Physicians’ Experiences With BRCA1/2 Testing in Community Settings

Abstract

Purpose

We surveyed a national sample of nonacademic physicians who ordered BRCA1/2 testing to understand their implementation of genetic testing and to assess recommendations for surveillance and cancer risk management of women with positive test results.

Patients and Methods

We surveyed physicians (N = 611 of 1,050; response rate, 58.2%) practicing in nonacademic settings who ordered BRCA1/2 testing during 2004 to 2005. We described physicians’ experiences with testing and used multivariable regression models to identify factors associated with more complete counseling and with recommendations for cancer risk management for a BRCA1 mutation carrier.

Results

Most physicians (68.2%) usually or always discussed six counseling items before testing. In adjusted analyses, physicians who were assisted by genetic counselors, nurse geneticists, or others (v counseling by themselves), those who spent more than 60 minutes in counseling, and medical oncologists (v surgeons or geneticists) were more likely to discuss all six items (all P < .05). A total of 61.4% of physicians would recommend bilateral prophylactic mastectomy to a 38-year-old BRCA1 mutation carrier who had completed childbearing. After adjustment, geneticists and gynecologists were less likely than medical oncologists and surgeons to recommend prophylactic mastectomy (P < .001), as were physicians in the Northeast versus those in other regions of the United States (P = .01).

Conclusion

Community-based physicians seem to be successfully incorporating BRCA1/2 testing into their practices. Physicians’ recommendations for surveillance of mutation carriers are generally consistent with practice guidelines, yet recommendations for preference-based procedures such as prophylactic mastectomy vary by physician characteristics such as specialty and geographic region. The providers whom patients see for testing may contribute to variations in prophylactic treatments.

INTRODUCTION

Genetic testing for mutations in the breast/ovarian cancer susceptibility genes BRCA1 and BRCA2 has become the most widely ordered genetic test in the United States since the test became commercially available in 1996. A germline mutation in either BRCA1 or BRCA2 confers greatly increased lifetime risks of breast and ovarian cancers.¹ Knowledge of BRCA1/2 mutation status guides recommendations for surveillance and risk-reducing interventions and may also influence therapeutic decisions for some newly diagnosed patients.

Most genetic testing in the United States is ordered by physicians, and therefore, the experiences that patients have with genetic testing likely depend in large part on the skills, attitudes, and practices of the physicians they see. As use of BRCA1/2 testing increases, more testing is being done outside of academic centers. Academic medical centers currently account for less than 50% of all BRCA1/2 testing performed by Myriad Genetics Laboratory (Myriad; Salt Lake City, UT), the only commercial provider of BRCA1/2 testing. Given the complexity of cancer genetics, risk assessment, and management of patients informed by genetic test results,² it is important to understand how genetic testing is being implemented in the community. Physicians ordering genetic testing must provide the requisite counseling and education and make recommendations to patients based on their genetic test results.^3,4 Although recommendations and guidelines for assessment, counseling, and follow-up care of women at high-risk for breast cancer have been available for some time,^3-5 it is not known to what extent these have diffused into clinical practice.

We surveyed a national sample of nonacademic physicians who had ordered BRCA1/2 testing to understand how they were implementing genetic testing in their practices. We also assessed their recommendations for surveillance and management of women with positive test results.

PATIENTS AND METHODS

Study Population

This physician survey is part of a larger study of patients’ experiences with BRCA1/2 testing. Myriad distributed surveys to physicians ordering BRCA1/2 tests for patients in the larger study. The patient cohort included all individuals testing positive for a deleterious mutation as well as the next three individuals with nonpositive (negative or variant of unknown significance) results. The study focused on testing in nonacademic centers; therefore, patients whose tests were sent from 130 academic medical centers were excluded. The physician cohort included all physicians who ordered BRCA1/2 tests for eligible patients. Physicians were surveyed only once regardless of the number of genetic tests they ordered during the study period. Myriad provided information about each physician's sex, practice geographic location, and number of BRCA1/2 tests ordered in the past year. The study protocol was approved by the Dana-Farber/Harvard Cancer Center institutional review board.

