Investigating the Impact of Screen-Sharing Visual Aids during Genomic Results Disclosure via Telehealth in Diverse Families in the TeleKidSeq Pilot Study

Prior to the COVID-19 pandemic, studies on the use of telehealth (or, remote service delivery models, such as telephone or videoconferencing) in genetic counseling across the globe demonstrated that telehealth can improve patient access to genetics services by reducing costs and travel time, especially in rural communities with limited access to genetics care [1, 2]. Although genetic counselors have traditionally used telephones as an alternative to in-person visits for patient intake and genetic results disclosure, videoconferencing has become substantially more widespread in recent years [3, 4]. As a mode of delivery, video televisits have shown to be comparable to in-person genetic counseling visits in terms of patient satisfaction, knowledge, psychological support, and other patient-centered outcomes [1, 2, 5‒9].

Although the growth of digital health tools represents a profound opportunity to advance quality and equity in healthcare [10], health disparities among sociodemographic groups have persisted in the USA [11, 12]. While the use of video televisits in genetic counseling has been well studied in remote rural communities [1, 2], urban communities have not traditionally been the focus of telehealth research due to geographic proximity to large academic medical centers. However, it is important to study vulnerable urban communities as they may also be at risk of healthcare disparities stemming from the “digital divide,” or the unequal distribution of information and technology within society [13].

Schifeling et al. [14] observed that the digital divide significantly affects the geriatric population. Although video visits were associated with longer visits and more diagnoses compared to telephone visits, minority patients, Medicaid recipients, and patients for whom English is a second language were less likely to participate in video visits [14]. Sociodemographic factors that influence the differential uptake of telehealth include age, race, ethnicity, education, preferred language, insurance type, and household income [15‒21].

Video televisit platforms are also becoming more sophisticated, with new features available for use in virtual visits, such as “screen-sharing” to display visual aids. In this study, we defined screen-sharing as the ability to display visual content, such as genomic test reports and illustrations of genetic concepts, in real-time during a videoconference to support patient education and assist in explaining genomic test results. While validation of the effectiveness of visual aids in genetic counseling is evolving, a recent study by Pederson et al. [22] suggests that visual aids play a critical role in enhancing the genetic counseling process, particularly by improving knowledge of genetic concepts. While these additional features have the potential to improve communication of clinical test results, sociodemographic characteristics and other patient factors may influence patients’ experiences and must also be assessed to optimize their application.

Patient understanding is an important metric commonly used to evaluate the effectiveness of interventions aimed at enhancing genetic counseling practice [23‒25]. Understanding of genetic results is a complex and multidimensional construct comprising patients’ perceptions of their understanding and objective knowledge of their results [26]. To date, no studies have evaluated the use of screen-sharing and its impact on patients’ understanding of genome sequencing (GS) results and experience of receiving this information through telehealth, particularly among underrepresented and underserved communities who have historically been the last to benefit from advances in genetics research and technology [27].

We present the outcomes of TeleKidSeq, a pilot study that evaluated the impact of using video televisits to return GS results to parents/legal guardians of children with suspected genetic conditions across two health systems in New York City (NYC) [28]. TeleKidSeq was one of two studies in the NYCKidSeq program, a member-site of the CSER consortium [29]. We aimed to evaluate the use of video televisits and screen-sharing within a diverse, multilingual NYC community. The TeleKidSeq study was designed to compare the use of screen-sharing to no screen-sharing to disclose GS results through video televisits and evaluate its impact on parents’ perceived understanding of their child’s results and perceived confidence in explaining these results with others (primary outcomes). Based on prior studies on the use of educational tools incorporating visuals in genetic counseling [25, 30], we hypothesized that using screen-sharing to display visual aids and other materials including genomic reports would improve GS results disclosure experience. Additionally, TeleKidSeq assessed the effect of screen-sharing on parent’s objective understanding of results and their perceptions of satisfaction and usefulness of telehealth (secondary outcomes). We also investigated whether longer visit lengths added additional burden to provider workflow and on patient engagement.

TeleKidSeq, described previously by Sebastin et al. [28], provided GS and genetic counselors delivered results via telehealth to families who were randomized to either video televisits with screen-sharing (ScrS) or video televisits without screen-sharing (NScrS) groups. Ethics approval was obtained from the Institutional Review Boards at the Icahn School of Medicine at Mount Sinai (#20-01353) and the Albert Einstein College of Medicine (#2020-12292).

