Long-Term Adaptation Among Adolescent and Young Adult Children to Familial Cancer Risk

Data were drawn from a cross-sectional, mixed-methods study involving members of English-speaking households with mother-child dyads in which mothers (1) underwent clinical genetic counseling and testing that included BRCA 1 to 5 years previously, (2) were not currently receiving treatment of cancer and/or had not a recurrence, and (3) had disclosed their BRCA status to their AYA children (currently 12 to 24 year olds). If a pathogenic variant was present, then the children themselves must not have yet been genetically tested to be study eligible (ie, absence of cascade genetic testing).

Participants were recruited from 4 regional cancer centers in Scarborough, Maine; Boston, Massachusetts; Washington, District of Columbia; and Houston, Texas. Institutional review board approvals were obtained from each site. Sites discovered their own cohorts by accessing and reviewing institutional electronic medical records, genetic counseling records and pedigrees, and administrative and research databases to identify potentially eligible mothers.

Potentially eligible mothers were contacted by mail and telephone by local research staff explaining the purpose of the study, enumerating the household (including ages and genders of AYA offspring), screening for eligibility, and obtaining written informed consent/assent. One index child per household was chosen at random to participate (using the next birthday method¹⁶ ) to avoid selection bias. Enrolled participants then completed structured telephone interviews. All data were deidentified before analysis.

In total, N = 813 households with potentially eligible dyads were identified and approached: N = 252 (31%) could not be reached to confirm eligibility and N = 561 (69%) had ≥1 successful contact. In N = 7 households, the mother was deceased and N = 168 households were determined to be ineligible. The most common reasons for ineligibility were that the mother had not already discussed cancer genetic risk information with the AYA children (49%), there was no AYA in the target age range (29%), and/or the mother was undergoing active treatment of cancer or had experienced a recurrence (19%). This resulted in N = 386 eligible households, and from those N = 336 (87%) agreed to participate and N = 50 declined. The primary reason for declining the study was maternal and/or AYA reports of not being interested in the topic (75%) or being too busy to participate (23%). Of the N = 336 participating households, N = 301 mothers (90% of those consented) and N = 274 AYAs (82% of those consented/assented) successfully completed telephone interviews. There were no meaningful site differences observed at any stage of the recruitment cascade nor differences between the characteristics of households who consented or declined. The analytic sample was limited to N = 272 AYAs with complete data.

Tobacco and alcohol use and physical activity were measured by standard epidemiologic items from the Centers for Disease Control and Prevention's Youth Risk Behavior Surveillance System.¹⁷  Lifetime tobacco use was assessed by a yes/no item on ever having tried smoking, and lifetime alcohol use on the number of days participants had ever consumed an alcoholic beverage (0–100+). Physical activity was assessed by the number of days they had engaged in at least 60 minutes of physical activity in the past week.

Measures were not specific to BRCA-related cancers unless otherwise noted. Worries about maternal breast or ovarian cancer risk were assessed with the 3-item Children’s Cancer Worries Scale.^18,19  This scale covers the extent to which children think about and/or are disturbed by their mothers’ chances of developing cancer (1 = not at all or rarely, 4 = a lot; α = 0.75). To assess cancer risk perceptions, AYAs rated their beliefs (1 = no chance, 7 = certain to happen) about their own likelihood of developing cancer in adulthood compared with others their age.^20,21  Cancer knowledge was adapted from a previously validated measure²²  with 22 true/false items on 2 subscales: (1) general cancer prevention knowledge and (2) knowledge about the causes of cancer. Cancer prevention self-efficacy was derived using 7 items on a Likert scale (1 = not at all confident, 4 = very confident; α = 0.76).²³  Cancer screening intentions were assessed by face-valid items on their likelihood of engaging in age- and sex-appropriate cancer screenings (1 = not at all likely, 4 = very likely).²⁴  AYAs’ perceptions about the roles of genes and health behavior in cancer development, and the importance of learning about genes and health behavior, were assessed with 4 task-specific items on (1 = not at all/not at all important, 7 = completely/very important).²⁵ 

