Refining a Multifaceted Model of Perceived Utility of Genomic Sequencing Results

Effective integration of genomic medicine into routine clinical care requires an understanding of how patients value and act on the information received. Research on the utility of genomic sequencing in clinical settings has typically explored clinical and personal utility separately, without examining the potential for interaction and overlap between the two [1, 2]. Clinical utility captures the extent to which screening or diagnostic data (e.g., test results) inform clinical decision-making, such as adoption of preventative care or treatment recommendations [3]. Personal utility broadly addresses the “non-health” related value of receiving genetic test results, including the impact on a patient’s emotional state, knowledge, behavior, and social environment [2]. Recently, Smith et al. [4] integrated both clinical and personal utility into a multifaceted perceived utility model, resulting in five domains: clinical, emotional, behavioral, cognitive, and social – each with several subdomains. Here, we report a novel assessment of how well this multifaceted perceived utility model applies to a diverse, adult population aged 18–49 at risk for hereditary cancer and propose new considerations for the model.

The multifaceted perceived utility model drew on both existing literature and empirical research across the seven Clinical Sequencing Evidence-generating Research (CSER) consortium studies that encompass diverse but mostly pediatric populations and rare disease families, thus reflecting much of the literature to date on the utility of genomic sequencing [5]. Prior qualitative research in the pediatric setting has demonstrated that parents found utility in identifying genetic causes for diseases where prior testing could not because it could lead to personalized treatment, provide psychological relief or satisfaction of curiosity, and aid in planning for the future for both the child who received testing and other potential children [6, 7]. Research on adults with predisposition to risk but without a known genetic diagnosis is more limited [1, 2, 8]. For example, Shickh and colleagues demonstrated the clinical utility of exome and genome sequencing in an adult population with suspected hereditary conditions and prior negative or inconclusive results. They reported that the number of pathogenic or likely pathogenic variants identified in their sample was similar across patients with and without a cancer history [9]. Likewise, Kohler and colleagues demonstrated the personal utility of exome sequencing among healthy adults between 45 and 65 years of age. Using a Delphi approach, they identified the most agreed-upon domains as knowledge of their condition, enhancement of coping, improvements in self-knowledge, feeling good for helping others, and mental preparation for the future [8]. In a study of average-risk adults, participants reported utility of exome sequencing, noting feelings of relief, validation, closure, and even excitement, with a few mentioning an intention to make lifestyle changes, and all reporting they had shared their results with family or friends [10].

Furthermore, the multifaceted perceived utility model, like the existing literature that focuses on either clinical or personal utility, does not account for the potential interaction and overlap between the two. Given that a person’s ability to act on their results could influence multiple aspects of their life (e.g., ability to cope, outlook on health, how they prioritize major life decisions) and vice versa, understanding how patients conceptualize utility, while accounting for interaction and overlap between heuristic domains, could contribute to improved results disclosure processes and the systems that connect patients to follow-up care and other downstream services.

The purpose of the analysis in this paper is to evaluate the relevance and completeness of Smith et al.’s [4] multifaceted perceived utility model, developed primarily with pediatric and rare genetic disease populations, for a diverse adult population at risk for hereditary cancer who received clinical genomic sequencing results in the CSER Cancer Health Assessments Reaching Many (CHARM) study. While the Smith model was developed using data from ten participants in the CHARM study, those participants represented a small portion of the total study sample, leaving room for more in-depth investigation into the model’s relevance. We examine adult patients’ perceptions of the utility of genomic sequencing results and analyze how utility in the CHARM study mapped to the multifaceted perceived utility model, noting which domains were represented and which were absent, which domains were most salient to our study population, and refinements that would make the model more robust and appropriate for a younger (age 18–49) adult population including those who have been historically underrepresented in clinical research and may have limited resources.

The overall goal of the CHARM study was to overcome barriers to cancer genetic services among racially and ethnically diverse and historically medically underserved adult populations by evaluating key laboratory, clinical, and behavioral interventions in primary care settings [11]. CHARM recruited English- and Spanish-speaking participants aged 18–49 from primary care clinics at Kaiser Permanente Northwest (KPNW), an integrated health system in Oregon and southwest WA, and Denver Health (DH), an integrated safety-net health system in Denver, Colorado, between August 2018 and March 2020. Participants completed an electronic patient-facing family history risk assessment for Lynch syndrome and hereditary breast and ovarian cancer syndrome [12, 13]. Those at increased risk for these syndromes or with limited family history information were invited to participate in this multi-intervention study. Participants submitted a saliva sample for exome sequencing and were tested for cancer risk variants, optional medically actionable secondary findings, and optional carrier findings. Demographic characteristics were collected via survey [11].

