[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 35.175.191.168. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Figure 1.
Participant Flow
Participant Flow

A total of 1001 members of the ClinSeq cohort were assessed for eligibility for this trial. Ultimately, 462 individuals were randomized. NIH CC indicates National Institutes of Health Clinical Center.

Figure 2.
Noninferiority of Trial Outcomes
Noninferiority of Trial Outcomes

Mean group differences between educational arms (web-based platform − genetic counselor) noted immediately after (A), 1 month after (B), and 6 months after (C) education with respect to 2 primary outcomes (knowledge, decisional conflict) and 3 secondary outcomes (anxiety, risk worry, risk perception). Test-specific distress is not depicted because a severe floor effect observed at 1 and 6 months rendered parametric tests inappropriate. Noninferiority was tested for outcomes shown with 1-sided 97.5% CIs, and equivalence was tested for the risk perception outcome with a 2-sided 95% CI. The gray shaded portion denotes the noninferiority (or equivalence) range defined by the prespecified margins (δNI or δE), which determines rejection of the null hypothesis if not exceeded. For knowledge, the possible score was 0 to 8 (δNI = −1); anxiety, 6 to 24 (δNI = +2); decisional conflict, 15 to 75 (δNI = +6); risk worry, 1 to 7 (δNI = +1); and risk perception, 1 to 7 (δE = ±1).

