“Is it Really Worth it to Get Tested?”: Primary Care Patients’ Impressions of Predictive SNP Testing for Colon Cancer

Abstract

Despite significant progress in genomics research over the past decade, we remain years away from the integration of genomics into routine clinical care. As an initial step toward the implementation of genomic-based medicine, we explored primary care patients’ ideas about genomic testing for common complex diseases to help develop future patient education materials and interventions to communicate genomic risk information. We conducted a mixed-methods study with participants from a large primary care clinic. Within four focus groups, we used a semi-structured discussion guide and administered brief pre- and post- discussion quantitative surveys to assess participants’ interest, attitudes, and preferences related to testing and receipt of test results. Prior to the discussion, moderators presented a plain-language explanation of DNA and genetics, defined “SNP”, and highlighted what is known and unknown about the risks associated with testing for SNPs related to colorectal cancer risk. We used the NVIVO 8 software package to analyze the transcripts from the focus group discussions.

The majority of participants (75%) were “very” or “somewhat interested” in receiving information from a colon cancer SNP test, even after learning about and discussing the small and still clinically uncertain change in risk conferred by SNPs. Reported interest in testing was related to degree of risk conferred, personal risk factors, family history, possible implications for managing health /disease prevention and curiosity about genetic results. Most people (85%) preferred that genetic information be delivered in person by a healthcare or genetics professional rather than through print materials or a computer. These findings suggest that patients may look to genetic counselors, physicians or other healthcare professionals as gatekeepers of predictive genomic risk information.

INTRODUCTION

Since the completion of the Human Genome Project in 2003, the field of genomics has become one of the frontiers of biomedical research, in part due to genome-wide association studies (GWAS) that seek to identify statistical relationships between gene variations and disease (Norrgard, 2008). GWAS studies have identified hundreds of associations of common genetic variants with over 80 diseases and traits, including almost two dozen cancers (Chanock, 2011; Hindorff et al., 2009). Moreover, genomic research has facilitated pharmacogenomic testing for drug tolerance (e.g., polymorphisms in CYP2D6) (Green, Guyer, & National Human Genome Research Institute, 2011). Despite these advances, we are years away from routine use of genomic information for personalized risk assessments for common adult conditions in primary care (Chanock, 2011; Green et al., 2011). These delays are due in part to concerns about the clinical utility of genetic testing for single nucleotide polymorphisms (SNPs), the most common type of genetic sequence variation related to common disease (Epstein, 2004; Haga, Khoury, & Burke, 2003; Institute of Medicine, 2011; Khoury & Mensah, 2005; Khoury et al., 2009; Robson et al., 2010).

Many professionals in the medical and genomics communities have cautioned against the premature translation of SNP testing to clinical practice (Biankin & Chanock, 2011; Goodwin, 2008; Haga et al., 2003; Hunter, Altshuler, & Rader, 2008; Kraft et al., 2009; Lunshof, Pirmohamed, & Gurwitz, 2006). Nonetheless, commercial companies offer genomic testing via direct-to-consumer (DTC) marketing, including nutrigenomic, pharmacogenomic and SNP testing for risk prediction of common diseases. The availability of DTC genomic testing precedes research to identify best practices in communicating the potential risks, benefits, and limitations of such testing. Recent studies have documented substantial interest in predictive susceptibility testing for numerous health conditions within the general population (Cherkas, Harris, Levinson, Spector, & Prainsack, 2010; Neumann et al., 2012) and women at moderately increased risk for breast cancer (Graves, Peshkin, Luta, Tuong, & Schwartz, 2011;). In one survey of social network users, almost three-quarters of participants reported that they would consult with their physician about genomic test results, and most believed that physicians had a professional obligation to help patients interpret test results (McGuire, Diaz, Wang, & Hilsenbeck, 2009).

The commercial availability of genomic tests (Kolor, Liu, St Pierre, & Khoury, 2009) and rapid escalation of gene discovery research highlight the need to study how people seek and apply uncertain genomic information. Just as translational research informed integration of BRCA1/2 genetic counseling and testing into clinical practice (Schwartz et al., 2002; Schwartz et al., 2004), translational research is needed to explore uptake, understanding, and potential impact of genomic tests prior to its continued commercial diffusion. With a better understanding of the implications of this type of genetic information, including concerns about clinical utility, privacy/confidentiality, and behavioral and psychosocial responses to testing, the field of genomics could begin to impact public health.

We explored primary care patients’ interest, attitudes, and preferences related to genomic testing and receipt of test results for common complex diseases. Specifically, we aimed to collect in-depth information on general opinions about genomics, SNP testing for common complex diseases, and preferences for information delivery. Our work was guided by The Informed Choice Model (Michie, Dormandy, & Marteau, 2003) which outlines an approach for determining the knowledge necessary for making an informed choice about uncertain medical tests/procedures, including use of professional guidelines and consumer views of the test. We focused on primary care patients’ opinions and attitudes about SNP testing because primary care patients may represent potential consumers of SNP testing (McGuire et al., 2009) and because primary care providers may serve as conduits of personalized genetic medicine in the future (Burke, 2004; Cheng, Cohn, & Dover, 2008; Institute of Medicine, 2005). Our work was also shaped by The Ottawa Framework for Decision Support, which suggests that effective risk communication include factual information, options for the decision to test, pros and cons related to testing, and clarification of attitudes about testing (Edwards et al., 2004; O'Connor et al., 1998). Guided by these theoretical approaches, we sought to understand patients’ perspectives on SNP testing to guide efforts related to appropriate use of genomic information about disease risk.

