A Systematic Review on the Impact of Genetic Testing for Familial Melanoma II: Psychosocial Outcomes and Attitudes

It is estimated that 1% of primary melanoma cases occur within families with a very strong history of melanoma [1]. To date, approximately 10 high-penetrance genes conferring a significantly increased risk of melanoma have been identified; however, they explain only around 22% of the familial melanoma cases [2]. By far the most common causative gene for familial melanoma is CDKN2A, a gene implicated in cell cycle regulation [3]. Pathogenic variants in CDKN2A incur a 52% lifetime risk of melanoma and an increased risk of pancreatic cancer [4]. The penetrance of CDKN2A for pancreatic cancer risk is not as well established, with European and North American studies estimating an 18% lifetime risk [5, 6], while Australian cohort studies have found no evidence of increased risk [6]. Possible explanations for the discrepancy may be due to lower smoking rates in Australia and/or differing CDKN2A mutations in Australian populations [6].

Although genetic testing using a panel of known susceptibility genes is commercially available [7], clinical implementation has been restrained as utility in improving prevention is unclear [8]. Additionally, there are concerns about causing psychological distress [9], and a number of complex ethical and public health issues are debated [10]. Previous systematic reviews on predictive genetic testing for various diseases have reported limited evidence of adverse psychosocial impact [11-13], and a modest impact on increasing preventative behaviour [14-16]. There have been no previous systematic reviews focusing solely on outcomes relevant to genetic testing for familial melanoma. This is the second in a 2-part systematic review series, where the first reported impact on preventative behaviour (see part I, the other paper of this series, DOI: 10.1159/000513919), while this review focuses specifically on psychosocial outcomes. This review explores a range of complex psychosocial outcomes beyond psychological distress, including attitudes towards genetic testing and testing of minors, as well as factors influencing uptake.

This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [17] and was preregistered with PROSPERO (International Prospective Register of Systematic Reviews), and it can be accessed here: https://www.crd.york.ac.uk/prospero/display_record.asp?ID = CRD42020186881.

Inclusion criteria comprised any studies where (i) participants were perceived to be at high-risk of developing melanoma, (ii) genetic testing (or hypothetical scenarios) was offered, and (iii) psychosocial outcomes were available. Only original research published in English was assessed for eligibility. The following databases were searched: Pubmed, Embase, CINAHL, PsycINFO and the Cochrane Library using terms related to genetic testing, psychosocial impact and melanoma risk; search strings are provided in online supplementary Table S1 (for all online suppl. material, see www.karger.com/doi/10.1159/000513576). A date range of January 1995 (discovery of CDKN2A as a melanoma gene) until June 2020 was applied. Screening was performed by reviewing titles/abstracts. Eligibility was assessed by reviewing full text. A subsequent backwards and forwards citation search was completed. The QualSyst assessment tool [18] was used to review the quality of studies included in the review. Screening, eligibility, quality and risk assessment was performed by author C.P., and a second author (T.Y.) screened 10% of the records and assessed quality in 25%. Data were extracted using a standardised report form, collecting: citation, study design, study population, outcomes and follow-up time points. Narrative synthesis was used to describe results, as the differences in methods and outcomes measured prevented result pooling [19].

A glossary of commonly used terms throughout this review is provided in Table 1.

In total, 2,352 articles were identified in the database and citation search, and 1,697 remained after removing duplicates. A further 1,590 articles were removed after screening titles/abstracts, leaving 49 articles assessed for eligibility by full-text review (Fig. 1). Thirty articles were excluded for reasons described in Figure 1. The final review included 21 articles describing outcomes from 12 unique studies, which included a total of 1,137 study participants. Two studies used hypothetical scenarios to explore psychosocial responses, and the remaining studies used CDKN2A genetic testing to provide genetic risk assessment. No other familial melanoma risk genes were assessed. The follow-up period ranged from cohort studies to 2 years after test result disclosure. Table 2 describes characteristics of studies, including outcomes assessed. Interrater reliability scored highly between authors for screening and quality assessment, producing a kappa score of 0.88 for screening and 0.91 for quality assessment. Quality assessment found all studies rated moderate to high in quality (online suppl. Table S2).

