Heart disease risk information, encouragement, and physical activity among Mexican-origin couples: Self- or spouse-driven change?

Wives reported increased physical activity levels if their husbands received increased heart disease risk feedback and health behavior recommendations.

Abstract

Family health history is an accessible, clinically-recommended genomic tool that improves health risk evaluation. It captures both genetic and modifiable risk factors that cluster within families. Thus, families represent a salient context for family health history-based interventions that motivate engagement in risk-reducing behaviors. While previous research has explored how individuals respond to their personal risk information, we extend this inquiry to consider how individuals respond to their spouse’s risk information among a sample of Mexican-Americans. One hundred and sixty spouse-dyads within Mexican-heritage households received a pedigree or a pedigree and personalized risk assessments, with or without behavioral recommendations. Analyses of Covariance (ANCOVAs) were conducted to assess the relationship between risk feedback, both personal and spouse, and self-reported physical activity levels at 3-month and 10-month assessments, controlling for baseline levels. The effect of being identified as an encourager of spouse’s healthy weight was also evaluated. Personal feedback had no effect on participants’ physical activity at either 3- or 10-month assessments. However, husbands’ risk information was associated with wives’ physical activity levels at 3-month assessment, with women whose husbands received both increased risk feedback and behavioral recommendations engaging in significantly higher physical activity levels than all other women. At 10-month follow-up, physical activity levels for both husbands and wives differed depending on whether they encouraged their spouse’s healthy weight. Spousal risk information may be a stronger source of motivation to improve physical activity patterns than personal risk information, particularly for women. Interventions that activate interpersonal encouragement among spouses may more successfully extend intervention effects.

Implications

Practice: Including spouses in discussions about family history-based disease risk can motivate improved health behaviors for both partners.

Policy: Shifting from patient-centered care to family-centered care affords opportunities for providers to broaden the reach of risk-reducing recommendations from individuals to families.

Research: Future research is needed to more fully understand the interpersonal mechanisms among spouses that shift motivations towards engagement in healthful behaviors.

INTRODUCTION

Family health history is an accessible, clinically-recommended genomic tool that is well-established in its ability to improve health risk evaluation [1, 2]. In addition, there is a nascent but growing field of inquiry into understanding how family health history-based risk information may influence health behavior change [3–7]. The current study integrates this area of research with substantial evidence demonstrating the positive influence spouses have on one another’s health behaviors [8–10]. While previous research has explored how individuals respond to their own family health risk information [11–13], we extend this inquiry to consider how individuals respond to risk information about one’s spouse among a sample of Mexican-origin families potentially at risk of heart disease.

THE IMPORTANCE OF FAMILY HEALTH HISTORY FOR MOTIVATING BEHAVIOR CHANGE

Several studies point to the potential for family history-based interventions to motivate individuals and families to initiate screening behaviors and reduce risk behaviors in both clinical and community settings [6, 14–16]. For example, using messages tailored to individuals’ family health history-based risk and current health behaviors, the Family Healthware Impact Trial demonstrated that patients in primary care were significantly more likely to improve their fruit and vegetable consumption and physical activity levels, compared to patients who received generic health-related messages [17]. However, messages from the Family Healthware Impact Trial included both personalized risk assessments, based on patients’ family health history, and behavioral recommendations, tailored to patients’ risk assessment, lifestyle habits, and screening behaviors, limiting our ability to disentangle what underlies observed behavior change. As well, for ethnic minority and immigrant populations, culturally-tailored family health history risk information has been shown to successfully promote family communication, is associated with intentions to improve diet and physical activity, and resulted in engagement in risk reducing behaviors [14, 18, 19]. Given these results and recent research suggesting that a healthy lifestyle can reduce heart disease risk by nearly 50% for those at increased risk [20], family health history-based risk information has the potential to improve health outcomes by motivating engagement in healthy behaviors.

INFLUENCE OF THE SPOUSAL RELATIONSHIP ON HEALTH BEHAVIORS

Marriage has been repeatedly shown to be consequential for physical health and well-being, and this influence extends to the health behaviors performed by members of the couple [21, 22]. For example, Falba and Sindelar demonstrated that a spouse’s behavior was a salient influence on individuals’ behavior change related to smoking, drinking, exercising, cholesterol screening, and flu vaccination [23]. Furthermore, in qualitative work, a sample of African-American men reported that wives’ food decisions were often driven by a desire to improve husbands’ diets [24]. Indeed, in heterosexual couples, it may be the wives, in particular, whose functioning and behavior are influenced by the health status of their husband [25]. Although nonconsanguineous spouses do not share family health histories, the social roles that individuals inhabit within the family system may, in fact, lend personal meaning to a spouse’s family history-based risk information, which in turn could motivate improved lifestyle habits.