MATERIALS AND METHODS

Setting and Participants

We conducted a mixed-methods cross-sectional study at the Lombardi Comprehensive Cancer Center at Georgetown University and the Division of General Internal Medicine at Georgetown University Hospital (GUH) in 2008. Our mixed-methods approach involved both focus groups and administration of a pre- and post-discussion survey. Focus groups provide an excellent way to gather information about an emerging area of research (Morgan, 1998) and we complemented this approach by embedding written surveys to quantitatively assess interest in SNP testing during the group discussion (Papp et al., 2004). Participants were recruited from the primary care clinic within the Division of General Internal Medicine at GUH. All study procedures were approved by the Institutional Review Board at Georgetown University/MedStar Health.

Men and women aged 18 years and older were eligible for this study. Exclusion criteria were (a) inability to read or understand English or (b) cognitive impairment that precluded informed consent. Potential participants’ ability to provide informed consent was determined by their self-reported medical history and the clinical determination of trained study personnel.

Procedures

Prior to beginning recruitment, study personnel reviewed IRB-approved recruitment scripts, conducted practice “mock recruitment” with the study PI and other research staff and observed the PI recruiting potential participants. Trained study personnel approached potential participants in a primary care clinic waiting room to verify age, eligibility, provide a study flier, and assess willingness to learn more about the study. In the summer and fall of 2008, we used convenience sampling by approaching all potential participants in the waiting room of a primary care clinic in an academic medical center, varying the days of the week and times during the day for our recruitment efforts. We did not assess the reason for participants’ medical appointments with their primary care physician. Study personnel provided interested patients a copy of consent materials; consents were signed either at the time of study recruitment or immediately prior to focus group participation. Interested participants were then contacted again to identify dates and times convenient for their schedule; we scheduled up to 12 participants per group.

Focus Groups

Discussion Guide Development

We developed a semi-structured discussion guide to elicit participants’ thoughts and opinions about genetic testing and their degree of interest in SNP testing for risk of cancer and other conditions (e.g., diabetes, heart disease, Alzheimer’s disease). Based on concepts in the Informed Choice Model and the Ottawa Framework for Decision Support, two members of the research team (KDG and BNP) drafted the discussion guide. We then elicited feedback of team members from various disciplines: genetic counseling, behavioral science, public health, bioethics, and medicine and revised the discussion guide accordingly. We made further revisions to the guide after the first focus group to clarify how we presented background information about SNPs. The final guide and accompanying pre- and post-discussion surveys assessed participants’ interest in and attitudes about testing and preferences for receiving test results (e.g., from a genetics professional, printed booklet, or computer; see Appendix).

Moderation of Group

Two study team members, a clinical psychologist and a certified genetic counselor, moderated the group discussion. At the beginning of each group, we obtained written informed consent, explained issues regarding confidentiality, and asked participants to complete brief demographic questionnaires. Moderators explained basic guidelines for the discussion (e.g., only one speaker at a time) and participants were encouraged to share their opinions freely and openly. Moderators then began the discussion by asking participants what they know and think about cancer risk due to genetics. In the first focus group, participants were informed of the moderator’s backgrounds as a clinical psychologist and genetic counselor. Following this first group, the study team decided to not specify the training of each of the moderators to reduce any potential influence on participants’ responses (e.g., such as preferences for the delivery of SNP testing information from a genetic counselor, a physician, etc). As noted by Giacomini and colleagues (Giacomini & Cook, 2000), iterative adjustment to the methods used in the collection of interview and focus group data is an appropriate technique and even a recommended way to gather qualitative data.

Overview of Genetics and Brief Pre- and Post-Discussion Survey Administration

Following the initial discussion of participants’ thoughts about genetics and cancer risk, the moderators gave a plain-language slide presentation to explain DNA and genetics, defined “SNP”, briefly described the Genetic Information Nondiscriminatory Act (GINA) and highlighted what is known and unknown about the risks associated with SNPs (see Appendix). We presented information in the context of two colorectal cancer SNPs because colorectal cancer affects both men and women and has established screening methods. After the overview of genetics, we administered a brief written survey (“pre-discussion survey”). Participants completed the survey individually; we then reviewed responses as a group to generate discussion about interest in genetic testing, factors that affect interest, and preferences for communication of test results. The survey assessed interest in SNP testing, factors that impact interest and preferences for delivery of SNP information (see Appendix).

Hypothetical Interest in Genetic Testing for Colon Cancer Risk

We assessed hypothetical interest in testing for SNPs that indicate increased and/or decreased risk based on SNP test results with two items: one that assessed interest in a SNP associated with increased risk and one that assessed interest in testing for a SNP associated with decreased risk of colorectal cancer (see Appendix). Response options were made on a 5-point Likert Scale ranging from ‘not at all interested’ to ‘very interested.’

Hypothetical Interest in Genetic Testing for Other Health Conditions

Using a single item, we assessed interest in SNP testing for other health conditions: ‘What health conditions would you be most interested in having SNP testing done to determine your risk?’ Response options were cancer, heart disease, diabetes, Alzheimer’s disease and a write-in response for other health conditions.