Table 1.  
Participant Demographics and Clinical Characteristics
Participant Demographics and Clinical Characteristics
Table 2.  
Variable Distributions
Variable Distributions
Table 3.  
Communication of Results
Communication of Results
1.
Facio  FM, Eidem  H, Fisher  T,  et al.  Intentions to receive individual results from whole-genome sequencing among participants in the ClinSeq study.  Eur J Hum Genet. 2013;21(3):261-265.PubMedGoogle ScholarCrossref
2.
Wilson  KL, Czerwinski  JL, Hoskovec  JM,  et al.  NSGC practice guideline: prenatal screening and diagnostic testing options for chromosome aneuploidy.  J Genet Couns. 2013;22(1):4-15.PubMedGoogle ScholarCrossref
3.
Hoskovec  JM, Bennett  RL, Carey  ME,  et al.  Projecting the supply and demand for certified genetic counselors: a worforce study [published online October 17].  J Genet Couns. doi:10.1007/s10897-017-0158-8Google Scholar
4.
Resta  R, Biesecker  BB, Bennett  RL,  et al; National Society of Genetic Counselors’ Definition Task Force.  A new definition of genetic counseling: National Society of Genetic Counselors’ Task Force report.  J Genet Couns. 2006;15(2):77-83.PubMedGoogle ScholarCrossref
5.
Biesecker  B, Austin  J, Caleshu  C.  Response to a different vantage point commentary: psychotherapeutic genetic counseling, is it?  J Genet Couns. 2017;26(2):334-336.PubMedGoogle ScholarCrossref
6.
McInerney-Leo  A, Biesecker  BB, Hadley  DW,  et al.  BRCA1/2 testing in hereditary breast and ovarian cancer families: effectiveness of problem-solving training as a counseling intervention.  Am J Med Genet A. 2004;130A(3):221-227.PubMedGoogle ScholarCrossref
7.
Athens  BA, Caldwell  SL, Umstead  KL, Connors  PD, Brenna  E, Biesecker  BB.  A systematic review of randomized controlled trials to assess outcomes of genetic counseling.  J Genet Couns. 2017;26(5):902-933.PubMedGoogle ScholarCrossref
8.
Green  MJ, Biesecker  BB, McInerney  AM, Mauger  D, Fost  N.  An interactive computer program can effectively educate patients about genetic testing for breast cancer susceptibility.  Am J Med Genet. 2001;103(1):16-23.PubMedGoogle ScholarCrossref
9.
Green  MJ, McInerney  AM, Biesecker  BB, Fost  N.  Education about genetic testing for breast cancer susceptibility: patient preferences for a computer program or genetic counselor.  Am J Med Genet. 2001;103(1):24-31.PubMedGoogle ScholarCrossref
10.
Green  MJ, Peterson  SK, Baker  MW,  et al.  Effect of a computer-based decision aid on knowledge, perceptions, and intentions about genetic testing for breast cancer susceptibility: a randomized controlled trial.  JAMA. 2004;292(4):442-452.PubMedGoogle ScholarCrossref
11.
Green  MJ, Peterson  SK, Baker  MW,  et al.  Use of an educational computer program before genetic counseling for breast cancer susceptibility: effects on duration and content of counseling sessions.  Genet Med. 2005;7(4):221-229.PubMedGoogle ScholarCrossref
12.
Haga  SB, Barry  WT, Mills  R,  et al.  Impact of delivery models on understanding genomic risk for type 2 diabetes.  Public Health Genomics. 2014;17(2):95-104.PubMedGoogle ScholarCrossref
13.
Biesecker  LG, Mullikin  JC, Facio  FM,  et al; NISC Comparative Sequencing Program.  The ClinSeq Project: piloting large-scale genome sequencing for research in genomic medicine.  Genome Res. 2009;19(9):1665-1674.PubMedGoogle ScholarCrossref
14.
Lewis  KL, Han  PK, Hooker  GW, Klein  WM, Biesecker  LG, Biesecker  BB.  Characterizing participants in the clinseq genome sequencing cohort as early adopters of a new health technology.  PLoS One. 2015;10(7):e0132690.PubMedGoogle ScholarCrossref
15.
Moher  D, Hopewell  S, Schulz  KF,  et al.  CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials.  Int J Surg. 2012;10(1):28-55.PubMedGoogle ScholarCrossref
16.
Hooker  GW, Babu  D, Myers  MF, Zierhut  H, McAllister  M.  Standards for the reporting of Genetic Counseling Interventions in Research and Other Studies (GCIRS): an NSGC Task Force report.  J Genet Couns. 2017;26(3):355-360.PubMedGoogle ScholarCrossref
17.
Bennett  RL, Hart  KA, O’Rourke  E,  et al.  Fabry disease in genetic counseling practice: recommendations of the National Society of Genetic Counselors.  J Genet Couns. 2002;11(2):121-146.PubMedGoogle ScholarCrossref
18.
Gross  SJ, Pletcher  BA, Monaghan  KG; Professional Practice and Guidelines Committee.  Carrier screening in individuals of Ashkenazi Jewish descent.  Genet Med. 2008;10(1):54-56.PubMedGoogle ScholarCrossref
19.
Langfelder-Schwind  E, Kloza  E, Sugarman  E,  et al; National Society of Genetic Counselors Subcommittee on Cystic Fibrosis Carrier Testing.  Cystic fibrosis prenatal screening in genetic counseling practice: recommendations of the National Society of Genetic Counselors.  J Genet Couns. 2005;14(1):1-15.PubMedGoogle ScholarCrossref
20.
Pletcher  BA, Bocian  M; American College of Medical Genetics.  Preconception and prenatal testing of biologic fathers for carrier status.  Genet Med. 2006;8(2):134-135.PubMedGoogle ScholarCrossref
21.
Prior  TW; Professional Practice and Guidelines Committee.  Carrier screening for spinal muscular atrophy.  Genet Med. 2008;10(11):840-842.PubMedGoogle ScholarCrossref
22.
Kincaid  JP, Fishburne  RP, Rogers  RL, Chissom  BS. Derivation of new readability formulas (automated readability index, fog count, and Flesch reading ease formula) for Navy enlisted personnel: Research Branch Rep. 8–75. Memphis, Tennessee: Chief of Naval Technical Training, Naval Air Station Memphis; 1975.
23.
Bell  CJ, Dinwiddie  DL, Miller  NA,  et al.  Carrier testing for severe childhood recessive diseases by next-generation sequencing.  Sci Transl Med. 2011;3(65):65ra4.PubMedGoogle ScholarCrossref
24.
Berg  JS, Adams  M, Nassar  N,  et al.  An informatics approach to analyzing the incidentalome.  Genet Med. 2013;15(1):36-44.PubMedGoogle ScholarCrossref
25.
Stenson  PD, Mort  M, Ball  EV,  et al.  The Human Gene Mutation Database: towards a comprehensive repository of inherited mutation data for medical research, genetic diagnosis and next-generation sequencing studies.  Hum Genet. 2017;136(6):665-677.PubMedGoogle ScholarCrossref
26.
Ng  D, Johnston  JJ, Teer  JK,  et al; NIH Intramural Sequencing Center (NISC) Comparative Sequencing Program.  Interpreting secondary cardiac disease variants in an exome cohort.  Circ Cardiovasc Genet. 2013;6(4):337-346.PubMedGoogle ScholarCrossref
27.
Kaphingst  KA, Facio  FM, Cheng  MR,  et al.  Effects of informed consent for individual genome sequencing on relevant knowledge.  Clin Genet. 2012;82(5):408-415.PubMedGoogle ScholarCrossref
28.
Cella  D, Hughes  C, Peterman  A,  et al.  A brief assessment of concerns associated with genetic testing for cancer: the Multidimensional Impact of Cancer Risk Assessment (MICRA) questionnaire.  Health Psychol. 2002;21(6):564-572.PubMedGoogle ScholarCrossref
29.
O’Connor  AM.  Validation of a decisional conflict scale.  Med Decis Making. 1995;15(1):25-30.PubMedGoogle ScholarCrossref
30.
Marteau  TM, Bekker  H.  The development of a six-item short-form of the state scale of the Spielberger State-Trait Anxiety Inventory (STAI).  Br J Clin Psychol. 1992;31(pt 3):301-306.PubMedGoogle ScholarCrossref
31.
DeMarco  TA, Peshkin  BN, Mars  BD, Tercyak  KP.  Patient satisfaction with cancer genetic counseling: a psychometric analysis of the Genetic Counseling Satisfaction Scale.  J Genet Couns. 2004;13(4):293-304.PubMedGoogle ScholarCrossref
32.
Schwartz  MD, Valdimarsdottir  HB, Peshkin  BN,  et al.  Randomized noninferiority trial of telephone versus in-person genetic counseling for hereditary breast and ovarian cancer.  J Clin Oncol. 2014;32(7):618-626.PubMedGoogle ScholarCrossref
33.
Henneman  L, Oosterwijk  JC, van Asperen  CJ,  et al.  The effectiveness of a graphical presentation in addition to a frequency format in the context of familial breast cancer risk communication: a multicenter controlled trial.  BMC Med Inform Decis Mak. 2013;13(1):55.PubMedGoogle ScholarCrossref
34.
Albada  A, van Dulmen  S, Spreeuwenberg  P, Ausems  MG.  Follow-up effects of a tailored pre-counseling website with question prompt in breast cancer genetic counseling.  Patient Educ Couns. 2015;98(1):69-76.PubMedGoogle ScholarCrossref
35.
Schuirmann  DJ.  A comparison of the two one-sided tests procedure and the power approach for assessing the equivalence of average bioavailability.  J Pharmacokinet Biopharm. 1987;15(6):657-680.PubMedGoogle ScholarCrossref
36.
R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013.
Original Investigation
March 2018

Web Platform vs In-Person Genetic Counselor for Return of Carrier Results From Exome Sequencing: A Randomized Clinical Trial

Author Affiliations
  • 1Social and Behavioral Research Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland
  • 2Medical Genomics and Metabolic Genetics Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland
JAMA Intern Med. 2018;178(3):338-346. doi:10.1001/jamainternmed.2017.8049
Key Points

Question  Is a web-based platform noninferior to a genetic counselor in returning carrier results from exome sequencing?

Findings  In a randomized noninferiority trial of 462 adults, return of results by a web-based platform was noninferior to return by a genetic counselor. Noninferiority was assessed by the lack of significant difference in arms by 1-sided t tests of knowledge of recessive inheritance, test-specific distress, and decisional conflict about choosing to learn results.

Meaning  Return of carrier results from exome sequencing by a web-based platform may be an acceptable, cost-effective alternative to a genetic counselor.

