Abstract
Uncertain genetic information such as variants of uncertain significance (VUS) are often encountered by patients in clinical cancer genetic testing. Although healthcare providers facilitate patient’s understanding of VUS-associated empirical risk and its medical implications, patients’ understanding and perceptions of risk often differ and may be based on subjective evaluations such as their perception of provider’s epistemic authority (EA). This study examines the hypothesis that individuals attribute greater EA to genetic counselors (GCs) (compared to gastrointestinal oncologists) and to providers who recommend more active VUS-related recommendations (compared to inactive). In a factorial experiment, 652 adult participants recruited on Amazon Mechanical Turk were block randomized to read one of 10 different types of VUS-related scenarios in the context of colon cancer (5 recommendation types x 2 provider types). GCs were attributed higher EA than gastrointestinal oncologists (p=<0.001). Active recommendations (comprehensive, check back, wrong) were attributed lower EA (M=3.67, SD=0.79) compared to the inactive (stand by, disregard) (M=3.89, SD=0.67) (p-value=<0.001). The wrong recommendation was attributed lowest EA compared to the four correct recommendations (mean difference= −0.34, −0.45, −0.35, −0.44 respectively; p=0.002), which when dropped from the analysis, showed no difference between the correct active and inactive recommendations (3.78 vs 3.89, p=0.095). The higher EA attributed to GCs is encouraging and possibly explained by increased public awareness of the genetic counseling profession. The lack of difference in EA attributed to various correct, yet incomplete forms of VUS-related recommendation indicates that individuals may be unaware of and thus completely rely on providers for complex medical topics like VUS. Communicating VUS-related uncertainty warrants caution and further research to elucidate best practices and outcomes.
Keywords: Variants of Uncertain Significance, VUS, Epistemic authority, genetic counseling
BACKGROUND
Uncertainty and ambiguity is inherent to genomic information and is increasingly prevalent in genomic medicine due to variants of uncertain significance (VUS), a type of genetic alteration with nebulous relationship with disease and an unknown impact on health (Richards et al., 2015). Providers themselves can be the source of uncertainty in genomic medicine if strategies used for uncertainty communication negatively influence patient experiences of VUS (Makhnoon, Shirts, & Bowen, 2019). Uncertainty communication, although ethically desirable for transparency, is challenging as it may cause confusion and worry in patients and lead to unwarranted medical management decisions (Macklin, Jackson, Atwal, & Hines, 2019). Hesitancy in acknowledging informational uncertainty to patients is not uncommon among physicians (Etkind, Bristowe, Bailey, Selman, & Murtagh, 2017) and this may stem from fear that miscommunication of uncertainty will diminish patients’ perception of provider’s epistemic authority (EA). The term EA is applied to a source of information on whom an individual may rely to acquire knowledge and whose knowledge in a given domain is considered a priori accurate to the extent that it does not require validation (Arie W. Kruglanski, Orehek, Dechesne, & Pierro, 2010). In fact, EA is strongly connected to provider-patient relationship and positively associated with patient’s adherence to medical recommendations (Piette, Heisler, Krein, & Kerr, 2005). Investigation of the role of patients’ perception of providers’ EA in communication of uncertain genetic test results is needed to inform effective methods for returning results from genomic testing in clinical care.
Various VUS management recommendations – active vs inactive, and correct vs incorrect, may differently impact patient’s perception of provider’s EA and their willingness to follow a provider’s recommendation. As previously described by Barnoy (Barnoy, Ofra, & Bar-Tal, 2012) and Stasiuk (Stasiuk, Bar-Tal, & Maksymiuk, 2016), active recommendations advocate for action (e.g., to undergo screening, to check back for reclassification updates) whereas inactive or passive recommendations refer to advice against treatment, or to wait and observe. Accuracy of VUS recommendations refer to their alignment with the American College of Medical Genetics and Genomics (ACMG) clinical management guidelines (Richards et al., 2015). Patients’ perception of various VUS uncertainty communication strategies is an issue that demands substantial research attention due to the (1) prevalence of VUS in clinical genomic testing (Rosenthal, Bernhisel, Brown, Kidd, & Manley, 2017), (2) high frequency of unwarranted recommendations given by providers in response to VUS, and (3) mismanagement of VUS (Macklin et al., 2019). Experimental research in medicine has shown that patients attribute higher EA to physicians who give active recommendations compared to physicians who advise against treatment (Barnoy et al., 2012). This bias is particularly salient to VUS as genomic practice guidelines recommend that VUS should not be used for medical recommendations (Richards et al., 2015), i.e., recommendations are mostly prohibitive rather than prescriptive. Echoing the ACMG recommendations, in oncology, the National Comprehensive Cancer Network (NCCN) recommends basing medical management for individuals with a VUS result in a cancer related gene on family history (NCCN, 2014). Thus in practice, a patient with a VUS may be given one of more of the following management recommendations: (1) Do not use VUS for medical management such as surgery and screening, (2) Do not test family members for clinical purposes, (3) Check back for updates regarding VUS reclassification, and (4) Consider participating in VUS reclassification study. Active VUS-related recommendations, such as to participate in a VUS reclassification study or to periodically check back for reclassification also exist (Makhnoon, Shirts, Bowen, & Fullerton, 2018) and may cause patients to attribute higher EA to providers, increase their belief in a provider’s ability to help, and increase their adherence to VUS-related recommendations. However, there is substantial variability in the frequency with which these recommendation are offered to patients (Makhnoon et al., 2018) and we do not understand whether active versus inactive VUS-related management recommendations affects patients’ perception of providers’ EA or not.
