University of Southern California Dissertations and Theses

The effects of health-related spousal influence on couples coping with chronic heart failure: an application of the actor-partner interdependence model

(USC Thesis Other)

The effects of health-related spousal influence on couples coping with chronic heart failure: an application of the actor-partner interdependence model

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1
!

The Effects of Health-Related Spousal Influence on Couples Coping with Chronic Heart
Failure: An Application of the Actor-Partner Interdependence Model
Jennifer L. Geren
University of Southern California

"!
Acknowledgements
I gratefully acknowledge my advisor, Bob G. Knight, for his guidance, his
encouragement, and his resolute dedication to getting me through this program. I am indebted to
Michael Rohrbaugh and Varda Shoham who generously shared their data with me and made this
dissertation possible. I would also like to thank Brian Baucom, Gayla Margolin, Mara Mather,
and Florence Clark for serving on my dissertation committee. My sincerest appreciation goes to
my family and friends, whose unconditional love and support sustained me through this
collective journey.
#!
Table of Contents
Acknowledgements 2
List of Tables 4
List of Figures 5
Abstract 7
Introduction 9
Methods 27
Results 40
Discussion 54
References 71
Appendices 114
$!
List of Tables

Table 1: Descriptive Summary of Patients 84

Table 2: Within-Partner Pearson Correlations for Study Variables 85

Table 3: Between-partner correlations for study variables 86

Table 4: Comparison between patients and spouses on demographic 87
and study variables

Table 5: APIM of the effect of spousal influence attempts on treatment 88
compliance at follow-up and with moderation by baseline marital quality

Table 6: APIM of the effect of spousal influence attempts on treatment 89
compliance at follow-up and with moderation by closeness variability

Table 7: APIM of the effect of treatment compliance at follow-up on 90
CHF symptom level at follow-up

Table 8: APIM of the effect of spousal influence attempts on general physical 91
health at follow-up and with moderation by baseline marital quality

Table 9: APIM of the effect of spousal influence attempts on general physical 92
health at follow-up and with moderation by closeness variability

Table 10: APIM of the effect of spousal influence attempts on general mental 93
health at follow-up and with moderation by baseline marital quality

Table 11: APIM of the effect of spousal influence attempts on general mental 94
health at follow-up and with moderation by closeness variability

Table 12: APIM of the effect of daily influence attempts on next day marital 95
closeness and with moderation by ratio of positive to negative influence
%!

List of Figures
Figure 1: The Actor-Partner Interdependence Model (APIM) 96

Figure 2: Conceptual APIM model of the indirect effect of spousal influence 97
attempts on CHF symptoms via patient compliance

Figure 3: Conceptual APIM model of the effect of spousal influence attempts on 98
patient and spouse general physical health

Figure 4: Conceptual APIM model of the effect of spousal influence attempts on 99
patient and spouse mental health

Figure 5: Conceptual APIM model with moderation of the indirect effect of 100
spousal influence attempts on CHF symptoms via patient compliance
by marital quality and closeness variability

Figure 6: Conceptual APIM model with moderation of the effect of spousal 101
influence attempts on patient and spouse general physical health
by marital quality and closeness variability

Figure 7: Conceptual APIM model with moderation of the effect of spousal 102
influence attempts on patient and spouse mental health by marital
quality and closeness variability

Figure 8: Conceptual model of repeated measures APIM of the effect of 103
providing and receiving influence attempts on next day marital
closeness

Figure 9: Conceptual model of repeated measures APIM with moderation of 104
the effect of providing and receiving influence attempts on next day
marital closeness by relative positivity to negativity of the influence

Figure 10: APIM of spousal influence attempts on patient compliance at 105
follow-up

Figure 11: Moderation of the effect of patient report of positive influence on 106
patient report of follow-up compliance by patient closeness variability

Figure 12: APIM of patient compliance on CHF symptoms 107

Figure 13: APIM of the effect of positive and negative influence attempts on 108
patient and spouse general physical health at follow-up

&!
Figure 14: Moderation of the effect of spouse report of positive influence on 109
spouse general physical health at follow-up by patient closeness variability

Figure 15: APIM of the effect of positive and negative influence attempts on 110
patient and spouse mental health at follow-up

Figure 16: Moderation of the effect of spouse report of positive influence on 111
patient mental health at follow-up by spouse closeness variability

Figure 17: Repeated measures APIM of the effect of patient report of daily 112
influence received on patient report of next day marital closeness

