the decision to test in women receiving genetic counseling for brca1 and brca2 mutations

P1: KEF

Journal of Genetic Counseling [jgc] PH266-jogc-486128 May 7, 2004 18:19 Style file version Nov 28th, 2002

Journal of Genetic Counseling, Vol. 13, No. 3, June 2004 (C© 2004)

The Decision to Test in Women Receiving GeneticCounseling forBRCA1and BRCA2Mutations

Kimberly Kelly, 1,6 Howard Leventhal,2 Michael Andrykowski, 1

Deborah Toppmeyer,3 Judie Much,3 James Dermody,4

Monica Marvin, 5 Jill Baran, 3 and Marvin Schwalb4

Functions of genetic counseling include provision of risk information and provisionof support in an effort to assist with decision making. This study examines (1) therelationship among intentions to test, self-reported provision of blood sample, andreceipt of test results; (2) the impact of genetic counseling on distress specific togene status, perceived risk of developing breast and ovarian cancer in the contexthaving BRCA1/2 mutations (mutations predisposing to increased risk of breast–ovarian cancer), and perceived risk factors for breast cancer; and (3) the clinicalprofile of those receiving/not receiving results. Intentions to test for BRCA1/2mutations, self-reported provision of blood sample immediately after counseling,and receipt of test results were statistically different but highly correlated, andintentions to test increased from pre- to postcounseling. A repeated measuresANOVA found distress specific to gene status and perceived risk factors decreasedfrom pre- to postcounseling. Further, two clinical profiles of consultands emerged:(1) those receiving results with change in intentions to test having lower levelsof distress and (2) those not receiving results and those receiving results withno change in intentions to test with higher levels of distress. Our findings areconsistent with the function of genetic counseling—to provide information andsupport to those with familial cancer, as well as to assist in decision making. The

1Department of Behavioral Science, University of Kentucky, Lexington, Kentucky.2Institute for Health, Health Care Policy, and Aging Research, Rutgers University, New Brunswick,New Jersey.

3LIFE Center – Cancer Risk Assessment and Genetic Counseling Program – Breast Oncology, CancerInstitute of New Jersey, New Brunswick, New Jersey.

4Center for Human and Molecular Genetics, University of Medicine and Dentistry of New Jersey, TheNew Jersey Medical School, Newark, New Jersey.

5Spectrum Health, Grand Rapids, Michigan.6Correspondence should be directed to Kimberly Kelly, PhD, University of Kentucky College ofMedicine, Department of Behavioral Science, College of Medicine Office Building, Lexington,Kentucky 40536-0086; e-mail: [email protected].

237

1059-7700/04/0600-0237/1C© 2004 National Society of Genetic Counselors, Inc.

P1: KEF


238 Kelly et al.

provision of support is important as distress specific to gene status may impedeflexible decision making about genetic testing.

KEY WORDS: BRCA1/2testing; genetic testing; breast cancer; ovarian cancer; genetic counseling;risk perception; psychosocial factors.

INTRODUCTION

Women carrying mutations inBRCA1/2have as high as an 85% risk of breastcancer and as high as a 60% risk of ovarian cancer by age 70 (Eastonet al., 1995;Oddouxet al., 1996). In the Ashkenazi Jewish population, the frequency of threeBRCA1andBRCA2mutations (the Ashkenazi Jewish panel) is approximately 2.6%(Roa et al., 1996), whereas the general population frequency is approximately0.1%. (Struewinget al., 1997). The presence of risk factors, such as personalor family history of breast or ovarian cancer and multiple generations affectedincreases the risk of having a mutation (Fitzgeraldet al., 1996; Oddouxet al.,1996). Hence, these mutations occur at a higher frequency in the Ashkenazi Jewishpopulation than in the general population and represent a significant health threatdue to their association with cancer.

The National Society for Genetic Counselors recommends that genetic testingbe conducted in the context of genetic counseling and defines the functions of ge-netic counseling as providing information and providing support to families regard-ing genetic disorders (National Society of Genetic Counselors, 2003). In addition,genetic counselors adhere to a professional code of nondirectiveness, a code thatis value-neutral and promotes patient autonomy in decision making (Djurdjinovic,1998; Kessler, 1992). Providing information and providing support in a value-neutral environment are critical functions of genetic counseling because optimaldecision making about screening for cancer and testing for cancer-predisposing ge-netic mutations requires knowledge about genetic disorders without the hindranceof excessive worry. Although these functions are critical for genetic counseling, thefield needs studies that investigate the relationship of perceived risk of cancer, dis-tress, and the receipt of test results in response to genetic counseling forBRCA1/2mutations. As a result of receiving genetic counseling forBRCA1/2mutations,individuals should understand more about their risks of having a mutation anddeveloping cancer, and they should be better able to manage worry about havinga mutation and developing cancer.

Distress and Risk in the Context of Genetic Counseling

Two meta-analyses of distress and risk perception in the context of geneticcounseling have been conducted (Butowet al., 2003; Meiser and Halliday, 2002).

P1: KEF


Decision to Test forBRCA1/2 239

Some studies have found that genetic counseling increases overall mental health(Cull et al., 1999) and decreases anxiety (Cullet al., 1999; Watsonet al., 1999),whereas other studies found no impact of genetic counseling on overall mentalhealth (Watsonet al., 1999), depression (Lermanet al., 1996a; Randallet al., 2001),intrusive thoughts of breast cancer (Randallet al., 2001), or anxiety (Randallet al.,2001). However, and most significantly, no studies found that genetic counselingincreaseddistress. Further, a number of studies have examined perceived risk ofcancer and perceived risk of having a mutation before and after genetic counseling(e.g., Burkeet al., 2000; Cullet al., 1999; Evanset al., 1994; Lermanet al., 1995,1997; Watsonet al., 1999). Most of these studies examine the change in accuracyof perceived risk from pre- to postcounseling. Generally, these studies found thatalthough individuals decrease in their perceived risk of breast cancer in the directionof increased accuracy (Burkeet al., 2000; Cullet al., 1999; Evanset al., 1994;Lermanet al., 1995; Watsonet al., 1999), perceived risk of breast cancer continuesto be overestimated (Burkeet al., 2000; Lermanet al., 1995; Watsonet al., 1999).

Although these studies have examined change in accuracy of perceived risk ofcancer and perceived risk of having a mutation, we identified no studies examiningchange in perceived risk factors for breast cancer (e.g., personal history of breastcancer, history of breast feeding), change in perceived risk of ovarian cancer, orchange in perceived risk of breast and ovarian cancer in the context ofBRCA1/2mutations. Further, we identified no studies examining change in distress specificto gene status, a more proximal measure of distress than worry about cancer. We donot know (1) if perceptions of risk factors change in response to genetic counseling,(2) what an individual believes her risk of breast and ovarian cancer to be if shedoes or does not have aBRCA1/2mutation, or (3) if distress specific to gene statusis lowered as a result of genetic counseling.

