- © 2008 by American Society of Clinical Oncology
Decision Making Regarding Prophylactic Mastectomy: Stability of Preferences and the Impact of Anticipated Feelings of Regret
- From the Department of Medical Decision Making and Department of Medical Psychology, Leiden University Medical Center, Leiden; Department of Radiation Oncology and Department of Epidemiology, Biostatics, and Health Technology Assessment, Radboud University Nijmegen Medical Center, Nijmegen, the Netherlands
- Corresponding author: Sandra van Dijk, PhD, Department of Medical Psychology, Leiden University Medical Center, Pieter de la Court Building, Postbox 9555, 2300 RB Leiden, the Netherlands; e-mail: dijk{at}lumc.nl
Abstract
Purpose Women who test positive for a BRCA1/2 mutation face difficult choices to manage their breast cancer risk; one of these choices is whether to opt for prophylactic mastectomy. Few data are available about this decision-making process. The current study provides data regarding the stability of risk-management preferences over time and the factors that are associated with these preferences.
Patients and Methods We analyzed data from 338 women who opted for breast cancer antigen (BRCA) testing. First, we prospectively assessed preferences of 80 BRCA mutation carriers at five different points in time ranging from 1 week after blood sampling up to 9 months after BRCA-test disclosure. Second, we applied univariate and multivariate regression analyses to examine which medical, sociodemographic, and psychological factors are related to a preference for prophylactic mastectomy.
Results Ninety percent of the women already indicated a preference regarding risk management at baseline. Moreover, most women had stable preferences over time. Furthermore, anticipated feelings of regret in case of a hypothetical breast cancer diagnosis in the near future were strongly related to risk-management preference (odds ratio = 8.93; P < .0001).
Conclusion Women seem to decide at a relatively early stage about their risk-management preferences. Many of them may be sensitive to the possibility of regret in case of a bad outcome. We discuss whether possible regret in the future is a rational reason for opting for prophylactic mastectomy, or whether it signifies an emotional coping process or strategy in which the future costs are no longer fully considered.
INTRODUCTION
Women who have a high risk of breast cancer due to their inheritance of a BRCA1 or BRCA2 mutation can opt for prophylactic surgery. Several studies have reported on the high efficacy of bilateral prophylactic mastectomy,1-4 and contralateral prophylactic surgery5,6 in reducing breast cancer risk. Furthermore, quite high levels of satisfaction and psychological benefits are reported by women who opted for prophylactic mastectomy.7-9 Even though prophylactic mastectomy may be an acceptable option, the decision to have it has far-reaching and irreversible consequences, and requires careful deliberation at several moments in time.
It has been suggested that preferences to undergo prophylactic mastectomy are not stable, and will not translate into actual behavior if a BRCA mutation is actually detected.10 To date, few data are available about the stability of preferences for risk management,11 whereas several studies have addressed medical and sociodemographic factors that influence decision making.12,13 The current study provides prospective data about preferences and uptake regarding prophylactic mastectomy at various points in time; that is, from before BRCA test disclosure until 9 months later. Furthermore, we inspect which medical, sociodemographic, and psychological factors are related to a preference for prophylactic mastectomy. We focused on the decision between prophylactic mastectomy and regular breast screening, given that some of the current choices for risk management—such as chemoprevention—were not available in the Netherlands during the study period.
PATIENTS AND METHODS
DNA Testing Procedure for Breast Cancer
Referrals for genetic counseling were based on national guidelines, which specify the number and age of affected family members as a prerequisite for referral. Before DNA testing, women usually received two counseling sessions with a geneticist or a genetic counselor. In these sessions, a standard protocol was applied in which a family pedigree was made, and information was provided about genetic risk and the psychosocial consequences of DNA testing. Risk-management options were mentioned briefly, with prophylactic surgery described as an option available for BRCA mutation carriers only. DNA testing was usually offered to individuals from families in which a pathogenic BRCA mutation was previously detected, and to individuals in whom the probability of detecting a mutation was greater than approximately 10%, usually because of an affected family member. All DNA test applicants received an invitation for an in-person disclosure session when the results became available. If women eventually learned that they carry a BRCA1/2 mutation, they received detailed information about the personal implications, and about the advantages and disadvantages regarding the risk-management options, such as prophylactic surgery or surveillance of breasts and/or ovaries. During the study period, chemoprevention was not an available option.