Figure 18: Moderation of the effect of patient report of daily influence on patient 113
report of daily marital closeness by the ratio of positivity to negativity
of the influence received
'!
Abstract
Coping with chronic illness is increasingly being viewed as a relational process rather than an
individual-based phenomenon. Although spouses serve as each others’ primary source of support
and assistance, relatively little is known about how spousal involvement in illness management
affects outcomes for patients and spouses. The aims of this project were to (1) identify the
distinct effects of two types of health-related spousal influence attempts (positive and negative)
on patient compliance and subsequent chronic heart failure (CHF) symptoms, (2) examine the
potential broader consequences of these influence attempts for patients’ and spouses’ general
physical and mental health, (3) investigate how the marital context (specifically marital quality
and variability in closeness) influences the above associations, and (4) explore the day-to-day
effects of providing and receiving influence attempts on feelings of marital closeness using a
daily diary methodology. Analyses for Aims 1 through 3 were conducted using an actor-partner-
interdependence model (APIM) with a short-term longitudinal dataset (n = 60 couples). Data
revealed that negative spousal influence attempts at baseline were associated with lower
treatment compliance for patients six months later and in turn greater CHF symptom levels,
whereas positive influence attempts had the opposite effect when feelings of marital closeness
were stable. Spouses exhibited a negative association between positive influence attempts and
physical health, but no association between influence attempts and physical health was found for
patients. Negative influence attempts were negatively associated with patients’ mental health six
months later; however, no association was found for spouses’ mental health. Lastly, in the daily
diary study (n = 26 couples) using lagged-time analyses in a repeated measures APIM (Aim 4) , I
found that receiving influence from one’s spouse served to increase patients’ feelings of marital
closeness, particularly when the influence was more positive than negative. However, providing
influence had no effect on spouses’ feelings of marital closeness. These findings demonstrate
(!
that spouses’ involvement in illness management is indeed consequential for patients with CHF,
with distinct consequences of positive and negative influence attempts. Benefits of positive
influence attempts were dependent on the marital context, whereas costs of negative influence
remained regardless of the marital context. Efforts to influence compliance also have effects on
feelings of marital closeness. Findings suggest future directions for research and reveal potential
areas for intervention.
)!
The Effects of Health-Related Spousal Influence on Couples Coping with Chronic Heart
Failure: An Application of the Actor-Partner Interdependence Model
Coping with chronic illness is increasingly being viewed as a relational process rather
than an individual-based phenomenon (e.g., Berg & Upchurch, 2007; Coyne & Fiske, 1992;
Revenson, 2003). Although spouses serve as each other’s primary source of support and
assistance, relatively little is known about how spousal involvement in illness management
affects outcomes for patients and spouses. Berg and Upchurch’s (2007) developmental-
contextual model of couples coping with chronic illness conceptualizes dyadic coping along a
continuum of spousal involvement ranging from uninvolvement of the spouse to
overinvolvement. Support behaviors of the involved spouse include emotional and instrumental
support, which can encompass joint problem solving, collaboration, influence and control.
Spousal involvement targeted at improving treatment adherence or assisting in illness
management is critically important for chronic health conditions like chronic heart failure (CHF).
However, little attention has been directed towards instrumental aspects of coping with illness
management in CHF.
Failure to adhere to the treatment regimen in CHF can result in worsening of symptoms
and even death of the patient (Coyne & Fiske, 1992). Furthermore, spouses are frequently
engaged in helping the patient with illness management, making it a salient part of the couples’
coping efforts (Franks, Wendorf, Gonzalez, & Ketterer, 2004; Gallant, Spitze, & Prohaska,
2007). Spousal involvement in illness management is likely to involve attempts to influence and
motivate the patient to follow the treatment regimen. Given the significance of treatment
adherence for patient survival and the complex nature of illness management in CHF,
understanding how couples cope with these more instrumental tasks of disease management is
*+!
critically important.
Research and theory suggest that the nature of these influence attempts is likely to have
important consequences for the patients as recipients of these attempts and for the caregiving
spouses as providers of the influence. In particular, developmental theories of marriage suggest
that the behaviors in which couples engage when coping with stressors are consequential not
only for adjustment to the stressor itself, but also for future marital quality (Karney & Bradbury,
1995; Cutrona, Russell, & Gardner, 2005).
The current project examined the specific types of influence tactics used by spouses in
helping CHF patients with illness management and the consequences of those influence attempts
for both patient and spouse physical and psychological health. Additionally, I examined the
effect of provision and receipt of influence on the marital relationship itself. Below, I review the
literature on couples coping, health-related spousal support and chronic heart failure, which was
used as the basis for the study aims and design.
Couples Coping with CHF
Chronic heart failure (CHF) is a life-threatening condition in which maintenance of
normal blood circulation is precluded by impairment of the heart muscle (McCall, 1995). CHF
frequently represents the end stage of many forms of cardiovascular disease including coronary
artery disease, hypertension, myocardial infarction and valvular diseases (Hunt et al., 2001).
CHF is primarily a disease of the elderly affecting 6-10% of individuals over age 65 and 2% of
the United States population as a whole (Lloyd-Jones et al., 2009). The prevalence of CHF is
increasing due to the increasing age of the population as well as to successes in the treatment of
acute myocardial infarction. Despite advances in treatment, the prognosis for patients diagnosed
with CHF remains poor, with mortality rates of 20% at the end of the first year post diagnosis
**!
and up to 50% at 5 years post diagnosis. Repeated hospitalizations are common in CHF,
contributing to its high economic burden. In fact, CHF accounts for over one third of all
Medicare spending, more than any other diagnosis (Setoguchi & Stevenson, 2009).
In addition to the burden placed on the healthcare system, CHF presents patients and their
spouses with many challenges. Functional disability stemming from CHF includes shortness of
breath during activity and sleep, edema in the ankles and feet, and general weakness and fatigue.
Management of the illness is complex and imposes strict demands requiring multidrug regimens,
significant changes in diet and exercise, careful monitoring of symptoms, and other lifestyle
modifications. Research and theory has increasingly suggested that coping with chronic illness
can be better understood as a dyadic, rather than individualistic process (e.g., Berg & Upchurch,
2007; Coyne & Fiske, 1992; Revenson, 2003). Spouses serve as each other’s primary source of
support and assistance (Revenson, 1994), and the prominence of the support provided by the
marital relationship is likely to increase in older adulthood as social networks begin to narrow
(e.g., Carstensen, 1993; Lang & Carstensen, 2002; Pinquart, 2003) and the frequency and
severity of chronic illness increases (Siegler, Bosworth, & Poon, 2003).
Examinations of couples coping with myocardial infarction, a risk factor for CHF, have
focused on aspects of emotional support (e.g. protective buffering) and collaboration (e.g. active
engagement; Coyne & Smith, 1991, 1994). Investigations of CHF patients and spouses have
highlighted the significance of psychosocial factors like psychological distress and contextual
variables like the quality of the marital relationship for predicting patient health outcomes
(Coyne et al., 2001; Rohrbaugh et al., 2002; Rohrbaugh, Shoham, Cleary, Berman, & Ewy,
2009). However, instrumental aspects of coping with illness management and subsequent effects
on adjustment to CHF have yet to be examined.
*"!
The Role of Spouses in Illness Management
Marriage has been repeatedly shown to be associated with lower mortality rates and
better mental and physical health (e.g., Johnson, Backlund, Sorlie, & Loveless, 2000; Manzoli,
Villari, Pirone, & Boccia, 2007). The health benefits of marriage are thought to occur in part
through social support and control processes that serve to reduce distress, increase engagement in
healthy behaviors, and discourage unhealthy behavior (Burman & Margolin, 1992; Kiecolt-
Glaser & Newton, 2001; Umberson, 1992). The association between marriage and health is
particularly relevant in the context of chronic illness given that spouses are the primary providers
of health-related support and control. As such, considerable research attention has been directed
at understanding the impact of social support and control on patient health outcomes.
Researchers have suggested that the mechanism by which social support and control processes
influence health outcomes is through the effect on patient adherence (DiMatteo, 2004; Dunbar-
Jacob, Schlenk, Baum, Revenson, & Singler, 2001).
Although the specific nature of patients’ adherence to a treatment regimen varies
somewhat from illness to illness, it frequently involves patients taking medications exactly as
prescribed, reducing health compromising behaviors like smoking or drinking, following diet
restrictions and controlling weight, and monitoring and reporting symptoms to medical
personnel. DiMatteo (2004) conducted a meta-analysis examining the effect of various types of
social support on patient adherence and found significant quantitative evidence that social
support has substantial effects on patient adherence. This analysis explored differences in the
effect of structural support (e.g., marital status, living alone or with others), emotional support
(e.g., expressions of caring, empathy) and instrumental support (e.g., provision of assistance,
reminders, support for specific behaviors) on patient adherence. Instrumental support yielded, on
*#!
average, significantly higher effects than emotional support or structural support, with odds of
adherence being 3.6 times higher among those who receive instrumental support relative to those
who do not compared to 1.35 and 1.27 times higher for emotional and structural support
respectively. However, closer examination of the 29 studies investigating the effect of
instrumental support on adherence shows correlations ranging from -.22 to .75. The variation in
these findings suggests that it may be important to assess the type of assistance provided to the
patient, as not all instrumental support is beneficial. Several more recent investigations shed light
on some of the specific types of behaviors spouses engage in that are targeted at illness
management or treatment adherence.
Social control is a concept that has been theorized to promote health in two ways: (a)
indirectly via the sense of obligation toward others (i.e. spouse) and responsibility to take care of
oneself and (b) directly via the interactions between partners that regulate, influence, or constrain
health behavior (Umberson, 1987; Lewis & Rook, 1999). The current project, given its focus on
specific spousal behaviors and consequences of those behaviors for each partner, is examining
this latter, direct aspect of social control.
Direct social control has been operationalized in a variety of ways. Definitions include
spousal attempts to assist, motivate, urge, advise and influence the patient to practice good health
behavior or adhere to a particular treatment regimen. Although research and theory on social
support and social control have been somewhat distinct within the literature, Helgeson and
colleagues (2004) noted that interactions defined as social control are likely to have been
captured in studies of instrumental support (e.g., providing concrete assistance) and emotional
support (e.g., encouraging specific health behavior). This overlap in constructs suggests that
research may benefit from the use of more clearly defined spouse behaviors.
*$!
Positive and Negative Influence Attempts. Research has recently highlighted the
overlapping nature of support and control and has begun to focus on identifying specific types of
spousal support and control and examining the differing effects of these spousal behaviors on
patients’ health and treatment adherence. For example, spouses’ attempts to aid and reinforce the
patients’ efforts to implement needed changes in health behavior through their use of tactics like
complimenting, motivating and encouraging have been distinguished from attempts to induce
needed changes through the use of tactics like criticizing, nagging, and warning the patient of
negative consequences of non-compliance (Fekete, Stephens, Druley, & Greene, 2006; Franks et
al., 2006; Helgeson, Novak, Lepore, & Eton, 2004; Stephens et al., 2009; Stephens, Rook,
Franks, Khan, & Iida, 2010; Tucker, Orlando, Elliot, & Klein, 2006).
Recent work has revealed distinct effects on patient health outcomes for these different
types spousal behaviors. Among patients trying to adhere to a diabetic diet, cross-sectional
analyses revealed that positive spousal influence behaviors like providing encouragement were
positively associated with patient adherence, whereas negative behaviors like warning the patient
of negative consequences of non-adherence were negatively associated with dietary adherence
(Stephens et al., 2010). In a prospective analysis of patients participating in a cardiac
rehabilitation program, spousal influence attempts that included assisting, validating and
encouraging were not associated with patient health behavior change six months later. In
contrast, spousal influence attempts that included reminding, warning and controlling were
negatively associated with health behavior change prospectively (Franks et al., 2006). Lastly,
among a sample of prostate cancer patients, greater urging by the spouse for the patient to adhere
to the post-surgical or post-radiation medical recommendations was associated with worse health
behavior both cross-sectionally and prospectively (Helgeson et al., 2004).
*%!
The deleterious effects of some types of the spousal influence tactics stand in stark
contrast to the original hypotheses set forth by social control theory, which suggested that social
control should facilitate engagement in healthy behaviors and discourage engagement in
unhealthy behaviors (Lewis & Rook, 1999; Umberson, 1987). However, these original
hypotheses were examined in relatively young, healthy individuals and couples. The current
research suggests that when coping with chronic and debilitating health conditions, the nature of
spousal influence is consequential for patient health outcomes.
The Role of Spousal Influence Attempts on Broader Outcomes: Patient’s Mental Health
and Spouse’s Physical and Mental Health
Although the intended effect of spousal support and control may be to impact the
patient’s physical health, research shows that these spousal behaviors also impact the patients’
mental health. For example, osteoarthritis patients who had undergone knee replacement surgery
demonstrated increased positive emotion in response to their spouses’ attempts to encourage
treatment adherence (Stephens et al., 2009). In contrast, influence attempts that included
reminding, warning, controlling, and criticizing were negatively associated with patient
psychological health both concurrently and prospectively (Fekete et al., 2006; Franks et al.,
2006). Although psychological distress among CHF patients has been shown to be related to
illness severity, it can also be a risk factor for future health complications and death (Carney,
Freedland, Rich, & Jaffe, 1995). Therefore, examining the impact of spouses’ support and
control behaviors on patients’ mental health is important for understanding patient adjustment to
CHF.
The couples-health literature has greatly enriched our understanding of coping with
chronic illness by focusing on coping from an interpersonal perspective. Although the patient has
*&!
understandably been the person of interest when it comes to predicting health outcomes, the
impact of dyadic coping efforts on spousal caregivers should not be overlooked (Coyne & Fiske,
1992; Revenson, 1994). A substantial literature suggests that caregivers are at risk for poorer
physical and mental health outcomes compared to their non-caregiving counterparts (see reviews
by Bookwala, Yee, & Schulz, 2000; Lavela & Ather, 2010; Pinquart & Sorensen, 2003, 2007),
with investigations focusing on the role of appraisals of caregiver burden, coping styles (e.g.,
avoidant versus approach oriented coping), availability of social support for the caregiver, and
variables specific to the care recipient (e.g., illness type, degree of patient frailty). However, the
specific behaviors spouses engage in when trying to aid and influence the patient to follow
treatment guidelines and implement necessary lifestyle changes may also have an impact on the
spouses’ physical and psychological well being. For example, spouses who attempt to influence
the patient by joining them in following a healthy diet and establishing an exercise routine may
reap physical health benefits from this approach. In contrast, spouses who argue, complain, or
criticize the patient for not following the treatment recommendations may suffer psychologically
over time.
Only one study to date has examined the consequences for spouses of engaging in health-
related control. Among a group of patients with type II diabetes, spouses’ use of social control to
influence the patients’ dietary adherence was associated with greater spouse burden (August,
Rook, Stephens, & Franks, 2011). Although the study was cross-sectional in nature, this
preliminary look suggests that attempts to influence patients’ health behavior are consequential
for spouses. Moreover, adjustment of spousal caregivers is consequential for future coping
efforts and thus a critically important aspect to examine.
*'!
Summary. Although spousal involvement in illness management with CHF patients
remains largely unexplored, the broader literature on couples coping with chronic illness
suggests that spousal involvement in illness management and treatment adherence is
consequential for patient physical and mental health outcomes. The current literature
characterizes spousal involvement in a variety of ways, most often labeling it as spousal support
or spousal control. This is problematic because the research is not consistent in how it defines
and measures support and control, thus muddying the notion imparted by the label itself that one
