Editor—The localisation of the two breast cancer susceptibility genes BRCA1 andBRCA2 made possible the use of mutation detection as a susceptibility test for people who wish to learn whether they carry a risk conferring mutation.1-4 Several studies have assessed attitudes to genetic testing for breast cancer susceptibility,5-11 most of which involved either community samples or women with just one first degree relative with breast cancer. The objective of our study was to assess attitudes to genetic testing for breast cancer susceptibility in a large sample of women at high risk of developing hereditary breast cancer on the basis of family history. The majority of women included in our sample (80%) had a family history consistent with a dominantly inherited predisposition to breast cancer (lifetime risk of 1 in 4 to 1 in 2),12 and the remainder (20%) was at moderately increased risk of developing breast cancer (lifetime risk of 1 in 8 to 1 in 4).12

The findings reported here are based on a sample of 461 unaffected women with a family history of breast cancer. Women who approached one of 14 familial cancer clinics and six associated outreach clinics in five Australian states between November 1996 and January 1999 were eligible for participation. Women were considered ineligible for study participation if they had a previous diagnosis of ovarian or breast cancer, were unable to give informed consent, or had limited literacy in English, since data were collected using self-report questionnaires. The study was approved by 16 institutional ethics committees.

Familial cancer clinic staff invited women to participate in the study during the preclinic telephone call, where possible. Questionnaires, consent forms, and reply paid envelopes were then mailed out by the coordinating research centre. Women were subsequently telephoned by the central research staff and given further information about the study and issues of informed consent. Women were asked to return the completed questionnaire and consent form before attending the familial cancer clinic, if possible. Reminder calls were made as required.

Sex, age, educational level, marital status, number and sex of biological children, and referral source were assessed. To provide an estimate of objective risk, clinic staff were asked to make a judgment on whether a participant's family history was either consistent or not consistent with a dominantly inherited predisposition to breast cancer, and participants were thus classified as being at “high risk” or “moderately increased risk”, respectively. Following the risk assessment interview at the familial cancer clinic and once pedigree information13 14 and relatives' diagnoses confirmed by medical records were available, clinic staff categorised participants into objective risk groups. For women from high risk families, clinic staff made a judgment on whether the participant was at either 50% or 25% mutation carrier risk. Risk of being a mutation carrier, rather than the estimated lifetime risk of developing breast cancer, was used as a measure of objective risk. The expert opinion of clinical geneticists was used as a gold standard, since there are currently no universally accepted standards to estimate breast cancer risk in high risk women. The number of first and second degree relatives who developed breast or ovarian cancer were collected from study participants.

One item asked participants to select their approximate perceived lifetime breast cancer risk from the following response options: 1%, 4%, 8%, 12%, 16%, 25%, 33%, 50%, 85%, and 100%. Risk was expressed both as a percentage and as odds (for example, 1 in 8).

One item measured perceived likelihood of being a mutation carrier, with five response options ranging from “Certain that I will not have gene” to “Certain I will have gene”. The phrase “have the gene” was used, because of the common misconceptions regarding heterozygosity among the lay population and to facilitate ease of understanding.

The Impact of Event Scale is a 15 item, validated scale that measures anxiety responses about a specific stressful event.15 16Subjects at increased risk of developing hereditary breast cancer may construe their being at risk as a continuous, rather than specific, trauma. Although the scale has not previously been specifically validated among subjects at high risk of developing breast cancer, it has been used in several studies as a measure of breast cancer anxiety7 17 and has previously been shown to be predictive of interest in genetic testing.18 In the current study the particular stressor was concern about being at risk of developing breast cancer. Participants were asked to rate symptoms of anxiety (for example, “I had strong waves of feelings about being at risk of breast cancer”) on a scale ranging from “Not at all” to “Often”.

The Monitoring-Blunting Style Scale is an eight item, validated scale measuring individual differences in coping styles in threatening situations.19 The scale measures a person's tendency either actively to seek out threatening information (“monitoring”) or to ignore it and distract oneself (“blunting”). The Monitoring-Blunting Style scale was included to assess individual differences in responding to genetic risk information.

