Patients’ perceptions of gene expression profiling in breast cancer treatment decisions

results identify a need to better support patient un derstanding of the test and its limitations. Findings illustrate the deep emotional investment patients make in gep test results and the impact of that invest ment on their treatment decisions. of distant disease recurrence and predictive of the benefit of adjuvant chemo therapy for patients with estrogen receptor–positive disease 6–10 . The resultant recurrence scores classify patients into groups with poor or good prognosis (likelihood of recurrence outside the breast in the next years). Those with low recurrence scores ABSTRACT Determining the likely benefit of adjuvant chemother apy for early-stage breast cancer patients depends on estimating baseline recurrence risk. Gene expression profile ( gep ) testing of tumours informs risk predic-tion, but evidence of its clinical utility is limited. We explored patient perceptions of gep testing and the impact of those perceptions on chemotherapy decisions. Methods We conducted group ( and individual interviews ( n with patients who used gep recruited through at hospitals in Ontario. Data were analyzed using content analysis and constant comparison techniques.


INTRODUCTION
Current guidelines for the management of early-stage breast cancer result in thousands of women receiving chemotherapy without benefit 1 . Clinical guidelines suggest that 90% of the 22,600 Canadian patients with human epidermal growth factor receptor 2negative, node-negative, hormone receptor-positive disease should be offered adjuvant chemotherapy [1][2][3][4] . Yet it is estimated that only 15% of such cancers will recur, suggesting that about 8500 Canadian patients will be treated without benefit each year [2][3][4][5] .
Distinguishing patients who likely will benefit from those who might not depends on baseline recurrence risk. Traditionally, recurrence risk assessments are based on several prognostic factors, including tumour size, tumour grade, nodal involvement, and hormone receptor status 2,3,5 . Gene expression profiling (gep) tests measure the expression levels of a set of prognostically relevant genes to predict the likelihood of distant recurrence at 10 years and the likely benefit from chemotherapy. In retrospective studies, several gep tests have been validated as prognostic of distant disease recurrence and predictive of the benefit of adjuvant chemotherapy for patients with estrogen receptor-positive disease [6][7][8][9][10] . The resultant recurrence scores classify patients into groups with poor or good prognosis (likelihood of recurrence outside the breast in the next 10 years). Those with low recurrence scores

Introduction
Determining the likely benefit of adjuvant chemotherapy for early-stage breast cancer patients depends on estimating baseline recurrence risk. Gene expression profile (gep) testing of tumours informs risk prediction, but evidence of its clinical utility is limited. We explored patient perceptions of gep testing and the impact of those perceptions on chemotherapy decisions.

Methods
We conducted one focus group (n = 4) and individual interviews (n = 24) with patients who used gep testing, recruited through clinics at two hospitals in Ontario. Data were analyzed using content analysis and constant comparison techniques.

Results
Patients' understanding of gep testing was variable, and misapprehensions were common. Patients valued the test because it provided them with certainty amidst confusion, with options and a sense of empowerment, and with personalized, authoritative information.
They commonly believed that the test was better and fundamentally different from other clinical tests, attributing to it unique power and truth-value. This kind of "magical thinking" was derived from an amplified perception of the test's validity and patients' need for reassurance about their treatment choices. Despite misperceptions or magical thinking, gep was widely considered to be the deciding factor in treatment decisions.

Conclusions
Patients tend to overestimate the truth-value of gep testing based on misperceptions of its validity. Our have a low likelihood of recurrence and will likely derive little or no benefit from chemotherapy; those with high risk recurrence scores have a higher likelihood of recurrence and will likely derive high benefit from adjuvant chemotherapy. The likelihood of benefiting from chemotherapy among those receiving intermediate recurrence risk scores is unknown and is the subject of the tailorx study 11 .
Thus, gep testing might further inform baseline risk prediction, potentially reducing unnecessary treatment, exposure to toxicity, and considerable health care costs 12 . Health technology assessment agencies in the United Kingdom, the United States, and Canada have evaluated gep tests [13][14][15] . Although the assessments have raised questions about the predictive and prognostic validity of gep tests 13,14,16 , the tests have been recommended for clinical practice to complement conventional risk-of-recurrence assessments and are increasingly being integrated into clinical practice in the United States and Canada 3,5,17 .
Despite the diffusion of gep tests into clinical practice, relatively little is known about their impact on chemotherapy decisions. Several studies have demonstrated that gep results are an important factor in the treatment recommendations made by oncologists [18][19][20][21] . Studies also indicate that patients are interested in receiving gep tests and that, in surveys posing hypothetical questions, gep results would affect patient treatment preferences to a greater extent than do standard recurrence measures. However, patients' understanding of the results varies depending on their knowledge and health literacy scores [22][23][24][25][26][27][28] . Yet, in this era of shared decision-making, no study has addressed patient perceptions and experiences of gep testing in informing actual treatment decisions in breast cancer or cancer contexts. Moreover, there is a paucity of evidence on why and how the tests inform treatment decisions made by patients, taking into account the range of clinical and contextual factors at play during the decision-making process. The aim of the present study was to gain insight into patient perceptions of the value of gep testing and the perceived impact of the results on chemotherapy treatment decisions.