Confirmation and action bias may also influence patients’ perception of provider’s EA and their trust in VUS-related recommendations. In the absence of any other causal factor, it may be easy to incorrectly believe that any genetic alteration present in a gene is causative of disease, despite the variant being a VUS of no known consequence. An incorrect recommendation that confirms patients’ bias may result in perception of higher EA as people tend to ascribe greater EA to those experts whose advice confirms people’s opinions. In addition, under conditions of high uncertainty, people may be inclined to act in order to gain a sense of control over a situation or eliminate the uncertainty (Patt & Zeckhauser, 2000). As a result, incorrect VUS-related recommendations from inexpert providers with little VUS experience may confirm patient’s opinion and yield higher EA. In addition, patients may attribute higher EA to genetic specialists (e.g., genetic counselors) compared to non-genetic experts (e.g., gastrointestinal oncologists) regarding VUS recommendations. It is important to understand the impact of provider specialty on perceived EA to inform provider-patient uncertainty communication in the era of precision medicine.
It is necessary to examine which types of directed advice and communication from which provider specialty yield higher perception of providers’ EA and therefore are likely to have the greatest positive impact on patient adherence to VUS-related management recommendations. This study will examine the potential magnitude of EA in response to different VUS communication strategies. The aims of the current study were to examine the effect of provider specialty and the effect of different VUS-related recommendations on perception of providers’ EA. We therefore tested the following hypotheses: Active VUS-related recommendations would elicit higher perception of EA compared to inactive VUS-related recommendations. In addition, gastrointestinal oncologists, i.e., seemingly non-genetic experts based on their title, will be perceived to have lower EA compared to genetic counselors.
MATERIALS AND METHODS
Sample population and recruitment
on June 22, 2019, we recruited a random sample of adults from an opt-in voluntary panel of internet users via Amazon’s Mechanical Turk (MTurk). MTurk is a crowdsourced internet marketplace that enables individuals and business to coordinate use of human intelligence to perform specific tasks and has been used extensively in psychology and other social sciences research to produce demographically diverse and reliable data (Buhrmester, Kwang, & Gosling, 2011). The survey was made available to MTurk workers via a link to Qualtrics and participants were each paid $0.50 upon completion of the survey. Eligibility criteria for the study were: (1) adults (at or above 18 years of age), (2) a ‘worker’ in Amazon’s Mechanical Turk website, (3) 95% Human Intelligence Task or HIT approval rate, and (4) with an US-based IP address. All participants provided informed consent prior to their inclusion in the study. This study involved a non-invasive intervention and anonymous survey with adult participants was designated as minimal-risk and exempted from review by the MD Anderson Cancer Center institutional review board.
Study Design
The study utilized a 5 (VUS recommendation type) x 2 (provider specialty) between participant factorial design. Participants were block randomized and were assigned to one of five recommendation conditions (Comprehensive, Disregard, Check back, Stand by, and Wrong) and two provider specialty conditions (Genetic Counselor or Gastrointestinal Oncologist). Values of factors (recommendation type and provider specialty) were randomly allocated across vignettes which simplified the modeling of the effects on judgments. In order to avoid fatigue, boredom and satisficing, each respondent was presented with a single vignette followed by a series of questions about that vignette. After reading the randomly assigned vignette, participants completed several survey measures.
The recommendation conditions included in this study was limited to the three VUS-related managements only (disregard, check back, and stand by) as the fourth management of participating in reclassification research studies is just offered as a suggestion, not a strong recommendation. Additionally, some patients may be ineligible to participate in such studies or a study may not be available to them. These recommendations were adopted from a review of >1000 genetic counseling summary reports abstracted from electronic medical records (as part of a separate project) where genetic counselors summarize the implications of a VUS result to patients. Recommendations where the patient is recommended to take an action of any kind is described as active (comprehensive, check back, and wrong), whereas passive recommendations where the patient is not required to take any action is described as inactive (disregard and stand by). The final vignettes were reviewed by a genetic counselor (MM) and piloted in a small convenience sample (n=5). Text analysis conducted using Flesh Kincaid grade level (that considers word length, sentence length and complexity) showed that the vignettes were comparable in complexity and word use (data not shown).