Cancer-specific distress was assessed using the well-validated Impact of Events Scale,²⁶  which measures subjective distress for a specific life event (ie, the possibility of their mothers developing cancer [again]). The scale contains 10 items (0 = not at all, 3 = often; α = 0.91). General life stress was assessed using the Perceived Stress Scale,²⁷  a 14-item measure of past 30-day stress appraisals reflecting how unpredictable, uncontrollable, and overloaded respondents find their lives (0 = never, 4 = often; α = 0.81). Depressive symptoms were assessed using the 20-item Center for Epidemiologic Studies Depression Scale (0 = none of the time, 3 = most of the time; α = 0.87).²⁸  Anxious symptoms were assessed using the 20-item state portion of the State-Trait Anxiety Inventory (1 = not at all, 4 = very much so; α = 0.90).²⁹ 

Age, sex, and race/ethnicity of AYAs and their mothers were self-reported, along with maternal educational attainment and marital status. Maternal clinical characteristics, including cancer status, BRCA pathogenic variant status, and family proband status were reported by mothers and confirmed by chart review.

Means and standard deviations, or frequency counts and percentages, were computed for the overall sample and compared between the 4 groups of interest by (1) child age (<18 years vs ≥18 years), (2) child sex (male vs female), 3) maternal cancer status (previously affected survivors, previously unaffected), and (4) maternal BRCA pathogenic variant status (BRCA+ vs BRCA-/uninformative). Two-tailed, independent samples t tests were used for continuous cancer cognition and quality of life outcomes, and χ² tests of independence were conducted with cancer risk behavior outcomes and other discrete covariates. Parsimonious multivariable regression analyses were then run for outcomes demonstrating significant associations with child and/or maternal factors.

Among the N = 272 AYAs, N = 47 (17.3%) had mothers with a known BRCA pathogenic variant, and N = 225 (82.7%) had mothers who either tested negative for a known familial BRCA pathogenic variant or received an uninformative negative BRCA result (Table 1). There was an approximately even split between adolescents (49.6%) and young adults (50.4%), with a mean age of 17.6 (SD = 3.5) years. AYAs were mostly female (67.6%) and non-Hispanic White (84.6%), with college-educated mothers (78.3%) in partnered relationships (84.9%) serving as family probands (85.2%) with a personal history of breast orovarian cancer (76.1%; cancer survivors). Mothers with BRCA+ pathogenic variants were less likely to be the family proband (65.2% vs 89.3%; P < .001). No other differences were observed.

Approximately one-quarter of the sample had used cigarettes in their lifetimes, more than one-third reported ever consuming at least 20 alcoholic drinks, and almost 90% did not meet public health guidelines for physical activity, which recommend at least 60 minutes of physical activity in this age group (Table 1).³⁰  Importantly, there were no observable differences in these outcomes when examined by maternal cancer history, maternal BRCA status, or child sex. Age differences in cigarette use were observed in the expected direction: young adults were more likely than adolescents to have tried smoking cigarettes, χ² (1, N = 272) = 70.1, P < .001. Significant age differences were also observed for alcohol use, such that young adults were also more likely to have consumed alcohol than adolescents, χ² (2, N = 272) = 146.7, P < .001.

Data on AYAs’ cancer cognitions and quality of life are given in Table 2 (overall and stratified by age and sex) and Table 3 (stratified by maternal cancer history and BRCA status).