Abstract

Importance  A critical bottleneck in clinical genomics is the mismatch between large volumes of results and the availability of knowledgeable professionals to return them.

Objective  To test whether a web-based platform is noninferior to a genetic counselor for educating patients about their carrier results from exome sequencing.

Design, Setting, and Participants  A randomized noninferiority trial conducted in a longitudinal sequencing cohort at the National Institutes of Health from February 5, 2014, to December 16, 2016, was used to compare the web-based platform with a genetic counselor. Among the 571 eligible participants, 1 to 7 heterozygous variants were identified in genes that cause a phenotype that is recessively inherited. Surveys were administered after cohort enrollment, immediately following trial education, and 1 month and 6 months later to primarily healthy postreproductive participants who expressed interest in learning their carrier results. Both intention-to-treat and per-protocol analyses were applied.

Interventions  A web-based platform that integrated education on carrier results with personal test results was designed to directly parallel disclosure education by a genetic counselor. The sessions took a mean (SD) time of 21 (10.6), and 27 (9.3) minutes, respectively.

Main Outcomes and Measures  The primary outcomes and noninferiority margins (δNI) were knowledge (0 to 8, δNI = −1), test-specific distress (0 to 30, δNI = +1), and decisional conflict (15 to 75, δNI = +6).

Results  After 462 participants (80.9%) provided consent and were randomized, all but 3 participants (n = 459) completed surveys following education and counseling; 398 (86.1%) completed 1-month surveys and 392 (84.8%) completed 6-month surveys. Participants were predominantly well-educated, non-Hispanic white, married parents; mean (SD) age was 63 (63.1) years and 246 (53.6%) were men. The web platform was noninferior to the genetic counselor on outcomes assessed at 1 and 6 months: knowledge (mean group difference, −0.18; lower limit of 97.5% CI, −0.63; δNI = −1), test-specific distress (median group difference, 0; upper limit of 97.5% CI, 0; δNI = +1), and decisional conflict about choosing to learn results (mean group difference, 1.18; upper limit of 97.5% CI, 2.66; δNI = +6). There were no significant differences between the genetic counselors and web-based platform detected between modes of education delivery in disclosure rates to spouses (151 vs 159; relative risk [RR], 1.04; 95% CI, 0.64-1.69; P > .99), children (103 vs 117; RR, 1.07; 95% CI, 0.85-1.36; P = .59), or siblings (91 vs 78; RR, 1.17; 95% CI, 0.94-1.46; P = .18).

Conclusions and Relevance  This trial demonstrates noninferiority of web-based return of carrier results among postreproductive, mostly healthy adults. Replication studies among younger and more diverse populations are needed to establish generalizability. Yet return of results via a web-based platform may be sufficient for subsets of test results, reserving genetic counselors for return of results with a greater health threat.

Trial Registration  clinicaltrials.gov Identifier: NCT00410241

Introduction

Genomic sequencing is increasingly used by medical practitioners in the care of their patients. Its use is primarily in identifying the cause of rare, undiagnosed conditions. Yet sequencing can generate multiple types of clinically relevant results, including carrier results that predict risks to future offspring, which is information that adults have interest in learning.1 As this technology gains favor in clinical practice, it will be challenging to uphold the standard for test results to be returned in person by a knowledgeable practitioner, typically a genetic counselor or medical geneticist.2 Not only is this face-to-face encounter impractical owing to workforce limtations, but increasing use of sequencing will migrate into mainstream medicine and primary care practitioners have significant constraints on their time to add discussions of multiple results. As such, less resource-intensive alternative delivery modes are needed for return of carrier results.3 Such a resource would enable more primary care physicians to effectively use this emerging technology.

Genetic counseling comprises 2 related but distinct functions: the provision of genetic information4 and psychological counseling about managing the threat of living at risk.5 In the design of studies to assess independent effects on outcomes, the education and counseling components can be separated.6 In a recent systematic review of randomized clinical trials (RCTs) reporting outcomes of genetic counseling for predictive genetic testing,7 3 studies compared education by a genetic counselor with pretest education by a web-based platform8-12 and found equivalence or noninferiority between the intervention arms. However, to our knowledge, no published RCTs in genetic counseling have been reported that assessed differences following receipt of results.

Commercial testing companies promote the use of expanded carrier testing and have developed online platforms for returning results. To our knowledge, no reports on the evidence of the effectiveness of these interventions have been published, particularly when compared with clinical genetic counseling. Use of expanded carrier testing by practitioners is increasing and, to our knowledge, no RCT of interventions returning carrier results from exome sequencing has been reported.

We conducted a novel RCT to return results to participants in a postreproductive exome sequencing cohort. Our cohort expressed interest in learning their carrier results for themselves and the benefit of their adult children.1 We selected the return of carrier results because they were desired and deemed low risk for adverse clinical outcomes due to participants’ postreproductive status.

Our objectives were to evaluate the efficacy of a web-based platform for educating patients by assessment of noninferiority compared with a genetic counselor and determine whether observed differences between educational arms were affected by subsequent counseling. We hypothesized that the web-based platform would be noninferior to a genetic counselor in knowledge of recessive inheritance, test-specific distress, and decisional conflict about choosing to learn results.

Methods
Eligibility Criteria

Eligible participants were primarily healthy adults from the ClinSeq cohort13,14 (eTable 1 in Supplement 1) who (1) had completed a baseline survey, (2) were heterozygous for a variant confirmed in a Clinical Laboratory Improvement Amendments–approved laboratory in at least 1 gene causing a phenotype inherited in an autosomal-recessive pattern, and (3) had not received prior genetic test results from ClinSeq.

The National Human Genome Research Institute Institutional Review Board approved this study. Participants provided informed verbal consent; they did not receive financial compensation. CONSORT guidelines15 and the National Society of Genetic Counselors guidelines for reporting studies were used to guide preparation of this article.16 The study protocol is available in Supplement 2.