Vignettes
Independent Variable: Manipulation of physician’s recommendation
General instructions and introductory text for all vignettes
Imagine that several members of your family have had colon cancer. About 10% of colon cancers are inherited, meaning that they are caused by genetic changes. You get genetic testing to see if you have a genetic variant that raises your risk of cancer. Your test results show that you have a variant of uncertain significance (VUS) in a gene called MLH1.
VUS is considered to be an inconclusive result
This means that the laboratory has insufficient evidence to say whether this genetic variant is damaging or harmless. There are many variants in MLH1 gene that can increase risk for colon cancer. However, it is unclear whether the variant identified in you is related to an increased cancer risk. In other words, the variant identified in you may or may not increase your risk for colon cancer.
You visit Dr. Smith, a Gastrointestinal Oncologist, a physician specialized in colon cancer (or Counselor Smith, a Genetic Counselor with a Masters in medical genetics and counseling) to get recommendation about how to manage your cancer risk.
Recommendation 1: Comprehensive
Dr. Smith (Counselor Smith) recommends, “Do not change your medical decision because of the VUS result as there is not enough known about the variant to give you a diagnosis. I will use your personal and family history to help you make decision about how to prevent cancer”.
If researchers find more information about this variant in future, the laboratory may change the interpretation of the variant, i.e., “reclassify” it. Dr. Smith (Counselor Smith) says, “The laboratory will alert me if it reclassifies your VUS result to either a deleterious or harmless variant in future. The time it takes to do this can range from months to years or longer. If your VUS is reclassified, I will inform you about it. Even though a VUS may not provide a clear answer, you can still stay in touch with me over the years by calling or scheduling an appointment to check for reclassification.”
Recommendation 2: Disregard
Dr. Smith (Counselor Smith) recommends, “Do not change your medical decision because of the VUS result as there is not enough known about the change to give you a diagnosis. I will use your personal and family history to help you make decision about how to prevent cancer”.
Recommendation 3: Check back
Dr. Smith (Counselor Smith) does not make any recommendation based on the VUS. However if researchers find more information about these variants in future, laboratories may change the interpretation of the variant, i.e., “reclassify” it.
Dr. Smith (Counselor Smith) says, “The laboratory will alert me if it reclassifies your VUS result to either a deleterious or harmless variant in future. The time it takes to do this can range from months to years or longer. Even though the VUS may not provide a clear answer right now, I encourage you to still stay in touch with me over the years by calling or scheduling an appointment to check for reclassification”.
Recommendation 4: Stand by
Dr. Smith (Counselor Smith) does not make any recommendation based on the VUS.
However, Dr. Smith (Counselor Smith) says “If researchers find more information about these variants in future, laboratories may change the interpretation of the variant, i.e., “reclassify” it. The laboratory will alert me if it reclassifies your VUS result to either a deleterious or harmless variant in future. The time it takes to do this can range from months to years or longer. If your VUS is reclassified, I will inform you about it”.
Recommendation 5: Wrong
Dr. Smith (Counselor Smith) says, “We cannot be sure whether VUS causes colon cancer or not. However, because there is a chance that it could cause colon cancer, I recommend that from now on, you get colonoscopies every 6 months instead of the standard recommendation of once every 10 years”.
Measures
Sociodemographic variables (education, income) were assessed along with several outcome variables and potential moderators. We also collected clinical variables such as self-reported history of genetic testing and personal/family history of genetic disorder.
Outcome variables and Potential mediators
Epistemic Authority
EA attributed to the healthcare provider in each scenario was measured using a six-item measure of EA adapted from Barnoy (Barnoy et al., 2012) and Stasiuk (Stasiuk et al., 2016). Items included “Do you think the doctor (or genetic counselor) is an expert in genetics? Do you accept what he/she says as correct?, Do you think that other expert doctors (or genetic counselors) would recommend the same?, Do you think that her/his recommendation is based on well-verified knowledge?, Does the doctor (or genetic counselor) make you feel certain? and Do you feel that the doctor (or genetic counselor) is a credible source of information?” Participants responded to these items on a 5 point Likert scale (definitely to definitely not). Overall EA was computed by summing responses within a person and calculating an average.
Subjective health literacy
Subjective Health literacy was measured using an item from the 3-item health literacy screening measure developed by Chew and colleagues (Chew et al., 2008): “How often do you have someone (like a family member, friend, hospital/clinic worker, or caregiver) help you read health materials?” A 5-point Likert scale was used, with endpoints labeled “None of the time” and “All of the time”; scores were reverse-coded so that higher scores indicated higher health literacy.
