Illustrating Cancer Risk: Patient Risk Communication Preferences and Interest regarding a Novel BRCA1/2 Genetic Risk Modifier Test

Women with pathogenic variants of BRCA1/2 (i.e., BRCA1/2 carriers) are at increased risk for hereditary breast and ovarian cancer compared to the general population. However, not all BRCA1/2 carriers will necessarily develop cancer over a lifetime (incomplete penetrance), and some will develop different types of cancer (variable expressivity). These variations in penetrance and expressivity are likely due to genetic and environmental factors. Incomplete penetrance results in wide-ranging cancer risk estimates; the average breast cancer risk estimates for BRCA1/2 carriers by age 70 range from 44 to 78% and from 31 to 56%, respectively [1], although breast cancer risks may be as high as 85% for carriers with especially strong family histories [2].

Despite this substantial uncertainty, female BRCA1/2 carriers are asked to consider risk management strategies with varying benefits, harms, and degrees of protection, including enhanced screening, chemoprevention, and prophylactic surgery [3, 4]. As highlighted by recent media attention (e.g., “Angelina Jolie effect” [5-7]), the decision regarding prophylactic mastectomy is often most challenging. Although prophylactic mastectomy substantially reduces a BRCA1/2 carrier’s breast cancer risk [8], it is an invasive and irreversible strategy with physical and psychological ramifications. Thus, it is recommended that BRCA1/2 carriers treat adoption of prophylactic mastectomy as a preference-sensitive decision [4]. Many carriers have difficulty making a prophylactic mastectomy decision [9-15], in part because they must evaluate the benefits and harms of surgery with an incomplete understanding of their future breast cancer risk.

Recent research has demonstrated that multiple common genetic variants (i.e., single nucleotide polymorphisms) modify the penetrance of BRCA1/2 mutations [16-19] and also contribute independently to breast cancer risk. These genetic risk modifiers each have a small quantitative effect on the subsequent risk of breast cancer and could be used to generate a polygenic risk score reflecting a refined estimate of a BRCA1/2 mutation carrier’s future cancer susceptibility [20]. Existing data strongly suggest that these genetic risk modifiers can provide information to meaningfully discriminate between higher- and lower-risk BRCA1/2 carriers in a manner that may assist clinical decision-making [21, 22]. Thus, the potential exists for incorporating genetic risk modifiers into a clinical genetic testing panel (i.e., genetic risk modifier testing [GRMT]) to more accurately predict breast cancer risks among BRCA1/2 carriers.

GRMT may have substantial clinical utility because it could reduce some of the uncertainty surrounding cancer risk estimates for women with BRCA1/2 mutations and aid their decision-making regarding prophylactic mastectomy versus other less invasive risk management strategies. However, for women to benefit from GRMT, effective approaches for communicating this novel risk information are needed. Information about GRMT is complex and ambiguous in nature because it requires an understanding of gene-gene interactions, is supported by a small but growing body of scientific evidence, and the revised risk estimates remain probabilistic in nature. Recommended best practices in risk communication suggest that various strategies may be effective in promoting comprehension of this novel genetic risk information, including the use of numerical expressions such as percentages and frequencies, or visual presentations such as risk ladders, icon arrays, bar charts, and survival curves [23, 24].

Basic cognitive psychological research on information processing demonstrates that images are especially effective at drawing attention to messages, facilitating understanding of health information, and increasing recall, suggesting particular value in pairing written, numerical risk information with graphical presentations [23, 25-29]. In addition, using a visual aid to communicate risk can benefit vulnerable patient subgroups such as older adults, immigrant populations, and patients with low health literacy [23]. However, the effectiveness and patient acceptability of different graphical risk communication formats varies across studies and health contexts. For example, in one qualitative study evaluating strategies to communicate absolute lifetime risk across various health conditions including breast cancer, simple bar charts were found to be preferable over other images such as line graphs, thermometer graphs, icon arrays, and survival curves [30]. In another qualitative study evaluating preferences for communicating genomic lifetime risk for melanoma, participants expressed the strongest preference for pie charts, followed by icon arrays, bar charts, scale diagrams, and box plots [31]. When the goal is to compare risks or convey an individual risk in context, bar charts, risk ladders, scales, and icon arrays appear quite effective [24]. Icon arrays – pictographs consisting of a field of 100 shapes with shading to depict the likelihood of an outcome – excel at portraying percentages as distinct visual units and part-to-whole ratios, work with individuals with varying levels of numeracy, and overcome some cognitive biases [32-35]; they have been adopted in several commercial and research settings to convey genetic risk [36-39]. Although no single risk communication approach is universally recommended, it is advised that message developers consider the specific communication goal (e.g., present a comparison, show a time-based trend), minimize cognitive effort by providing less information not more, convey absolute risk rather than relative risk information, and provide comparative risk information when appropriate [40-43].