Open-ended semi-structured qualitative interviews were conducted by L.R., C.G., M.C., and G.J. with a group of CHARM participants within 1 month of results disclosure; a subset of these participants (primarily those who had clinically significant findings and responded to request for a follow-up interview) were interviewed again at 6 months post-results disclosure until data saturation was reached [14]. These timepoints aligned with survey timepoints in the overall CHARM study [11]. The interviews were designed to be open-ended (ethnographic) explorations of individuals’ experiences with sequencing in the CHARM study to elucidate the workings of a communication intervention [15]. As such, we explored perceptions of counseling communication and interest in and ability to engage in the counseling process. Interview participants were distinct from the ten CHARM participants interviewed to develop the Smith model [4], and the interview guide was novel. Thus, our interviews did not specifically target perceived utility as the Smith interviews did; nevertheless, in our coding process, utility emerged as an important theme. Questions about the following topics provided insights into utility: general approach to health and healthcare; familiarity and prior experience with genetics; perception of genetic counseling communication; preparation for and understanding of results; and attitudes toward recommended care and follow-up with primary care or specialist providers (see online supplementary material for interview guides used at each timepoint; for all online suppl. material, see https://doi.org/10.1159/000531782). We used purposive sampling to oversample for Spanish-speaking participants and those who received pathogenic or likely pathogenic results to better represent the diversity of the overall CHARM cohort and test results that included enough pathogenic results to reach saturation (only 1–2% of the CHARM participants received pathogenic variants).

Descriptive statistics were used to detail demographic characteristics. Interviews were audio recorded, transcribed verbatim, and analyzed using a modified grounded theory approach [16]. Our research team collectively developed a codebook based on initial review and open coding of the 1-month interview transcripts and further refined it with subsequent coding. Codes were added upon review of the 6-month interview transcripts to account for additional topics discussed, such as communication with family and cascade testing. Each transcript was coded by two researchers (D.D., L.R., C.G., M.C., and H.L.) using Dedoose qualitative data analysis software to ensure consistent application of the codes [17]. Discrepancies were resolved through consensus. Coded data and emerging themes were discussed with the full analysis team. Initial analyses indicated substantial (direct and indirect) discussions of utility, broadly defined. Therefore, we conducted additional reviews of coded data where views about utility were expressed to analyze how participants conceptualized utility and how our data aligned or did not align with the Smith framework. Our analysis also examined differences based on result type (pathogenic or likely pathogenic), carrier (unrelated to cancer), variant of uncertain significance (VUS) in a gene related to cancer risk, and normal results (negative for all categories selected by participant), age (comparing participants older and younger than the approximate mean age of 35, which was used as a proxy for whether they would be at or near screening age), and healthcare and other resources available to participants.

Thirty-three CHARM participants completed the 1-month interviews; 11 were interviewed again at 6 months. Fourteen received pathogenic or likely pathogenic cancer risk variants (ATM, BRCA1, BRCA2, CHEK2, FANCM, MSH6, MYBPC3, MUTYH) and other medically actionable findings (e.g., DSG2, TSC1); 8 received VUS results (ATM, BRCA2, CHD1, CHEK2, MSH6, PALB2, PMS2, POLE); 6 received normal (negative) results; and 2 received carrier results unrelated to cancer (GJB2, G6PD). Three participants had more than one result and were grouped based on their highest risk result (i.e., a participant with both pathogenic and VUS results was categorized as pathogenic). Demographic characteristics are displayed in Table 1. In the sections that follow, we describe how each of the five Smith et al. [4] domains are represented in our data and how our data demonstrates that these domains are overlapping rather than discrete categories.