Study Design

We used a 2 × 2 between-participant factorial design. All participants were randomized to 1 of 4 study arms using an online resource (Research Randomizer, version 4.0, https://www.randomizer.org/) education by web-based platform only, education by counselor only, education by web-based platform followed by genetic counseling, and education by counselor followed by genetic counseling. Study flow is documented in Figure 1. Herein, we report outcomes across the educational arms because randomization to counseling had no effect on our primary outcomes as reported within. Analyses of the counseling sessions are planned for later publication.

Educational Arms

The content for the educational arms was developed using existing resources on carrier status and recessive inheritance and professional guidelines on reporting carrier results to patients.17-21 Each education arm conveyed the same information: what it means to be a carrier, autosomal-recessive inheritance, carrier status for children and grandchildren, the participant’s personal carrier results report, and testing limitations. Concepts were illustrated using identical visual aids in both arms. The individualized carrier results report included information text bubbles that defined the headings for the results in the web-based platform; the counselor explained the results in the other arm. The unique individualized risks for disease among participants’ adult children and grandchildren (low in both cases but for different reasons) and manifestations of the conditions identified in a carrier state were described in both educational arms. The information (absent the names of genetic conditions) scored at a 9.2 grade level as assessed by the Flesch-Kincaid scale.22

The web-based platform was piloted with volunteer ClinSeq participants, and improvements were made to the description of carrier status in response to interview feedback (eAppendix in Supplement 1). Given the high level of education in the cohort and the ease of use for the pilot participants, no focus groups were conducted. If the web-based platform intervention is used in future studies of more diverse and younger populations, the reports may need to be simplified and assessed using cognitive interviewing among the target population.

The education provided by the genetic counselor (one of us, K.L.L.) was not scripted but was designed to convey the same information as the the web-based platform. The counselor deferred any counseling issues that arose in the education session.

All educational sessions were audiorecorded and transcribed. A total of 106 of 228 (46.5%) of the transcripts from the genetic counselor arm were analyzed to assess whether the major topics were addressed consistently and that ancillary topics were not introduced. This approach was key to ensuring fidelity by maintaining content similarity between study arms.

Genetic Counseling Arm

All genetic counseling sessions were conducted by the same genetic counselor (K.L.L.). Genetic counseling was distinguished from educational by considering counseling as any participant concerns related to one’s carrier results that were not information based, which were deffered if raised in the education arm.

Genetic Testing Reports

Eligible variants included those in genes associated with disorders inherited in an autosomal-recessive pattern23,24 if there was no credible evidence of a heterozygote phenotype. Variants also (1) met quality cutoffs and a minor allele frequency cutoff of 0.15 and (2) were splice-site, stop-gain, frameshift, or missense variants previously reported in the Human Gene Mutation Database.25 Pathogenicity assessments were generated for all variants by 1 of us (J.J.J.) and then reviewed by a panel (including several of us, K.L.L., J.J.J., and L.G.B.). A modified 6-point scale was used to classify variants as benign, likely benign, variant of uncertain significance (VOUS)-low, VOUS-high, likely pathogenic, or pathogenic based on several factors: predicted variant effect (loss of function, missense), incidence in affected individuals, frequency of variant in control populations, and functional data.26 Variants classified as pathogenic, likely pathogenic, or VOUS-high were Clinical Laboratory Improvement Amendments (CLIA) validated and reported to participants. Participants received a CLIA report of their carrier results at the end of their visit and a letter 1 month later.

Quantitative Outcomes

Primary trial outcomes were selected from the theory of planned behavior and published RCTs: relevant knowledge, test-specific distress, and decisional conflict. Secondary outcomes included anxiety, risk worry, perceived risk, communication of results to at-risk relatives, and satisfaction. Participants completed surveys immediately after education and immediately after counseling based on randomization, 1 month later, and 6 months later. The surveys included the following scales.

Knowledge of recessive inheritance was measured at all points using 4 novel, true-false items. If both parents are carriers of a mutation associated with a recessive genetic disease, the chance that their pregnancy will be affected by that disease is 25%; if 1 parent is a carrier of a variant in a gene and the other is not, the chance that each of their children is a carrier is 25%; only 25% (1 of 4) of people’s genome sequence is inherited from their mother; and a person can be a carrier for a disease even if no one else in the family has the disease. Items were responded to using a 5-point response scale coded with precedent from a validated scale27 as 0, definitely no, probably no, and uncertain; 1, probably yes; and 2, definitely yes. Items 2 and 3 are false and so reverse scored. Summed total scores range from 0 to 8, with higher scores indicating greater knowledge (test-retest reliability: r = 0.62, P < .001).

Test-specific distress was measured at 1 and 6 months using the distress subscale of the Multidimensional Impact of Cancer Risk Assessment28 adapted for genetic test results. Six items address the frequency with which participants have experienced a distressing emotion in the past 2 weeks on a 4-point scale: 0, never; 1, rarely; 3, sometimes; and 5, often. Summed total scores range from 0 to 30, with higher scores indicating greater test-specific distress (Cronbach α = 0.75).

Decisional conflict was measured at 1 and 6 months using the Decisional Conflict Scale,29 which is a rating scale of 15 items. Summed total scores range from 15 to 75, with higher scores indicating greater decisional conflict about the decision to learn one’s carrier results (Cronbach α = 0.94).

Disclosure rates were assessed at 1 and 6 months by asking participants whether they had told their partner or health care professional about their results. In addition, participants were asked to indicate how many biological daughters, sons, sisters, and brothers they had and how many they had told of their results. Responses were dichotomized into having told at least 1 child or sibling or not.

Risk worry was measured at baseline and at 6 months using a single, Likert-type item (How worried are you that your relatives could be affected with a genetic condition that you have passed on?) with a 7-point response scale ranging from 1, not at all worried; to 7, extremely worried (test-retest reliability: r = 0.35, P < .001).

Perceived risk was measured at baseline and at 6 months using a single item (I feel like my relatives could be affected by a genetic condition that I have passed on) with a Likert-type response scale ranging from 1, strongly disagree; to 7, strongly agree (test-retest reliability: r = 0.42, P < .001).

