Deaf Genetic Testing and Psychological Well-Being in Deaf Adults

Limited data suggest that enhanced self-knowledge from genetic information related to non-medical traits can have a positive impact on psychological well-being. Deaf individuals undertake genetic testing for deaf genes to increase self-knowledge. Because deafness is considered a non-medical trait by many individuals, we hypothesized that deaf individuals receiving a genetic explanation for why they are deaf will experience increased psychological well-being. We report results from a prospective, longitudinal study to determine the impact of genetic testing (GJB2, Cx26; GJB6, Cx30) on perceived personal control (PPC), anxiety, and depression in deaf adults (N = 209) assessed following pre-test genetic counseling as well as 1-month and 6-months following test result disclosure. Participants were classified as Cx positive (n = 82) or Cx negative/inconclusive (n = 127). There was significant evidence for Cx group differences in PPC and anxiety over time (PPC: Cx group*time interaction p = 0.0007; anxiety: Cx group*time interaction p = 0.002), where PPC scores were significantly higher, and anxiety scores were significantly lower for the Cx positive group relative to the negative/inconclusive group following test result disclosure. Compared to pre-test, PPC scores increased at 1-month (p = 0.07) and anxiety scores decreased at 6-months (p = 0.03) for the Cx positive group. In contrast, PPC scores decreased (p = 0.009, p < 0.0001) and anxiety scores increased (p = 0.09, p = 0.02) for the Cx negative/inconclusive group at 1- and 6-months post test result disclosure. Genetic testing for deaf genes affects the psychological well-being of deaf individuals. Increasing deaf adults’ access to genetic testing may potentially enhance self-knowledge and increase psychological well-being for those who receive a genetic explanation, which could offer downstream health benefits.


Introduction
Genetic testing for deaf genes 1 is widely available and there is evidence that deaf 2 individuals are interested in such testing to enhance self-knowledge . Genetic information has the potential to produce psychological sequelae; however, there are no empirical data on the psychological outcomes of deaf genetic testing in deaf adults. This is a significant deficit of information because deaf individuals face higher mental health burdens than hearing individuals (Kvam et al. 2007), due to factors such as access to effective communication, childhood trauma including abuse, socioeconomic issues, and experiences with stigma and discrimination (Fellinger et al. 2012). This mental health disparity heightens concerns about the potential psychological impact of genetic information particularly in light of the scarcity of mental health services tailored for individuals who use a signed language (Fellinger et al. 2012). The purpose of the current study is to examine the psychological impact of genetic testing for deafness on deaf adults. Identifying the psychological outcomes of genetic information is important for designing genetic counseling strategies, for providing anticipatory genetic counseling, and for assessing the relationship to health outcomes.
There has been considerable research on psychological outcomes of predictive and diagnostic genetic testing in adults. This body of research has reliably demonstrated that individuals who test "positive" for a disease-causing variant experience more psychological distress than those who test "negative" (Broadstock et al. 2000;Hamilton et al. 2009;Shaw et al. 1999;Vansenne et al. 2009). However, these studies have been conducted on adults with or at risk for a variety of medical conditions, i.e., conditions which require treatment by medicine. It is not entirely clear that those results can be generalized to genetic testing for deafness because for many individuals deafness is considered a non-medical trait (Robertson 2003) that does not require medical intervention.
The view of deafness as a non-medical trait is mainly held by Deaf individuals and hearing individuals affiliated with the Deaf community. Deaf communities have developed throughout the world over the last 200 years. These communities have a distinct culture with their own beliefs, customs, attitudes, language, and behavioral norms that are actively transmitted across generations (Padden 1989;Padden and Humphries 1988). American Sign Language (ASL) is the primary language of the Deaf community in the United States; and when including individuals who use ASL as a second language, this is one of the most frequently used languages in the US after English and Spanish (Mitchell et al. 2006). Deaf communities are vibrant, closely-knit communities comprised of individuals who embrace deafness as a valued personal characteristic (Jacobs 1989), not a disability or medical condition.
To date, there has been very little research on the impact of genetic information for non-medical traits on individuals from which to generate hypotheses about the psychological impact of deaf genetic testing on deaf adults. However, there is some evidence that enhanced self-knowledge from genetic information related to deafness would have a positive impact on the psychological well-being of deaf individuals. In a study examining the extent to which stigma is attached to the genetic basis of a variety of conditions (Sankar et al. 2006), the authors found that compared to individuals with medical conditions such as breast cancer, Deaf individuals felt less stigma and made more positive statements about a genetic or hereditary basis to their trait because of the importance of deafness to their cultural group identity and individual identity. Although that study did not offer or study outcomes of genetic testing, the results suggest that deaf individuals' psychological well-being will be influenced in part by genetic information, and that those who receive a genetic test result that explains why they are deaf will experience enhanced psychological well-being.
Studies of psychological outcomes of genetic testing in adults almost uniformly assess individuals' levels of anxiety and depression before and after testing. In the assessment of genetic testing for non-medical traits in adults, perceived personal control-the sense that one has behavioral, cognitive, and decisional control over aspects of their life (Berkenstadt et al. 1999)-may also be an important psychological outcome. In a study of muscle-related non-medical 1 Use of the terms "deaf genes" and "deaf genetic testing" is intentional. "deaf genes" and "deaf genetic testing" are the American Sign Language-deaf-centric equivalent of the English language-hearingcentric concepts of "genes for deafness"/"genetic testing for deafness". This is analogous to the idea that deaf people do not use people-first terminology to describe themselves (i.e., it is appropriate to say "deaf person" and inappropriate to say "person who is deaf" or "person with deafness"). Genetic counselors will be exposed to the terms "deaf genes" and "deaf genetic testing" when they provide genetic counseling to deaf adults interested in this testing. 2 We use the term deaf (with a lower case d) to refer simply to an audiologic phenotype of hearing loss. Deaf (with a capital D) is used to refer to individuals who are members of the Deaf community, a distinct cultural group. The term Deaf is distinct from deaf; individuals within the Deaf community may be deaf or hard-of-hearing (Senghas and Monaghan 2002). traits, individuals who received neutral genotypes for these traits (analogous to a "negative" result) experienced improved psychological well-being, measured by self-concept and health perception, compared to individuals who received positive genotype information (Gordon et al. 2005). Although this result seems counter-intuitive, the authors hypothesized that those with neutral genotypes experienced positive well-being because they felt a greater sense of personal control over the ability to change their bodies through exercise than individuals with positive genotypes in which "genetics" held more control over their bodies. Thus, it appears that the psychological impact of genetic information on non-medical traits may depend on whether the genetic information augments or diminishes perceived personal control and highlights the importance of assessing this psychological variable.
In the current study, we examine the effect of receiving a GJB2/GJB6 test result on deaf adults' levels of perceived personal control, anxiety, and depression with data collected from a prospective, longitudinal genetic counseling and testing study (Deaf Genetics Project). Given that deaf individuals may feel positively toward a genetic or hereditary basis to deafness, we hypothesized that deaf individuals receiving a genetic explanation for why they are deaf will experience enhanced psychological well-being compared to those who do not receive a clear genetic explanation for why they are deaf.