In the overall sample, AYAs reported relatively low levels of worry about their mothers’ risk of breast or ovarian cancer (M = 4.1 of 12 maximum), although they perceived moderate risk of developing cancer themselves (M = 4.2 of 7 maximum). Knowledge about cancer was high (M = 19.3 of 24 maximum), with AYAs knowing about both cancer prevention (M = 8.0 of 9 maximum) and cancer’s causes (M = 11.3 of 15 maximum). AYAs felt confident in their ability to take steps to prevent cancer in general (M = 22.6 of 28 maximum) and endorsed strong intentions to undergo age- and sex-appropriate cancer screenings (M = 3.5 of 4 maximum). AYAs thought that genes were moderately important in determining cancer risk (M = 4.9 of 7 maximum) and similarly to learn about genetic risk factors for developing cancer (M = 5.4 of 7 maximum). They also reported health behavior as moderately important in determining cancer risk (M = 5.1 of 7 maximum), and of learning about health behavior (M = 5.8 of 7 maximum). As a group, AYAs exhibited low cancer-specific distress (M = 8.2 of 45 maximum), moderate general life stress (M = 23.6 of 56 maximum), low depressive symptoms (M = 11.5 of 60 maximum), and low anxiety (M = 31.8 of 80 maximum) that were within ranges considered normal for the population.

Regarding AYA age and sex effects, young adults perceived higher risks of developing cancer than did adolescents (P = .04). Similarly, females perceived their own cancer risk as higher than males did (P < .001). Young adults also reported being more knowledgeable about cancer than did adolescents (P = .01), especially its causes (P = .002). Conversely, adolescents were more confident in their abilities to prevent cancer than was reported by young adults (P = .02), whereas young adults had greater intentions to undergo cancer screening than did adolescents (P = .002). Finally, adolescents reported more cancer-specific distress than did young adults (P = .002), and female AYAs reported more general stress than AYA males(P < .001).

There were also significant associations with maternal status variables: AYAs with BRCA+ mothers reported more worry about their mothers’ chances of developing cancer than did those whose mothers had received negative or uninformative genetic test results (P = .03). AYAs whose mothers were cancer survivors also perceived greater personal risks of developing cancer themselves than those whose mothers had not had cancer (P = .001). Finally, AYAs of cancer-surviving mothers were more knowledgeable about cancer than were children whose mothers had not had cancer (P = .02); this finding was especially salient for understanding cancer’s causes (P = .001). Relatedly, AYA children of BRCA+ mothers believed more strongly that genes determine cancer risk (P = .03) and that it is important to learn about genes (P = .001) relative to AYA children of mothers with negative or uninformative genetic test results. No additional differences were observed as significant.

As shown in Table 4, a multivariable model including child age, child gender, and maternal cancer status indicated an overall significant association with AYAs’ perceived risk of developing cancer (P < .001). Specifically, females and AYA children of cancer survivors perceived themselves to be at higher personal risk of cancer than did males and children of mothers who were not previously affected.

A second model including child age and maternal cancer status indicated an overall significant association with total cancer knowledge (P = .005), such that young adults and children of cancer survivors showed greater knowledge about cancer than did adolescents and children of mothers not previously affected. A similar finding emerged when examining AYAs’ understanding of cancer’s causes (P ≤ .001).

Genetic testing for pathogenic variants in BRCA may influence family members’ psychosocial adaptation, as measured by their cancer risk behavior, beliefs about cancer, and quality of life, especially among AYAs growing up in households where testing is pursued and with cancer-surviving parents.³  In this sample of AYAs whose mothers were tested for BRCA and who were told about their mothers’ BRCA status, including BRCA+ in 17.3% of cases, outcome data collected 1 to 5 years after testing failed to indicate that such disclosure was significantly associated with either health-promoting or health-compromising cancer risk behavior in these children; in addition to lack of statistical significance, differences between groups were small. That is, children who had learned of their family’s high genetic risk for breast or ovarian cancer were no more or less likely to act in ways that could prevent cancer (or protect their overall health) compared with those without a BRCA pathogenic variant. This finding mirrors that seen elsewhere, whereby breast or ovarian cancer family history information alone is insufficient to motivate healthy behavior change,³¹  even without the added influence of genetic risk information.