is better than the other (i.e. support better than control). The current study examined spouses’
attempts to influence the patient to comply with the treatment regimen, thus I will refer to these
behaviors as health-related spousal influence attempts.
Furthermore, spousal caregivers may also be affected by the nature of their influence
attempts regarding disease management. Given spouses’ prominent role in helping the patient
with demands of CHF disease management and the importance of treatment adherence for
patient survival, a fine-grained assessment of the nature of health-related spousal influence
attempts and the consequences of these behaviors for both patients and spouses is needed.
The Role of Marital Quality for Couples Coping with CHF
Current research on couples coping with CHF has taken a contextual, interpersonal
perspective in trying to understand patient adjustment in CHF (Rohrbaugh et al., 2009). One
important construct is marital quality. Although marital quality has been operationalized in many
different ways in the literature, definitions frequently represent marital quality using a
multifaceted approach. High quality marriages are thought to reflect marital happiness and
satisfaction, adaptive communication patterns, and high positive and low negative affectivity
*(!
during marital interaction (e.g., Bookwala, 2005; Coyne et al., 2001). Research in the health
literature has begun to examine marital quality as a factor in patient adjustment to chronic illness.
Among a sample of 189 CHF patients and their spouses, marital quality was shown to
predict patient survival over the next eight years independently of baseline illness severity
(Rohrbaugh, Shoham, & Coyne, 2006). Other work with this sample of CHF patients and
spouses examined patient self-efficacy and spouse confidence in the patient’s ability to follow
the treatment regimen. Results showed that spouse confidence predicted patient health change,
however spouse confidence emerged as a correlate of marital quality and was no longer a
significant predictor of patient health when marital quality was controlled (Rohrbaugh, et al.,
2004). Marital quality has been similarly implicated in adjustment to chronic illness in the health
literature more broadly, with relationship quality predicting outcomes for diabetes, colorectal
cancer, and chronic obstructive pulmonary disease patients (Ashmore, Emery, Hauck, &
MacIntyre, 2005; Hagedoorn et al., 2011; Trief, Ploutz-Synder, Dee Britton, & Weinstock, 2004;
Trief et al., 2006; Unger, Jacobs, & Cannon, 1996).
The notion that marital quality shapes the nature of coping efforts and in turn, health
outcomes has been posited by many (e.g., Coyne & Fiske, 1992; Karney & Bradbury, 1995;
Revenson, 1994, 2003), yet robust examinations of this effect in the health literature are rather
limited. One study of spousal caregivers of dementia patients found that pre-illness relationship
quality influenced the quality of the care provided. Using retrospective reports of marital quality,
caregivers who reported greater pre-illness relationship quality were less likely to engage in
potentially harmful behavior (e.g. yelling, insulting, swearing, handling roughly) compared to
caregivers who reported lower pre-illness relationship quality (Williamson & Shaffer, 2001).
*)!
Other research has suggested that marital quality plays a moderating role in the
association between spousal influence and health behavior. In a study of health-related social
control in older adults, influence attempts by spouses that included a mixture of supportive and
controlling behaviors targeted at general health behavior elicited differing affective and
behavioral responses depending on the quality of the marital relationship. Older adults who
experienced more spousal influence in the context of high relationship satisfaction reported less
negative affect and less frequent hiding of unhealthy behaviors. In contrast, greater spousal
influence in the context of low relationship satisfaction was associated with more negative affect
and more frequent hiding of unhealthy behavior (Tucker, 2002). This research suggests the
possibility that although some influence attempts may be annoying or counterproductive for the
recipient, when the marriage is viewed positively these spousal behaviors are perceived as
reflective of caring and therefore less detrimental to mood and more beneficial to health. In
contrast, when the marriage is negatively appraised, these behaviors are likely to be viewed more
negatively and thus result in negative outcomes.
Static Versus Dynamic Views of Marital Quality. Taken together, research on couples
coping with CHF, and other health conditions more broadly, suggests that the marital
relationship serves as a contextual backdrop. This perspective emphasizes that historical marital
quality is important for understanding adaptation to chronic illness. However, a historical view
suggests a static quality to marital functioning. The literature on continuity and change in marital
quality over the life span suggests that marital quality in older long-term couples remains stable
and potentially increases in later life (Anderson, Russell, & Schumm, 1983; Guilford &
Bengtson, 1979; Orbuch, House, Mero, & Webster, 1996; Weishaus & Field, 1988). This
positive view of marriages in older adulthood is based on cross-sectional and longitudinal studies
"+!
that have primarily focused on mean-level marital satisfaction across different ages or time,
which overlooks potential changes occurring at the individual- or couple-level. Studies that
demonstrate general mean-level stability or improvement are often interpreted as stability for all
of the individuals or couples in that sample. This approach neglects the possibility that some may
remain stable while others may fluctuate or change over time, and precludes our understanding
of factors that influence stability or variability within the marriage.
Intraindividual variability describes within-person variability over time and refers to
short-term fluctuations (Nesselroade, 1991). A recent investigation of intraindividual variability
in relationship satisfaction revealed that married women whose relationship satisfaction
fluctuated more widely from week to week tended to have lower mean-level satisfaction and
higher levels of depressive symptoms (Whitton & Whisman, 2010). Although mean-level
relationship satisfaction is a well-established predictor of depression, the study found that
variability was a significant predictor of prospective depression over and above mean-level
satisfaction. However, variability in depression did not predict relationship satisfaction
prospectively. This study highlights the utility of assessing short-term fluctuations in marital
quality for understanding mental health outcomes beyond what can be garnered from more
traditional mean-level assessments.
The prominence of spousal involvement in the practical and instrumental tasks of disease
management in CHF presents couples with interactions that could result in changes to each
spouse’s feelings about the relationship. Intraindividual variability in marital quality over shorter
time scales may be a better reflection of the marital context in which coping is occurring. As
described earlier, available research suggests that the quality of the marriage is consequential for
how couples manage and ultimately adapt to the illness. However, the mechanism by which the
"*!
marital context impacts CHF patients’ health has not been examined directly. Existing studies of
health-related spousal support in the context of chronic illness often include marital quality as a
covariate in the analysis, limiting our understanding of the potential moderating effect of marital
quality on illness management. The current project examined the moderating role of marital
quality on the associations between spousal influence attempts and health outcomes for both
patients and spouses using both a traditional, global measure of marital quality as a well as a
measure of day-to-day variability in feelings of marital closeness.
Marital Quality as an Outcome: Day-to-Day Impacts of Chronic Illness Management
A dyadic view of coping with chronic illness posits that, for couples coping with this
stressor, relationship quality is influential in adjustment to the illness context. Although research
in this area is somewhat limited, the literature reviewed in the preceding sections attests to the
importance of considering relationship quality when studying couples coping with chronic
illness. Although the primary goal of the current project is to examine how marital quality
impacts coping processes and outcomes, consideration of how the marital relationship is
impacted by the illness context is also warranted. Developmental theories of marriage have
posited a transactional relationship between coping processes and marital quality over time
(Karney & Bradbury, 1995; Cutrona, Russell, & Gardner, 2005). These theories suggest that how
couples cope with stressors is consequential not only for adjustment to the stressor itself, but also
for future marital quality.!Although many theorists agree that chronic illness is one such stressor
that can affect marital quality, research examining the affect of specific coping behaviors on
marital quality is scarce, as marital quality is seldom treated as a longitudinal outcome for
couples’ coping with chronic illness.
""!
Most existing research on couples coping with chronic illness is cross-sectional in nature.
These investigations provide an understanding of associations between coping behaviors and
health outcomes, however, this approach does not capture the interpersonal process that couples
are engaging in around disease management (Laurenceau & Bolger, 2005). Although health-
related spousal influence may be well-intended and motivated by a desire to improve patient
health, it is reasonable to consider that these behaviors may affect the marriage over time.
Repeated measures offered by daily diary methodology allow for the examination of important
questions pertaining to spouses’ day-to-day experience of illness management and the effect of
this process on relational well being.
Recent investigations within the marital literature have utilized daily diary methodology
in an effort to understand this dynamic process, and have revealed that coping processes are
indeed interwoven with relationship-level variables like marital closeness. For example,
examinations of support provision and receipt among younger couples managing acute stressors
have shown that support provision is beneficial for providers’ and recipients’ relationship
closeness (Gleason, Iida, Shrout, & Bolger, 2008). Although these findings are encouraging, the
effects of support provision and receipt may differ when the nature of the support enacted by the
spouse involves attempts to influence the patient to follow stringent medical recommendations.
In an effort to provide a more nuanced view of couples’ day-to-day experience of illness
management in CHF, the current project examined the effect of provision and receipt of health-
related spousal influence on feelings of marital closeness.
A Dyadic Perspective on Couples Coping with CHF
Although current research has increasingly shifted toward the relational processes of
coping with chronic illness by examining spousal involvement in the process, the measurement
"#!
and statistical techniques used have not truly captured the relational phenomena, but rather have
remained at the individual level. The Actor-Partner Interdependence Model (APIM; see Figure
1) is an approach to analyzing dyadic data that permits investigators to estimate (simultaneously
and independently) the effect that an individual’s score on a predictor variable has on their own
score on an outcome variable (know as the actor effect, path a) and on their partner’s score on an
outcome variable (known as the partner effect, path p) (Cook & Kenny, 2005; Kenny, Kashy, &
Cook, 2006). Very often researchers adopt an actor-oriented perspective and assume that a
person’s outcomes are a function of that person’s characteristics only. Also common in the
health-related support and control literature is the partner-oriented model using spouses’ reports
of support behaviors to predict patient outcomes. Although the actor-oriented and partner-
oriented models may accurately reflect the phenomenon, these assumptions must be tested using
data collected from both partners and analyzed using a technique that will estimate both
individual and dyadic factors.
Another important consideration when examining topics like support, compliance, and
health outcomes is - whose report should be used? There is often debate regarding who provides
the most accurate report. Are patients’ reports of their own compliance inherently biased? Or
does their first-person account provide the most accurate reflection of compliance levels?
Likewise, are spouses likely to overestimate their supportive behaviors and underestimate their
more critical ones? Past work has often used the average score of partners’ reports in an effort to
adjust for potential biases, however this approach obscures the reality that patients and spouses
have distinct perceptions of these behaviors. Additionally, taking the sum or average of two
partners scores could result in a couple in which the patient and spouse report moderate levels of
support provided and received looking identical to a couple in which the patient reports receiving
"$!
little support from their spouse and the spouse reports providing a lot of support (Christensen &
Arrington, 1987). Most would agree that these two couples are quite different, however creating
“dyad” scores may well result in a loss of information. The APIM allows for the distinct
perspectives of dyad members to be examined directly and thus can help clarify some of the
questions in the couples-health literature regarding whose report to use.
The current project utilized the APIM to examine the study aims and hypotheses outlined
below. The APIM could be used in this study because all study variables were collected from
both partners, a necessity for dyadic data analysis.
The Current Study
The first goal of this project (Aim 1) was to examine the associations between spouses’
efforts to increase patients’ treatment compliance and patients’ subsequent CHF symptom level
(see Figure 2 for conceptual model.) Two types of health-related spousal influence attempts were
examined: positive influence and negative influence. Positive influence attempts included
praising the patient for following the treatment regimen and joining them in their efforts to elicit
the desired health behavior. Negative influence included arguing, complaining, or criticizing the
patient for not following the regimen and applying negative pressure like warning the patient of
negative consequences or insisting they feel worse than they will admit. I predicted that positive
influence attempts would be related to greater compliance with the treatment regimen and in turn
lower CHF symptom level for patients at 6-month follow-up (Hypothesis 1a). I predicted that
negative influence attempts would be related to lower compliance with the treatment regimen
and in turn higher CHF symptom level for patients at follow-up (Hypothesis 1b).
Aim 2 was to examine the effects of spousal influence attempts for patients’ physical and
mental health more broadly (see Figures 3 and 4 for conceptual models.) I predicted that positive
"%!
influence attempts would be associated with better physical health and better mental health at
follow-up for patients (Hypothesis 2a). I expected that negative influence attempts would be
associated with worse physical health and worse mental health at follow-up for patients
(Hypothesis 2b). Additionally under this aim, I examined the effects of engaging in health-
related influence attempts for spouses. I predicted that positive influence attempts would be
associated with better physical health and better mental health at follow-up for spouses
(Hypothesis 2c), and that negative influence attempts would be associated with worse physical
health and worse mental health at follow-up spouses (Hypothesis 2d).
For Aim 3, which examined how marital quality influences the associations between
influence attempts, compliance and health, marital quality was conceptualized in two ways. First,
I used a composite measure of marital quality that reflected marital satisfaction as well as marital
communication patterns. Second, in an effort to move beyond static measures of marital quality,
I used a measure of the day-to-day variability in feelings of marital closeness. Greater variability
in closeness was thought to capture volatility or instability in the relationship, and thus was
important to examine in addition to and separately from global measures of marital quality. The
following hypotheses represent predictions about the specific context in which the associations
predicted under Aims 1 and 2 were likely to occur. (See Figure 5, 6, and 7 for conceptual
models.) I predicted that the indirect effect of positive influence attempts on CHF symptoms via
patient compliance would be most evident when marital quality was high and/or closeness
variability was low, and least evident when marital quality was low and/or closeness variability
was high (Hypothesis 3a). I also expected that the indirect effect of negative influence attempts
on CHF symptoms via patient compliance would be most evident when marital quality was low
and/or closeness variability was high, and least evident when marital quality was high and/or
"&!
closeness variability was low (Hypothesis 3b). I predicted that the effect of positive influence
attempts on patients’ and spouses’ general physical and mental health would be most evident
when marital quality was high and/or closeness variability was low, and least evident when
marital quality was low and/or closeness variability was high (Hypothesis 3c; see Figures 6 and
7). I expected that the effect of negative influence attempts on patients’ and spouses’ general
physical and mental health would be most evident when marital quality was low and/or closeness
variability was high, and least evident when marital quality was high and/or closeness variability
was low (Hypothesis 3d).
The fourth and final aim of this study was to examine marital quality as an outcome.
Specifically, I used a daily diary methodology to test the effects of providing and receiving
health-related spousal influence on patients’ and spouses’ day-to-day feelings of marital
closeness based on daily reports of influence and closeness (see Figure 8 for a conceptual
model). I predicted an association between reports of influence attempts and marital closeness
reported on the next day for patients as receivers of the influence and spouses as providers of the
influence (Hypothesis 4a). However given the importance of the nature of the influence attempts,
the association between influence and next-day closeness (lagged effect) would be moderated by
the relative positivity to negativity of the influence attempts such that marital closeness would
increase in response to influence when the influence was more positive than negative and
decrease in response to influence when the influence was more negative than positive
(Hypothesis 4b; see Figure 9 for a conceptual model).