A related consideration involves the types of information that individuals may want to receive when obtaining results from GRMT, given that various risk communication strategies differ in the information they can potentially convey. Several recent studies suggest that there are limits to both the amount and complexity of information that users want to receive in the context of genetic testing. For example, a study comparing different risk communication strategies for presenting Oncotype DX genomic test results regarding risk for breast cancer recurrence found that participants preferred and had better comprehension of a simpler presentation that included less information (i.e., a recurrence risk stated in plain language, a risk continuum graphic, and confidence interval) than a standard report that included multiple information elements (i.e., a recurrence score, recurrence risk, graph, confidence interval, plain language risk categories, an assay description, and information about the test and company) [44]. Similarly, an analysis of verbal communication occurring in cancer genetic counseling sessions observed a mismatch between the types of information provided in the session and the information that patients preferred to receive [45]. Whereas the information provided frequently focused on details of risk assessment, biology, and technical aspects of testing, patients wanted to receive more information about their personal cancer risks and how test results directly applied to them. Thus, test users may not perceive all available information, or the same information that is prioritized by genetics experts, as being relevant or useful.

Before GRMT can be translated into patient care, it is critical to first develop and systematically evaluate GRMT risk communication materials to ensure that they are interpretable and consistent with the preferences of future test users. Thus, with the present study we sought to develop risk communication materials that would effectively convey results of this novel test to individuals. We developed materials using various risk communication strategies (e.g., icon arrays, bar charts) to present different types of potentially relevant information including the number of genetic risk modifier variants for which one had tested positive, prevalence of genetic risk modifiers in the population of BRCA1/2 carriers, implications of genetic risk modifiers for future breast cancer risks, and limitations and scientific uncertainties regarding GRMT. Then, using “gold standard” cognitive interviewing procedures [46], we obtained feedback from BRCA1/2 carriers about their perceptions, comprehension, and preferences regarding sample risk communication materials conveying hypothetical GRMT results to identify a refined approach to communicating about GRMT.

Eligible study participants were identified by their physician and included English-speaking women aged 25 to 80 years who had a BRCA1/2 mutation and were being treated at the Memorial Sloan Kettering Cancer Center Special Surveillance Breast Clinic.

Development and evaluation of GRMT risk communication materials proceeded in two phases.¹ In phase 1, the educational material was designed by our study team of experts in clinical genetics, health psychology, risk communication, and qualitative methodology to describe how genetic risk modifiers affect breast cancer risk. This educational material explained how each genetic risk modifier has different variants that enhance, reduce, or have a neutral effect on a BRCA2 carrier’s overall risk for breast cancer. This material included multiple analogies to describe the process by which genetic risk modifiers affect breast cancer risk among women with BRCA2 mutations (analogies included a written description of how balls of different sizes could be added to a cup of water to raise the water level with illustrations, a written description of how small cracks and a hole created in a layer of ice could interact to cause ice to break apart on a frozen lake, and a written description of how smoking behavior could interact with genetic modifiers to cause lung cancer to develop in some smokers but not others). To determine whether the educational material was perceived as comprehensible by the target user population, a team member (J.G.H.) conducted brief cognitive interviews with participants. Participants reviewed the educational material and provided verbal feedback including responses to true/false knowledge items and their perspectives about the clarity of the information. Interviews were audio-recorded, transcribed, and reviewed by members of the study team. The first round of cognitive interviews (n = 7) indicated several areas for improvement, including a preference for illustrations to accompany the analogies, confusion on several key points, and no clear preference for any of the analogies. The educational material was revised to include new analogies (including a written description of how weights of different sizes could be added to a scale/balance with illustrations, a written description of how cups of different sizes could be used to add or remove water from a larger cup to change the water level with illustrations, and a written description of how smoking behavior could interact with genetic modifiers to cause lung cancer to develop in some smokers but not others with illustrations) and then evaluated in a second round of cognitive interviews (n = 6). Participant feedback indicated a strong preference for the analogy involving a scale/balance, a high level of knowledge following review of the material, and few suggestions for additional revision. The final educational material (online suppl. material, see www.karger.com/doi/10.1159/000505854) was used in phase 2 of the study.