Research Design and Research Team
The Deaf Genetics Project is a prospective, longitudinal study to examine the impact of genetic counseling and genetic testing (GJB2 and GJB6 genes) on deaf adults and the deaf community. The Deaf Genetics Project was designed by a multiinstitutional, multi-disciplinary team of Deaf, hard of hearing and hearing investigators. The team employed a collaborative research model that integrated the Deaf cultural perspective, the hearing cultural perspective, the academic cultural perspective, and the community service perspective into the research design and implementation. In addition, three certified ASL/English interpreters were members of the project staff. These bilingual, bicultural interpreters were certified by the Registry of Interpreters for the Deaf (RID), a U.S.-based organization that provides interpreters with a code of conduct and ethical guidelines (Registry of Interpreters for the Deaf 2005). All members of the research team, including the genetic counselors, attended approximately 10 h of Deaf cultural sensitivity workshops to enhance their genetic counseling interactions with deaf and hard-of-hearing individuals. To facilitate accurate and consistent explanations of complex genetic information, the staff interpreters and Deaf researchers consulted with several Deaf and hearing academics who had familiarity with ASL terminology for genetics terms and concepts to ensure that appropriate signs were used and the interpreters were trained extensively in the relevant genetics topics. Details on the participants and study protocol are described in this section and also have been published elsewhere (Baldwin et al. 2012;Boudreault et al. 2010).
Although there are many genes with variants that can lead to deafness, the Deaf Genetics Project focused on GJB2 and GJB6. This is because nearly half of individuals with non-syndromic genetic deafness will have a recessive variant in the GJB2 gene encoding the Connexin 26 protein (Denoyelle et al. 1997;Kelsell et al. 1997;Rabionet et al. 2000;Zelante et al. 1997), making this the most common worldwide explanation for hereditary deafness (Prasad et al. 2000) and the most commonly recommended genetic test for non-syndromic sensorineural deafness prior to the availability of next generation sequencing (NGS) technologies. Furthermore, it is also known that a large deletion in the GJB6 gene encoding the Connexin 30 protein, del(GJB6-D13S1830), coupled with a GJB2 deafness-causing variant can result in deafness (Pallares-Ruiz et al. 2002), and is often tested along with GJB2. From this point on we use the more commonly used abbreviations Cx26 and Cx30 in place of GJB2 and GJB6, and refer more generally to Cxrelated deafness.

Participants
Individuals who were at least 18 years old with an unexplained sensorineural deafness since an early age (defined as birth to age 6 years) were eligible to participate. Participants were recruited over a period of 25 months from the Los Angeles, Bay Area, and Riverside areas of California through a variety of venues using deaf-friendly recruitment materials. Interested individuals initiated contact with project personnel via point-to-point webcam communication (e.g., videophone), TTY, email, or voice telephone.

