Psychosocial and Clinical Factors Associated with Family Communication of Cancer Genetic Test Results Among Women Diagnosed with Breast Cancer at a Young Age

Abstract

Genetic test results have medical implications beyond the patient that extend to biological family members. We examined psychosocial and clinical factors associated with communication of genetic test results within families. Women (N=1080) diagnosed with breast cancer at age 40 or younger completed an online survey; 920 women that reported prior cancer genetic testing were included in analysis. We examined the proportion of immediate family members to whom they communicated genetic test results, and built multivariable regression models to examine clinical and psychosocial variables associated with the proportion score. Participants were most likely to communicate test results to their mother (83%) and least likely to their son (45%). Participants who carried a BRCA mutation (OR=1.34; 95% CI = 1.06, 1.70), had higher interest in genomic information (OR=1.55; 95% CI = 1.26, 1.91) and lower genetic worry (OR=0.91; 95% CI = 0.86, 0.96) communicated genetic test results to a greater proportion of their immediate family members. Participants with a BRCA1/2 mutation shared their genetic test results with more male family members (OR=1.72; 95% CI = 1.02, 2.89). Our findings suggest that patients with high worry about genetic risks, low interest in genomic information, or receive a negative genetic test result will likely need additional support to encourage family communication.

Introduction

Genetic test results have medical implications beyond the patient that extend to the biological family. For family members, receiving information about a mutation allows for personal decisions regarding additional medical care and seeking genetic counseling and testing, as well as receiving social support from the family (Taber, 2014; Wiseman et al, 2010). Although healthcare providers can help facilitate communication of genetic test results, patients ultimately decide whether or not to communicate this essential information to family members (Cheung et al., 2010).

In their review, Gaff and colleagues (2007) found that up to 64% of family members who are informed of a genetic risk from a patient seek genetic testing. Prior studies indicate that first-degree family members are most likely to receive genetic risk information. Findings from McGivern and colleagues (2004) show information about a BRCA 1/2 gene mutation was discussed with 88% of first-degree family members as compared to 41–54% of extended or distant relatives. Finlay and colleagues (2008) similarly found differences in rates of disclosure between first- and second-degree family members. Koehly and colleagues (2003) reported that both first-degree relatives and spouses are important when discussing genetic counseling and testing. Research has shown that more female family members receive this information from the patient than males, especially for more distant relatives (McGivern et al., 2004). Cheung and colleagues (2010) found that patients most often discuss genetic information with family members within the same generation, while Forrest and colleagues (2003) reported that patients felt responsible for telling younger generations about genetic risk. However, prior research has also demonstrated that some family members may never receive information about genetic test results due to reluctance from the patient or lack of contact. Family rules about communication can inform choices such as who is responsible for informing specific family members, and decisions about disclosure (Gaff et al., 2007; Forrest et al., 2003; Hamilton et al., 2005).

While prior research has examined whether and to whom genetic test results are communicated within families, much less is known about the clinical and psychosocial factors that affect the communication of genetic test results to family members. Examination of this latter question is critical to identifying potentially modifiable factors that may be effective intervention targets for improving family communication. We addressed this research gap among women diagnosed with breast cancer at a young age. One prior study examined clinical predictors of communication among young breast cancer patients (Cheung et al., 2010). These women are more likely to carry germline gene mutations than women diagnosed at an older age (de Sanjose et al., 2003; Young et al., 2009), making family communication of genetic test results particularly salient. In addition, these patients may have concerns that are unique when compared to other breast cancer patients, such as the younger age of their children. For women diagnosed with breast cancer at a young age, Guth and colleagues (2015) found the majority of their children to be primary school age (35%) or baby/toddler age (21%) and there was a desire for additional children among the women. Patients diagnosed with breast cancer at a young age are more likely to have living parents and younger siblings and they, and their family members, have more years at risk for other primary cancers when compared to older patients with breast cancer. Additional focused research is therefore needed to understand the communication patterns of these families.

