Recent scientific breakthroughs in the genetics of breast cancer may have had effects on women’s perceptions of risk and subsequent worry about breast cancer. Here, we present the rates of interest in counseling among women identified from diverse sources, their levels of cancer worry and perceived risk, and predictors of their agreement to participate in breast cancer risk counseling. Women were identified through breast cancer cases and through media offers. They completed a telephone survey and were ultimately either entered or not entered into a counseling trial. Overall, almost half (46%) of cases who were approached responded to the contact letter asking for information about potentially interested relatives. A total of 588 women responded to the brief media solicitations over a 15-week period. Participants recruited from media contacts reported slightly but significantly higher levels of worry about getting cancer, compared to case-recruited participants. Cancer worry negatively and significantly predicted entry into the counseling project. The results presented here may have implications for recruiting women in the general population with a family history of breast cancer for counseling about their risk for the disease.

Breast cancer is currently the focus of substantial research attention. Many national health policy-making bodies and advocacy groups, including the American Cancer Society and the National Women’s Health Network, have placed a high priority on public action to reduce breast cancer mortality (1, 2). This attention to breast cancer has had the effect of putting this issue on the center stage for United States women and has probably resulted in increased media coverage, both nationally and locally. Recent scientific breakthroughs in the genetics of breast cancer, such as the discovery of the BRCA1 and BRCA2 genes (3, 4), have only heightened the attention from both the public and the media.

This increased attention may have had unintended effects on women’s perceptions of risk and subsequent worry about breast cancer. Various studies have shown that women may either over- or underestimate their risk. For example, our own data indicate that 25% of women with and without family histories, recruited from a local health maintenance organization, underestimate their risk for breast cancer (5). In other studies, 25–45% of women of moderate risk, defined as having a single first-degree relative with breast cancer, overestimated their personal risk for breast cancer; these women reported higher levels of worry about cancer than women of a similar risk profile with more accurate levels of risk perception (6). In several studies of women at increased risk for breast cancer, clinical levels of anxiety and depression have been reported in one-third of the women at high risk (7). Collectively, these data demonstrate a need for information and support for women regarding breast cancer risk, especially women with a family history of the disease.

There are several models for meeting consumer needs for breast cancer education. Voluntary organizations and hospitals often distribute simple informational pamphlets and materials describing risk information. The Cancer Information Service of the National Cancer Institute maintains a toll free number for the public to call to obtain free information on a variety of cancer-related topics. However, some women need or request more intensive support than these approaches offer. The primary care setting is a logical source of information about cancer risk and follow-up. Breast care centers and mammography providers are another source of information on risk, screening, and follow-up care. For women concerned about inherited risk, genetic counselors can provide personalized familial risk assessment and interpretation. Support groups often take a different approach, focusing less on the provision of facts and risk information and more on coping with emotional distress and obtaining social support. Determining women’s interest in obtaining risk counseling is one method of assessing the public’s need for such services. One study reporting on interest in participating in counseling activities found that many women wished to receive counseling services, although some were at relatively low risk for breast cancer (8).

Here, we offered breast cancer risk counseling to women with a mild to moderate family history of the disease. We focused on women at moderate risk because they might be experiencing the most uncertainty and confusion about their risk. We wanted to determine the proportion of these women identified through different sources who would be interested in obtaining either genetic counseling or short-term support group counseling. These data will be useful in understanding and defining the potential need for information and counseling services regarding breast cancer risk. Also, future research projects could use these data to predict rates of response to research project solicitations. Here, we present the rates of interest in counseling among women identified from diverse sources, their levels of cancer worry, and perceived risk. We will then use these variables as predictors of their agreement to participate in a breast cancer risk counseling study at a later date. We hypothesized that women recruited through self-identification methods, such as media contracts, will report higher levels of worry about cancer, compared to women who are contacted by us to participate via their relatives with cancer. We further hypothesized that, regardless of initial contact method, worry about cancer would predict later entry into the counseling trial.

Overview

Participants initially either called the study staff after seeing a media cue or were called by study staff because their relative with cancer gave us their name and contact information. All participants provided eligibility information and several descriptive variables during a telephone survey. These descriptive variables served as predictors of later entry into the counseling trial. At the telephone survey stage, potential participants were told about the requirements of the study (i.e., baseline survey, randomization, counseling participation, and follow-up surveys). Eligible participants received a baseline survey to complete at home and mail back to the study staff. Eligible participants with a completed baseline were entered into the counseling trial. This entry variable (1, yes; 0, no) served as the outcome in regression analyses.

