Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jul 1.
Published in final edited form as: Health Psychol. 2018 May 10;37(6):553–561. doi: 10.1037/hea0000579

Associations of Perceived Norms With Intentions to Learn Genomic Sequencing Results: Roles for Attitudes and Ambivalence

Allecia E Reid 1, Jennifer M Taber 2, Rebecca A Ferrer 2, Barbara B Biesecker 3, Katie L Lewis 3, Leslie G Biesecker 3, William M P Klein 2
PMCID: PMC5962407  NIHMSID: NIHMS951206  PMID: 29745680

Abstract

Objective

Genomic sequencing is becoming increasingly accessible, highlighting the need to understand the social and psychological factors that drive interest in receiving testing results. These decisions may depend on perceived descriptive norms (how most others behave) and injunctive norms (what is approved of by others). We predicted that descriptive norms would be directly associated with intentions to learn genomic sequencing results, whereas injunctive norms would be associated indirectly, via attitudes. These differential associations with intentions versus attitudes were hypothesized to be strongest when individuals held ambivalent attitudes toward obtaining results.

Methods

Participants enrolled in a genomic sequencing trial (n=372) reported intentions to learn medically actionable, non-medically actionable, and carrier sequencing results. Descriptive norms items referenced other study participants. Injunctive norms were analyzed separately for close friends and family members. Attitudes, attitudinal ambivalence, and sociodemographic covariates were also assessed.

Results

In structural equation models, both descriptive norms and friend injunctive norms were associated with intentions to receive all sequencing results (ps<.004). Attitudes consistently mediated all friend injunctive norms-intentions associations, but not the descriptive norms-intentions associations. Attitudinal ambivalence moderated the association between friend injunctive norms (p≤.001), but not descriptive norms (p=.16), and attitudes. Injunctive norms were significantly associated with attitudes when ambivalence was high, but were unrelated when ambivalence was low. Results replicated for family injunctive norms.

Conclusions

Descriptive and injunctive norms play roles in genomic sequencing decisions. Considering mediators and moderators of these processes enhances ability to optimize use of normative information to support informed decision making.

Keywords: genomic sequencing, descriptive norms, injunctive norms, attitudes, ambivalence

The Precision Medicine Initiative (“All of Us”), which aims to engage over 1 million Americans in genomic sequencing, highlights the potential for using genomic information to improve population health (The White House, 2015). Advancements in sequencing technology have reduced the cost and increased availability of testing (Biesecker & Green, 2014). Genomic sequencing can be used to identify the cause of undiagnosed disease (Daoud et al., 2016), to test for medically actionable secondary findings (Green et al., 2013), and for broad carrier screening (Haque et al., 2016). Results from genomic sequencing can be particularly useful in the case of actionable results, to reduce future health risks by pursuing screening or changing one’s lifestyle behaviors, and in carrier screening, to make informed reproductive choices.

However, the novelty of genomic sequencing introduces challenges for individuals faced with decisions about undergoing and receiving the results of testing. In considering the value of the potential information, recipients must weigh the likelihood of an uncertain result. These are findings that may or may not represent pathogenic variants in the absence of further data. As such, decisions to receive genomic sequencing results are complicated by consideration of the consequences of a range of uncertainties (Han et al., in press). For patients, genetic discrimination has been a long-standing concern (Green, Lautenbach, & McGuire, 2015). Individuals and clinicians making these decisions do so with limitations in information and social considerations. Given the potential health benefits and the associated risks, it is important to understand the factors that contribute to individuals’ decisions not only to undergo genomic sequencing, but also to ultimately receive the results of genomic sequencing tests.

Research has focused primarily on intrapersonal predictors (e.g., risk, worry) in the context of genomic testing and sequencing (Sweeny, Ghane, Legg, Huynh, & Andrews, 2014). Owing to little experience with this novel testing, individuals may look to others during decision making. Whereas descriptive norms reflect how most others behave in a situation, injunctive norms reflect what is approved or disapproved by most others (Cialdini, Reno, & Kallgren, 1990). Associations have been documented between injunctive norms and genetic testing intentions (Sweeny et al., 2014), with less attention to descriptive norms. As genomic sequencing has become more widely utilized (Lim et al., in press) and is projected to increase in availability (Johansen Taber, Dickinson, & Wilson, 2014), research is needed to provide a head start in understanding how psychosocial factors might drive decision making.

In the present research, participants enrolled in a genomic sequencing trial indicated intentions to learn genomic information for medically actionable diseases and non-actionable diseases, reflecting genes that predispose to a disease that can versus cannot be treated. Intentions were also rated for carrier status, genes which may affect their biological children’s health but not their own. Intentions to receive results indeed vary, as individuals enroll in sequencing trials not only to receive their results, but also to contribute to scientific knowledge (Facio et al., 2011). We examined associations of descriptive and injunctive norms with intentions for receiving sequencing results and considered whether these associations were mediated by attitudes and moderated by attitudinal ambivalence.

Understanding the mediators and moderators of putative determinants of behavior is important for designing efficacious interventions. With respect to mediators, descriptive and injunctive norms are likely to operate via different pathways. It has been hypothesized that descriptive norms have a direct influence on behavior, whereas injunctive norms have an indirect influence, altering beliefs or cognitions en route to behavior (Cialdini, 2003). Indeed, experiments support that injunctive norms operate through evaluations and attitudes (Cialdini, 2003; Reid & Aiken, 2013). However, to our knowledge, whether descriptive norms also have an indirect effect through evaluations has not been examined.