Children, 0–21 years of age, who had an undiagnosed neurological, immunologic, or cardiac condition with a suspected genetic etiology were referred to the study at two health systems, Mount Sinai Health System (MS) and Albert Einstein College of Medicine/Montefiore Medical Center (EM). Children were excluded if they had a molecular diagnosis, an obvious genetic diagnosis based on phenotype and clinical criteria, or a bone marrow transplant. At least one parent or legal guardian (participant) who spoke English or Spanish was required to attend all study visits and complete surveys. The study aimed to enroll at least 60% of families from underrepresented and underserved populations to achieve the CSER-wide objective of evaluating the utility of GS in ancestrally diverse and medically underserved groups [31, 32]. To ensure this objective was met, TeleKidSeq set a study-wide definition of diversity based on self-reported race and ethnicity (non-White/European American ancestry) and/or lived in a medically underserved area designated by the Health Resources and Services Administration (HRSA) [33].

Participants and children were recruited and enrolled into TeleKidSeq between December 2020 and July 2021, completing three study visits by May 31, 2022. Study visits included baseline enrollment and survey (BL), results disclosure genetic counseling and survey (ROR1), and 6-month follow-up survey (ROR2). The flow of participation and attrition from eligibility screening to study completion is summarized in Figure 1. Study teams at MS and EM assessed 830 children and participants for inclusion in the study, those eligible (N = 768) were approached for recruitment. As part of recruitment screening, participants were asked if they had access to a device with video-calling capabilities (smartphone, tablet, laptop, or computer), stable Wi-Fi connection or data plan, and a private location to complete the virtual study visits. If participants expressed any concerns, the study team facilitated the remote visit(s) at their respective health system using devices provided by the study. Ultimately, 452 participants were enrolled and randomized (ScrS, N = 224; NScrS, N = 228) based on institution (MS, EM) and primary indication (neurologic, cardiac, immunologic) via the REDCap randomization module [34, 35], and completed the BL survey. REDCap was used to facilitate electronic written informed consent provided by participants and children 18 years of age or older who were cognitively able. Of those who completed the BL survey, 432 attended the baseline pre-test counseling visit, and 409 children and available biological parent(s) provided specimens for diagnostic GS. As TeleKidSeq was designed to accommodate a fully remote genomic testing experience, saliva kits were offered to families for sample collection. If children were unable to provide a saliva sample, they were scheduled for a blood draw. Of these 409 children, 51 provided saliva and 358 provided blood samples.

GS was completed on 409 children and their available biological parental samples (95 singleton, 120 duo, 194 trio). The child’s genetic counselor received the diagnostic GS report, reviewed the findings and applied a case-level interpretation of positive, likely positive, uncertain, or negative. Details on the GS procedures and case-level interpretation have been previously described in Abul-Husn et al. [36] and Bonini et al. [37] and are briefly described in online supplementary Methods (for all online suppl. material, see https://doi.org/10.1159/000542444).

Following GS, 403 of 409 participants completed the results disclosure genetic counseling session via Zoom (a HIPAA-compliant videoconferencing platform). Subsequently, 396 of them completed the ROR1 follow-up survey within 1 month of their results disclosure visit, and 370 completed the ROR2 survey approximately 6 months later resulting in a 90.4% (370/409) completion rate for the study. Participants were compensated up to USD 80 for completing all study visits. Additional information on these study procedures can be found in Sebastin et al. [28].

All families met with an assigned genetic counselor to receive pre-test genetic counseling via Zoom to review the risks, benefits, and limitations of GS, including the option to receive secondary findings [28]. The study consent and procedures were also reviewed with participants and children during this session, and genetic counselors obtained informed consent for GS. The intervention (ScrS) or control (NScrS) took place during the results disclosure genetic counseling session. Participants in both arms received counseling that involved the genetic counselor reviewing the GS results and next steps for the child’s care. Depending on the result type, the counselor might also discuss the specific condition, inheritance pattern, implications for family members, and provide the family with resources. Additionally, the genetic counselor generated a personalized genomic test report for all children using the Genomic Understanding, Information, and Awareness (GUÍA) application [38].

During results disclosure video visits, genetic counselors had the option to utilize visual aids, including the GS report, personalized GUÍA report, and images conveying genetics concepts at their discretion. In the ScrS arm, genetic counselors could use the screen-sharing feature to display visual aids, whereas, in the NScrS arm, genetic counselors could show physical copies of these visual aids by raising these materials to their computer’s camera. After results disclosure, participants in both arms received a GUÍA PDF, a copy of the GS report, and any additional resources (such as condition-specific support websites or contact information for follow-up specialties) discussed with the family.