In phase 2, GRMT result materials were developed to depict sample hypothetical results of a GRMT panel comprised of 14 genetic risk modifiers (consistent with the number of identified genetic risk modifiers at the time this study was developed [17, 18]) for a woman with a BRCA2 mutation. We developed four sample GRMT result presentations, informed by risk communication best practices (Fig. 1). These presentations differed in the risk communication strategies used to depict the results as well as the specific pieces of information conveyed by the presentation. Presentation 1 depicted a panel of genes to convey information about the specific number and types of genetic risk modifier variants carried by an individual. Presentation 2 used icon arrays to convey contextual social comparison information regarding how an individual’s GRMT result compared to those of other BRCA2 carriers (integrating “restroom gender” icons as per Zikmund-Fisher et al. [47], with images created with Iconarray.com [48]). Presentation 3 included a verbal risk estimate stating an individual’s average breast cancer risk and the upper and lower confidence intervals of the estimate based on her GRMT result. Presentation 4 involved a graphical risk estimate with a bar chart showing the breast cancer risks of the average woman and average BRCA2 carrier, as well as the estimated range of breast cancer risk for the individual given her GRMT result using a “blurred bar” feature to convey the uncertainty of the specific estimate [49]. Each presentation also included 2 to 5 relevant limitations of the GRMT results (e.g., “These estimates could change in the future as more information is learned about these genetic modifiers and cancer”). We created three versions of each presentation that reflected low, moderate, or high breast cancer risk GRMT results.

To evaluate these GRMT result materials, we recruited 30 participants (from a pool of 72 eligible patients, of whom 11 refused and 31 could not be contacted/scheduled) to complete an in-person individual cognitive interview with a study team member (J.G.H., M.G.G., E.S.). Participants were randomized to receive hypothetical GRMT results corresponding to one of the three levels of breast cancer risk (10 participants per version). Participants were first shown the educational material developed in phase 1 and told to “Imagine that as a woman with a BRCA2 mutation, your doctor has offered you a new test that checks for the 14 genetic modifiers for breast cancer risk. You take the test and get these results.” Participants were shown each of the four sample result presentations one at a time (in the order of presentation 1: panel of genes, presentation 2: icon array, presentation 3: verbal risk estimate, presentation 4: graphical risk estimate) and instructed to imagine each presentation was the only results that they had seen. For each presentation, participants were asked to provide qualitative feedback about the presentation and their comprehension of the results based on a semi-structured interview guide, and were also asked to indicate their perceived lifetime breast cancer risk based on the results using a 5-point Likert scale (1 = very unlikely to 5 = very likely). After reviewing all four presentations individually, participants were asked to examine them side-by-side and provide feedback about their overall impressions of the materials and their anticipated interest and utility of GRMT. Interviews lasted approximately 60 min and were audio-recorded and transcribed. Following the interview, participants completed a survey assessing demographics and numeracy (using the validated Subjective Numeracy Scale [50, 51]). Participants received USD 50 for their contribution.