Participant Recruitment

Women were identified through different sources, each producing respondents with potentially unique psychological and risk characteristics. Each is described below.

Recruitment through Breast Cancer Cases.

We identified breast cancer cases through an ongoing case-control study of breast cancer in young women at the FHCRC,3 through a study of mammographic practices at the FHCRC, and through the practice of an oncologist based at the University of Washington. For all sources, we mailed a letter signed by the principal investigator and cosigned by either an investigator from the case-control study or the oncologist. The letter invited the women to provide contact information for any 18–74-year-old female relative living in western Washington who might want help understanding her risk for breast cancer. We enclosed a form for providing contact information for relatives and an addressed stamped envelope for easy return. Cases approached via the mammography study received a follow-up phone call if their materials were not returned within 2 weeks. We hypothesized that recruitment through breast cancer cases would provide a good estimate of interest in counseling among women with increased breast cancer risk.

Recruitment through Mass Media.

A general press release announcing the study was sent to all broadcast and print media outlets in the metropolitan Seattle area. This resulted in public service announcements on several popular radio stations, a news story on a local television network affiliate, and small articles in a variety of weekly and daily newspapers, including one of the two major daily Seattle papers. We also contacted the employee newsletter editors for the largest employers in the region and were able to include an announcement of the study in all of these publications. We hypothesized that women recruited through media outlets would report more worry and higher perceived risk than women recruited through breast cancer cases.

Eligibility

Eligibility requirements for the study were: age of 18–74 years, residence within 60 miles of the research institute, at least one relative with breast cancer (degree of relationship not specified), no personal history of breast or ovarian cancer, expressed interest in participating in any of the three study conditions, willingness to complete questionnaires, and no report of a family history of cancer that is indicative of a possible autosomal dominant inherited predisposition to breast cancer. Those included in this latter “genetic high-risk” group had two or more first- or second-degree relatives in the same biological line diagnosed with breast cancer before age 50 or ovarian cancer at any age. Women who met this criteria were not considered appropriate candidates for a randomized trial of moderate risk counseling and were referred to genetic counseling services at the University of Washington.

Measures

Questions measuring key psychological variables were included in the telephone screening survey completed by all potential counselees. Case-recruited women were mailed the survey to complete and mail back in a postage-paid envelope. Media-recruited women completed the survey by phone. We measured cancer worry and risk perception using questions from the Cancer Genetic Studies Consortium Core Item list. The Cancer Genetic Studies Consortium is a group of investigators funded by the NIH to study the psychosocial issues surrounding genetic testing related to cancer risk. One of the products of the consortium was a common core of instruments, to be used by the consortium members for pooled analyses. The cancer worry scale is a four-item scale developed by Lerman et al. (6) to measure the level of chronic worry about cancer in healthy individuals. This scale has been widely used to study cancer-related distress in a variety of genetic counseling studies. Risk perception was measured by a single item asking women to compare their risk of getting breast cancer to “most women” using eight categories, ranging from 1 (“little or no chance”) to 8 (“much higher than average”). Additional questions measured personal history of breast or ovarian cancer, age (in years), and initial interest participating in either of the counseling options offered (individual or group) or a delayed counseling comparison group using five categories, ranging from 1 (“not at all interested”) to 5 (“very interested”). Potential participants were also asked to provide their number of blood relatives diagnosed with breast or ovarian cancer.

Stages of Assessment

Regardless of source, potential participants completed a brief telephone survey to determine eligibility for the research program and initial interest in participating in counseling.

After participants completed the screening questionnaire and eligibility confirmation, study staff mailed each potential participant a baseline questionnaire. This baseline questionnaire contained several measures of anxiety, depression, breast cancer screening frequency, quality of life, and knowledge about breast cancer and genetic mutation testing. If a woman mailed back a completed questionnaire, study staff called her to fully describe study procedures, confirm willingness to participate in all aspects of the study, and invite her to participate. Women who completed this stage were randomized into one of the counseling conditions and scheduled for participation. At 6 and 24 months, postrandomization participants completed follow-up surveys by mail and telephone.