Previous work suggests possible factors that might affect family communication of genetic test results in this patient population. The Common Sense Model, which is based on self-regulation theory (Leventhal et al., 1997), was developed to explain the response and management of health threats that motivate behavior change. This model has been applied to genetic risk information (Marteau, & Weinman, 2006), and suggests that beliefs about the causation of breast cancer might influence family communication (Marteau, & Weinman, 2006). When cancer survivors develop causal beliefs related to health behaviors, this can prompt communication of healthy lifestyle changes with first-degree relatives (Rabin, & Pinto, 2006).

In addition to causal beliefs, prior research suggests other genetics-related cognitions that might be important psychosocial predictors of family communication about genetic test results (Ashida et al., 2011; Diefenbach, & Leventhal, 1996). Differences in genetic knowledge might affect inter-generational transfer of genetic information. Ashida and colleagues (2011) have found that genetic knowledge is lower among older individuals. McBride and colleagues (2009) discuss interest in genomic information which may drive communication within a family about shared genetic risk. Genetic self-efficacy, or an individual’s perceived ability to understand and explain genetics (Parrott et al., 2004), has been associated with perceived familiarity and perceived importance of family health history (Ashida et al., 2012) and which might facilitate family communication. Two types of worry might influence family communication about genetic test results. Worry about cancer among cancer survivors relates to additional cancer related tests, new primary cancers and recurrence (Jensen et al., 2010; Gotay & Pagano, 2007). Cancer worry more generally has been associated with risk perception (Hay et al, 2005; Shiloh et al., 2013), depression and anxiety (Deimling et al., 2005) and a greater interest in genetic testing (Cameron & Reeve, 2006). Genetic worry relates to how people feel in regards to the impact of one’s genetics on her health and relatives’ disease risks and could affect communication.

This study investigates whether communication of cancer-related genetic test results to immediate family members by women diagnosed with breast cancer at a young age was associated with clinical factors, psychosocial factors, and family members’ gender. Based on prior evidence we hypothesized participants would communicate genetic test results to more types of immediate family members when: (1a) they carry a BRCA1/2 gene mutation and/or (1b) have a strong family history of breast cancer. Additionally, we hypothesized participants would communicate genetic test results to more types of immediate family members when they had greater: (2a) interest in genomic information, (2b) genetic self-efficacy, (2c) genetic causal belief for breast cancer, (2d) informational norms related to healthy habits, (2e) genetic worry, and (2f) cancer worry. In addition, we examined whether these factors were different based on the gender of the family members. Understanding the factors that affect family communication about genetic test results is critical so that strategies can be developed to support these processes.

Methods

Participants

The participants in this study were recruited from an existing nationwide cohort (Young Women’s Breast Cancer Program; YWBCP; http://www.siteman.wustl.edu/ywbcp.aspx) comprised of women diagnosed with breast cancer at the age of 40 or younger. Before any contact, we searched the Social Security Death Index and removed those individuals who were deceased. We mailed invitation letters to 1778 YWBCP members inviting them to participate in a survey online, by telephone, or by mail, followed by an email invitation that contained a link to the online survey. We sent two follow-up emails with links to the survey to those who did not opt-out of further contact, followed by a paper version of the survey by mail. Sixteen participants without an email address were sent a paper version of the survey, with a follow-up paper version. Of those contacted, 1080 (61%) women completed the survey. As the purpose of this study was to examine family communication about genetic test results, we limited the analysis to those participants who reported having had cancer genetic testing. Of the 1080 participants that completed the survey, 920 (85.2%) met these criteria. All participants reviewed a consent information sheet or were read the information by phone. The survey took about 20–30 minutes to complete and participants received a $10 gift card. The study procedures were approved by the Washington University School of Medicine Human Research Protection Office.