"'!
Methods
Participants
Participants for this study come from the Arizona Family Heart Project, a short-term
longitudinal study examining couples coping with CHF in one spouse and examining predictors
of patient adjustment (see Rohrbaugh, Mehl, Shoham, Reilly, & Ewy, 2008; Rohrbaugh et al.,
2009). The sample of 60 patients with chronic heart failure (CHF), 43 men and 17 women, and
their opposite-sex spouses were recruited from University of Arizona cardiology clinics. All
patients carried a confirmed CHF diagnosis and had a left ventricular ejection fraction ! 40 (M =
29.1, SD = 8.7), usually documented by echocardiogram during the previous six months. Patients
had been diagnosed with CHF on average 4.8 years earlier (SD = 5.1) and with some type of
heart problem 11.5 years earlier (SD = 9.8). Forty-two percent of the patients had experienced
myocardial infarction, 32% were diabetic, and 25% had hypertension. Although 50% of the
patients had been hospitalized and 42% had made emergency room visits in the previous six
months, all participants were outpatients at the time of the initial assessment.
Mean ages of patients and spouses were 67 (SD = 11.7) and 65.6 (SD = 10.7)
respectively, and couples had been married an average of 34.8 (SD = 16.7) years. The patient
sample was predominately non-Hispanic White (85%), well-educated (40% college graduates),
and at the 65
th
percentile of national household income. See Table 1 for a summary of the patient
characteristics.
Data Collection
The initial assessment involved in-home interviews, conducted by two research
assistants, in which patients and spouses were interviewed together and separately. In the
conjoint interview, patients and spouses reported on the patient’s history of heart problems,
"(!
current CHF symptom level and medical utilization over the past 6 months (i.e. hospitalizations
and emergency room visits). The individual patient interview included detailed assessments of
the patient’s perceptions of the frequency and type of their spouse’s attempts to influence their
health behavior and adherence to the medical regimen, psychological distress and marital quality.
The individual spouse interview assessed the spouse’s perceptions of the frequency and type of
attempts used to influence the patient’s health behavior and compliance with medical regimen,
psychological distress and marital quality.
Approximately six weeks after the initial assessment, 26 of the 60 couples participated in
the daily diary portion of the study. The diary study consisted of 14 consecutive days in which
participants called in to a voicemail recorder before noon each day and recorded their answers to
the questions about their experiences from the day before. Items were rated on a 0 to 10 scale,
with anchors depending on the nature of the question. Questions used for the current analysis
were those pertaining to spouse and patient reports of provision and receipt of health-related
spousal influence attempts and marital closeness.
Follow-up assessments occurred six months after the baseline assessment in separate
telephone interviews with the patient and spouse. Patient CHF symptom level was collected from
both patients and spouses. Patients and spouses also reported on their own general physical
health and mental health. One patient died during the six-month follow-up interval and two other
couples were not able to participate in the six-month assessment, therefore the prospective
analyses relating health-related spousal influence attempts to health outcomes were based on 57
couples.