Counseling Interventions

Eligible women were randomly assigned to one of three arms: individual genetic counseling, group psychosocial counseling, or delayed counseling (control group). The genetic counseling intervention was adapted from traditional genetic counseling protocols for women with a strong family history of breast cancer. It included pedigree analysis, review of genetic and nongenetic risk factors, discussion of screening, and current research in medical genetics. The psychosocial intervention was based on cognitive behavioral theories, counseling protocols for cancer patients, and studies of health promotion. It consisted of four 2-h sessions combining health education, skills training, and social support. Topics included risk and risk factors for breast cancer, screening technique, social support, and stress management. Women in both counseling interventions received an individualized risk estimate based on the models of Gail et al. (9) and Claus et al. (10). Women randomized to the delayed counseling condition were asked to wait to receive counseling until completion of all data collection for the study.

Analyses

We first calculated levels of response at each stage of the recruitment process, both overall and by recruitment source. We then conducted analyses of cancer worry and risk for each recruitment source. We calculated the bivariate relationship between descriptive, reproductive, and psychological variables and entry into the study. For these and all other bivariate comparisons, we used t tests and χ2 tests as appropriate, to identify significant differences. Finally, we created a logistic regression model predicting entry into the study using initial interest in participating in the counseling project, worry, and actual and perceived risk as predictor variables.

Recruitment.

Table 1 shows the rates of recruiting potential counselees through breast cancer cases. Overall, almost half (46%) of cases who were approached responded to the contact letter asking for information about potentially interested relatives. Approximately half of the cases who responded, or 24% of the original number of cases, reported at least one relative who met the simple eligibility requirements. Each case responding with any eligible relative provided one to three relatives. Therefore, the total yield of relatives was 44%. The relative yield varied slightly, but not significantly, by the type of case approached. For example, the proportions of breast cancer cases who responded to the contact letter from the oncologist’s office, the case-control study, and the mammography study were 45, 40, and 53%, respectively. The proportions of cases reporting eligible relatives were also different (22, 18, and 38%, respectively), but again, these differences were not significant.

The proportion of potential participants who agreed to continue with screening and go on to participate in counseling is shown in Table 2 for each of the stages of screening. A total of 588 women responded to the brief media solicitations over a 15-week period. The 205 responding relatives of breast cancer cases brought the total number of potential participants to 793. Table 2 shows the number of potential contacts by source of woman, the number of eligible women, and the percentage yield at telephone survey and final entry stage. Eligibility in this table is defined as reporting the data needed at each stage, meeting the eligibility requirements of the study, and expressing continued interest in participating in breast cancer risk counseling. The proportion of women eligible at each stage varied by participant source. Three-fourths of women who called the counseling line as a result of media attention (78%) were eligible. Reasons for ineligibility for media-recruited women at the telephone survey stage included personal history of breast or ovarian cancer (3%), not having a relative with breast cancer (18%), not living within 60 miles of the research center (4%), age (1%), lack of interest in counseling (13%), and meeting the criteria for referral to genetics services (67%). Fifty % of the case-recruited relatives were eligible to receive a baseline questionnaire. For case-recruited women, reasons for ineligibility included personal history of breast or ovarian cancer (10%) and lack of interest in counseling (90%). Information provided by the cases allowed us to screen for proximity to the research center, age, and family history of cancer before contacting their relatives.

The differences in response between media- and case-recruited women continued through to the entry stage, in which participants were randomized into the study. The total yield of women interested in counseling after the screening process was 45%, including a 53% yield for media-recruited women and 23% for case-recruited women.

Telephone Survey Results.

There were no differences in demographic variables for case- and media-recruited women in this study. Also, there were no demographic differences in these variables among sources of case-recruited relatives, although the small sample sizes make such comparisons unreliable. We examined the levels of cancer worry and risk for participants recruited from breast cancer cases and from the general public media recruitment. Participants recruited from media contacts reported slightly but significantly higher levels of worry about getting cancer [6.33 ± 1.9 and 5.79 ± 1.5 (means ± SD) for media and case recruits, respectively; P < 0.01]. These media-recruited participants also reported slightly but significantly higher levels of perceived personal risks for breast cancer (6.54 ± 1.4 and 6.13 ± 1.6 for media and case recruits, respectively; P < 0.01).