Study Protocol
The study protocol had three stages ( Fig. 1) and an interpreter was available for all participants during the course of the research protocol unless the participant opted to use spoken English without an interpreter. The staff interpreters often also functioned as the primary contact for many of the deaf participants in the project, helping to demonstrate the project's commitment to cultural and linguistic sensitivity. For all three stages, individuals selected one of four locations for their participation: University of California Los Angeles, California State University Northridge, California School for the Deaf-Fremont, or California School for the Deaf-Riverside.
Stage I: Interested individuals contacted study personnel and completed a brief screening questionnaire to determine initial eligibility given that Cx-related deafness is typically an early-onset, apparently nonsyndromic sensorineural deafness. Our screening questionnaire assessed age at diagnosis of deafness, and if the individual had ever been told that their deafness was associated with Waardenburg syndrome, Pendred syndrome, Jervell-Lange-Nielsen syndrome, Alport syndrome, Branchio-Oto-Renal syndrome, Treacher Collins syndrome, Usher syndrome, Stickler syndrome, or childhood meningitis. The screening questionnaire also assessed if deaf genetic testing (including Cx26 and Cx30) had ever been performed and the results. Individuals whose responses indicated that they had become deaf after age 6 years, had a syndromic form of deafness, had a genetic explanation for why they are deaf, or had already had full Cx26 sequencing and Cx30 deletion testing were ineligible to participate. Individuals who screened eligible were invited to undergo an audiology evaluation to confirm the presence of sensorineural deafness. The purpose of the screening questionnaire and audiology evaluation was to ensure that Cx26 and Cx30 genetic testing was offered only to individuals for whom it was potentially relevant, i.e., those with early-onset, apparently nonsyndromic sensorineural deafness.
Stage II: Participants met with one of four board-certified genetic counselors for an in-person pre-test genetic counseling session. During the semi-structured pre-test genetic counseling sessions, the genetic counselor explained the remaining study protocol and participants were informed that the overall goal of the study was to "learn what deaf/hard-of-hearing individuals think about genetic counseling and testing, what it means for their lives, and what it means for the Deaf community" to allay potential concerns that the study had a medical or disability focus. The genetic counselor provided information on the general etiologies of deafness, basic genetics concepts, and general information about genetic testing and genetic services. During the pretest genetic counseling session, information about Cx26, Cx30, and their autosomal recessive mechanism of inheritance was discussed. The limitation that the study was offering genetic testing for only two of the more than 60 known nonsydromic deaf genes (Shearer and Smith 2012) were discussed in detail.
Participants interested in pursuing genetic testing signed a consent form. The genetic counselor then queried the participant about deaf, hard-of-hearing, and hearing relatives using a standardized protocol and generated a detailed 3-generation pedigree. A buccal sample was also obtained from the participant for genetic analysis. The buccal sample was sent to the CLIA-certified UCLA Orphan Disease Testing Center for Cx26 sequencing and testing of the 309 kb deletion [del(GJB6-D13S1830)] in Cx30 (del Castillo et al. 2002). Genetic counseling sessions were standardized as much as possible, using a genetic counseling notebook with visual aids and identifying key concepts to routinely discuss.
Stage III: Participants returned for another in-person genetic counseling session when the genetic test results were available. The genetic counselor explained the genetic test results, put them in the context of the participant's family and medical history, and answered participants' questions. All participants received a copy of their genetic test report and a genetic counseling summary letter. In some cases, participants received additional information about genetics Fig. 1 Deaf Genetics Project. Study protocol, sample sizes, and questionnaire response rates clinics in their area, either because they were interested in continuing to try to learn why they are deaf, or because something of clinical importance with a strong genetic component that put their health at risk was noted in the family history, e.g., early onset breast cancer. We did not collect information on whether or not participants pursued additional genetics evaluations.
Participants determined eligible to participate in Stages II and III were asked to complete a series of four questionnaires. The first questionnaire was completed immediately following audiologic confirmation of eligibility for the genetic counseling and testing stage of the study (baseline questionnaire). The second questionnaire was completed immediately following the pre-test genetic counseling session (pre-test questionnaire). One month and 6 months after participants received their genetic test results, they were asked to complete the third questionnaire (1-month posttest questionnaire) and the final questionnaire (6-months post-test questionnaire), respectively. All four study questionnaires assessed nearly identical information to examine the effect of genetic information in a longitudinal framework.
This study was approved by the relevant institutional review boards. Genetic counseling and genetic testing were provided at no charge to participants in the study. All participants provided informed consent, and as part of this process they were informed that they could pursue genetic counseling and genetic testing outside of this study on their own. All DNA samples were destroyed at the end of the study.

Instrumentation
The study questionnaires assessed demographic factors, perceived personal control, anxiety, and depression, as well as attitudes toward genetic testing, knowledge and understanding of genetics and genetic testing, deaf identity, and a variety of behaviors. All questionnaire items were translated into ASL and Spanish and back-translated to ensure accuracy and equivalency of meaning (Brislin 1970). Study questionnaires were available online in ASL only, English only, ASL/English, and Spanish only, and participants were able to complete the questionnaires in the language of their choice. To assess the impact of genetic test results on participants' psychological well-being over time, we focus on responses to psychological measures collected at three timepoints: immediately following pre-test genetic counseling, and 1-month-and 6-months post-test results.

Primary Outcome Variables
Standard scales were used to assess psychological outcomes of genetic testing. Perceived personal control (PPC) was assessed using a modified version of the PPC scale designed and validated for use in studies of the psychological impact of genetic counseling (Berkenstadt et al. 1999). The PPC scale contains nine statements covering behavioral, cognitive, and decisional control. Behavioral control refers to the availability of an instrumental response, cognitive control addresses the ability to process information so as to reduce stress, and decision control captures the opportunity to choose among various courses of action. Respondents rated each item on a five-point Likert scale of strongly agree to strongly disagree. The overall PPC score is recommended for use in analyses (Smets et al. 2006) and has good reliability with Cronbach's α ranging from 0.78 to 0.83 (Aalfs et al. 2007;Berkenstadt et al. 1999;Smets et al. 2006). As adapted for this study, scores range from 9 to 45 with higher scores indicating higher perceived personal control. A PPC score was generated for each respondent at each time point. In this sample, Cronbach's alpha for the PPC items across the assessment time points ranged from 0.80 to 0.86.
Anxiety was assessed using the State component of the Spielberger State-Trait Anxiety Inventory (STAI-S) (Spielberger 1993), which assesses situation specific level of anxiety. The inventory is composed of 20 statements that a respondent assesses on a four-point Likert scale of "not at all" to "very much so". Scores range from 20 to 80 with higher scores indicating greater anxiety. Scores ≥40 are commonly taken to indicate clinically significant symptoms (Addolorato et al. 1999;Forsberg and Bjorvell 1993;Knight et al. 1983). An STAI-S score was generated for each respondent at each time point using the instrument's algorithm. This instrument has been validated in numerous studies and has been widely used in studies of the psychological impact of genetic testing (Kasparian et al. 2007). In this sample, Cronbach's alpha for the STAI-S ranged from 0.91 to 0.94 across the assessment time points.
Depression was assessed using the Center for Epidemiological Studies-Depression (CES-D) scale (Radloff 1977). This instrument consists of 20 items that measure the frequency of symptoms during the last week. Each symptom is evaluated on a four-point scale of "rarely or none of the time" to "all of the time". Scores range from 0 to 60 with higher scores indicating higher depression. Scores ≥16 are commonly taken to indicate depression. This instrument has been validated in numerous studies and has been widely used in studies of the psychological impact of genetic counseling and testing (Kasparian et al. 2007). A CES-D score was generated for each respondent at each time point. In this sample, Cronbach's alpha for the CES-D scale items across the assessment time points ranged from 0.87 to 0.90.