With regard to moderators, individuals may be most likely to look to others for guidance on whether to receive genomic sequencing results when the best course of action is ambiguous (Cialdini & Trost, 1998). Attitudinal ambivalence occurs when individuals simultaneously hold positive and negative attitudes about an object (Conner & Sparks, 2002). Ambivalence weakens reliance on personal beliefs in decision making (Conner & Sparks, 2002) and may therefore heighten reliance on normative information. Ambivalence is common in genetics decision making, as people hold positive attitudes about certain aspects (e.g., knowledge gained) but negative attitudes about others (e.g., discrimination; Haga et al., 2013; Sapp et al., 2010). Whether the effect of descriptive norms is moderated by attitudinal ambivalence has not been examined, and the literature on injunctive norms is mixed (e.g., Conner, Povey, Sparks, James, & Shepherd, 2003; Hohman, Crano, Siegel, & Alvaro, 2014). Consistent with Hohman et al. (2016), previous research has not considered that ambivalence should heighten the indirect influence of injunctive norms on attitudes Conversely, because descriptive norms do not operate through attitudes, high ambivalence should enhance the direct association of descriptive norms with intentions but should not foster an indirect influence via attitudes.

Descriptive and injunctive norms are differentially motivated by the goals of accurate decision making versus relationship formation and maintenance, and therefore differ in which reference group is most influential (Cialdini & Trost, 1998). For descriptive norms, the behavior of others who have experienced the same situation (i.e., others who have the opportunity to learn sequencing results) indicates the most effective course of action and is most influential (Goldstein, Cialdini, & Griskevicius, 2008). For injunctive norms, we are most influenced by potential receipt of approval from others with whom we desire a relationship (i.e., close others; Christensen, Rothgerber, Wood, & Matz, 2004; Neighbors et al., 2008).

We examined associations of descriptive and injunctive norms with intentions to learn genomic sequencing results. Participants were enrolled in a National Institutes of Health genomic sequencing trial (ClinSeq®) aimed at identifying variants related to heart disease and carrying out whole genomic sequencing. Prior to receiving any results, participants reported intentions for learning results for medically actionable diseases, non-actionable diseases, and carrier status. Building on the reference groups identified as important in previous research, descriptive norms referenced other study participants. Injunctive norms were examined for two separate referents—close friends and family members. It would make less conceptual sense for people’s sequencing decisions to be driven by approval from other study participants (people with whom they are not acquainted), or by perceptions of whether family and friends seek out sequencing results, given the relatively low frequency of enrollment in sequencing studies.

We hypothesized that whereas descriptive norms would have a direct association with intentions to learn sequencing results (see Figure 1), the effect of injunctive norms would be mediated by attitudes. Moreover, high attitudinal ambivalence was expected to strengthen the associations between descriptive norms and intentions and between injunctive norms and attitudes, but have no effect on the relationship between descriptive norms and attitudes. Given previous research, we made no predictions regarding whether ambivalence would moderate the association between injunctive norms and intentions. As it is unclear whether the role of norms varies in the context of more versus less threat (i.e., medically non-actionable versus actionable results), our predictions did not differentiate the three types of sequencing results.

Figure 1.

Figure 1

Open in a new tab

Hypothesized relationships among descriptive norms, injunctive norms, attitudes, intentions and ambivalence. Solid lines reflect relationships hypothesized to be significant, dashed lines reflect relationships hypothesized to be non-significant, and bold lines reflect relationships for which we did not make predictions.

Method

Participants and Procedure

The National Human Genome Research Institute’s IRB approved the parent study. Participants provided informed consent before enrolling in the study. All participants provided DNA samples and completed a baseline survey before receiving any genetic information. They were informed that genetic information may be available in the future. Other manuscripts have examined predictors of intentions to learn sequencing results but have not examined norms, ambivalence, or attitudes (Ferrer et al., 2015; Taber, Klein, Ferrer, Han, et al., 2015; Taber, Klein, Ferrer, Lewis, Biesecker, et al., 2015; Taber, Klein, Ferrer, Lewis, Harris, et al., 2015).

Participants (N = 540) were recruited from the greater Bethesda, Maryland area. At enrollment, participants were required to be over age 45, to maximize likelihood of detecting heart disease, and under 65, to allow sufficient time for studying long-term outcomes in the parent study. Inclusion in analyses required that participants provided complete data on all predictors (n = 394). Our analytic approach for addressing non-independence due to 54 spousal pairs enrolled in the trial, described below, resulted in a final sample of 372 observations.

Measures

Participants responded to items referencing the results of genomic sequencing for three different types of outcomes—medically actionable, non-medically actionable, and carrier results. Medically actionable disease results referenced “a gene variant that predisposes you to a disease that can be prevented or treated.” Non-medically actionable disease results referred to “a gene variant that predisposes you to a disease that cannot be prevented or treated.” Carrier status results referenced “a gene variant that does not affect your health, but that may be important to the health of other relatives, such as your children.”

Intentions

Intentions to learn the three types of sequencing results were each measured by two items (adapted from Dormandy, Hankins, & Marteau, 2006). Items included “I intend to learn such a result” rated on a scale from 1 (definitely no) to 5 (definitely yes) and “How likely is it that you will choose to learn about such a result?” rated on a scale from 1 (extremely unlikely) to 7 (extremely likely). Correlations between the two items ranged from .22 to .72 (medically actionable: r = .22, p < .001; non-medically actionable: r = .72, p < .001; carrier status: r = .31, p < .001). The modest correlations for medically actionable and carrier results are likely due to range restriction. Over 80% of respondents selected the top two response options on both sets of items. The high correspondence in ratings, and the high correlation for the non-medically actionable items, suggests that the items are tapping the same construct.

Given the different scales for the two items, items were standardized and then averaged to form independent scales for each of the three types of results. A log transformation was applied to normalize the distributions (medically actionable: original skew = -1.64 and kurtosis = 2.17, transformed skew = -1.09 and kurtosis = -0.20; non-medically actionable: original skew = -1.54 and kurtosis = 1.93, transformed skew = -0.84 and kurtosis = -0.66; carrier: original skew = -1.94 and kurtosis = 4.43, transformed skew = -1.04 and kurtosis = -0.05). Scores were reverse-scored to preserve the original direction of the items (Tabachnick & Fidell, 2013; see also, e.g., Taber, Klein, Ferrer, Lewis, Biesecker, et al., 2015 for similar treatment of intentions). Dichotomizing the intentions measures (highest possible score vs. all others) to address skew and ceiling effects produced the same results. The continuous measures are therefore reported.