Risk Aversion
Risk aversion was measured using the 6-item Pearson Risk Attitude (PRA) scale (α = .61) (Pearson et al., 1995); exemplary items include “I enjoy taking risks,” and “I try to avoid situations that have uncertain outcomes.” A 5-point Likert scale was used, with endpoints “Strongly Disagree” and “Strongly Agree.”
Need for Closure
Need for Closure, a related construct assessing an “individual’s desire for a firm answer to a question” and aversion to general uncertainty, was measured using selected items of the Need for Closure (NFC) Scale (α = .85) (A. W. Kruglanski & Webster, 1996); exemplary items include “I don’t like situations that are uncertain,” and “I do not usually consult many different opinions before forming my own view.” A 5-point Likert scale was used, with endpoints “Strongly Disagree” and “Strongly Agree.”
Data analysis
To test experimental effects on attributed EA, we conducted factorial analysis of variance (recommendation by provider specialty) with the EA as dependent variable.
Independent variables were recommendation conditions (Comprehensive, Disregard, Check back, Stand by, and Wrong), and provider specialty conditions (genetic counselor or gastrointestinal oncologist). If significant main effects or interactions were observed, Tukey’s multiple comparison test or independent t-test was performed depending on the results of ANOVA. We created multivariable regression to examine the association between EA, independent variables and potential mediators. Continuous variables were presented as mean ± standard deviation when normally distributed or as median and interquartile ranges if not. For analyses of recommendation and specialty and EA, we stratified groups of participants with different active and inactive recommendation conditions. We used a t-test statistic to compare means. All analysis was conducted in R and statistical significance was assessed at p≤.05.
RESULTS
A total of 652 completed surveys were received with number of responses per condition ranging between 61 and 69. Sample population characteristics are shown in Table 1. Survey participants were predominantly White (76.1%), of higher educational attainment with 41.6% having a Bachelor’s degree or equivalent, and high income levels with 61.3% earning greater than $45,000 per year. 19.8% of the sample indicated having undergone genetic testing and 14.3% indicated that they had a family or personal history of a genetic disorder. A summary of additional baseline characteristics collected from the survey participants is outlined in supplementary table S1.
Table 1.
Demographic and clinical characteristics of MTurk participants (N=652).
| Variable | Categories | N | % |
|---|---|---|---|
| Race (n = 652) | |||
| White or Caucasian | 496 | 76.1 | |
| Black or African American | 71 | 10.9 | |
| Asian | 46 | 7.1 | |
| Mixed/Other | 25 | 3.8 | |
| Educational attainment (n = 652) | |||
| Never attended school | 0 | 0.0 | |
| Grade school (grades 1 to 8) | 0 | 0.0 | |
| Some high school (grades 9 to 12) | 9 | 1.4 | |
| High school graduate or GED | 66 | 10.1 | |
| Post high school training other than college | 20 | 3.1 | |
| Some college | 178 | 27.3 | |
| Bachelor’s degree or equivalent | 271 | 41.6 | |
| Master’s degree (MS, MBA, MFA, etc.) | 85 | 13.0 | |
| Doctoral or other professional degree (PhD, MD, JD or other) | 18 | 2.8 | |
| Marital Status (n = 652) | |||
| Now married | 305 | 46.8 | |
| Widowed | 13 | 2.0 | |
| Divorced | 48 | 7.4 | |
| Separated | 20 | 3.1 | |
| Never married | 261 | 40.0 | |
| Income (n = 652) | |||
| Less than $15,000 | 45 | 6.9 | |
| $15,000 to $29,999 | 98 | 15.0 | |
| $30,000 to $44,999 | 104 | 16.0 | |
| $45,000 to $59,999 | 121 | 18.6 | |
| $60,000 to $89,999 | 155 | 23.8 | |
| $90,000 to $149,999 | 95 | 14.6 | |
| $150,000 to $199,999 | 18 | 2.8 | |
| $200,000 or above | 11 | 1.7 | |
| Genetic testing (n = 652) | |||
| Had genetic testing before | 129 | 19.8 | |
| Never had genetic testing | 497 | 76.2 | |
| Don’t know/Decline to answer | 21 | 3.2 | |
| Personal/Family history of genetic disorder (n=652) | |||
| Yes | 93 | 14.3 | |
| No | 523 | 80.2 | |
| Don’t know/Decline to answer | 31 | 4.8 | |