Indeed, an individual may have a contextually embedded, contingency-basedestimate of risk. In other words, because individuals are in thecontextof having afamily history of breast–ovarian cancer and in thecontextof genetic counseling forhereditary breast–ovarian cancer and preparing for genetic testing forBRCA1/2mutations, and because they are considering thecontingencyof a positive or nega-tive result, it is important to understand what they believe their risk will be if theirresults are positive or negative. Individuals should believe that a positive test resultis associated with a risk of breast and ovarian cancer well above general populationrates and a negative result is associated with a risk of breast and ovarian cancer ator slightly above the general population risk. One could argue that the receipt oftest results has little meaning if there is no expected change in perceived risk ofcancer as the result of receiving a positive or negative result. Indeed, the perceptionthat risk of cancer could be lowered by receiving a negative result may serve as amotivator to test. Prior methods of assessing risk have only examined perceivedrisk of breast or ovarian cancer in general and have not made the distinction of thecontingencyof whether one has a positive or negative result.

P1: KEF


240 Kelly et al.

Further, we advocate for a method of assessing distress that is contextuallyembedded. Methods of assessing distress that examine more distal, generic distressor that focus on worry about cancer do not capture a critical element of the ge-netic counseling experience: distress specific to gene status. Indeed, prior studies(Blumanet al., 1999; Kellyet al., 2003) have found that individuals had grosslyoverestimated risk of having aBRCA1/2mutation. We argue that assessing dis-tress specific to having a mutation inBRCA1/2is equally important as (1) geneticcounseling endeavors to assist individuals in managing worry and (2) elevatedlevels of distress specific to gene status may be associated with the decision totest, in spite of equivalent levels of objective risk. Therefore, it is critical that moreproximal measures of distress in the context of genetic counseling be utilized, suchas distress specific to gene status.

Intentions to Test

Although not in the context of genetic counseling, studies examining the re-lationship between distress and intentions to test find that higher scores on theCancer Worry Scale adapted for breast (Gwynet al., 2003) and ovarian (Lermanet al., 1994) cancers, concern about colon cancer (Petersenet al., 1999), the Impactof Events Scale adapted for cancer (Kinneyet al., 2001) and the Profile of MoodStates (Lermanet al., 1994) were associated with greater intentions to test. Stud-ies examining the relationship of perceived risk with intentions to test (not in thecontext of genetic counseling) are less consistent than those examining distressand intentions to test. Four studies found that those with higher perceived riskof ovarian cancer (Lermanet al., 1994; Schwartzet al., 2000), higher perceivedrisk of having aBRCA1mutation (Struewinget al., 1995), and higher perceivedrisk of having a hereditary mutation (Andrykowskiet al., 1997) were more likelyto desire testing. Two studies found no relationship between intentions to testand perceived risk of hereditary cancer at precounseling (Cappelliet al., 1999)or change in perceived risk of having aBRCA1mutation from pre- to postcoun-seling (Lermanet al., 1997). One study found no relationship between perceivedrisk of breast cancer at postcounseling and the receipt of test results (Leeet al.,2002).

We located only one study examining the change in intentions to test for aBRCA1mutation in response to genetic counseling. This study found no changein intentions to test and that higher precounseling intentions to test and greaterfamily history were predictors of higher postcounseling intentions to test, and dis-tress and risk were explored as predictors of intentions to test (Lermanet al.,1997). As the provision of information about risk and the provision of sup-port are critical to genetic counseling, it is important to explore their impacton intentions to test and on receipt of test results in the context of geneticcounseling.

P1: KEF



Genetic Testing

A number of studies have independently examined intentions to test for hered-itary mutations that predispose to cancer (e.g., Gwynet al., 2003; Kinneyet al.,2001; Leeet al., 2002; Lermanet al., 1994) and self-report of receipt ofBRCA1test results (Lermanet al., 1996a). However, our search resulted in only one studythat included a combination of two decision to test measures: intentions to testand provision of blood sample. In their study, Lermanet al. (1997) found that61% of women intended to test for aBRCA1mutation before counseling, 71%of women intended to test after counseling, and 52% provided a blood sample.We located no studies examining relationships among intention to test, provisionof blood sample, and actual receipt of genetic test results. It is helpful to knowto what extent conclusions about intentions to test and self-reported provision ofblood sample can be generalized to actual receipt of test results.

This study was designed to examine (1) the impact of genetic counseling ondistress specific to gene status, perceived risk of cancer in the context of havingmutations, and perceived risk factors for breast cancer; (2) the relationship amongintention to test, self-reported provision of blood sample, and receipt of test results;and (3) the clinical profile of those receiving and not receiving results. We hypoth-esized that (1) distress specific to gene status, perceived risk factors, and perceivedrisk of cancer would be associated with precounseling intention to test and receiptof results forBRCA1/2mutations; (2) intention to test, self-reported provision ofblood sample, and the receipt of test results would be highly positively correlated;(3) distress specific to gene status, perceived risk factors, perceived risk of cancer,and intention to test would decrease following counseling; and (4) those not re-ceiving results would have low levels of distress and perceived risk of cancer, andthose receiving results with intention to test not impacted by genetic counselingwould have high levels of distress and perceived risk.

METHODS

Participants

Women (N = 106) from approximately 70 families were recruited for thestudy through newspaper articles, letters to local physicians (oncologists and gy-necologists), and presentations at relevant community groups (Jewish organiza-tions and breast cancer interest groups). All women self-identified as of AshkenaziJewish descent. For women with a family history of cancer (whether or not they hada personal history of cancer), the line of transmission was traced to an AshkenaziJewish grandparent. For individuals with a personal history with no family history,at least one grandparent was of Ashkenazi Jewish descent. Additionally, partic-ipants met one of the criteria noted below to increase the probability of finding

P1: KEF


242 Kelly et al.

Table I. Demographic Characteristics of the Study Sample (N = 106)

Percent

Cancer historyPrior personal cancer history 50.0

Breast only 41.5Ovarian only 1.9Endometrial only 0.9Prostate only 0.9Breast and ovarian 3.7Breast and colon 0.9

Family history only 50.0Test result

Positive (mutation present) 19.0Informative negative (mutation located in family not present) 5.7Uninformative negative (no mutation located but may be other) 63.8No test 11.4

Marital statusMarried 78.6Single, never married 7.8Divorced/separated 5.8Widowed 7.8

Highest level of educationSome graduate school through postgraduate school 52.0Some college through college graduate 41.1High school (12th grade) 6.9

Annual family income$100,000 or greater 43.0$50,000–99,000 39.0$20,000–49,999 15.0$19,999 or less 3.0

Agea Range= 18–83Mean= 49.0SD= 13.4

aPercentage frequencies were not reported for age.

a BRCA1or BRCA2mutation. The demographics of the sample are shown inTable I.