Descriptive norms

Descriptive norms were assessed with two items paralleling those for intentions. Items assessed whether other ClinSeq® participants “intend to learn their results” rated from 1 (definitely no) to 5 (definitely yes) and would likely “choose to learn their results” rated from 1 (extremely unlikely) to 7 (extremely likely). The items were standardized and averaged (r = .65, p < .001).

Injunctive norms

Separate items assessed participants’ perceived approval to learn results from their doctor, partner, sibling(s), child(ren), and close friends. Close friends were selected as the primary referent because participants would seek approval from friends, and like the “other ClinSeq® participants” referent used for descriptive norms, was one for whom sequencing results have no bearing on personal health outcomes. Injunctive norms included three items, one for each type of sequencing results. Participants responded to, “My closest friends think that I should learn such a result” on a 1 (definitely no) to 5 (definitely yes) scale. To maintain the same level of specificity as descriptive norms, the three items were averaged (α = .86) for participants who responded to at least two items. As results may be consequential for biological family members, their approval could arguably exert only a direct effect on intentions, rather than an indirect effect through attitudes. To demonstrate generalizability of results, we conducted a sensitivity analysis examining family injunctive norms, reflecting mean level of approval from biological family members—siblings and children.

Attitudes

Attitudes were assessed with 18 items total, six for each of the three types of sequencing results. Items were adapted from previous research (Biesecker, 2012; Dormandy et al., 2006; Montenegro et al., 2011; Solomon et al., 2012). For each type of sequencing result, participants rated on a 1 to 7 scale the extent to which “For me, learning such a result would be:” (1) a bad thing vs. not a bad thing, (2) not beneficial vs. beneficial, (3) harmful vs. not harmful, (4) not a good thing vs. a good thing, (5) not worthwhile vs. worthwhile, and (6) unimportant vs. important. Consistent with the treatment of injunctive norms, all items were averaged (α = .92).

Attitudinal ambivalence

Attitudinal ambivalence was assessed with six items total, two for each type of sequencing result (adapted from Lipkus, Green, & Marcus, 2003). Items reflected felt, rather than potential, ambivalence. Both assessment approaches are common, and neither has been deemed preferable for psychological research (Conner & Armitage, 2008). Participants indicated to what extent they “have mixed feelings about whether to receive this type of sequencing result” and are “torn about whether to learn this type of sequencing result” on a 1 (strongly disagree) to 5 (strongly agree) scale. The six items were averaged (α = .90).

Demographics

Participants reported demographic variables that may be associated with intending to learn the results of genetic testing. These included age, gender, race and ethnicity (coded as White vs. non-White). Participants also reported educational level (1 = high school, 2 = some college, 3 = college graduate, 4 = post-graduate) and household income, ranging from 1 (less than $25,000/ year) to 5 (more than $100,000/ year).

Data Analysis

Fifty-four pairs of spouses were enrolled in the parent study, potentially violating the assumption of independence of observations. Among the primary variables, only injunctive norms were correlated between husbands and wives (r = .39, p = .01; Kenny, Kashy, & Cook, 2006). As recommended by Kenny et al. (2006), we addressed nonindependence by using the average of the spouse’s scores on all variables in place of the husband’s and wives’ individual scores, resulting in an analytic sample of 372. Sample descriptives and bivariate correlations among the variables of interest were examined. Given their associations with favorability toward genetic testing, gender, age, and education were included as covariates (Aro et al., 1997; Collins, Ryan, & Truby, 2014; Vermeulen, Henneman, van El, & Cornel, 2014). Ethnicity and income, which often covary with education, were also included as covariates.

All hypotheses were tested in structural equation models, analyzed in Mplus 7 (Muthén & Muthén, 2015). Direct and indirect effects were examined simultaneously in three separate models, one for intentions to learn each type of sequencing result. Norms and all covariates predicted attitudes, and all variables, including attitudes, predicted intentions. Mediation analyses utilized bias-corrected bootstrapped confidence intervals (MacKinnon, 2008). Interactions between ambivalence and both norms were added to these base models. Variables were mean centered prior to forming interactions; significant interactions were probed at 1 standard deviation above and below the mean of ambivalence (Aiken & West, 1991). In the sensitivity analysis, all analyses were conducted with “family injunctive norms” inserted in place of “friend injunctive norms.”

Results

Participants averaged 60.7 years of age at survey completion (SD = 5.51). They were predominantly male (57%), White (93%), had a college or post-graduate degree (89%), and earned more than $100,000 per year (78%), a common demographic profile for people engaged in genomic sequencing research and practice (Hensley Alford et al., 2011). Participants retained in analyses did not differ from those who failed to provide complete data (n=146) on demographic factors, descriptive norms, or intentions (ps ≥ .08). However, retained participants had more positive attitudes toward (p = .03) and were less ambivalent in their attitudes about learning their sequencing results (p = .01). Most individuals (77%) were excluded because they indicated that at least two of the three friend injunctive norms items were “not applicable.”

Descriptive statistics for and bivariate correlations between the primary predictors and outcomes are presented in Table 1. There was only a small correlation between descriptive and injunctive norms (r = .14), supporting that they are distinct constructs. Descriptive and injunctive norms both exhibited small to moderate correlations with intentions to learn each type of sequencing result (rs = .18 - .22). Attitudes were more strongly correlated with all three intentions outcomes (rs = .30 - .52) than were either descriptive or injunctive norms.

Table 1.