Despite the considerable evidence of the impact of spouses on each other’s health behaviors, behavior change interventions that jointly engage wives and husbands have been largely unsuccessful [26]. Lewis and colleagues proposed that, for couples, a number of individual and dyadic characteristics, as well as interpersonal mechanisms (e.g., a couple’s communication style), may affect whether spouses respond collectively to a shared health stressor, such as one’s personal or partner’s disease risk [27]. Individual characteristics, such as gender, may be particularly important, as women tend to be instrumental in facilitating the adoption of household or family-level strategies aimed at reducing risk [28]. As well, interpersonal mechanisms, such has behavioral encouragement may be particularly important in motivating healthful behaviors.

Behavioral encouragement is an interpersonal mechanism that seeks to influence behaviors through social support provision; such support may occur through communication pathways, social facilitation, or social modeling [29, 30]. An individual’s poor health practices could prompt encouragement by others [31, 32]. In turn, individuals’ perceived receipt of behavioral encouragement is associated with both intentions to engage and self-reported engagement in screening and improved lifestyle behaviors, including physical activity [14, 19, 30, 33–37]. However, there are few examples that investigate how encouraging others may influence the encourager’s own behavior. Moreover, the majority of research investigating behavioral encouragement as an interpersonal influence mechanism has focused on broader family influences or parent–child interactions; there are few examples that consider behavioral encouragement among spousal couples. To fill these gaps, we investigate whether individuals change their behavior in response to their spouses’ family history-based risk information in addition to their own. As well, we assess whether individuals’ behavioral responses are greater when they are identified by their spouses as actively encouraging their spouses to engage in healthful behaviors.

THE CURRENT STUDY

Here, we investigate whether the provision of family history-based risk assessments to Mexican-heritage spouses influences engagement in healthful lifestyle habits associated with heart disease risk. Heart disease persists as the leading cause of mortality in the USA, accounting for nearly a quarter of total deaths in 2010 [38, 39]. Hispanic Americans are at elevated risk for several heart disease risk factors [40]. And, compared to non-Hispanic whites, Mexican-Americans, the largest subgroup of Hispanic Americans, are more likely to exhibit such heart disease risk factors as overweight or obesity, diabetes, and physical inactivity [41]. These disparities likely emerge from the combination of genetic, lifestyle, and environmental factors which often contribute to one’s family history.

The impact of an individual’s family history-based disease risk on his or her spouse’s risk-reducing behaviors has not been examined. Thus, the current study investigates whether individuals of Mexican heritage engage in increased physical activity in response to heart disease risk information relevant to themselves or their spouse; our study design allows us to evaluate whether provision of behavioral recommendations along with disease risk information influences spouses’ behavioral response. In addition, we investigate whether individuals who encourage their spouse to engage in healthful behaviors are more likely to improve their own behaviors. Uniquely, we use spousal reports of perceived encouragement received, rather than the individual’s self-report of perceived encouragement provided. This approach considers that influence tactics may be perceived by the recipient as discouraging or manipulative. Here, by assessing perceived encouragement received, we focus on only those tactics that are perceived by the recipient as encouraging. We thus hypothesize that an important aspect of being perceived as an encourager is engaging in the behavior that is being encouraged. As such, it may be that assessing perceived encouragement receipt as opposed to provision could isolate more directly the aspects of that encouragement process that are ultimately associated with spousal health behavior change. If we find evidence that such interpersonal mechanisms impact behavior, the efficacy of public health interventions to engage at-risk communities in physical activity and reduce heart disease risk may be improved by leveraging spousal relationships. We hypothesize, based on theory and extant literature, that individuals will respond to both their own and their spouse’s heart disease risk and that this response will be strengthened when behavioral recommendations accompany risk assessment and if individuals are identified as encouragers of their spouses’ healthful behaviors.