Anxiety was measured after education and counseling using the short-form State-Trait Anxiety Inventory,30 which consists of 6 items on a 4-point response scale ranging from 1, not at all; to 4, very much. Summed total scores range from 6 to 24, with higher scores indicating greater anxiety (Cronbach α = 0.80).

Satisfaction was measured at 6 months using a modified version of the Genetic Counseling Satisfaction Scale,31 which consists of 3 items (ie, I feel better about my health after getting my result[s] back, the result session was the right length of time, and the result session helped me to process the information about my result[s]), allowing for responses on a 4-point scale ranging from 1, strongly disagree; to 4, strongly agree. Summed total scores range from 3 to 12, with higher scores indicating greater satisfaction (Cronbach α = 0.86).

Power Analysis

Power calculations were based on sample sizes and SDs of the outcome measures assuming 80% power and 2-sided hypothesis tests using a P < .05 α-level criterion. For the immediate outcomes, the minimum detectable differences between arms were 0.61 units in knowledge and 0.79 units in anxiety. For the 6-month outcomes, the minimum detectable differences were 0.65 units in knowledge, 0.32 units in risk worry, 0.49 units in perceived risk, and 2.11 units in decisional conflict. The minimum detectable differences in proportions of dichotomous outcomes at 6 months were 32.1% in distress, 31.8% for disclosure to spouses, 37.9% for children, 43.2% for siblings, and 53.2% for health care professionals. Power at 1 month was comparable to power at 6 months.

Statistical Analysis

Differences between arms at baseline were assessed using χ2 analysis for categorical variables and analysis of variance for continuous variables. The mean difference between educational arms (web-based platform vs counselor) and 1-sided 97.5% CI were calculated for 3 primary outcomes and 2 secondary outcomes to test for noninferiority, which was supported if the CI did not exceed the prespecified noninferiority margin (δNI): −1 in knowledge; +1 in test-specific distress, +6 in decisional conflict, +1 in risk worry, and +2 in anxiety. The δNI for test-specific distress was determined based on a published RCT12 that found significant difference in distress (measured with the Multidimensional Impact of Cancer Risk Assessment) between those who were and were not at increased lifetime risk for type 2 diabetes; the δNI for decisional conflict was based on a margin from a previous noninferiority trial of telephone counseling32; the δNI for anxiety was determined based on a published RCT33 that found no significant differences in Spielberger State-Trait Anxiety Inventory scores between graphic display and frequency format of lifetime risk for breast cancer and replicated in a more recent publication.34 Without precedent to determine a δNI for the novel knowledge scale, the margin was set at the smallest incremental change on the scale (ie, a single point), which corresponded to approximately half an SD at baseline. For single-item measures with categorical response scales (risk worry, perceived risk), the margins were also set at the smallest incremental change. This is a conservative approach applied in noninferiority analyses where there is a lack of precedent in use of a novel scale.32

For the two 1-sided t test procedure,35 the mean difference between educational arms (web-based platform vs counselor) and 95% CI were calculated to test for equivalence for risk perception for which provision of genetics information aims to make more accurate, but collectively not to increase or decrease. Equivalence was supported if the interval did not exceed the equivalence margin (δE): ±1 in perceived risk.

In secondary analyses, 2-way analysis of variance was used to assess the effect of a counseling session if significant differences were detected in primary analyses. Sensitivity analyses were conducted among those with and without children given the differential implications of receipt of results. Differences between arms in satisfaction and disclosure rates were assessed using a t test and the Fisher exact test, respectively. Analyses were based on available data at each point, and both intention-to-treat and per-protocol analyses were applied. For outcomes exhibiting nonnormality, robustness of the findings was verified with nonparametric tests. Parametric analyses were conducted using SPSS, Macintosh version 20.0 (IBM Corp), and nonparametric analyses were conducted using the package36 in R with pairwise CI.

Qualitative Outcomes

Responses to the open-ended question, what, if any, information do you feel was missing from the [genetic counselor/computer] session? were independently coded by 2 of us (I.M.M., A.R.H.) using NVivo 11 (QSR International). Both investigators used the same codebook to analyze the responses by thematic analysis and reconciled most discrepancies through discussion. Intercoder reliability calculated by percent agreement was 98.8% and 99.0% in the web and counselor arms, respectively.

This study was conducted from February 5, 2014, to December 16, 2016, and ended because the 6-month response rate suggested that the target sample size would be achieved.

Results
Participants

All participants completed baseline surveys assessing psychological variables after enrollment in the ClinSeq cohort study and before enrolling in this RCT. As such, time from taking the baseline survey to participation in the RCT varied: time elapsed ranged from 4 months to 4 years (mean [SD], 1.9 [0.7] years). This duration was approximately normally distributed and not significantly associated with any covariates and thus was not contolled for in the analyses. Barring the 3 individuals excluded after randomization, all participants completed in-trial surveys. A total of 398 (86.1%) participants returned 1-month surveys and 392 (84.8%) returned 6-month surveys; nonresponders did not differ significantly from the 462 participants in any demographic variables. All 462 of 571 eligible participants (80.9%) provided informed verbal consent to participate in the RCT.

Overall, this sample was predominantly married, well-educated, postreproductive, and non-Hispanic white; these characteristics were not significantly different from the full ClinSeq cohort.13 Mean (SD) age of the participants was 63 (63.1) years; other demographic and session characteristics are reported in Table 1. The randomization was effective as there were no significant differences in any variables at baseline.

Fidelity to the Intervention

Fidelity to the counselor arm ranged from 83% to 100% (mean [SD], 95% [5.6%]) across the 8 central topics and their subdomains (eTable 2 in Supplement 1). As such, the content of the information conveyed in both educational arms was highly consistent with the exception of information tailored to the patient’s personal variant results, as designed.