Factorial ANOVA showed two significant overall main effects of provider specialty and VUS recommendation on EA attribution and no significant interactions (Table 2). Genetic counselors were attributed higher EA than gastrointestinal oncologists F(1, 642)=11.02, ηp2 = 0.015, p=<0.001. The second main effect was for type of VUS recommendation F(4, 642)=7.88, ηp2 = 0.046, p=<0.001). T-test between mean of the three active compared to sum of the two inactive recommendations showed that the active recommendations were perceived to have lower EA (M=3.67, SD=0.79) compared to the inactive (M=3.89, SD=0.67) (t=−3.7, df=619.5, p=<0.001, d=0.30). Excluding the incorrect recommendation from analysis, mean EA attributed the two correct active recommendations was not significantly different to the mean EA attributed to the two correct inactive recommendations (3.78 vs 3.89, t=−1.6, df=516.9, p=0.095, d=0.16). A post hoc Tukey’s test showed that recommendation 5 different significantly from each of the other recommendations at p<0.05. Providers who offered wrong VUS-related recommendations were perceived to have significantly lower EA compared to each of the other four correct recommendations 1 to 4 (Table 3). However, there was no difference in attributed EA among the four types of correct VUS-related recommendations. To further test the effect of the degree of active message framing, we compared the recommendations Comprehensive to Check Back. Although both provide the option to check in for updates regarding VUS reclassification, the first frames that possibility as an option while the latter states that checking back is actively recommended by the provider. We find that EAs attributed to these recommendations are not significantly different (M=3.78 and 3.79 respectively, t=−0.12, df=257.98, p=0.91, d=0.01).
Table 2:
Provider’s Epistemic Authority: results of Analysis of Variance (ANOVA)
| Variable | Df | Sum sq | Mean Sq | F value | ηp2 | Pr(>F) |
|---|---|---|---|---|---|---|
| Provider’s specialty | 1 | 6 | 5.95 | 11.02 | 0.015 | <0.001 |
| Recommendation | 4 | 17 | 4.25 | 7.87 | 0.046 | <0.001 |
| Provider’s specialty × Recommendation | 4 | 1.9 | 0.47 | 0.87 | 0.005 | 0.47 |
| Residuals | 642 | 347 | 0.54 |
Table 3.
Difference between subgroups of recommendation types and provider specialties compared by Tukey’s test.
| Groups compared | Mean difference | P-value |
|---|---|---|
| Recommendation types | ||
| 2 vs 1 | 0.110 | 0.743 |
| 3 vs 1 | 0.011 | 1.000 |
| 4 vs 1 | 0.106 | 0.772 |
| 5 vs 1 | −0.338 | <0.001 |
| 3 vs 2 | −0.100 | 0.815 |
| 4 vs 2 | −0.004 | 1.000 |
| 5 vs 2 | −0.448 | <0.001 |
| 4 vs 3 | 0.095 | 0.839 |
| 5 vs 3 | −0.349 | 0.002 |
| 5 vs 4 | −0.444 | <0.001 |
| Provider Specialties | ||
| GO vs GC | 0.191 | 0.001 |
Significant differences are in bold; GO: gastrointestinal oncologist; GC: genetic counselor
In multivariable regression, genetic counselors were attributed higher EA compared to gastrointestinal oncologists (β=0.19, p=<0.001) and each correct VUS related recommendation elicited higher EA compared to the incorrect VUS related recommendation (Table 4). Having a personal or family history of genetic disorder was associated with higher EA perceptions (β=0.29, p=0.002). However, experience with genetic testing was not associated with EA perceptions (p=0.70). People with higher subjective health literacy were likely to have lower EA perceptions (β=−0.06, p=0.019).
Table 4:
Multivariable regression of perceived epistemic authority among MTurk participants.
| Variable | β | Std Error | P-value | |
|---|---|---|---|---|
| Education | 0.002 | 0.025 | 0.938 | |
| Income | −0.019 | 0.019 | 0.306 | |
| Provider specialty | Gastrointestinal Oncologist | ref | ||
| Genetic Counselor | 0.192 | 0.058 | <0.001 | |
| Type of VUS recommendation | ||||
| Comprehensive | 0.330 | 0.092 | <0.001 | |
| Disregard | 0.445 | 0.092 | <0.001 | |
| Check Back | 0.330 | 0.093 | <0.001 | |
| Stand by | 0.433 | 0.092 | <0.001 | |
| Wrong | Ref | |||
| Risk Aversion | 0.041 | 0.045 | 0.362 | |
| Need for Closure | −0.05 | 0.063 | 0.396 | |
| Subjective health literacy | −0.065 | 0.03 | 0.019 | |
| Objective health numeracy | Low | Ref | ||
| High | 0.097 | 0.101 | 0.335 | |
| Experience with genetic testing | None/decline to answer | Ref | ||
| Yes | −0.031 | 0.082 | 0.707 | |
| Personal/family history of genetic disorder | None/decline to answer | Ref | ||
| Yes | 0.290 | 0.094 | 0.002 | |
Significant results are in bold
DISCUSSION
This experimental study tested whether communicating various types of VUS-related recommendations by genetic counselors and gastrointestinal oncologists through hypothetical scenarios altered participants’ perceptions of providers’ epistemic authority. To our knowledge, it is the first study to examine these questions in the context of the increasingly important issue of genomic medicine and VUS. This study’s findings are clearly preliminary given its use of hypothetical scenarios and non-patient study participants; however, they have several important implications for the practice of genomic medicine and research efforts.