Primary Predictor Variables
Cx group-To assess the impact of genetic test results, participants were classified into one of two groups: Cx positive or Cx negative/inconclusive. Individuals were classified as Cx positive if their genetic test result clearly explained why they are deaf, i.e., they have two known Cx26 or Cx30 deafness-causing variants. Individuals were classified as Cx negative/inconclusive if their genetic test result clearly indicated that they do not have a Cx-related deafness, or if their genetic test result did not provide enough information to determine if they have Cx-related deafness. The Cx negative and Cx inconclusive groups are treated as a single group because preliminary analyses demonstrated that they had similar patterns of PPC, STAI-S, and CES-D scores (p's>0.05).

Demographic Characteristics
Age, sex, ethnicity/race, income (in $15 K increments), highest level of education achieved, linguistic preference during the audiology and genetic counseling and testing sessions, family history of deafness, and cultural affiliation were also assessed. Positive family history was defined two different ways: (1) Individuals with at least one first or second degree relative with early onset deafness; (2) individuals with at least one deaf parent. For comparability with previous publications from the Deaf Genetics Project (Baldwin et al. 2012;Boudreault et al. 2010), we report only the results from the first definition because they do not substantively differ from the second definition.
In addition, enrollment site (four sites), genetic counselor (four counselors), and interpreter were recorded for each participant. The latter variable was coded with four categories (one for each of the three interpreters plus a category for no interpreter) to capture potential nuances in communication during the genetic counseling sessions. Since our study protocol did not require participants to receive their test results at a pre-specified interval following buccal sampling/consent, we also assessed the interval between sample collection and results disclosure.

Data Analysis
Subjects without data on both the 1-month and 6-month questionnaires are excluded from all analyses. To maximize the remaining sample size, we used simple imputation to fill in missing data on individual PPC, STAI-S, and CES-D scale items (affecting ≤1 % of the PPC, STAI-S, and CES-D scale items), with the exception that scale scores were not computed for the ≤1 % of subjects with substantial missing data.
Descriptive statistics were produced and reviewed for the presence of outliers and data errors. PPC, STAI-S, and CES-D difference scores between pre-test and 1-month, and pretest and 6-month were computed, and subjects with difference scores >3 SDs from the mean were excluded from the relevant analyses as outliers. Potential confounders with the primary outcome variables or the Cx grouping variable were assessed using Chi-square or Fisher's Exact, analysis of variance, or Pearson's product correlation.
For the main analyses, we first assessed whether there were between-group differences in psychological effect due to the genetic test result over time. To address this question we performed repeated measures regression with the interaction between Cx group and assessment timepoint as the primary predictor variable, Cx result and timepoint as the corresponding main effects, and confounders found to correlate with at least one of the psychological measures as covariates. Analyses were performed separately for the PPC, STAI-S, and CES-D scales. When the interaction term was significant, post-hoc regression analysis was performed to determine the assessment timepoints at which the two Cx groups differed in their scores.
We then assessed whether there were within-group differences in psychological effect of the genetic test result over time. To address this question we performed within-groups repeated measures regression analysis with assessment timepoint as the primary predictor variable and relevant demographic variables and confounders as covariates. We also quantified the magnitude and direction of change experienced by each Cx group by computing the difference between the pre-test questionnaire scores and the 1-month or 6-month scale scores and performed paired t-tests to determine the statistical significance of these differences. To facilitate interpretation and comparison of effect sizes with other studies, we computed Cohen's d within each group as the difference between the means from pre-test to follow-up divided by the standard deviation (Hamilton et al. 2009). Following convention, d≤0.20 is considered a small effect size, d=0.5 medium, and d≥0.8 large (Cohen 1988). Analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC). Statistical significance was set at α=0.05.