Quantitative Outcomes

Means (SDs) of study variables by education arm are given in Table 2. Bivariate analyses resulted in no significant differences between educational arms in any of these variables at baseline. The web-based platform was noninferior to the genetic counselor in terms of knowledge assessed immediately after education and all primary outcomes assessed 1 month later: knowledge, test-specific distress, and decisional conflict. The main analysis at 6 months yielded consistent results. There were no significant differences at 6 months between educational arms in knowledge (mean group difference, −0.18; lower limit of 97.5% CI, −0.63; δNI = −1), test-specific distress (median group difference, 0; upper limit of 97.5% CI, 0; δNI = +1), or decisional conflict (mean group difference, 1.18; upper limit of 97.5% CI, 2.66; δNI = +6). The web-based platform was also noninferior to the genetic counselor for anxiety at the immediate follow-up and for risk worry and equivalent for perceived risk at 6 months. These results are represented in Figure 2. Because test-specific distress data exhibited a floor effect at both 1- and 6-month follow-ups, violating normality assumptions for parametric tests, differences between educational arms were assessed with nonparametric tests. No significant differences were detected at either time point nor did the nonparametric CIs exceed the noninferiority margin (δNI = +1), which supports the hypothesis of noninferiority.

Based on an observed statistically significant difference between the educational arms (although not clinically significant by our margin for anxiety, δNI = +2), analysis of variance testing was used to evaluate differences in anxiety immediately following counseling or no counseling, based on randomization (eFigure in Supplement 1). The interaction effect was significant (F1,448 = 6.94, P < .009), suggesting that the difference in anxiety between educational arms resulted from whether counseling followed education.

All analyses were run separately for those with and without children and the results were consistent for each of the outcomes except decisional conflict at 6 months: the mean difference between educational arms was 2.43 (95% CI, 0.77 to 4.09) among those with at least 1 child, whereas the mean difference was −1.79 (95% CI, −5.69 to 2.12) among those with no children. Thus, parents reported statistically significantly greater decisional conflict when educated by the web-based platform (not exceeding δNI = +6)—an effect not observed among participants without children.

Satisfaction was high overall (mean [SD], 9.86 [2.12]) but significantly lower in the web-based platform arm at 6 months: the mean difference between educational arms was 1.11 (95% CI, 0.71-1.52; P < .001). As reported in Table 3, there were no significant differences observed between educational arms in rate of disclosure to spouses, children, siblings, or health care professionals at 1 or 6 months; however, the power to detect differences was low.

Qualitative Outcomes

Immediately after the educational intervention, 174 of 225 (77.3%) participants in the counselor arm and 96 of 193 (49.7%) in the web-based platform arm answered that none or nothing was missing from the educational sessions. More participants from the web-based platform arm (31 of 193 [16.1%]) than from the counselor arm (5 of 225 [2.2%]) requested additional information specific to their results, such as disease treatment options, risk of disease, and testing options for family members, as well as the frequency and prevalence of their variant in the general population.

Discussion

This study addresses a critical conundrum of clinical genomics: the need for less resource-intensive results delivery modes apparently conflicts with the need to maintain current standards of practice. In-person delivery of individual test results by a health care professional is the standard of care and presumed to be superior to other modes. Our data demonstrate noninferiority of a web-based platform in knowledge of recessive inheritance, test-related distress, and decisional conflict about choosing to learn results. There are important service delivery implications of these results as they suggest that carrier results can be returned to certain populations via a web-based platform that conveys relevant information with sufficient gains in knowledge and no evidence of adverse psychological well-being. These results are consistent with those of 3 other RCTs returning single genetic variant results comparing in-person with computer interventions.8-12 With additional supporting evidence, in-person genetic counseling may be reserved for individuals receiving results that are more health threatening than carrier results.5 Given the limits of the genetics health care workforce,3 such evidence-based alternative delivery modes will be needed. In addition, effective web-based tools for supporting sequencing would allow health care professionals to comfortably and responsibly use this new technology in their practice. This evidence can also help to inform a responsible approach to the results delivery from genome sequencing to address 1 of the challenges faced by large-scale sequencing efforts, such as the National Institutes of Health All of Us Research Program (https://www.nih.gov/research-training/allofus-research-program).

Our results demonstrate that a strong interest in learning carrier results at baseline1 translated to downstream uptake, which differs from past research8 and suggests that participants perceived potential value of their results for their family members. Although this sample was a well-educated group, it remains heartening that we found no indication of distress or other potential psychological harms that may arise from learning one’s carrier status. Parents randomized to the web-based platform expressed greater decisional conflict about learning results, which is not unexpected in that the results pertain to risks to their grandchildren. Those randomized to the web-based platform were less satisfied with the session than were those ramdomized to a genetic counselor. In light of the noninferiority assessments for our primary outcomes and high satisfaction scores overall, the difference may not be clinically meaningful, making it difficult to justify the expense of in-person results delivery. Yet, in response to our qualitative findings, use of a web-based platform should include links to more detailed information on the specific diseases identified and risks to family members for those who desire additional information or reinforcement of the information gained.

Limitations

Our participants are of postreproductive age and early adopters of technology14 who are capable of articulating areas of need and concern related to return of sequence information. As such, results from this study may not generalize to other populations. A replication study is planned for a more diverse, newly recruited cohort. It is also important to replicate these findings among younger adults who may use the information for reproductive decision making.

We chose carrier results for this study because they are common and numerous, but also because they have limited direct health influence on our participants. This was an important consideration for participant safety. These results are not necessarily generalizable to exome or genome sequencing results relating to primary findings for an underlying genetic disease or to secondary findings where the current standard of disclosure by a genetics health care professional should be followed.

Conclusions

Overall, our findings suggest that use of alternative delivery modes in the return of carrier results from genome sequencing should be considered in the face of limited professional resources and the ever-present imperative to reduce health care costs. This approach could also facilitate the use of exome and genome sequencing by nongenetics health care professionals by providing a responsible approach to routine results return that does not place high demands on the ordering clinician. We speculate that similar approaches for return of other sequencing results that are nonthreatening to personal health (eg, pharmacogenetics) may be appropriate. This study provides initial evidence for the effectiveness of carrier information provision by a web-based platform in an older population, which can support the wider use of genomic testing by clinicians and allow genetics health care professionals to focus on more pressing clinical needs for which standard genetic counseling is paramount.