Our first hypothesis was not supported. In fact, contrary to our first hypothesis, we found that participants had higher perceptions of EA in response to inactive recommendations compared to active recommendations which was driven by the lower ratings of the wrong recommendation within the active group. Once the wrong recommendation is excluded, there was no difference in perception of providers’ EA in response to the four correct VUS-related recommendations which very understandably, may indicate the general public’s unfamiliarity with VUS, lacking a frame of reference for a reasonable VUS-related recommendation. From an epistemic point of view, expert judgements are considered reliable and trustworthy, which in combination with people’s unfamiliarity with VUS might explain why all correct recommendations were received equally well. Therefore, active vs. inactive framing of VUS recommendations have little consequence for the general population who trust medical experts and are likely unfamiliar with VUS. Interestingly, for both specialty types, participants had significantly lower perception of providers’ EA when they were presented with a wrong VUS-related recommendation. The wrong recommendation in this study was positively framed (active recommendation), which according to the cognitive bias literature, people are more likely to comply with. Yet, the large magnitude of change in screening frequency (10 years to 6 months) mentioned in the recommendation may have invoked the number-size framing effect causing people to doubt the wrong recommendation and attribute it the lowest EA.
We found that genetic counselors were perceived to have higher EA compared to gastrointestinal oncologists, which supports our second hypothesis. However, the effect size was small and its impact (if any) downstream clinical outcomes such as patient adherence to VUS-related recommendations warrants further investigation. Advances in genomic technology, reduction in testing costs, and increased public awareness have led to a growing demand for genetic services in both clinical and direct-to-consumer spaces (Hoskovec et al., 2018). The 88% growth of the genetic counseling profession between 2006 and 2016 (“National Society of Genetic Counselors: 2016 Professional Status Survey Executive Summary.,”) has likely been accompanied by similarly increased awareness of the genetic counseling profession and their roles in genomic medicine among in the general population. Thus, the small increase in EA attributed to genetic counselors in this study could indicate a genuine trust in GCs that is certainly well-deserved when it comes to clinical genetics. GCs have greater number of years of genetics training compared to physicians, who on average only receive 36 hours of instruction on medical genetics (Plunkett-Rondeau, Hyland, & Dasgupta, 2015). In addition, GCs are uniquely trained to manage the psychosocial challenges that come with genetic test results in general, including VUS (Biesecker, 2001). In addition, how individuals process knowledge acquired from experts also depends on their so-called self-epistemic authority, i.e., perception of own knowledge in a certain field (A. W. Kruglanski et al., 2005). Indeed, participants who reported having a personal/family history of genetic disorder attributed higher EA to correct VUS recommendations and GCs in this study. However, awareness of the genetic counseling profession is still far from 100% in the general population (Ahram, Soubani, Abu Salem, Saker, & Ahmad, 2015; Maio, Carrion, Yaremco, & Austin, 2013) and it is possible that the most people have no particular explicit or implicit beliefs regarding healthcare provider’s role in genomic medicine and are equally likely to trust any expert in the field. Indeed, participants with higher health literacy who likely possess better self EA reported lower EA perceptions in this study.
This study should be considered in light of its limitations. Our study manipulated multiple VUS recommendations at once (for example, clinical action based on VUS information and exaggerated colonoscopy frequency in recommendation 5); while this approach is representative of real-world circumstances of providing inaccurate VUS recommendation and thus has ecological validity, it prohibits isolation of the effects of individual components of the recommendation on epistemic authority. While the lack of embedded manipulation checks within our experiment increases its ecological validity, we cannot be certain that the participants were aware of and attentive to the independent variables. One particularly important issue is that the representation of active/inactive advice in the vignettes themselves may be overly simplistic as it does not distinguish between the various types of active and inactive recommendations that exist. For example, the advice to check back is less actively communicated in the Comprehensive recommendation compared to the Check Back - as such the active/inactive classification of this recommendation depends on the perception of the participants. Advice “for treatment” and that “one may undergo treatment” are both considered active recommendations but may have different implications. In future experiments, adding information about the frequency of checking back, as is done in some clinical practice, may make these recommendations more clearly active. The preliminary findings from this study are not generalizable to individuals at-risk for cancer who would be eligible to undergo clinical genetic testing and thus may have a different reaction to VUS and uncertainty. In addition, family history of cancer may influence individual’s feelings towards VUS. Furthermore, every vignette in this study was accompanied with a complete general explanation of VUS which may not be offered in all clinical settings. For example, non-genetic practitioners may be unable to provide patients with clear explanations of VUS due to informational deficit as a result of difficulties in staying on top of rapidly expanding new developments in clinical genomics.