BRCA mutation carriers were offered follow-up consultations with a multidisciplinary team involved in the Family Cancer Clinics of Nijmegen, Groningen, and Maastricht (the Netherlands), generally consisting of a medical oncologist, a gynecologist, and a surgeon. The participants were advised to have annual mammography screening and a breast examination by a physician, and to perform monthly breast self-examination. Occasionally, annual magnetic resonance imaging screening was offered as well.
Study Procedure
The first part of this article focuses on all women who learned that they carry a BRCA1 or BRCA2 mutation. It reports on questionnaire results at five different points in time: a baseline assessment 1 week after DNA testing (T1), 4 weeks after DNA testing (T2), 2 weeks after disclosure of a positive DNA test result (T3), 3 months after disclosure (T4), and 9 months after the positive DNA test result (T5). Only BRCA mutation carriers completed T3 through T5. The second part of the article also reports on the group of BRCA mutation carriers before and after disclosure, but predominantly focuses on predisclosure data (T2) from the overall sample of women who provided a blood sample for BRCA testing.
Data were collected in a longitudinal, randomized trial on shared decision making. The effects of providing decision aids have been described elsewhere.14,15 The study was approved by the local research ethics committees. All women who provided a blood sample for BRCA testing at one of the university hospitals in the eastern part of the Netherlands (Family Cancer Clinics of Nijmegen, Groningen, and Maastricht), and who were eligible from 1999 to 2001 were asked for informed consent. Women were excluded if they had insufficient knowledge of the Dutch language, had undergone bilateral mastectomy, were treated for breast cancer less than 1 month before, or had distant metastases. Women were randomly assigned to receive a first decision aid (DA) after T1 or after T2. This DA consisted of a brochure and video with information on screening and prophylactic surgery. After T3, BRCA mutation carriers were randomly assigned to either regular care or a second DA. This DA consisted of a value assessment and individualized treatment information.
Measures
Sociodemographics and medical history.
Information was obtained with regard to sociodemographics (age, marital status, education level, and presence and age of children), and medical history (personal history of breast cancer, time since diagnosis and breast cancer treatment if applicable [ie, breast-conserving therapy or mastectomy], having had biopsies, and having undergone prophylactic mastectomy).
Risk-management preference.
In all questionnaires, women indicated their intended breast cancer risk-management choice if they were found to be a BRCA mutation carrier. The options were prophylactic mastectomy, regular breast screening, and undecided.
Anxiety.
Anxiety was measured using the 20 items of the State Anxiety scale of the Spielberger State-Trait Anxiety Inventory.16 The reliability of the scale was good (Cronbach's α = .84).
Risk perception.
Women were asked about their breast cancer risk of being a BRCA mutation carrier (without having had prophylactic mastectomy) compared with the population risk of developing breast cancer. Response options ranged from 1 (much lower), to 4 (the same), and up to 7 (much higher).
Anticipated regret.
Women were asked to imagine having decided in favor of breast surveillance instead of prophylactic mastectomy. Subsequently, they rated how much regret they would feel about that decision if diagnosed with breast cancer several years later. Response options ranged from 1 (not at all) to 4 (very much).
Data Analysis
We used the SPSS 11.5 package (SPSS Inc, Chicago, IL) for data analysis. Frequencies (means, standard deviations), t tests, and χ2 analyses were used to describe the study population and to test whether women who tested positive differed from those who tested negative.
Part I: preference and actual uptake for BRCA mutation carriers.
To assess changes in preferences for either breast surveillance or prophylactic mastectomy among BRCA mutation carriers, we conducted Cochran Q test to detect longitudinal changes over time (T1 to T5). Additionally, we examined eventual changes over time by conducting McNemar tests for each possible combination of time points (T1 to T5; n = 10 combinations). We could not include the women who were undecided as a third category in these analyses because the number of women in this group was small.
Part II: associations with preference for undergoing prophylactic mastectomy or being undecided.