METHOD

Study design and procedures

The data for the current report come from (Project Risk Assessment for Mexican Americans [RAMA]). A total of 162 multigenerational households participated; households were recruited in a southwestern metropolitan city and surrounding area, through the population-based (Mano a Mano cohort) cohort of Mexican-origin families maintained by (the University of Texas MD Anderson Cancer Center) [42]. Three to four adult members from each household participated, with 160 households including married couples.

Baseline data were collected during in-home interviews with bilingual interviewers using tablet computers. Each participant independently provided their family health history regarding first- and second-degree relatives, as well as their personal disease status for four diseases: heart disease, diabetes, colon cancer, and breast cancer. Questions used to obtain participants’ family health history mirrored those used within the Surgeon General’s My Family Health Portrait [43] and the Centers for Disease Control and Prevention’s Family Healthware [44]; questions included relatives’ kinship role, disease diagnoses with the option of “don’t know” responses, and age of diagnosis. The family history information obtained is consistent with the information patients bring to their healthcare providers; participant and relatives’ reported diagnoses were not verified through medical records. The current report focuses on family history of heart disease as heart disease is the second leading cause of death for those of Hispanic origin [40].

Households were randomized to one of four feedback conditions characterized by the following two factors: (i) the number of household participants receiving family history based personalized risk assessments (one participating household member or all participating household members) and (ii) whether behavioral recommendations aimed at reducing risk accompanied the personalized risk assessments. Family Healthware, a tool developed by the Centers for Disease Control and Prevention (CDC), was used to generate personalized risk assessments and behavioral recommendations; the risk algorithm and message content have been detailed elsewhere [14, 44, 45]. Generated message content was revised by the study team to a 8th grade reading level and translated into Spanish [44]. Participants received feedback packets, either in English or Spanish based on participant preference, within 1 week of completing their baseline assessments. Follow-up measurements occurred at 3 and 10 months postfeedback receipt on average and were conducted through telephone interview by bilingual interviewers blinded to household random assignment. Participants received a $20 gift card each time they completed an assessment. The CONSORT diagram is presented in Fig. 1, with 159 wives and 151 husbands completing the 3-month assessment and 156 wives and 145 husbands completing the 10-month assessment.

Fig 1.

Study CONSORT diagram

The current paper focuses on the participants that completed all assessments, including baseline, 3-month (first follow-up) assessment, and 10-month (second follow-up) assessment. Those with missing data on key variables across the three assessments were dropped resulting in a final sample of 137 wives and 126 husbands. There were no significant differences between those retained and those dropped due to missing data on heart disease risk, baseline physical activity levels, BMI, or most demographic characteristics. However, wives born in Mexico were more likely to have complete data for analysis compared to those born elsewhere.

Both spouses in all households received a pedigree, or family tree, representing their self-reported family health history, a primer on how to interpret their pedigree, and neighborhood resources for health care and related support. In households where only one participating member received personalized risk assessments, the recipient was randomly assigned to be either the wife or husband of the spousal couple. Thus, for each household, either one or both spouses received personalized risk assessments or personalized risk assessments and behavioral recommendations. For households in which only one spouse member received personalized risk information, the other received only their pedigree with interpretation primer and neighborhood resources.

Measures

Personalized risk feedback and behavioral recommendations

The Family Healthware algorithm resulted in three risk categories for heart disease: weak risk (average, population risk), moderate risk, and strong risk [45]. Due to sample size constraints, those spouses who received no personalized risk assessment (pedigree only) and those spouses receiving a weak risk assessment in absence of behavioral recommendations were combined into a “no risk assessment/average risk” category. Similarly, the moderate and strong risk categories were collapsed into one overarching “increased risk” category, consistent with previous analyses [46]; participants’ with a personal diagnosis of heart disease were also included in this increased risk category. Behavioral recommendations were tailored to individuals’ risk and baseline health behavior, including screening for heart disease risk factors (i.e., hypertension and high cholesterol) and engagement in risk reducing behaviors (e.g., being physically active, eating a healthy diet, and reducing/maintaining a healthy weight through these activities) [6, 14, 44]. Thus, four risk feedback categories could be observed for each wife and husband based on household randomization and the personalized content of each spouse’s feedback: (i) no risk feedback/average risk, no recommendations; (ii) average risk, with recommendations; (iii) increased risk, no recommendations; and (iv) increased risk, with recommendations.

