Decision Aids From Genetics to Treatment of Breast Cancer: Title and subTitle BreakLong-term Clinical Utility or Temporary Solution?

Medical science is a victim of its own success. Clinical and translational investigation has resulted in a bounty of relevant and specialized knowledge that is adopted into routine clinical use. While this 21st-century science is hurtling quickly into the clinical arena, the relative growth of clinicians in some specialized areas, who are facile in using 21st-century scientific knowledge to practice 21st-century medicine, has not increased at a rate to keep pace with that of recent scientific advances. Compounded with the growing trend of viewing patients as partners in medical decision making, necessitating educating and helping patients with choices in a field that changes weekly, the medical community is facing a crisis. Two articles published in this issue of JAMA try to address this problem by exploring the utility and limitations of patient decision aids in breast cancer genetic risk assessment¹ and breast cancer surgery.²

Aptly, this year marks the 10th anniversary of the identification of BRCA1, a major susceptibility gene for hereditary breast and/or ovarian cancer syndrome.³ The first susceptibility gene for an inherited cancer syndrome was isolated approximately 20 years ago⁴ and ushered an explosion of knowledge about cancer predisposition. Complex algorithms have been proposed to help clinicians decide when to refer patients to cancer genetics professionals and to help health care professionals accurately assess cancer risk and offer genetic testing when appropriate.⁵ Because genetic testing is widely available in the United States, the exponential demand for cancer genetics professionals (both cancer genetics counselors and physician clinical cancer geneticists) has already outpaced supply. Unlike traditional clinical subspecialties, a typical cancer genetics consultation requires 2 to 3 hours of face-to-face time. There are approximately 400 cancer genetic counselors in the United States (Bea Leopold, MA, National Society of Genetic Counselors, written communication, July 1, 2004), perhaps a fraction of whom have expertise in cancer genetic counseling. The news is even more bleak for their physician counterparts: there remains a paucity of formally trained clinical cancer geneticists. Even if training for such cancer genetics professionals started en masse today, it would still take at least a decade to feel the impact. Therefore, something "out of the box" needs to be done now.

In this issue of JAMA, Green and colleagues¹ report the results of a randomized controlled trial evaluating the efficacy of alternative educational methods to the face-to-face education that is currently part of traditional genetic counseling. The investigators enrolled 211 women with personal or family history of breast cancer from outpatient clinics offering cancer genetic counseling at 6 US-based medical centers. These patients were randomly assigned to receive one-on-one genetic education as part of traditional genetic counseling (counselor group, n = 105) or education by a computer program (computer group, n = 106) followed by genetic counseling. Outcome measures were those that test factual knowledge, those that assess decision making and its related parameters, such as risk perception and intent to have genetic testing, and those that measure emotional reactions (eg, anxiety, conflict, satisfaction). All outcome measures were stratified by high vs low risk of carrying a germline BRCA1 or BRCA2 mutation.

Knowledge scores increased significantly for both groups (P<.001), irrespective of risk status, although the increase in knowledge in the computer group was slightly greater than that in the counselor group (P = .03) for low-risk women. The intent to undergo BRCA1/2 testing decreased in both groups in low-risk but not high-risk women. The counselor group had lower scores on decisional conflict and anxiety assessments and higher satisfaction scores. From their observations, the authors conclude that an interactive computer system was more effective than standard genetic counseling for increasing knowledge of breast cancer and genetic testing among women at low risk of carrying a mutation but that genetic counseling was more effective in allaying anxieties and facilitating accurate risk perceptions. Thus, they suggest that such a computer program could serve as an educational tool for low-risk women or could be a useful adjunct to genetic counseling for those at high risk.

In the context of workforce shortage in cancer genetics counseling, it would appear that Green and colleagues have provided a "right answer." However, are their conclusions valid? Is the study well controlled and is the design well thought out? For example, as pointed out by the authors, one explanation for why the counselor group may have had lower anxiety scores and higher satisfaction scores is that the genetic counselors but not the computer provided individualized cancer risk estimates for the probability of carrying a BRCA1/2 mutation and of developing breast cancer. The counselors also provided psychosocial support to address emotional concerns if the latter arose during a counseling session. Furthermore, despite that the authors carefully tried to ensure uniformity among counselors and counseling sessions, this is virtually impossible to control, as would be the case when individuals from different centers are involved in a relatively new clinical field. Despite these and other caveats, the findings of Green and colleagues should stimulate further studies in patient decision making and the controlled search for alternative nontraditional adjuncts to cancer risk assessment and counseling in the face of the existing workforce.