Questionnaire and Survey Administration

The survey instrument was designed to collect information about physicians’ experiences with BRCA1/2 testing and recommendations for additional treatment of patients with a positive test result. Questions were developed after literature review and discussions with physicians, and the questionnaire was revised after cognitive testing and debriefing with eight clinician experts.

Physicians were asked whether patients are ever referred to them by other providers specifically for BRCA1/2 counseling and possible testing, whether they have ordered genetic tests other than BRCA1/2, how often they schedule specific visits for BRCA1/2 counseling before testing, the number of minutes spent counseling for BRCA1/2 testing, assistance with counseling from others, whether the practice bills for professional counseling time, and the topics discussed during counseling (based on American Society of Clinical Oncology statement on genetic testing⁶). Physicians were also asked whether they typically provide positive, negative, or variant results in person, by phone, or by mail and whether results are included in patients’ medical records. They were asked what surveillance testing they would recommend for a 35-year old woman with a BRCA1 mutation who desires no prophylactic surgery. Finally, they were asked whether they would recommend prophylactic surgery and other interventions for a 38-year-old healthy BRCA1 mutation carrier who has completed childbearing. Physicians also provided their specialty and medical school graduation year.

During August 2004 through December 2005, Myriad mailed surveys with a $20 prepaid cash incentive to 1,086 physicians who had ordered BRCA1/2 tests. Physicians not responding within 3 weeks were mailed a second survey. No further contacts were made. Of 1,086 physicians surveyed, 36 physicians were ineligible because they had never ordered genetic testing (n = 27), were affiliated with an academic center (n = 7), or practiced outside the United States (n = 2). Of the remaining 1,050 physicians, 611 responded (response rate, 58.2%). Women were more likely to respond than men (65% v 54%; P = .001), as were physicians who ordered more BRCA1/2 tests in the past year (68.3% of high-volume, 54.8% of medium-volume, 53.2% of low-volume physicians; P = .003), but response rates did not differ by US census region (P = .82) or urban practice location (P = .88).

Analysis

Testing and counseling.

We described physicians’ experiences with BRCA1/2 testing. We dichotomized items about counseling before testing as “always” or “usually” versus “sometimes” or “rarely.” These six items (benefits and limitations of close surveillance if positive test, benefits and limitations of prophylactic mastectomy if positive test, benefits and limitations of prophylactic oophorectomy if positive test, possibility of psychological reaction, sharing test results with family members, and confidentiality and privacy issues) were among the elements of informed consent to be discussed with all patients before testing as stipulated in the American Society of Clinical Oncology statement on Genetic Testing.⁶ We summed the number of the six for which providers responded “always” or “usually” and dichotomized this summary variable at six versus five or less.

We used χ² tests to identify bivariate associations between clinician characteristics and whether they usually/always discussed all six items. Specifically, we examined provider sex, specialty, years since medical school graduation, whether they schedule a specific counseling visit, time spent counseling, whether a nurse or genetic counselor assists them with counseling, whether they bill for counseling visits, geographic region of practice, whether other providers refer patients to them (a measure of expertise perceived by others), and number of BRCA1/2 tests ordered in the past year (a measure of testing experience). We used logistic regression to assess factors associated with discussing all six items, including in the model all variables with P ≤ .20 in unadjusted analyses.

Recommendations for women with BRCA1/2 mutation.

We described recommendations for surveillance testing and management of women with BRCA1/2 mutations. We used χ² tests to identify bivariate associations of provider characteristics with recommendations for prophylactic bilateral mastectomy for a 38-year-old BRCA1 mutation carrier who had completed childbearing after dichotomizing the responses as recommended for (somewhat or strongly) versus recommended against (somewhat or strongly). Specifically, we examined sex, specialty, number of BRCA1/2 tests ordered in the past year, whether other physicians refer patients to them for testing, and geographic region. We used logistic regression models to assess whether provider characteristics were associated with recommendations for prophylactic mastectomy, including all variables with P ≤ .20 in unadjusted analyses.

Analyses were conducted using SAS statistical software, version 9.1 (SAS Institute, Cary, NC).

