Health System-Led Early Consent and Direct Contact of At-Risk Relatives: Pilot Study Results Open Access

Hereditary breast and ovarian cancer syndrome (caused by variants in the BRCA1 or BRCA2 genes) and Lynch syndrome (caused by variants in the MLH1, MSH2, MSH6, or PMS2 genes) increase the risk for colorectal, endometrial, ovarian, and other cancers. When genetic testing finds a pathogenic variant in these genes in one individual (a proband), there are often immediate implications for the proband’s biological relatives, who may carry the same variant and be at increased risk of cancer. Cascade testing is the process of case finding within families by notifying at-risk relatives and inviting them to consider genetic testing. Current standard of care recommends that at-risk relatives of these probands be notified and referred for cascade testing for the identified variant because early intervention and surveillance for people with the variant can lead to improved morbidity and mortality [1]. However, cascade testing rates are low, representing missed opportunities to reduce the morbidity and mortality associated with hereditary cancers [2‒6]. Only half of relatives of people with Lynch syndrome receive genetic testing [7], while uptake of BRCA testing among eligible relatives is 20%–30% [8]. A 2022 systematic review and meta-analysis of 87 included studies and 33,223 at-risk relatives found that 48% (95% CI, 38–58) of at-risk relatives underwent cascade genetic counseling and 41% (95% CI, 34–48) underwent cascade genetic testing [9]. Further, across 34 studies (n = 11,711 at-risk relatives), 70% (95% CI, 60–78) had been informed of their risk [10], suggesting that more complete familial risk notification could increase at-risk relatives’ participation in cascade testing and potentially life-saving or risk-reducing interventions.

Health system-facilitated direct contact of at-risk relatives is a promising approach to improving at-risk relative notification and cascade testing and providing additional support to probands [11]. Studies outside the USA have found that direct contact interventions are acceptable to patients and families and notify more relatives than patient-led contact alone [12‒17]. We developed an intervention for health system-led direct contact of at-risk relatives in a US setting, informed by the preferences of patients and families. These preferences included early contact of at-risk relatives before the proband is tested, representing a novel approach that optimizes relative autonomy about receiving familial risk information [11]. This report describes the reach, acceptability, and limited efficacy of the program. Qualitative results have been reported separately [18].

We conducted a single-arm, prospective mixed-methods intervention pilot study at one US integrated health system’s medical genetics service. We evaluated the reach, acceptability, harms, and limited efficacy of a health system-led direct contact program at baseline and 6–8 weeks’ follow-up [19].

The intervention was based on our previous work, where we used qualitative methods to assess patient and family needs and preferences for direct contact intervention in a US setting. We conducted two phases of focus groups using interactive human-centered co-design methods. In the first phase, we assessed patient and family preferences for direct contact in focus groups with a diverse sample of patients and families. In the second phase, we assessed reactions to a prototype direct contact program using 1:1 interviews. The result of this two-phase process was the intervention presented here and a set of program design requirements that could apply to multiple US settings. These requirements included: using multiple communication channels and attempts to reach relatives; getting patients consent to contact their relatives and early consent of relatives before proband testing; applying a programmatic approach to direct contact to minimize workload of clinical teams; using direct contact to complement, not replace, patient-family communication; and allowing relatives to control information flow (e.g., respect their right not to know genetic information) [11].

Per these requirements, we designed this intervention to serve as a complement, not a replacement, to either usual care received by Kaiser Permanente Washington (KPWA) genetic counseling patients or patient-led family communication. We designed the process to include early consent of relatives before the proband testing to address strong preferences for respecting relatives’ right to know about potential risk and to make decisions about their preferences in the absence of a specific opportunity to be notified or invited to testing. We included gradual sharing of information with relatives (an initial letter followed by phone follow-up where more information was provided), to give relatives multiple chances to decline to learn information. The intervention complies with emerging interpretations of the HIPAA privacy rule in the context of genomic medicine, in which direct contact of at-risk relatives is acceptable with the patient’s consent [20].