Smith and colleagues characterized clinical utility as genomic sequencing results that are clinically useful for either the patient or the patient’s family. Five subdomains were identified in their framework, including diagnostic remarkability (establishing, confirming, or ruling out a diagnosis), appropriateness of follow-up care, informed clinical management, referral to clinical trial, and monitoring for early disease detection [4]. Our interview participants expressed that their results were valuable for informing the management of their clinical care, including determining follow-up care and monitoring for early disease detection. Participants consistently referenced the value of having their results in the medical record so that they could be quickly and easily shared with their primary care clinician. Younger participants appreciated having their results in their medical record, even when results would not be clinically actionable for several years. One transgender participant did not originally intend to discuss his pathogenic cancer result with his primary care provider, but later recognized that doing so would facilitate a pathway to risk-reducing surgeries, which could be viewed as preventative care in the event that gender-affirming surgeries would not be covered by insurance:

Another participant with a pathogenic variant for hereditary cancer found it reassuring to have a previous identification of Lynch syndrome confirmed by the genomic sequencing in CHARM. Several participants also conveyed the value of the genomic sequencing results in guiding their cancer screening and prevention procedures (e.g., mammography and colonoscopy); those with normal or VUS results mentioned their intention to continue with routine screening, and those with pathogenic or likely pathogenic results planned to pursue earlier or more frequent screening.

Several participants who had either pathogenic, likely pathogenic, or VUS results discussed how the results were important for helping family members identify a need for more frequent or regular screening or to gain access to cascade testing. When asked whether their family members received testing, one participant said, “Yeah, actually. All three of them have now gotten their results, so we know where the gene comes from.” (age 31, likely pathogenic result). Others noted that while their relatives planned to get tested, they hadn’t yet, with one saying, “I think the rest of my sisters probably will at some point… when they feel like it” (age 29, pathogenic result). Those with pathogenic or likely pathogenic results who had children too young to be tested understood the risks for their children and planned to inform them and their pediatricians – and encourage genetic testing – once they were older.

Smith and colleagues defined the cognitive utility of genomic sequencing results as the value the information holds apart from its connection to any specific recommendations for follow-up care. Their framework identified three subdomains, including value of knowing information, perceived health risk, and satisfaction of curiosity [4]. Our participants discussed the value of knowing the information, both in terms of learning about any health risk or lack thereof, or simply satisfying their curiosity. Participants generally described the value of having the information in and of itself, seeing it as important and interesting. While these views were shared across participants with various result types, they were most common among those with pathogenic, likely pathogenic, and/or VUS results. Notably, these views were also somewhat more widely held by participants ages 35 and over compared to those under 35. Some specifically pointed to the importance of having a clear explanation of the results and what to do next. For example, one participant noted:

A few participants with VUS results also mentioned how the results informed them of genetic markers that they should be aware of and look out for in the future, once research progressed. Across all groups, participants also spoke about how their results satisfied their curiosity about their risk specifically or about the science generally, with comments such as “Well, knowing that it runs on my mother’s side of the family, you know, I was curious” (age 42, pathogenic result) or “I have always felt like I first wanted to get the genetic testing for the sake of science. Aren’t you curious about this?” (age 33, pathogenic result).

Most participants found value in learning about their risk since they had undergone testing due to a family history of cancer or an unknown family history. As one participant said,

Following result disclosure, most participants with normal or VUS results acknowledged that they were still at risk (some at average risk, some at higher than average risk) due to various factors such as family history or environmental exposures. Others noted that they appreciated having a clearer picture of their overall health and the possibility of being more prone to certain kinds of cancer.

In the model developed by Smith and colleagues, behavioral utility provides value in terms of modifying non-clinical behaviors or preparing for the future. They identified six subdomains, including insurance coverage, health habits (diet, exercise, smoking, substance use), information seeking, future planning (estate, financial, career choices), parenting decisions, and reproductive decision-making [4]. Our participants found value in how the results influenced their health habits, parenting decisions, and information seeking. However, some participants acknowledged that access to resources such as health insurance presented barriers which hindered their ability to act on the information. Participants from each result group mentioned that they would make lifestyle changes or take better care of themselves based on their results. A participant with VUS results commented that it was valuable “just to know that I need to continue taking care of myself and my body and knowing that whatever I do now can have consequences later on in my life” (age 34, VUS result).

