1/2 mutation. This may result in an overestimate because BRCA2 mutations are less prevalent than BRCA1 mutations. Recently available models such as the BRCAPRO may provide a more comprehensive and objective risk estimate.36 A prospective study with systematic documentation of genetic risk categories would have provided more valuable and relevant information regarding how genetic risk categories might have an impact on genetic test utilization. Finally, clinical cancer genetics is a rapidly evolving field, not only in terms of medical developments, but also in its social, ethical, and legal aspects. The passage of legislation to protect against genetic discrimination, increasing willingness of insurance companies to reimburse for genetic testing, and emerging data on cancer preventive measures26, 37–40 might have an impact on testing uptake. Our experience in the first 6 months of 2001 (not included in the analysis), however, has been fairly concordant with what has been reported in this study, with 33 of 107 (31%) of eligible probands undergoing testing (60% full sequencing, 40% screening panel). Nonetheless, a follow-up study to evaluate potential changes in testing utilization with time would be warranted.
We report a low utilization rate of genetic testing for BRCA1/2 mutations in a large clinical series. Potential obstacles to testing include cost, fear of insurance discrimination, and need to involve an affected family member in the testing process. Factors that previously have been reported to be important in hypothetical and research situations, such as education level, reproductive status, involvement in screening, healthy lifestyle, and high perceived risk of breast carcinoma, did not have an impact on test utilization in our study. Our study was, however, not a prospective one designed to look at motivators and barriers to genetic testing specifically. We thus lack the ability to examine the relative importance of factors in any individual patient's decision making. Prospective studies examining these factors are warranted.
The authors thank all the patients who participated in the Breast and Ovarian Surveillance Service for providing their personal and family history information to make our database possible. They are also grateful to Dr. Florence Houn, Karen A. Johnson, Susan E. Appling, and Jennifer Bucholtz for providing clinical services, Linda Thompson for providing administrative support, and Alison Klein for designing the clinical database.
- 7ASCO Subcommittee on Genetic Testing for Cancer Susceptibility. Statement of the American Society of Clinical Oncology. Genetic testing for cancer susceptibility. J Clin Oncol. 1996;14: 1730–6.