Sample Demographics
Figure 1 provides details on sample sizes and questionnaire response rates at each stage. Participation rate was high with 241/263 (91.6 %) eligible individuals completing all protocol stages and receiving their genetic test results. Questionnaire response rates also were high, ranging from~98 to 99 % for the baseline and pre-test questionnaires to~76 % for the 1-month and 6-months post-test questionnaires. Importantly, among the 241 subjects who received their genetic test results, 100 % completed the baseline and pre-test questionnaires, 209 completed at least one of the two post-test questionnaires [159 (66 %) completed both (complete responders), 50 (20.7 %) completed only one (partial responders)], and 32 (13.3 %) did not complete any post-test questionnaires (nonresponders). Comparisons of complete, partial, and nonresponders did not reveal statistically significant differences on any of the demographic variables evaluated in this study, in the distribution of Cx test results, or as a function of enrollment site, genetic counselor, interpreter, or interval between sample and result disclosure (p's>0.05).
Responses from the 209 subjects who completed at least one post-test questionnaire (86.7 % of those who participated in the test result disclosure session) were analyzed, and demographic characteristics of this group are provided in Table 1. In this study sample, 39.2 % were classified as Cx positive (n=82) and 60.8 % as Cx negative/inconclusive (n=99, n=28, respectively). The Cx groups were compared on the demographic variables in Table 1 and found to be comparable on all but the family history variable (p's> 0.05). Not surprisingly, the two Cx groups differed in terms of the presence of deaf relatives where 75.6 % of those in the Cx positive group had at least one closely related deaf relative compared to 41.7 % of those in the Cx negative/inconclusive group (p < 0.0001). When focusing only on parents, 31.7 % of those in the Cx positive group had at least one deaf parent compared to 18.1 % of those in the Cx negative/inconclusive group (p=0.03).
Previous studies have found that age, gender, education level, ancestry, and family history are associated with psychological distress in the deaf population (Fellinger et al. 2012;Sheppard and Badger 2010) or in patient populations undergoing genetic testing (Gritz et al. 2005;Vernon et al. 1997). Thus, we assessed whether these demographic items as well as enrollment site, genetic counselor, interpreter, cultural affiliation, and interval between sample and result disclosure should be included as covariates in our main analyses. We did not evaluate preferred language as a separate covariate because it is highly correlated with cultural affiliation in this sample (p<0.0001). In this study sample, age, cultural affiliation, ancestry, enrollment site, and genetic counselor were not significantly associated with PPC, STAI-S, or CES-D scores at pre-test (p's>0.05). Consistent with previous research, we found that education level, gender, and family history should be included as covariates. Compared to participants with a college degree, those without a college degree had lower PPC scores (p= 0.0009), higher STAI-S score (p<0.0001) and higher CES-D scores (p<0.0001). Males had higher PPC scores than females (p=0.03), and participants with no closely related deaf relatives had higher CES-D scores than those with at least one deaf relative (p=0.003). Interpreter was also significantly associated with pre-test PPC scores (p=0.02), and interval between sample and results was significantly associated with 1-month anxiety scores (p=0.04); both were included as covariates.
Mean scores on the PPC, STAI-S, and CES-D scales at each timepoint are provided in Table 2. The mean sample scores on the STAI-S were within normal ranges, and 15.1 %, 18.8 %, and 20.6 % of participants met criteria for a clinical cut-off of ≥40 for the STAI-S at pre-test, 1 month-, and 6-months post test result, respectively. The mean sample scores on the CES-D also were within normal ranges, with 10.7 %, 13.2 %, and 11.8 % of participants exceeding the cut-off of ≥16 for depressive symptoms at each assessment timepoint, respectively.

Between-Group Analyses
First we assessed whether the psychological effect of genetic test results over time differed between the two Cx groups.

Perceived Personal Control
Repeated measures regression analysis demonstrated a statistically significant interaction between Cx group and assessment timepoint (F (2,345) =7.49, p=0.0007), even after controlling for the main effects of Cx group (F (1,190) =7.68, p=0.006) and timepoint (F (2,345) =2.76, p=0.06), and the covariates family history (p=0.69), education (individuals without a 4 year college degree had lower PPC scores, p= 0.02), sex (females had lower PPC scores, p=0.0009), interpreter (p=0.50), and interval between sample and result disclosure (p=0.89). Figure 2 plots participants' PPC least squares estimate scores before and after receiving their genetic test results. Post-hoc analyses demonstrated that pre-test PPC scores did not differ between the two Cx groups (p=0.51). However, PPC scores differed significantly between the two groups at 1-month (p=0.002) and 6months (p=0.003) post-test results, and in each case, the Cx positive group exhibited higher PPC than the Cx negative/inconclusive group.

Anxiety
Repeated measures regression analysis demonstrated a statistically significant interaction between Cx group and assessment timepoint (F (2,342) =6.4, p=0.002), even after controlling for the main effects of Cx group (F (1,190) = 1.86, p=0.17), timepoint (F (2,342) =0.16, p=0.85), family history (individuals without deaf relatives had higher STAI-S scores, p=0.004), education (individuals without a 4 year college degree had higher STAI-S scores, p= 0.0001), sex (p=0.12), interpreter (p=0.63), and interval between sample and result disclosure (p=0.15). Figure 3 plots participants' STAI-S least squares estimate scores before and after receiving their genetic test results. Posthoc analyses demonstrated that the two Cx groups did not differ significantly on their STAI-S scores at pre-test (p=0.80) or 1-month post-test results (p=0.24). However, STAI-S scores differed significantly between the two groups at 6-months post-test results (p=0.04), where the Cx positive group exhibited lower anxiety than the Cx negative/inconclusive group.

Depression
Repeated measures regression analysis did not demonstrate a statistically significant interaction between Cx group and assessment timepoint (F (2,326) =0.97, p=0.38) while controlling for the main effects of Cx group (F (1,180) =0.21, p= 0.65), timepoint (F (2,326) =0.10, p =0.90), family history (individuals without deaf relatives had higher CES-D scores, p=0.0003), education (individuals without a 4 year college degree had higher CES-D scores, p<0.0001), sex (p =0.80), interpreter (p=0.72), and interval between sample and result disclosure (p=0.95). Thus, there is no evidence that CES-D scores differed between the two Cx groups either before or after learning their genetic test results. Figure 4 plots participants' CES-D least squares estimate scores at the three assessment timepoints.