We separately examined the effects of number of affected blood relatives on cancer worry and perceived risk in media participants. Fifty-six % of the participants reported having two or more relatives diagnosed with breast or ovarian cancer, whereas 40% reported having only one affected relative. There was not a significant difference in cancer worry between the women with one affected relative and those with two or more affected relatives. However, there was a significant difference in perceived risk between these two groups (P < 0.01), with mean perceived risk scores of 6.2 ± 1.4 and 6.9 ± 1.1 for one and two or more relative participants, respectively.

Table 3 presents the relationships between the predictor variables and the outcome of entry into the study. There was a significant difference in participation by initial interest level with a higher average interest level among those who participated in the study. The source of recruitment also significantly differed with a higher percentage of participants resulting from the media recruitment. Those who participated in the study also had a significantly higher average perceived risk score and a higher percentage of participants had two or more relatives with breast or ovarian cancer, compared to those with only one relative. There was no significant difference in the average age or perceived risk score between the women who participated and those who did not.

Predicting Entry into the Counseling Trial.

Table 4 presents the results of a logistic regression predicting entry in the counseling project (1, participation; 0, no participation). The independent variables in the equation were age, perceived risk, cancer worry, source of recruitment (2, case; 1, media), and initial interest in counseling as measured on the telephone survey. Neither age nor perceived risk were significant predictors of eventual participation in the counseling project. Cancer worry negatively and significantly predicted participation in the counseling project, in that the likelihood of participation decreased with increasing levels of cancer worry. The source of recruitment predicted participation, with the likelihood of participating higher for media participants Finally, initial expression of interest was the strongest predictor of counseling participation, with higher levels of interest increasing the ultimate likelihood of participation. Interactions with source of recruitment were tested for all other predictors of counseling participation using the likelihood ratio test. No interactions were found to be significant predictors nor did any of the interactions improve the model at the 0.05 significance level.

Here, we present data from women who are comparable in their cancer worry to others that have participated in this type of research (11, 12). In these prior studies, the average cancer worry scores of between 5 and 6.5 on the Lerman scale, are comparable to the values found in our study. Lerman et al.(6) used the two items that measure the impact of cancer worry on mood and functioning to define “moderate” as between 3 and 4. We doubled the values (e.g., ≥6) because of the full four-item scale used in this study. If we use this cutoff point, most of the women in this study have moderate levels of cancer worry. This raises a concern for women in the general public, in that elevated levels of worry or concern about cancer risk could interfere with general functioning and quality of life. This should be the focus of future research efforts.

It is not surprising that women approached through the media differ from those approached through relatives with breast cancer with respect to interest in participating in a risk counseling study and with respect to perceived personal risk for breast cancer. There could be several reasons for this: (a) women responding to media advertisements have already self-selected themselves from a pool of persons with all levels of interest and perceived risk; (b) the media-recruited women had “stronger” family histories of breast cancer than case-recruited women; or (c) individuals who volunteer to participate in research studies are different from those who do not.

The results presented here may have implications for counseling women in the general population with a family history of breast cancer about their risk for the disease. These implications include the idea that women with higher perceived risk will be more likely to volunteer for help and that initial expression of interest is a good, but not perfect, indicator of actual participation. This is a topic of increasing interest, as the identification of genetic sequences conferring susceptibility to breast cancer may lead to calls for large-scale population testing and/or screening, especially if preventive interventions can be advised for individuals at risk. Already, the identification of a mutation in BRCA1 that at first appeared specific to Jewish women has led some to suggest that general testing of this population group is now feasible (13) and desirable (14). Prior to generalized population testing, women with a personal and/or a family history of breast cancer may be perceived as reasonable targets for genetic susceptibility testing.

The populations studied in this research are not necessarily representative of the general population, in which women with no family history of breast cancer likely comprise the majority of individuals (15). In addition, all participants were recruited using approaches designed to target individuals with a particular interest in participating in breast cancer research. Despite these limitations, examination of data from the media-recruited group may provide insights that are relevant to women in the general population. The media-recruited group is most relevant to women seeking risk counseling in the general population because this sample was not selected through prior participation in research studies (participants recruited through existing FHCRC studies) or for relation to someone undergoing care for breast cancer (participants recruited through oncologist’s practice). Also, the media-recruited group was “self-referred,” in contrast to all other participants in which a breast cancer case was used as a resource for participant names. The case-recruited group may be most relevant to women or families offered counseling through specialized medical services, such as a cancer genetics counseling practice setting. Still other limitations of the present study included the measurement of perceived risk with only a single item and the disparity in sample sizes across cells in some of the analyses, which could lead to difficulties in analysis. Replication and extension of the research questions would give more confidence in definitive answers.