Survey measures were developed to collect study-related outcomes, demographics, and other participant characteristics of interest [28]. In addition to study-specific measures, TeleKidSeq surveys included measures harmonized across the CSER consortium to explore shared outcomes of interest, such as clinical utility, across the network [39]. Bilingual study coordinators administered surveys in the participants’ preferred language (English or Spanish) at the BL, ROR1, and ROR2 survey time points (shown in Fig. 1).

TeleKidSeq’s primary outcomes were participants’ perceived understanding of their child’s GS results and perceived confidence in explaining the results to others, which were captured on all participants at ROR1 and ROR2 using two Likert-style items developed for this study. Perceived understanding was measured by asking participants, “How much did you understand about the results that were given to you? Rate your understanding on a scale of 1 (“very little or none of it”) to 5 (“almost all of it”).” Similarly, for perceived confidence, participants were asked, “If you needed to explain your child’s genetic test results to someone else, how confident would you feel doing so? Answer on a scale of 1 (“not confident at all”) to 5 (“completely confident”)” (see online suppl. Table 1 for survey measures and time points).

Secondary outcome measures consisted of participants’ objective understanding of their child’s GS results and telehealth experience. Objective understanding was measured using four items developed for this study that were answered by participants at ROR1 and ROR2 and by the genetic counselors following results disclosure. We captured participants’ experience with telehealth at ROR1 using six items adapted from the 26-item Telemedicine Satisfaction and Usefulness Questionnaire (TSUQ) [40]. The “brief TSUQ” includes questions about satisfaction, ease of use, convenience, and privacy with participants responding on a 5-point Likert scale of 1 (“strongly disagree”) to 5 (“strongly agree”). Refer to online supplementary Table 1 for details on primary and secondary outcome measures.

Participant and child characteristics were collected at BL and included age, biological sex, relationship to the child, history with genetic testing/counseling, insurance type, education level, race and ethnicity, income and number of people supported (see online suppl. Table 2). Self-reported race and ethnicity (population group), a harmonized CSER measure adapted from the 2020 Census, allowed participants to select from ten options: American Indian, Native American, or Alaska Native; Asian; Black or African American (AA); Native Hawaiian/Pacific Islander; White or European American (EA); Middle Eastern or North African/Mediterranean, Hispanic/Latino(a) (H/L), “other,” “prefer not to answer,” and “none of these fully describe me” (see online suppl. Table 2) [39]. Population group responses were transformed for analysis with prioritization of H/L such that those who selected H/L were categorized as H/L regardless of any other population designations made. Participants that selected more than one population group, excluding H/L, were recategorized as “more than one race” and those that made only one population group selection remained as originally answered (see online suppl. Methods). Participants were also asked four questions that were adapted from the Computer-Email-Web (CEW) Fluency Scale [41] to capture details on their level of internet use, comfort with using the internet, satisfaction with their skills, and general comfort using computers (see online suppl. Table 2). Additional characteristics were collected by the study team including health system (MS, EM), survey language (English, Spanish), primary phenotype of the child (neurologic, cardiac, immunologic), case-level interpretation of the GS result (positive/likely positive, uncertain, negative), as well as duration of the results disclosure visit, location of participants during results disclosure (at home, on hospital campus, or other), type of device used (smartphone, computer/laptop/tablet), and if any technical issues were encountered during the session (see online suppl. Table 2).

Two analytic samples were separately assessed in this study (ROR1 and ROR2). Participants were included in analyses if they attended the results disclosure visit, completed either ROR1 only or both ROR1 and ROR2 surveys, and received the assigned study intervention. The ROR1 sample consists of 392 participants who answered the BL and ROR1 surveys and were present for the results disclosure visit (N = 17 excluded). The same criteria was applied for the ROR2 sample (N = 360), which consisted of participants who completed the BL, ROR1, and ROR2 survey (N = 32 excluded, see online suppl. Methods). Descriptive statistics were used to determine frequency of responses at BL, ROR1, and ROR2. Equivalency between NScrS and ScrS arms was evaluated using chi-square or Fisher’s exact test for categorical variables and t tests for continuous variables. All statistical analyses were performed using SAS software version 9.4 (SAS, Cary, NC).

Ordinal logistic regression examined the differences between arms for the primary outcomes of perceived understanding and confidence. The analysis controlled for parental education level, parental age, survey language, insurance type, genetic counselor, and case-level interpretation of GS results. For further details on definitions for case-level interpretation used for the study, refer to Abul-Husn et al. [36].