Within-Group Analyses
We next performed within-groups repeated measures regression analysis separately for each psychological scale and Cx group to determine which groups experienced change in their scores over the three assessment timepoints. To further assess the effect of genetic information on psychological well-being, we quantified the magnitude and direction of change experienced by each Cx group at 1-month and 6months post-test result relative to the pre-test genetic counseling session (Table 3). Compared to their pre-test scores, individuals who received a positive Cx result experienced a nearly significant increase in PPC at 1-month (d= 0.21, p=0.07), and a non-significant increase 6-months (d= 0.12, p=0.32) post test result. The initial increase and subsequent decline in PPC scores in this group over time explains the non-significant 'time' variable in the withingroups repeated measures analysis (Fig. 2, p=0.25). Individuals in this group also experienced a non-significant decrease in STAI-S at 1-month (d=−0.14, p=0.22), and significant decrease in STAI-S scores at 6-months (d = −0.26, p=0.03) post test results. The decrease in STAI-S scores over time is captured by nearly significant 'time' variable in the within-groups repeated measures analysis (Fig. 3, p=0.09). There was no evidence for substantive change in CES-D scores at 1-month (d=−0.07, p=0.59), 6months post-test results (d=−0.1, p=0.40), or over all three assessment timepoints (Fig. 4, p=0.74).
Compared to their pre-test scores, individuals who received a negative or inconclusive Cx result experienced a significant decrease in PPC scores 1-month (d=−0.25, p= 0.0009) and 6-months (d =−0.40, p < 0.0001) after they learned their test results. The continued decline in PPC scores in this group over time is captured by the withingroups repeated measures analysis (Fig. 2, p < 0.0001). Moreover, following receipt of their genetic test results, the Cx negative/inconclusive group experienced a nearly significant increase in STAI-S scores at 1-month (d=0.16, p=0.09) and a significant increase in STAI-S scores at 6months post test result (d=0.22, p=0.02). The increase in STAI-S scores over time is captured by the within-groups repeated measures analysis (Fig. 3, p=0.01). There was no evidence for substantive change in CES-D scores at 1month (d=0.10, p=0.34), 6-months post-test results (d= 0.08, p = 0.41), or over all three assessment timepoints (Fig. 4, p=0.

Major Findings
We report results from the first prospective, longitudinal study to evaluate the psychological impact of deaf genetic testing in deaf adults (n=209). Our data demonstrate that deaf genetic testing has a significant impact on deaf individuals' levels of perceived personal control and anxiety, but does not have an effect on levels of depression. Furthermore, the direction of the psychological effect of genetic information, that is, whether this information results in psychological well-being or psychological distress over time, depends on whether the test result is positive or negative/inconclusive. Our overall finding is that individuals who received a positive Cx result experienced increased psychological well-being, while those who received a negative Cx result experienced psychological distress. We hypothesized that deaf adults receiving a genetic test result that explains why they are deaf would experience psychological well-being compared to those who do not receive an explanation for why they are deaf. Our results support this hypothesis as there was evidence that deaf individuals who received a positive Cx result had higher perceived personal control scores and lower anxiety scores following test result disclosure than those with a negative/inconclusive result. Perceived personal control and anxiety have been found to be correlated with the extent to which genetic counseling clients' needs are fulfilled (Pieterse et al. 2005). Thus, we speculate that individuals who received a positive Cx result experienced psychological well-being because their result addressed their motivations for genetic testing. For deaf adults, testing motivations include enhancing self-knowledge, helping family members learn why they are deaf, learning if they can have deaf children, and strengthening the deaf community . Moreover, because perceived personal control is an important element of the concept labeled empowerment-a set of beliefs that enable an individual with a genetically-defined trait to feel that they have some degree   )-we speculate that deaf genetic testing empowers deaf adults who receive a positive result. Not only was there was no evidence that the Cx positive group experienced psychological distress following test result disclosure, but these results support our hypothesis that genetic selfknowledge from deaf genetic testing enhances the psychological well-being of deaf adults with a Cx positive result.
Our results also demonstrated that the group who received negative or inconclusive Cx genetic test results experienced psychological distress in terms of decreased perceived personal control and increased anxiety scores following test result disclosure. This result contrasts with studies of adults undergoing genetic testing for a medical condition, where in that context the group that experiences psychological distress typically is made up of the individuals who receive positive genetic test results, although these individuals have been found to experience some benefit as well (Williams et al. 2010). Overall, our results suggest that the psychological impact of genetic testing on adults for human traits and conditions may be influenced by the social construction of that human condition as medical or non-medical.
This study also found that levels of anxiety and perceived personal control did not necessarily return to baseline by 6months post test result, in contrast to many studies on the effects of genetic testing on adults for medical conditions in which the psychological impact of test results appears to be short-lived (Broadstock et al. 2000;Hamilton et al. 2009;Vansenne et al. 2009). Rather, for those with a positive Cx result, the largest decrease in anxiety scores occurred at 6months, and the overall trend was suggestive of an increasing trajectory of psychological well-being. Similarly, the largest increase in anxiety scores and decrease in perceived personal control scores occurred at 6-months for those with a negative/inconclusive result and suggested an increasing trajectory of psychological distress. These results suggest that genetic testing for non-medical traits may produce qualitatively different psychological outcomes from testing for medical conditions, and more specifically that individuals who receive a negative/inconclusive Cx test result may need additional follow up to address feelings of anxiety or decreased sense of control. In addition, longer term outcome studies of deaf genetic testing are needed to determine if and when these psychological effects of genetic information resolve, e.g., return to baseline or if the trajectory would continue to change.
Our study provides data that suggest that the effect sizes for the psychological effects of genetic self-knowledge for deafness are on the order of small to medium for anxiety (ranging from −0.26 to 0.22), and medium to large for perceived personal control (ranging from −0.40 to 0.21), depending on the Cx result group and assessment timepoint. These effect sizes are consistent with those observed in studies of the psychological impact of genetic counseling or genetic testing for medical conditions. As one example, studies of the psychological impact of reproductive or cancer genetic counseling found effect sizes for perceived personal control in the medium to large range (0.31 to 0.41) (Pieterse et al. 2005;Smets et al. 2006). As another example, a meta-analysis of disclosure of genetic test results for hereditary breast and ovarian cancer that included symptomatic and unaffected individuals found that the largest effect sizes for state anxiety were in the small to medium range (−0.33 to 0.22), depending on the result group and assessment timepoint (Hamilton et al. 2009). These results suggest that the magnitude of the psychological effect of genetic testing is generalizable across nonmedical traits and medical conditions. Deaf individuals are vulnerable to mental health issues, due to factors such as access to effective communication, childhood trauma including abuse, socioeconomic issues, type of deafness, and experiences with stigma and discrimination (Fellinger et al. 2012). Furthermore, there are data to suggest that rates of anxiety and depression are higher in deaf populations compared to hearing populations (Kvam et al. 2007), thereby heightening concerns about potential psychological impact of genetic information. Thus, another important finding from this study is that the Cx positive and Cx negative/inconclusive group mean scores on the CES-D scale for depression and the STAI-S scale for anxiety did not fall into ranges suggestive of clinically significant symptoms of depression or anxiety following test result disclosure. Moreover, the group mean CES-D and STAI-S scores observed in the current study are comparable to or lower than mean scores found in genetic testing studies of medical conditions (Gritz et al. 2005;Hamilton et al. 2009;van Roosmalen et al. 2004). Hence, the current study expands to the realm of non-medical traits the observation that psychological distress created by genetic information does not necessarily indicate clinically significant distress.
Although it is reassuring that genetic self-knowledge does not affect level of depression in deaf individuals, we note that 10.7 %-13.2 % of subjects had CES-D scores ≥16 clinical cutoff before and after genetic testing. Though a non-trivial percentage, it is lower than that observed in genetic testing studies of disease traits such as Lynch syndrome [e.g., 17.1 % (Esplen et al. 2007);25.8 % (Gritz et al. 2005)], and lower than that observed in a population based study of slightly older (presumably hearing) individuals with a variety of health conditions as well as individuals with no health conditions (29 %-39.7 % and 18.5 % had scores ≥4 clinical cutoff on a shortened version of the CES-D, respectively) (Wikman et al. 2011). Given the admittedly limited evidence that rates of depression are higher in deaf populations than hearing populations (Fellinger et al. 2012), one explanation for the relatively low rate in this sample is that the CES-D does not provide an adequate measure of depression in the deaf population and yielded an underestimate of depressive symptoms in the sample. This is unlikely, however, because depressive symptoms are the same in deaf and hearing populations (Sheppard and Badger 2010), a translation-back translation procedure was used to develop an ASL version of the CES-D, and we found high internal reliability of the CES-D items in our sample. Another possible explanation for our study's low percentage of clinical depression symptoms is that the sample is not representative of the deaf population. Because the ability to effectively communicate is associated with depression (Fellinger et al. 2012), and individuals unable to provide informed consent due to poor language abilities were not eligible to participate in our study, there may have been a participation bias toward mentally healthy individuals. Hence, generalizability of our findings should be done with caution and may not apply to deaf individuals who have not had access to effective communication. Regardless, these results suggest that genetic counselors should be prepared to address mental health issues in some deaf clients requesting deaf genetic testing as they would for any client.
Finally, the current study witnessed a high level of participation in genetic testing (91.6 % of eligible individuals). Not only does this participation rate provide further evidence regarding deaf individuals' interest in genetic selfknowledge, it also demonstrates a high level of consistency between interest and action that is not often observed in genetic testing studies of medical conditions. Across a variety of studies of medical conditions, the percentage of individuals declining genetic testing or learning their results has been as high as 40 % (Foster et al. 2004;Keogh et al. 2004;Peterson et al. 2002), and one of the primary reasons for declining testing is apprehension over the test result (Foster et al. 2004). Results from the current study suggest either that deaf genetic testing does not produce the type of apprehension that genetic testing for medical conditions produces, or that deaf individuals who felt apprehension self-selected against participation at the outset. Other factors that may have contributed to the high participation rate include the use of a culturally and linguistically appropriate research protocol, as well as providing genetic testing at no charge to our study participants. Studies of genetic testing for medical conditions have found mixed results on the effect of test cost on test uptake (Keogh et al. 2004;Peterson et al. 2002) so it is unclear the extent to which this factor played a role in our participation rate.