")!
Measures
CHF Severity. The New York Heart Association (NYHA Criteria Committee, 1994)
functional classification was used as the measure of illness severity. This classification system
places patients in one of four categories based on the degree of shortness of breath or angina pain
patients experience during physical activity. Classification ranges from Class I, which includes
patients with cardiac disease but with no symptoms or limitation during ordinary physical
activity like walking, to Class IV, which includes patients with severe symptoms and limitations
even while at rest. At the time of the baseline assessment, mean New York Heart Association
(NYHA) functional class was 2.3 (SD = 0.8) on a I to IV scale, with 13.3%, 55%, 20%, and
11.6% of the patients in classes I, II, III, and IV respectively. Illness severity was used as a
covariate in the analyses due to its significant association with patient physical and mental health
outcomes in current dataset.
Psychological Distress. Both patients and spouses completed the 25-item version of the
Hopkins Symptom Checklist-25 (HSCL-25; Heshbacher, Downing, & Stephansky, 1978) during
the baseline assessment. The HSCL-25 is a symptom inventory that measures symptoms of
anxiety and depression on a 4-point scale (1 = not at all, 2 = a little, 3 = quite a bit, 4 =
extremely). Item responses are summed to generate a total score ranging from 25 to 100. The
HSCL-25 has been used in previous studies of couples coping with CHF and has shown good
predictive and discriminant validity for identifying individuals with high levels of anxious and
depressive symptoms (e.g., Rohrbaugh et al., 2002; Rohrbaugh et al., 2009). Scores above 43
have been shown to be a suitable cutoff for predicting psychiatric diagnosis (Hough, Landsverk,
Stone, & Jacobson, 1982). Patients were significantly more distressed than spouses with mean
scores of 43.1 (SD = 11.11) for patients and 39.4 (SD = 9.88) for spouses (t(58) = 2.47 p = .016).
#+!
Thirty-seven percent of patients and 20% of spouses scored above the clinical cutoff associated
with a diagnosis of depression or anxiety. Internal consistency of this measure for the current
sample of patients and spouses was .92 and .94 respectively. See Appendix A for a complete list
of items.
Previous research with this sample of couples found that spouse baseline psychological
distress (as measured by the HSCL-25) was associated with patients’ physical and mental health
symptoms at 6-month follow-up over and above the effect of the patient’s own baseline distress
(Rohrbaugh et al., 2009). Therefore, the current study used spouse psychological distress as a
covariate in the analyses.
Health-Related Spousal Influence Attempts. For this investigation, two scales
reflecting health-related spousal influence attempts were developed from a pool of 12 items
assessing spouses’ attempts to influence the patient to follow medication, diet and exercise
regimens and to monitor their symptoms. Positive influence attempts (e.g. helping, praising,
engaging in desired behavior change with the patient) and negative influence attempts (e.g.
arguing, criticizing, warning) were each assessed using six items. The item content was derived
from prior research on health-related social support and control and was tailored to specifically
address health and treatment concerns of CHF patients (Cohen & Lichtenstein, 1990).
The distinction between positive and negative types of influence being used in the
proposed study is consistent with current research on health-related support and control among
older couples coping with chronic illness such as diabetes, osteoarthritis, and prostate cancer, as
well as research on couples managing general health concerns (e.g., Fekete et al., 2006; Lewis &
Rook, 1999; Stephens et al., 2010; Tucker & Mueller, 2000). However, the proposed measure
samples a broader array of behaviors than previous research in this area (see Appendix B for
#*!
complete list of items). Additionally, past research has only assessed support and control
behaviors from one person’s perspective and has varied in whose report is used as the predictor.
Some have found evidence supporting the use of spouse reports (e.g., Stephens et al., 2010;
Khan, 2010), others have found support for using patient reports (e.g., Franks et al., 2006), and
still other studies have found null results. The current study assessed spouse behaviors from both
the patients’ and spouses’ perspective and implemented a statistical model that allows for both
reports to be utilized and for their unique effects to be estimated.
Spouses indicated how often they used a particular type of influence tactic on a 5-point
scale (1 = not at all to 5 = very often). Patients indicated how often their spouse engaged in a
particular type of influence tactic on the same scale. Items for the positive and negative influence
scales were summed separately, with a potential range in scores of 6 to 30 for each type of
influence. The mean for positive influence attempts was 17.98 (SD = 5.29; range = 9-30; ! =
.69) based on patients’ report and 16.11 (SD = 5.32; range = 6-29; ! = .68) based on spouses’
report. The mean for negative influence attempts was 15.29 (SD = 4.63; range = 6-24; ! = .68)
based on patients’ report and 13.04 (SD = 4.43; range = 6-24; ! = .67) based on spouses’ report.
Patients reported significantly greater amounts of positive (t(57) = 2.61 p = .011) and negative
(t(57) = 3.12 p = .003) spousal influence attempts than did spouses. See Appendix B for a
complete list of items.
Compliance with Treatment Regimen. The measure of patient compliance was
collected at six-month follow-up from patients and spouses separately. The seven-item measure
included queries regarding taking medications exactly as prescribed, maintaining a healthy diet,
and pacing oneself to avoid overexertion, which were rated on a five-point scale (1 = rarely or
never to 5 = always). Patients’ ratings reflected the patients’ perception of the extent to which
#"!
they followed the treatment regimen. Spouses’ ratings reflected the spouses’ perceptions of the
patient’s compliance with the regimen. Mean compliance score as reported by patients was 3.75
(SD = 0.58, range = 2.0-5.0, ! = .70) at follow-up. Mean compliance score as reported by
spouses was 3.68 (SD = 0.78, range = 2.0-5.0, ! = .70) at follow-up. See Appendix C for a
complete list of items.
Marital Quality. Previous research with the current sample of CHF couples has utilized
a composite measure of marital quality comprised of two highly correlated measures of marital
satisfaction and marital communication patterns (Rohrbaugh et al., 2008, 2009). This same
composite measure of marital quality was used in this study. The composite score was generated
by standardizing the scores on the individual measures and then averaging the standardized
scores. The two measures comprising the composite are described below.
Marital satisfaction was assessed using the Relationship Assessment Scale (RAS:
Hendrick, 1988). The RAS is a brief, well-validated assessment of marital quality requiring
patients and spouses to independently rate seven items on a 1 to 5 scale with anchors that varied
depending on the question. Sample items include “How well do you feel your partner meets your
needs?”, “In general, how satisfied are you with your relationship?”, and “How many problems
are in your relationship?” Items four and seven were reverse scored and then the seven items
were averaged for a total score for each spouse. Mean RAS scores were 4.40 (SD = 0.61, range =
2.43 – 5, ! = .84) and 4.36 (SD = 0.68, range = 2.64 – 5.0, ! = .90) for patients and spouses
respectively at baseline, which reflects a relatively high level of marital satisfaction among these
couples. Patient and spouse scores were significantly correlated (r = 0.59, p < .01). See
Appendix D for a complete version of the RAS.
##!
The Constructive Communication Scale is a well-validated assessment of marital
communication patterns during conflict (CCS; Heavey, Larson, Zumtobel, & Christensen, 1996).
The CCS consists of 7 items rated on a 9-point scale (1 = very unlikely to 9 = very likely).
Spouses indicate the extent to which each item reflects what typically occurs in their relationship
during conflict or problem solving discussions. Three items assess constructive behaviors that
include both partners engaging in the discussion, expressing feelings, and suggesting solutions
and compromises. Four items assess destructive communication behaviors that include blaming,
threatening and verbal aggression. The CCS is scored by subtracting the average of the four
destructive items from the average of the three constructive items. Positive scores reflect greater
constructive behavior than destructive behavior, whereas negative scores reflect the opposite.
Mean CCS scores were 3.05 (SD = 1.95, range = -1 – 6.0) and 2.55 (SD = 2.16, range = -1.5 –
6.0) for patients and spouses respectively at baseline, which reflects greater constructive
communication than destructive communication during conflict and problem solving among
these couples. Patient and spouse scores were significantly correlated (r = 0.56, p < .01). See
Appendix E for a complete version of the CCS.
The RAS and CCS were significantly correlated for patients (r = .61, p < .001) and
spouses (r = .64, p < .001). The standardized composite measure ranged from -2.42 to 1.12 (M =
0, SD = .81) for patients and -1.65 to 1.36 (M = 0, SD = .79) for spouses. This composite
measure of marital quality was entered into the models as a potential moderator in the models
predicting patient and spouse outcomes from health-related spousal influence.
Variability in Marital Closeness. A measure of intraindividual variability was
calculated for each partner using the daily measures of marital closeness during the 14-day diary
portion of the study. Patients and spouses rated daily how “close or connected” they felt to their
#$!
partner on a 0 to 10 scale (0 = not at all, 10 = extremely close and connected). The standard
deviation of the 14 occasions of measurement of marital closeness for each partner was utilized
as the measure of variability (Whitton & Whisman, 2010). Higher scores reflect greater day-to-
day variability in feelings of marital closeness. Mean variability scores were 1.01 (SD = 0.67,
range = 0 – 3.08) and 1.08 (SD = 0.73, range = 0 – 2.76) for patients and spouses respectively.
Variability was entered into the models as a potential moderator in the models predicting patient
and spouse outcomes from health-related spousal influence.
Patient CHF Symptom Level. A measure of patients’ CHF symptom level was collected
at six-month follow-up. Both patients and spouses reported on patient symptom level
independently. The measure of CHF symptoms reflected the extent to which patients had
experienced eight specific symptoms in the previous month (Rohrbaugh et al., 2008, 2009). The
symptoms were (a) fatigue or lack of energy for normal activities; (b) difficulty breathing,
especially with exertion; (c) waking up breathless at night; (d) swelling in ankles and feet; (e)
chest pain; (f) heart flutter (fibrillation); (g) dizziness or fainting; and (h) nausea, with abdominal
swelling or tenderness. Symptoms were rated on a three-point scale (1 = not at all, 2 = some, 3 =
a lot) and the average was taken of the eight ratings to obtain an overall symptom level. Mean
CHF symptom score at follow-up was 1.41 (SD = 0.31, range = 1-2.8, ! = .84) based on patients’
ratings and 1.44 (SD = 0.37, range = 1-2.5, ! = .81 based on spouses’ ratings.
Patient and Spouse General Physical and Mental Health Outcomes. The Short Form
(SF-36) Health Survey was used to assess patients’ and spouses’ general physical and mental
health at follow-up. The SF-36 (McHorney, Ware, & Raczek, 1993; Ware, 2000) is a widely
used measure that yields a norm-based physical health composite score (PCS) and mental health
composite score (MCS). Scores range from 0 to 100 and higher scores represent better health.
#%!
The physical health subscale assesses the degree to which respondents experienced problems or
limitations in activities due to their health over the past four weeks. The mental health subscale
assesses respondents’ feelings over the past four weeks with items such as “Have you felt
downhearted and blue?” and “Have you been a happy person?” For patients, the mean PCS at
follow-up was 49.2 (SD = 24.2, ! = .80) and mean MCS at follow-up was 67.0 (SD = 21.1, ! =
.71). Spouses’ PCS mean at follow-up was 64.6 (SD = 25.5, ! = .81) and their MCS mean was
74.3 (SD = 19.6, ! = .84). As expected, spouses scored significantly higher than patients on the
PCS (t(57) = 3.10 p = .003) , which is reflective of better physical health among spouses. See
Appendix F for the complete SF-36.
Daily Diary Measures. Patients and spouses independently responded to 15 single-item
measures that were rated on a 0 to 10 scale, with anchors depending on the nature of the
question. Items were modeled after measures from the baseline assessment and were worded to
assess what had occurred on the previous day. Participants called in daily to report responses for
the previous day on a voicemail recorder. Patients and spouses indicated the extent to which the
spouse tried to influence the patient yesterday to keep a healthy diet, get the right amount of
exercise, or take medications as prescribed. Marital quality was assessed by asking patients and
spouses to indicate how close and connected they felt to their partner yesterday. The data
pertaining to provision and receipt of health-related spousal influence was only collected from a
subsample of the couples. Data from a total of 26 couples was used from the daily diary study to
analyze hypotheses pertaining to associations between spousal influence and marital closeness.
Analytic Plan
Dyadic Analysis Using Multilevel Modeling. Given the dyadic nature of the data,
multilevel modeling (MLM) for mixed independent variables was used to construct the APIM
#&!
for distinguishable dyads. (Kenny & Kashy, 2010; Kenny, Kashy, & Cook, 2006). As the name
multilevel modeling implies, there are multiple levels within the data that can be analyzed. The
focus in the current project is couples, therefore two levels exist in the current data, individuals
(level 1) who are nested within a dyad (level 2). In dyadic data analysis, the outcome measure or
dependent variable is always a level 1 variable thus allowing each individual’s outcome to be
predicted independently in the model. Predictor variables or independent variables on the other
hand, can be between-dyads variables (level 2 variables), within-dyads variables (level 1
variables) or mixed variables. Between-dyads variables differ from dyad to dyad, thus both
members of the dyad have the same score on the variable. For example, marital duration is a
between-dyads variable because couples vary on how long they have been married, but each
partner in the couple has the same marital duration. Within-dyads variables differ between the
two members within the dyad, but not between dyads. Gender is a prototypical within-dyads
variable in heterosexual couples because each dyad has a man and a woman, but when you
average across gender each dyad has an identical average score. Mixed independent variables
vary both between and within dyads (Kenny & Cook, 1999). Marital satisfaction is an example
of a mixed independent variable because the two spouses may differ from one another on how
satisfied they are, and couples on average may be more or less satisfied than others. The analyses
of study hypotheses used mixed independent variables with some between-dyads variables
included as covariates.
Distinguishability of Dyad Members. In addition to determining the nature of the
independent variables, distinguishability of the dyad members must be established. Dyad
members are considered distinguishable if there is a meaningful factor that can be used to order
the two individuals. The designation of a “meaningful” factor can be done empirically, based on
#'!
actual differences in the data, or theoretically. In the current study, I distinguished dyad members
based on their role as either patient or spouse, as the primary goal of this project was to examine
differences by role. These dyads could also be distinguished by gender because the data consist
of heterosexual couples. And while gender differences are a salient topic in the marital literature,
the current sample has few female patients (28% of the sample). Post-hoc analyses attempted to
identify potential gender differences by testing the interaction of role and gender, but it should be
noted that the power to detect these effects is limited by small sample size.
Nonindependence in Dyadic Data. MLM was used to account for the nonindependence
in dyadic data. Dyadic nonindependence occurs when the two scores from the two members of
the dyad are more similar to one another than are two scores from two people who are not
members of the same dyad. Ignoring the nonindependence of dyadic data and conducting
analyses using the individual as the unit of analysis can lead to inaccurate estimates of standard
errors, which can lead to both Type I and II errors. It should be noted that the determination of
nonindependence is both a theoretical and empirical question. Given that the current study
examined married partners who have chosen to be in a relationship as opposed to individuals
who are randomly paired together, there is a theoretical reason for addressing issues of
nonindependence. From an empirical standpoint, the degree of nonindependence can be
examined statistically. Because the dyads in this project are distinguishable, nonindependence
was assessed by computing the Pearson product-moment correlation coefficient. The correlations
between patient and spouse variables in the current project are small to large in size (see Table 2
and 3). Given these nonzero correlations, the data are nonindependent, and thus the dyadic
dependency must be included in the analysis. In the APIM model (see Figure 1), both partners’
characteristics are used to predict their own and each others’ outcomes. To account for the
#(!
nonindependence, the model allows for a correlation between the two errors of the patient and
spouse’s outcomes. The correlation between the residuals controls for the additional sources of
nonindependence not explained by the actor and partner effects in the model. In addition to
correlated residuals, the APIM includes correlations between the independent variables. This
correlation ensures that if either partner’s independent variable predicts an outcome variable, it is
done while controlling for the other partners’ independent variable. Thus, actor effects are
estimated controlling for partner effects, and partner effects are estimated controlling for actor
effects.
Test of the Indirect Effect. The conceptual model for Aim 1 predicts an indirect effect of
positive and negative influence on CHF symptom level via the intervening variable of treatment
compliance (see Figure 2). Analysis of this mediation model was conducted in two steps (Krull
& MacKinnon, 2001; MacKinnon, Fairchild, & Fritz, 2007). In the first step, the associations
between positive and negative influence attempts at baseline and treatment compliance at follow-
up were tested. In the second step, the associations between treatment compliance and CHF
symptom level at follow-up were tested. Presentation of the results and figures reflect this two-
step approach.
Test of Moderation. Lastly, we can test moderator effects using the APIM. A moderator
variable influences the size of the effect of an independent variable on a dependent variable. In
the APIM, the moderator can be an actor variable or a partner variable. (e.g., if patient marital
quality and spouse marital quality moderate the effect of spousal influence attempts on
compliance). For example, we might find that the patient actor effect of negative influence on
patient compliance is moderated by the patient’s own level of marital quality (i.e., actor-
moderated actor effect). Or the patient actor effect of negative influence on compliance could be
#)!
moderated by the spouse’s level of marital quality (i.e., partner-moderated actor effect).
Although both actor and partner moderators can be tested simultaneously in APIM using
interaction terms, due to the moderate sample size, actor and partner moderator variables were
analyzed in separate models (Garcia, Kenny, & Ledermann, 2012).
Test of Models. All analyses were done using the MIXED procedure in SPSS 17.0 and
used the APIM two-intercept approach (Raudenbush, Brennan, & Barnett, 1995). The two-
intercept model was used in order to obtain separate parameter estimates for patient and spouse.
All continuous variables were grand mean centered prior to analysis to minimize problems
related to multicollinearity among predictors that results from the creation of interaction terms.
Actor and partner effect estimates are shown as unstandardized regression coefficients (Aiken &
West, 1991). Therefore the beta values shown represent the rise in compliance, CHF symptoms,
general physical health, or mental health scores associated with a one-point rise in the predictor
variable. Significant interactions were plotted using outcome values estimated at one standard
deviation above and one standard deviation below the mean of the two predictor variables
(Dawson & Richter, 2006). I utilized an estimate of overall model fit based on the proportional
reduction of residual error variance, often referred to as “pseudo R
2
” (Luke, 2004; Snijders &
Bosker, 2012). This statistic was calculated by first estimating the error variance for the
unrestricted or empty model (s
2
e
!), i.e., the model without any predictors, and the error variance
for the predicted model (s
2
e
). These error terms were then entered into the following equation:
s
2
e
! - s
2
e
= s
2
e
!
This percent reduction in variance can be interpreted as the percent of variance in the outcome
variable explained by the predictor variables in the model.
$+!
Repeated Measures Analysis. The same analytic approach was taken in analyzing the 14-
day diary data to test Aim 4. Although this repeated measures data appear to have three levels
(i.e. time nested within person, and person nested within dyad), the dyad members are measured
at the same time points making the level of repeated measure the same for both members of the
dyad. Thus time and person are crossed, not nested (West, 2012). When the three-level model is
assumed, the correlation between dyad members’ scores at each time point is assumed to be zero,
which in this case is incorrect. As such, I used a two-level cross-classified model so that the
residual errors between partners at each time point and within each partner across time could be
estimated. The standard APIM was combined with the stability and influence model (e.g. earlier
measures of marital closeness is expected to predict marital closeness at a later point in time) to
obtain a cross-lagged APIM. This model estimates the effect of the patient’s report of receiving
influence and the spouse’s report of providing influence on patient and spouse marital closeness
the next day, while controlling for the effect of previous day’s closeness. This model also
included a test for moderation by the nature of the influence attempts (i.e. ratio of positivity to
negativity of the influence attempts) provided in the baseline assessment, as combining multiple
methods like diary data and global self-report data can increase understanding of the
phenomenon (Laurenceau & Bolger, 2005).
Results
Descriptive Statistics of Patient-Spouse Dyads
All variables were summarized using descriptive statistics (means and standard
deviations; see Table 4). Paired t-tests were used to examine differences between patient and
spouse demographic characteristics and study variables. On average, patients were older than
spouses, more distressed at baseline, had worse physical health at six-month follow-up, and
$*!
reported more positive and negative influence attempts by spouses than spouses reported about
themselves. There were no significant differences between patients and spouses in baseline
marital quality, variability in marital closeness, reports of patient compliance, reports of patient
CHF symptoms at follow-up, mental health at follow-up, and daily diary variables.
Correlations Among the Variables and Nonindependence
Correlations among the key study variables and covariates are presented in Tables 2 and
3. In Table 2, correlations for patients are reported below the diagonal, and correlations for
spouses are reported above the diagonal. Correlations between each partners’ report of the same
variable (i.e., measure of nonindependence) are presented along the diagonal. Table 3 shows the
correlations of patients’ and spouses’ reports of all study variables. As can be seen in these
tables, the three physical and mental health outcomes of interest in this study were correlated
with illness severity and baseline psychological distress. Previous analyses of the current data
found that spouses’ psychological distress at baseline was negatively associated with patient
health outcomes at follow-up over and above the effect of the patients’ own baseline distress. As
such, spouse distress and illness severity were included as covariates in all models examining
health outcomes
1
. These covariates were not found to be significant predictors in every model
tested, but were consistently used in all models so as to allow for accurate comparison between
models.
It should be noted that patients and spouses did not differ significantly in marital quality
at the baseline assessment or in their day-to-day variability in marital closeness. However for
spouses, closeness variability was significantly correlated with marital quality (r = -.29, p = .03)
such that lower general marital quality for spouses was associated with greater day-to day
variability in marital closeness. In contrast, patients’ marital quality at baseline was not
$"!
correlated with their variability in marital closeness. Examination of the partial correlation of
marital quality and closeness variability while controlling for gender revealed that this difference
between patients and spouses was not due to gender differences between the groups. Given this
difference in correlation pattern based on role, marital quality was included as a covariate in the
models that tested closeness variability as a moderator in order to isolate the effect of closeness
variability.
The Effects of Health-Related Spousal Influence on CHF Symptom Level
Figures 2 and 5 depict the models analyzed to test for the indirect effect of spousal
influence attempts on patient CHF symptom level at follow-up via the intervening variable of
follow-up treatment compliance (Aim1). The APIM was used to identify actor effects (which
estimate the effects of each person’s report of influence on their own report of patient
compliance and subsequent report of patient symptom level) and partner effects (which estimate
the effects of each person’s report of influence on their partner’s report of patient compliance
and subsequent report of patient symptom level). There are two sets of actor and partner effects
in the proposed model, first the effects of influence on compliance and then the effects of
compliance on CHF symptoms, which are tested in separate APIM models (Kenny, West, &
Garcia, 2012). The proposed model also includes moderation by marital quality and closeness
variability on the association between influence attempts and compliance. Therefore, first I
present the actor and partner effects of spousal influence attempts on treatment compliance, then
the results of the moderation analysis of the effects of influence on compliance (interaction
effects), and lastly I present the actor and partner effects of compliance on CHF symptom level.