Correlations Among Norms, Attitudes, Ambivalence, and Intentions

Descriptive norms a Friend injunctive norms Family injunctive norms Attitudes Ambivalence Intentions: medically actionable results a Intentions: non-medically actionable results a Intentions: carrier results a
Friend injunctive norms .14**
Family injunctive norms .14** .76***
Attitudes .16** .43*** .48***
Ambivalence −.13** −.25*** −.34*** −.51***
Intentions: medically actionable results .21*** .17*** .12* .30*** −.31***
Intentions: non-medically actionable results .20*** .18*** .20*** .52*** −.41*** .57***
Intentions: carrier results .22** .19** .23** .50** −.40** .63*** .86***
Means (SD) 4.74 (0.73) 4.16 (0.79) 4.35 (0.67) 6.41 (0.70) 1.60 (0.71) 5.57 (0.70) 5.28 (1.04) 5.64 (0.66)
Possible/ observed range 1-6/ 2-6 1-5/ 1.33-5 1-5/ 2-5 1-7/ 3.41-7 1-5/ 1-5 1-6/ 2.5-6 1-6/ 1-6 1-6/ 1-6
Open in a new tab
**

p ≤ .01;

***

p ≤ .001

a

To aid interpretation, mean, standard deviation, and possible and observed ranges are reported using the average of the raw item scores, prior to standardizing items and transforming scale scores.

With attitudes excluded from the models, there were main effects of both descriptive and injunctive norms over and above demographic factors when predicting intentions to learn medically actionable, non-medically actionable, and carrier sequencing results (see Table 2; all ps ≤ .004). Descriptive and injunctive norms explained 7%, 6%, and 7% of the variance in intentions to learn medically actionable, non-medically actionable, and carrier sequencing results, respectively. Given their overlapping confidence intervals (Cumming, 2009), the standardized coefficients indicated no difference in the magnitude of the associations of descriptive and injunctive norms with intentions for obtaining all results (medically actionable: βs = .20 vs. .15; non-medically actionable: βs = .17 vs. .16; carrier: βs = .19 vs. .16).

Table 2.

Main Effects of Norms and Covariates on Intentions

Intentions to learn medically actionable results Intentions to learn non-medically actionable results Intentions to learn carrier results
B SE p B SE p B SE p
Descriptive norms 0.04 0.01 .001 0.04 0.01 .002 0.04 0.01 .001
Friend injunctive norms 0.03 0.01 .004 0.04 0.01 .003 0.04 0.01 .002
Gender −0.02 0.02 .21 0.02 0.02 .24 0.02 0.02 .26
Ethnicity 0.04 0.04 .31 0.02 0.04 .56 0.06 0.04 .08
Age 0.00 0.00 .86 −0.00 0.00 .48 0.00 0.00 .84
Education 0.02 0.01 .03 −0.00 0.01 .96 0.01 0.01 .38
Income −0.00 0.01 .88 −0.01 0.01 .27 −0.01 0.01 .53
Open in a new tab

Mediation Analyses

We hypothesized that effects on intentions of injunctive norms but not descriptive norms would be mediated by attitudes. Injunctive norms (B = 0.35, SE = 0.04, p < .001) were significantly associated with attitudes. Despite the consistent main effects observed of injunctive norms on intentions, injunctive norms no longer predicted any of the intentions outcomes when attitudes were added to the models (medically actionable: B = 0.01, SE = 0.01, p = .49; non-medically actionable: B = -0.02, SE = 0.01, p = .20; carrier: B = -0.01, SE = 0.01, p = .32). However, attitudes significantly predicted intentions to learn medically actionable (B = 0.07, SE = 0.02, p ≤ .001), non-medically actionable (B = 0.17, SE = 0.02, p ≤ .001), and carrier sequencing results (B = 0.13, SE = 0.01, p ≤ .001). Consistent with this evidence of mediation, bias-corrected bootstrapped confidence intervals indicated that the effect of injunctive norms was significantly mediated by attitudes for intentions to learn medically actionable (mediated effect = .03; 95% CI [.01, .06]), non-medically actionable (mediated effect = .04; 95% CI [.02, .07]), and carrier sequencing results (mediated effect = .05; 95% CI [.03, .06]). As reported above, after controlling for attitudes, the direct effects of injunctive norms on intentions to learn non-medically actionable and carrier sequencing results were negative but non-significant. We therefore computed the sum of the absolute values of the mediated and direct effects and examined the proportion of this total effect that was mediated by attitudes ([mediated effect / (|mediated effect| + |direct effect|)]; MacKinnon, 2008). Attitudes explained 76% of the relationship between injunctive norms and intentions to learn medically actionable results, 77% of the injunctive norms to non-medically actionable result relationship, and 82% of the injunctive norms to carrier result relationship.

Descriptive norms were also associated with attitudes (B = 0.08, SE = 0.04, p = .03), and as reported above, attitudes were associated with all intentions outcomes. However, as expected, associations between descriptive norms and intentions were largely unchanged after controlling for attitudes. Descriptive norms remained a significant predictor of intentions to learn medically actionable (B = 0.03, SE = 0.01, p = .002), non-medically actionable (B = 0.03, SE = 0.01, p = .02), and carrier sequencing results (B = 0.03, SE = 0.01, p = .004). Bias-corrected bootstrapped confidence intervals further supported that the effect of descriptive norms was not mediated by attitudes when predicting intentions for learning medically actionable (mediated effect = .01; 95% CI [.00, .01]), non-medically actionable (mediated effect = .01; 95% CI [.00, .03]), and carrier sequencing results (mediated effect = .01; 95% CI [.00, .02]).

Moderator Analyses

Finally, we examined whether attitudinal ambivalence moderated the associations of descriptive and injunctive norms with intentions and attitudes (see Figure 1 for hypotheses). Ambivalence was significantly negatively associated with attitudes and each of the intentions measures, while controlling for attitudes (all p’s <.03). Contrary to hypotheses, attitudinal ambivalence did not moderate the effect of descriptive norms on intentions to learn medically actionable (B = 0.01, SE = 0.01, p = .70), non-medically actionable (B = 0.02, SE = 0.01, p = .09), or carrier sequencing results (B = 0.02, SE = 0.01, p = .11). The injunctive norms by ambivalence interactions were not significant when predicting intentions to learn non-medically actionable (B = 0.03, SE = 0.02, p = .09) or carrier sequencing results (B = 0.01, SE = 0.02, p = .42) but was significant for medically actionable results (B = 0.05, SE = 0.02, p = .01; R2change= .02). Simple slopes indicated a positive association of injunctive norms with intentions when ambivalence was high (B = 0.04, SE = 0.02, p = .03). When ambivalence was low, injunctive norms were unrelated to intentions (B = -0.02, SE = 0.02, p = .13).