Women with a previous diagnosis of cancer were eligible if they (1) had breastcancer diagnosed before the age of 50; (2) had breast cancer at any age and had atleast one first or second degree relative with either breast cancer diagnosed beforethe age of 50 or ovarian cancer at any age; or (3) had been diagnosed with ovariancancer at any age. Women without a previous diagnosis of cancer were eligibleif they (1) had one first degree relative with breast cancer before the age of 50,ovarian cancer at any age, and another first or second degree relative with breastcancer before the age of 60 or ovarian cancer at any age; (2) had both a paternalsecond degree relative with breast cancer before the age of 50 or ovarian cancerat any age and a paternal second degree relative with breast cancer before the ageof 60 or ovarian cancer at any age; (3) were first degree relatives of a male withbreast cancer at any age; (4) were first or second degree relatives of an individualwith a documentedBRCA1or BRCA2mutation.

P1: KEF



Procedure

Eligibility was determined by a board certified or board eligible genetic coun-selor. Once determined eligible, participants were mailed a questionnaire packetthat included a consent form and the precounseling questionnaire. The consentform asked for participation in (1) counseling, (2) tape recording, and (3) inter-viewing. The precounseling questionnaire contained questions from a larger studyinvestigating the response to counseling and testing forBRCA1/2mutations. Dataon perceived risk factors, perceived risk of cancer, distress specific to gene status,and genetic testing collected as part of this study are reported here. After com-pleting the precounseling survey, participants were instructed to phone the geneticcounselor to schedule a genetic counseling session. Participants received free ge-netic counseling and testing in return for their participation. Participation did notrequire tape recording of counseling sessions.

The consent form and precounseling questionnaire were returned at the coun-seling session. The counselor reviewed the consent form with the participant andanswered questions prior to beginning the counseling session. Counseling wasconducted by a board certified or board eligible genetic counselor (N = 3) con-sistent with an outline of a standard cancer genetic counseling session providedby the Familial Cancer Risk Counseling Special Interest Group of the NationalSociety of Genetic Counselors (National Society of Genetic Counselors, 2003).Genetic counseling topics included mechanisms of cancer inheritance, meaningof a positive and a negative result, risks and benefits associated with mutationtesting, efficacy of health interventions to reduce cancer risk, and cancer risks as-sociated with mutations. Sessions lasted 1–2 hr. A second informed consent formwas completed for participants opting to be tested forBRCA1/2mutations. Bloodwas drawn from individuals electing to be tested.

One to 2 days following the counseling session, a member of the researchteam other than the genetic counselor conducted the postcounseling telephoneinterview. The participants’ decision to proceed with testing was obtained by thegenetic counselor in a subsequent phone call. Testing was done by a CLIA approvedlaboratory by standard methods for the Ashkenazi Jewish panel (Toninet al., 1996).For those desiring genetic test results, genetic test results were provided in a face-to-face meeting with a genetic counselor.

Measures

Demographic Data

Questionnaire items covering demographic data included personal historyof cancer (yes/no), gender, age, years of education, and annual family income.These questions were assessed at precounseling. Further, pedigrees (e.g., fam-ily members, relationships of family members, cancer history, age of onset of

P1: KEF


244 Kelly et al.

cancer) constructed during genetic counseling were entered into the Cancer Genesoftware. On the basis of pedigrees, BRCAPRO estimates of objective risk ofhaving a mutation were calculated for women in the study (Parmigianiet al.,1998).

Distress Specific to Gene Status

We tailored six items from the Profile of Mood States (POMS; Lebo andNesselroade, 1978; McNairet al., 1971) to focus on distress specific to genestatus, and we designed three additional items to assess issue-oriented distress.Three items (nervous, anxious, and tense) were chosen from the POMS to rep-resent anxious mood. These were then adapted to focus on gene status (e.g.,“nervous about gene status”). Three items (depressed, sad, and hopeless) werechosen from the POMS to represent depressed mood. These were then adaptedto focus on gene status (e.g., “depressed about gene status”). Further, from ourdiscussions about the clinical issues surrounding those with and at risk for breastcancer, we identified three key concerns of patients: disfigurement, death, andpain. These three concerns led to the development of three items (worried aboutdisfigurement following breast cancer [as it relates to your gene status], afraidof dying [as it relates to your gene status], and concerned about pain [as it re-lates to your gene status]) to assess issue-oriented distress related to gene sta-tus. These nine items used the same 5-point scale as used in the POMS: a 5-point Likert-type scale (1= not at all, 2= a little bit, 3= moderately, 4= quitea bit, 5= very much). Distress specific to gene status was assessed at pre- andpostcounseling.

Perceived Risk Factors

Nine items assessed to what extent the individual felt a particular risk factorincreased or decreased her chance of developing breast cancer. The factors includedwere risk factors culled from a review of the literature (e.g., Beral and Reeves, 1993;Colditzet al., 1993; Fiorettiet al., 1999; Gailet al., 1989; McTiernan and Thomas,1986); however it was noted that we do not know the extent to which these factorspose risk for cancer in those with hereditary mutations inBRCA1/2genes. Risk fac-tors included personal history of breast cancer, early menarche, family history ofbreast or ovarian cancer, late menopause, history of breast feeding (protective), useof birth control pills, use of estrogen replacement therapy (ERT), delayed childbirth(after 30), and having no full-term pregnancies. Each risk factor was rated using a5-point scale (1= decrease greatly, 2= decrease, 3= not change/not applicable,4= increase, and 5= increase greatly). Items in this section were included at boththe pre- and postcounseling assessments.

P1: KEF



Perceived Risk of Cancer

Six questions assessed how likely women perceived their likelihood of devel-oping a specific type of cancer with and withoutBRCA1orBRCA2gene mutations.Those with a prior history were asked to estimate their chance of getting canceragain. Items included perceived risk of breast cancer (again) withoutBRCA1orBRCA2mutations, perceived risk of ovarian cancer (again) withoutBRCA1orBRCA2mutations, perceived risk of breast cancer (again) with aBRCA1mutation,perceived risk of ovarian cancer (again) with aBRCA1mutation, perceived risk ofbreast cancer (again) with aBRCA2mutation, and perceived risk of ovarian cancer(again) with aBRCA2mutation. Participants were asked to respond with a per-centage (0–100%). These items were included at both the pre- and postcounselingassessments.

Decision to Test

To assess intentions to test, participants were asked, “At this point in time, whatdo you think your decision will be about taking a genetic test for breast cancer?”The response scale ranged from−10 to 10 and was anchored with “absolutelywill not” and “absolutely will,” respectively. Intentions to test were assessed atboth the pre- and postcounseling assessments. One item assessed the self-reportedprovision of blood sample: “After the counseling session, did you leave a bloodsample to be analyzed for the gene mutations?” Participants could respond yesor no. This item was included in the postcounseling assessment. Hence, the itemto assess self-reported provision of blood sample would not capture those whoprovided a blood sample after the postcounseling assessment. Finally, informationregarding receipt of test results was provided by the genetic counselor. Those whoreturned for a genetic results session to receive results were coded as receivingresults, and those who did not return for a genetic results session were coded asnot receiving results.