$#!
The effects of health-related spousal influence on treatment compliance
Actor and Partner Effects. Effect estimates and significance levels are presented in
Table 5 (column 1) and Figure 10. Results indicated a significant actor effect for patients such
that negative influence attempts as reported by the patient at baseline were negatively associated
with the patient’s report of compliance at follow-up. A significant partner effect for spouses was
found such that positive influence attempts as reported by the patient were positively associated
with the spouse’s report of compliance at follow-up.
This model included illness severity and spouse psychological distress as covariates,
which were not significantly associated with either patient or spouse report of compliance at
follow-up. The model tested explained 15.7% of the variance in the patients’ report of treatment
compliance at follow-up and 18.1% of the variance in the spouses’ report of the patients’
treatment compliance at follow-up.
Marital Quality and Variability in Marital Closeness as Moderators. Patient and spouse
marital quality did not moderate any of the actor or partner effects of influence attempts on
treatment compliance (see Table 5, columns 2 and 3).
When patient and spouse closeness variability were included in the model, some direct
and moderated effects were found (see Table 6). Results indicated that one interaction was
statistically reliable; patients’ variability in marital closeness had a moderating effect on the
association between influence attempts and compliance. Specifically, patient closeness
variability moderated the patient actor effect of positive influence on follow-up compliance (i.e.
actor-moderated actor effect). Note that patients’ report of positive influence was positively
associated, although not significant at the .05 level, with their report of compliance (as indicated
by the positive effect estimate for Pt pos infl " Pt comply). This effect was moderated by patient
$$!
closeness variability such that at low levels of patient closeness variability patients’ report of
positive influence is positively associated with patients’ report of compliance at follow-up. In
contrast, at high levels of variability in patients’ feelings of closeness, there is no relationship
between positive influence and compliance at follow-up (see Figure 11).
Including patient and spouse closeness variability also resulted in some additional main
effects. A significant actor effect for spouses emerged such that negative influence attempts as
reported by the spouse at baseline were negatively associated with the spouses’ report of
compliance at follow-up. Additionally, spouse closeness variability was positively associated
with their own report of compliance at follow-up. Illness severity and spouse psychological
distress were included in these models as covariates. Only spouse distress was found to be
significantly and positively associated with patient compliance at follow-up as reported by both
patients and spouses. The actor-moderated model (Table 6, column 1) explained 20.8% of the
variance in the patients’ report of treatment compliance at follow-up and 19.2% of the variance
in the spouses’ report of the patients’ treatment compliance at follow-up.
The effects of treatment compliance on CHF symptom level
Actor and Partner Effects. Results showed a significant actor effect for patients such that
greater compliance as reported by the patient at follow-up was associated with lower CHF
symptom level as reported by the patient (see Table 7 and Figure 12). Spouses’ report of patient
compliance was not associated with CHF symptom level as reported by either the patient or the
spouse. Illness severity and spouse psychological distress were included in these models as
covariates. Spouse distress was associated with greater CHF symptoms as reported by both
patients and spouses at follow-up. Illness severity was also associated with greater CHF
symptoms at follow-up, but only for patients’ report of CHF symptoms. The model explained
$%!
35.1% of the variance in the patients’ report of CHF symptom level at follow-up and 22.0% of
the variance in the spouses’ report of the patients’ CHF symptom level at follow-up.
In summary, I found partial support for Aim 1 as positive and negative spousal influence
attempts at baseline were indirectly associated with CHF symptom level at follow-up via the
intervening variable treatment compliance. As predicted, negative influence attempts were
associated with lower compliance at follow-up and higher CHF symptom level (Hypothesis 1b).
However, positive influence attempts were associated with greater compliance at follow-up and
lower CHF symptom level only when patient closeness variability was low, providing partial
support for Aim 3 (Hypothesis 3a). Lastly, the use of APIM revealed that the associations
between influence, compliance, and CHF symptoms were mostly driven by the patients’ report
of the variables, yielding primarily an actor-actor model. Partner effects were very limited with
patients’ report of positive influence relating to spouses’ report of compliance. Furthermore, the
significance of spouses’ reports were rather limited, as only one effect was found for spouses’
report of negative influence relating to their report of compliance when closeness variability was
controlled in the model.
The Effects of Health-Related Spousal Influence on General Physical Health
Figures 3 and 6 depict the conceptual models tested to examine the effect of spousal
influence attempts on patient and spouse general physical health at follow-up. First, I present the
actor and partner effects of spousal influence attempts on general physical health. Then, I present
the results of the moderation analysis of the effects of influence on physical health by marital
quality and closeness variability.
Actor and Partner Effects. No significant actor or partner effects were found for positive
and negative influence on general physical health at follow-up for patients or spouses (see Table
$&!
8, column 1 and Figure 13). Covariates of illness severity and spouse distress at baseline were
associated with worse physical health for patients only. The model tested explained 39.1% of the
variance in the patients’ general physical health at follow-up and 9.7% of the variance in the
spouses’ general physical health at follow-up.
Marital Quality and Variability in Marital Closeness as Moderators. Including patient
and spouse marital quality as a moderator did not alter these findings (see Table 8, columns 2
and 3). However, including patient and spouse closeness variability in the model resulted in one
significant interaction effect (see Table 9). Patient closeness variability was shown to moderate
the effect of spouse report of positive influence on spouse physical health at follow-up. Recall
that spouses’ report of positive influence was negatively associated with their physical health (as
indicated by the negative effect estimate for Sp pos infl " Sp Physical Health). Although this
actor effect was not statistically significant, including patient closeness variability as a moderator
resulted in a significant interaction (see Figure 14). At low levels of positive influence reported
by spouses, spouse physical health was significantly better when patient closeness variability
was low. At high levels of positive influence reported by spouses, patient closeness variability
had no effect on spouse physical health (i.e. partner-moderated actor effect). Covariates of illness
severity and spouse distress at baseline were associated with worse physical health for patients
only. The partner-moderated model (Table 9, column 2) explained 39.6% of the variance in the
patients’ general physical health at follow-up and 19.1% of the variance in the spouses’ general
physical health at follow-up.
In summary, I did not find support for the predictions in Hypotheses 2a and 2b pertaining
to general physical health for patients. Nor did I find support for the predictions in Hypotheses 2c
and 2d for physical health for spouses. I also did not find support for the predicted moderation of
$'!
influence attempts by marital quality and closeness variability for physical health outcomes in
Hypotheses 3c and 3d. Contrary to my predictions, when patient closeness variability was low,
low levels of positive influence by spouses were associated with better physical health outcomes
for spouses, whereas high levels of positive influence were associated with worse health for
spouses.
The Effects of Health-Related Spousal Influence on Mental Health
Figures 4 and 7 depict the conceptual models tested to examine the effect of spousal
influence attempts on patient and spouse general mental health at follow-up. First, I present the
actor and partner effects of spousal influence attempts on mental health at follow-up. Then, I
present the results of the moderation analysis of the effects of influence on mental health by
marital quality and closeness variability.
Actor and Partner Effects. Results showed a significant actor effect for patients such that
patients’ report of negative influence was negatively associated with their mental health at
follow-up (see Table 10, column 1 and Figure 15). No partner effects were found for influence
attempts and mental health at follow-up. Covariates of illness severity and spouse distress were
included in the model, but only spouse distress was negatively associated with mental health at
follow-up for both patients and spouses. The model tested explained 37.4% of the variance in
patients’ mental health at follow-up and 18.8% of the variance in spouses’ mental health at
follow-up.
Marital Quality and Variability in Marital Closeness as Moderators. Including patient
and spouse marital quality as a moderator did not alter these findings (see Table 10, columns 2
and 3). However, including patient and spouse closeness variability in the model did result in one
significant interaction effect, the interaction between spouse closeness variability and spouse
$(!
report of positive influence (see Table 11). Recall that spouses’ report of positive influence was
negatively associated with patients’ mental health at follow-up (as indicated by the negative
effect estimate for Sp pos infl " Pt Mental Health). Although this partner effect was not
statistically significant, including spouse closeness variability as a moderator resulted in a
significant interaction (i.e. partner-moderated partner effect). When spouse closeness variability
is low, patient mental health is better at low levels of positive influence (see Figure 16). When
spouse closeness variability is high, spouse report of positive influence has no effect on patient
mental health.
Spouse distress remained a significant predictor of mental health at follow-up for
patients, but was not for spouses when spouse closeness variability was included in the model.
Additionally, illness severity was negatively associated with patient mental health at follow-up
when patient and spouse closeness variability was included in the model. The partner-moderated
model (Table 11, column 2) explained 60.5% of the variance in patients’ mental health at follow-
up and 11.7% of the variance in spouses’ mental health at follow-up.
In summary, I found minimal support for the predictions pertaining to mental health for
patients. As predicted in Hypothesis 2b, negative influence as reported by patients was
negatively associated with patient mental health at follow-up. However, no association was
found for positive influence and patient mental health (Hypothesis 2a). I did not find support for
the predictions in Hypotheses 2c and 2d pertaining to mental health for spouses, as positive and
negative influence were not associated with spouse mental health at follow-up. I also did not find
support for the predicted moderation of influence attempts by marital quality and closeness
variability for mental health outcomes in Hypotheses 3c and 3d. Contrary to my predictions, low
levels of positive influence by spouses (as opposed to high levels of positive influence), in the
$)!
context of low spouse closeness variability, was associated with the best mental health outcomes
for patients.
The Effects of Health-Related Spousal Influence on Daily Marital Closeness
Figures 8 and 9 depict the conceptual models tested to examine the effect of providing
and receiving spousal influence on day-to-day marital closeness. The analysis was performed
using repeated measures data from the 14-day daily diary study. As such, influence attempts
reported by patients and spouses on day (t) were used to predict patient and spouse marital
closeness on day (t + 1), while controlling for marital closeness on day (t). Patients and spouses
did not report on the nature of the daily influence attempts, so the ratio of relative positivity to
negativity of the influence, as characterized by the couples during the baseline assessment, was
utilized as a potential moderator of the association between influence attempts and next day
marital closeness. First, I present the actor and partner effects of spousal influence attempts on
marital closeness. Then I present the results of the moderation analysis of the effects of influence
on marital closeness by the relative positivity to negativity of the influence attempts.
Actor and Partner Effects. Results showed a positive and significant actor effect for
patients of receiving influence on their own next day marital closeness (see Table 12, column 1
and Figure 17). No partner effects were found for daily marital closeness for patients or spouses.
Previous day marital closeness was significantly associated with next day closeness for both
patients and spouses. The model tested explained 51.9% of the variance in patients’ feelings of
marital closeness and 67.3% of the variance in spouses’ feelings of marital closeness.
Moderation by the relative positivity to negativity of influence attempts
provided/received. When patient and spouse ratings of the relative positivity to negativity of
influence attempts (pos:neg ratio) were included in the model, some direct and moderated effects
%+!
were found (see Table 12, columns 2 and 3)). Results indicated that one interaction was
statistically reliable; patients’ pos/neg ratio of influence attempts had a moderating effect on the
association between day t receipt of influence attempts and day t + 1 marital closeness for
patients (i.e. actor-moderated actor effect). Recall that patients’ report of receiving influence on
one day was positively associated with patients’ feelings of marital closeness on the next day (as
indicated by the positive effect estimate for Pt infl " Pt marital closeness). This effect was
moderated by the patients’ characterization of the influence attempts received such that when the
influence was more positive than negative (2:1), receiving greater amounts of influence from
their spouse was associated with greater feelings of marital closeness for patients the next day.
However, when the influence was more negative than positive (1:2), the patients’ report of
receiving influence was not associated with their feelings of closeness the next day (see Figure
18). Including patients’ and spouses’ ratings of the relative positivity to negativity of influence
attempts also resulted in a significant main effect for spouses. Spouses’ characterization of the
influence attempts provided was positively associated with their own report of marital closeness.
So spouses who characterized their influence attempts as more positive than negative also
reported greater marital closeness. The model tested (Table 12, column 2) explained 52.8% of
the variance in patients’ feelings of marital closeness and 70.2% of the variance in spouses’
feelings of marital closeness.
In summary, I found support for Hypotheses 4a and 4b as influence attempts were
associated with next day marital closeness, and this effect was moderated by the nature of the
influence attempts. As predicted, next day marital closeness for patients increased in response to
their report of the previous day’s influence attempts when the influence was more positive than
negative. However, the predicted decrease in marital closeness in response to the previous day’s
%*!
influence attempts that were more negative than positive was not supported, rather no association
was found. Furthermore, spouse marital closeness was not associated with providing influence as
was predicted, but rather was associated with the nature of their influence attempts.
Post-Hoc Analyses
Attempts were made to assess the importance of other relevant factors that were not the
focus of the current study hypotheses, but that may offer additional insights. The following
concepts and variables were assessed in post-hoc analyses: interaction between positive and
negative influence attempts, gender, age, and marital duration. Given parameter limitations
imposed by the current sample size, these variables were tested as moderators on the APIM
models of positive and negative influence, but not on models with moderation by marital quality
or closeness variability.
Interaction of Positive and Negative Influence Attempts. The interaction between
positive and negative influence attempts was included in the models testing the effects of
influence on compliance, general physical health, and mental health at follow-up. Results
showed that the pattern of independent effects of positive influence and negative influence on
these outcomes remained unchanged with the inclusion of the interaction term. Additionally,
there were no significant interactions of positive and negative influence in any models testing
outcomes of compliance, physical health, or mental health for patients or spouses (for
compliance - "s ranged from -.002 - .001, p values ranged from .59 – .92; for physical health - "s
ranged from -.163 - .201, p values ranged from .27 – .88; for mental health - "s ranged from -
.002 - .118, p values ranged from .36 – .99).
Gender Differences. Gender was examined as a potential moderator in the models
testing the effects of influence on compliance, general physical health, mental health, and daily
%"!
marital closeness. Results showed that the pattern of independent effects of positive influence
and negative influence on these outcomes remained unchanged with the inclusion of gender as a
moderator. Gender was not directly associated with any of the outcomes and there were no
significant interactions of gender with positive or negative influence in any models tested (for
compliance - "s ranged from -.063 - .125, p values ranged from .12 – .93; for physical health - "s
ranged from -1.48 – 6.84, p values ranged from .28 – .99; for mental health - "s ranged from -
2.08 – 1.26, p values ranged from .18 – .99; for daily marital closeness - "s ranged from -2.94 –
.023, p values ranged from .22 – .98).
Age Differences. Age was examined as a potential moderator in the models testing the
effects of influence on compliance, general physical health, mental health, and daily marital
closeness. Results showed that the pattern of independent effects of positive influence and
negative influence on these outcomes remained unchanged with the inclusion of age as a
moderator. Age was not directly associated with any of the outcomes and there were no
significant interactions of age with positive or negative influence in any models tested (for
compliance - "s ranged from -.001 - .003, p values ranged from .20 – .77; for physical health - "s
ranged from -.581 – .382, p values ranged from .19 – .89; for mental health - "s ranged from -
.309 – .247, p values ranged from .32 – .87; for daily marital closeness - "s ranged from -.051 –
.010, p values ranged from .11 – .99).
Marital Duration. Marital duration was examined as a potential moderator in the models
testing the effects of influence on compliance, general physical health, mental health, and daily
marital closeness. Results showed that the pattern of independent effects of positive influence
and negative influence on the outcomes remained unchanged with the inclusion of marital
duration as a moderator. Marital duration was not directly associated with any of the outcomes
%#!
and there were no significant interactions of age with positive or negative influence in the
models examining compliance, physical health, and mental health (for compliance - "s ranged
from -.004 - .003, p values ranged from .29 – .97; for physical health - "s ranged from -.080 –
.203, p values ranged from .14 – .96; for mental health - "s ranged from -.240 – .141, p values
ranged from .24 – .99). However, there was a significant interaction for spouses between
patients’ daily report of receiving influence and marital duration (" = .007, p = .008). For
spouses in marriages of long duration, the patients’ report of receiving influence from the spouse
was positively associated with spouses’ feelings of marital closeness the next-day. In contrast,
for spouses in marriages of shorter duration, the patients’ report of receiving influence from the
spouse was negatively associated with next-day feelings of marital closeness for spouses.
Additionally, given the high degree of covariance of error terms in the APIM models
testing influence on compliance and compliance on CHF symptom level, models were analyzed
again using standard regression procedures with couple-level outcome variables. Results of the
regression of patient and spouse reports of positive and negative influence on the average of
patient and spouse reports of compliance at follow-up revealed a distinct pattern of results from
the APIM model. The previously found negative association between the patient report of
negative influence and patient report of compliance at follow-up was not found when the using
the couple-level measure of compliance (" = -.021, p = .910). Results of the regression of the
couple-level measure of compliance on the couple-level measure of CHF symptom level was
similar to the results of the APIM, with greater compliance being associated with lower symptom
level (" = -.122, p = .023). Both analyses suggest that meaningful information was gathered from
the APIM analysis, as we found that patients’ reports of influence, compliance, and symptoms
are what drive the finding, and thus the APIM results will be the focus of the discussion.
%$!
Discussion
The aims of the current study were to (1) identify the distinct effects of two types of
spousal influence attempts (positive and negative) on patient compliance and CHF symptoms,
(2) examine the potential broader consequences of these influence attempts for patients’ and
spouses’ general physical and mental health, (3) investigate how the marital context (specifically
marital quality and variability in closeness) influences the above associations and (4) explore,
using a daily diary methodology, the day-to-day effects of providing and receiving influence
attempts on feelings of marital closeness.
Under Aim 1, as predicted the study revealed that negative influence attempts were
associated with lower treatment compliance for patients six months later and in turn greater
symptom levels. One measure of the marital context, a measure that tapped the day-to-day
variability in feelings of marital closeness was found to moderate the associations between
influence attempts and compliance. Although marital quality
2
, as measured by marital
satisfaction and marital communication patterns, surprisingly did not moderate the associations
between influence attempts and compliance. As predicted, there was a positive association
between positive influence and compliance in the context of stability of marital closeness, with
high day-to-day variability nullifying the positive effect. Unexpectedly, the negative associations
between negative influence and compliance were not buffered by greater stability of marital
closeness, but rather remained regardless of the marital context.
When looking at the consequences of spousal influence attempts for general physical
health more broadly (Aim 2), there was surprisingly no association between positive or negative
influence attempts at time 1 and patient health at time 2. However, spouse physical health was
negatively associated with their positive influence attempts when patient feelings of closeness
%%!
were stable. Also contrary to predictions, there were no negative consequences of engaging in
negative influence attempts for spouses’ physical health at follow-up. When looking at the
relationship between spousal influence attempts and general mental health, as expected negative
influence at time 1 was associated with worse mental health at follow-up for patients. However,
patients’ mental health did not evidence benefits from positive influence attempts as expected,
but rather was negatively associated with greater positive influence when spouses’ feelings of
closeness were stable. Also surprising was the finding that neither positive nor negative influence
was associated with mental health at follow-up for spouses.
Lastly, Aim 4 used daily diary data to examine the day-to-day effect of providing and
receiving influence on feelings of marital closeness and found that receiving influence from
one’s spouse served to increase patients’ feelings of marital closeness, particularly when the
influence was more positive than negative. However, providing influence had no effect on
spouses’ feelings of marital closeness. Yet, the nature of the influence (relative positivity to
negativity) was directly related to spouses’ marital closeness, with greater positive relative to
negative influence being associated with greater feelings of marital closeness for spouses.
Relations Between Constructs: Positive versus Negative Influence
When distinguishing spousal influence attempts that appear more positive (e.g. praising
for exercising, join in following healthy diet) from ones that seem more negative (e.g. criticizing
for not following medication regimen, warning to attend to symptoms), a distinct pattern of
positive associations for positive attempts and negative associations for negative attempts
emerged. First and foremost these findings demonstrate that spouses’ involvement in illness
management is indeed consequential for patients with CHF. While the starkly contrasting
findings between positive and negative influence attempts highlight the importance of the nature
%&!
of the behaviors spouses engage in around illness management, an interesting distinction in their
effects emerged. The results showed a consistent negative relationship for negative influence
attempts that was not moderate by the marital climate in which the influence was occurring,
whereas positive influence attempts were only associated with outcomes under certain
circumstances. These findings suggest that negative influence attempts should be avoided, as
their negative effect is buffered neither by positive relationship characteristics nor by positive
influence attempts themselves.
Past research in spousal support and control has been muddled by idiosyncratic
definitions of support and control, with much overlap between the constructs (Helgeson, et al.,
2004). More recent work has attempted to improve clarity by assessing specific spousal
behaviors as opposed to general notions of what is supportive and what is controlling (e.g.,
Fekete et al., 2006; Franks et al., 2006; Stephens et al., 2010). However, most of this work has
been cross-sectional in nature. The results of these cross-sectional studies demonstrate that at a