The use of media sources proved to be a useful recruitment method for this counseling study of breast cancer risk, providing ≈71% of the total potential contacts in a very short period of time. This suggests that there is a great deal of interest in the topic of breast cancer and personal risk in the general population of western Washington, among women with a family history of breast cancer. This may be especially true for genetically high-risk women because they comprised over half of the media-recruited ineligibles. The effectiveness of the media recruitment strategy also suggests that use of various media resources may be an effective way to disseminate information to women about breast cancer risk in general.

Media-recruited women participating in this study were very interested in obtaining counseling and personalized risk information. It is notable that this group of participants expressed more worry about getting breast cancer and had higher levels of perceived personal risk for breast cancer than other study participants. However, among the media-recruited women, those with a stronger family history had significantly higher perceived risk but no significant difference in cancer worry.

Further research may illuminate the reasons for these observations. For example, it is not possible to know whether these observations would hold for women without a family history of breast cancer. Media recruitment strategies may have contributed to this effect, by attracting women especially concerned about breast cancer and wanting assistance understanding their personal risk. In contrast, in recruitment via breast cancer cases, patients decide which relatives’ names to provide, and the relatives aren’t volunteering to participate. However, the increase in worry and higher perceived risk exhibited by media-recruited participants may indicate an increased need for assistance in understanding their risk for breast cancer for individuals seeking risk counseling. There are many uncertainties regarding how best to deliver cancer risk information in this time of increasing complexity of genetic and environmental risk information and increasing health care cost containment. Data from this randomized controlled trial of two counseling interventions will be valuable in addressing these questions.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

        
1

This study was funded by a grant from the National Human Genome Research Institute. This work was reviewed and approved by the Investigational Review Board of the Fred Hutchinson Cancer Research.

                
3

The abbreviation used is: FHCRC, Fred Hutchinson Cancer Research Center.

Table 1

Identification of breast cancer case relatives

Total no.Yield (%)a
Total no. of cases contacted 558  
Total no. of case responders 237 46 
No. of case responders with eligible relatives 132 24 
No. of eligible relatives provided 243 44 
Total no.Yield (%)a
Total no. of cases contacted 558  
Total no. of case responders 237 46 
No. of case responders with eligible relatives 132 24 
No. of eligible relatives provided 243 44 
a

Potential counselees divided by total number of cases at each stage.

Table 2

Recruitment of potential counselees

StagePotential contacts (total n = 793)No. eligible% yielda
MediaCaseTotalMediaCaseTotalMediaCaseTotal
Telephone 288 205 793 458 101 559 78 50 70 
Entry into the study 309 47 356 309 47 356 53 23 45 
StagePotential contacts (total n = 793)No. eligible% yielda
MediaCaseTotalMediaCaseTotalMediaCaseTotal
Telephone 288 205 793 458 101 559 78 50 70 
Entry into the study 309 47 356 309 47 356 53 23 45 
a

Percentage eligible at telephone and entry stage of original potential contacts.

Table 3

Levels of predictor variables by eventual participation in the study

ParticipationP
Yes (n = 357)No (n = 364)
Initial interest (x̄ ± SD) 4.33 ± 0.87 3.59 ± 1.34 0.00 
Source of recruitment    
 media, n (%) 310 (87) 269 (74) 0.00 
 case, n (%) 47 (13) 95 (26)  
Perceived risk score (x̄ ± SD) 6.6 ± 1.2 6.3 ± 1.5 0.02 
Cancer worry score (x̄ ± SD) 6.1 ± 1.7 6.3 ± 2.0 0.33 
Age (x̄ ± SD) 42.4 ± 11.5 41.5 ± 13.4 0.35 
ParticipationP
Yes (n = 357)No (n = 364)
Initial interest (x̄ ± SD) 4.33 ± 0.87 3.59 ± 1.34 0.00 
Source of recruitment    
 media, n (%) 310 (87) 269 (74) 0.00 
 case, n (%) 47 (13) 95 (26)  
Perceived risk score (x̄ ± SD) 6.6 ± 1.2 6.3 ± 1.5 0.02 
Cancer worry score (x̄ ± SD) 6.1 ± 1.7 6.3 ± 2.0 0.33 
Age (x̄ ± SD) 42.4 ± 11.5 41.5 ± 13.4 0.35 
Table 4