Transcripts were analyzed through thematic content analysis, an inductive qualitative data analysis method that seeks to identify and interpret recurring conceptual patterns directly from the data through intensive reading, coding, and interpretation [52-56]. This analytic approach [55, 57] involved four coders (J.G.H., M.G.G., J.S.W., E.S.) and iterative rounds of consensus analysis to ensure reliability of the findings [58, 59]. ATLAS.ti was used to facilitate analysis [60]. First, team members independently read the same transcripts, identified notable narrative content, created descriptive and interpretive codes capturing the essence of such content, and assigned the codes to the relevant narrative segments; team members subsequently met to reach consensus on codes, their meanings, and assignment to the narratives. Through this process a consensus codebook was developed. The team subsequently used the codebook to code future transcripts, which was refined and modified throughout this process of independent and collaborative coding. Upon coding of all 30 transcripts, the coders then engaged in a secondary analysis to interpret and synthesize the coded narrative content representing key themes related to risk communication material preferences as well as perceived interest in and response to GRMT that had emerged during the first stage of consensus coding. The final stage involved team collaboration to identify commonalities across each coder’s thematic observations and reach agreement regarding salient themes and subthemes and how they should be best described. We selected the most illustrative participant quotations from the interview data to support our key findings. Additionally, participants’ responses were quantified in instances where the interview data allowed it (e.g., Likert scale responses regarding perceived breast cancer risk, yes/no or categorical response options to interview probes). Descriptive statistics were computed for demographic and numeracy data.

The characteristics of phase 2 participants are presented in Table 1. Participants ranged in age from 28 to 74 years (mean = 46.3 years), and the majority were white (86.7%), non-Hispanic (93.3%), and had completed postgraduate education (63.3%). The numeracy level in this sample was generally high (mean = 4.8 out of 6).

As shown in Figure 2, participants across all three hypothetical risk versions expressed a preference for the graphical risk estimate presentation (n = 27, 90%). Analysis of the qualitative data (Table 2 for all themes, subthemes, and participant quotations) indicated that participants perceived this risk communication presen-tation to be clear and straightforward. Furthermore, participants across all three risk versions noted that the graphical risk presentation provided comparative risk data that enable personalization of risk, including valuable contextual information that assists with comprehension of one’s personal breast cancer risk. A few participants shared their impressions about how uncertainty associated with the risk estimate was depicted in the graphical risk estimate presentation, whereas some found this feature to be a useful way to communicate the probabilistic nature of risk estimates, and others found this to be confusing or unsatisfying because of a desire for a specific number.

Although most participants preferred the graphical risk estimate presentation, half of the sample suggested that it should be used in combination with the verbal risk estimate presentation (n = 15; 50%). The verbal risk estimate presentation was described as having specific strengths, including being both clear and succinct, and requiring no cognitive interpretation to understand one’s personal breast cancer risk. Participants recommended combining these two presentations to provide the most comprehensive information regarding a woman’s personal breast cancer risk. Several participants also suggested labeling the bar graph used in the graphical risk estimate presentation with percentage gradations to improve comprehension (e.g., 0 to 100 with increments of 10).

In contrast, participants generally expressed dissatisfaction with the panel of genes presentation and icon array presentation. With the panel of genes presentation, participants often described uncertainty about what the results would mean for their personal breast cancer risk. Some participants were confused about how to interpret this presentation, raising questions about the size of the effect of a genetic risk modifier variant or whether interactions occurred between the copies of genetic risk modifiers inherited from each parent. For others, the information conveyed by the panel of genes presentation was understood but perceived to be unnecessary. In response to the icon array presentation, many participants felt confusion and frustration and had difficulty in interpreting its meaning. Participants frequently described the presentation as incomplete or complex, which led to difficulty in interpreting the results. Other participants correctly interpreted the meaning of the icon array presentation but found the information overwhelming or irrelevant to their personal situation. Therefore, for both the panel of genes presentation and the icon array presentation, participant feedback suggested that these presentations were either difficult to understand or conveyed information deemed to be unimportant.

We explored how each of the different GRMT result presentations may have influenced participants’ perceived breast cancer risk. Specifically, for each GRMT result presentation we conducted ANOVAs to assess whether perceived breast cancer risk differed across the hypothetical risk versions (low vs. moderate vs. high). Post hoc analyses were conducted using Fisher’s least significant difference test. As presented in Figure 3, the perceived lifetime risk of developing breast cancer significantly differed among participants shown the three hypothetical risk versions with the graphical risk estimate presentation (F_{(2, 26)} = 8.09, p = 0.002). Specifically, in response to the graphical risk estimate presentation, participants shown the high-risk version perceived greater risk for breast cancer than did participants shown the moderate-risk version, who in turn perceived greater risk for breast cancer than did those shown the low-risk version. Similarly, the perceived breast cancer risk significantly differed among participants shown the three risk versions with the verbal risk estimate presentation (F_{(2, 26)} = 10.65, p < 0.001). However, no differences in perceived risk were observed across risk versions in response to either the panel of genes or icon array presentations.