Several participants felt the results were important for understanding possible actions to take regarding their children’s health. These participants had pathogenic, likely pathogenic, and VUS results. An older participant with a VUS in a colon cancer-related gene and whose father had colon cancer seemed to consider it a positive result, saying, “When my son gets older, I’m going to share with him and let him know this is what I have and get him a genetic test.” Another participant with a VUS felt relieved she wouldn’t pass breast cancer genes to her daughters; in contrast, another said the “unclear” result probably wouldn’t be very useful information for her family members. Those with negative results acknowledged that their children may still have other risk factors (including if the other parent was a carrier).

A few participants with pathogenic and carrier results talked about how acting on their results would depend on their insurance status. A younger participant with a pathogenic result in a breast cancer gene said she would have to wait for insurance before getting future examinations done, saying, “The lump that I’ve had in my boob for the last however long, when my insurance kicks in tomorrow, I’ll make an appointment to check it.” (age 28, pathogenic result) Another participant noted potential barriers to getting recommended follow-up care, explaining,

Smith and colleagues described emotional utility as covering a range of emotions both positive and negative. Two subdomains were identified in their framework including adverse response (anxious feelings, confusion, depressive symptoms, disappointment, fear, frustration, guilt, sadness, worry) and positive response (empowerment, gratitude, hope, relief) [4]. When asked about how prepared they felt to receive their results, a few of our participants mentioned feeling a sense of restlessness or nervousness before the results were disclosed. Adverse responses to receiving the results included feelings of anxiety and shock. One participant with a pathogenic result remarked that they were hesitant to share their results with their family and needed more time to process.

However, several participants expressed feeling excited to receive their results. Our findings also indicate almost all participants noted a sense of ease or acceptance after receiving their results, despite experiencing adverse anticipatory emotions. A few participants across all groups felt grateful, happy, or good about getting the information. A younger participant with negative results said that the results provided peace of mind and alleviated anxiety about their health, while two participants over 35 with pathogenic and VUS results mentioned that having the information empowered them. Participants across all groups mentioned feeling some sense of relief and confidence to make more informed decisions about their health moving forward.

Participants also mentioned being appreciative of the way the information was delivered by the genetic counselor and how the process made them feel reassured and comforted. Some across all groups noted that at points in time they were worried – or questioned whether they should be more concerned – but ultimately felt good thanks to the calm disposition of the genetic counselor. In contrast to how some participants’ emotions changed from negative to positive after speaking with the genetic counselor, some participants’ emotions became more negative after learning their results. One participant with likely pathogenic results mentioned, “It was still a little bit shocking coming away from that, especially as my mind spiraled a little bit after we were off the phone, but as far as like a first point of contact, like someone to go over it with me, I felt very calm during the process.” (age 31, likely pathogenic result).

According to Smith and colleagues, social utility describes how genomic results influence people’s interactions with family and schools, their involvement in advocacy or support groups, and the quality of their relationship with care providers. They identified seven subdomains: (1) advocacy activities, (2) blame, (3) access to support services, (4) degree of social support, (5) discrimination (employment, schooling, insurance), (6) quality of relationship with care providers, and (7) social stigma [4].

Our participants discussed the social utility of their results less frequently than other forms of utility. However, a few mentioned concerns about financial resource limitations, insurance discrimination, their relationship with their primary care clinician, and their relative degree of social support from family. For instance, a younger participant with pathogenic results described that their decision to seek out preventative surgery was constrained by their role as the household’s sole provider. A few participants cited concerns with getting insurance or increasing rates based on their results. As one put it, “there’s a fear that… like if it’s on your record that people in the future could see you’re a high risk of cancer and aren’t going to want to insure you or that it’ll increase your insurance rates” (age 29, pathogenic result). Two younger participants with pathogenic results also commented on a desire to change primary care clinicians due to poor relationships, with one noting,

One participant acknowledged complications when sharing results with their family members. When asked if they spoke with their sibling, the participant responded, “No, well because he doesn’t know anything about my results, I don’t want to burden him… And I’m waiting because he can only get the test done when he turns 18” (age 21, pathogenic result).

Although it is not included in the Smith et al. [4] model of perceived utility, several participants spoke about the social value of supporting research for the benefit of others (e.g., generating new information, developing better or more accessible screening, etc.). This perspective was only shared by those with pathogenic, likely pathogenic, or VUS results and was more common among those 35 and over compared to those under 35.