Statistical Analysis

Means, standard deviations, and frequencies were generated to describe thedata and ensure suitability for parametric statistical analysis. Three Principal Com-ponent Analyses were conducted to examine the underlying factor structure of (1)the nine-item distress specific to gene status scale, (2) the nine-item perceived riskfactors scale, and (3) the six-item perceived risk of cancer scale. Eigenvalues>1and factor loadings>.45 were sought. Three Cronbach alphas were computed for(1) the distress specific to gene status items, (2) the perceived risk factors items,and (3) the perceived risk of cancer items. Cronbach alphas>.6 were sought. Three

P1: KEF


246 Kelly et al.

mean scores were computed: one for the distress specific to gene status items, onefor the perceived risk factors items, and one for the perceived risk of cancer items.

A hierarchical regression was conducted for intention to test at precounselingwith age, years of school, annual family income, and personal history of cancerentered in the first block and distress specific to gene status, perceived risk factors,and perceived risk of cancer at precounseling entered in block two. Power fora multiple hierarchical regression with 4 control variables (r 2 = .5) and threepredictor variables (r 2 = .05) was adequate (N = 106,λ = 11.33,β = .20,α =.05) at 80%. A logistic regression was conducted for receipt of test results withage, years of school, annual family income, and personal history of cancer enteredin the first block and distress specific to gene status, perceived risk factors, andperceived risk of cancer at postcounseling entered in block two. Again, power wasadequate for this regression (N = 106,ORnull = 0.05,ORalternative= 1.5,r 2 = .05,β = .20,α = .05) at 80%.

Chi square analyses explored differences between the intentions to test atpre- and postcounseling, intentions to test postcounseling and self-reported pro-vision of blood sample postcounseling, and self-reported provision of blood sam-ple and receipt of test results. Pre- and postcounseling intention to test were di-chotomized such that those who reported intentions of “10” were considered asone group at each time point, and those with lower levels of intentions wereconsidered as another group at each time point. Power for a chi square anal-ysis (N = 106, χ2 = 9.54, w= .30, β = .13, α = .05) was acceptable at 87%.The relatedness of intention to test, self-reported provision of blood sample, andthe receipt of test results was assessed with correlations. Specifically, a standardPearson product–moment correlation was computed for pre- and postcounselingintentions to test; point biserial correlations were computed for intention to testwith self-reported provision of blood sample and intention to test with receiptof test results; and phi was computed for the comparison of provision of bloodsample and receipt of test results. Power was acceptable to examine relatedness(N = 106, r = .3, β = .12, α = .05) at 88%.

Within-participantst tests were computed to determine change in distressspecific to gene status, perceived risk factors, perceived risk of cancer, and in-tentions to test over time. Power for the within-participantst test with a meandifference of 0.5 was high (N = 106, precounselingσ = 1.00, postcounselingσ = 2.00, β = 0.05, α = 0.05) at 95%. Finally, a repeated measures analysis ofvariance was conducted to determine change from pre- to postcounseling for dis-tress specific to gene status, perceived risk factors, perceived risk of cancer with thebetween-participants factor of decision to test (no test, test with no change in in-tention to test, and test with change in intention to test) with BRCAPRO estimatedrisk of having a mutation entered as a covariate. Because of the small sample size insome cells, the 12 groups delineated in Table II were collapsed into 3. Group 2 hadno change in intentions to test from pre- to postcounseling and received test results.

P1: KEF



Tabl

eII.

Gro

ups

asa

Fun

ctio

nof

Inte

ntio

nsto

Test

,Sel

f-R

epor

ted

Pro

visi

onof

Blo

odS

ampl

e,an

dR

ecei

ptof

Test

Res

ults

%of

Inte

ntio

nsto

test

atS

elf-

repo

rted

prov

isio

nof

Rec

eipt

ofG

roup

Sam

ple

pre-

and

post

coun

selin

gbl

ood

sam

ple

post

coun

selin

gte

stre

sults

No

grou

p1%

Unk

now

nY

esY

es#2

Rec

eivi

ngre

sults

withn

och

ange

inin

tent

ions

38%

Gav

era

tings

of10

atpr

e-an

dpo

stco

unse

ling

Yes

Yes

1%N

och

ange

inra

tings

atpr

e-an

dpo

stco

unse

ling

Yes

Yes

1%N

och

ange

inra

tings

from

pre-

topo

stco

unse

ling

No

Yes

#1R

ecei

ving

resu

ltsw

ithch

ange

inin

tent

ions

42%

Incr

ease

inra

tings

from

pre-

topo

stco

unse

ling

Yes

Yes

2%D

ecre

ase

inra

tings

from

pre-

topo

stco

unse

ling

Yes

Yes

2%In

crea

sein

ratin

gsfr

ompr

e-to

post

coun

selin

gN

oY

es1%

Dec

reas

ein

ratin

gsfr

ompr

e-to

post

coun

selin

gN

oY

es#0

Not

rece

ivin

gte

stre

sults

1%N

och

ange

inra

tings

from

pre-

topo

stco

unse

ling

Unk

now

nN

o1%

Incr

ease

inra

tings

from

pre-

topo

stco

unse

ling

Unk

now

nN

o2%

No

chan

gein

ratin

gsfr

ompr

e-to

post

coun

selin

gN

oN

o5%

Incr

ease

inra

tings

from

pre-

topo

stco

unse

ling

No

No

3%D

ecre

ase

inra

tings

from

pre-

topo

stco

unse

ling

No

No

P1: KEF


248 Kelly et al.

Group 1 had a change in intentions to test from pre- to postcounseling and receivedtest results. Group 0 did not receive test results. Power for a repeated measuresanalysis with three levels of a between-participants factor (n1 = 35, n2 = 35, n3 =35, df= 1,102, σ = 20.00, effect size= .40, β = .04, α = .05), two levels ofa within-participants factor (N = 105, df= 1,102, σ = 4.00, effect size=.80,β > .001, α = .05), and an interaction effect (N = 105, df= 1,102, σ = 4.00,effect size= .80,β > .001, α = .05) was near 100%.