During the interview, participants were asked how interested they would be in receiving GRMT if testing were available. Most (n = 23; 76.7%) participants would be interested in having their own GRMT. A minority (n = 4; 13.3%) were ambivalent or uncertain about their personal interest. Finally, a few participants (n = 3; 10.0%) were uninterested in GRMT.

Several common themes regarding the utility of GRMT were described by participants across all three hypothetical risk versions. Participants frequently explained that as BRCA1/2 carriers, they were continuously negotiating the ambiguities of their cancer risks. Thus, the potential ability of GRMT to clarify their understanding of their breast cancer risk was highly valued. Participants viewed the prospect of having additional knowledge as empowering and valuable for informing risk management decisions and interactions with healthcare providers. Many expressed a belief that GRMT would be valuable regardless of the specific results gained from testing. Although participants explained that they might take different health-related actions upon knowing that they were at higher or lower risk, the inherent value of the GRMT itself would generally remain high.

Participants’ feedback suggested that the perceived utility of GRMT was tied to the actionability, precision or accuracy, and personalization of the results. In general, participants would ascribe higher value to GRMT if the results could provide clear direction to inform their next steps regarding health decisions and could offer concrete, precise, and personalized risk information. If GRMT could not provide such information, then the value of the testing would be diminished. In general, participants recognized that scientific understanding is evolving as research is conducted over time and that exact or definitive risk estimates are not likely to be attainable; yet, obtaining as much precision as possible was deemed to be important. Although expressed among participants across all three risk versions, such concerns about the specificity of GRMT results were most frequently raised by those viewing low or moderate risk results.

Participants were queried about the impact that GRMT results would have on their cancer risk management decisions. A majority (65.5%) anticipated that GRMT would make their decision-making processes easier, some (31.0%) were unsure how this information would affect the difficulty of their decision-making, and only one participant (3.5%) anticipated that GRMT would make her decisions harder. Furthermore, participants believed different GRMT results would have specific implications for risk management decisions. Participants anticipated that learning low-risk GRMT results would lead them to continue heightened breast cancer surveillance practices and assist in making choices about foregoing more radical preventative measures (e.g., prophylactic mastectomy). Participants anticipated that learning moderate-risk GRMT results would not significantly alter their present understanding of their risk. Thus, these results would have a limited impact on risk management practices, particularly surveillance, and would not likely lead them to pursue prophylactic surgery. Participants felt that learning high-risk GRMT results would enhance interest in more invasive preventative measures and might lead some to intensify screening.

As participants evaluated the risk communication materials, many anticipated that their previous experiences, including family history of cancer and initial emotional response to receiving a pathogenic BRCA1/2 result, would influence their affective reactions to GRMT results. When participants had strong family cancer histories, they often reported feeling underwhelmed by the GRMT results. These participants tended to perceive their breast cancer risk as high and felt that the GRMT results did little to alter these beliefs. Nonetheless, most of these women ascribed value to the GRMT results since they regarded receiving more information as better. Across all risk versions and presentations, participants discussed how they felt that their emotional response to GRMT results would not be nearly as strong as their initial response to learning of their BRCA1/2 mutation. These individuals perceived that the GRMT results would offer additional insight into their cancer risk, but would unlikely carry the emotional weight or negative impact of their BRCA1/2 genetic test result.