When examining whether ambivalence moderated associations of norms with attitudes, results supported our hypothesis that the relationship of descriptive norms to attitudes would not depend on level of ambivalence (B = 0.04, SE = 0.03, p = .16) but that ambivalence would moderate the relationship of injunctive norms to attitudes (B = 0.25, SE = 0.04, p ≤ .001; R2change= .04; see Figure 2). Injunctive norms were strongly associated with attitudes when individuals were highly ambivalent (B = 0.43, SE = 0.06, p ≤ .001) but were unrelated to attitudes when individuals were low in ambivalence (B = 0.07, SE = 0.05, p = .14). These results support the idea that attitudinal ambivalence did not foster greater consideration of others’ behaviors (descriptive norms) to inform one’s own attitudes but did enhance incorporation of what is approved of by others (injunctive norms) into personal attitudes.

Figure 2.

Figure 2

Open in a new tab

Interaction between attitudinal ambivalence and friend injunctive norms predicting attitudes toward learning sequencing results.

Sensitivity Analysis

All analyses were re-examined for “family injunctive norms,” reflecting approval from children and siblings (n = 417)1. As expected, the mediated effects fully replicated. Attitudes mediated the associations of family injunctive norms with intentions to learn medically actionable (mediated effect = .03; 95% CI [.02, .05]), non-medically actionable (mediated effect = .06; 95% CI [.04, .08]), and carrier sequencing results (mediated effect = .05; 95% CI [.03, .08]), explaining 94%, 100%, and 87% of these associations, respectively. Likewise, descriptive norms predicted each intention outcome over and above attitudes (all ps ≤ .004). As descriptive norms were unrelated to attitudes in this context (B = 0.05, SE = 0.03, p = .17), attitudes could not mediate the association of descriptive norms with intentions.

With respect to moderation by ambivalence when predicting intentions, there were no interactions between family injunctive norms and attitudinal ambivalence for any of the intentions outcomes (all ps > .11). The descriptive norms by ambivalence interaction was non-significant for medically actionable and carrier intentions (ps > .08) but was significant for non-medically actionable intentions (B = 0.03, SE = 0.01, p = .04; R2change= .01). The association between descriptive norms and intentions was evident when ambivalence was high (B = 0.05, SE = 0.01, p < .001) but not when ambivalence was low (B = 0.01, SE = 0.01, p = .54). In comparison with the primary analyses, the significant friend injunctive norms by ambivalence interaction predicting medically actionable intentions failed to replicate, and the descriptive norms by ambivalence interaction predicting non-medically actionable intentions was only significant in the context of family injunctive norms. Thus, the norms by ambivalence interactions when predicting intentions were inconsistent across the two sets of analyses.

When predicting attitudes, the interactions of family injunctive norms and descriptive norms with ambivalence were fully consistent with the results for friend injunctive norms. Ambivalence moderated the association between family injunctive norms and attitudes (B = 0.25, SE = 0.07, p < .001; R2change= .03), but not the association between descriptive norms and attitudes (B = 0.05, SE = 0.05, p = .32). Simple slopes again supported that family injunctive norms were associated with attitudes when ambivalence was high (B = 0.44, SE = 0.07, p < .001) but not when ambivalence was low (B = 0.09, SE = 0.06, p = .13).

Discussion

Understanding when and why individuals opt to learn genetic risk information is increasingly important, given the growth in availability of this type of information. As most have little precedent for whether genetic risk information is likely to be beneficial or harmful, it is unsurprising that these decisions may be influenced by perceptions of others’ likely behaviors and attitudes. Our findings demonstrated the unique roles that descriptive and injunctive norms play in decisions to receive genomic sequencing results, an increasingly more accessible type of genetic testing that has the potential to benefit health by providing information about one’s entire genome. Controlling for one another’s effects, descriptive norms (reflecting how other study participants were likely to behave) were directly associated with intentions to learn sequencing results, whereas the effect of injunctive norms (reflecting approval from close others) was transmitted through personal attitudes toward learning sequencing results. Moreover, highlighting the necessary circumstances, injunctive norms were consistently unrelated to attitudes when individual’s attitudes were univalent and presumably strongly held.

It is worth noting that the observed effects of social norms emerged in a sample of middle-aged to older (Mage = 60), highly educated adults (89% had college or post-graduate degrees). Individuals in this middle-adulthood range, approximately aged 45-65, perceive themselves as highly knowledgeable and therefore hold strong attitudes about many topics (Visser & Krosnick, 1998). As a result, middle-aged adults are more resistant to changing their attitudes than younger and older adults (Visser & Krosnick, 1998), potentially thwarting an influence of norms. In addition, individuals raised in more educated households are less likely to shift their behavior in response to descriptive norms than are working-class individuals (Na, McDonough, Chan, & Park, 2016). Norms may therefore have an even stronger association with genomic sequencing intentions and attitudes in younger, older, or less educated populations.

Consistent with previous theorizing (Cialdini, 2003), our results indicated a direct effect of descriptive norms on intentions to learn genomic sequencing results and an indirect effect of injunctive norms, through attitudes. The results for injunctive norms support that conforming to what is approved of by others requires some degree of cognitive processing. Health decisions vary in personal relevance and importance, resulting in variation in the degree to which they invoke cognitive consideration. Injunctive norms may therefore be especially influential in the context of fairly deliberate decisions, such as obtaining genomic sequencing results, but may be less important for decisions marked by more automaticity, such as established tobacco use.