The most commonly reported psychological outcomes assessed included: generalised anxiety and depression [20-26]; melanoma-specific distress [20-28]; cancer (melanoma and pancreatic) worry/concern [24, 26-28]; and decision (for genetic testing) regret [20, 24-26]. Most studies utilised validated, widely accepted surveys such as the Hospital Anxiety and Depression Scale which includes two 7-item subscales to measure both anxiety and depression, and is widely used in research [29]; the Impact of Events Scale, which comprises 15 items to assess intrusive and avoidant cognitions and behaviours in respondents and has been used to measure disease-specific distress [30]; and the Multidimensional Impact of Cancer Risk Assessment scale, a 25-item survey that is commonly used to assess distress and uncertainty in people undergoing genetic testing [31]. Other validated questionnaires were used to capture cancer worry [32-34], decision regret [35], mental health [36], anxiety and depression in adolescents [37, 38], and lastly open-ended and semi-structured interviews were used in 3 studies [25, 26, 28].

Generalised distress was reported in 9 publications from 7 separate studies [20-26, 39, 40]. All studies used validated measures to quantitatively assess outcomes related to generalised anxiety and depression [20-26, 39, 40]. Within studies, results were compared between a variety of subgroup populations, such as mutation carriers, non-carriers, test acceptors, test decliners, those previously diagnosed with melanoma (affected) and those without (unaffected). Across all studies, mean distress scores were low and below clinically relevant thresholds. No studies reported an increase from baseline measures in generalised anxiety or depression following test disclosure. Four studies reported no significant change in anxiety in carriers or non-carriers over time [21, 25, 26, 40]. One study (n = 119) reported no association between generalised distress scores at baseline (before testing) and personal melanoma history [23]. At 12 months after testing, this study measured a significant decrease in anxiety for both acceptors and decliners of testing [20]. Another study (n = 60), which compared affected carriers, unaffected carriers and non-carriers, found a significant decrease in anxiety across all groups over time [24]. One cross-sectional study which surveyed participants prior to testing reported low levels of anxiety and depression for both acceptors and decliners of testing; however, anxiety was slightly higher in decliners (mean anxiety score 2.5; SD 2.8) compared to acceptors (mean anxiety score 1.7; SD 1.9) [22, 39]. Two studies reported significant decreases in depression over time, which was not dependent on carrier status [20, 26]. One study found carriers reported “sadness” levels significantly greater than non-carriers; however, this difference was existing at baseline and remained unchanged at 6 months of follow-up [25].

Disease-specific distress or worry for melanoma was reported in 8 studies [20-28] and for pancreatic cancer in 3 studies [22, 24, 27].

Overall, melanoma-specific distress was relatively low and stable over time, with no significant changes from before to after genetic testing [20, 21, 24]. Group comparisons in one study found no significant differences in baseline distress between acceptors and decliners of testing [22]. Another study reported no differences in melanoma-specific distress between carriers, test acceptors or test decliners at either baseline or 12 months of follow-up [20]. Three studies reported significantly higher distress in specific subgroups including: affected participants [23], affected carriers [24] and in carriers (both affected and unaffected) [25], with no significant change in distress from pre- to post-testing surveys.

Four studies reported on melanoma-specific worry [24, 26-28]. A study evaluating responses to hypothetical scenarios found that a personal or family history of melanoma was associated with calmly accepting increased-risk results [28]. In comparison, those without any melanoma history anticipated feeling increasingly worried, scared and/or shocked [28]. Two later studies, which provided genetic testing, also considered personal history and reported that affected carriers had significantly greater worry than unaffected carriers and non-carriers, with no significant change over time [24, 27]. A study, which tested 18 minors, reported a significant decrease in melanoma-related worry over time (1 year), for both carriers and non-carriers, with no significant difference between groups [26].