Our findings demonstrate that the utility of genomic sequencing that participants perceived spanned the domains established by Smith and colleagues. While these domains can be used to organize perceptions of utility, they are not necessarily discrete. We found that when some participants discussed how they would use the results or what the results meant for them, they spoke in a way that broached multiple domains simultaneously. For our participants, the utility of genomic sequencing sometimes overlapped across domains, most commonly across the clinical, cognitive, and behavioral domains. For instance, one participant described being able to rule out a diagnosis (clinical utility) while simultaneously talking about the value of the information (cognitive utility) and the way the results impacted parenting decisions (behavioral utility):

Another participant spoke to the value of the information and their health risk (cognitive utility), how the results could inform their care in the future (clinical utility), and how the results did not cause undue concern (emotional utility):

Another said,

For this person, participating in the study and confirming their risk could directly influence their health habits and communication with their family, as well as contribute to the greater social good. One participant who spoke to the value of confirming their Lynch syndrome diagnosis (clinical utility) also said that their participation was motivated by their curiosity (cognitive utility):

Our findings demonstrate that the multifaceted perceived utility model developed by Smith and colleagues [4] is appropriate for a diverse adult population aged 18–49 at risk for hereditary cancer who underwent sequencing due to a family history of cancer or a lack of family history knowledge. Importantly, conducting research among this population raised two important issues that should be considered as additions to the Smith model: (1) an individual’s perception of utility may change over time, and (2) an individual’s perception of utility may be influenced by the resources available to them. Further, our findings suggest that the model should account for the evident overlap in domains of utility.

Overall, the Smith model serves as a useful tool for parsing the nuances of how our study population values sequencing results, and our application of the model to this population reinforces the need for a holistic approach to utility. Our findings demonstrate support for all domains of the model, although several subdomains that may relate more to pediatric contexts/patients (e.g., referral to clinical trials, advocacy activities, blame, access to support services, and social stigma) were not represented in our study population. For our participants, clinical utility was most relevant to those who received pathogenic or likely pathogenic results and those old enough to start screening. Younger participants and those with more limited access to care could anticipate clinical utility and potential barriers to follow-up care, respectively. While participants with normal or VUS results did not describe much clinical utility, they did speak to the value of their results in other domains. For instance, a few with VUS results discussed how their results could influence their health behaviors. Our findings also indicated various degrees of perceived utility for family members, with some swiftly acting on the participant’s results while others had not acted as of the 6-month follow-up interview.

Cognitive utility was widely applicable for our participants, with those who received all types of results noting the importance of having the information in and of itself, how their results shed more light on their family’s health history, and how they helped to satisfy curiosity. Behavioral utility was also widely applicable, with many mentioning their intention to make lifestyle changes or take better care of themselves based on their results. This aligns with findings from prior studies, including those that examined genetic and genomic testing in other contexts (e.g., direct-to-consumer studies) [18‒20]. Also notable in this domain was the importance participants placed on the results for their family’s health, especially as they pertained to reproductive and parenting decisions. Our findings provide additional support for the utility perceived by families whose children do not have access to family history information, such as adoptees or those whose biological parents were egg or sperm donors [21, 22].

Regarding emotional utility, our findings demonstrate the role of time in how participants value their results. Several of our participants discussed negative emotions in anticipation of receiving their results but later felt a sense of ease or acceptance after they had received them. This temporal element is deserving of additional research, particularly with younger populations, as one’s perception of their results could change over the life course. Lastly, discussion of social utility arose the least, with some participants discussing barriers to recommended follow-up care due to competing financial priorities and insurance concerns and others commenting on the role of their relationship with their provider. Though not included in the Smith et al. [4] model as a subdomain of social utility, some of our participants pointed to the value of research participation in benefitting others by generating new information or leading to the development of better or more accessible screening.

Insights from our interviews also suggest that additional efforts to direct patients to services beyond the standard clinical follow-ups, such as mental health or social services, could provide additional utility for those undergoing genome sequencing. For instance, receiving results could have repercussions for an individual’s mental health (e.g., emotional distress), particularly when results indicate that they or their family members are at higher risk of disease. Mental health counseling and practical support in accessing recommended follow-up care could decrease disparities in care and prove critical to the effective implementation of genomic sequencing in primary or other care settings that serve a broad population.