We first assessed univariate associations with risk-management preferences using logistic regression analysis. For this purpose, risk perception and anticipated regret were dichotomized using a mean-split, whereas anxiety, which had a large range, was divided into three categories. Because the study variables—hospital location and DA—were significantly related to risk-management preference, we controlled for these in the first step of all logistic regression analyses. Correlates with preference were examined using data from T2 variables for two reasons. First, the T2 questionnaire had been specially developed to assess risk-management decision-making; second, the number of respondents of the overall sample was sufficiently large to draw reliable conclusions from multivariate analyses. We investigated whether the same univariate associations were observed for the subgroup of BRCA mutation carriers at T2. In addition, we assessed which variables predicted risk-management preference 9 months after disclosure (T5). Within the overall group, a multivariate logistic regression analysis was applied, including all univariate associations with a significance level of P < .10.
All analyses focused on women who had a preference for surveillance or prophylactic mastectomy. Because we were also interested in women who were undecided at T2, univariate associations with being either decided or undecided were examined as well.
RESULTS
Participants
At the beginning of the study, 453 women were eligible and 390 gave informed consent (response rate, 86%). For 22 women, risk-management preference was not applicable because they had no breast tissue due to previous surgery. We excluded these women from the analyses. Furthermore, 32 women did not complete the T2 questionnaire or did not provide a valid response on the question about preference for risk management. Eighty of the 338 women who were eligible tested positive for a BRCA mutation. They completed follow-up postdisclosure questionnaires.
Participant Characteristics
Table 1 lists sociodemographic and medical characteristics of the study sample at T2. The women who eventually learned that they carried a BRCA mutation (n = 80) differed in some respects from the women who did not receive a positive BRCA test result (n = 258). With regard to psychological variables, including risk management preference, no differences were observed.
Women's Characteristics at T2 by a Subsequent Positive or Negative BRCA Test Result
Part I: Preference and Actual Uptake for BRCA Mutation Carriers
In Figure 1, preferences over time are displayed for all BRCA mutation carriers (77 of 80 women, with three missing) who provided valid responses to all five questionnaires (note that the preferences at T1 to T2 were conditional, whereas the preferences and uptake at T3 to T5 were reported after actually having received a positive BRCA test result). Before DNA testing at T1, 32% of the women (25 of 77) indicated that they intended to have a prophylactic mastectomy if they received a positive test result. Nine months after test result disclosure (T5), 40% of the women (31 of 77) indicated that they either intended to have prophylactic mastectomy (n = 17), or that they actually underwent surgery (n = 14). Among the 14 women who underwent prophylactic mastectomy, seven had a personal history of breast cancer. Among the 17 women who intended to have prophylactic mastectomy, 13 planned to opt for a prophylactic mastectomy within the next 2 years. The other four women planned to have the surgery within 2 to 10 years.
Preferences and actual uptake over time for complete sets of responses of women who learn that they carry a BRCA mutation (n = 77). pm, prophylactic mastectomy.
Of the women who preferred surveillance or prophylactic mastectomy, most had stable preferences over time (T1 to T5, Cochran Q = 5.94; P = .20). Moreover, there were no significant changes between any of the possible 10 combinations of time points T1, T2, T3, T4, and T5 (McNemar P ≥ .109). Of the 46 women who preferred screening at T1, 38 still preferred screening at T5. Of the 25 women who preferred prophylactic mastectomy at T1, 22 indicated the same preference at T5. Finally, we compared the DA groups on combinations of time points to check the potential influence of providing DAs on stability and did not find correlations (T1 to T5, Cochran Q; P > .07).
Part II: Associations With Preference for Undergoing Prophylactic Mastectomy or Being Undecided
Of the overall sample of women (n = 338) who completed our T2 questionnaire, 124 women preferred prophylactic mastectomy, 182 women opted for surveillance, and 32 women were undecided about their preference. To assess the correlations of preference for risk management, we excluded the women who were undecided. Two women did not complete the questionnaire measures for anxiety and anticipated regret.