Regarding secondary outcomes, the assessment of overall objective understanding was calculated by totaling the number of matched responses between the participant and genetic counselor to the four binary questions, producing a summary score ranging 0 to 4 (refer to online suppl. Methods). The objective understanding instrument was previously described by Suckiel et al. [42] A higher summary score indicated better objective understanding. Poisson regression was used to examine the effect of the intervention on the summary score. Telehealth experience was evaluated by generating a summary score using the brief TSUQ instrument, with a higher score indicating greater satisfaction with the experience. The brief TSUQ produced a Cronbach’s alpha of 0.92, indicating high internal consistency among the questions. Multivariate linear regression was performed to assess whether brief TSUQ varied by arm at 3 months. Objective understanding and telehealth experience outcomes were analyzed with education level, parental age, survey language, insurance type, health system or genetic counselor, and case-level interpretation of GS results controlled for in their respective models.

A longitudinal repeated measures analysis was conducted to assess changes in participants’ perceived understanding and confidence, as well as objective understanding over time, using responses from both ROR1 and ROR2 totaling 752 participant observations: 392 from ROR1 and 360 from ROR2. The cumulative odds ratios (ORs) were generated via generalized estimating equations for correlated observations. Alternating logistic regression with a fully exchangeable working correlation structure was used to analyze perceived understanding and confidence. The summary score for objective understanding was modeled using Poisson regression with an unstructured covariance matrix.

Further exploratory analyses were conducted to assess if screen-sharing visual aids had a differential impact on participants’ perceived understanding, perceived confidence, and objective understanding. Repeated measures analyses were further stratified by the three largest population groups (AA [N = 57], EA [N = 92] and H/L [N = 189]), case-level interpretation (positive/likely positive [N = 73], uncertain [N = 165], negative [N = 154]), and device used by the participant at results disclosure (smartphone [N = 212]; and computer/laptop/tablet [N = 160]). In Suckiel et al. [38], significant differences in the effect of GUÍA on understanding outcomes were identified among H/L participants who did not use an interpreter. Based on these findings, we further stratified the H/L population by those who used a Spanish interpreter (N = 81) versus those that did not use an interpreter (N = 108) during results disclosure.

Additional analyses were conducted using a general linear model approach (SAS GLM procedure) to explore the mean differences in length of results disclosure visit and telehealth experience using the brief TSUQ score among randomization groups, select population groups, case-level interpretation, participant device used during disclosure, Spanish interpreter use, and technical issues encountered. Pearson correlation coefficient was used to assess the relationship between length of results disclosure and brief TSUQ score.

Out of the 409 children and participants who were enrolled in the TeleKidSeq study, 392 participants met criteria to be included in this analysis. The characteristics of the ROR1 participants and children, with 192 in the ScrS and 200 in the NScrS arm, are summarized in Table 1. The cohort was balanced across both arms with regard to demographics and other characteristics of interest. The mean age of participants was 40.5 years (range: 21.3–71.7), 85.1% were mothers (N = 334), and the largest population groups were H/L (N = 189, 48.2%), EA (N = 92, 23.5%), and AA (N = 57, 14.5%).

In the analytic sample, 17.1% (N = 67) of participants reported having less than high school education; 45.9% (N = 180) reported having completed a high school/GED, technical school, or associate degree; and 36.8% (N = 144) reported having completed college or postgraduate education. The majority of children (62.8%, N = 246) had public insurance (Medicaid), and 45.9% (N = 180) of families met or fell 200% below the federal poverty level for New York City. Approximately one-quarter of participants (27.3%, N = 107) reported having previous genetic testing or counseling for themselves or a family member. For 23.0% (N = 90), the ROR1 survey was conducted in Spanish. Upon completion of the GS, 18.4% (N = 72) of children had a positive or likely positive case-level interpretation, 42.1% (N = 165) had an uncertain interpretation, and 39.5% (N = 155) had a negative interpretation.

At baseline, almost all participants reported high levels of internet use. On average, 97.2% (N = 380) reported daily internet use. With regard to comfort with technology use, 84.7% (N = 331) of participants were at least somewhat comfortable with using computers, 90.3% (N = 354) were at least somewhat comfortable using the internet, and 88.0% (N = 344) were at least somewhat satisfied with their current skills for using the internet (Table 1).