The primary intervention components were (1) genetic counselor offer to contact at-risk relatives directly, (2) direct outreach to each at-risk relative requested by the enrolled proband, and (3) direct notification of proband test results to consenting relatives. The study genetic counselor (J.M.Z.) conducted all pretest counseling appointments and returned results to enrolled probands.

During genetic counseling appointments for eligible patients, the genetic counselor assessed personal and family cancer history and recommended genetic testing per KPWA clinical guidelines. If the proband was eligible and decided to pursue testing, the genetic counselor reviewed the three-generation family pedigree taken as part of the risk assessment and identified at-risk relatives who might benefit from cascade testing in the case of pathogenic or likely pathogenic results.

The genetic counselor then offered to contact each at-risk relative directly to share the proband’s results when available. For each relative for whom the proband requested direct notification, the genetic counselor recorded the relative’s name, relationship to the proband, contact information, and proband’s preference about whether to have their name mentioned in the relative’s invitation letter or to be referred to simply as “your relative.”

The study team informed the proband of relatives who declined to participate. For probands wishing to notify at-risk relatives themselves, the genetic counselor offered to provide a letter for probands to share with relatives. Since relatives joined the study before the proband’s test results were known, we chose to return all proband results (pathogenic/likely pathogenic, negative, or variants of uncertain significance [VUS]) to participating relatives, rather than only pathogenic or likely pathogenic results. We based our approach on usual clinical practice for returning results to probands. There was no limit on the genetic counselor’s number of attempts to reach relatives for the purpose of returning results.

To return pathogenic or likely pathogenic results to relatives, the genetic counselor contacted each consented relative by phone. For relatives who were KPWA members, the genetic counselor provided a referral to KPWA Genetic Services for counseling and testing. For non-KPWA members, the genetic counselor directed relatives to resources for pursuing genetic counseling and testing outside of KPWA.

To disclose VUS results, the genetic counselor also contacted consenting relatives by phone. During the phone call, the genetic counselor informed the relative of the finding, explained why clinical care should be based on personal and family history of cancer, and did not recommend that relatives be tested for VUS.

To disclose benign or likely benign test results to relatives, the genetic counselor sent a letter by mail with the results. The letter contained an offer to speak with the genetic counselor by phone. All results were returned to probands and relatives prior to the participant’s 6–8 week post-enrollment data collection survey.

Our population of interest was probands seeking genetic counseling for hereditary cancer risk assessment for possible hereditary breast or ovarian cancer or Lynch syndrome and their first- or second-degree relatives. Eligible probands were currently enrolled at Kaiser Permanente Washington (KPWA), an integrated health system in the US Pacific Northwest that serves approximately 700,000 members. We used administrative data to identify potentially eligible probands. Adult KPWA members with an upcoming scheduled appointment with KPWA Genetic Services for pretest genetic counseling for hereditary breast or ovarian cancer or Lynch syndrome were eligible for recruitment. Clinical genetics practice at KPWA is provided by certified genetic counselors and medical geneticists. Risk assessment was based on KPWA guidelines in place at the time of the study. Study genetic counseling activities took place between September 2021 and October 2022. KPWA care guidelines are based on NCCN guidelines for assessment of genetic or familial risk for breast, ovarian, and pancreatic cancer or colorectal cancer [21, 22].

Eligible relatives were first- or second-degree relatives of enrolled probands who were (a) identified by an enrolled proband who provided permission for the study to contact them; and (b) identified as potentially eligible for cascade testing by the study genetic counselor at the proband’s pretest counseling. Relatives did not have to be a KPWA member. Because of resource constraints for this pilot study, we did not include relatives living outside of the USA. First-degree relatives include biological parents, siblings, and children; and second-degree relatives include biological grandparents, grandchildren, uncles, aunts, nephews, nieces, and half-siblings.

Inclusion criteria for all participants included age 18 or older and able to complete the consent process and study surveys in English. We excluded individuals currently receiving hospice care.