Study Design
This qualitative study is part of a larger, mixedmethods study to examine the value of the gep test for breast cancer patients and medical oncologists. The first phase was the present qualitative study, which used a focus group and interviews. In the second phase, a quantitative survey that included a discrete choice experiment to estimate the utility of the gep test relative to other factors was completed by a nationally representative sample of oncologists, women with a history of breast cancer, and adult women from the public, with survey sampling performed using an online survey panel (manuscripts forthcoming). The research ethics boards at St. Joseph's Hospital, Sunnybrook Health Sciences Centre, and Princess Margaret Hospital approved the study. All participants provided informed consent before taking part in this study.

Sample Recruitment
We recruited early-stage (stage i-ii) breast cancer patients who had completed surgical treatment and used gep testing in their chemotherapy decisions. We targeted patients who were offered the gep test after initiation of public funding in Ontario in March 2010. Participants in the tailorx trial 11 were ineligible, because that trial began before initiation of public funding for gep testing, creating inherent differences in how patients accessed and used the gep test.
Eligible patients who were scheduled for routine follow-up visits were identified from clinical records by participating oncologists and designated site coordinators at two academic hospitals in Toronto, Ontario. Coordinators attended follow-up clinics and offered information about the study to eligible patients. Interested patients were invited to contact the qualitative researcher to discuss the study, arrange participation, and give consent. The qualitative researcher contacted eligible patients who indicated interest but who did not call the qualitative researcher within a few weeks. At one of the clinics, the study was introduced to patients by their oncologist at their follow-up visit, and the site coordinator then provided further information if the patient was interested in participating in the study.

Data Collection
We conducted a focus group and interviews to accommodate patient schedules and to encourage maximum participation. A semistructured discussion guide developed through literature review and clinical consultation elicited details of patient experiences of decision-making about treatment and the role that the gep test played in that process (Appendix a). Areas of discussion included feelings previously held about chemotherapy, treatment discussions with the oncologist and other health care providers, attitudes toward and understanding of the gep test, the treatment decision and factors influencing that decision, and impact of the gep test on the decision-making process. Demographic data were collected in questionnaires administered before the interviews and the focus group.

Data Analysis
All qualitative data (focus group and interviews) were digitally audio-recorded for verbatim transcription.
All corrected transcripts were then merged into a single dataset and entered into a software program to code for both anticipated and emergent themes. Qualitative data were analyzed using content analysis and constant comparison techniques, which involved collecting data until saturation of themes was reached and included searches for disconfirming evidence 29 . Data were validated through peer-debriefing, during which developing themes and analyses were identified and discussed with the study team.

Patient Demographics
A total of 28 patients participated in the study from 2010-2011 (focus group: n = 4; telephone interviews: n = 24). Most of the patients in our sample were married (61%), had children (67%), were highly educated (79%), and did not undergo chemotherapy (68% ; Table I). In addition, many participants appeared to have been overwhelmed by the results, providing confused accounts of numbers, charts, and graphs showing recurrence risks that they found difficult to interpret. There were also misapprehensions about the laboratory analyses involved. For example, one woman believed that the test involved injecting tumour samples with chemotherapy drugs to determine response:

Patient Understanding and Misunderstanding of GEP
My understanding is they actually take a sample of the tumour ... and they run it against the actual chemo to see how it responds.
-Patient 17 in interview

The Value of GEP in Treatment Decisions
Patients described emotionally and socially complex reasons why they valued gep testing in making their treatment decisions. Those reasons were often shaped by pre-existing beliefs and expectations of chemotherapy, which derived from sources such as prior experience of caring for a friend or relative with cancer, media representations of cancer treatment, and Internet or book-based research. Some held negative views of chemotherapy and anticipated undesirable effects on their work, lifestyle, and family. Others were open to taking chemotherapy because they wanted to feel that they had done everything possible to fight their cancer. Expectations about chemotherapy were also informed by the rapport with providers, often in relation to the willingness of the patient's oncologist to be consultative and the degree of confidence instilled as a result. It was against this backdrop of pre-existing beliefs about chemotherapy and interactions with providers that patients valued the gep test, because it provided them with certainty amidst confusion, with options, and a sense of empowerment, and with personalized, authoritative information.