Measures

Outcome variables

Family Communication

The types of first-degree relatives to whom genetic test results were communicated was assessed using a proportion measure used in previous research (Ashida et al., 2012). The numerator of the proportion score was the number of types of first-degree relatives to whom the patient had communicated genetic test results. To calculate this, participants were first asked “have you ever shared your genetic test results with any family members?” If the participant responded yes, they were asked to indicate the family members with whom the shared the information (i.e., mother, father, sister(s), brother(s), son(s), daughter(s)). If the participant selected a particular type of family member, a value of one was assigned; a 0 was assigned if not selected. The values were then summed to assess how many different types of first-degree relative to whom the participant had shared genetic test results (range 0–6). To control for family size and types of relatives available with whom to communicate, the denominator of the proportion score was based on participants’ immediate family structure. We asked whether the participant had any sister(s), brother(s), son(s), or daughter(s). All participants were scored as having a biological mother and father with whom to communicate since we did not have data regarding whether these individuals were alive at the time of breast cancer diagnosis or genetic testing. Thus, for the denominator we generated a family structure score of 2 to 6, reflecting the number of types of family members with whom the participant could have shared their genetic test results. We then created the proportion score, which reflected how many types of first-degree relatives with whom the participant communicated genetic test results out of the possible types (range 0–1).

Gendered Family Communication

We created two gendered proportion scores for sharing with male (i.e., father, brother, son) or female (i.e., mother, sister, daughter) immediate family members, following the same procedures. The gendered proportion scores similarly controlled for the number of types of either male or female family members available with whom to communicate. We created a three-level variable that assessed whether participants: (1) shared with more males than females; (2) shared with the same proportion of males and females; or (3) shared with more females than males.

Predictor Variables

Interest in Genomic Information

Participants were asked “how important is it to you to learn more about how your genes affected your chance of getting breast cancer” (McBride et al., 2009). Response options were on a seven point Likert-type scale from not at all important to extremely important. For analysis, the responses were categorized into 3 groups: 7 (extremely important), 5 or 6 (important), or 0–4 (somewhat important or less.)

Genetic Self-Efficacy

Parrott and colleagues (2004) developed a three-item measure to assess genetic self-efficacy (e.g., “I understand how to assess the role of genes for health”). Response options for each item were on a five point Likert-type scale ranging from strongly disagree to strongly agree (Parrott et al., 2004). We averaged the responses to create a genetic self-efficacy score (Cronbach’s alpha of 0.90), which was treated continuously in analysis.

Genetic Worry

We assessed genetic worry with three items (e.g., “your genes put you at increased risk for developing a common disease, like heart disease or diabetes”; “you already have a health condition that was caused primarily by your genes”; “your relatives could be affected with a genetic condition that you have passed on”; Biesecker et al., 2009). The seven point Likert-type response option scales ranged from not at all worried to extremely worried. We calculated an average genetic worry score (Cronbach’s alpha of 0.74), which was treated continuously in analysis.

Cancer Worry

We included the three items from the Assessment of Survivor Concerns (ASC; Gotay & Pagano, 2007) Cancer Worry subscale, which assessed worry related to future tests, another type of cancer, and reoccurrence based on a four point Likert-type scale ranging from not at all to very much (Gotay & Pagano, 2007). We calculated an average cancer worry score (Cronbach’s alpha of 0.78), which was treated continuously in analysis.

Genetic Causal Belief

We assessed genetic causal belief for breast cancer with one item: “how much do you think genetic make-up determines whether or not a woman will get breast cancer?” (McBride et al., 2009). We have used the adapted item in previous studies (Ashida et al., 2011; Ashida et al., 2012). The item was answered on a five point Likert-type scale from not at all (1) to completely (5). For analysis, responses were dichotomized as high (4–5) or somewhat to low (1–3).

Informational Norms

We assessed informational norms with two items: “the people who mean the most to me think I should learn more about ways I can keep myself healthy” and “how motivated you would say you are to do what these people want you to do?” (Hay et al., 2012). Participants responded on seven point Likert-type scales. For analysis, the responses were categorized as scores of 5–7 (agree), 4 (neither agree or disagree), or 1–3 (disagree) for the first item and 5–7 (motivated) or 1–4 (somewhat motivated or less) for the second item.