Related to this, our findings also indicate that an individual’s perception of the utility of their results may be influenced by the resources available to them to take recommended actions. Specifically, participants reported situations in which their perception of their results was influenced by having limited access to financial or healthcare resources. This was most clearly evidenced by the participants who spoke of needing additional screening (i.e., a mammogram) but not being able to afford the procedure, and one participant who mentioned needing time to pay off existing healthcare bills before getting follow-up care and incurring new charges. Researchers and clinicians have an ethical responsibility to not only inform participants about their risks but to enable them to act appropriately in response to them. Without the latter, participants may not be able to benefit or may be harmed from undergoing sequencing.

Additionally, while the concept of disutility was not discussed by our participants, further research is needed to explore the notion of disutility raised by Smith et al. [4] and others [23]. As efforts in genomic medicine aim to make access more equitable, understanding the perceived disutility of genomic sequencing could provide added insight into the barriers that could limit diversity such as privacy concerns and insurance discrimination. Future research should continue to investigate these aspects of utility to inform interventions that promote equitable care for all patients in genomic medicine.

Some limitations should be considered when interpreting our findings. Our study was relatively small and focused on an adult population aged 18–49 with either a family history of cancer or an unknown family history, so the implications of our findings may not be applicable to all adult populations, nor to pediatric populations. Given our finding that utility can change over time, especially for a younger population, our findings only capture a snapshot within the first 6 months post-results disclosure, and more robust follow-up would be important in future studies. Furthermore, because we largely focused on those with pathogenic results in the 6-month interviews, we did not fully explore the potential utility (or harm) of receiving normal or VUS results at that timepoint.

This study demonstrates the value of integrating multiple domains and subdomains into a single holistic framework that accounts for the various ways in which individuals perceive value in obtaining sequencing results. Our findings provide clear evidence that younger, unaffected adults value genomic sequencing in multiple ways. Further, better understanding of the various aspects of utility could improve services for the historically underserved populations who undergo sequencing to gain information related to hereditary cancer syndromes or other genetic conditions. More research is needed to assess how an individual’s perceptions of utility can change over time or due to circumstances such as socioeconomic status or insurance coverage. Our findings illustrate that utility is not fixed but rather can be dynamic and contingent on numerous factors. Overall, insights from this study may prove beneficial for developing approaches to genomic sequencing among unaffected adults that could incorporate clinical and non-clinical needs and improve access to follow-up care.

The authors acknowledge the efforts of the entire CHARM study team without whom this study would not have been possible. We would also like to acknowledge the invaluable input provided by our Patient Advisory Committees, our Patient Representatives, and all CHARM participants. We thank Jill Pope for editing and Christopher “Sam” Peterson for formatting the manuscript.

All participants provided written informed consent, and all procedures were approved by the Kaiser Permanente Northwest Institutional Review Board (Protocol #000733).

The authors have no conflicts of interest to declare.

This work was supported by a grant from the National Human Genome Research Institute (U01HG007292), with additional support from U24HG007307 (Coordinating Center). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conception of work: Devan M. Duenas, Leslie Riddle, Claudia Guerra, and Galen Joseph. Data collection: Leslie Riddle, Claudia Guerra, Mikaella Caruncho, and Galen Joseph. Data analysis: Devan M. Duenas, Leslie Riddle, Claudia Guerra, Mikaella Caruncho, Hannah Lewis, and Galen Joseph. Drafting of manuscript: Devan M. Duenas, Leslie Riddle, Claudia Guerra, Mikaella Caruncho, and Galen Joseph. Critical review and revision of manuscript: Devan M. Duenas, Leslie Riddle, Claudia Guerra, Mikaella Caruncho, Hannah Lewis, Kathryn M. Porter, Stephanie Kraft, Katherine P. Anderson, Barbara Biesecker, Marian J. Gilmore, Jamilyn Zepp, Michael C. Leo, Benjamin S. Wilfond, and Galen Joseph. Final approval of manuscript: Devan M. Duenas, Leslie Riddle, Kathryn M. Porter, Stephanie Kraft, Katherine P. Anderson, Barbara Biesecker, Marian J. Gilmore, Jamilyn Zepp, Michael C. Leo, Benjamin S. Wilfond, and Galen Joseph.

Data are not publicly available due to ethical reasons. Further inquiries can be directed to the corresponding author.