Demographics

Demographic characteristics including age, place of birth, educational attainment, and health insurance status were obtained through self-report at baseline assessment. Body mass index (BMI) was calculated based upon participants self-reported height and weight at baseline [47].

Physical activity

Physical activity was measured using items from the CDC’s Family Healthware tool, which were used to tailor behavioral recommendations; questions were slightly modified to resonate more strongly with our study population [44]. At each assessment point, participants were asked, “On average, how many times per week do you participate in physical activity, such as: Walking, Mowing the Lawn, Running, Gardening, Exercise Classes, Dancing, Bicycling, Soccer, Swimming?” (Never; less than once a week; one to two times a week; three to four times a week; five or more times a week) and “On average, how long do you do these activities each time?” (Less than 10 min; 10–19 min; 20–29 min; 30–39 min; 40 or more minutes). Responses to these two questions were multiplied based on median values within each response range to obtain an approximation of minutes of physical activity per week.

Encourager of healthful behavior

Established procedures for measuring ego-centered networks [48] were employed to identify encouragers of healthful behaviors. At baseline, each spouse enumerated important people in his or her life, with opportunity to modify these lists at each consequent follow-up assessment. From the list of enumerated people, spouses identified those who encouraged them to maintain a healthy weight. A wife was considered an encourager if her husband indicated that she had encouraged him to maintain a healthy weight; similarly, a husband was considered an encourager if his wife indicated that he encouraged her to maintain a healthy weight. Two different encourager variables were constructed for each wife and husband: baseline encourager, dichotomized as the participant did/did not provide encouragement to his/her spouse, based on the spouse’s report, at baseline, and new encourager, dichotomized as the participant did/did not provide new encouragement to his/her spouse at follow-up. A new encourager was a spouse who did not provide encouragement at baseline, but encouraged his/her spouse at the 3-month and/or 10-month follow-up assessment, based on spousal report.

Statistical analysis

Stratified analyses, rather than a dyadic analysis, were conducted to address the proposed research questions for three reasons. First, physical activity levels were uncorrelated amongst spouses (baseline: r < .01; 3-month: r = −.04; 10-month: r = .03), indicating no within couple dependence on the outcome. Second, dyadic analysis would reduce the sample size further due to missing data, thereby impacting inferential power. Finally, given possible equifinality—that is, the potential for husbands and wives to achieve similar outcomes through different mechanisms—investigation of such variable effects would require fitting three-way interactions within a dyadic analysis; such interactions can be better investigated through stratified analysis [49].

To investigate differences in mean physical activity at 3-month and 10-month follow-up assessments due to personal and spouses’ personalized risk feedback, Analyses of Covariance (ANCOVAs) were conducted using SPSS version 21.0. The full model, with main effects of personal and spouses’ risk feedback, along with the interaction effect between personal and spouses’ risk feedback was fitted; the interaction was subsequently dropped from the model because it was not significant, and a main effect only model was fitted. Given the moderate sample size, covariates were limited to variables significantly associated with the 3-month and/or 10-month physical activity measures—BMI and baseline physical activity. For significant main effects or interaction effects, post-hoc analyses were conducted using the Least Significant Difference (LSD) procedure.

Similarly, ANCOVAs were conducted to test for a significant moderating effect of being an encourager on the relationship between spouse’s risk feedback and physical activity at 3-month and 10-month follow-ups. A hierarchical model was fitted, which included the main effects of personal and spouse’ risk feedback content, the main effects of baseline and new encouragement, as well as the interaction between encouraging healthful behaviors and spouses’ risk feedback.

RESULTS

Descriptives

Table 1 summarizes sample characteristics and the distribution of feedback content. With respect to physical activity, at baseline assessment, wives reported an average of 76.1 min per week of physical activity (SD = 67.2), compared to an average of 97.9 min per week reported by husbands (SD = 76.6). Over the course of the intervention, wives demonstrated a significant increase in their average amount of weekly physical activity compared to baseline assessment (3-month: 95.0 min [SD = 70.5], p = .002; 10-month: 102.5 min [SD = 72.0], p < .001). However, husbands did not (3-month: 102.7 min [SD = 72.4], p = .56; 10-month: 109.0 min [SD = 73.8], p = .15).

Table 1.