RESULTS

Scale Formation

A Principle Components Analysis was performed to determine the underlyingfactor structure of the nine items in the distress specific to gene status scale. Aone-factor solution was achieved (eigenvalue= 5.29) with factor loadings rangingfrom .46 (worried about disfigurement following breast cancer) to .89 (anxious).Internal consistency for the nine-item scale was high (α = .90). A second PrincipleComponents Analysis was performed to determine the underlying factor structureof the nine items in the perceived risk factors scale. A three-factor solution wasachieved (eigenvalues> 1.0). Personal history of breast cancer, family historyof breast or ovarian cancer, and history of breast feeding did not load with theremaining six items. The remaining six items had factor loadings ranging from .47(use of birth control pills) to .82 (delayed childbirth). However, internal consistencyfor the nine-item scale was acceptable (α = .60), and it was decided to considerthe nine perceived risk factor items as a single scale. A single scale was chosenas three scales would decrease power in subsequent analyses. Finally, participantsmade six probability judgments of perceived risk of cancer. A third PrincipleComponents Analysis was performed to determine the underlying factor structureof the six items in the perceived risk of cancer scale. Two factors had eigenvalues>1; however two items were common to both factors. The factor analysis was thenrecomputed with a forced one-factor solution. A one-factor solution was achieved(eigenvalue= 3.53) with factor loadings ranging from .56 (risk of ovarian cancerwithout aBRCA1/2mutation) to .86 (perceived risk of ovarian cancer with aBRCA2mutation). Internal consistency for the six-item scale was good (α = .86).

Relationship of Intention to Test, Self-Reported Provisionof Blood Sample Postcounseling, and Receipt of Test Results

We hypothesized that distress specific to gene status, perceived risk factors,and perceived risk of cancer would be associated with precounseling intentionto test and receipt of results forBRCA1/2mutations. A hierarchical regressionwas conducted for intention to test at precounseling with age, years of school,

P1: KEF



annual family income, and personal history of cancer entered in the first block anddistress specific to gene status, perceived risk factors, and perceived risk of cancerentered in the second block. This model was not significant. A logistic regressionwas conducted for receipt of test results with age, years of school, annual familyincome, and personal history of cancer entered in the first block and distress specificto gene status, perceived risk factors, and perceived risk of cancer entered in blocktwo. This model was also not significant.

Table II shows individuals in the study as a function of their intention totest, self-reported provision of blood sample, and receipt of test results. Chi squareanalyses explored differences between the intentions to test at pre- and postcounsel-ing, intentions to test postcounseling and self-reported provision of blood samplepostcounseling, and self-reported provision of blood sample and receipt of testresults. Intentions to test at pre- and postcounseling differed,χ2(1, N = 102)=17.06, p < .001, such that 40% of participants went from intentions to test of“−10” to “9” at precounseling to “10” at postcounseling. Intentions to test at post-counseling and self-reported provision of blood sample postcounseling differed,χ2(1, N = 103)= 39.45, p < .001, such that 8% of participants reported inten-tions to test of “−10” to “9” at postcounseling but reported leaving a blood sample,and 4% of participants reported intentions to test of “10” but did not leave a bloodsample. Finally, self-reported provision of blood sample at postcounseling andreceipt of test results differed,χ2(1, N = 95)= 57.52, p < .001, such that 5%reported that they did not provide a blood sample after the counseling session butreceived test results.

We also hypothesized that intentions to test, self-reported provision of bloodsample postcounseling, and the receipt of test results would be highly positivelycorrelated. A standard Pearson product–moment correlation was computed to de-termine the relationship of pre- and postcounseling intentions to test. Point biserialcorrelations were computed to determine the relationships of intentions to test withself-reported provision of blood sample and intentions to test with receipt of testresults. Phi (correlation) was computed to determine the relationship betweenself-reported provision of blood sample and receipt of test results. As shown inTable III, all of the variables were highly correlated, and receipt of test results wasmost highly correlated with self-reported provision of blood sample.

Change Over Time and the Decision to Test

To test the hypothesis that distress specific to gene status, perceived riskfactors, perceived risk of cancer, and intentions to test would decrease followingcounseling, two separate analyses were conducted. First, within-subjectst testswere conducted on the three distress specific to gene status subscales (anxiety,depression, and issue-oriented), the nine perceived risk factors, the six perceivedrisk of cancer items, and intentions to test forBRCA1/2mutations. Table IV shows

P1: KEF


250 Kelly et al.

Table III. Correlation Table of Decision to Test Variables

Precounseling Postcounseling Self-reported Receiptintention intention provision of blood of test

to test to test sample postcounseling results

Precounseling intention to test —Postcounseling intention to test r = .67∗ —Self-reported provision of blood rpb = .54∗ rpb = .61∗ —

sample postcounselingReceipt of test results rpb = .53∗ rpb = .54∗ 8 = .78∗ —

Note.Receipt of test results was most highly correlated with provision of blood sample. Precounselingintentions to test and receipt of results were the least highly correlated. Correlations of precounselingintentions to test with receipt of results and postcounseling intentions to test with receipt of test resultswere similar.∗ p < .01.

that the three distress specific to gene status subscales (anxiety, depression, andissue-oriented distress), seven of the nine perceived risk factors (personal historyof breast cancer, history of breast feeding, use of birth control pills, use of es-trogen replacement therapy, delayed childbirth, having no full-term pregnancies,and early menarche) and three of the six perceived risk of cancer items (perceivedrisk of breast cancer without mutations, perceived risk of ovarian cancer withoutmutations, and perceived risk of ovarian cancer withBRCA2mutation) decreasedfrom pre- to postcounseling. However, intentions to test increased.

Second, a repeated measures analysis of variance was conducted to determinechange from pre- to postcounseling for distress specific to gene status, perceivedrisk factors, perceived risk of cancer with the between-subjects factor of decisionto test (no test, test with no change in intentions to test, and test with changein intentions to test) with BRCAPRO estimated risk of having a mutation en-tered as a covariate. This analysis also tested our fourth hypothesis that thosenot receiving results would have low levels of distress and perceived risk of can-cer, and those receiving results with intentions to test not impacted by geneticcounseling would have high levels of distress and perceived risk. Main effectsemerged for change over time,F(3, 82)= 6.09, p = .001, and the decision totest,F(3, 83)= 3.29, p = .03. As illustrated in Table IV, distress specific to genestatus decreased from pre- to postcounseling, and perceived risk factors decreasedfrom pre- to postcounseling.