Back to top
Article Information

Corresponding Author: Barbara B. Biesecker, PhD, Social and Behavioral Research Branch, National Human Genome Research Institute, National Institutes of Health, 31 Center Dr, Room B1B36, Bethesda, MD 20892 (barbarab@mail.nih.gov).

Accepted for Publication: November 22, 2017.

Published Online: January 22, 2018. doi:10.1001/jamainternmed.2017.8049

Author Contributions: Dr B. Biesecker and Ms Lewis had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: B. Biesecker, Johnston, L. Biesecker.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: B. Biesecker, Umstead, Turbitt, L. Biesecker.

Critical revision of the manuscript for important intellectual content: B. Biesecker, Lewis, Umstead, Johnston, Turbitt, Fishler, Patton, Miller, Heidlebaugh.

Statistical analysis: B. Biesecker, Umstead.

Obtained funding: L. Biesecker.

Administrative, technical, or material support: Lewis, Johnston, Fishler, Patton, Heidlebaugh, L. Biesecker.

Study supervision: B. Biesecker, Johnston, Fishler, L. Biesecker.

Conflict of Interest Disclosures: Dr L. Biesecker serves as an uncompensated consultant to the Illumina Corp. No other conflicts were reported.

Funding/Support: Ms Lewis, Dr Johnston, Ms Fishler, Mr Patton, Mr Miller, and Dr L. Biesecker were supported by National Institutes of Health (NIH) grants HG200359 08 and HG200387 03.

Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Cristofer Price, PhD (Abt. Associates), and Niraj Trivedi, PhD (National Human Genome Research Institute [NHGRI]/NIH]), provided statistical consultation; Julia Fekecs, BS, and Mark Fredriksen, BS, developed the web platform graphics; Gillian Hooker, PhD, ScM, provided input into study design and constructs assessed in ClinSeq; Ashlee Hulbert, BS, assisted in the laboratory; and William Fix, BS (NHGRI/NIH) assisted in piloting the study; William Klein, PhD (National Cancer Institute/NIH), and Paul Han, MD, MA, MPH (Maine Medical Center Research Institute), contributed to the constructs assessed in ClinSeq; David Ng, MD (NHGRI/NIH), served on the panel that assigned pathogenicity to variants returned; and Tyra Wolfsberg, PhD, Anh-Dao Nguyen, MS, and Mark Fredriksen, BS programmed the web platform. The contributors did not receive financial compensation outside of salary.

Disclaimer: The opinions expressed here represent those of the authors and do not necessarily represent the opinions or policies of the institutions with which they are affiliated.