To recruit probands, the study team sent each proband a mailed study information sheet, invitation letter, and opt-out information. A study team member (JE) then followed up by phone, explained the study, and provided a paper or electronic consent form to those interested in participating. For individuals declining to participate, the study team asked about reasons why and recorded answers in the study REDCap database. Recruitment was conducted between January 9, 2021 and January 10, 2022.

Since study recruitment happened before the genetic counseling appointment, if the proband and counselor decided not to pursue testing or testing was not indicated, the proband was still eligible to continue with all study data collection activities, including their views on the acceptability of direct contact programs. However, relative recruitment was not applicable for these probands.

To recruit relatives, the genetic counselor recorded the contact information for each at-risk relative who the proband wished the study team to contact. A study team member (J.J.E.) mailed a study information sheet and letter inviting the relative to participate in the study. The letter stated that a study team member will follow up by phone, and provided opt-out information. In the follow-up phone call, the study team member explained the study and provided a paper or electronic consent form to those interested. The study team member attempted to reach relatives up to 3 times for recruitment.

Outcomes of interest were reach, acceptability, and limited efficacy of the direct contact intervention. Data sources included electronic health records, genetic counseling visit notes, participant surveys, and qualitative interviews. At 2 time points (enrollment and 6–8 weeks later), participants completed a self-administered survey (surveys are in the online suppl. materials; for all online suppl. material, see https://doi.org/10.1159/000545404). Participants received a USD 20 incentive upon completion of the follow-up survey. As part of the follow-up survey conducted 6–8 weeks post-enrollment, we invited all probands and relatives to participate in a semistructured interview about their experience with the intervention. We contacted interested participants to schedule an interview; the results of these interviews are reported elsewhere [18].

We recorded all study data into a REDCap database, including data on reach and reasons for declining participation. The study genetic counselor also added notes and entered data into this same REDCap database.

Reach. Our measure of reach included the number of probands and relatives who responded to study recruitment, consented to participate, and reasons for refusal to participate. We also collected the number of probands who consented to the direct contact process, the number and type of relatives who the proband identified for direct contact, and the number of relatives successfully reached and enrolled in the study.

Acceptability. We measured intervention acceptability using both a follow-up survey conducted 6–8 weeks post-enrollment and semistructured interviews with a subset of participants who agreed to participate. The follow-up survey includes 10 acceptability items created de novo for this study based on Sekhon et al. [23] theoretical framework of acceptability [24] that included attitude toward receiving results; effort required to receive results; fit with values; perceived effectiveness/impact; self-efficacy; opportunity costs; and overall acceptability. We also included 5 items from Brehaut et al. [25] decision regret scale with wording adapted for clarity. For these items, response options are on a 5-point Likert scale ranging from “strongly disagree” to “strongly agree.” As an exploratory measure of potential harm and impact of direct contact on family communication, we also measured family dynamics both pre- and post-intervention. Both the baseline and follow-up surveys included 6 items on family communication from Bowen et al. [26], as well as 7 items on family cohesion and 8 items on family satisfaction from Olson’s Family Adaptability and Cohesion Evaluation Scale (FACES) [27].

Efficacy. Our primary efficacy outcome was the proportion of at-risk relatives notified. This outcome reflects our recognition of the importance of relatives’ autonomy in decision-making about pursuing testing. We also assessed relatives with whom the proband shared their genetic test results and the number and types of relatives who subsequently had genetic counseling and/or testing. For relatives only, we assessed whether the relative had or intended to have genetic testing after talking with the study genetic counselor, the relative’s genetic test results (if tested), and 8 items adapted from Seiffert et al. [28] about whether the relative took specific actions after talking with the study genetic counselor, such as contacting their doctor or changing their diet or exercise habits.

Demographics. Data on participant characteristics, including demographics, health history, family health history, and previous genetic testing, were assessed in the baseline survey.