Predictors of counseling trial participation

Predictor variablesParameter estimateSEPOdds ratio
Initial interesta 0.58 0.08 0.000 1.8 
Source of recruitmentb −0.38 0.22 0.08 0.68 
Perceived risk score 0.07 0.06 0.23 1.1 
Cancer worry score −0.14 0.04 0.002 .9 
Age 0.004 0.007 0.53 1.0 
Predictor variablesParameter estimateSEPOdds ratio
Initial interesta 0.58 0.08 0.000 1.8 
Source of recruitmentb −0.38 0.22 0.08 0.68 
Perceived risk score 0.07 0.06 0.23 1.1 
Cancer worry score −0.14 0.04 0.002 .9 
Age 0.004 0.007 0.53 1.0 
a

Five-point scale ranging from “not at all interested” to “very interested.”

b

2, case; 1, media.

1
Breast Cancer Technical Committee. Breast Cancer in Washington State: Status of the Problem and Recommendations for Intervention Washington State Department of Health and Fred Hutchinson Cancer Research Center Seattle  
1990
.
2
National Cancer Institute, Office of Cancer Communications. National Cancer Institute Breast Cancer Research: Highlights National Cancer Institute Bethesda  
1988
.
3
Miki Y., Swensen J., Shattuck-Eidens D., Futreal P. A., Harshman K., Tartigian S., Lui Q., Cochran C., Bennett L. M., Ding W. A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1.
Science (Washington DC)
,
266
:
66
-71,  
1994
.
4
Wooster R., Bignell G., Lancaster J., Swift S., Seal S., Mangion J., Collins N., Gregory S., Gumbs C., Micklem G. Identification of the breast cancer susceptibility gene BRCA2.
Nature (Lond.).
,
378
:
789
-792,  
1995
.
5
Fries, E. A., White, K. S., and Bowen, D. J. Impact of behavior-specific versus trait social desirability on health-enhancing and health-compromising behaviors. Basic Appl. Soc. Psychol., in press, 1999.
6
Lerman C., Trock B., Rimer B. K., Jepson C., Brody D., Boyce A. Psychological effects of breast cancer screening.
Health Psychol.
,
10
:
259
-267,  
1991
.
7
Kash K. M., Holland J. C., Halper M. S., Miller D. G. Psychological distress and surveillance behaviors of women with a family history of breast cancer.
J. Natl. Cancer Inst.
,
84
:
24
-30,  
1992
.
8
Rimer B. K., Schildkraut J. M., Lerman C., Lin T. H., Audrain J. Participation in a women’s breast cancer risk counseling trial. Who participates? Who declines?.
Cancer (Phila.)
,
77
:
2348
-2355,  
1996
.
9
Gail M. H., Brinton L. A., Byar D. P., Corle D. K., Green S. B., Schairer C., Mulvihill J. J. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually.
J. Natl. Cancer Inst.
,
81
:
1879
-1886,  
1989
.
10
Claus E. B., Risch N., Thompson W. D. Autosomal dominant inheritance of early onset breast cancer.
Cancer (Phila.)
,
73
:
643
-651,  
1994
.
11
Lerman C., Daly M., Sands C., et al Mammography adherence and psychological distress among women at risk for breast cancer.
J. Natl. Cancer Inst.
,
85
:
1074
-1080,  
1993
.
12
Lerman C., Schwartz M. D., Miller S. M., Daly M., Sands C., Rimer B. K. A randomized trial of breast cancer risk counseling: interacting effects of counseling educational level and coping style.
Health Psychol.
,
15
:
75
-83,  
1996
.
13
Struewing J. P., Abeliovich D., Peretz T., et al The carrier frequency of the BRCA1 185delAG mutation is approximately 1% in Ashkenazi Jewish individuals.
Nat. Genet.
,
11
:
198
-200,  
1995
.
14
Schulman J. D., Stern H. J. Genetic predisposition testing for breast cancer.
Cancer J. Sci. Am.
,
2
:
244
-249,  
1996
.
15
Mettlin C., Croghan I., Natarajan N. The association of age and familial risk in a care-control study of breast cancer.
Am. J. Epidemiol.
,
131
:
973
-983,  
1990
.