References
1.
Facio  FM, Eidem  H, Fisher  T,  et al.  Intentions to receive individual results from whole-genome sequencing among participants in the ClinSeq study.  Eur J Hum Genet. 2013;21(3):261-265.PubMedGoogle ScholarCrossref
2.
Wilson  KL, Czerwinski  JL, Hoskovec  JM,  et al.  NSGC practice guideline: prenatal screening and diagnostic testing options for chromosome aneuploidy.  J Genet Couns. 2013;22(1):4-15.PubMedGoogle ScholarCrossref
3.
Hoskovec  JM, Bennett  RL, Carey  ME,  et al.  Projecting the supply and demand for certified genetic counselors: a worforce study [published online October 17].  J Genet Couns. doi:10.1007/s10897-017-0158-8Google Scholar
4.
Resta  R, Biesecker  BB, Bennett  RL,  et al; National Society of Genetic Counselors’ Definition Task Force.  A new definition of genetic counseling: National Society of Genetic Counselors’ Task Force report.  J Genet Couns. 2006;15(2):77-83.PubMedGoogle ScholarCrossref
5.
Biesecker  B, Austin  J, Caleshu  C.  Response to a different vantage point commentary: psychotherapeutic genetic counseling, is it?  J Genet Couns. 2017;26(2):334-336.PubMedGoogle ScholarCrossref
6.
McInerney-Leo  A, Biesecker  BB, Hadley  DW,  et al.  BRCA1/2 testing in hereditary breast and ovarian cancer families: effectiveness of problem-solving training as a counseling intervention.  Am J Med Genet A. 2004;130A(3):221-227.PubMedGoogle ScholarCrossref
7.
Athens  BA, Caldwell  SL, Umstead  KL, Connors  PD, Brenna  E, Biesecker  BB.  A systematic review of randomized controlled trials to assess outcomes of genetic counseling.  J Genet Couns. 2017;26(5):902-933.PubMedGoogle ScholarCrossref
8.
Green  MJ, Biesecker  BB, McInerney  AM, Mauger  D, Fost  N.  An interactive computer program can effectively educate patients about genetic testing for breast cancer susceptibility.  Am J Med Genet. 2001;103(1):16-23.PubMedGoogle ScholarCrossref
9.
Green  MJ, McInerney  AM, Biesecker  BB, Fost  N.  Education about genetic testing for breast cancer susceptibility: patient preferences for a computer program or genetic counselor.  Am J Med Genet. 2001;103(1):24-31.PubMedGoogle ScholarCrossref
10.
Green  MJ, Peterson  SK, Baker  MW,  et al.  Effect of a computer-based decision aid on knowledge, perceptions, and intentions about genetic testing for breast cancer susceptibility: a randomized controlled trial.  JAMA. 2004;292(4):442-452.PubMedGoogle ScholarCrossref
11.
Green  MJ, Peterson  SK, Baker  MW,  et al.  Use of an educational computer program before genetic counseling for breast cancer susceptibility: effects on duration and content of counseling sessions.  Genet Med. 2005;7(4):221-229.PubMedGoogle ScholarCrossref
12.
Haga  SB, Barry  WT, Mills  R,  et al.  Impact of delivery models on understanding genomic risk for type 2 diabetes.  Public Health Genomics. 2014;17(2):95-104.PubMedGoogle ScholarCrossref
13.
Biesecker  LG, Mullikin  JC, Facio  FM,  et al; NISC Comparative Sequencing Program.  The ClinSeq Project: piloting large-scale genome sequencing for research in genomic medicine.  Genome Res. 2009;19(9):1665-1674.PubMedGoogle ScholarCrossref
14.
Lewis  KL, Han  PK, Hooker  GW, Klein  WM, Biesecker  LG, Biesecker  BB.  Characterizing participants in the clinseq genome sequencing cohort as early adopters of a new health technology.  PLoS One. 2015;10(7):e0132690.PubMedGoogle ScholarCrossref
15.
Moher  D, Hopewell  S, Schulz  KF,  et al.  CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials.  Int J Surg. 2012;10(1):28-55.PubMedGoogle ScholarCrossref
16.
Hooker  GW, Babu  D, Myers  MF, Zierhut  H, McAllister  M.  Standards for the reporting of Genetic Counseling Interventions in Research and Other Studies (GCIRS): an NSGC Task Force report.  J Genet Couns. 2017;26(3):355-360.PubMedGoogle ScholarCrossref
17.
Bennett  RL, Hart  KA, O’Rourke  E,  et al.  Fabry disease in genetic counseling practice: recommendations of the National Society of Genetic Counselors.  J Genet Couns. 2002;11(2):121-146.PubMedGoogle ScholarCrossref
18.
Gross  SJ, Pletcher  BA, Monaghan  KG; Professional Practice and Guidelines Committee.  Carrier screening in individuals of Ashkenazi Jewish descent.  Genet Med. 2008;10(1):54-56.PubMedGoogle ScholarCrossref
19.
Langfelder-Schwind  E, Kloza  E, Sugarman  E,  et al; National Society of Genetic Counselors Subcommittee on Cystic Fibrosis Carrier Testing.  Cystic fibrosis prenatal screening in genetic counseling practice: recommendations of the National Society of Genetic Counselors.  J Genet Couns. 2005;14(1):1-15.PubMedGoogle ScholarCrossref
20.
Pletcher  BA, Bocian  M; American College of Medical Genetics.  Preconception and prenatal testing of biologic fathers for carrier status.  Genet Med. 2006;8(2):134-135.PubMedGoogle ScholarCrossref
21.
Prior  TW; Professional Practice and Guidelines Committee.  Carrier screening for spinal muscular atrophy.  Genet Med. 2008;10(11):840-842.PubMedGoogle ScholarCrossref
22.
Kincaid  JP, Fishburne  RP, Rogers  RL, Chissom  BS. Derivation of new readability formulas (automated readability index, fog count, and Flesch reading ease formula) for Navy enlisted personnel: Research Branch Rep. 8–75. Memphis, Tennessee: Chief of Naval Technical Training, Naval Air Station Memphis; 1975.
23.
Bell  CJ, Dinwiddie  DL, Miller  NA,  et al.  Carrier testing for severe childhood recessive diseases by next-generation sequencing.  Sci Transl Med. 2011;3(65):65ra4.PubMedGoogle ScholarCrossref
24.
Berg  JS, Adams  M, Nassar  N,  et al.  An informatics approach to analyzing the incidentalome.  Genet Med. 2013;15(1):36-44.PubMedGoogle ScholarCrossref
25.
Stenson  PD, Mort  M, Ball  EV,  et al.  The Human Gene Mutation Database: towards a comprehensive repository of inherited mutation data for medical research, genetic diagnosis and next-generation sequencing studies.  Hum Genet. 2017;136(6):665-677.PubMedGoogle ScholarCrossref
26.
Ng  D, Johnston  JJ, Teer  JK,  et al; NIH Intramural Sequencing Center (NISC) Comparative Sequencing Program.  Interpreting secondary cardiac disease variants in an exome cohort.  Circ Cardiovasc Genet. 2013;6(4):337-346.PubMedGoogle ScholarCrossref
27.
Kaphingst  KA, Facio  FM, Cheng  MR,  et al.  Effects of informed consent for individual genome sequencing on relevant knowledge.  Clin Genet. 2012;82(5):408-415.PubMedGoogle ScholarCrossref
28.
Cella  D, Hughes  C, Peterman  A,  et al.  A brief assessment of concerns associated with genetic testing for cancer: the Multidimensional Impact of Cancer Risk Assessment (MICRA) questionnaire.  Health Psychol. 2002;21(6):564-572.PubMedGoogle ScholarCrossref
29.
O’Connor  AM.  Validation of a decisional conflict scale.  Med Decis Making. 1995;15(1):25-30.PubMedGoogle ScholarCrossref
30.
Marteau  TM, Bekker  H.  The development of a six-item short-form of the state scale of the Spielberger State-Trait Anxiety Inventory (STAI).  Br J Clin Psychol. 1992;31(pt 3):301-306.PubMedGoogle ScholarCrossref
31.
DeMarco  TA, Peshkin  BN, Mars  BD, Tercyak  KP.  Patient satisfaction with cancer genetic counseling: a psychometric analysis of the Genetic Counseling Satisfaction Scale.  J Genet Couns. 2004;13(4):293-304.PubMedGoogle ScholarCrossref
32.
Schwartz  MD, Valdimarsdottir  HB, Peshkin  BN,  et al.  Randomized noninferiority trial of telephone versus in-person genetic counseling for hereditary breast and ovarian cancer.  J Clin Oncol. 2014;32(7):618-626.PubMedGoogle ScholarCrossref
33.
Henneman  L, Oosterwijk  JC, van Asperen  CJ,  et al.  The effectiveness of a graphical presentation in addition to a frequency format in the context of familial breast cancer risk communication: a multicenter controlled trial.  BMC Med Inform Decis Mak. 2013;13(1):55.PubMedGoogle ScholarCrossref
34.
Albada  A, van Dulmen  S, Spreeuwenberg  P, Ausems  MG.  Follow-up effects of a tailored pre-counseling website with question prompt in breast cancer genetic counseling.  Patient Educ Couns. 2015;98(1):69-76.PubMedGoogle ScholarCrossref
35.
Schuirmann  DJ.  A comparison of the two one-sided tests procedure and the power approach for assessing the equivalence of average bioavailability.  J Pharmacokinet Biopharm. 1987;15(6):657-680.PubMedGoogle ScholarCrossref
36.
R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013.
×