single point in time, spouses who use techniques like encouraging the patient are more likely to
be paired with patients who are more compliant. Similarly, these studies show that spouses who
use techniques like warning the patient of negative consequences of non-adherence are more
likely to be paired with patients who are less compliant. However, these findings are difficult to
interpret with regard to directionality and beg the question of whether spouses engage in more
negative influence attempts because their patient-spouse is non-compliant. The current study’s
use of a time-lagged design does not eliminate this possibility as these couples were sampled in
the midst of this coping process, however it provides stronger support for the ideas posited in
previous research and the current research that spousal influence attempts have an impact on
patient illness outcomes.
%'!
The Role of the Patient’s Perspective
An additional step taken in the current study was to use both patient and spouse reports of
the study variables. This approach revealed that it was the patients’ perspective of the process
that drove the outcomes. Specifically, patients’ report of the influence attempts predicted the
patients’ report of their own compliance six months later, which in turn was associated with the
patients’ report of their CHF symptom level. The spouses’ perspective on their own influence
attempts was minimally predictive of their view of patient compliance, with patients’ perspective
of influence predicting the spouses’ report of compliance more consistently. Additionally,
spouses’ perspective of patient compliance was not associated with either the patients’ or the
spouses’ view of CHF symptoms at follow-up. These findings demonstrate that although spouses
are intimately involved in this process, the spouses’ view does not independently contribute to
CHF-specific treatment compliance and symptom outcomes. Given that it is the patient who is
suffering the symptoms, enduring the treatment regimen, and experiencing their spouses’ efforts
to “assist” with illness management, the findings suggest that the patients’ perceptions become
paramount in the context of a chronic illness like CHF.
In addition to revealing the patients’ view as the relevant predictor in the models, these
findings may also help to clarify some of the inconsistent findings in the literature with regard to
spousal support and control. Past research has most often utilized the spouses’ report of the
support and control behaviors to predict patient outcomes (e.g., Fekete et al., 2006; Franks et al.,
2006). Such literature has at times found significant effects of spouse involvement but other
times found no association. While it seems intuitive to use the spouses’ perspective on the
support behaviors, as they are the ones engaging in the behaviors, the current study demonstrates
that it is the perspective of the recipient that is predictive of outcomes. Additionally, teasing
%(!
apart patient and spouse perspectives reveals that the use of couple-level variables (averaging
spouse and patient reports), an approach often found in the literature, could also obscure
findings.
The role of the patient report is demonstrated again when looking beyond CHF-specific
outcomes. First, only the patients’ mental health at follow-up was associated with influence
attempts; no effects were found for spouse mental health. The patients’ report of negative
influence was negatively associated with patients’ mental health six months later, a finding
consistent with past research of patients in a cardiac rehabilitation program (Franks et al., 2006).
Additionally, patients’ mental health was negatively associated with spouse reports of positive
influence, suggesting that too much involvement from spouses even when it is positive in nature
may not be beneficial. This is consistent with some findings in the marital literature
demonstrating negative effects of “over provision” (Brock & Lawrence, 2009) and affective
costs to recipients of support, as it calls attention to their need for support and may challenge
their sense of self-efficacy (e.g., Bolger, Zuckeman, & Kessler, 2000; Shrout, Herman, & Bolger,
2006). It is somewhat surprising that spouses’ mental health did not benefit from engaging in
positively laden influence attempts, particularly when existing literature shows affective benefits
of providing support to a spouse in need (Brown, Nesse, Vinokur, & Smith, 2003; Liang, Krause,
& Benner, 2001). However, this discrepancy may be a result of measurement issues. Research
demonstrating the emotional benefits of providing support has typically assessed these benefits
on the basis of single-item measures of mood or positive/negative affect. The current study’s
mental health measure assessed mood (e.g. depression, anxiety, happiness), but also captured a
broader range of psychological symptoms including energy and fatigue, concentration, and
%)!
engagement in social activities. The current results suggest that spouses’ use of positive
influence attempts does not influence this broader assessment of psychological well-being.
Also surprising is the finding that spouses’ did not evidence any negative consequences
of their own when engaging in negatively laden influence attempts, given the sizeable couples
literature demonstrating significant costs of negative and conflictual behavior (see Kiecolt-Glaser
& Newton, 2001 for a review). One possible explanation for this unexpected finding is the
relationship between spouse variables. Spouses’ reports of the variables, including their baseline
psychological distress, negative influence attempts, marital measures, and the mental health
follow-up measure were significantly correlated, which may have made examination of their
independent contributions difficult. When controlling for spouses’ baseline psychological
distress in the models, negative influence attempts did not independently predict follow-up
mental health for spouses. However, when removing spouse psychological distress from the
model, negative influence attempts remained a non-significant predictor of spouse mental health
at follow-up. Taken together, the findings demonstrate that spouses’ general mental health is
closely related to their level of psychological distress, and their use of negatively laden influence
attempts does not independently contribute to worse mental health. Furthermore, it appears that
for distressed spouses in this sample, negative mood was relatively stable over the course of the
study and warrants investigation of other factors beyond the spouses’ involvement in the
patients’ health condition that contribute to spouse mental health in this context. The literature on
caregiving spouses of dementia patients has focused on the caregivers’ perceptions of burden of
the illness and caregiving demands, as well as on coping styles of caregivers (e.g., Pinquart &
Sörensen, 2003), which is distinct from the nature of their caregiving behaviors as was examined
in the current study. Future research on couples coping with chronic illness should incorporate
&+!
measures of burden in addition to caregiver behavior as burden may be a more relevant predictor
of psychological distress in caregivers.
The prominence of the patients’ experience also extends beyond health outcomes to
relational outcomes. Based on 14 consecutive days of measurement, patients’ report of receiving
influence was positively associated with their day-to-day feelings of marital closeness when the
influence was characterized as more positive than negative. Again, spouses’ daily marital
closeness was not related to their report of providing influence or the patients’ report of receiving
influence. However, spouses’ daily marital closeness was associated with their perception of the
nature of the influence attempts (i.e. relative positivity to negativity), with greater positive
relative to negative influence being associated with greater feelings of marital closeness for
spouses. The significance and implications of these findings will be discussed in greater detail in
the section on marital consequences.
One divergence from these patient-focused findings emerged in the case of general
physical health. When looking beyond the effects of influence attempts on CHF symptoms to
patient physical health more generally, we did not find associations between influence attempts
and general physical health at follow-up for patients. This finding suggests that the influence
attempts intended to improve compliance with CHF treatment regimen do not result in broader
health effects for patients. While it may be surprising that behaviors related to lower levels of
CHF symptoms would not also be associated with improved general health, several factors may
be contributing to this finding. First, the influence attempts assessed in the current study are very
specific to the CHF treatment regimen and therefore appear to have more of a targeted effect.
Second, the outcome measure used to assess general physical health is a broad measure of health
that is heavily weighted with items measuring functional health. Therefore, it is possible that
&*!
modest improvement in the severity of CHF symptoms does not translate into gains in functional
abilities like moderate to vigorous activities including walking several blocks or lifting heavy
objects.
The findings for general physical health outcomes were the only results demonstrating
consequences of influence attempts for spouses. However, these results were also contrary to
predictions. The hypothesized benefits of positively laden influence attempts like joining with
the patient in maintaining a healthy diet and exercise routine were not only lost on spouses, but
were found to be negatively associated with spouse physical health in the context of low patient
closeness variability. This finding might suggest some level of caregiver burnout, as the more
spouses do to try to improve patient compliance, even when their attempts are positive, the worse
their physical health (Pinquart & Sorensen, 2007). While this finding may suggest that for
spouses doing less is better for their own physical health, this association is only found when
patients report stable marital closeness. Meanwhile, in couples with low patient-reported
stability, any level of positive influence attempt was associated with worse physical health for
spouses. This may indicate that marital stability or other unmeasured variables are more
important predictors of spouse health for these couples.
The Moderating Role of the Marital Context on Associations between Influence and Health
Outcome
The study findings support the notion that the marital context in which influence attempts
are occurring is indeed important to consider when characterizing this process and understanding
the outcomes. However, the results clearly show that how you assess the marital context matters.
Patient and spouse ratings of marital quality, based on a commonly used approach of
characterizing overall marital quality by assessing marital satisfaction and marital
&"!
communication patterns, did not moderate any of the effects of spousal influence on the
outcomes examined in the study, including treatment compliance, physical health, and mental
health. However, a measure that tapped the day-to-day variability in feelings of marital closeness
did moderate the relationship between influence attempts and the outcomes examined.
One possible explanation for the lack of findings with the current measure of marital
quality stems from the nature of the sample. As a group, these couples rated their marital quality
as quite high on average, which may have limited its utility as a moderator. In fact, the findings
suggest that even among highly satisfied couples, negative influence attempts can be detrimental
to physical and mental health outcomes. However, it is also possible that patient and spouse
reports of marital quality were biased in some way. The items comprising the marital quality
measure did not specify a time frame (e.g., In the past month, how satisfied have you been
with…), thus leaving it open for individual interpretation. Participants may have reflected on
their marriage more generally and taken a historical perspective of their relationship, as opposed
to evaluating its current level of functioning. This kind of retrospective report is likely to be
subject to memory biases. There is an extensive literature demonstrating that older adults have a
tendency to emphasize the positive and de-emphasize the negative, particularly when it pertains
to meaningful interpersonal relationships (e.g., Lang & Carstensen, 2002; Story et al., 2007),
which may have resulted in higher ratings of marital quality among this sample of older, long-
married couples.
In contrast, significant results were found when using a measure that captured the within-
person variability in feelings of marital closeness. High variability in marital closeness
experienced by patients served to nullify the positive association between positive influence and
compliance. This finding suggests that the beneficial effect of positive influence for patients is
&#!
dependent on stability in their feelings of marital closeness, whereas the detrimental effect of
negative influence attempts remains regardless of the marital context. This finding again
demonstrates that it is the patients’ variable that is important in the process, as it was patients’
closeness variability that moderated the relationship between influence attempts and treatment
compliance.
So why is the measure of day-to-day variability relevant when the global measure is not?
Proponents of experience sampling research posit that repeated sampling over shorter time
intervals reduces some of the error inherent in retrospective self-reports (Bolger, Davis, Rafaeli,
2003, Laurenceau & Bolger, 2005). It is possible that the daily reports of marital closeness
provided by patients and spouses were a more accurate reflection of the current marital climate
than the baseline assessment of marital quality. The daily measure may allow for more honest
responding, as it may be easier for a person to openly acknowledge a bad day in a relationship
when it is just one day to reflect upon. In contrast, asking a person to reflect broadly on the
quality of arguably one of the most significant relationships of their life may trigger a form of
social desirability turned inward and lead to a biased report in an effort to protect themselves.
Additionally, recent literature suggests important differences in the construct of marital quality
and marital closeness or satisfaction. In particular, scholars are noting that measuring marital
happiness or levels of closeness and cohesion focuses solely on personal satisfaction, whereas
assessments of marital quality capture more of the purpose or functionality of the marital
relationship and are thus quite distinct (Knapp & Lott, 2010).
However, one key difference between the two measures is that the variability measure
captures the stability or instability of marital feelings as opposed to a global, static assessment of
the marriage. Given this difference, the results suggest that for patients with CHF, consistency in
&$!
the marriage is more important for CHF health outcomes than overall quality. However, when
looking at physical and mental health outcomes a more complex pattern emerges. The findings
suggest that perhaps in a more stable marriage, doing more (even when the attempts are positive)
is more like doing too much. And this “doing too much” contributes to worse outcomes for
patients’ mental health and spouses’ physical health. In contrast, in a marriage that is less stable,
it is possible that positive influence attempts at any level become irrelevant, and that other
individual- or couple-level factors unmeasured here come to the fore. An additional
interpretation of these findings is that high levels of positive influence attempts is somewhat
expected or the norm for these highly satisfied couples, as no differences in outcomes were
found at high levels of positive influence attempts. Rather, it is at the lower levels of positive
attempts that you see differences based on instability in closeness, which may be a result of other
individual- or couple-level factors. Future research that samples from a broader range of couples
may help to clarify these questions.
Consequences for the Marital Relationship of Providing and Receiving Spousal Influence
The literature on couples coping with chronic illness has acknowledged that the marital
context is important to consider, and the current study examined this idea specifically. Given the
findings that instability in feelings of marital closeness is consequential for health outcomes, the
logical next step would be to consider how the coping process might affect the marital
relationship. Developmental theories of marriage posit that marital quality is shaped over time by
how couples manage or cope with stressful events (Karney & Bradbury, 1995) and the current
study shows that coping with chronic illness is one such context that can impact feelings about
the marriage. Specifically, in a time-lagged model, receiving influence was positively associated
with next-day feelings of marital closeness for patients, particularly when the influence was
&%!
characterized as more positive than negative. It should be noted that it is the amount of positive
influence relative to negative influence that is important for patient closeness. Patients benefited
from receiving influence the previous day even when the influence received was equally as
positive as it was negative (i.e. ratio of 1:1), and only lose the benefit when the amount of
negative influence is twice that of the positive influence attempts. This suggests that patients
appear to “take the good with the bad” when it comes to how they might interpret spouses’
attempts to influence their treatment compliance. Furthermore, it demonstrates that even among
older patients in marriages that have spanned decades, marital feelings can indeed be affected by
how the couple engages around illness management.
In contrast, spouses’ feelings of marital closeness did not benefit from attempting to
influence the patient. This finding may appear inconsistent with findings of affective and
relational benefits of providing spousal support (e.g., Brown et al, 2003). However much of this
work pertains to spouses providing emotional support during times of acute stress and often is
found when there is reciprocation of support (e.g., Kleiboer, Kuijer, Hox, Schreurs, & Bensing,
2006), which is an aspect of support not addressed in the current project. In the current sample,
spouses’ daily feelings of marital closeness were associated with the nature of their influence
attempts, with greater use of positive-relative-to-negative attempts being positively associated
with marital closeness. While it appeared that spouses’ feelings of marital closeness were not
directly affected by the provision of influence to patients, post-hoc analyses revealed that marital
duration moderated this process for spouses. The additional analyses show that for spouses in
marriages of long duration, providing influence was positively associated with next-day feelings
of marital closeness for spouses. In contrast, for spouses in marriages of shorter duration,
providing influence was negatively associated with next-day feelings of marital closeness for
&&!
spouses. It should be noted that although marital duration was highly correlated with age, the
results for moderation by marital duration were significant while controlling for age in the
analyses. Additionally, age itself did not moderate the relationship between providing influence
and marital closeness.
The contrasting findings for marriages of shorter and longer durations suggest a
developmental process may be at work within these couples. It is possible that spouses who have
been married for 50+ years have had more experience in coping with stressors and thus have
more confidence in their attempts to assist the patient compared to spouses in marriages of 20
years. Its also possible that coping with a chronic and life threatening illness like CHF is more
normative or more easily accepted for these spouses of 50+ years, where as spouses of 20 years
may not have anticipated such an experience at this stage in their marriage. However, more
research is needed examining the developmental stages of marriage and how they interact with
the coping process.
Implications for Intervention
The study findings offer clinical implications for healthcare providers to intervene with
couples who are coping with CHF. First and foremost, healthcare providers should be aware of
the finding that it is the patients’ perspective of how they are treated with regard to their
compliance, as well as their level of treatment compliance that is associated with patients’
physical and mental health outcomes. Second, given the contrasting results between positive and
negative influence attempts, healthcare providers can role model for spousal caregivers how to
speak to patients about treatment compliance. By demonstrating an emphasis on positive
influence attempts over negative ones, healthcare providers can offer caregivers an approach to
being involved in helping the patient with illness management that is associated with positive
&'!
outcomes for both patients and the caregiving spouses. The findings also suggest that negative
and punitive tactics are not only ineffective for patients with CHF, but are detrimental for patient
outcomes. As such, patients who are struggling to adhere to the treatment regimen may benefit
more from motivational interviewing approaches to dealing with the challenges to adhering to
the treatment regimen. Motivational interviewing is a non-judgmental and non-adversarial
approach that attempts to (1) increase the patient’s awareness of the potential problems,
consequences, and risks faced as a result of the behavior in question and (2) help patients think
differently about their behavior and ultimately to consider what might be gained through change
(Miller & Rollnick, 1991). Physicians could model this strategy for spouses and provide an
example of an alternative technique for helping the patient improve adherence. Recent research
has demonstrated the effectiveness of motivational interviewing for increasing CHF patients
readiness for physical activity (Brodie, Inoue, & Shaw, 2008). Also, patients and spouses may
benefit from providers engaging them in a conversation with the goal of identifying what patients
feel would be most helpful and encouraging for them in their efforts to comply. Providers may
also want to attend to the couples’ marital relationship if they see signs of distress, as feelings
about the marital relationship are consequential for patient outcomes. Lastly, caregiving spouses’
should be monitored for elevated levels of psychological distress, as this is associated with
negative outcomes for spouses’ physical and mental health. Spouses’ health and well being
should not be neglected, given the important and influential role they play in CHF illness
management and subsequent patient outcomes.
Limitations
The current study has certain limitations that should be noted and used to inform future
research. First, the longitudinal nature of the current study provides stronger support for
&(!
inferences about causality compared to cross-sectional studies, but nonetheless caution should be
used when drawing conclusions about causality as the length of the study was relatively short.
Second, patients in the sample varied rather widely in time since diagnosis (newly diagnosed –
10 years since diagnosis), which may have resulted in a muddying of effects. Additionally, given
the long term nature of these marriages and the years the couples had already spent coping with
this illness, life span influences of the individual- and couple-level development are likely
important in understanding the findings. Future work in this area may benefit from enrolling
study participants more uniformly at time of diagnosis or in the early stages of illness to get a
clearer picture of how spousal involvement influences patient behavior and outcomes. Third, the
examination of impact of spousal influence attempts on the marriage was limited to a narrow
snapshot of daily feelings of marital closeness and may not translate to longer-term changes in
marital quality. The current study would have benefited from a repeat of the baseline assessment
of marital quality at the six-month follow-up to see if the effects seen in the shorter time scale of
the diary study would manifest in the broader measure of marital quality. Fourth, the
generalizability of these findings may be somewhat limited, as the participants represented a
rather homogeneous convenience sample composed of patients and spouses who presented to a
university cardiology clinic. Lastly, although post-hoc analyses failed to find gender differences
in the findings, it is possible that the analyses lacked enough statistical power to identify
differences as the sample was comprised mostly of male patients. Future work in this area would
benefit from attempts to sample more diversely with regard to gender, as well as other
demographic factors to improve generalizability and identification of potential group differences
for targeted interventions.
&)!
While this study has certain limitations, it makes a number of important contributions to
the literature on couples coping with chronic illness. First, the exclusive focus on patient
outcomes in the existing literature was broadened by investigating the health consequences of
engaging in health-related influence attempts for spouses. Second, while the existing literature
had suggested that marital quality could influence this process, past research has most often
controlled for the effects of marital quality. The current study directly examined the potential
moderating role of marital quality in couples’ efforts to manage CHF using a baseline measure of
marital quality as well as a more novel, repeated measure of variability in marital closeness.
Third, this study shed light on the potential long term effects that coping with a chronic illness
may have on the marital relationship itself by examining the effect of providing and receiving
health-related spousal influence on patients’ and spouses’ day-to-day feelings of marital
closeness. Lastly, this study is the first to my knowledge to examine coping with chronic illness
using the Actor Partner Interdependence Model, which bears relevance in terms of the relative
importance and influence of each partners’ perspective of the coping process and subsequent
outcomes.
Conclusion
In sum, the findings from this study make a number of important contributions to the
literature on couples coping with chronic illness. First and foremost, spousal involvement in
CHF illness management is consequential both for patient and spouses, suggesting that engaging
both partners in the coping process and assessing support behaviors is important for improving
illness-specific and broader health outcomes. Second, while coping with chronic illness is indeed
a relational process, the patients’ perspective of illness specific issues is paramount and should
be treated as such. Third, dyadic coping occurs within a marital context. Improving our
'+!
understanding of marital functioning, both how it affects coping and how it is affected by coping,
is important for helping couples that are managing chronic illness.