Three studies assessed impact of genetic testing on pancreatic cancer-specific worry. One study reported that approximately half of the participants were both surprised and anxious to learn about pancreatic cancer risk [22]. Another study found that carriers had significantly greater pancreatic cancer worry at both baseline (before testing) and 6 months after disclosure [27]. Similarly, another study found both carriers and those with a pancreatic cancer family history recorded significantly greater worry [24]. Furthermore, mean pancreatic cancer worry scores for all participants increased slightly at 1 month, then decreased significantly from 6 months to 2 years after test disclosure [24].

Ten publications, from 7 unique studies, reported outcomes on participants’ perceived risk of melanoma [20, 23, 27, 40-44] and perceived control of subsequent melanomas [45, 46]. Four longitudinal studies found no sustained significant changes in perceived risk from pre- and post-testing, with follow-up ranging from 4 months to 2 years [27, 40, 42, 43]. Carriers and those with a personal history of melanoma reported higher perceived risk compared to non-carriers and unaffected individuals [23, 27, 41, 43]. A study comparing acceptors and decliners of testing found no significant baseline differences in reported risk perception [23]. When comparing objective risk to perceived risk, one study revealed that on average, carriers underestimated risk, while non-carriers overestimated their risk [40]. Likewise, Aspinwall et al. [43] found that unaffected carriers consistently underestimated their risk, non-carriers overestimated, while affected carriers were more accurate. Furthermore, Aspinwall et al. [43] reported that while non-carriers recorded decreased perceived risk immediately after test disclosure, long-term follow-up from 1 month to 2 years found levels returning to previous levels of overestimated risk perceptions. Lastly, one study reported that understanding and acceptance of risk information was higher in those who received genetic testing, compared to those who received equivalent counselling without genetic testing [44].

Two studies reported on perceived control of developing subsequent primary melanoma(s) [45, 46]. In one study, 45% of participants reported increased perceived control 2 years after genetic testing, while 38% reported no change, and 17% reported decreased perception of control [45]. In the second study, individuals who reported greater personal responsibility over their health were also found to have higher perceived control over melanoma risk [46], while religiosity, including belief that cancer was “God’s will,” was not associated with perceived control [46].

Minimal decisional regret over uptake of genetic testing was reported in 4 studies [20, 24-26]. Using the Decision Regret Scale, Kasparian et al. [20] reported minimal regret among carriers and non-carriers. Out of a possible score of 100, with higher scores representing great regret, the mean score for carriers was 15, compared to 6 for non-carriers [20]. A study offering testing for minors surveyed participants’ mothers with open-ended questions and found no self-reported regret in consenting to testing for their children [26].

An early Australian study, using hypothetical testing scenarios, reported 68% of participants would seek genetic testing should it become available [47]. Intentions to take up testing was greater in participants with higher empiric risk. For example, 90% of participants with both family and personal history of melanoma would pursue testing if available, compared to 30% of participants with no melanoma history [47]. Additionally, the study used hypothetical testing scenarios of different CDKN2A penetrance models (100, 90, 30%) and found little difference in uptake between 100 and 90%; however, uptake declined significantly in the 30% penetrance scenario [47].

Four studies reported the percentage of contacted participants who proceeded with genetic testing, which included 21 [20], 66 [48], 70 [21] and 77% [22]. Kasparian et al. [20] reported that intentions to have genetic testing at baseline were 67%. However, at the conclusion of the study only 21% of participants had undergone genetic testing [20]. The low uptake of testing was attributed to study design factors including participants required to approach a clinical genetic service themselves to request testing and needing to meet the clinics criteria for publicly funded genetic testing [20].

Two studies reported participants’ reasons for declining testing, which included: “happy with life how it is,” “not relevant to me” and “can’t tell me if I’ll definitely get melanoma” [20, 39]. Factors positively associated with uptake included: female gender, family history, older age, having children, perceived risk, higher empiric pretest risk, lower pretest anxiety and the gene penetrance [20, 22, 39, 47, 48].