Factors Influencing Uptake of Genetic Testing

We asked participants to identify the factors they considered most important regarding SNP testing for cancer risk using a list of options adapted from prior research of interest in SNP testing (Cherkas et al., 2010; Graves et al., 2011) (see Appendix). We also asked participants to indicate their willingness to pay for SNP testing.

Preferences for Information Delivery

We assessed preferences for delivery of pre-testing education and results disclosure by asking participants how they would want both pre-SNP test and SNP result information provided to them (e.g., in-person, print materials, computer; see Appendix).

Discussion of Pre-Discussion Survey and Administration of Post-Discussion Survey

After participants completed the pre-discussion survey individually, the moderators prompted participants to share their responses and elicited discussion about each survey item. Following this discussion, moderators again asked participants to complete the same brief written survey to capture whether opinions about SNP testing had changed because of the group discussion (“post-discussion survey”). Each survey took less than 10 minutes to complete.

Group Procedures

We conducted a total of four groups, with four to eight participants per group (see Table 1 for demographics). We maximized our ability to achieve data saturation by following the recommendations of Giacomini and colleagues (Giacomini & Cook, 2000): 1) use of an iterative process of data collection and data analysis, 2) use of two trained raters to code data, 3) review of identified themes by multiple study investigators to establish that coders appropriately interpreted participants’ viewpoints and 4) use of ‘member checking’ to compare participant responses between quantitative surveys and the qualitative transcripts. Each focus group lasted 2 hours and was audiotaped with participants’ permission and later transcribed verbatim. A study team member took field notes during the focus groups to capture contextual factors such as participants’ expressions and/or body language (Pope & Mays, 1995). We reimbursed participants for costs related to transportation and parking. Participants received a $30 gift card to thank them for their time.

Table I.

Participant Characteristics (N=24)

Characteristics	N (%)
Age (Range 32 – 77 years)
(M = 54.3, SD = 12.3 years)
< 50 years	9 (37.5%)
≥ 50 years	15 (62.5%)
Race
Caucasian	9 (37.5)
African-American	10 (41.7)
Other	5 (20.8)
Education
<College	11 (45.8)
≥College	13 (54.2)
Household Income
<$75k	14 (58.3)
≥$75k	4 (16.7)
Did not respond	6 (25.0)
Married/living as married
Yes	8 (33.3)
No	16 (66.6)
Personal cancer history
Yes	5 (20.8)
No	19 (79.2)
Family cancer history
Yes	13 (54.2)
No	11 (45.8)

Quantitative Analyses

Quantitative analyses were used to analyze the pre- and post-discussion survey data. Specifically, we generated descriptive statistics to characterize sociodemographics, family history, psychosocial and medical variables (Table 1). We also generated descriptive statistics to characterize level of interest in genetic testing. Finally, we conducted t-tests and χ² tests to examine relationships between participant characteristics and interest in SNP testing (high vs. low interest).

Qualitative Analyses

We analyzed the focus group transcripts using thematic analysis and qualitative description approaches (Braun & Clarke, 2006; Sandelowski, 2000). Analyses were conducted with the NVIVO 8 software package, a qualitative research analysis program (NVIVO Software, QSR International). Guided by our theoretical models (Michie et al., 2003; O'Connor & O'Brien-Pallas, 1989), empirical research (Graves et al., 2011) and clinical experiences related to genetics and cancer risk (Peshkin, DeMarco, Brogan, Lerman, & Isaacs, 2001), we searched for themes related to interest in SNP testing, factors that influence interest, and general understanding of SNP testing and SNP risk. First, trained research staff reviewed the content and separately identified preliminary themes found in the transcripts. Second, two trained raters independently identified a priori and data-driven themes to describe logical relationships among the participants’ responses. Third, we revised and refined the identified themes using an iterative process until raters and study team investigators reached agreement about major coding themes and subthemes. We organized our results by theme and subtheme and extracted exemplary quotes.

RESULTS

Participants

Of 54 eligible primary care patients approached to participate, 24 participants took part in this study (44% response rate). The mean age of participants was 54.3 years (Range: 24–78 years). More than half (62%) of the participants were female and 54% had attended college or beyond. Less than half (41%) were married and 42% were African-American. Five participants reported a personal history of cancer [breast (n=2) carcinoid (n=1), throat (n=1), and colon (n=1) cancer] and seven reported a family history of cancer (see Table 1). The demographic characteristics of participants across the four focus groups did not differ.

Quantitative Results

Pre-SNP Risk Discussion Survey

In the pre-discussion survey, 75% of participants reported interest in SNP testing for a SNP associated with increased risk of colorectal cancer: 54.2% of participants reported that they were “very interested” and 20.8% were “somewhat interested.” Fewer participants (33.3%) reported that they were “very interested” in SNPs associated with a decreased risk. Cancer survivors reported less interest in SNP testing (χ²=7.14, p=0.03) compared to individuals without a personal history of cancer. Demographic factors and family history of cancer were unrelated to interest in SNP testing. Pre-discussion, participants reported they would be willing to pay between $150 and $200, on average, for a SNP test related to colorectal cancer risk; older age was related to a greater willingness to pay more money for a SNP test (r = .45, p =0.04).