Clinical predictors

BRCA 1/2 mutation status was collected from the YWBCP database and characterized as positive (i.e., has a pathogenic variant in either BRCA1 or BRCA2), negative (i.e., no variant identified in either BRCA1 or BRCA2), variant (i.e., has a variant of unknown clinical significance), or unknown (i.e., history of genetic testing was unavailable). Because our main interest was the effect of carrying a known pathogenic variant on communication, we combined “negative” and “variant” in the analysis. Family history of breast cancer was scored by an experienced genetic counselor (JI) and classified as strong (i.e., one first- or second-degree relative diagnosed <50, two relatives diagnosed at any age, or male relative diagnosed), moderate (i.e., one first- or second-degree relative diagnosed ≥50) or low (i.e., no first- or second-degree relatives diagnosed/ too little information to characterize family history, i.e., adopted).

Covariates

We collected data on the following clinical covariates from the YWCBP database: diagnosis with more than one primary cancer, current age, and time since diagnosis. We assessed educational attainment, marital status, and household income on the survey.

Data Analysis

We examined bivariate associations between the two outcomes (family communication and gendered family communication) and the psychosocial and clinical predictors. To analyze the family communication outcome, we used generalized linear models with a binomial probability distribution. We built two multivariable generalized linear models, one with significant clinical predictors (Table 3, Model 1) and one including the addition of significant psychosocial factors (Table 3, Model 2). We also used generalized logit models to analyze the gendered family communication outcome using significant predictor variables from the bivariate analyses (Table 4). Marital status was included as a covariate in the model. Statistical significance was assessed as p < 0.05. All statistical analyses were conducted using SAS/STAT version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Characteristics of Participants

As shown in Table 1, about 29% of participants were classified as having a strong family history of breast cancer and 13% carried a deleterious mutation in BRCA1 and/or BRCA2. The mean current age was 45 years old; mean time since diagnosis was 9.5 years. The majority of participants were married (80%) and had a college degree or higher level of education (82%). As shown in Table 2, respondents were most likely to communicate test results to a mother (83%) and least likely to a son (45%). On average, participants had discussed their genetic test results with most but not all possible family members (proportion of 0.7).

Table 1.

Characteristics of respondents

Characteristics	N	%
BCRA1/2 Mutation Status (N = 920)
Positive	119	12.9%
Negative / Variant	694	75.4%
Unknown	107	11.6%
Family History of Breast Cancer (N = 918)
Unknown or Low	483	52.6%
Moderate	169	18.4%
Strong	266	29.0%
Diagnosis with 1+ Primary Cancer (N = 918)	131	14.3%
Educational Attainment (N = 919)
Some college or less	170	18.5%
College degree	328	35.7%
More than college degree	421	45.8%
Marital Status (N = 919)
Married/Living as Married	732	79.7%
Widowed/Divorced/Separated	103	11.2%
Never Married	84	9.1%
Household Income (N = 809)
< $100,000	377	46.6%
Interest in genomic information (N=917)
Somewhat important or less	175	19.1%
Important	319	34.8%
Extremely important	423	46.1%
Genetic causal belief (N=920)
Low	664	72.2%
Social influences on learning more (N=920)
Disagree	383	41.6%
Neither agree nor disagree	204	22.2%
Agree	333	36.2%
Social influences affecting motivation (N=910)
Somewhat motivated or less	399	43.9%
Motivated	511	56.2%
	M (SD)	Range
Age, years (N = 920)	44.6 (8.1)	26 – 76
Time since diagnosis, years (N = 830)	9.5 (7.0)	0 – 49
Interest in genomic information (N=917)	5.9 (1.4)	1 – 7, scale max 7
Genetic self-efficacy (N=920)	3.2 (0.7)	1 – 5, scale max 5
Genetic worry (N=920)	4.0 (1.5)	1 – 7, scale max 7
Cancer worry (N=920)	2.8 (0.8)	1 – 4, scale max 4
Genetic causal belief (N=920)	3.2 (0.7)	1 – 5, scale max 5
Social influences on learning more (N=920)	3.7 (2.0)	1 – 7, scale max 7
Social influences affecting motivation (N=910)	4.8 (1.5)	1 – 7, scale max 7

Table 2.