Study Limitations
This study has several limitations. First, although this study provides much-needed data on the psychological impact of deaf genetic testing on deaf adults, the extent to which our results generalize beyond Cx-related deafness is currently unknown. Second, while we successfully recruited a sample with diverse ancestral backgrounds, the majority of the sample was non-Hispanic Caucasian and thus generalizability to other ethnic/racial groups should be done with caution.
A third limitation is that this study pre-dates the widespread availability of NGS technologies and focused only on testing for Cx-related deafness since it is the most common genetic explanation for deafness. This testing strategy had several ramifications that may limit generalizability of our findings. First, audiology assessment was required for eligibility determination, which may have deterred participation. To address this issue we developed a culturally and linguistically sensitive protocol and offered participation in convenient, deaf-friendly locations. Second, because Cxrelated deafness is nonsyndromic, we limited eligibility to individuals with apparent nonsyndromic deafness. Thus, the extent to which our results apply to the psychological outcomes of genetic testing for syndromic forms of deafness, which often include significant medical conditions are not known and deserve investigation. Third, a negative or inconclusive Cx result does not rule out a genetic explanation for deafness. For that reason, individuals who received a negative or inconclusive result were informed that they could schedule a genetics clinic evaluation (outside of the research protocol) if they were interested in pursuing information to learn why they are deaf. We did not collect information on whether or not participants pursued this option. Thus we are not able to comment on if or how their level of psychological well-being changed during the course of the study as a result of finding a different genetic cause for their deafness. However, it is important to keep in mind that we followed participants for only a period of 6 months after they learned their genetic test results, and that may have been too short a period for individuals to actually go through all of the steps from obtaining a referral for a genetics evaluation to learning new genetic test results. Thus the results on the psychological outcomes of our Cx26 negative/inconclusive group are unlikely to have been impacted by their seeking additional genetic information outside of the research protocol. If additional testing did in fact occur, it would make the two groups (positives versus negatives/inconclusives) more similar in their well-being and in fact bias the results toward the null of no effect of test result. Hence our study design is conservative. However, it is possible that the psychological distress observed in the Cx negative/inconclusive group would not generalize to the case of NGS because a negative/inconclusive NGS result could provide a greater level of certainty.