Post-SNP Risk Discussion Survey

Participants reported the same level of interest in a SNP test for increased risk before and after the risk communication discussion (p<0.05). Similarly, interest decreased but was not statistically significantly different from pre- to post-discussion for a test related to decreased risk after the risk communication discussion (p< 0.05; Table II). Participants also ranked reasons for their interest in SNP testing. The highest ranking reasons, followed by the percentage of participants who selected it as their first choice, were: the amount the SNP increases or decreases my risk for cancer (17%), my own risk factors/family history of cancer (21%), and whether the test results would affect my cancer screening or risk reduction options or recommendations (13%). The pre- and post-discussion reasons participants indicated they would have a SNP test were not significantly different. Although the average amount participants indicated they would be willing to pay as an out-of-pocket cost decreased to $50 – $100 post-discussion, this difference was not statistically significance (paired t(19) = 1.32, p = .20; Table II). Most participants (85%) indicated a first choice preference to have SNP testing information delivered to them in-person by either a physician (35%) or a genetic counselor (50%). Preferences related to delivery of SNP information did not differ between the individual focus groups.

Table II.

Interest in and Delivery Preferences for SNP Testing for Colorectal Cancer Risk

Variable	Pre-Discussion	Post-Discussion
Interest in SNP testing for SNP that increased risk of CRC	75%	75%
Interest in SNP testing for SNP that decreased risk of CRC	65%	59%
Interest in having SNP testing to determine risk for these conditions
Cancer	71%	69%
Heart Disease	82%	81%
Diabetes	47%	44%
Alzheimer’s disease	71%	81%
Factors influencing interest in SNP testing: First Choice Ranking
Amount of risk conveyed	17%	17%
My own risk factors/family history of cancer	25%	21%
If results affect screening / risk reduction recommendations	13%	13%
Preferences for delivery of SNP pre-test information: First Choice Ranking
In-Person by a physician	35%	35%
In-Person by a genetic counselor	45%	50%
By Internet / computer	7%	8%
Printed Materials	17%	17%
Telephone call with a genetic counselor or physician	0%	0%
Preferences for delivery of SNP results information: First Choice Ranking
In-Person by a physician	45%	48%
In-Person by a genetic counselor	55%	42%
By Internet / computer	0%	10%
Printed Materials	0%	10%
Telephone call with a genetic counselor or physician	0%	0%
Average out-of-pocket expense willing to spend on SNP testing for CRC risk	$100–$150	$50–$100

Notes: CRC = Colorectal cancer

Qualitative Results

The following themes emerged from the analysis of focus group transcripts: reasons to have SNP testing, reasons not to have SNP testing, utility of SNP test results, potential outcomes following SNP testing, and information delivery preferences regarding SNP test results. Themes were consistent across all groups. Representative quotes for each theme are displayed in Table 3.

Table III.

Themes in Qualitative Results

Themes	Representative quotes
Reasons to Have SNP Testing
Personal Benefit of Knowing Risk	I think it’s better to know than not to know. But I’d like to know. Before I get it, I’d like to know.
Gathering Information for Family	That’s why I told my children that they need to be checked for heart disease. If there’s a SNP that would signal that, that would be great. It may help them prevent from getting a heart attack.
Other Reasons to Test	Sure, if someone offers it up, why not?
Reasons Not to Have SNP Testing
Concerns about Insurance and Privacy	What are the implications of this? If you know this and you got an increased risk and it goes into your medical history, life insurance companies take those medical histories into account when they’re coming up with plans. Does that impact what your rate is going to be?
Lack of Interest due to Small Change in Risk / Focus on One Disease	If you’re talking about a significant investment of money and you’re looking at it’s only going to make it change a little bit on this graph, do I really care?
Clinical Utility of SNP Results	It depends on if my doctor thinks if I might be at risk. That might convince me to do [SNP testing].
Anticipated Behavioral and Psychosocial Outcomes Following SNP Testing
Impact on Health Behaviors and Screening	If it tells me that I’m more prone to have that cancer at least I will go on and do what I have to get it better. I think education is everything.
Emotional Responses
Perspectives from cancer survivors	For people like ourselves who already had cancer, attitude is everything I think. And when you look at patients, to me, that’s the biggest single factor of why one person does better than another person, in general.
Perspectives from people without a personal cancer history	Worrying won’t do any good. But I think that there’s a significant benefit to knowing something certain. I’d like to know if I’m on the road to developing some form of cancer and if I am which form of cancer.

Reasons to Have SNP Testing

Consistent with the survey responses, many participants discussed being interested in SNP testing during the focus groups. Interestingly, none of the participants had any personal experiences with genetic testing, although many of them reported that they had heard of genetic testing for breast cancer risk through the media. Specific to SNP-testing, two primary subthemes emerged regarding why participants would be interested, despite the known limitations to SNP testing: the personal benefit of knowing risk and gathering information for family.

Personal Benefit of Knowing Risk

The most commonly cited reason for interest in SNP testing was to be informed and have knowledge about personal risk: “I’d like to know if that’s where I’m headed. Maybe there’s nothing I could do about it—that doesn’t matter. I think there’s a benefit to knowing as much as you can about yourself.” Other participants discussed benefits to knowing which aspects of their health they should attend to more than others. For example, one participant noted, “When it comes to genetic testing, in general, I would like to know what to worry about and what not to worry about.” Some participants commented that they felt SNP results would provide insight about prioritizing healthcare and screening for potential disease: “My doctor could pay more attention to appropriate testing.” Participants who wanted as much information as possible about their health appeared similar to individuals referred to as information seekers or high monitors – people who scan for relevant threat / health risk information (Miller, 1987).