During the results disclosure visit, 57.0% (N = 212) of participants used a smartphone to join the video televisits while 43.0% (N = 160) used a computer, laptop, or tablet. Data on the type of device used were not available for 20 participants. Most participants completed the visit remotely from their home (88.0%, N = 345) or in other settings (10.7%, N = 42), with eight having an on-site virtual visit (1.3%, N = 5). On average, these sessions lasted 28.7 min (standard deviation [SD]: 15.4, range: 6–100 min). Genetic counselors reported technical issues in 16.1% (N = 63) of participants’ results disclosure sessions. The majority of challenges reported by the genetic counselors involved participants experiencing difficulty in connecting to the session, logistical challenges in involving a Spanish interpreter, internet stability during the visit, and distractions from other members of the household.

The TeleKidSeq pilot study was designed to capture three dimensions of participants’ understanding: (1) perceived understanding of their child’s genomic test results; (2) perceived confidence in their ability to explain these results to others; and (3) their objective understanding of these results. In our cohort, 64.5% (N = 253) of participants reported that they understood “all or almost all” of their child’s results (level 5, possible range: 1–5) after results disclosure (ROR1; mean time from results disclosure: 0.16 days, range: 0–13 days). Approximately 6 months following results disclosure (ROR2; mean time from results disclosure = 6.4 months, range: 5.1–10.8 months), perceived understanding declined, with 49.4% (N = 178) selecting level 5. Similarly, with regard to perceived confidence, 53.7% (N = 210) of participants reported that they felt “completely confident” (level 5, possible range: 1–5) in their ability to explain their child’s results to someone else at ROR1 and 37.5% (N = 135) at ROR2. For objective understanding, over half of participants matched with the genetic counselor on all four questions at ROR1 (55.9%, N = 219) and ROR2 (51.8%, N = 186). The mean objective understanding summary score across all participants was 3.19 (SD = 1.10) at ROR1 (N = 392) and 3.06 (SD = 1.17) at ROR2 (N = 359).

Although genetic counselors had the option to use visuals in both the ScrS and NScrS arms (although not through screen-sharing in the NScrS arm), visuals were used more frequently in the ScrS arm: 99% of participants had visuals shared in the ScrS arm versus 56% in NScrS. We examined the impact of screen-sharing visual aids on primary outcomes of perceived understanding and confidence using ordinal logistic regression, controlling for education level, parental age, survey language, insurance type, genetic counselor/health system, and case-level interpretation of GS results. There was no significant association between screen-sharing and participants’ perceived understanding of their child’s GS results at either time point (ROR1, p = 0.76; ROR2, p = 0.15) or in repeated measures analysis (p = 0.32). There was no significant association between screen-sharing and participants’ confidence in explaining their child’s results at ROR1 (p = 0.83) or in repeated measures (p = 0.14); at ROR2, however, we observed that participants in the ScrS arm were more likely to have lower confidence than those in the NSCrS arm (OR = 0.65, 95% confidence interval [CI] [0.44, 0.96], p = 0.03) (Table 2). For objective understanding, we did not observe any significant associations between screen-sharing and the objective understanding sum score at any time point (ROR1, p = 0.93; ROR2, p = 0.71) or in repeated measure analysis (p = 0.94) (Table 3). Notably, on repeated measures, all understanding outcomes decreased significantly over time from ROR1 to ROR2: perceived understanding (OR = 0.50, CI [0.40, 0.63], p < 0.0001), perceived confidence (OR = 0.48, CI [0.38, 0.59], p < 0.0001), and objective understanding (OR = 0.95, CI [0.92, 0.99], p = 0.013) (Tables 2, 3).

To further explore how participant characteristics may have had a differential impact on screen-sharing impact and experience, we stratified our repeated measures analysis by the three largest population groups (H/L, AA, and EA), interpreter use among H/L participants, case-level interpretation, and device (computer/laptop/tablet or smartphone) during results disclosure (shown in Fig. 2). When stratifying by self-reported population group and interpreter use, we did not find any significant associations between screen-sharing and perceived understanding or confidence among these groups. For objective understanding, we found that AA participants in the ScrS arm were more likely to demonstrate lower objective understanding than those in the NScrS arm (OR = 0.83, CI [0.71, 0.97], p = 0.01).