Study Variables
The preference for prophylactic mastectomy varied among women from different clinics, with the number of women preferring prophylactic mastectomy being lower in one of the three hospitals (χ2 = 9.39; P = .009). Also, the first DA affected the women's preference. Of the women who received a brochure and a video about the risk-management options before T2, 47% preferred prophylactic mastectomy, whereas 35% of the women who received a DA after T2 preferred prophylactic mastectomy (χ2 = 4.83; P = .028). Therefore, all univariate (Table 2), and multivariate (Table 3) logistic regression analyses were controlled for these study variables.
Univariate Logistic Regression Analyses Regarding Preference for Undergoing Prophylactic Mastectomy
Multivariate Logistic Regression Analysis Regarding Preference at T2 for Undergoing Prophylactic Mastectomy Within the Overall Group of Women (n = 304)
Sociodemographic and Medical Variables
Women who had young children (ie, younger than age 13 years), were more likely to prefer prophylactic mastectomy in case of carriership. Also, women with a personal history of breast cancer were more likely to prefer prophylactic mastectomy (in the current sample, having a personal history of breast cancer is confounded with being the first test applicant in the family). Among the women with a history of breast cancer, those having undergone a unilateral radical mastectomy were more in favor of having a prophylactic mastectomy (for their contralateral breast) than women who had undergone a breast-conserving procedure (Table 2). For the remaining sociodemographic or medical variables, no differences were observed (ie, age, education, marital status, employment status, number of affected family members, [number of] breast biopsies for unaffected women, and time since cancer diagnosis for affected women).
Psychological Variables
Prophylactic mastectomy was preferred more frequently among women who rated their hypothetical breast cancer risk as a BRCA mutation carrier as being high. Furthermore, higher levels of anxiety were associated with a preference for prophylactic mastectomy. Finally, the amount of anticipated regret was strongly related to risk-management preference. Women who anticipated strong feelings of regret about not having had prophylactic mastectomy in the case of a breast cancer diagnosis were more likely to prefer prophylactic mastectomy (Table 2).
Multivariate Associations With Preference for Risk Management for the Overall Group
In the multivariate analysis, the level of anxiety was no longer significantly associated with preference (Table 3). If we included breast cancer treatment among women with a personal history of breast cancer, treatment remained a significant predictor (odds ratio = 4.13; P = .003).
BRCA Mutation Carriers
We observed a similar pattern of univariate associations for the subgroup of BRCA mutation carriers before disclosure (T2) and 9 months after disclosure (T5; Table 2). However, having a personal breast cancer history was not significant in these subgroup analyses. Women with a personal history of breast cancer did not prefer prophylactic mastectomy more often than unaffected women. At T2, 40% of the affected women opted for prophylactic mastectomy, whereas 31% of the unaffected women preferred prophylactic mastectomy (P = .5). At T5, 45% of the affected women opted for prophylactic mastectomy, whereas 40% of the unaffected women preferred prophylactic mastectomy (P = .68). Also, anxiety was no longer significantly related to a preference in the subgroup of BRCA mutation carriers (T2, P = .067; T5, P = .27).
Undecided Status
We assessed whether the 32 women who were undecided at T2 were different from the women who, at T2, had already indicated a risk-management preference for either surveillance or prophylactic mastectomy. We first inspected the study variables. The first DA tended to help women in becoming somewhat more decided about their risk-management preference (χ2 = 3.58; P = .058), whereas hospital location was not associated with being undecided (χ2 = .18; P = .91). If we controlled for having received the first DA, women who were undecided perceived the breast cancer risk of BRCA mutation carriers as being somewhat higher than women who were decided already (odds ratio = 2.45; P = .044). No additional differences were observed. With regard to the subgroup of BRCA mutation carriers, we observed that women who were undecided at T1, T2, T3, and/or T4 made similar choices at T5 when compared with women who were decided all along (χ2 = .04; P = .84).
DISCUSSION
It has been suggested that women seem to refrain from choosing prophylactic mastectomy if they actually face the decision after a positive BRCA test result.10 In that study, only 3% of the unaffected carriers obtained prophylactic mastectomy during a 1-year follow-up period. In contrast, our data do not suggest that the disclosure of a BRCA mutation greatly influences decision making. Most women have already formed their personal risk-management preference before DNA test disclosure, and hold on to that preference after a positive test result.