Further, as shown in Fig. 1, distress specific to gene status was found todiffer as a function of decision to test,F(2, 84)= 4.75, p = .01. At precounseling,individuals who tested with a change in intentions to test had lower distress specificto gene status than those receiving results with no change in intentions to test(p = .008), and a trend emerged such that individuals not receiving results hadhigher distress specific to gene status than those receiving results with a changein intentions to test (p = .07). At postcounseling, individuals who tested with a

P1: KEF



Table IV. Change Over Time in Distress Specific to Gene Status, Perceived Risk Factors, PerceivedRisk of Cancer, and Intentions to Test

Precounseling PostcounselingMean (SD) Mean (SD) Significance

Distress specific to gene status 2.2 (0.84) 1.9 (0.74) F(1, 84)= 10.73, p = .002Anxiety specific to gene status 2.4 (1.1) 2.1 (1.0) t(1, 91)= 3.95, p = .001Depression specific to gene status 1.9 (0.84) 1.7 (0.84)t(1, 102)= 2.92, p = .004Issue-oriented distress specific to 2.3 (0.87) 1.9 (0.84)t(1, 102)= 4.55, p = .001

gene statusPerceived risk factors 3.5 (0.38) 3.3 (0.38) F(1, 84)= 9.02, p = .004Personal history of breast cancer 3.8 (1.0) 3.6 (0.77)t(1, 99)= 2.35, p = .02Family history of breast or ovarian 4.4 (0.98) 4.2 (0.73)ns

cancerHistory of breast feeding 2.5 (0.61) 2.6 (0.61)t(1, 100)= −2.20, p = .03Use of birth control pills 3.4 (0.77) 3.1 (0.66) t(1, 99)= 3.64, p = .001Use of estrogen replacement therapy 3.6 (0.84) 3.2 (0.50)t(1, 100)= 4.39, p = .001

(ERT)Delayed childbirth (after 30) 3.4 (0.62) 3.2 (0.46) t(1, 100)= 3.84, p = .001No full-term pregnancies 3.4 (0.66) 3.1 (0.51)t(1, 100)= 4.18, p = .001Early menarche 3.5 (0.68) 3.2 (0.65) t(1, 95)= 4.13, p = .001Late menopause 3.1 (0.74) 3.1 (0.47)nsPerceived risk of cancer 45.8 (21.2) 41.2 (19.3) nsBreast cancer withoutBRCA1or 34.5 (23.6) 26.6 (23.7) t(1, 88)= 3.07, p = .003

BRCA2mutationsOvarian cancer withoutBRCA1or 25.9 (21.9) 16.2 (19.3) t(1, 84)= 4.17, p = .001

BRCA2mutationsBreast cancer withBRCA1mutation 60.5 (27.6) 62.4 (26.7) nsBreast cancer withBRCA2mutation 59.0 (27.7) 59.0 (27.3) nsOvarian cancer withBRCA1mutation 47.2 (26.8) 44.1 (24.1) nsOvarian cancer withBRCA1mutation 46.4 (27.9) 39.4 (24.4) t(1, 87)= 2.37, p = .02Intention to test 7.2 (3.7) 8.7 (3.4) t(1, 101)= −4.94, p = .001

change in intentions to test had lower distress specific to gene status than thosereceiving results with no change in intentions to test (p = .01), and a trend emergedsuch that individuals who did not test had higher distress specific to gene status thanthose receiving results with a change in intentions to test (p = .07). No differencesemerged between those not receiving results and those receiving results with nochange in intentions to test. No interaction or main effects emerged for BRCAPROestimated risk of having a mutation.

DISCUSSION

This study examined: (1) the relationship of precounseling intentions to test,postcounseling intentions to test, self-reported provision of blood sample, andreceipt of test results; (2) change in distress specific to gene status, perceived riskfactors, and perceived risk of cancer from pre- to postcounseling; and (3) the clinicalprofile of those receiving and not receiving results. Our new scales (distress specific

P1: KEF


252 Kelly et al.

Fig. 1. Distress specific to gene status as a function of time and decision to test.

to gene status, perceived risk factors, and perceived risk of cancer) demonstratedacceptable statistical properties, especially those for distress specific to gene statusand perceived risk of cancer.

We hypothesized that distress specific to gene status, perceived risk factors,and perceived risk of cancer would be associated with precounseling intentionto test and receipt of results forBRCA1/2mutations; however, our models werenot significant. Our findings are not consistent with studies finding that distressand risk were associated with intentions to test (Butowet al., 2003; Meiser andHalliday, 2002) and we believe this inconsistency is due to our high-risk AshkenaziJewish, highly motivated sample compared to lower risk samples in prior studies.Our findings are consistent with Leeet al. (2002), that perceived risk of cancerand receipt of results were not associated.

A series of chi square analyses evidenced that intentions to test, the self-reported provision of blood sample, and the receipt of test results were separateconcepts and mark distinct steps in the genetic testing process. Despite similaritieson regression analyses for intentions to test and receipt of test results and thecorrelations of the decision to test variables, care must be used in the extrapolationof information on studies that use the proxy outcomes of interest in testing andself-reported provision of blood sample when examining the receipt of test results.Further, we hypothesized that intentions to test, the self-reported provision ofblood sample, and the receipt of test results would be highly correlated. Thishypothesis was confirmed. The variables that were most highly correlated wereself-reported provision of blood sample and the receipt of test results. Intentions

P1: KEF



to test at pre- (r = .53) and postcounseling (r = .54) were similarly related to thereceipt of test results. The relationship of health intentions with health behavioris similar to that reported in the larger psychosocial literature (e.g., cannabis useintentions and behavior,r = .51; Sheppard, 1989; meta-analysis of condom useintentions and behavior,r = .45; Albarracinet al., 2001). We suspect that in less-motivated samples, precounseling intentions to test, postcounseling intentions totest, self-reported provision of blood sample, and receipt of results will be lesshighly correlated.

As predicted, decreases from precounseling to postcounseling were notedin three distress specific to gene status subscales (anxiety, depression, and issue-oriented distress), seven of the nine perceived risk factors (personal history ofbreast cancer, history of breast feeding, use of birth control pills, use of estro-gen replacement therapy, delayed childbirth, having no full-term pregnancies, andearly menarche) and three of the six perceived risk of cancer items (perceived riskof breast cancer without mutations, perceived risk of ovarian cancer without mu-tations, and perceived risk of ovarian cancer withBRCA2mutation). The findingsthat distress specific to gene status and perceived risk of cancer declined reaffirmedthe function of genetic counseling to support and inform patients. Women in thestudy found most of the risk factors to have a negligible effect on their personalbreast cancer risk, indicating that either (1) the factor was not personally relevantor (2) if the factor was relevant, it did not increase their personal risk of breastcancer. Apparently, genetic counseling decreased the risk related to somatic (en-vironmental) factors, possibly due to the increased salience of hereditary cancerrisk.

Our hypothesis that intentions to test would decrease was not supported: incontrast to the findings of Lermanet al. (1997), we found that intentions to testincreased. Nearly 40% of our sample rated their precounseling intention to test as“definitely will,” and this was not changed by genetic counseling. However, over40% of our sample increased in intentions to test from pre- to postcounseling. Toour knowledge, these testing rates (89%) are higher than those previously citedin the literature; although, a recent study of high-risk Norwegians reported ratesof 63% (Boddet al., 2003). We believe our hightesting rates were due to (1) theavailability of free genetic counseling and genetic testing, (2) family membersrecruiting other family members and hence increased likelihood of informativeresults, and (3) our highly educated sample of Ashkenazi Jewish women whovalued research (based on anecdotal reports).