Notwithstanding these limitations, the current study provides important preliminary evidence to guide further research and has several strengths including its large sample size. MTurk workers are a generalizable population with regards to health status and behaviors (Walters, Christakis, & Wright, 2018). In addition, the cross-sectional survey experiment method combines the benefits of sample surveys and experiments to offer greater generalizability, valid causal inference and validity. Our findings offer no support that active vs. inactive framing impacts perceptions of EA and GCs are attributed marginally higher EA compared to oncologists. Thus, this dimension of uncertainty communication is unimportant as healthcare providers convey uncertain information to patients in clinical oncology. In addition, practical constraints caused by increasingly popularity of direct-to-consumer testing including testing of clinically significant variants (“BRCA Basics,”), and the predicted GC workforce shortage (Hoskovec et al., 2018) may necessitate oncologists as well as other provider specialties (e.g., nurse practitioners) to counsel patients without consulting GCs. There is a pressing need to understand how various healthcare providers can convey VUS-related uncertainty in a way that boosts patients trust and ensures informed decision making. The current study provides a basis for future work to address this crucial need.
Supplementary Material
What is known about this topic
VUS-related uncertainty communication is desirable but challenging as it can lead to confusion, worry, and unwarranted medical management decisions.
What his paper adds to the topic
This paper tests the influence of VUS recommendation framing and provider specialty on participants’ perceptions of providers’ epistemic authority.
Acknowledgement
We thank Diana Urbauer from MD Anderson Cancer Center who conducted statistical power calculations for this work that greatly improved this manuscript.
Funding:
Supported in part by a research training grant award from the Cancer Prevention and Research Institute of Texas – CPRIT (Award# RP170259) (SM), and grant from the NIH/NCI under award number P30CA016672 that used the Shared Decision Making Core (RJV).
Footnotes
Conflict of Interest:
Sukh Makhnoon, Maureen Mork, Banu Arun, Robert Volk, and Susan Peterson declare that they have no conflict of interest.
Human Studies and Informed Consent
This study conforms to human subjects protection standards as defined by the Declaration of Helsinki and the US Federal Policy for the Protection of Human Subjects. The study was approved by the UT MD Anderson Cancer Center Institutional Review Board.
Animal Studies
No non-human animal studies were carried out by the authors for this article
Data Availability Statement
The datasets generated during and/or analyzed during the current study are not publicly available as publications are still underway from the dataset but are available from the corresponding author on reasonable request.
References
- Ahram M, Soubani M, Abu Salem L, Saker H, & Ahmad M (2015). Knowledge, Attitudes, and Practice Regarding Genetic Testing and Genetic Counselors in Jordan: A Population-Based Survey. J Genet Couns, 24(6), 1001–1010. doi: 10.1007/s10897-015-9839-3 [DOI] [PubMed] [Google Scholar]
- Barnoy S, Ofra L, & Bar-Tal Y (2012). What makes patients perceive their health care worker as an epistemic authority? Nurs Inq, 19(2), 128–133. doi: 10.1111/j.1440-1800.2011.00562.x [DOI] [PubMed] [Google Scholar]
- Biesecker BB (2001). Goals of genetic counseling. Clin Genet, 60(5), 323–330. [DOI] [PubMed] [Google Scholar]
- BRCA Basics. Retrieved from https://www.23andme.com/brca/
- Buhrmester M, Kwang T, & Gosling SD (2011). Amazon’s Mechanical Turk: A New Source of Inexpensive, Yet High-Quality, Data? Perspect Psychol Sci, 6(1), 3–5. doi: 10.1177/1745691610393980 [DOI] [PubMed] [Google Scholar]