When stratifying by type of device (computer/laptop/tablet and smartphone), we observed that, among participants using a computer, laptop, or tablet during results disclosure, participants in the ScrS arm were more likely to have lower perceived confidence (OR = 0.56, CI [0.32, 0.99], p = 0.046) than those in the NScrS arm. We did not see an association for those that used smartphones. For perceived understanding, we found a similar but borderline significant association among participants using a computer, laptop, or tablet (OR = 0.58, CI [0.33, 1.02], p = 0.057). For objective understanding, no significant associations with screen-sharing and understanding were detected for either device type. When stratifying by case-level interpretation (positive/likely positive, uncertain, and negative results), we did not see any significant effect of screen-sharing on perceived understanding, perceived confidence, or objective understanding within these different result types.

We evaluated participants’ experience with telehealth using the summary score of the brief TSUQ, designed to measure overall telehealth satisfaction and ease of use. The brief TSUQ was administered at ROR1 as described in the Methods above. For the study population overall, the mean brief TSUQ score was 27.3 (SD = 3.44, range: 8–30). In a multivariate linear regression (estimate [β] = 0.56, CI [−0.11, 1.23], p = 0.10), no significant association was observed between screen-sharing and brief TSUQ summary score. For a breakdown of mean score for each brief TSUQ item by arm, refer to online supplementary Table 3.

We compared means across the whole cohort among different population groups to observe whether specific participant characteristics had an impact on brief TSUQ scores. When comparing the mean brief TSUQ scores among the largest population groups, H/L participants who did not use an interpreter reported significantly higher mean brief TSUQ scores than AA participants (28.3, SD = 3.1 vs. 26.5, SD = 4.0; p = 0.0015), but not when compared to EA participants (27.4, SD = 2.9; p = 0.07). In addition, among H/L participants, those who did not use an interpreter had higher mean brief TSUQ scores than those who used an interpreter (28.3, SD = 3.1 vs. 26.6, SD = 3.7; p = 0.0012). The mean brief TSUQ score also differed significantly by case-level interpretations: participants whose child received uncertain results reported lower mean brief TSUQ scores than those who received negative results (26.6, SD = 3.9 vs. 27.7, SD = 3.1; p = 0.004) and positive results (26.6, SD = 3.9 vs. 28.0, SD = 2.7; p = 0.005) (shown in Fig. 3). There was no difference in mean brief TSUQ between device types, computer/laptop/tablet versus smartphone (p = 0.50). Lastly, participants for whom genetic counselors reported technical issues during the results disclosure did not report a difference in mean brief TSUQ compared to those for whom technical issues were not encountered (p = 0.11).

We further investigated participants’ characteristics and their impact on the length of the results disclosure visit. Overall, we found that the means of results disclosure visit length differed significantly between the ScrS and NScrS arms, with the mean visit length for NScS visits being 27.1 (SD = 15.6) minutes and for ScrS, 30.3 (SD = 15.1) minutes (p = 0.04). In comparing the means of different case-level interpretation types, it was observed that the visit length also differed among positive, negative, and uncertain results (all p < 0.0001). On average, the length of a results disclosure visit for a negative result was approximately 21.0 min (SD = 11.3), 40.6 min (SD = 15.7) for positive results, and 30.5 min (SD = 14.8 for uncertain results. With regard to device type, there was a significant difference (p = 0.04) between the mean length of visit between computer/laptop/tablet (26.8 min, SD = 14.2) as compared to smartphone (30.1 min, SD = 16.4).

In addition, we investigated whether the experience of technical issues during results disclosure as reported by the genetic counselor affected length of visit. Participants encountering technical issues (mean = 33.9 min, SD = 17.0) had significantly longer visits (p = 0.0035) than those for whom technical issues were not reported (mean = 27.7 min, SD = 15.0). Importantly, the length of results disclosure visit was not correlated with participants’ brief TSUQ score (Pearson correlation coefficient = −0.02, p = 0.69) (shown in Fig. 4).

In this timely pilot study conducted in the midst of the COVID-19 pandemic, we explored whether the use of screen-sharing to display visual aids during GS results disclosure affected participants’ understanding and experience of receiving their child’s results. We found that the use of screen-sharing did not have a significant impact on participants’ understanding of and confidence in explaining their child’s GS results, or on their perceptions of telehealth satisfaction and ease of use.

Based on previous studies on the use of visuals in genetic counseling [30, 43], we hypothesized that displaying visual aids and other materials including genomic reports would enhance participants’ results disclosure experience. Although we saw that screen-sharing did not have a significant effect on understanding or telehealth experience over time, the majority of participants in both arms reported high levels of perceived understanding and confidence. Contrary to our hypothesis, we note that, at the ROR2 time point, participants’ confidence in explaining their child’s results to others was lower among participants in the ScrS arm as compared to those in NScrS. Although we cannot exclude the effect of multiple comparisons, we also found that among participants who used a computer, laptop, or tablet, a moderate decrease in perceived confidence was seen in the ScrS arm as compared to the NScrS arm using repeated measure analysis.