Communication to immediate family members: Types of family members, gender family communication and family communication proportion scores

	Have family member type		Communicated genetic test results to
	N	%	N	%
Type of family member
Mothera	920	100%	760	82.6%
Fathera	920	100%	603	65.5%
Sister(s)	597	64.9%	561	94.0%
Brother(s)	600	65.2%	439	73.2%
Daughter(s)	476	51.7%	307	64.5%
Son(s)	454	49.4%	203	44.7%
	N		%
Shares more with females	397		43.15
Shares more with males	130		14.13
Shares equally	393		42.72
	Mean		SD
Family communication proportion	0.73		0.24
Male communication proportion	0.64		0.37
Female communication proportion	0.82		0.26

^aParticipants were scored as having a living biological mother and father in this analysis.

Effect of Clinical and Psychosocial Factors on Family Communication

In bivariate analysis examining the associations between clinical characteristics and family communication, carrying a BRCA1/2 mutation (X²= 10.21, p <.001) was significantly associated with more family communication but family history of breast cancer was not significantly related to this outcome. In bivariate analysis, the following psychosocial characteristics were significantly associated with a higher proportion of family communication: higher interest in genomic information (X²= 12.11, p =.001), higher genetic self-efficacy (X²= 6.60, p = .01), higher informational norm for motivation to be healthy (X²= 4.87, p = .03), and lower genetic worry (X²= 3.49, p = .06). Genetic causal beliefs and cancer worry were not significantly related to family communication (data not shown).

We first built a multivariable model examining the association of clinical characteristics with family communication (see Table 3, Model 1). We then added psychosocial factors to the model (Table 3, Model 2). In this latter model, those with lower genetic worry (OR=0.91; 95% CI = 0.86, 0.96) and higher interest in genomic information (OR=1.55; 95% CI = 1.26, 1.91) communicated about genetic test results to more types of family members. In addition, carrying a BRCA1/2 mutation (OR=1.34; 95% CI = 1.06, 1.70) was associated with communication to more types of family members. Never having been married (OR=1.45; 95% CI = 1.05, 1.98) was associated with communication with more types of family members and longer time since diagnosis (OR=0.98; 95% CI = 0.97, 0.99) was associated with communication with fewer types.

Table 3.

Clinical and psychosocial predictors of family communication in multivariable models

	Model 1 (n=829)			Model 2 (n=826)
Predictor variable	OR	95% CI	p-value	OR	95% CI	p-value
BRCAa
Positive	1.24	0.99 – 1.57	0.066	1.34	1.06 – 1.70	0.016*
Unknown	0.91	0.72 – 1.14	0.414	0.96	0.76 – 1.21	0.703
Interest in Genomic Informationb
Extremely Important	----	----	----	1.55	1.26 – 1.91	<0.001*
Important	----	----	----	1.39	1.12 – 1.71	0.002*
Genetic Worry	----	----	----	0.91	0.86 – 0.96	<0.001*
Marital Statusc
Never married	1.47	1.08 – 2.02	0.016*	1.45	1.05 – 1.98	0.022*
Widowed/Divorced/Separated	1.00	0.79 – 1.26	0.997	0.96	0.76 – 1.21	0.723
Time since diagnosis	0.98	0.97 – 0.99	<0.001*	0.98	0.97 – 0.99	<0.001*