RESULTS

Of 611 respondents, 39% were women and the mean number of years since medical school graduation was 23.1 (standard deviation = 9.5; Table 1). More than half (53%) were medical oncologists, 15% were surgeons, 8% were geneticists, and 11% were gynecologist/obstetricians. The remaining physicians were primary care physicians (9.3%) and physicians of other specialties (3.6%). Nearly two thirds of physicians reported that they counsel and test patients referred to them by other physicians for BRCA1/2 testing, suggesting that their colleagues consider them to have expertise in this area. Many (42%) have ordered genetic tests other than BRCA1/2, and the median number of BRCA1/2 tests ordered in the past year was eight (interquartile range, two to 24).

Table 1.

Characteristics of Physicians Ordering BRCA1/2 Tests

Characteristic	Physicians (N = 611)
Sex, %
Male	60.6
Female	39.4
Specialty, %
Medical oncologist	53.2
Surgeon	15.1
Geneticist	7.7
Gynecologist/obstetrician	11.1
Other*	12.9
No. of years since graduation from medical school†
Mean	23.1
SD	9.5
US Census region, %
Northeast	21.3
Midwest	21.1
South	33.1
West	24.6
Urban-rural residence, %
Large metropolitan area	52.4
Smaller metropolitan area	38.0
Nonmetropolitan area	9.7
No. of patients seen per week‡
Mean	77.8
SD	46.9
Counsels and tests patients who are referred by other physicians for possible BRCA1/2 testing, %	61.7
Has ordered genetic tests other than BRCA1/2, %	42.2
No. of BRCA1/2 tests ordered during study period of July 2004 through December 2005
Median	8
Interquartile range	2-24

Abbreviation: SD, standard deviation.

^*Twelve physicians with missing information on specialty were included in the other category.

^†Thirty physicians were missing information on year of graduation from medical school.

^‡Seventeen physicians were missing information on number of patients seen per week.

Testing and Counseling

When testing for BRCA1/2, most physicians always (41.9%) or usually (17.5%) schedule specific visits for which counseling is the primary reason for the visit (Table 2). More than half of physicians spend more than 30 minutes counseling patients about the test before testing, and physicians are frequently assisted by genetic counselors, nurse geneticists, nurse practitioners, physician assistants, or other counselors (Table 2).

Table 2.

Physicians’ Reported Practice Patterns When Providing Genetic Testing

Practice Pattern	% of Physicians (N = 611)
Counseling
Schedules specific visits for counseling
Always	41.9
Usually	17.5
Sometimes	31.3
Never	9.0
Unknown	0.3
Total No. of minutes of counseling before BRCA1/2 testing
≤ 30	47.0
31-60	35.8
> 60	16.0
Unknown	1.2
Staff member(s) counseling
Physician only	33.7
Physician assisted by genetic counselor/nurse geneticist/NP/PA	35.4
Physician assisted by other counselor	26.7
Unknown	4.3
Practice bills for professional time associated with counseling	60.9
When counseling patients before testing, always or usually discusses the following:
Benefits and limitations of close surveillance	92.0
Benefits and limitations of prophylactic mastectomy	83.8
Benefits and limitations of prophylactic oophorectomy	82.2
Possibility of a psychological reaction	81.0
Sharing test result with family members	88.1
Confidentiality and privacy issues	91.3
Providing results
Inform patients of positive results
In person	74.0
By phone	23.2
By mail	0.5
Missing	2.3
Inform patients of variant results
In person	65.8
By phone	29.0
By mail	1.2
Missing	4.1
Inform patients of negative results
In person	50.8
By phone	45.5
By mail	2.1
Missing	1.6
Documents results of tests in medical record
Always	62.2
Usually	15.9
Sometimes	11.5
Never	8.2
Missing	2.3

Abbreviations: NP, nurse practitioner; PA, physician assistant.

Physicians typically inform patients of positive or variant results in person, but often provide results by phone, especially negative results (Table 2). Most physicians always (62.9%) or usually (15.9%) document test results in the patient's medical record (Table 2).