In contrast, several cofactors were associated with these children's long-term psychosocial adaptation. First, AYAs whose mothers survived cancer, along with female AYAs, perceived themselves to be among those at the greatest risk of developing cancer; effect sizes were medium. Given women's breast cancer risks compared with men's³²  and their odds of BRCA-related cancers in the presence of deleterious variants,¹³  such concerns seem warranted. However, greater concern over personal cancer risk in the absence of known mutations, even among children of survivors, may be inaccurate. Thus, it is important to provide AYAs in such families with clear information about the role of genes in health, both to reduce unnecessary worry³³  and to provide tailored medical management recommendations.¹³  This patient education strategy can also be decoupled from an immediate offer of genetic testing and may be viewed as more preparatory to future cascade testing upon reaching adulthood.¹ 

Second, young adults and AYAs whose mothers had cancer exhibited the greatest understanding about general cancer causes; effect sizes were medium and consistent with their maturity levels and previous experience with maternal cancer.³⁴  Interestingly, adolescents reported higher confidence in their ability to prevent cancer, whereas young adults reported stronger intentions to undergo recommended cancer screenings; effect sizes were small to medium. That adolescents felt more confident in their ability to prevent cancer relative to young adults (who reported higher intentions to engage in cancer screening and knew more about cancer’s causes) may reflect cognitive maturation in thinking about cancer risk.³⁴  Although these constructs were not specific to BRCA-related cancers, it could ultimately translate into increased awareness of BRCA, proactive cancer prevention behavior in the long-term, and the pursuit of cascade genetic testing in families of known mutation carriers.

Finally, and with respect to children of BRCA+ mothers, they were among the most concerned about the possibility of their mothers developing breast or ovarian cancer with a medium effect size, which is consistent with epidemiologic and clinical data.³⁵  Likewise, these same children believed most strongly in genes determining cancer risk and the importance of learning about such risk, with female adolescents exhibiting particularly high levels of concern, and commensurate with their stage of life.³⁶  All effect sizes were medium. Taken together, these findings also underscore that AYAs generally adapt well to learning of their mothers’ BRCA genetic risk, but subgroup differences indicate a need for more individualized educational interventions.

The predominantly non-Hispanic White sample limits generalizability to members of underrepresented racial or ethnic groups. Future studies could incorporate more diverse samples and oversample mothers with pathogenic BRCA variants. Additional research could also include other comparison groups (eg, AYAs of men/fathers tested for BRCA, healthy controls at average cancer risk). Furthermore, all mothers underwent comprehensive pre- and posttest genetic counseling. Thus, these findings may not extend to those receiving genetic testing without such services. Finally, other associations may exist among cancer risk behavior, beliefs about cancer, and quality of life outcomes to further inform this area of research.

Younger at-risk relatives in familial cancer kindreds are an understudied group. In this novel investigation, we found that although multiple child and maternal characteristics were mildly associated with children's cancer cognitions and quality of life, cancer risk behavior and psychological distress-related outcomes were highly resilient to BRCA+ notification. These data indicate that it is safe and appropriate for mothers to share their BRCA genetic test results with AYA children. These data also suggest that early disclosure of maternal BRCA+ results to AYAs about hereditary cancer and familial risk could set the stage for more open and ongoing communication about genetic testing. Additionally, these conversations could inform AYAs’ interest in and use of such testing (ie, cascade genetic testing) and their engagement in other risk-reducing measures. Presently, rates of BRCA cascade testing are relatively low, despite the issuance of guidelines recommending such testing.³⁷  Such needs could be met through family-based genetic education/counseling strategies, with interventions directed toward mothers who are hesitant and/or unsure of how to communicate about genetic testing and familial risk with their offsprin g.^9,38  Other interventions could target AYAs themselves, encouraging discussions with their high-risk parent. These efforts to increase the uptake of cancer risk assessments, cascade testing, and age-appropriate cancer screening practices among those who are eligible will require the development and evaluation of new tools and resources delivered as adjuncts to cancer genetic counseling in light of workforce availability,³⁹  and the nascent state of genetic education across the lifespan. Such interventions may improve AYAs long-term psychological well-being and encourage proactive behaviors that reduce morbidity and mortality from preventable cancers.

We are grateful to Patrick McAlary, MSN, and the clinicians, research staff, and participant families involved with the project.

AYA

adolescent and young adult