^aComparison group is negative/variant

^bComparison group was somewhat important or less

^cComparison group is married/living with partner

^*p<0.05

Effect of Gender on Family Communication

Participants shared with more types of female than male family members (proportions of 0.82 and 0.64, respectively, p<0.001). We separately examined factors associated with communicating genetic test results with more types of male family members and factors associated with more communication with female family members. In bivariate analysis, carrying a BRCA1/2 mutation (X² = 9.06, p =.011) and family history of breast cancer (X² = 10.94, p =.027) were significantly associated with discussing genetic test results with more types of male family members. For the psychosocial factors, high interest in genomic information (X² = 7.25, p = 0.03) was associated with communication with more types of male family members. In bivariate analysis, none of the clinical or psychosocial factors were significantly associated with communication with more types of female family members.

We then built a multivariable model to examine the clinical and psychosocial variables associated with gendered family communication (Table 4). Carrying a BRCA1/2 mutation (OR=1.72; 95% CI = 1.02, 2.89) was significantly associated with higher odds of discussing genetic test results with more types of male family members as compared to sharing with an equal proportion of male and female types. Never having been married was associated with discussing genetic test results with more types of female family members as compared to an equal proportion (OR=0.45; 95% CI = 0.27, 0.76).

Table 4.

Predictors of gendered family communication in multivariable models (n = 919)

	Shared with more female categoriesc			Shared with more male categoriesc
Predictor variable	OR	95% CI	p-value	OR	95% CI	p-value
BRCAa
Positive	0.72	0.46 – 1.12	0.143	1.72	1.02 – 2.89	0.043*
Unknown	1.50	0.96 – 2.36	0.078	1.37	0.71 – 2.62	0.349
Marital Statusb
Never married	0.45	0.27 – 0.76	0.002*	0.63	0.31 – 1.29	0.205
Widowed/ Divorced/ Separated	1.18	0.75 – 1.87	0.475	1.44	0.78 – 2.64	0.245

^aComparison group is negative/variant

^bComparison group is married/living with partner

^cComparison group is equal proportion of male and female familial categories

^*p<0.05

Discussion

In this study, we examined communication of genetic test results within families of women diagnosed with breast cancer at a young age. We focused on investigating both clinical and psychosocial factors that affect patterns of family communication for this patient population. Specifically, we investigated the extent to which different types of family members such as parent, siblings, and children receive the information and potential differences based on the gender of these members.

We found partial support for our first hypothesis. Carrying a BRCA1/2 mutation was significantly associated with communication of genetic test results to more types of family members but family history of breast cancer was not associated with this outcome. Knowledge about the presence or absence of a BRCA1/2 mutation assists not only the patient but also the entire family (McGivern et al., 2004), and a patient may feel obligated to spread the genetic information to family members they feel might be at risk (Wiseman et al., 2010). However, communication with family members about a negative genetic test result is less common. Knowing about a negative genetic test result can help family members with screening guidelines and other medical decisions. Many of these family members may still be at increased risk for cancer due to their family history, despite a negative BRCA result. Our findings suggest that patients and families might need support from the health care provider to understand the meaning of a negative genetic test result, and how to communicate it to family members.

We also found support for the importance of genetics-related cognitions in family communication; higher interest in genomic information and lower genetic worry were associated with greater family communication. Participants who are interested in genomic information believe it is important to learn more about genetics (McBride et al., 2009). Individuals with a higher interest in genomic information might assume that other family members are also interested in the information, potentially increasing sharing. Contrary to our hypothesis, we did not find that cancer worry was associated with family communication. This result is consistent with Cheung and colleagues (2010), who also found that cancer worry, specifically for mood and daily activities, was not significantly related to family communication of genetic test results. However, in our study, those with higher genetic worry had less family communication about genetic test results. Because genetic worry relates to both the patient’s and family members’ disease risks, it may be that participants with high genetic worry are concerned about the possibility of causing worry among family members by sharing genetic test results or may cope with their own worry by not talking about the information. McGivern and colleagues (2004) found that sharing genetic test results with family is selective and genetic information might not be shared with family members if it is a tough topic within the family (McGivern et al., 2004). Genetic casual beliefs were not significantly related to family communication. For this population of young breast cancer survivors, causal beliefs for the development of breast cancer might not be as relevant in predicting family communication about genetic test results as other genetics-related cognitions.