Baseline characteristics of participants, physical activity levels, and contents of risk feedback

	Wives (n = 137)	Husbands (n = 126)
Demographics
Mean (SD) age (years)	47.9 (9.1)	49.7 (10.0)
% born in Mexico	85.4	81.7
% at most high school diploma/GED	85.1	84.1
% without health insurance	41.6	36.5
% obese (BMI > 30)	55.5	42.9
Physical Activity Levels
Mean (SD) physical activity (minutes/week)—Baseline	76.1 (67.2)	97.9 (76.6)
Mean (SD) physical activity (minutes/week)—3-month assessment	95.0 (70.5)	102.7 (72.4)
Mean (SD) physical activity (minutes/week) —10-month assessment	102.5 (72.0)	109.0 (73.8)
Risk Feedback Packet Contents
% No risk assessment/average risk assessment, no recommendations	43.1	39.7
% Average risk assessment, with recommendations	16.8	21.4
% Increased risk assessment, no recommendations	22.6	20.6
% Increased risk assessment, with recommendations	17.5	18.3

BMI body mass index; GED general educational development test.

Impact of risk feedback on self-reported physical activity

There was no significant interaction between personal and spouses’ personalized risk feedback on physical activity levels at 3-month or 10-month assessment for husbands or wives, thus, the term was removed from the model. Surprisingly, personal risk assessments had no effect on wives’ physical activity at either 3-month (F[3,128] = 0.64, p = .59; partial η² = .015) or 10-month follow-up assessment (F[3,128] = 0.50, p = .68; partial η² = .012), controlling for baseline levels and BMI. A main effect, however, was observed for husbands’ risk feedback on wives’ physical activity levels at 3-month follow-up assessment (F[3,128] = 4.00, p = .01; partial η² = .086). Post-hoc analyses indicated that women who were partnered to men receiving an increased risk assessment, along with behavioral recommendations, engaged in significantly higher levels of physical activity than all other women (estimated marginal mean differences: 32.8, 38.3, and 54.4 more minutes than those whose husbands received no risk assessment/weak assessment, an increased risk assessment, or a weak risk assessment with recommendations, respectively). At 10-month follow-up, however, the main effect of husbands’ risk information on wives’ physical activity levels was no longer significant (F[3,128] = 1.73, p = .17; partial η² = .039).

Likewise, there was no significant main effect of personal risk feedback on husbands’ physical activity levels at 3-month (F[3,117] = 0.19, p = .90; partial η² = .005) or 10-month follow-up (F[3,117] = 1.26, p = .29; partial η² = .031), controlling for baseline levels. In contrast to their wives’ response, there was no effect of wives’ personalized risk feedback on husbands’ physical activity levels at 3-month (F[3,117] = 0.10, p = .96; partial η² = .002) or 10-month follow-up (F[3,117] = 0.18, p = .91; partial η² =.005).

Moderating effect of being a behavioral encourager

Test statistics evaluating main effects and interaction effects involving behavioral encouragement are presented in Table 2. At 3-month follow-up, there was no main effect of being an encourager at baseline or a new encourager on the physical activity levels of wives or husbands. However, at 10-month follow-up, there was a significant main effect of wife being a new encourager on wives’ physical activity (partial η² = .033); wives whose husbands reported receiving spousal encouragement engaged in significantly higher levels of physical activity than when no encouragement was reported (estimated marginal mean of 128 min vs. estimated marginal mean of 93 min). For husbands, there was a significant interaction at 10-month follow-up between wives’ risk feedback and wives reported receipt of new encouragement (partial η² = .088). Fig. 2 provides estimated marginal means for this interaction. Post-hoc analyses indicate that when wives received an increased heart disease risk message with behavioral recommendations, husbands who were new encouragers of their spouse’s healthy weight engaged in significantly higher levels of physical activity (estimated marginal mean of 180 min) at 10-month follow-up than husbands who did not encourage their wives (estimated marginal mean of 84 min, p = .014). A similar result was observed for the husbands who were new encouragers, comparing those whose wives received an increased risk assessment with behavioral recommendations to those whose wives received a no feedback/average risk assessment, without recommendations (estimated marginal mean of 55 min, p < .01) or those who received an average risk assessment with behavioral recommendations (estimated marginal of 73 min, p = .03).

Table 2.