Gathering Information for Family

Some participants felt that having information about disease risk based on SNP testing would positively influence the health and health-related behaviors of their family members. One participant remarked:

“One of the main reasons why I’m looking into this is because whatever’s going to happen to me will happen to me. But, I’m really looking at how I can best advise my grandchildren to get early detection. I want my son to exercise more…to get himself checked out, to get a colonoscopy regularly.”

Many participants felt additional information provided by SNP test results would impact their family’s health decision-making. For example, one woman commented: “I want to know because if it is genetic I have to warn my children…. It would be good for my family to take care of themselves.”

Participants reported interest in testing not only to gather information for their family, but also because of a family history of cancer. Participants associated the likelihood of having a colon cancer SNP with having a family history of cancer (i.e.: “I definitely would [get genetic testing] because my father died of colon cancer at age 50”). This association was true even after the group discussion in which we emphasized that we do not yet know the clinical implications of having both a family history and specific risk-related SNPs.

Other Reasons to Test

Participants who were not particularly interested in SNP testing to gather information for themselves or their family members noted altruism and a physician recommendation as other reasons to test. For example, one participant noted that she might consider testing if it was conducted in a research context: “With those numbers [the small increase in risk conveyed by a single SNP] the only reason I can justify having it would be to help the research come along. I can’t think of any other reason to do it.” Similarly, one woman commented about engaging in health research, “I like to participate in the empirical analysis of what’s going on with me. I wouldn’t let it worry me. It’s just part of engaging.”

Another participant relayed how a physician recommendation for the test would make him very interested in having it: “If in the routine physicals that we get every year, if my physician thought that this was a useful bit of extra information I would be very interested in it.”

Reasons Not to Have SNP Testing

In contrast to many participants who indicated interest in SNP testing, a few participants cited reasons for not wanting to test. Reasons not to test largely focused around concerns about privacy and insurance and the small amount of change in risk typically conveyed by SNP results.

Concerns about Insurance and Privacy

A few participants were reassured by the information presented during the initial education session on GINA which was newly signed by President Bush at the time of the focus groups in 2008. For example, one woman commented: “I think the legislation that the President signed was a good idea, especially considering that as genetic testing becomes more and more prevalent, there are going to be some unintended consequences.” In contrast, several others remained skeptical that these protections were enough, saying:

“I understand that there has been legislation signed, that piece of paper, but that being said, I don’t feel confident that that information doesn’t fall into the wrong hands and by that I mean insurance companies and places of employment.”

A number of participants expressed concerns about privacy issues and noted mistrust related to whether genetic information could be kept out of the hands of health insurance companies:

“I’m very torn with the idea of genetic testing just because I think sometimes the information that you give out today, it’s out there and sometimes you don’t know where it’s going to go, who’s going to use it, to what end and if this information is going to actually benefit you or someone else….who’s going to pay your insurance premiums if they find out that you have this kind of illness and if your children have this kind of illness are they going to want to cover them? It becomes a cost-benefit thing for me. Is it really worth it to get tested?”

Lack of Interest due to Small Change in Risk / Focus on One Disease

One woman commented about how the small increase in risk typically conveyed by SNPs would not be worth the emotional costs to her:

“I think the 1% seems like too little of a difference to make it worth the agony. … There’s not enough incentive to put myself through something like that. If it was a higher percentage—probably, definitely, close to sure based on all these tests we have done—that you will know how likely you are to get this. I would probably reconsider it. But for 1.5%, 1 %?”

Likewise, another woman commented, “Well, if it went from 6% to 20%, I would be a little more concerned. But 1.5%? I mean, I guess that is significant. But I don’t know that I’d be losing sleep over that.” This same participant went on to note that she may not be interested in SNP testing for only one disease at a time: “…if you’re just going to choose one test, it doesn’t make any sense for me. If you do a whole screening, that makes more sense to me.”

Clinical Utility of SNP Results

A few participants expressed doubt about how useful this type of testing might be, indicating uncertainty about whether SNP results would impact their care: “I felt like it depended on my specific case and how that information would be used in either a treatment or monitoring plan. So that was why I wasn’t sure about how I felt about it.” Similarly, having been informed about the limitations of SNP testing, some participants questioned its clinical utility, especially for individuals who are already following health guidelines: “What is the reason for telling them [SNP results] if you already know that getting a colonoscopy, eating right, maintaining a healthy weight, exercising...if you’re doing all of the cancer prevention anyway?”

Another participant noted that risk information associated with SNPs should be given only in the context of information about specific behavioral responses that stem directly from receipt of the SNP risk information.

“If you’re going to give them a test and as a result of that test you say, ‘This test tells you are at risk for colon cancer. Here are the things you need to do over the next five years or whatever—the rest of your life—to mitigate that risk.’ But telling them that they’re at risk should only be the first part of the sentence. The second part of the sentence ought to be ‘here’s what you need to do.’ If you don’t know what to do—if you can’t answer the second part of the sentence—don’t give them the first part.”

Similarly, a different participant noted, “What would you do with the information [from SNP test results]? That’s the key.”