Further, a repeated measures analysis found that distress specific to gene statusand perceived risk factors decreased following counseling, whereas there was nochange noted in the perceived risk of cancer. Our hypothesis that those not receivingresults would have low levels of distress and perceived risk of cancer and thatthose receiving results with no change in intentions to test would have high levelsof distress and perceived risk of cancer was partially supported. There were no

P1: KEF


254 Kelly et al.

significant differences in perceived risk of cancer or perceived risk factors amongthose not receiving results, those receiving results with no change in intentionsto test, and those receiving results with change in intentions to test. However,these three groups did differ in distress specific to gene status. Those receivingresults with change in intentions to test had lower levels of distress specific to genestatus than those not receiving results and those receiving results with no changein intentions to test. The U-shaped curve depicted in Fig. 1 could not be found instudies that use linear or binary logistic regression models to predict decision totest (e.g., Lermanet al., 1997). Although our methodology does not allow us todetermine causality and the absolute differences in the levels of distress amongthe three groups are small, we suspect that those presenting for counseling withlower distress specific to gene status are able to learn more from counseling andcan make more informed decisions about genetic testing.

Certain limitations of the study should be noted. One limitation to consideris the independence of data gleaned in this study. As noted in the participantssection, once one individual was found to be positive, other family members wereencouraged to have genetic counseling. Frequently, two or more family memberspresented for the initial genetic counseling together. Hence, some individuals inthe study were from families where three or more relatives tested. Also, concernsabout the generalizability of the findings should be considered. Women in thisstudy were of Ashkenazi Jewish descent, well-educated, and had a high income.Concern may be raised about the usefulness of findings about risk perceptionsthat were gleaned from such a highly educated population. Yet, it is important toremember that particular areas were troublesome (e.g., perceived risk of cancer) forthese women, and these areas may need to be specially targeted in future educationsessions in women with lower levels of education. Specifically, ovarian cancer riskwith mutations and the role of risk factors in the development of cancer couldbe targeted as areas for education. With this potential increase in knowledge andaccuracy, the best possible decision can be made about testing for these mutations.

In sum, genetic counseling decreased gene status specific distress and per-ceived risk factors. Further, genetic counseling increased intentions to test. Thus,our findings are consistent with the function of genetic counseling—to provideinformation and support to those with familial cancer. Our data indicate that largenumbers of individuals who receive test results may present for genetic counselingalready knowing that they want to test. Although this may evidence nondirec-tiveness on the part of the genetic counselors involved, the finding is somewhatconcerning in that increased knowledge of risk and decreased distress specific togene status did not appear to greatly influence the decision to test. Further, ourresults indicate that measures of intentions to test and self-reported provision ofblood sample are not perfectly correlated with receiving test results. Thus, morestudies should be conducted to investigate receipt of test results and the role ofdistress specific to gene status in diverse populations.

P1: KEF



ACKNOWLEDGMENTS

This study was supported by grants from the New Jersey Commission onCancer Research and the Mid-Atlantic Regional Human Genetics Network. Anearlier version of this manuscript was submitted in fulfillment of the first author’smaster’s thesis. The first author thanks Drs Gretchen Chapman and Ann O’Learywho served on her thesis committee and provided guidance on the development ofthe thesis and therefore this manuscript. In addition, the authors thank Drs SusanBrownlee, Chantal Robitaille, Yael Benyamini, and Linda Patrick-Miller for theirassistance in survey development and the individuals who participated in this studywho made this study possible.

REFERENCES

Albarracin, D., Johnson, B. T., Fishbein, M., & Muellerleile, P. A. (2001). Theories of Reasoned Actionand Planned Behavior as models of condom use: A meta-analysis.Psychol Bull, 127(1), 142–161.

Andrykowski, M. A., Lightner, R., Studts, J. L., & Munn, R. K. (1997). Hereditary cancer risk notifi-cation and testing: How interested is the general population?J Clin Oncol, 15(5), 2139–2148.

Beral, V., & Reeves, G. (1993). Childbearing, oral contraceptive use and breast cancer.Lancet, 34,1102.

Bluman, L. G., Rimer, B. K., Berry, D. A., Borstelmann, N., Iglehart, J. D., Regan, K.,et al. (1999).Attitudes, knowledge, and risk perceptions of women with breast and/or ovarian cancer consideringtesting for BRCA1 and BRCA2.J Clin Oncol, 17(3), 1040–1046.

Bodd, T. L., Reichelt, J., Heimdal, K., & Møller, P. (2003). Uptake of BRCA1 genetic testing in adultsisters and daughters of known mutation carriers in Norway.J Genet Couns, 12(5), 405–417.

Burke, W., Culver, J. O., Bowen, D., Lowry, D., Durfy, S., McTiernan, A.,et al. (2000). Geneticcounseling for women with an intermediate family history of breast cancer.Am J Med Genet, 90,361–368.

Butow, P. N., Lobb, E. A., Meiser, B., Barratt, A., & Tucker, K. M. (2003). Psychological outcomesand risk perception after genetic testing and counselling in breast cancer: A systematic review.Med J Aust, 178(2), 77–81.

Cappelli, M., Surh, L., Humphreys, L., Verma, S., Logan, D., Hunter, A.,et al. (1999). Psychologicaland social determinants of women’s decisions to undergo genetic counseling and testing for breastcancer.Clin Genet, 55,419–430.

Colditz, G. A., Willett, W. C., Hunter, D. J., Stampfer, M. J., Manson, J. E., Hennekens, C. H.,et al.(1993). Family history, age, and risk of breast cancer. Prospective data from the Nurses HealthStudy.JAMA, 270,1563–1568.

Cull, A., Anderson, E. D. C., Campbell, S., Mackay, J., Smyth, E., & Steel, M. (1999). The impact ofgenetic counselling about breast cancer risk on women’s risk perceptions and levels of distress.Br J Cancer, 79(3/4), 501–508.

Djurdjinovic, L. (1998). Psychosocial counseling. In D. L. Baker, J. L. Schuette, and W. R. Uhlmann(Eds.),A guide to genetic counseling(pp. 128–170). New York: Wiley-Liss.

Easton, D. F., Ford, D., Bishop, T., & Breast Cancer Linkage Consortium. (1995). Breast and ovariancancer incidence in BRCA1-mutation carriers.Am J Hum Genet, 56,265–271.

Evans, D. G. R., Blair, V., Greenhalgh, R., Hopwood, P., & Howell, A. (1994). The impact of geneticcounselling on risk perception in women with a family history of breast cancer.Br J Cancer, 70,934–938.

Fioretti, F., Tavani, A., Bosetti, C., La Vecchia, C., Negri, E., Barbone, F.,et al. (1999). Risk factorsfor breast cancer in nulliparous women.Br J Cancer, 79(11/12), 1923–1928.