Analyses were descriptive and exploratory, with the intent of preparing for a future, larger trial of direct contact interventions. We conducted descriptive univariate analyses in both probands and relatives. To assess any possible differences in acceptability or decision regret by specific population groups, we conducted bivariate analyses of acceptance of direct contact and acceptability outcomes. We also assessed the proband sample by whether the proband had accepted or declined the offer of direct contact of relatives. We reported the frequency of endorsing either agree or strongly agree to each item. We also assessed changes between family cohesion and satisfaction from baseline to follow up. We assessed the frequency of pathogenic or likely pathogenic findings and subsequent relative notification actions at the family level.

The program’s reach is shown in Figure 1. We approached 148 probands. Of these, 37% (n = 55) consented to the study. In all, 46 people (31%) actively declined to participate, and 4 were determined ineligible because they did not have any family members in the USA. We were unable to reach 43 people. The most common refusal reason was a belief that the program would not benefit the proband, either because relatives had already received genetic testing (n = 3); relatives were outside the USA or not English-speaking (n = 2); the proband had no relatives (n = 1) or believed relatives would not participate (n = 1); or did not want to involve relatives (n = 1). Ten people (7%) planned to talk to their relatives directly about their results. Eight probands (5%) declined due to burden associated with research participation. Two people (1%) indicated that health system direct contact of relatives was discordant with their values. Among study participants, response rates to the follow-up survey were 69% (38/55) for probands and 91% (39/43) for relatives. The most common proband-reported perceived barriers at baseline to informing relatives of their results were: lost touch with relatives (n = 20, 38%) and difficult family relationships (n = 14, 27%). Others included that information would not be useful to relatives (n = 11, 22%) and concern about upsetting relatives (n = 7, 13%). All 55 probands who consented to the study met with the study genetic counselor. Of these 55, 31 completed genetic testing. In 15 of those untested, the study genetic counselor determined that there was no indication for testing. Other reasons for not receiving testing included leaving KPWA before testing could be completed (n = 1), declining testing (n = 6), or not providing sample (n = 2) (Fig. 1).

Twenty-nine (53%) enrolled probands requested direct contact of one or more relatives and provided consent to contact a total of 101 relatives (mean 3.5 relatives per proband), 87% (n = 89) of whom were first-degree relatives. We attempted to contact all 101 relatives, and 44% (n = 45) consented. We were unable to contact 36% (n = 37). Eleven people (11%) started the consent process but we did not receive complete consent forms. Nine people (9%) declined participation.

Table 1 shows the demographic characteristics of the enrolled probands and relatives. Enrolled probands (n = 55) were mean age 58: 75% female (n = 41) and 88% White race (n = 45). The majority (69%, n = 37) were married; 58% (n = 31) reported college education or more. The enrolled proband population was similar in age, sex, and race/ethnicity both to the sampling frame and to the group of individuals who were not enrolled (data not shown). Nine study participants indicated LGBTQ+ status, but these are not shown in Table 1 for small cell size.

The relatives identified by probands were 88% first-degree relatives (44 siblings, 33 children, and 12 parents) and 42% male. The consenting relatives were also primarily first-degree relatives (87%, 17 siblings, 17 children, and 5 parents) and 64% female.

Enrolled relatives (n = 45) were similar to enrolled probands in most demographic characteristics, family history, and history of genetic testing. The relative sample was 64% female, similar to the proband sample (75%). Fewer relatives than probands reported personal cancer history (19% vs 40%). Eleven relatives (25%) were Kaiser Permanente members (KPWA or other KP region), and 23 (51%) resided in the same state (Washington) as the proband. Reasons for relative non-enrollment, where provided, included incompletion of written consent (n = 11), burden (n = 2), perceived effectiveness (n = 1), and values discordance (n = 1). No probands who requested direct contact asked for their name to be redacted during the direct contact process.

Probands who accepted versus declined direct contact were similar in age, sex, marital status, education, employment status, family size, family cancer history, and self-reported previous genetic or direct to consumer testing. However, probands who accepted direct contact were less likely to report excellent health than those who declined (45% vs 72%). Probands who accepted direct contact were less likely to report previous genetic or direct to consumer testing in family members (28% vs 73%) and were more likely to report personal cancer history than probands who declined direct contact (55% vs 22%).