Three studies reported outcomes on participants’ attitudes towards testing minors [26, 47, 49]. When asked about attitudes towards testing minors for familial melanoma, approximately half of the participants (53%) were supportive generally, a third stipulated only under medical recommendation, and 10% opposed melanoma genetic testing in minors [47]. A later, quantitative study of adults who received testing for CDKN2A found 87% expressed support for testing availability for children, with greater support in individuals diagnosed with melanoma at a younger age [49]. Participants’ perceived benefits for testing children included raising risk awareness and improving primary and secondary prevention [26, 49]. The perceived minimum testing age ranged from 0 to 25 years, with 12 being the most common answer [49]. A study which provided genetic testing for 18 minors (mean age of 12.4 years), interviewed parents of participants 1 month and 1 year after testing, and found all mothers rated the experience beneficial, due to promoting risk awareness (90%) and increasing sun protection behaviour (82%) [26].

Seven studies provided participant-reported outcomes on potential benefits and limitations of genetic testing for familial melanoma [20, 21, 24, 40, 41, 47, 50], 2 of which used only hypothetical scenarios of testing [41, 47]. Overall, benefits were far more frequently reported than limitations, for example, in one study using open-ended questions, nearly 95% of participants provided at least one benefit at each assessment, compared to 16% of all study participants who provided a perceived limitation of testing [24]. Across the studies, the most commonly cited benefits reported included: increased awareness/certainty of own risk, learn about children’s risk, motivation for sun-protective behaviour and screening, assisting research and relief for non-carriers. The most commonly identified disadvantages included: negative psychosocial reaction, concern about effect on family, possible test inaccuracy, reduced protective behaviour (for non-carriers) and insurance discrimination. Furthermore, 3 studies reported results separately for carriers and non-carriers [24, 40, 50]. Carriers most commonly rated motivational benefits (e.g., for sun protection), while non-carriers rated emotional relief regarding personal and children’s risk as most important. Both carriers and non-carriers frequently reported informational benefits. Predictably, increased worry for personal risk and children’s risk were often reported by carriers as disadvantages of testing.

Based on our systematic review there is limited evidence that genetic testing for familial melanoma causes significant psychological distress. The evidence across the studies, which are based on well-validated multi-item inventories, indicates that generalised distress does not appear to be impacted by testing outcomes, carrier status or personal history. Although measures of psychological distress across the studies were, on average, relatively low and below clinically relevant thresholds, significant differences were detected between subgroups. Presence of higher pretest anxiety in some studies suggests that the testing process itself may cause a short-term increase in generalised distress, which resolves quickly thereafter. In comparison, disease-specific distress seems associated with carrier status and/or personal history. Quantitative measures show that carriers and those with personal history, on average, report significantly higher levels of disease-specific distress, with little change over time. Elevated disease-specific distress in these groups is present at baseline, prior to testing, and therefore not the result of testing outcomes and possibly due to personal or family history. Incidence of higher disease-specific distress in carriers and those previously diagnosed with melanoma is expected and is potentially important in motivating risk-reducing behaviour. Common psychological constructs predict that while excessive levels of distress may have a detrimental impact on health behaviour uptake, moderate levels of distress promote risk reduction [51].

Somewhat surprisingly, there was no evidence that genetic risk assessment had any impact on participants’ perceived risk of subsequent melanomas. While it would appear illogical that personalised genetic risk feedback would not influence perceived risk, it is however consistent with other multifactorial diseases [16]. One reasoning for unchanging risk perception would be if pre-existing beliefs were accurate. On closer examination, it appears while carriers with a previous melanoma diagnosis held fairly accurate risk perceptions, unaffected carriers more often underestimated their risk. When applied to psychological models on health behaviour such as the Protective Motivation Theory [52], this lack of impact on perceived risk may translate to a lack of adapting risk-reducing behaviours. It is important to consider the possible differences in how genetic risk results are disclosed, and how much risk education is provided. This is often not described in detail in studies, and therefore, conclusions on efficacy cannot be drawn. However, future investigation into genetic risk counselling and education methods may be beneficial [53].

When reviewing participants’ attitudes and beliefs regarding wider availability of genetic testing for familial melanoma, feedback was largely positive, and participants were also in favour of making testing available to minors. The most frequently nominated benefits included greater awareness of own risk and children’s potential risk, and motivation for protective behaviour and screening. For those testing negative for a familial mutation, relief rated highly. Concerns regarding impact on relatives and insurance discrimination were also voiced. Both topics are complex issues and require meaningful discussion in pretest counselling, allowing an informed decision to proceed with testing [53]. Positively, several countries have introduced policies or legislation to protect against the use of genetic test information in medical and/or life insurance underwriting [54].