This study examined the preferences and attitudes of BRCA1/2 carriers regarding newly developed risk communication materials for a novel GRMT that has the potential to refine estimates of their future breast cancer risks. Participant feedback revealed a common preference for combining graphical and verbal risk estimate presentation as a means of communicating GRMT results. Participants perceived these presentations as including clear and comprehensible risk communication strategies as well as providing valuable, personalized information about their cancer susceptibility. This feedback is consistent with psychological research demonstrating the utility of pairing written, numerical information with graphical presentations of risk [23, 25-29]. Furthermore, through analysis of participants’ quantitative ratings of their perceived lifetime breast cancer risks in response to each sample presentation and across risk versions, we observed that the graphical and verbal risk estimate presentations uniquely affected participants’ risk beliefs. Based on both the qualitative and quantitative data, we conclude that a refined risk communication approach integrating graphical and verbal risk estimate presentations will not only meet future patients’ preferences, but may effectively convey the meaning of GRMT results.

In contrast to past research and recommended risk communication best practices supporting the utility of icon arrays [32-35, 40], participants were particularly displeased with this presentation. It is notable that contrary to how icon arrays are typically used, which is to graphically convey a proportion or percentage reflecting the likelihood of a specific event (e.g., experiencing a side effect) for an individual, we attempted to use this strategy to convey the prevalence of genetic risk modifiers in the population of BRCA2 carriers. Our objective was to provide social comparison information about how an individual’s GRMT result compared to those of other BRCA2 carriers. Many participants found this application of the icon array difficult to understand, and even those who correctly understood the intended meaning were dissatisfied with the information. Participants were similarly dissatisfied with the panel of genes presentation, although a minority believed this information was interesting or could be offered in a supplemental nature to patients who wanted to learn more about their results. It is possible that participants’ preferences for specific ways of presenting GRMT results were guided by their preferences for specific pieces of information. Across presentations, varying risk communication strategies were used to convey distinct types of information that could be revealed by or be relevant to GRMT results. Ultimately, participants’ predominant concern was understanding their personal cancer risk; consequently, presentations that offered information about the number of single nucleotide polymorphisms identified through GRMT or the population prevalence of these results were deemed unsatisfactory since such information could not easily be converted into insight about their own susceptibility.

We observed substantial interest in the prospect of undergoing GRMT, with three-quarters of participants expressing interest in testing. This level is consistent with a recent study wherein 85% of interviewed BRCA1/2 carriers were interested in receiving refined cancer risk estimates [61]. Similarly, a study of breast cancer polygenic risk information among women from breast cancer families with uninformative genetic testing results found high interest (86.5%) and moderate actual uptake (42.1 to 61.8%) of a polygenic risk score [62]. This finding is also consistent with a study examining interest in genetic testing for genes related to modest changes in breast cancer risk (e.g., moderate penetrance gene mutations) [63]; in this survey, 77% of women at moderate to high risk for breast cancer were interested in testing, with interest being greater for tests that conveyed more cancer risk. However, in the present study, high levels of interest existed across the hypothetical risk versions (i.e., low, moderate, and high risk). Participants’ interest in GRMT appears to have been largely shaped by the desire to gain a deeper understanding of their future cancer risk, a general belief in the empowering nature of information, and the potential for these results to inform, and possibly minimize the difficulty of, risk management decisions. Previous research has demonstrated that reducing uncertainty regarding cancer risk and decision-making is a key motivator for undergoing genetic testing, and patients might experience ongoing distress when their uncertainty is not reduced following testing [64-67]. This is not only true for individuals who receive uninformative genetic test results (e.g., variants of uncertain significance) [67-69], but also for individuals who receive pathogenic BRCA1/2 genetic test results indicating heightened cancer risk. These individuals can struggle with managing sustained, residual uncertainty regarding the prospect of cancer development and the potential efficacy of risk management options, and information seeking can be a common coping strategy [70]. Consequently, women with BRCA1/2 mutations could be very receptive to GRMT given its potential to provide information that can resolve at least one source of uncertainty, namely the probability of developing cancer in the future [71]. However, our results suggest that resolving probabilistic uncertainty is not entirely sufficient for establishing the utility of GRMT; study participants also placed high value on the ability of GRMT to produce risk information that is precise and accurate. Such concerns reflect another source of uncertainty – ambiguity (i.e., uncertainty regarding the strength, validity, consistency, or adequacy of risk estimates or risk information [71]) – that patients contend with when interpreting genetic risk information.