Table 2 shows the acceptability and decision regret of probands by acceptance or decline of the direct contact offer, and for relatives who completed direct contact. Overall acceptability of the idea of the study and study participation were high for both groups of probands, though overall comfort with the study was higher in the group accepting direct contact (100% vs 71%, p = 0.014). Measures of burden and opportunity (talking with genetic counselor; participating made life easier; and reporting that study participation was worth the time) suggested low reported burden in both groups. Probands who accepted direct contact were statistically significantly more likely to endorse measures of perceived effectiveness (being in the study helped the proband or family; helped family members manage risk; relatives only learned of their risk because of the study) compared to those who declined.

Probands who accepted direct contact were significantly less likely to report decision regret about participating in the study than those who declined (summary measure 6.9 vs 23.0, p = 0.002; lower scores indicate lower regret). In particular, probands who accepted direct contact were more likely to report making the right decision (100% vs 71%, p = 0.014) and that they would make the same decision over again (100% vs 64%, p = 0.004). No participants in either group reported harm related to their decision. Acceptability findings were generally similar between relatives and probands. For example, 90% of relatives (n = 35) reported overall comfort with the idea of direct contact, low regrets, and no reports of harm. Sixty-three percent of relatives (n = 24) reported that being in the study would help them or family members manage their cancer risk. The most common behaviors reported after talking with the study genetic counselor were talking with biological family members (n = 31, 39%) and looking up information about cancer risk (n = 6, 13%).

Proband results-sharing behavior with relatives, friends, and primary care provider was similar across both accepter and decliner probands at the follow-up (Table 3). Accepters were more likely to have reported sharing genetic testing results with another provider and were more likely to report discussing cancer risk with their children than decliners. Genetic counseling and genetic testing in relatives were similar for accepters and decliner probands.

Measures of family cohesion and satisfaction were stable between baseline and follow-up (data not shown). For probands, the family cohesion summary score was 24.3 (SD 6.8) at baseline and 26.3 (SD 6.1) at follow-up. The proband family satisfaction summary score was 21.7 (6.4) at baseline and 23.8 (5.1) at follow-up. For relatives, family cohesion was 24.1 (7.9) at baseline and 26.0 (7.0) at follow-up and family satisfaction was 20.7 (6.5) at baseline and 22.7 (6.8) at follow-up.

Figure 2 shows the limited efficacy of the direct contact of the program as measured by the number of at-risk relatives notified and who reported testing. Of the 45 consented relatives, 43 completed contact with the study genetic counselor and received the proband’s test results (1.5 relatives per proband agreeing to direct contact). This included 28 mailed negative results and 7 VUS results returned via phone. The study genetic counselor returned 6 likely pathogenic or pathogenic findings to participating probands. In these 6 families, probands requested direct contact of a total of 21 relatives, 8 of whom had enrolled in the study. The study genetic counselor contacted these 8 people by phone and reported the pathogenic or likely pathogenic finding; the genetic counselor reached 6 of these and mailed certified letters to the remaining 2. At 6–8 weeks’ follow-up, 4 of these relatives had completed the study follow-up survey: 2 reported that they had received genetic testing, 1 reported that they were planning to have testing, and 1 reporting not receiving testing.

We conducted a nonrandomized, single-arm pilot study examining the reach, acceptability, and limited effectiveness of a health system-led program for offering direct contact of relatives for people receiving genetic testing for hereditary cancer risk assessment. Our study resulted in direct contact and notification of the genetic testing results of 29 probands provided to 43 relatives, including several pathogenic results.