We also examined which variables were associated with a preference for prophylactic mastectomy. Remarkably, risk-management preference was most strongly correlated with anticipated feelings of regret; that is, the amount of regret women think they would have if they were diagnosed with breast cancer after rejecting the option of prophylactic mastectomy. Women with strong feelings of anticipated regret were more inclined toward prophylactic mastectomy than women who expected to have less intense feelings of regret. Recent reports have indeed suggested that anticipated negative emotions may have an unrecognized and profound impact on decision making in other domains, such as smoking initiation17 and purchase of lottery tickets.18
Several limitations should be noted. First, most women had stable preferences for prophylactic mastectomy. This obviously does not mean that such preferences will fully translate into actual action. Women who have not yet undergone surgery may change their minds. In addition, our study sample of BRCA mutation carriers was small, which diminished the ability to detect actual changes. Another limitation of our study is that it was designed to evaluate DAs. By actually studying decision making and the effects of DAs, we inevitably influenced the decision-making process. For example, after receiving the first DA, women more often preferred prophylactic mastectomy. Furthermore, our conclusions regarding the associations with risk-management preferences are mainly based on multivariate, pretest-disclosure analyses within the overall group of DNA test applicants. However, it is likely that these findings can be generalized to the group of BRCA mutation carriers who actually decide, because most women had stable preferences, and we also observed similar associations within the group of BRCA mutation carriers (however, a personal history did not seem to be associated with a preference for prophylactic mastectomy among BRCA mutation carriers). Yet another limitation is that the above-described associations were found for preferences instead of actual action. Future studies with larger samples, longer follow-up time, and well-validated measures are needed to confirm whether the actual uptake of prophylactic mastectomy is associated with the same set of variables.
The results from this study may add to our understanding of the decision-making process regarding prophylactic mastectomy. Whereas many women may be comfortable enough with screening, others may choose the safest option to avoid potential future self-blame and anxiety. For others, an important part of this anxiety is the fear of leaving young children motherless.19 This study suggests that women who prefer prophylactic mastectomy may be very responsive to regret anticipation in case of a bad outcome. This may occur because it is common practice to leave the decision to the patient, thus triggering a sense of personal responsibility for future bad outcomes.20,21 Put differently, our data suggest that women who opt for prophylactic mastectomy feel bound to do all that is possible now to prevent future harm. This may be considered an understandable and rational reason to choose prophylactic mastectomy. However, the impact of anticipated regret may be interpreted as a maladaptive coping strategy to cope with the present uncertainty. We call it maladaptive because—in decisions that carry a heavy emotional burden—people act more rigidly and tend to neglect trade-offs between costs and benefits.22
Health care professionals should pay special attention to the possibility that women may simply neglect adverse consequences of prophylactic mastectomy to cope with the cancer threat. In addition, health care professionals should be aware that most women seem to decide at a relatively early stage of genetic counseling about their future risk management. Therefore, a careful discussion about risk-management preferences and the way women cope with their cancer risk should not be postponed until DNA test disclosure.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
AUTHOR CONTRIBUTIONS
Conception and design: Sandra van Dijk, Wilma Otten, Peep F.M. Stalmeier
Financial support: Peep F.M. Stalmeier
Collection and assembly of data: Mariëlle S. van Roosmalen, Peep F.M. Stalmeier
Data analysis and interpretation: Sandra van Dijk, Wilma Otten, Peep F.M. Stalmeier
Manuscript writing: Sandra van Dijk, Mariëlle S. van Roosmalen, Wilma Otten, Peep F.M. Stalmeier
Final approval of manuscript: Sandra van Dijk, Mariëlle S. van Roosmalen, Wilma Otten, Peep F.M. Stalmeier
Acknowledgments
We thank the women who participated in this study, and research assistants M. Oude Elberink and I. Bakker. We acknowledge the support of the participating Family Cancer Clinics, led by J.C. Oosterwijk in Groningen, N. Hoogerbrugge in Nijmegen, and U. Moog in Maastricht, the Netherlands.
Footnotes
-
Supported by Dutch Cancer Society Grant No. 98-1585.
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
- Received December 22, 2006.
- Accepted January 23, 2008.