Two thirds (68.2%) of physicians reported that all six counseling items are usually or always discussed with their patients before testing. In unadjusted analyses, geneticists and medical oncologists were more likely than other specialists to report discussing all six items (Table 3). Physicians who test patients referred by other physicians, those who do more testing, and those who schedule specific visits for counseling were more likely than others to report discussing all counseling items, as were physicians who spend more time counseling and those who are assisted by genetic counselors, nurse geneticists, or nurse practitioners/physician assistants.

Table 3.

Factors Associated With Usually or Always Discussing All Six Counseling Items

Characteristic	Unadjusted Proportion Discussing All Six Items (%)	P	Discussion of All Six Items*		P
			Adjusted Odds	95% CI
Sex		.10			.73
Male	65.7		Reference
Female	72.1		1.08	0.72 to 1.61
Specialty		< .001			.06
Medical oncologist	73.5		Reference
Surgeon	64.0		0.54	0.31 to 0.93
Geneticist	78.7		0.41	0.18 to 0.96
Gynecologist/obstetrician	53.9		0.62	0.33 to 1.17
Other/missing	56.0		0.62	0.33 to 1.15
No. of years since medical school graduation		.40			—
≤ 15	68.7		—
16-23	68.0		—
24-29	70.1		—
≥ 30	64.0		—
Unknown	82.1		—
No. of BRCA1/2 tests sent during past year		< .001			.20
1-3	57.9		Reference
4-15	64.7		0.85	0.51 to 1.40
≥ 16	80.9		1.36	0.75 to 2.5
Counsels and tests patients who are referred by other physicians for possible BRCA1/2 testing		< .001			.53
Yes	75.7		Reference
No/missing	55.8		0.86	0.53 to 1.38
Schedules specific visits for counseling		< .001			.09
Always/usually	76.8		Reference
Never/sometimes	55.9		0.69	0.45 to 1.06
Total No. of minutes of counseling before BRCA1/2 testing		< .001			< .01
≤ 30	55.0		Reference
31-60	75.2		1.48	0.91 to 2.42
> 60	89.8		3.70	1.66 to 8.29
Staff member(s) counseling		< .001			.05
Physician only	58.5		Reference
Physician assisted by genetic counselor/nurse geneticist/NP/PA	82.5		2.0	1.19 to 3.37
Physician assisted by other counselor	64.3		1.37	0.86 to 2.18
Unknown	50.0		0.89	0.36 to 2.21
Practice bills for professional time associated with counseling		< .001			.39
Yes	74.3		Reference
No/missing	58.6		0.83	0.54 to 1.27
US Census region		.78			—
Northeast	68.6		—
Midwest	69.5		—
South	65.6		—
West	70.3		—

Abbreviations: NP, nurse practitioner; PA, physician assistant.

^*Using logistic regression to control for all variables in the table except for number of years since medical school graduation and geographic region. Analyses excluded 14 physicians who did not respond to the section on counseling, five physicians missing information on no. of minutes counseling, and two physicians missing information on whether they schedule a specific visit for counseling. Wald CIs are given, and P values are based on the likelihood ratio χ² test.

In adjusted analyses, surgeons and geneticists were less likely than medical oncologists to report always/usually discussing all six counseling items (Table 3). Physicians assisted by genetic counselors, nurse geneticists, or other physician extenders, compared with physicians who counseled without assistance, were more likely than others to report that counseling discussions included all six items, as were physicians who (alone or in conjunction with staff) spent more than 60 minutes in counseling. Physician sex, number of BRCA1/2 tests sent during the past year, whether the physician counsels and tests patients referred from other physicians, and whether the practice bills for professional time associated with counseling were not significantly associated with the number of counseling items discussed.