We found high acceptability and low decision regret in both probands and relatives. The most common reasons for not requesting direct contact were about usefulness to the family or their own planned self-disclosure of results, consistent with our intent of having the intervention complement, not replace, family communication. Of note, probands who declined direct contact unexpectedly reported significantly more regret than probands who requested direct contact. This finding may further suggest acceptability of direct contact if probands who declined direct contact experienced decisional regret. Further, enrolled probands selected individual relatives for direct contact, suggesting thoughtful consideration about how direct contact might integrate into their own family communication plans. Overall, survey results indicated that the program was acceptable even for people who did not choose to use the program.

Our results are consistent with our previous work [11, 29] and with other studies examining the acceptability of direct contact programs. A qualitative study in Sweden found high acceptability among patients who had received a direct contact letter, and that families used direct contact as a complement to patient-mediated notification [30], often preferring direct contact as a complement rather than a replacement to familial communication. In a public survey conducted in Australia (n = 1030), the majority (68%) reported wanting to be contacted directly by a provider about genetic risk for conditions that can be prevented/treated early with privacy concerns limited to use of personal contact information. Almost 50% would prefer that a family member contacted them before the letter was sent, consistent with our study design [31]. Direct contact approaches have also been acceptable to patients with familial hypercholesterolemia in the USA [32].

We intentionally focused on capturing any harms associated with direct contact, based on the relatively new nature of US-based direct contact programs and the traditional focus of US healthcare on individual doctor-patient interactions. We noted very few values-based reasons for refusing study participation and no reports of harm. These prospective findings are consistent with our early work that found few privacy or other hypothetical concerns [29]. Further, measures of family cohesion and satisfaction did not indicate family stress associated with the direct contact program.

Our findings are consistent with other studies, suggesting that direct contact approaches can reach at-risk relatives and lead to cascade testing. A 2022 systematic review and meta-analysis synthesized the evidence for direct contact approaches for relative notification compared to patient-mediated approaches on uptake of relative genetic counseling and testing uptake. Direct relative contact approaches resulted in statistically significant higher uptake in relatives (63%; 95% CI, 49–75) than patient-mediated disclosure (35%; 95% CI, 24–48) across 17 studies, and similar benefit on relative genetic testing uptake. However, few studies including direct contact were conducted in the USA (3 studies of genetic counseling uptake and 7 studies of genetic testing uptake) [17]. The pooled estimate of relative uptake of genetic counseling was 68% (95% CI: 45–87) [9]. However, all three studies were limited to patients who had received a pathogenic variant, while in our study relatives consented to the study before knowing the proband’s result.

Two unique features of our program are worth noting. First, our study consented relatives before the proband’s testing and includes relative notification of all proband results, rather than limit relative outreach to those probands with pathogenic test results. This focus was consistent with our preliminary work, which found clear preference for relative outreach before the proband’s testing [11]. This pretest relative contact approaches optimizes relative right not to know and control of information: relatives had multiple opportunities to decline to hear more information: initial outreach, phone contact with study team, choice to engage with the study genetic counselor, and choice to decline results after talking with the genetic counselor. In our pilot, we identified people before their genetic counseling appointment, and many people were not in the end eligible for genetic testing. Future implementations could establish best practices for identifying probands who are known to be eligible for genetic testing. Also, an early consent approach involves outreaching to more families, which has resource implications. Second, because of the early consent design, we returned all results to participating relatives. Our qualitative interviews with program participants suggested that receiving negative results was meaningful for relatives [18]. We designed both of these features to be responsive to patient and family preferences, but they do differ from previous studies of direct contact, which have limited contact of relatives to families with known pathogenic variants. Both features also raise the question of resources when thinking of larger scale implementation. Future research can explore how to balance patient preference for early consent with logistical considerations, such as the workload required to identify and contact at-risk relatives. Formal valuation of clinical workflow impacts, such as hours of genetic counselor time, was out of scope for this pilot study. However, we designed and completed the intervention in close partnership with our care delivery partners, and the study team conducted all nonclinical activities, including initial outreach, consent, and outcome data collection. Systems considering implementing such programs may consider allocating nonclinical staff to support the program. Further research can assess the resource implications of direct contact programs, balancing the volume of genetic counseling work with the benefits of direct contact relative health outcomes achieved through increased cascade testing and intervention for those identified at increased risk for hereditary cancers.