Attitudinal ambivalence primarily moderated the association between injunctive norms and attitudes. When predicting intentions, there was very limited support for interactions between ambivalence and both norms. Thus, on the whole, associations between norms and intentions are not likely to depend on attitudinal ambivalence. Rather, constructs capturing ambiguity and uncertainty that are more proximal to behavior, such as decisional conflict (O’Connor, 1995), may moderate associations between descriptive norms and genomic sequencing intentions. To the extent that decisional conflict reflects preoccupation with and emotional distress surrounding genetic testing decisions, information about how others have proceeded may serve as a mechanism for helping to resolve this conflict. More broadly, our results support that the utility of normative information is likely to be dependent on the strength of patients’ own beliefs. If genome sequencing preferences are strongly held, normative information is likely to be less consequential than in situations where the best course of action has not yet been decided (Hohman et al., 2016).

The present results have broader implications for how providers and genetic counselors might approach decisions surrounding genomic sequencing. Going beyond personal values and beliefs to consider important others’ views may help the client and counselor to better understand the factors contributing to ambivalence. Addressing these factors may lead to patients making more informed choices and experiencing less decisional conflict and regret. Educating counselors about the role and importance of psychosocial constructs presents the ideal means for utilizing psychological research to benefit population health. Clinicians and medical students have been successfully trained to elicit and respond to patient statements indicating negative attitudes and norms toward engaging in health promoting behaviors (Chisholm et al., 2016; Fisher et al., 2004), providing a model for how genetic counselors might be similarly trained.

Our results specifically suggest an important role for patient stories, which provide information about the choices made by similar patients and have been identified as an important component of patient decision aids (Elwyn et al., 2006). Likewise, for patients who express ambivalence about whether to undergo testing or how to proceed given the results, exploring perceived approval from close others for various options might be useful for resolving ambivalence and nudging behavior in one direction (O’connor, Jacobsen, & Stacey, 2002). Individuals may also sensor their own beliefs, especially in situations where the results of sequencing are threatening and require making difficult decisions (e.g., non-medically actionable results). Toward the goal of achieving shared decision making between patients and practitioners (Makoul & Clayman, 2006), participants’ views of others’ behaviors or level of approval may provide insight into their own likely actions and attitudes.

The present research should be considered in light of the study limitations. First, we relied on cross-sectional data. However, as bivariate associations typically become weaker over time, it seems unlikely that attitudes would mediate the descriptive norms-intentions relationship in longitudinal data. Moreover, given support in experimental data (Hohman et al., 2016; Reid & Aiken, 2013), we expect that mediation of injunctive norms by attitudes would also hold in longitudinal data. Second, although intentions are proximal to behavior, demonstrating that only descriptive norms directly influence behavior would provide additional support for our hypotheses. In the context of genomic sequencing, assessing behavior is difficult as provision of results depends on the availability of such results. Third, generalizability of results to individuals who decline enrollment in sequencing trials, and tend to be less positive toward genomic sequencing (Robinson et al., 2016), is unclear. However, as decliners are likely to be highly ambivalent, descriptive and injunctive norms may both play strong roles in decision making.

In sum, genomic sequencing holds substantial promise for improving health through early identification of those at greatest risk. Sequencing is currently being used in specialty clinics to diagnose rare conditions and to identify genetic contributions to autism and developmental disorders (Lim et al., in press; Sawyer et al., 2016). Genomic sequencing will increasingly become part of mainstream medicine, with projections of routine use to inform pharmaceutical use to treat disease. Further delineation of the role of psychosocial factors in whether individuals both seek out and choose to learn sequencing results will enhance our ability to guide early adoption into practice (Gray et al., 2014), and ultimately, to fully realize the benefits of sequencing for improving population health.

Acknowledgments

This research was supported by the Intramural Research Program of the National Human Genome Research Institute.

Footnotes

1

In an additional sensitivity analysis, doctor injunctive norms replicated the pattern of effects reported for family injunctive norms.