Recommendations for Women With BRCA1/2 Mutation

When asked about recommendations for surveillance of a 38-year-old woman who tested positive for a BRCA1 mutation, most physicians reported care that was consistent with published recommendations available at the time of the survey (Table 4). Nearly all physicians (95.6%) recommended monthly breast self-examination. Approximately three quarters (74.3%) recommended clinical breast examinations every 6 to 12 months, with 23.8% recommending them quarterly. Nearly all physicians recommended mammography every 6 to 12 months (95.3%) and pelvic examination every 6 to 12 months (95.9%). Most physicians also recommended pelvic ultrasound every 6 to 12 months (84.6%) and CA-125 testing every 6 to 12 months (74.0%) Although not part of guideline recommendations at the time of the study, nearly half of physicians (51.7%) recommended breast magnetic resonance imaging (MRI) annually, and another 5.7% recommended it twice yearly.

Table 4.

Recommendations for BRCA1 Carriers

Recommendations	% of Physicians (N = 611)
Surveillance recommendations for 35-year-old woman with BRCA1 mutation who desires no prophylactic surgery. Recommend … *
Monthly breast self-examination	95.6
Clinical breast examination every 6-12 months	74.3
Mammogram every 6-12 months	95.3
Pelvic examination every 6-12 months	95.9
Pelvic ultrasound every 6-12 months	84.6
CA-125 testing every 6-12 months	74.0
Recommend surveillance breast MRI …
No	34.7
Every 6 months	5.7
Yearly	51.7
Missing	7.9
38-year-old woman with BRCA1 mutation who has completed childbearing
Recommend bilateral prophylactic mastectomy	61.4
Recommend bilateral salpingo-oophorectomy	84.9
38-year-old woman with BRCA1 mutation who has completed childbearing and underwent bilateral prophylactic oophorectomy with hysterectomy
Recommend estrogen	6.6
Recommend tamoxifen	71.2
Recommend raloxifene	33.7

Abbreviation: MRI, magnetic resonance imaging.

^*Time intervals for testing were categorized based on recommendations in published articles and American Society of Clinical Oncology recommendations.

More than half of physicians (61.4%) recommended bilateral prophylactic mastectomy for a 38-year old woman who had completed childbearing, and 85% would recommend bilateral salpingo-oophorectomy. Assuming the woman had undergone bilateral oophorectomy, only 6.6% would recommend estrogen, but most (71.2%) would recommend tamoxifen, and a third (33.7%) would recommend raloxifene.

Table 5 presents unadjusted and adjusted associations of physician characteristics with recommendations for bilateral prophylactic mastectomy. In unadjusted analyses, medical oncologists and surgeons were more likely than other specialists to recommend this procedure. Physicians practicing in the Northeast were least likely to recommend this procedure, and physicians with a moderate volume of BRCA1/2 testing were more likely than low- or high-volume testers to recommend prophylactic mastectomy.

Table 5.

Factors Associated With Recommendations for Prophylactic Mastectomy for a 38-Year-Old Woman With a BRCA1 Mutation Who Has Completed Childbearing

Characteristic	Unadjusted Proportion Recommending Prophylactic Mastectomy (%)	P	Recommendation of Prophylactic Mastectomy*		P
			Adjusted Odds	95% CI
Sex		.10			.26
Male	64.9		Reference
Female	58.2		0.81	0.56 to 1.17
Specialty		< .001			< .001
Medical oncologist	69.0		Reference
Surgeon	75.6		1.44	0.83 to 2.49
Geneticist	26.7		0.15	0.07 to 0.31
Gynecologist/obstetrician	47.1		0.43	0.24 to 0.76
Other	52.6		0.58	0.33 to 1.02
No. of years since medical school graduation		.43			—
≤15	61.7		—
16-23	63.7		—
24-29	58.3		—
≥ 30	66.1		—
Unknown	50.0		—
Counsels and tests patients who are referred by other physicians for possible BRCA1/2 testing		.26			.60
Yes	64.1		Reference
No	59.5		0.89	0.57 to 1.38
No. of BRCA1/2 tests sent during past year		.04			.38
1-3	58.0		Reference
4-15	69.4		1.36	0.84 to 2.19
≥ 16	59.6		1.06	0.63 to 1.77
Geographic region		.04			.01
Northeast	52.7		Reference
Midwest	63.2		1.65	0.97 to 2.78
South	68.3		2.17	1.34 to 3.51
West	61.7		1.85	1.11 to 3.08

^*Using logistic regression to control for all variables in the table except for number of years since medical school graduation. Wald CIs are given, and P values are based on the likelihood ratio χ² test. Analysis did not include nine physicians who did not respond to the item about treatment recommendations for the 38-year-old woman with a BRCA1 mutation who has completed childbearing.