In summary, these results suggest that women with BRCA1/2 mutations would be highly receptive to GRMT and that the minimal amount of necessary information to be included in GRMT result risk communication materials includes graphical and verbal estimates of future breast cancer risk. Communicating cancer risk through these presentations appears to be acceptable to patients, influences their breast cancer risk perceptions, and may assist in risk management decision-making. However, several study limitations must be acknowledged. This sample was small, racially and ethnically homogenous, and generally well educated with high numeracy. Further, participants were treated at one institution and recruited from a high-risk cancer surveillance clinic. Thus, results may not be generalizable to the broader population of BRCA1/2 carriers treated in other settings or who are less actively involved in managing their cancer risks. Though participants had direct experience with BRCA1/2 genetic testing, the GRMT results were hypothetical and were presented in the order described (i.e., panel of genes, then icon array, verbal risk estimate, and finally graphical risk estimate). This ordering of exposure to the result presentations may have influenced participants’ perspectives; for example, the significant difference observed in perceived risk for the graphical risk estimate presentation may be due to participants having seen the verbal risk estimate previously. Therefore, findings related to the impact of the result presentations on perceived risk should be interpreted cautiously. Nonetheless, each participant viewed all four pre-sentations side-by-side toward the conclusion of the interview and was solicited for her overall preferences and impressions. We also did not assess additional individual characteristics that might have influenced both presentation preferences and overall interest in GRMT. For example, patient characteristics such as intolerance for uncertainty, a dispositional tendency to respond negatively on a cognitive, emotional, and behavioral level to an unknown outcome [72, 73], have been associated with responses to genetic risk information [68] and may have contributed to patients’ perceptions of the GRMT result presentations. Future studies should investigate how patients’ attitudes toward and experiences with different sources of uncertainty affect not only their use of GRMT, but their subsequent reactions to the information provided by this testing. Overall, this study design allowed for an in-depth, thorough analysis of the perspectives of patients similar to those who are likely to be initial users of a future GRMT, a gold standard initial step toward maximizing the comprehensibility and acceptability of novel risk communication strategies in research and clinical settings.

These findings can be used by our team and other investigators to inform ongoing efforts to translate GRMT into cancer care in a manner consistent with patient preferences and information needs [39, 61, 74]. Genomics research offers promise that GRMT and calculation of polygenic risk scores can be leveraged to refine estimates of BRCA1/2 carriers’ cancer risks [16-19, 21, 22]; this possibility was a primary driver of participants’ interest in GRMT, with test results being deemed as valuable insomuch as they could help patients to contextualize and clarify uncertainty about their cancer susceptibility. However, as this testing moves into clinical practice, it will be crucial for future studies to investigate patients’ actual uptake of GRMT and to explore how results subsequently shape patients’ cognitive perceptions of risk and emotional reactions, and to examine what influence GRMT ultimately has on medical decision-making.

The authors are extremely grateful to all participating patients.

The Memorial Sloan Kettering Cancer Center Institutional Review Board deemed this work exempt research; therefore, all -participants confirmed verbal agreement with their voluntary willingness to participate. All study procedures and materials were approved by the Memorial Sloan Kettering Cancer Center Institutional Review Board.

The authors have no conflicts of interest to declare. The content is solely the responsibility of the authors and does not necessarily represent the official views of any funder.

This research was supported by an award from the Breast Cancer Research Foundation (principal investigator: M.E. Robson), the Robert and Kate Niehaus Center for Inherited Cancer Genomics, the Andrew Sabin Family Foundation, and NCI P30 CA008748. J.G. Hamilton was also supported by a Mentored Research Scholar Grant in Applied and Clinical Research (MRSG-16-020-01-CPPB) from the American Cancer Society and by the National Cancer Institute of the National Institutes of Health under award number R21 CA230879.

J.G. Hamilton, PhD, MPH, M. Genoff Garzon, MA, I.H. Shah, BA, K. Cadet, MPH, E. Shuk, MA, J.S. Westerman, BA, J.L. Hay, PhD, K. Offit, MD, MPH, and M.E. Robson, MD contributed to the design and implementation of the research, to the analysis of the results, and to the writing of the manuscript.

The data that support the findings of this study are available from the corresponding author upon reasonable request.