P1: KEF


256 Kelly et al.

Fitzgerald, M. G., MacDonald, D. J., Krainer, M., Hoover, I., O’Neil, E., Unsal, H.,et al. (1996).Germ-line BRCA1 mutations in Jewish and non-Jewish women with early-onset breast cancer.N Engl J Med, 334(3), 143–149.

Gail, M. H., Brinton, L. A., Byar, D. P., Corle, D. K., Green, S. B., Schairer, C.,et al.(1989). Projectinginidividualized probabilities of developing breast cancer for white females who are being examinedannually.J Natl Cancer Inst, 81,1879–1886.

Gwyn, K., Vernon, S. W., & Conoley, P. M. (2003). Intention to pursue genetic testing for breast canceramong women due for screening mammography.Cancer Epidemiol Biomarkers Prev, 12, 96–102.

Kelly, K., Leventhal, H., Toppmeyer, D., Much, J., Dermody, J., Marvin, M.,et al. (2003). Subjectiveand objective risk of carrying a BRCA1/2 mutation in individuals of Ashkenazi Jewish descent.J Genet Couns, 12(4), 351–371.

Kessler, S. 1992. Psychological aspects of genetic counseling. VII. Thoughts of directiveness.J GenetCouns, 1, 9–17.

Kinney, A. Y., Croyle, R. T., Dudley, W. N., Bailey, C. A., Pelias, M. K., & Neuhausen, S. L. (2001).Knowledge, attitudes, and interest in breast–ovarian cancer gene testing: A survey of a largeAfrican-American kindred with a BRCA1 mutation.Prev Med, 33,543–551.

Lebo, M., & Nesselroade, J. (1978). Intraindividual differences dimensions of mood change duringpregnancy identified in five P-technique factor analyses.J Res Pers, 12,205–224.

Lee, S., Bernhardt, B. A., & Helzlsouer, K. J. (2002). Utilization of BRCA1/2 genetic testing in theclinical setting: Report from a single institution.Cancer, 94(6), 1876–1885.

Lerman, C., Biesecker, B. B., Benkendorf, J. L., Kerner, J., Gomez-Caminero, A., Hughes, C.,et al.(1997). Controlled trial of pretest education approaches to enhance informed decision-making forBRCA1 gene testing.J Natl Cancer Inst, 89(2), 148–157.

Lerman, C., Daly, M., Masny, A., & Balshem, A. (1994). Attitudes about genetic testing for breast–ovarian cancer susceptibility attitudes about genetic testing for breast–ovarian cancer susceptibil-ity. J Clin Oncol, 12(4), 843–850.

Lerman, C., Lustbader, E., Rimer, B., Daly, M., Miller, S., Sands, C.,et al. (1995). Effects of individ-ualized breast cancer risk counseling: A randomized trial.J Natl Cancer Inst, 87(4), 286–292.

Lerman, C., Narod, S., Schulman, K., Hughes, C., Gomez-Caminero, A., Bonney, G.,et al. (1996a).BRCA1 testing in families with hereditary breast–ovarian cancer.JAMA, 275(24 SP-1885), 1892.

Lerman, C., Schwartz, M., Miller, S., Daly, M., Sands, C., & Rimer, B. (1996b). A randomized trial ofbreast cancer risk counseling: Interacting effects of counseling, education level, and coping style.Health Psychol, 15(2), 75–83.

McNair, D., Lorr, M., & Droppleman, L. (1971).Profile of mood states. San Diego, CA: Educationaland Industrial Testing Service.

McTiernan, A., & Thomas, D. B. (1986). Evidence of a protective effect of lactation on risk of breastcancer in young women. Results from a case control study.Am J Epidemiol, 124(3), 353–358.

Meiser, B. & Halliday, J. L. (2002). What is the impact of genetic counselling in women at increased riskof developing hereditary breast cancer? A meta-analytic review.Soc Sci Med, 54(10), 1463–1470.

National Society of Genetic Counselors. (2003). Retrieved from [Web Page] www.nsgc.orgOddoux, C., Struewing, J. P., Clayton, C. M., Neuhausen, S., Brody, L. C., Kaback,et al. (1996).

The carrier frequency of the BRCA2 6174delT mutation among Ashkenazi Jewish individuals isapproximately 1%.Nat Genet, 14,188–190.

Parmigiani, G., Berry, D. A., & Aguilar, O. (1998). Modeling risk of breast cancer and decisions boutgenetic testing.Am J Hum Genet, 62,145–148.

Petersen, G. M., Larkin, E., Codori, A. M., Wang, C. Y., Booker, S. V., Bacon, J.,et al.(1999). Attitudestoward colon cancer gene testing: Survey of relatives of colon cancer patients.Cancer EpidemiolBiomarkers Prev, 8(4, P. 2), 337–344.

Randall, J., Butow, P., Kirk, J., & Tucker, K. (2001). Psychological impact of genetic counselling andtesting in women previously diagnosed with breast cancer.Intern Med J, 31,397–405.

Roa, B., Boyd, A., Volcik, K., & Richards, S. (1996). Ashkenazi Jewish population frequencies forcommon mutations in BRCA1 and BRCA2.Nat Genet, 14,185–187.

Schwartz, M. D., Hughes, C., Roth, J., Main, D., Peshkin, B. N., Isaacs, C.,et al. (2000). Spiritualfaith and genetic testing decisions among high-risk breast cancer probands.Cancer EpidemiolBiomarkers Prev, 9, 381–385.

P1: KEF



Sheppard, M. A. (1989). Adolescents’ perceptions of cannabis use by their peers: Does it have anythingto do with behavior?J Drug Educ, 19(2), 157–164.

Struewing, J. P., Hartge, P., Wacholder, S., Baker, S. M., Berlin, M., McAdams, M.,et al. (1997). Therisk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews.N Engl J Med, 336(20), 1401–1408.

Struewing, J. P., Lerman, C., Kase, R., Giambarresi, T., & Tucker, M. (1995). Anticipated uptakeand impact of genetic testing in hereditary breast and ovarian cancer families.Cancer EpidemiolBiomarkers Prev, 4, 169–173.

Tonin, P., Weber, B., Offit, K., Couch, F., Rebbeck, T. R., Neuhausen, S., andet al. (1996). Frequencyof recurrent BRCA1 and BRCA2 mutations in Ashkenazi Jewish breast cancer families.Nat Med,2(11), 1179–1183.

Watson, M., Lloyd, S., Davidson, J., Meyer, L., Eeles, R., Ebbs, S.,et al.(1999). The impact of geneticcounselling on risk perception and mental health in women with a family history of breast cancer.Br J Cancer, 79(5/6), 868–874.

the decision to test in women receiving genetic counseling for brca1 and brca2 mutations

Documents