There are limitations to this study which must be acknowledged. Of note, in both the ScrS and NScrS arms, we were unable to monitor how large or legible the visuals and text were displayed on participants’ screens. It is possible that participants in the ScrS arm may have been overwhelmed or distracted when interacting with both the genetic counselor and reading the contents shown on their screen. In addition, although we were limited in our ability to track what the participant saw during the visit, a lack of screen-sharing may have positively affected participant’s experience as the participant was able to see and engage with the genetic counselor throughout the entire video visit.

It is also important to note that visuals were offered in both the ScrS and NScrS arms although not through screen-sharing in the NScrS arm, and visuals were used more frequently in the ScrS arm. Since genetic counselors in both arms could use visual aids at their discretion, there is the potential for provider bias influencing when visual aids were utilized. Decisions to use visuals could have been influenced by factors such as strength of the participant’s internet connection, type of device used by the participant, and result type. This study was designed to assess screen-sharing as a tool rather than the impact of the types of visual aids themselves. Additional research should investigate if there are specific types of visual or text-based information shared through screen-sharing that most enhance patient experience in genomic results disclosure, as well as whether this effect on confidence persists in disclosure using all device types, among various population groups, and in other clinical contexts. Ultimately, enrollment was constrained by funding and clinical research challenges posted by the COVID-19 pandemic. As a result, this limitation may affect the ability to detect smaller effects or make robust conclusions about subgroup differences. Future studies with larger, power-based sample sizes would be needed to strengthen these findings.

For the TeleKidSeq study, we developed a survey instrument to assess objective understanding of the implications of GS results, as opposed to general recall of the case-level interpretation (“positive,” “negative,” or “uncertain”), which is more commonly used in genetics research. This instrument was designed to capture the nuances of GS results communication which may be missed in assessing understanding based on result type only; results communication may vary significantly from case to case, despite having the same overall case-level interpretation. Across the TeleKidSeq cohort, participants demonstrated high levels of participant-genetic counselor response agreement in line with what was seen in the NYCKidSeq RCT [42]. When stratifying by population group, we observed that, among AA participants, we saw a modest decrease in the odds of higher objective understanding in the ScrS arm as compared to the NScrS arm. Further research is warranted to understand the effects of sociodemographic characteristics on patient’s understanding of their GS results to ensure that telehealth in clinical genomics serves people of all backgrounds.

In using the brief TSUQ instrument adapted for TeleKidSeq, we observed that participants had overall positive perceptions of telehealth satisfaction and ease of use across both arms. Although the COVID-19 pandemic necessitated telehealth use, telehealth will likely be a mainstay within genomic medicine and healthcare at large. The mainstreaming of telehealth was evidenced by the majority of our participants reporting prior use of telehealth, with only 26% reporting that being part of this study was their first time using telehealth. When comparing mean brief TSUQ scores across population and interpreter use groups, we found that mean brief TSUQ scores were significantly higher in H/L participants who did not use an interpreter compared to H/L participants who did use an interpreter. While it is possible that participants who used an interpreter experienced information overload during results disclosure, future research should investigate the effect of interpreter use on participants’ telehealth experience. Additionally, we recognize the potential for digital apps to streamline results disclosure and support patients and providers as they navigate the genomic testing process.

To our knowledge, there have been no studies comparing telehealth satisfaction among patients with differing genetic test result types. We saw that participants who received positive/likely positive and negative results reported significantly higher mean scores than those who received uncertain results, which is reflective of current literature around uninformative results [42, 44] and their psychological impact [45]. The TeleKidSeq study was not designed to disentangle if telehealth satisfaction and perceptions of ease of use differed between case-level interpretations due to the effect of telehealth or the result types themselves. Future research should explore whether case-level interpretation may influence patients’ experiences with telehealth.

To approximate the burden of telehealth on patients and providers, we examined the length of GS results disclosure visits. Longer visit lengths may have represented additional strain on provider workflow and on patient engagement and attention. Despite the additional time incurred by screen-sharing, there was no significant difference in perceptions of telehealth satisfaction and ease of use between the two arms, demonstrating that, from the participants’ perspective, the length of visit did not detract from overall telehealth experience. Taken as a whole, our findings support that video televisits with or without the use of screen-sharing of visual aids is a convenient and effective mode of delivery for genetic test results disclosure in diverse, multilingual patients.