We also found marital status and time since diagnosis were associated with family communication. Interestingly, participants who had never been married reported communication with more types of first-degree relatives than those with spouses or partners. It may be that participants who were never married relied more on their biological family members for support compared with married or partnered individuals. We also found that time since diagnosis was inversely related to family communication. This could reflect a trend toward more open communication about genetic test results in recent years or could be due to recall bias, as more recent conversations are remembered to a greater extent (McGivern et al., 2004). Longitudinal studies are needed to explore changes in family communication patterns over time.

Consistent with prior research, we found genetic test results may be reaching more female than male immediate family members (e.g., Bradbury et al., 2012; McGivern et al., 2004). Our analysis added an examination of the clinical and psychosocial factors that affect this difference by gender. We found participants who carried a BRCA1/2 mutation shared to a greater extent with male family members (as compared to an equal number of male and female family members) than those who were not known to be carriers. A positive test result may highlight the importance of communicating the result to all biological family members, including males, compared to negative or indeterminate results. This finding may have important implications for intervention, as a prior study has shown that when at-risk male family members are informed of the positive mutation, none of the males sought genetic testing (Finlay et al., 2008). These families may need support in determining appropriate follow up for male family members.

It is important to consider how the age of these breast cancer survivors impacted the reported family communication patterns. Their children were likely younger than for many women diagnosed with breast cancer, which could impact the amount of information communicated, as well as the timing and the content of the discussions. As seen with Koopman and colleagues (2004), children have different levels of understanding of illness as they develop cognitively and the content of discussions with young children are different than those with teenagers (Koopman et al., 2004). Participants with young children might not communicate the genetic test results at all (Taber, 2014). Parents report talking to their children about BRCA1/2 genetic test results starting around ten years old (Bradbury et al., 2012). Longitudinal studies are needed to examine communication with children over time.

There were several limitations of the study. The majority of participants were Caucasian and college educated, so the results may not generalize to other populations of young breast cancer patients. In order to reduce participant burden, we did not collect data on the exact number of family members within each relational category to whom genetic test results were communicated. We did not have data regarding whether parents were living at the time of breast cancer diagnosis or genetic testing. It would also be useful to control for time since testing. If some of the participants received genetic testing shortly before the study was completed, they may not have had time to contact as many family members. In addition, we do not know the content of the conversations, including to what extent the information communicated was accurate. We examined self-report of past communication, which could differ from actual communication behaviors. We investigated communication between the patient and a family member, not additional communication between relatives. We did not analyze communication with extended family (e.g., aunts, uncles, cousins) or the differential communication effects of BRCA1/2 mutations on the maternal versus paternal side; both of these are important areas for future research. Other gene mutations that have implications for both men and women could be also investigated to determine if these clinical and psychosocial factors continue to impact family communication. Researchers could also investigate other variables that could impact family communication, such as family characteristics and perceptions of closeness. Future studies may focus on communication within families beyond the transmission of information, as a process that is influenced by a variety of factors including culture, life stage, and family roles (Chaturvedi et al., 2014; Mendez et al., 2016).

Despite the limitations of the study, our findings add to our understanding of communication of genetic information within families. Interest in genomic information and genetic worry were associated with the patterns of family communication about genetic test results, and our findings suggest that strategies to increase or improve communication might be needed for those with low interest in genomic information or high levels of genetic worry. In addition, strategies to support communication might be needed for those who receive negative or indeterminate results. Communication with family members about genetic risk information is valuable for all related family members (Taber, 2014). Identification of factors affecting family communication processes are essential to developing effective strategies to support this type of communication of genetic test results among families of breast cancer patients.