Anticipated Behavioral and Psychosocial Outcomes Following SNP Testing

When asked to consider the possible implications of SNP testing, participants highlighted both positive and negative potential outcomes. Positive outcomes included increased motivation for making changes in lifestyle habits like diet and exercise and gaining additional information to help tailor health screenings. Negative outcomes cited by participants largely centered on potential negative emotional responses, including increased anxiety or worry.

Impact on Health Behaviors and Screening

Participants expressed mixed views on whether SNP test results would meaningfully change their health habits or screening behaviors. Individuals who felt they were already doing everything they could to help prevent cancer or detect it early through screening, appeared less interested in SNP testing. In contrast, people who perceived themselves to be at increased risk due to lifestyle factors indicated that SNP test results would either help motivate them to engage in healthier behaviors or influence their cancer screening frequency. In discussing why she indicated interest in SNP testing for specific diseases, one woman noted:

“I checked heart disease and diabetes for the main reason that I thought that there was a behavioral aspect to it. Maybe those two conditions, there’s certain things I could do that would impact how severe of a condition I would have. So, knowing that might push me towards exercising a little bit more, eating better, those types of things.”

Another participant considered the impact of information from SNP testing for colon cancer within the context of other risk factors like her family history of colorectal cancer, remarking that it might impact her screening schedule: “I think my risk is higher since I already have some other things that indicate that. It [SNP test results] might be information that would allow me to have more colonoscopies—more frequently than 5 years.”

Emotional Responses

Participants noted a range of potential emotional responses – disinterest, relief from knowing about risk, greater worry. Anticipated emotional responses appeared to differ between people who had been diagnosed with cancer and those who had not been personally affected by cancer.

Perspectives from cancer survivors

Participants with a personal history of cancer (n=5) expressed either disinterest or anticipated worry in response to SNP results which indicated increased risk for disease. With regard to SNP testing for colon cancer risk, one survivor noted, “From a lay perspective, I don’t think I’m interested in it because I’ve already had colon cancer.” For this participant, SNP risk information about colon cancer did not have any added value. Another cancer survivor noted how managing the uncertainty related to SNP risk information may not be worth the additional worry:

“Having a family history and a personal history with cancer, it becomes a question of how much is this going to cost me, how much worrying will I do based on what these results are. I already know that I’m at an increased risk of having something. So to me, it’s more [a question of] maybe I have a risk here for something else, do I really want to know that?”

Similarly, another survivor noted how additional risk information may not be worth it personally, but recognized how people who are information seekers may want to know as much as possible about their risk, even if the information is uncertain.

“I have to get up and feel positive about the rest of my life as a cancer survivor. And I have had cancer twice so I have to make sure that my mental health is such that I can go forward and not feel duped. So that’s why I say, it’s not right for everybody. For people who want to get their hands on every single piece of information and think it makes a difference, yeah, it’s right for them.”

Perspectives from people without a personal cancer history

Most, but not all, participants personally unaffected by cancer appeared less concerned about feeling worried in response to SNP results. For example, one unaffected participant noted how she might respond emotionally: “You’re definitely going to think about it [SNP test results] because once someone tells you something and it registers in your brain, you’re going to think about it. But to worry about it…well, that’s something different.”

Another participant with a stated interest in knowing as much as possible about his health noted, “Worrying won’t do any good. But I think that there’s a significant benefit to knowing something certain. I don’t think there’s anything worth worrying about—it won’t do you any good.” In contrast, one woman personally unaffected with cancer but who had a family member recently diagnosed with cancer was not interested in testing because the emotional costs might be too high for her: “Could I deal with the possibility of chance…of ignorance is bliss? So, that side of me [not wanting to know] usually prevails.”

Preferences for Information Delivery

The majority of participants expressed interest in receiving information about SNP testing in-person from a genetics professional or their physician. Participants were able to discern the differences in specialty training between genetics professionals and their physicians. For example, one participant commented,

“I would want to get the information from the best source. ….If it’s a doctor who specializes in genetics, that might make a difference. But, I really feel as though a person who has the Ph.D. in genetics would know a lot more than the average doctor.”

One participant noted that he preferred that a genetic counselor deliver information before the genetic test and a physician deliver the results:

“I had in person [as preference for information about SNP testing]. …For before [testing], I wanted to go to a genetic counselor—that was my first choice. My doctor was my second choice. I was just thinking that I would want my doctor involved in that process because the genetic counselor is not going to treat me for anything—my doctor is. So I want them aware of what was going on and talk about that with them.”

Another woman echoed the importance of getting genetic testing information in person: “I think everyone would prefer to have the person talking with them.” She also noted that she was unfamiliar with the genetic counseling profession but would be open to the idea if given a physician referral.

A few participants stated preferences for having the information delivered to them through print materials. For example, one woman commented: “I like to read about things first and digest the information at my own pace.” She elaborated further about her preferences, noting, “I think reading first, for me at least, and then through a genetic counselor because I feel like that person would have more experience than my physician.” Another woman also preferred print materials, noting, “I prefer printed [material] because I can understand better than through my doctor.”