In adjusted analyses, geneticists and gynecologist/obstetricians were significantly less likely than medical oncologists (and surgeons) to recommend prophylactic mastectomy (Table 5). Physicians in the Northeast (v other regions) remained less likely to recommend this procedure. Physician sex, whether others refer patients to them for counseling and testing (a marker of expertise perceived by others), and volume of BRCA1/2 testing were not associated with recommending prophylactic mastectomy in adjusted analyses.

DISCUSSION

We surveyed community-based physicians who ordered BRCA1/2 testing for patients during 2004 and 2005 to understand physicians’ practices regarding counseling and testing as well as their recommendations for women testing positive for BRCA1/2 mutations. Our sample of nonacademic physicians who have sent genetic tests was comprised primarily of medical oncologists; this is consistent with national estimates suggesting that oncologists are the biggest users of genetic susceptibility testing.⁷ Only 8% of our cohort were geneticists.

Guidelines recommend thorough counseling for women undergoing BRCA1/2 testing.^6,8,9 Although geneticists reported the highest rates of counseling about six important items in unadjusted analyses, when we also controlled for other factors, geneticists were less likely than medical oncologists to report always or usually counseling about all six items. Many physicians rely on the assistance of genetic counselors or others to assist with counseling, and those who have genetic counselors, nurse geneticists, or nurse practitioners/physician assistants assist are likely to have more thorough counseling discussions. Other data also suggest that nurses and genetic counselors can be effective at providing education to patients undergoing breast cancer genetic testing.¹⁰

Despite a prior report that 91% of physicians believe that patients with positive genetic test results are at risk for insurance discrimination,¹¹ studies of patients who have undergone BRCA1/2 testing have not found evidence of actual insurance discrimination.^12,13 In our cohort, most physicians reported documenting results of tests in the medical record, suggesting that physicians who are actually testing patients are not overly concerned about the potential for test-based discrimination.

Physicians varied in how they provide results of genetic testing to patients, consistent with patients’ reports in a prior study.¹⁴ Although most physicians reported that they inform patients of positive results in person, a substantial minority (23.2%) provided the results by phone, a practice that is increasing in frequency and is currently being studied.¹⁵ Physicians more often used telephone (and rarely mail) to inform patients of variant or negative results, although more than half of physicians gave these results in person.

Recommendations for surveillance testing at the time of our survey included breast self-examination monthly and clinical breast examination, pelvic examination, mammogram, pelvic ultrasound, and CA-125 testing every 6 to 12 months.^3,4,16 Most physicians suggested surveillance care for BRCA1 carriers that was consistent with these recommendations. Data suggesting benefits of breast MRI for BRCA1/2 carriers became available in late 2004^17,18 and is recommended in current treatment guidelines.⁹ More than half of physicians in our study were recommending breast MRI in late 2004 and 2005.

Prophylactic surgery is an option for women with BRCA1/2 mutations.¹⁹ Bilateral prophylactic mastectomy substantially reduces risk of breast cancer in women with a family history of breast cancer²⁰ and women with BRCA1/2 mutations,^21-23 as does contralateral mastectomy in BRCA1/2 carriers with breast cancer.²⁴ Premenopausal prophylactic bilateral salpingo-oophorectomy is associated with a decreased risk of breast and ovarian cancer in BRCA1/2 carriers,^25,26 likely leading to an increased life expectancy.²⁷

In our study, most physicians (85%) recommended bilateral prophylactic oophorectomy for a 38-year old BRCA1 carrier who had completed childbearing. However, fewer physicians (61.4%) recommended bilateral prophylactic mastectomy. Medical oncologists and surgeons were more likely to recommend prophylactic mastectomy than geneticists and gynecologists. Recommendations for this surgery also varied by geographic region, with physicians in the Midwest, South, and West more likely to recommend prophylactic mastectomy than physicians in the Northeast. This trend is consistent with preferences for mastectomy over breast-conserving surgery for breast cancer treatment in these areas.²⁸