Six items each assessed perceived benefits and limitations of genetic testing for breast cancer susceptibility. Figs 1 and 2 show items relating to benefits and limitations respectively. These items have been used in related studies.20 21 Women endorsed perceived importance of each benefit or limitation on three response options ranging from “Not at all important” to “Very important”.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Percentages of women endorsing perceived benefits of genetic testing.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Percentages of women endorsing perceived limitations of genetic testing.

Interest in genetic testing was the outcome variable and asked women whether they would be interested in a genetic test. Four response options ranged from “Yes, definitely” to “No, definitely not”.

The outcome variable “interest in genetic testing” was recoded as a binary variable. Only eight participants (2%) reported being “probably not” interested in genetic testing and none reported being “definitely not” interested. It is generally recommended to collapse categories which contain less than 5% of cases to achieve a more balanced distribution.22 Therefore, rather than collapsing the two “yes” and two “no” categories, the new variable was defined as “yes, definitely” versus “yes, probably”, “probably not” and “definitely not”.

To assess associations of predictor variables with the outcome variables, contingency table analysis with χ² tests were performed for categorical predictor variables, Mann-Whitney U, and Kruskal-Wallis tests for ordinal and non-normal interval variables and two sample t tests for normally distributed interval data. For the outcome variable “interest in genetic testing” hierarchical logistic regression was used to identify variables associated independently with interest in genetic testing. All independent variables with a bivariate association of p<0.1 with the outcome variable were entered as predictors and age and objective risk as covariates.

Of the 520 women who met the eligibility criteria, 59 women declined participation or never returned the questionnaire (response rate of 89%). Table 1 summarises sociodemographic and family history variables of the study sample.

The mean age of breast cancer onset in the youngest person in the family was 41 years (SD=9.6). The number of self-reported first and second degree relatives with a diagnosis of breast or ovarian cancer ranged from 1 to 18, with a median of three. Ninety five women (21%) had a family history which included ovarian cancer in addition to breast cancer. All women were assessed before receiving a genetic testing result. Seventy percent of women reported being “definitely”, 22% “probably”, and 2% “probably not” interested in genetic testing, with the remainder (6%) being unsure.

Bivariate analyses between predictor variables and interest in genetic testing showed that subjective carrier risk was significantly associated with interest: 86% of women who reported being “quite certain” or “certain” that they were carriers were “definitely interested” in testing, compared to 71% of those who were uncertain (χ²=8.20, p=0.017). Women who were “definitely interested” in genetic testing had significantly higher breast cancer risk perceptions (Z=−2.08, p=0.038). Neither educational level (χ²=0.82, p=0.37), referral status (χ²=0.32, p=0.57), having daughters or not (χ²=0.22, p=0.63), breast cancer anxiety (Z=−1.48, p=0.14), nor monitoring score (Z=−0.062, p=0.95) were significantly associated with interest in genetic testing. Table 2 shows the final regression model. Only perceived mutation carrier risk was significantly associated with interest in genetic testing (p=0.0017), in that women who were uncertain if they were mutation carriers were less likely to be definitely interested in genetic testing (OR 0.37, 95% CI 0.18-0.73, p=0.0044), compared to women who were quite certain or certain that they were carriers.

Fig 1 shows the percentages of participants who endorsed each perceived benefit as a “very important” factor in deciding about whether or not to have a genetic test. The item endorsed by the highest percentage of women as very important (87%) related to perceiving genetic testing as helpful in understanding what steps to take to reduce one's cancer risk. Learning about one's children's risk was the second most frequently endorsed item, with 77% of women who had children reporting it to be a very important factor.

Fig 2 shows the percentages of participants who endorsed each perceived limitation as a “very important” factor in deciding about whether to have a genetic test. The item endorsed by the highest percentage of women as very important (16%) related to concerns about the effect of genetic testing on the family. Interestingly, only 4% and 11% of women reported that worry about losing insurance was a very important or somewhat important factor, respectively.