References

  1. Aiken LS, West SG. Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage; 1991. [Google Scholar]
  2. Aro AR, Hakonen A, Hietala M, Lönnqvist J, Niemelä P, Peltonen L, Aula P. Acceptance of genetic testing in a general population: Age, education and gender differences. Patient Education and Counseling. 1997;32:41–49. doi: 10.1016/s0738-3991(97)00061-x. http://dx.doi.org/10.1016/S0738-3991(97)00061-X. [DOI] [PubMed] [Google Scholar]
  3. Biesecker LG. Opportunities and challenges for the integration of massively parallel genomic sequencing into clinical practice: lessons from the ClinSeq project. Genetics in Medicine. 2012;14:393–398. doi: 10.1038/gim.2011.78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Biesecker LG, Green RC. Diagnostic clinical genome and exome sequencing. New England Journal of Medicine. 2014;370:2418–2425. doi: 10.1056/NEJMra1312543. [DOI] [PubMed] [Google Scholar]
  5. Chisholm A, Hart J, Mann K, Perry M, Duthie H, Rezvani L, Peters S. Investigating the feasibility and acceptability of health psychology-informed obesity training for medical students. Psychology, health & medicine. 2016;21:368–376. doi: 10.1080/13548506.2015.1062523. [DOI] [PubMed] [Google Scholar]
  6. Christensen PN, Rothgerber H, Wood W, Matz DC. Social norms and identity relevance: A motivational approach to normative behavior. Personality and Social Psychology Bulletin. 2004;30:1295–1309. doi: 10.1177/0146167204264480. [DOI] [PubMed] [Google Scholar]
  7. Cialdini RB. Crafting normative messages to protect the environment. Current Directions in Psychological Science. 2003;12:105–109. [Google Scholar]
  8. Cialdini RB, Reno RR, Kallgren CA. A focus theory of normative conduct: Recycling the concept of norms to reduce littering in public places. Journal of Personality and Social Psychology. 1990;58:1015. [Google Scholar]
  9. Cialdini RB, Trost MR. Social influence: Social norms, conformity and compliance. In: Gilbert D, Fiske S, Lindzey G, editors. The Handbook of Social Psychology. 4th. Vol. 2. New York: McGraw-Hill; 1998. pp. 151–192. [Google Scholar]
  10. Collins J, Ryan L, Truby H. A systematic review of the factors associated with interest in predictive genetic testing for obesity, type II diabetes and heart disease. Journal of Human Nutrition and Dietetics. 2014;27:479–488. doi: 10.1111/jhn.12179. [DOI] [PubMed] [Google Scholar]
  11. Conner M, Armitage CJ. Attitudinal ambivalence. In: Crano WD, Prislin R, editors. Attitudes and attitude change. New York, NY: Taylor and Francis; 2008. pp. 261–286. [Google Scholar]
  12. Conner M, Povey R, Sparks P, James R, Shepherd R. Moderating role of attitudinal ambivalence within the theory of planned behaviour. The British Journal of Social Psychology. 2003;42:75. doi: 10.1348/014466603763276135. [DOI] [PubMed] [Google Scholar]
  13. Conner M, Sparks P. Ambivalence and attitudes. European Review of Social Psychology. 2002;12:37–70. doi: 10.1080/14792772143000012. [DOI] [Google Scholar]
  14. Cumming G. Inference by eye: Reading the overlap of independent confidence intervals. Statistics in Medicine. 2009;28:205–220. doi: 10.1002/sim.3471. [DOI] [PubMed] [Google Scholar]
  15. Daoud H, Luco SM, Li R, Bareke E, Beaulieu C, Jarinova O, Dyment DA. Next-generation sequencing for diagnosis of rare diseases in the neonatal intensive care unit. Canadian Medical Association Journal. 2016;188:E254–E260. doi: 10.1503/cmaj.150823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Dormandy E, Hankins M, Marteau TM. Attitudes and uptake of a screening test: The moderating role of ambivalence. Psychology & Health. 2006;21:499–511. doi: 10.1080/14768320500380956. [DOI] [Google Scholar]
  17. Elwyn G, O’Connor A, Stacey D, Volk R, Edwards A, Coulter A, Bernstein S. Developing a quality criteria framework for patient decision aids: Online international Delphi consensus process. Bmj. 2006;333:417–419. doi: 10.1136/bmj.38926.629329.AE. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Facio FM, Brooks S, Loewenstein J, Green S, Biesecker LG, Biesecker BB. Motivators for participation in a whole-genome sequencing study: Implications for translational genomics research. European Journal of Human Genetics. 2011;19:1213–1217. doi: 10.1038/ejhg.2011.123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ferrer RA, Taber JM, Klein WM, Harris PR, Lewis KL, Biesecker LG. The role of current affect, anticipated affect and spontaneous self-affirmation in decisions to receive self-threatening genetic risk information. Cognition and Emotion. 2015;29:1456–1465. doi: 10.1080/02699931.2014.985188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Fisher JD, Cornman DH, Osborn CY, Amico KR, Fisher WA, Friedland GA. Clinician-initiated HIV risk reduction intervention for HIV-positive persons: Formative research, acceptability, and fidelity of the Options Project. JAIDS Journal of Acquired Immune Deficiency Syndromes. 2004;37:S78–S87. doi: 10.1097/01.qai.0000140605.51640.5c. [DOI] [PubMed] [Google Scholar]
  21. Goldstein NJ, Cialdini RB, Griskevicius V. A room with a viewpoint: Using social norms to motivate environmental conservation in hotels. Journal of Consumer Research. 2008;35:472–482. [Google Scholar]
  22. Gray SW, Martins Y, Feuerman LZ, Bernhardt BA, Biesecker BB, Christensen KD, McGuire AL. Social and behavioral research in genomic sequencing: Approaches from the Clinical Sequencing Exploratory Research Consortium Outcomes and Measures Working Group. Genetics in Medicine. 2014;16:727–735. doi: 10.1038/gim.2014.26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Green RC, Berg JS, Grody WW, Kalia SS, Korf BR, Martin CL, Biesecker LG. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genetics in Medicine. 2013;15:565–574. doi: 10.1038/gim.2013.73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Green RC, Lautenbach D, McGuire AL. GINA, genetic discrimination, and genomic medicine. New England Journal of Medicine. 2015;372:397–399. doi: 10.1056/NEJMp1404776. [DOI] [PubMed] [Google Scholar]
  25. Haga SB, Barry WT, Mills R, Ginsburg GS, Svetkey L, Sullivan J, Willard HF. Public knowledge of and attitudes toward genetics and genetic testing. Genetic testing and molecular biomarkers. 2013;17:327–335. doi: 10.1089/gtmb.2012.0350. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Han PKJ, Umstead KL, Bernhardt BA, Green RC, Joffe S, Koenig B, Biesecker BB. A taxonomy of medical uncertainties in clinical genome sequencing. Genetics in Medicine. doi: 10.1038/gim.2016.212. (in press) [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Haque IS, Lazarin GA, Kang H, Evans EA, Goldberg JD, Wapner RJ. Modeled fetal risk of genetic diseases identified by expanded carrier screening. JAMA. 2016;316:734–742. doi: 10.1001/jama.2016.11139. [DOI] [PubMed] [Google Scholar]