- Chew LD, Griffin JM, Partin MR, Noorbaloochi S, Grill JP, Snyder A, … Vanryn M (2008). Validation of screening questions for limited health literacy in a large VA outpatient population. J Gen Intern Med, 23(5), 561–566. doi: 10.1007/s11606-008-0520-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Etkind SN, Bristowe K, Bailey K, Selman LE, & Murtagh FE (2017). How does uncertainty shape patient experience in advanced illness? A secondary analysis of qualitative data. Palliat Med, 31(2), 171–180. doi: 10.1177/0269216316647610 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hoskovec JM, Bennett RL, Carey ME, DaVanzo JE, Dougherty M, Hahn SE, … Wicklund CA (2018). Projecting the Supply and Demand for Certified Genetic Counselors: a Workforce Study. J Genet Couns, 27(1), 16–20. doi: 10.1007/s10897-017-0158-8 [DOI] [PubMed] [Google Scholar]
- Kruglanski AW, Orehek E, Dechesne M, & Pierro A (2010). Lay Epistemic Theory: The Motivational, Cognitive, and Social Aspects of Knowledge Formation. Social and Personality Psychology Compass, 4(10), 939–950. doi: 10.1111/j.1751-9004.2010.00308.x [DOI] [Google Scholar]
- Kruglanski AW, Raviv A, Bar-Tal D, Raviv A, Sharvit K, Ellis S, … Mannetti L (2005) Says Who?: Epistemic Authority Effects in Social Judgment. Vol. 37. Advances in Experimental Social Psychology (pp. 345–392). [Google Scholar]
- Kruglanski AW, & Webster DM (1996). Motivated closing of the mind: “seizing” and “freezing”. Psychol Rev, 103(2), 263–283. [DOI] [PubMed] [Google Scholar]
- Macklin SK, Jackson JL, Atwal PS, & Hines SL (2019). Physician interpretation of variants of uncertain significance. Fam Cancer, 18(1), 121–126. doi: 10.1007/s10689-018-0086-2 [DOI] [PubMed] [Google Scholar]
- Maio M, Carrion P, Yaremco E, & Austin JC (2013). Awareness of genetic counseling and perceptions of its purpose: a survey of the Canadian public. J Genet Couns, 22(6), 762–770. doi: 10.1007/s10897-013-9633-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- Makhnoon S, Shirts BH, & Bowen DJ (2019). Patients’ perspectives of variants of uncertain significance and strategies for uncertainty management. J Genet Couns. doi: 10.1002/jgc4.1075 [DOI] [PubMed] [Google Scholar]
- Makhnoon S, Shirts BH, Bowen DJ, & Fullerton SM (2018). Hereditary cancer gene panel test reports: wide heterogeneity suggests need for standardization. Genet Med, 20(11), 1438–1445. doi: 10.1038/gim.2018.23 [DOI] [PubMed] [Google Scholar]
- National Society of Genetic Counselors: 2016 Professional Status Survey Executive Summary. Retrieved from http://www.nsgc.org/page/whoaregeneticcounselors
- Patt A, & Zeckhauser R (2000). Action Bias and Environmental Decisions. Journal of Risk and Uncertainty, 21(1), 45–72. doi: 10.1023/A:1026517309871 [DOI] [Google Scholar]
- Pearson SD, Goldman L, Orav EJ, Guadagnoli E, Garcia TB, Johnson PA, & Lee TH (1995). Triage decisions for emergency department patients with chest pain: do physicians’ risk attitudes make the difference? J Gen Intern Med, 10(10), 557–564. doi: 10.1007/bf02640365 [DOI] [PubMed] [Google Scholar]
- Piette JD, Heisler M, Krein S, & Kerr EA (2005). The role of patient-physician trust in moderating medication nonadherence due to cost pressures. Arch Intern Med, 165(15), 1749–1755. doi: 10.1001/archinte.165.15.1749 [DOI] [PubMed] [Google Scholar]
- Plunkett-Rondeau J, Hyland K, & Dasgupta S (2015). Training future physicians in the era of genomic medicine: trends in undergraduate medical genetics education. Genet Med, 17(11), 927–934. doi: 10.1038/gim.2014.208 [DOI] [PubMed] [Google Scholar]
- Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, … Rehm HL (2015). Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med, 17(5), 405–424. doi: 10.1038/gim.2015.30 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rosenthal ET, Bernhisel R, Brown K, Kidd J, & Manley S (2017). Clinical testing with a panel of 25 genes associated with increased cancer risk results in a significant increase in clinically significant findings across a broad range of cancer histories. Cancer Genet, 218–219, 58–68. doi: 10.1016/j.cancergen.2017.09.003 [DOI] [PubMed] [Google Scholar]
- Stasiuk K, Bar-Tal Y, & Maksymiuk R (2016). The Effect of Physicians’ Treatment Recommendations on Their Epistemic Authority: The Medical Expertise Bias. J Health Commun, 21(1), 92–99. doi: 10.1080/10810730.2015.1049308 [DOI] [PubMed] [Google Scholar]
- Walters K, Christakis DA, & Wright DR (2018). Are Mechanical Turk worker samples representative of health status and health behaviors in the U.S.? PLoS One, 13(6), e0198835. doi: 10.1371/journal.pone.0198835 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.