Table 3 provides an overview of the results of bivariate analyses to identify associations between demographic and psychological variables and individual perceived benefits and shortcomings of testing. It shows that women who had daughters, compared to women who had sons only, were significantly more likely to report that learning about one's children's risk was an important factor (84% versus 61%, χ²=29.31, p<0.0001). Women who reported that not being able to handle the knowledge emotionally was a very or somewhat important factor had significantly higher breast cancer anxiety than women who reported that this was not at all important (Z=−6.91, p<0.0001). None of the other variables was significantly associated with interest in genetic testing.

This study found that interest in genetic testing for breast cancer susceptibility is very high in a familial cancer clinic population, with 92% of women indicating that they would definitely or probably be interested. This percentage is consistent with previous findings.7-9 23 24 Owing to the clinic based recruitment, the sample may not be representative of high risk women as a whole and is likely to be self-selected for interest in genetic testing. Thus, the high percentage of women reporting interest in genetic testing is perhaps not surprising. Equally, since interest in genetic testing was self-reported, it may not necessarily translate into actual use.

In contrast to findings from other studies,6 8 we found that breast cancer anxiety was not associated with interest in genetic testing. Our results suggest that interest in genetic testing in women with a strong family history of breast cancer is unlikely to be motivated by psychological distress as the primary factor, and contrast with those of related studies.9 18 We found that interest in genetic testing was associated with perceived likelihood of being a mutation carrier, but not objective risk, confirming earlier results from the two studies that assessed attitudes to genetic testing of people with a strong family history of breast cancer.23 24

Perceiving genetic testing as helpful in understanding what steps to take to reduce one's cancer risk and learning about one's children's risk were the most commonly reported reasons for considering testing. These findings are consistent with several studies which assessed reasons for undergoing genetic testing for breast cancer susceptibility,7 9 21 23 25-27 and underline the importance women attribute to both reasons.

Understanding what steps to take to reduce one's cancer risk was endorsed by 87% of women as a very important factor. In contrast, several surveys of attitudes to genetic testing for Huntington's disease identified wanting “to be certain” as the most commonly reported reason.28-31 In the breast cancer scenario, the salience attributed to understanding how to reduce one's breast cancer risk reflects the potential of genetic testing to provide women with a more informed basis for decision making about prophylactic strategies. Perhaps not surprisingly, women with daughters, compared to those with sons only, attributed greater importance to learning about one's children's risk. The greater risk to daughters, compared to sons, seems to be a strong motivator to undergo genetic testing.

Only 15% of participants endorsed worries about insurance as a somewhat or very important shortcoming of genetic testing, compared to 34% of subjects undergoing BRCA1 testing in the United States.21 Contrasting findings are likely to reflect differences in medicolegal factors and government policies for health and life insurance. Health insurance premiums in the United States are risk based, while there is universal government health coverage and even risk based private health insurance in Australia. These differences highlight the ethical issues faced by countries such as the United States, where insurance issues may be among the most important factors influencing a woman's decision concerning genetic testing.

On the whole, the degree of importance attributed to perceived shortcomings was much lower than that attributed to benefits, as has also been observed in related studies in the United States.7 25 This finding suggests that women believe that the benefits of genetic testing outweigh its risks. It also indicates that women may benefit if counsellors provide comprehensive information on the limitations of genetic testing to ensure that decisions about genetic testing are informed decisions. Full deliberation of positive and negative consequences of alternate choices is considered one of the most important features of informed decision making.32 33A randomised trial showed that a genetic counselling approach, but not a purely educational approach, achieved significant increases in perceived limitations and decreases in perceived benefits ofBRCA1 testing.32 Thus, an approach to genetic services provision that relies exclusively on conveying factual information may be less likely to succeed in achieving comprehensive consideration of both benefits and shortcomings of testing.