Our pilot study contributes to the body of evidence establishing direct contact as a promising approach to improving cascade testing in US settings, adding an early estimate of the expected reach of such a program in a real-world setting. Our findings surpassed the criteria we established for progressing to a larger trial (minimum 25% consent rate among eligible probands approached; 75% or more respondents endorse either “agree” or “strongly agree” to the survey item: “I am comfortable with the idea of this study.”) [19]. We further contextualize these findings with our qualitative inquiry of program participants, which found that benefits of direct contact included reduced burden for probands, increased family discussions about health, and notification of relatives who otherwise would not have learned of the proband’s results [18]. We will use these data to plan sample size and population selection for larger future studies.

Our results also identify several areas for future research. Future research can explore how to balance patient preference for very early consent with logistical considerations, such as the workload required to identify and contact at-risk relatives.

We note several limitations. As a single-arm pilot study, the study lacked a comparison group, so we cannot directly compare rates of cascade testing or relative notification in the absence of an offer of direct contact. We conducted this study at a single site with a predominantly White, majority college-educated population, so cannot assume that our findings would apply to all health systems and population groups, particularly minoritized groups who may have different experiences with US healthcare. Due to delays related to the COVID-19 pandemic, we were only able to assess followup outcomes at 6–8 weeks, a short time frame to assess relative genetic testing behavior. Our scope was intentionally limited to patient and family experiences and outcomes. Though we worked closely with our clinical partners, two of whom are coauthors on this report (K.A.L., J.M.Z.), we did not systematically evaluate health care stakeholder acceptability or evaluate resource implications. Further, due to IRB requirements, potential participants were informed that they would be contacted by a genetic counselor if they joined the study, which would likely not occur outside of a research setting. It is likely that some relatives understood this, were not interested, and therefore did not respond to our attempts to reach them. As such, we are limited in our ability to assess fully reasons for declining participation. These limitations are balanced by several strengths, including our mixed methods approach, our patient preference-informed intervention, and the real-world US clinical setting.

In this pilot study, we found that a health system-led direct contact program designed to honor patient and family preferences and complement family communication is feasible, acceptable, reaches at-risk relatives, and communicated genetic risk. Future research could include larger trials of direct contact approaches in US settings with longer-term follow-up time to observe health outcomes; assessment of resource implications related to implementation and clinical workflows; studies that include individuals from historically minoritized groups or groups with limited access to health care and genetic counseling; and further investigation into identifying individuals or families who may be most likely to benefit from direct contact programs. Building the evidence base will assist health systems with assessing tradeoffs between allocating resources to direct contact programs and improved health outcomes of relatives.

The authors acknowledge Melissa Rabelhofer and Irma Gomez for assistance in preparing the manuscript for submission. We also acknowledge Catherine Lim, Brooks Tiffany, and Marlaine Figueroa Gray, whose engagement in the early phases of the project contributed to the development of the intervention. We also acknowledge and thank the study participants.

This study protocol was approved by the Kaiser Permanente Washington Institutional Review Board (KPWA IRB) on June 4, 2021 (Approval No. 1707328-10). All study participants provided written informed consent.

The authors have no conflicts of interest to declare.

This study was funded by the National Human Genome Research Institute (R01HG010144, Henrikson). The funder had no role in the design, data collection, data analysis, and reporting of this study.

Substantial contributions to the conception or design of the work: N.B.H., K.A.L., A.S., J.D.R., and S.M.F. Acquisition, analysis, or interpretation of data for the work: J.M.Z., J.J.E., P.R.B., N.B.H., M.L.A., A.S., and J.G. Drafting the manuscript: N.B.H., P.R.B., and M.L.A. Reviewing the manuscript critically for important intellectual content: all authors. Final approval of the version to be published: all authors.

The data that support the findings of this study are not publicly available due to privacy reasons but are available from the corresponding author upon reasonable request.