Certainty Amidst Confusion
Many patients described discomfort in making decisions based on the opinions of their oncologistdespite considering those providers to be medically competent-because they had received conflicting opinions from other clinicians responsible for their care. For women in that position, gep test results were perceived as providing a more stable platform from which to act and a degree of clarity that had otherwise been absent: It's hard because you get different people explaining their area to you. It's so specialized, you know? ... Well, let's just say it didn't paint a very clear picture ... and that's why the genome testing is so important. -Patient 10 interview

Options and a Sense of Empowerment
In situations in which rapport between the patient and oncologist was poor, or in which the patient had lost confidence in her oncologist, the test was experienced as compensating for the failure of the relationship. For example, one woman contrasted the doctrinaire manner of a first oncologist with the approach of a second who used the test as part of a shared decision-making process in which a number of options were considered: [The gep test] also opened up a dialogue.... The first oncologist I talked to was just very cut-and-dry, like, ... "You have no options here." And so to deal with what you're dealing with, plus the fact that you have no options, no voice ... It was just overwhelming.... The two experiences were in stark contrast in terms of one oncologist saying, "Yeah, there are no other sources of information, you have no other options, you have to do this," versus, "Yeah, we do have options, we can do no chemo or we could do four rounds or we could do six rounds ... Let's gather our resources so we can figure out the right one for you." -Patient 14 interview As the preceding example illustrates, the gep test brokered a dialogue for some women where previously none had existed. As a consequence, it was perceived as providing options at a difficult moment in their lives, allowing them to feel like active agents rather than passive recipients of care, regardless of whether it changed their treatment decision.

Personalized, Authoritative Information
Patients also valued the gep test because they believed it provided them with more personalized, authoritative information than other clinical tests and medical opinions. Many expressed discomfort with the notion that treatment decisions based on traditional risk stratification markers (tumour size, tumour grade, nodal involvement, and hormone receptor status) would be protocol-driven or statistically-derived. They believed that the gep test reflected their unique circumstances, detached from statistics: It meant that I wasn't just going to be lumped in with statistics.... I certainly felt that with this kind of tailored or personalized kind of investigation ... "guesswork" was going to disappear. -Patient 9 interview Moreover, patients did not generally understand that the test is based on population-based estimates derived from sample cohort data. They believed that information unique to them was being returned, in contrast with the statistical estimates patients receive based on the pathology assessment of their tumours (size, grade, nodal involvement, and hormone recep- In addition, the test was valued for providing what was perceived to be higher-calibre information. The feeling that the gep test was state-of-theart, more scientific, and therefore more reliable was commonly expressed: He delineated them as fantastically precise measurement. They showed that there was absolutely no, no reason, a zero factor, to participate in chemo or radiation.... I said, "There couldn't have been a better gift in my life than that...." It was the greatest gift I'd ever been given.

-Patient 26 interview
This amplified perception of gep's scientific accuracy characterized a type of emotionally-invested "magical thinking" that underpinned perceptions of the test.

Magical Thinking
Perceptions of the gep test by patients were linked to the value and the symbolic importance that some patients ascribed to it. In this sense, a type of "magical thinking" underpinned their perceptions of the test, which was founded on a belief that gep testing had unique scientific power and, therefore, truth-value. The presumed validity of the results was a key feature of the test for many patients, with very few questioning or even discussing the test's potential limitations with their oncologist. When prompted to think about why they hadn't considered the possible limitations of the test, several participants identified emotional reasons for not doing so: I didn't ask, and it's something I would normally ask.... I think probably I just wanted to take it at face value. I don't know.... I didn't think about it 'til just now, but when you said that, I thought, "Well, why didn't I ask?" I don't know why I didn't. I think I was so happy about what I was seeing on that paper, and hearing from her, that I didn't question it. The suggestion that there might be cause to question the validity of the test led one woman to become angry. Like others, she admitted (in this case, unwittingly) that her emotional investment in the test had precluded any such questioning: The terminology here is very scientific.... It tells you how many patients that this research is based on, in what situation.... You even get graphs.... It's statistically significant information; nothing here is fluff the way I see it.... I would not even want anybody to question that, because for me this test is my reassurance that I did the right thing. I think it's bogus that people would question the validity of testing 21 genes. -Patient 10 interview In one exceptional case, a patient did conduct more detailed, independent research and discovered that the test was limited in ways that proved disappointing: The recurrence score was based on people who had larger tumours, or a higher grade, or had already taken the tamoxifen ... which I kind of didn't realize at the time-that it was a very limited population that they got their results from. -Patient 7 interview For some, an additional factor bolstering the presumed validity of the test was the public funding for it: I had no idea there was even another world out there that wouldn't be supporting the test.... The fact that's covered tells me that it's absolutely supported, right? -Patient 4 in focus group 1 A high emotional investment was placed in the test. The test was perceived as more emotionally significant than other conventional risk stratification marker information (tumour size, tumour grade, nodal involvement, and hormone receptor status) that they had received. The investment is reflected in the structure of narratives that frequently position the test as a beacon of hope in the darkest hour: Anything that happened after that [diagnosis] was like pulling me out of a pit, [and]  For these patients, the turnaround heightened the importance that they had attached to the test because of the sense of having been "rescued" from additional hardship.