Assuming that the clinical cancer genetics community embraces the suggestion of Green and colleagues that low-risk women can use the computer system alone as an educational intervention, an important issue is how to identify women at low risk even before an encounter with cancer genetics professionals. One possible way is to enlist the aid of primary caregivers and provide education for these clinicians, perhaps with an interactive computer adjunct for cancer risk assessment.⁶ Even with such an interactive tool, communicating risk to a patient remains an issue.^{7 - 8} For high-risk women, Green et al suggest using an interactive computer system as an educational adjunct together with genetic counseling. Given the current limited workforce supply in cancer genetics, could more and more of the education piece be placed in the computer? The answer is a qualified yes; increasing factual knowledge would facilitate this—ie, systematic data linking genotype to the prediction of cancer development need to be obtained for each heritable cancer syndrome. In addition, decision-making studies and high-risk communication studies in the context of molecular data should be encouraged. In the meantime, however, a concerted effort to train more genetic counselors who can provide cancer counseling and to train physicians formally in the subspecialty of clinical cancer genetics must occur.

In a second article in this issue of JAMA, Whelan and colleagues² highlight many of the same issues that confront clinicians and counselors providing information about genetic risk that also are faced by surgeons guiding patients through the difficult decisions surrounding local therapy of breast cancer. Indeed, many strategies to improve communication are shared. Whelan et al set out to provide patient educational material for women with early-stage breast cancer who were deciding between mastectomy and lumpectomy (breast-conserving surgery) and to make this decision aid useful in the offices of community surgeons. The McMaster University group constructed a decision board, basically a small flip chart covering 4 aspects of the initial surgery: treatment choices, adverse effects, functional and cosmetic results, and survival. The decision board was tested in Ontario community surgical practices and was well accepted by patients and their surgeons.⁹ In the current report, a randomized, prospective evaluation was undertaken among 20 surgical practices involving 201 patients, dividing surgical practices into those using usual preoperative counseling and surgeons using the decision board.

Immediately following counseling, at 6 months, and at 12 months, patients were queried about their choices of therapy, conflict about their decision, satisfaction, anxiety, and depression. More patients receiving the decision board chose breast-conserving procedures over mastectomy (94% vs 76%) and initially felt less conflicted about their decision. However, at 6 and 12 months after therapy, measures of decisional conflict and satisfaction were similar between the 2 groups. Scores of anxiety and depression were not different.

Community surgeons managing diverse diseases with shrinking resources have inadequate time and resources for thorough counseling prior to surgery for early-stage breast cancer. Keating et al¹⁰ documented a discrepancy between patients and their surgeons regarding whether both lumpectomy and mastectomy were presented and discussed. Satisfaction with treatment decisions appears to be higher among women who think they were offered options and participated in those decisions.¹¹ Decision aids and educational materials may extend the effective time physicians spend counseling women about surgical options. The materials may provide important factual information that may be omitted in brief clinical encounters. Because overall survival appears equivalent whether the patient receives a mastectomy or a lumpectomy,¹² the most basic goal of preoperative education is a realistic estimate of the risks of recurrence.

In common with women concerned about genetic predisposition to breast or ovarian cancer, overestimation of actual risks is a barrier to sensible decision making for women with early-stage breast cancer. Women who are worried about breast cancer risk may have intrusive thoughts and may practice excessive breast self-examination¹³ or have unwarranted anxiety about mammography.¹⁴ Women with ductal carcinoma in situ perceive a high risk of recurrence and metastasis, which may influence their decision making.¹⁵ Women seeking genetic counseling for hereditary breast cancer overestimate their chances of carrying a susceptibility gene.¹⁶ Educational aids may help. However, educational materials are proliferating on the Internet and in print form without evaluation and review.¹⁷ The scrutiny given to the interactive computer aid for genetic counseling and the decision board for surgical counseling is rare and raises the standard for educational materials. Indeed, would the scrutiny required for routine clinical decision aids, such as those used by Green and colleagues and Whelan and colleagues, overwhelm those "in the know," thus shifting the workforce shortage proximally? After all, there are many inherited cancer syndromes beyond hereditary breast and/or ovarian cancer syndrome, and there are many surgical scenarios requiring multiple decision boards. And if medical knowledge is exponentially increasing, who will ensure that all these well-scrutinized decision aids are kept up to date?

Decision aids are useful only if they accurately reflect current and changing information and clinical practice. Breast cancer, whether sporadic or heritable, is remarkable for its heterogeneous biology and behavior, which may be unpredictable. Clinical experience and modern molecular-based classification allow for individualized treatment of an increasing number of patients with breast cancer. For instance, many risk factors influence local recurrence after lumpectomy, which may include an extensive in situ component, grade, and patient age.¹⁸ Treatment is also changing. Preoperative chemotherapy after core needle biopsy may be used for breast cancers previously treated with surgery first. Large clinical trials of breast conservation vs mastectomy, undertaken in decades past, changed clinical practice and were the basis of the decision board. Similarly, even several years after the discovery of BRCA1 and BRCA2, mutation analysis for these genes offered to those in cancer genetics clinics were "on a research basis." However, as a result of new data showing that medical management could be drastically changed, such genetic testing slowly became accepted as part of the routine clinical armamentarium.

New results and a changing art of practice continue to provide a moving target. Educational aids are helpful, but an experienced clinician, in touch with changing concepts and data, remains indispensable for integrating and explaining tests and treatments offered to patients. Medicine, even 21st-century genomic medicine, remains very much an art as well as a science.