  28. Hensley Alford S, McBride CM, Reid RJ, Larson EB, Baxevanis AD, Brody LC. Participation in genetic testing research varies by social group. Public Health Genomics. 2011;14:85–93. doi: 10.1159/000294277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hohman ZP, Crano WD, Niedbala EM. Attitude ambivalence, social norms, and behavioral intentions: Developing effective antitobacco persuasive communications. Psychology of Addictive Behaviors. 2016;30:209–219. doi: 10.1037/adb0000126. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hohman ZP, Crano WD, Siegel JT, Alvaro EM. Attitude ambivalence, friend norms, and adolescent drug use. Prevention Science. 2014;15:65–74. doi: 10.1007/s11121-013-0368-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Johansen Taber KA, Dickinson BD, Wilson M. The promise and challenges of next-generation genome sequencing for clinical care. JAMA Internal Medicine. 2014;174:275–280. doi: 10.1001/jamainternmed.2013.12048. [DOI] [PubMed] [Google Scholar]
  32. Kenny DA, Kashy DA, Cook WL. Dyadic data analysis. New York: Guilford Press; 2006. [Google Scholar]
  33. Lim ET, Uddin M, De Rubeis S, Chan Y, Kamumbu AS, Zhang X, Walsh CA. Rates, distribution and implications of postzygotic mosaic mutations in autism spectrum disorder. Nature Neuroscience. doi: 10.1038/nn.4598. (in press) [DOI] [Google Scholar]
  34. Lipkus IM, Green LG, Marcus A. Manipulating perceptions of colorectal cancer threat: implications for screening intentions and behaviors. Journal of Health Communication. 2003;8:213–228. doi: 10.1080/10810730305684. [DOI] [PubMed] [Google Scholar]
  35. MacKinnon DP. Introduction to statistical mediation analysis. Mahwah, NJ: Erlbaum; 2008. [Google Scholar]
  36. Makoul G, Clayman ML. An integrative model of shared decision making in medical encounters. Patient Education and Counseling. 2006;60:301–312. doi: 10.1016/j.pec.2005.06.010. [DOI] [PubMed] [Google Scholar]
  37. Montenegro G, Powell E, Huang J, Speziani F, Edwards YJK, Beecham G, Züchner S. Exome sequencing allows for rapid gene identification in a Charcot-Marie-Tooth family. Annals of Neurology. 2011;69:464–470. doi: 10.1002/ana.22235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Muthén LK, Muthén BO. Mplus user’s guide. 7th. Los Angeles, CA: Muthén & Muthén; 2015. [Google Scholar]
  39. Na J, McDonough IM, Chan MY, Park DC. Social-class differences in consumer choices: Working-class individuals are more sensitive to choices of others than middle-class individuals. Personality and Social Psychology Bulletin. 2016;42:430–443. doi: 10.1177/0146167216634043. [DOI] [PubMed] [Google Scholar]
  40. Neighbors C, O’connor RM, Lewis MA, Chawla N, Lee CM, Fossos N. The relative impact of injunctive norms on college student drinking: The role of reference group. Psychology of Addictive Behaviors. 2008;22:576–581. doi: 10.1037/a0013043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. O’Connor AM. Validation of a decisional conflict scale. Medical Decision Making. 1995;15:25–30. doi: 10.1177/0272989X9501500105. [DOI] [PubMed] [Google Scholar]
  42. O’connor AM, Jacobsen MJ, Stacey D. An evidence‐based approach to managing women’s decisional conflict. Journal of Obstetric, Gynecologic, & Neonatal Nursing. 2002;31:570–581. doi: 10.1111/j.1552-6909.2002.tb00083.x. [DOI] [PubMed] [Google Scholar]
  43. Reid AE, Aiken LS. Correcting injunctive norm misperceptions motivates behavior change: A randomized controlled sun protection intervention. Health psychology. 2013;32:551–560. doi: 10.1037/a0028140. [DOI] [PubMed] [Google Scholar]
  44. Robinson JO, Carroll TM, Feuerman LZ, Perry DL, Hoffman-Andrews L, Walsh RC, Team, M. P. Participants and study decliners’ perspectives about the risks of participating in a clinical trial of whole genome sequencing. Journal of Empirical Research on Human Research Ethics. 2016;11:21–30. doi: 10.1177/1556264615624078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sapp JC, Hull SC, Duffer S, Zornetzer S, Sutton E, Marteau TM, Biesecker BB. Ambivalence toward undergoing invasive prenatal testing: An exploration of its origins. Prenatal diagnosis. 2010;30:77–82. doi: 10.1002/pd.2343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Sawyer SL, Hartley T, Dyment DA, Beaulieu CL, Schwartzentruber J, Smith A, Boycott KM. Utility of whole-exome sequencing for those near the end of the diagnostic odyssey: Time to address gaps in care. Clinical Genetics. 2016;89:275–284. doi: 10.1111/cge.12654. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Solomon BD, Hadley DW, Pineda-Alvarez DE, Kamat A, Teer JK, Cherukuri PF, Berkman BE. Incidental medical information in whole-exome sequencing. Pediatrics. 2012;129:e1605–e1611. doi: 10.1542/peds.2011-0080. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sweeny K, Ghane A, Legg AM, Huynh HP, Andrews SE. Predictors of genetic testing decisions: A systematic review and critique of the literature. Journal of Genetic Counseling. 2014;23:263–288. doi: 10.1007/s10897-014-9712-9. [DOI] [PubMed] [Google Scholar]
  49. Tabachnick BG, Fidell LS. Using multivariate statistics. 6th. Boston: Pearson; 2013. [Google Scholar]
  50. Taber JM, Klein WM, Ferrer RA, Han PK, Lewis KL, Biesecker LG, Biesecker BB. Perceived ambiguity as a barrier to intentions to learn genome sequencing results. Journal of behavioral medicine. 2015;38:715–726. doi: 10.1007/s10865-015-9642-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Taber JM, Klein WM, Ferrer RA, Lewis KL, Biesecker LG, Biesecker BB. Dispositional optimism and perceived risk interact to predict intentions to learn genome sequencing results. Health psychology. 2015;34:718–728. doi: 10.1037/hea0000159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Taber JM, Klein WM, Ferrer RA, Lewis KL, Harris PR, Shepperd JA, Biesecker LG. Information avoidance tendencies, threat management resources, and interest in genetic sequencing feedback. Annals of Behavioral Medicine. 2015;49:616–621. doi: 10.1007/s12160-014-9679-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. The White House. The precision medicine initiative. 2015 from https://obamawhitehouse.archives.gov/precision-medicine.
  54. Vermeulen E, Henneman L, van El CG, Cornel MC. Public attitudes towards preventive genomics and personal interest in genetic testing to prevent disease: A survey study. The European Journal of Public Health. 2014;24:768–775. doi: 10.1093/eurpub/ckt143. [DOI] [PubMed] [Google Scholar]
  55. Visser PS, Krosnick JA. Development of attitude strength over the life cycle: surge and decline. Journal of Personality and Social Psychology. 1998;75:1389–1410. doi: 10.1037//0022-3514.75.6.1389. [DOI] [PubMed] [Google Scholar]

RESOURCES