The Deciding Factor in Treatment Decisions
Some patients described the test as one of many factors considered in their decision, but the indication most widely shared was that the test results had been the deciding factor, over and above all other considerations. This was true for patients who were initially reluctant to undergo chemotherapy, but who were persuaded of its importance because of the test results, and for patients facing the opposite scenario of wanting reassurance that foregoing chemotherapy would be justified in their case despite the popular notion that it is the standard of care: In those cases, patients generally interpreted the results to mean whatever they wanted them to mean, aligning them with their pre-existing treatment preferences. That finding might provide a further indication of emotionally-informed magical thinking.

DISCUSSION
As one of the first qualitative studies of patients' perceptions of gep testing, this work presents novel insights into how and why patients valued the gep test in their treatment decisions. Patients often viewed their gep results as providing information that was more scientifically valid, uniquely personalized, and emotionally significant than any other information they had received. In many of the patient narratives, the test figured as a transformational element that empowered them, allowed them to feel confident in their decisions, and in many cases, rescued them from unnecessary chemotherapy.
Participants understood that the test would indicate whether chemotherapy would or would not be beneficial in their care, but they generally did not understand that their results were founded on a population-based derivative. Furthermore, the suggestion that the test should be viewed as critically as any other technology was very uncomfortable for them to consider. Some acknowledged that they had willingly suspended critique because they needed something to provide a sense of certainty in the midst of such a difficult situation.
Our study has several limitations. The participant group comprised mainly highly educated women, all of whom used gep testing and who were drawn from two tertiary centres. The findings might also reflect a degree of self-selection, in that some participants had seen their oncologist's original recommendation overturned by the test results, which might have provided motivation to share their story. Also, the subgroup of participants receiving intermediate results did not include anyone who experienced ongoing indecision after receiving their results.
This study did not interview matched sets of patients and oncologists. Nor was it designed to test patient comprehension or to explore provider efforts to explain the gep test. For those reasons, it is not possible to describe a direct causal relationship between oncologists' explanation of gep testing and what patients understood. However, our results are consistent with the literature demonstrating variability in how patients understand the tests and the interest of patients in using gep tests to inform treatment decisions [22][23][24][25][26][27][28] . Nonetheless, no study has addressed patient perceptions and experiences of gep testing in informing actual treatment decisions. Our study thus provides novel insights into how and why patients valued the test: it was seen as providing them with certainty amidst confusion, with options and a sense of empowerment, and with personalized, authoritative information. Ultimately, gep was widely considered to be the deciding factor in treatment decisions, despite the misperceptions or magical thinking the test engendered.

CONCLUSIONS
Our results provide insight into an important paradox: patients tend to overestimate the truth-value of gep testing based on misperceptions of its validity. These results identify a need for communication or decision aids to support patients' understanding of the test and its limitations. Finally, our results might also help to support provider awareness of the ways in which patients can be emotionally invested in their gep results and the impact of those investments on treatment decisions.

ACKNOWLEDGMENTS
We thank the patients who participated in this study for generously sharing their experiences with us. This work was conducted with the support of funding provided by Cancer Care Ontario and the Ontario Institute for Cancer Research (through funding provided by the Ministry of Health and Long-Term Care and the Ministry of Research and Innovation of the Government of Ontario).

CONFLICT OF INTEREST DISCLOSURES
YB was supported by a fellowship from the Canadian Institutes of Health Research. DAM is supported by a Canada Research Chair in Health Services and Systems Research. The remaining authors have no financial conflicts of interest to disclose.