Nearly three quarters of physicians in our cohort would recommend tamoxifen for an unaffected 38-year old BRCA1 carrier who had undergone prophylactic oophorectomy but not mastectomy, a rate higher than that of a 2002 survey of breast cancer physicians, for whom 57% recommended tamoxifen for a similar 40-year old BRCA1 carrier.²⁹ Limited data suggest that tamoxifen can decrease the risk of new breast cancers in BRCA1/2 mutation carriers who have had breast cancer.^30,31 The primary preventive benefits of tamoxifen may be limited to BRCA2 carriers, whose cancers are more likely to be estrogen receptor positive.^32,33 Among affected BRCA1/2 carriers, the risk-reducing benefits of tamoxifen in women after oophorectomy have been inconsistent,^30,31 and data are lacking to help guide recommendations for unaffected carriers. Potentially important differences in the management of BRCA1 versus BRCA2 carriers have only recently been recognized.

Although the Study of Tamoxifen and Raloxifene trial suggests that the reduction in invasive breast cancer with raloxifene in women at high risk of breast cancer is similar to that with tamoxifen,³⁴ few data are currently available about the use of raloxifene in BRCA1/2 carriers. A minority of physicians in our study recommended estrogen therapy after bilateral salpingo-oophorectomy, suggesting concern about a medication that may stimulate growth of breast cancers,³⁵ despite a decision analysis and some data suggesting that short-term hormone therapy may be a reasonable option for women.^27,36

Our findings should be viewed in light of several limitations. First, we surveyed physicians whose patients were tested for mutations by Myriad. Myriad provides the vast majority of clinical BRCA1/2 genetic analysis in the United States, and thus our sample of physicians closely reflects those physicians in nonacademic centers who are ordering BRCA1/2 genetic testing nationally. Some university and research laboratories provide limited testing, but these likely account for a small proportion of tests performed in the United States. Second, we studied only nonacademic physicians; therefore, our findings cannot necessarily be generalized to academic physicians, who may differ in their testing practices or recommendations. Third, we cannot be certain that physicians’ reported behaviors and responses to vignettes reflect reality. Nevertheless, others have found clinical vignettes to be a valid tool for measuring care delivered in clinical practice.³⁷ Moreover, we had limited information about additional details, such as the actual role of physicians and other professionals when physicians reported being assisted in counseling by others. Finally, the survey responses are subject to response bias, despite our relatively high response rate.

In conclusion, community-based physicians seem to be successfully incorporating BRCA1/2 testing into their practice, which may be particularly important with the increase of direct-to-consumer advertising of BRCA1/2 testing,³⁸ which is likely to expand physicians’ roles in counseling and testing. They often schedule specific counseling visits for patients and receive assistance with counseling from others in their practice. Medical oncologists and physicians who receive assistance from genetic counselors or other physician extenders tend to conduct the most complete counseling, but most physicians report that their patients are always or usually counseled about six topics that are recommended to be included in pretest counseling. Physicians’ recommendations for surveillance of mutation carriers are generally consistent with practice guidelines. Nevertheless, recommendations for preference-based procedures such as prophylactic mastectomy vary by characteristics such as physician specialty and geographic region, suggesting that the providers that patients see for testing may contribute to variations in prophylactic treatments.

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

AUTHOR CONTRIBUTIONS

Conception and design: Nancy L. Keating, Lisa DiGianni, Judy E. Garber

Financial support: Judy E. Garber

Collection and assembly of data: Kathryn A. Stoeckert, Lisa DiGianni

Data analysis and interpretation: Nancy L. Keating, Kathryn A. Stoeckert, Meredith M. Regan, Lisa DiGianni, Judy E. Garber

Manuscript writing: Nancy L. Keating, Kathryn A. Stoeckert, Meredith M. Regan, Lisa DiGianni, Judy E. Garber

Final approval of manuscript: Nancy L. Keating, Kathryn A. Stoeckert, Meredith M. Regan, Lisa DiGianni, Judy E. Garber