ANCOVA main and interaction effects of heart disease risk feedback and behavioral encouragement

	Wives (n = 137)	Husbands (n = 126)
	F(df), p-value	F(df), p-value
3-Month follow-up
Risk feedback (personal)	F(3,120) = 0.44, p = .72	F(3,109) = 0.22, p = .89
Risk feedback (spouse)	F(3,120) = 0.67, p = .57	F(3,109) = 0.30, p = .83
Baseline encouragement	F(1,120) = 0.20, p = .66	F(1,109) = 0.55, p = .46
New encouragement	F(1,120) = 0.01, p = .93	F(1,109) = 0.00, p = .98
Interactions:
Baseline encouragement × Risk feedback (spouse)	F(1,120) = 0.77, p = .52	F(1,109) = 0.27, p = .84
New encouragement × Risk feedback (spouse)	F(1,120) = 0.14, p = .94	F(1,109) = 1.35, p = .26
10-Month follow-up
Risk feedback (personal	F(3,120) = 0.29, p = .83	F(3,109) = 1.88, p = .14
Risk feedback (spouse)	F(3,120) = 1.00, p = .40	F(3,109) = 1.32, p = .27
Baseline encouragement	F(1,120) = 2.44, p = .12	F(1,109) = 0.18, p = .67
New encouragement (3- and 10-month)	F(1,120) = 4.10, p = .04	F(1,109) = 0.60, p = .44
Interactions
Baseline encouragement × Risk feedback (spouse)	F(3,120) = 1.78, p = .15	F(3,109) = 0.30, p = .83
New encouragement × Risk feedback (spouse)	F(3,120) = 2.30, p = .08	F(3,109) = 3.50, p = .02

Covariates: baseline physical activity, BMI.

Fig 2.

Model estimated mean minutes per week physical activity for wives at 10 months by husband’s risk feedback and wife’s encourager status (A). Model estimated mean minutes per week physical activity for husbands at 10 months by wife’s risk feedback and husband’s encourager status (B)

DISCUSSION

Even as sophisticated genomic technologies develop at an unprecedented pace, family health history remains an important and accessible clinical tool for evaluating patients’ risk and tailoring their clinical care. Increased availability of personalized medical information is accompanied by a need to understand how patients and their family members accept and respond to this information. The current study underscores the shortcomings of targeting personalized risk information toward individuals without first considering the potential influence of their family environment [50]. Moreover, this work highlights the potential benefit of individuals’ personalized risk information on close others. The family-based design of (Project RAMA) afforded the unique opportunity to assess the impact of both personal risk information as well as spousal risk information on physical activity, a health behavior key to the prevention and management of chronic diseases that tend to cluster within families. Our results suggest that spousal risk may be a more salient source of motivation to increase physical activity levels than an individual’s personal risk information.

Within many family systems, spouses function as important sources of support and influence in making health behavior changes. The interdependence model of communal coping and behavior change suggests that wives are especially motivated to activate communal coping strategies in the face of a family member’s health threat [27]. Communal coping represents a cooperative problem-solving process in which spouses appraise their personal or spouse’s disease risk as a joint problem requiring cooperative action to solve [51]. Lewis and colleagues posit that because women frequently inhabit caretaking roles within their family systems, they are likely to interpret a health threat as significant not only for the affected individual, but for the relationship as well. In turn, wives employ influence strategies in an effort to motivate husbands’ adoption of healthful behaviors. Literature on Latino family functioning further underscores the influential role of the wife–mother on behavior change within the family system [52]. Accordingly, in this study, when husbands received an increased risk assessment for heart disease, with behavioral recommendations, wives may have construed meaning for the couple (and/or family) opting to themselves engage in risk-reducing behaviors. Moreover, our results show that wives who began encouraging healthful behaviors in response to risk feedback, reported a substantial increase in physical activity almost a year later compared to those who did not initiate this new interpersonal behavior potentially indicating activation of a communal coping process.

As well, our results suggest that provision of behavioral recommendations in conjunction with risk information might be particularly important if the goal is to motivate behavior change. The provision of risk-reducing behavioral recommendations, including messages about healthy weight and physical activity, may empower spouses to encourage the behavior, especially when provided to a close family member who is at increased risk of heart disease [32]. While wives demonstrated a sustained increase in their physical activity levels when they activated behavioral encouragement processes, regardless of their husbands’ risk assessment, husbands exhibited such behavior change only when their wives were at increased risk. The temporal sequencing and connections between health risks and spousal encouragement are likely complex as these findings suggest, and may change over time [31]. Future research seeking to understand factors that promote the development of new encouragement pathways between spouses, in particular varied motivations between wives and husbands, can inform development of interventions that more effectively reach both members of the spouse couple.