Practice Implications
Genetics evaluation for deaf individuals has long been recommended for providing an etiologic diagnosis for deafness (American College of Medical Genetics 2002). Historically, however, there has been limited (or few) genetic counseling services tailored specifically for the deaf adult population outside of very unique settings such as Gallaudet University, a U.S. university for deaf students (Arnos et al. 1992(Arnos et al. , 1991b. Thus, although a number of studies have demonstrated that deaf individuals are interested in learning why they are deaf (Arnos et al. 1991b;Boudreault et al. 2010;Burton et al. 2006;Martinez et al. 2003;Withrow et al. 2009), it is still the case that few deaf individuals seek genetic counseling or genetic services (Enns et al. 2010;Middleton et al. 2010). This study underscores the importance of addressing the determinants of low utilization of genetic services by deaf adults because it provides empirical data that genetic information can enhance psychological well-being among individuals who receive a genetic explanation for why they are deaf.
We suggest a multi-pronged approach to address utilization of genetic services. First, this study supports broadening the standard objectives of genetics evaluations to include not only recurrence chance and medical management in instances where the genetic etiology could result in medical complications if unrecognized (e.g., cardiac manifestations in Jervell and Lange-Nielsen Syndrome), but also to enhance psychological well-being. Such a broadening of perspective among genetic counselors and trainees may have downstream beneficial outcomes. For example, because psychological distress is significantly associated with a variety of negative health outcomes (Janus et al. 2007), enhancing deaf individuals' psychological well-being could play a role in reducing health disparities in physical and mental health experienced by the deaf community (Barnett et al. 2011;Fellinger et al. 2012).
Under-utilization of genetic services may also stem in part from perceptions of cultural insensitivity or lack of information among hearing professionals regarding the Deaf community, and the common perception within the medical community that deaf people have a disability which needs intervention. In addition, the Deaf community has been subject to discrimination, non-acceptance, audism, 3 and even eugenics (Lane 1999;Schuchman 2004). As a result of this history, Deaf individuals may view the genetics community in particular with distrust. Thus a second practice implication is developing culturally and linguistically appropriate genetic counseling services to meet the needs and expectations of the deaf population (Arnos et al. 1991a;Baldwin et al. 2012;Boudreault et al. 2010;Enns et al. 2010). Since it is impractical for all deaf adults to receive genetic services from the very few specialized services that exist, genetic counselors can work within their own institutions to develop strategies for increasing awareness of and access to genetic services that are geared specifically to deaf individuals in their area.
A third practice implication is in the area of training. Training programs can encourage their students to develop ties with members of the local deaf community to raise awareness of genetic services and to learn effective ways to provide culturally and linguistically sensitive services to their deaf clients. Although a recent study of 158 practicing genetic counselors reported that 70 % received instruction in Deaf culture in their training program (Enns et al. 2010), there currently is no published information on the type or duration of instruction in Deaf culture or communication with deaf clients that students receive. Additional findings in Enns et al. suggest that improvements in instruction are warranted.
Finally, although members of the deaf community share commonalities, it is important to keep in mind that deaf individuals/families seen in genetic counseling are also unique and that their specific needs should always be addressed. As one example, an important role for genetic counselors is assessing their client's general well-being, including indications of distress which could be addressed during the session or may warrant referral to a mental health provider. Hence, a fourth implication for the practice and training of genetic counselors is the importance of addressing potential mental health concerns for their deaf clients. Although the level of psychological distress experienced by deaf individuals who did not receive a genetic explanation for why they are deaf did not rise to the level of clinical concern, some deaf clients may enter or leave a counseling session with significant mental health concerns. There currently are insufficient mental health services tailored for deaf individuals, particularly those who use a signed language (Fellinger et al. 2012). However, some mental health therapists specialize in serving deaf clients, and genetic counselors should become aware of possible resources in their area. In addition, training programs could better prepare their students for addressing mental health needs of their deaf clients by inviting these mental health specialists to give lectures on this topic.

Research Recommendations
This study has identified a number of avenues for future research. Importantly, studies that address the generalizability of these results are needed, including generalizability to minority groups within the deaf population, to individuals with syndromic forms of deafness, and to individuals who receive genetic information from NGS. Furthermore, studies on the psychological impact of genetic testing on hearing parents of a deaf child are warranted because the context of parents learning genetic information about their child differs from adults learning genetic information about themselves. It is also interesting to consider that in the clinical setting genetic counselors may convey expectations of a negative (or positive) genetic test result to their clients in part to manage their clients' expectations. Thus another area of research could focus on psychological outcomes of genetic test results as a function of a pre-test assessment of the likelihood of a particular test result.
Longer term follow-up studies of the impact of genetic testing for deafness in the deaf population are needed to evaluate if psychological sequelae return to baseline levels at some point, as well as to elucidate how to ameliorate feelings of anxiety that seem to be associated with a negative/inconclusive result. As we did not collect information on reasons for declining genetic testing or learning test results from the~8 % of participants who did not complete their study participation, additional research is needed to better understand the factors that affect deaf individuals' reasons for declining deaf genetic counseling and testing. Moreover, research to better define why deaf individuals experience increased psychological well-being from a positive genetic test result should be conducted. Finally, studies that evaluate the relationship between psychological wellbeing from genetic information and subsequent health outcomes are needed.

Conclusions
The data presented in this study shows that genetic testing provides positive psychological effects on deaf adults who learn why they are deaf. These results can be used to develop tailored counseling strategies and to provide anticipatory guidance for deaf clients. These results can also be used to promote genetic services to deaf individuals and increase utilization of genetic services by this population. Increasing deaf individuals' psychological well-being could offer downstream health benefits. Thus the results of this study underscore the importance of increasing deaf adults' access to culturally and linguistically appropriate genetic services.