DISCUSSION

The purpose of this study was to explore primary care patients’ attitudes about predictive SNP testing, examine factors that influence testing interest, and explore preferences for information delivery. Most participants expressed interest in testing for SNPs related to increased risk of disease. Interest in testing was stronger among two subgroups of people: information seekers—people who wanted to have as much information as possible about their health—and people who had not been personally affected by cancer. Information seekers in the present study appeared to value risk information even when that information is uncertain, limited in scope, or without actionable risk management strategies, characteristics similar to early adopters of technology (Rogers, 2003). These findings are consistent with a recent report about motivation for whole genome sequencing in which many participants wanted information about disease risk, regardless of uncertainty (Facio et al., 2011).

Although the number of individuals affected with cancer in our sample was small, opinions within this group were largely consistent in a lack of interest in SNP testing for cancer risk. This disinterest may be attributable to both cognitive and emotional factors. Cognitively, affected participants appeared to understand the limited added value that SNP risk information has on their future risk. Emotionally, affected individuals’ lack of interest in testing may represent specific coping mechanisms, such as avoidance of unnecessary information that might exacerbate worry about disease recurrence or development of a second primary cancer.

Of note, participants’ ratings of interest in SNP testing did not differ after explanation of the level of risk conveyed and limitations to testing. This finding highlights a key concern of those in the medical and genomics communities that interest in SNP testing is driven by unrealistically high expectations of the value of genomic information (Evans, Burke, & Khoury, 2010). Whether this cognitive dissonance is due to beliefs in genetic exceptionalism (Evans et al., 2010), difficulty interpreting complex and clinically ambiguous information, true tolerance for obtaining uncertain risk information, or some other bias is unclear. Unfortunately, with the ready and likely on-going availability of commercial SNP testing for various disease risks (Borry, Cornel, & Howard, 2010), misunderstandings about the limitations and risks of SNP testing may lead people to make uninformed decisions about testing.

A key factor that seemed to drive interest in SNP testing included gathering information for family members; this finding parallels reasons for seeking BRCA1/2 genetic testing (Hallowell et al., 2005) and whole genome sequencing (Facio et al., 2011). Among participants who expressed interest in SNP testing to obtain information for family, the level of risk predicted by SNP results and/or whether the information could be used as the impetus for behavior change seemed the most salient factors. Participants appeared hopeful that this type of genomic information could improve the health of their loved ones – anticipating that genetic results might motivate family members to engage in screening or adopt healthier lifestyle behaviors. Participants’ associations between SNPs and family history appeared driven by erroneous ideas that SNP testing was highly predictive of risk or had specific implications related to screening or health behavior change like diet and exercise. Perhaps this hope for family members is an indirect form of “therapeutic misconception,” or the idea that direct clinical benefit is possible, even when provided with evidence to the contrary (Facio et al., 2011; Klitzman, 2010; Sulmasy et al., 2010).

People who expressed less interest in SNP testing questioned whether results were actually meaningful relative to the small increases or decreases in risk. These results are similar to ideas expressed by genetics professionals who received results from whole genome scans as part of a research study (O'Daniel, 2010). Unique to the present study was our finding that some participants expressed less interest in SNP testing when the test was offered in isolation for one specific disease rather than as part of multiplex testing for a variety of disease risks. The element of choice—which and how many specific diseases to seek genomic risk information about at one time—is a topic that has begun to emerge (Facio et al., 2011) but deserves additional attention in future translational genomics research.

Most people preferred that genetic information be delivered in person by a healthcare or genetics professional rather than through print materials or a computer. Primary care practitioners likely will not have the time or expertise to appropriately convey predictive risk information (McGuire et al., 2009; Rafi et al., 2009); consequently, development of alternate models to deliver genetic and genomic information is important. Examples of alternative models include specialized training for nurse practitioners (Myers et al., 2011) or inclusion of genetics specialists within primary care settings (Battista, Blancquaert, Laberge, van Schendel, & Leduc, 2012). Although delivering individual SNP-level genomic results through genetic counselors is not tenable or likely even necessary at the population/public health level (O'Daniel, 2010; Pagon, 2002), arguments for including genetic counselors in the process include their unique ability to convey complex risk information, accurate assessment of family history, and provision of expert guidance in certain situations like whole genome sequencing (O'Daniel, 2010). In contrast to the preferences expressed by most of the participants in the present study, a number of DTC testing companies deliver SNP results via the internet without routine involvement of a health care professional, although this model may be changing (Borry et al., 2010).

Many people expressed concerns about whether genetic information could be kept confidential, despite the passage of GINA. Concerns over privacy of genetic information are similar to unease voiced in reports published prior to GINA (Rose, Peters, Shea, & Armstrong, 2005). In the present study, people were skeptical that medical institutions would be able to keep genetic information away from health and life insurance companies, even after discussion of the protections afforded by law. The distinctly private nature of genetic information and consequent worry over potential stigmatization or discrimination highlight the continued relevancy of ethical and legal issues surrounding future clinical applications of genomic tests.

A few participants said that SNP results might cause increased worry or anxiety, particularly because no current recommendations exist regarding risk management. People who expressed anticipatory worry related to SNP results appeared less interested in testing. Within the BRCA1/2 literature, pre-test anxiety or distress is frequently predictive of longer-term distress, regardless of whether the test result was positive or uninformative (Ertmanski et al., 2009; Graves et al., 2012). If SNP testing is ever incorporated into routine clinical practice, then attention to pre-testing levels of distress may be useful to help people make fully informed decisions about testing. Recent findings from a study of psychosocial and behavioral outcomes following DTC genomic testing among self-selected participants indicate that anxiety levels did not change from pre- to post-test and the majority of participants reported no test-related distress (Bloss, Schork, & Topol, 2011). Future work can explore distress outcomes following genomic testing among more diverse and/or population-based groups of participants.