The diminished effect of husbands’ risk feedback on wives’ physical activity at 10-month follow-up is consistent with a challenge common to many efforts aiming to promote long-term behavior change. Even those interventions that are successful in creating an initial response often have greater difficulty sustaining behavior change over time. However, many physical activity interventions to date have targeted individuals or have been tailored to individual characteristics, including current health behaviors or cognitive factors [53, 54]. Our findings highlight the importance of interpersonal relationships and the roles that individuals inhabit within the context of their family systems. Indeed, the potential to activate social influence processes through genomic information may offer continued reinforcement of the behavior over time thereby enhancing the salience and long-term effectiveness of the intervention.

At 10-month follow-up, wives who were new encouragers of their husband’s health behavior over the course of the intervention were engaging in higher levels of physical activity. This subgroup of wives demonstrated an investment in their spouses’ health, and thus may have been motivated to engage in physical activity to support and encourage their husbands. Alternatively, wives may have been modeling this behavior for their husbands and/or their children, whose risk is directly affected by their parents’ risk. Although we are not positioned to disentangle these possible explanations, we see a similar pattern among husbands at 10-month follow-up with those who had begun encouraging their wives over the course of the intervention and who were partnered to women with increased heart disease risk feedback engaging in the highest levels of physical activity. Our results suggest that interventions may more successfully extend the effects of a behavior change when they consider how couple members function in the context of their relationship, as well as the dynamic dyadic processes that unfold over time. However, future research is warranted to explore the mechanisms underlying these patterns of behavior change.

Although this study is one of the first to explore the impact of family history-based risk information on spouse-couples, there are several opportunities to strengthen subsequent studies. First, the current study population is ethnically and geographically homogenous. Without a comparison group, we are unable to identify whether there are cultural influences contributing to our findings, or whether the current results are generalizable to a more heterogeneous population. However, there are few studies of this type that have focused on Mexican-heritage populations, a largely understudied population at increased risk of heart disease and associated comorbidities. Future studies may consider other cultural contexts to assess whether these results generalize to other communities. Second, our measures of physical activity were collected through participant self-report at each measurement point. Although the measures were taken from the CDC’s Family Healthware tool, employing instruments to measure physical activity objectively, such as pedometers, would be a valuable strategy for validating the current results. As well, our measure of encouragement relied on a single item, because it was drawn from an ego-centered social network assessment. Future work that includes more robust measures of the social interactions related to encouragement will advance this work. Finally, we are cautious not to overextend our results, due to our modest sample size. Limited numbers receiving each type of risk feedback lead us to consider these results as hypothesis-generating rather than conclusive. We had slightly greater difficulty retaining husbands in the study compared to wives, suggesting that future studies may need to take extra measures to encourage men to participate over the full course of the study.

In addition to informing clinical utility of family health history, these findings have important implications for broader community settings. Current clinical and public health practice may neglect crucial opportunities to mobilize patients to engage in risk-reducing behaviors and encourage these behaviors among family members. The current results suggest the potential to transform motivation to change behavior through risk of close others, which may require a shift from a patient- or individual-centered approach to a family-centered approach. Thus, there may be greater opportunities for leveraging behavior change if clinicians and public health practitioners consider the relationship context in which this information is delivered.

Compliance With Ethical Standards

Conflict of Interest: The authors have no conflicts of interest.

Disclaimer: The authors have had full control over the primary data and we agree to allow the journal to review the data if requested. The findings reported in the current manuscript have not been previously published and the manuscript is not submitted for publication elsewhere. However, there have been several publications generated from the project that is the basis of the current paper [14, 19, 30, 37, 46, 55-60].

Ethical Approval: This research was conducted per the Ethical Standards for the Protection of Human Participants in Research following IRB review and approval at the National Human Genome Research Institute and the University of Texas, MD Anderson Cancer Center. No animals were involved in this research and as such the welfare of animals was not compromised during the implementation of this project.

Informed Consent: All participants provided written informed consent to participate in this study during baseline assessment and verbal consent at each follow-up assessment.