Measured uptake of genetic testing in research provides some indication of the potential demand from the familial melanoma cohort, if testing became more widely available. It is encouraging to see evidence that intentions to pursue genetic testing appear mostly aligned with empiric risk [47]. This may ease concerns that testing services may be overwhelmed by individuals deemed ineligible for testing. One study reported a much lower test uptake (21%) [20] than others. Initially, this appears to be an outlier; however, it may actually represent a more realistic model of uptake. In this study, participants were required to arrange testing themselves thereby reflecting real-life practice, compared to other study designs which followed more streamlined protocols. This is consistent with uptake of BRCA1 and BRCA2 testing, which was higher when offered to a highly invested research population (78%) [55], as compared to clinical cohorts of 30% in men and 62% in women [56]. The opportunity to learn more about children’s potential risk also rated highly as a motivator for uptake of genetic testing [22]. Reasons for declining testing included those relating to convenience, such as time and travel, and the sentiment “I’m happy with life how it is” [20]. A study from the Netherlands reported an association with generalised anxiety and disease-specific worry with declining genetic testing [22, 39]; however when measured in an Australian cohort, no difference in distress was reported between test acceptors and decliners [20].

There are limitations to consider for this systematic review. The heterogeneity of study methodology used and the inclusion of hypothetical scenarios to draw participant responses places difficulty in synthesising outcomes and limits subsequent conclusions. However, it was assuring to note that outcomes provided in the hypothetical studies were closely aligned with clinical study outcomes reviewed. Further limitations to generalisability include small, homogenous (white, well-educated) self-selected study populations who were highly motivated to participate in the study, and in some studies included several participants from a small number of families. It would be interesting to see whether studies offering genetic testing to a large number of unrelated individuals from differing socio-economic backgrounds yield similar psychosocial outcomes. The overall low sample sizes and lack of gold standard randomised controlled trials place restrictions on the strength of evidence drawn from this review [57]; however, the low incidence of mutation carriers and ethical issues of controlled randomisation are limiting. Therefore, as this is an emerging field with limited research conducted, the importance of capturing all available outcomes was prioritised, and subsequent conclusions are made with caution. While 5 studies followed participants for 1–2 years, a third of the studies in this review were cross-sectional, and another 3 included a follow-up period of 6 months or less, which limits conclusions regarding the long-term psychosocial consequences of genetic testing in familial melanoma. Lastly, grey literature and conference abstracts were not included in the search strategy, and therefore a publication bias risk exists.

This systematic review has found limited evidence of adverse psychological impacts of genetic testing or negative effects on risk perception. Study participants’ attitudes towards and perceptions of genetic testing were positive, with perceived benefits far outweighing perceived disadvantages. If larger studies, with more heterogeneous populations and long-term follow-ups produced comparably positive findings, this would be invaluable in informing future clinical practice.

Evidence suggests that genetic testing for familial melanoma does not result in adverse psychosocial outcomes.

This is a literature review which did not involve human subject recruitment. Thus, ethical approval was not required.

H.P.S. is a shareholder of MoleMap NZ Limited and e-derm consult GmbH, and undertakes regular teledermatological reporting for both companies. H.P.S. is a medical consultant for Canfield Scientific Inc., Revenio Research Oy and also a medical advisor for First Derm.

C.A.P. is funded by an Australian Government Research Training Programme Scholarship. A.M.M.-L. is funded via an NHMRC ECF APP1158111. H.P.S. holds an NHMRC MRFF Next Generation Clinical Researchers Program Practitioner Fellowship (APP1137127).

C.A.P. performed database search, screening, eligibility and quality assessment. T.Y. independently screened 10% of articles and assessed quality of 25% of articles. C.A.P. prepared the first draft of the manuscript. All co-authors contributed to writing, provided intellectual input and approved the final version of the paper.