Our study is among the first to explore primary care patients’ attitudes about SNP testing and, to our knowledge, the first to evaluate whether interest in SNP testing changes after a detailed discussion about its meaning, benefits and limitations. Of note, we did not find any differences in reported interest after the group discussion. Perhaps the consistency in participant responses stems from inherent and hard-to-change beliefs about personal disease risk or comfort with risk-taking among early adopters of new technologies (Rogers, 2003). Although our current sample was small, it was diverse with regard to age, race, ethnicity and cancer affected status. Future studies can examine whether the present results generalize to larger samples.

Results from this study should be interpreted within the context of the study limitations. Participants’ expression of interest in testing was done following presentation of hypothetical scenarios. Reported interest in testing overestimates true test uptake for genome tests (Sanderson, O'Neill, Bastian, Bepler, & McBride, 2010); however, participants’ discussion of the factors that influence interest are likely representative of their actual opinions and attitudes. Moreover, individuals who agreed to participate in a study of attitudes about genetic / genomic testing and cancer risk may not be representative of the general population and thus reported interest in testing may not be generalizable. We also assessed SNP testing interest primarily within the context of one specific disease. With rapid advancements in whole genome sequencing, exploring people’s opinions about bundled/multiplex genomic testing is a reasonable next step for future research. Finally, participants in our first focus group were aware that one of the group moderators was a genetic counselor; this knowledge could have swayed preferences for delivery of SNP-based genetic risk information, although we did not observe differences between the focus groups on this outcome.

Research Recommendations

Emerging research describes initial responses to SNP-based genomic risk information in highly-selected samples such as users of online genomic testing services (Bloss et al., 2011; Kaufman, Bollinger, Dvoskin, & Scott, 2012). Building on the results from the present study, future work can explore uptake of SNP-testing in broader populations such as primary care patients and examine the longer-term psychosocial and behavioral outcomes following receipt of SNP-based predictive risk information for a variety of health conditions. For example, genomic feedback influences smoking quit attempts (Lerman et al., 1997; McBride et al., 2002) but not sustained cessation (Audrain et al., 1997; Carpenter et al., 2007; Ito et al., 2006). Lack of long-term change in prior work may stem from a failure to integrate transient increases in motivation into a behavioral intervention. The potential of genomic feedback to influence cancer prevention outcomes beyond smoking cessation is largely unexplored. We do not yet know if genomic feedback will be useful for motivating a broader range of health behaviors, including cancer screening, physical activity or healthy nutrition (McBride et al., 2010).

Implications for Genetic Counseling Clinical Practice

Overall, study results underscore an apparent misunderstanding about the significance of the risk levels conferred by SNPs, despite the detailed discussion during the groups about the level and nature of the change in risk conferred by SNPs. Moreover, findings suggest that patients may look to genetic counselors, physicians or other healthcare professionals as gatekeepers and caretakers of predictive genomic risk information (Burke & Psaty, 2007). Thus, our study has two-tiered implications for the genetic counseling profession. First, genetics professionals must be prepared to integrate counseling and testing into the current infrastructure to address issues related to SNP and other forms of genomic testing. For example, for individuals seeking genetic counseling prior to SNP testing, a discussion about the clinical utility of such may be central to patient decision making about whether or not to pursue testing and how the information may impact outcomes, including health behaviors, satisfaction with decision-making, and psychological parameters such as worry. In addition, pre-test counseling provides an opportune time to provide risk assessment based on family history, which in many instances has a stronger bearing on risk assessment than SNP testing. During our focus groups, we did not have time to adequately explain and explore the aspects of family history that are central to risk assessment, and how the results of SNP testing do not factor in familial risks or other established risk factors for disease. Relatedly, the results of SNP testing, next generation sequencing panels, exome sequencing and/or whole genome sequencing may still be more useful to unaffected at-risk individuals if a relative affected with the condition in question to undergoes testing first. This is a well established paradigm in clinical genetics, and the results of our focus groups suggest that education and counseling about this issue will be needed. Further, patients may pursue genetic counseling after testing has been obtained through commercially companies (Kaufman et al., 2012; Kolor et al., 2009). This encounter also provides a valuable opportunity for genetic counselors to put test results in the context of established risk information (e.g., family history) and to address patients’ concerns about disease risk and management.

The second and more broad-based implication to the genetics community is that genetics professionals are well positioned to educate the public and healthcare providers about the potential benefits and limitations of predictive genomic information, and to address some of the ethical, legal and social issues that may be associated with the delivery of this type of risk information. Thus, the genetic counseling community can begin to build resources to respond to potential demand for educating health care providers and the general public (O'Daniel, 2010). The need for these educational resources and genetic counseling services will be particularly salient if future translational research efforts begin to document clinical utility of genomic information and/or as whole genome sequencing becomes more prevalent.

Appendix Note

Parts of the presentation were obtained from information on the website of the National Cancer Institute: http://www.cancer.gov. The specific website we used for portions of our focus group presentation was: http://nci.nih.gov/cancertopics/understandingcancer/geneticvariation.

Artwork originally created for the National Cancer Institute. Reprinted with permission of the artist, Jeanne Kelly. Copyright 2011.