Although the TeleKidSeq pilot study did not find a significant association between screen-sharing and understanding or telehealth experience across all participants, overall, participants demonstrated high perceived understanding, perceived confidence, objective understanding, and telehealth experience scores. These findings support the use of video televisits with or without screen-sharing in genomic results disclosure. Additionally, we identified modest effects of screen-sharing within select population groups that highlight the need for communication strategies that ensure that diverse, multilingual communities are deriving equitable benefit from genomic results disclosure when telehealth is used. These findings represent one of the first studies investigating the use of screen-sharing in genomic results disclosure. Future research should continue to build on our knowledge around genetic and genomic results disclosure, especially as digital tools with screen-sharing become more widely used in clinical practice, to support development of best practices.

The authors thank all the children, parents, and families who participated in this study. The authors also thank the New York Genome Center, Rady Children’s Institute for Genomic Medicine, the Mount Sinai Hospital Genomics Stakeholder Board members, and all referring physicians.

This study protocol was approved by the Institutional Review Boards of Icahn School of Medicine at Mount Sinai (Approval #20-01353) and Albert Einstein College of Medicine (Approval #2020-12292). Digitally signed, informed consent to participate was obtained from parents/legal guardians of all participating children as required by the Institutional Review Boards of both institutions.

Dr. Eimear Kenny has received speaker honoraria from Illumina, 23andMe, Allelica, and Regeneron Pharmaceuticals; received research funding from Allelica; and serves as a scientific advisory board member for Encompass Biosciences, Foresite Labs, and Galateo Bio. Dr. Noura Abul-Husn is an employee and equity holder of 23andMe and serves as a scientific advisory board member for Allelica. All other authors declare no conflict of interest.

Research reported in this publication was supported by the National Human Genome Research Institute (NHGRI) and the National Institute on Minority Health and Health Disparities (NIMHD) of the National Institutes of Health under Award Number U01HG009610. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was supported in part through the computational and data resources and staff expertise provided by Scientific Computing and Data at the Icahn School of Medicine at Mount Sinai and supported by the Clinical and Translational Science Award (CTSA) grant UL1TR004419 from the National Center for Advancing Translational Sciences.

All authors read and approved the final manuscript. Conceptualization: E.E.K., M.P.W., C.R.H., B.D.G., J.M.G., G.A.D., J.A.O., N.R.K., M.S., S.A.S., K.E.B., M.D.B., K.B., P.M., K.M.G., and M.A.R.; data curation: B.J.I., L.G., J.A.O., and N.R.K.; formal analysis: B.J.I., L.G., and J.A.O.; funding acquisition: E.E.K., M.P.W., C.R.H., and B.D.G.; investigation: M.S., S.A.S., K.E.B., M.D.B., K.B., P.M., K.M.G., M.A.R., J.E.R., N.Y., K.L.A., M.A.R., E.M., J.L., N.R.K., and M.A.R.; methodology: E.E.K., M.P.W., C.R.H., B.D.G., J.A.O., N.R.K, M.S., L.G., and B.J.I.; project administration: N.R.K. and M.A.R.; supervision: E.E.K., M.P.W., C.R.H., B.D.G., N.R.K, and M.A.R; visualization: B.J.I., L.G., N.R.K., and J.A.O.; writing – original draft: J.A.O., N.R.K., M.S., L.G., B.J.I.; and writing – review and editing: J.A.O., N.R.K., M.S., L.G., B.J.I, S.A.S., K.E.B., P.N.M., M.D.B., K.B., K.M.G., M.A.R., J.E.R., N.Y., K.L.A., M.A.R., E.M., J.L., R.E.Z., G.A.D., J.M.G., N.S.A.-H., L.J.B., B.D.G., C.R.H., E.E.K., and M.P.W.

Authors' present affiliation: Kaitlyn Brown, Illumina Incorporated, San Diego, CA, USA; Katie M. Gallagher, College Graduate Program in Human Genetics, Sarah Lawrence College, Bronxville, NY, USA; Noura S. Abul-Husn, 23andMe, Inc., Sunnyvale, CA, USA

There are restrictions on the availability of datasets due to participant-level sharing permissions. A subset of the dataset for those who gave permission to share with external researchers may be available from the corresponding authors upon request. The datasets supporting the current study have not been deposited in a public repository because the primary and secondary outcome measures were not part of the publicly shared CSER harmonized measures but may be requested from the corresponding authors.