'*!
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions.
Newbury Park, London, Sage.
Anderson, S. A., Russell, C. S., & Schumm, W. A. (1983). Perceived marital quality and family
life-cycle categories: A further analysis. Journal of Marriage and the Family, 45, 481-
492.
Ashmore, J. A., Emery, C. F., Hauck, E. R., & MacIntyre, N. R. (2005). Marital adjustment
among patients with chronic obstructive pulmonary disease who are participating in
pulmonary rehabilitation. Heart & Lung, 34, 270-278, doi:10.1016/j.hrtlng.2004.12.005
August, K. J., Rook, K.S., Stephens, M. A., P., & Franks, M. M. (2011). Are spouses of
chronically ill partners burdened by exerting health-related social control? Journal of
Health Psychology, 16, 1109-1119.
Berg, C. A., & Upchurch, R. (2007). A developmental-contextual model of couples coping with
chronic illness across the adult life span. Psychology and Aging, 133, 920-954.
Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived.
Annual Review of Psychology, 54, 579-616.
Bolger, N., Zuckerman, A., Kessler, R. C. (2000). Invisible support and adjustment to stress.
Journal of Personality and Social Psychology, 79, 953-961.
Bookwala, J. (2005). The role of marital quality in physical health during the mature years.
Journal of Aging and Health, 17, 85-104.
Bookwala, J., Yee, J.L., & Schulz, R. (2000). Caregiving and detrimental mental and physical
health outcomes. In G.M. Williamson, D.R. Shaffer & P.A. Parmelee (Eds.), Physical
illness and depression in older adults: A handbook of theory, research, and practice. (pp.
'"!
93–131). New York: Kluwer Academic/Plenum Publishers.
Brock, R. L., & Lawrence, E. (2009). Too much of a good thing: Underprovision versus
overprovision of partner support. Journal of Family Psychology, 23, 181-192.
Brodie, D. A.. Inoue, A., & Shaw, D. G. (2008). Motivational interviewing to change quality of
life for people with chronic heart failure: A randomised controlled trial. International
Journal of Nursing Studies, 45, 489–500.
Brown, S. L., Nesse, R. M., Vinokur, A. D., & Smith, D. M. (2003). Providing social support
may be more beneficial than receiving it: Results from a prospective study of mortality.
Psychological Science, 14, 320-327.
Burman, B., & Margolin, G. (1992). Analysis of the association between marital relationships
and health problems: An interactional perspective. Psychological Bulletin, 112, 39-63.
Carney, R. M., Freedland, K. E., Rich, M. W., & Jaffe, A. S. (1995). Depression as a risk factor
for cardiac events in established coronary heart disease: A review of possible mechanisms.
Annals of Behavioral Medicine, 17, 142–149.
Carstensen, L. L. (1993). Motivation for social contact across the life span: A theory of
socioemotional selectivity. In J. Jacobs (Ed.), Nebraska Symposium on Motivation (Vol. 40,
pp. 209-254). Lincoln: University of Nebraska Press.
Christensen, A., & Arrington, A. (1987). Research issues and strategies. In T. Jacob (Ed.),
Family interaction and psychopathology: Theories, methods, and findings (pp. 259-296).
New York: Plenum Press.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.) Hillsdale, NJ:
Erlbaum.
'#!
Cohen, S., & Lichtenstein, E. (1990). Partner behaviors that support quitting. Journal of
Consulting and Clinical Psychology, 58, 304-309.
Cook, W. L., & Kenny, D. A. (2005). The actor-partner interdependence model: A model of
bidirectional effects in developmental studies. International Journal of Behavioral
Development, 29, 101-109.
Coyne, J. C., & Fiske, V. (1992). Couples coping with chronic and catastrophic illness. In T. J.
Akamatsu, M. A. P. Stevens, S. E. Hobfoll, & J. H. Crawther (Eds.), Family health
psychology (pp. 129-149). Washington, DC: Hemisphere Publishing Corp.
Coyne, J. C., Rohrbaugh, M. J., Shoham, V., Sonnega, J. S., Nicklas, J. M., & Cranford, J. A.
(2001). Prognostic importance of marital quality for survival of congestive heart failure.
American Journal of Cardiology, 88, 526-529.
Coyne, J. C., & Smith, D. A. F. (1991). Couples coping with a myocardial infarction: A
contextual perspective on wives’ distress. Journal of Personality and Social Psychology,
61, 404-412.
Coyne, J. C., & Smith, D. (1994). Couples coping with a myocardial infarction: contextual
perspective on patient self-efficacy. Journal of Family Psychology, 8(1), 43-54.
Cutrona, C. E., Russell, D. W., & Gardner, K. A. (2005). The relationship enhancement model of
social support. In T. A. Revenson, K. Kayser, & G. Bodenmann (Eds.), Couples coping
with stress (pp. 3-23). Washington, DC: American Psychological Association.
Dawson, J. F., & Richter, A. W. (2006). Probing three-way interactions in moderated multiple
regression: Development and application of a slope difference test. Journal of Applied
Psychology, 91, 917 – 926.
'$!
DiMatteo, M. R. (2004). Social support and patient adherence to medical treatment: A meta-
analysis. Health Psychology, 23, 207-218, doi:10.1037/0278-6133.23.2.207
Dunbar-Jacob, J., Schlenk, E., Baum, A., Revenson, T. A., & Singler, J. E. (2001). Handbook of
health psychology. Mahwah, NJ: Erlbaum.
Fekete, E. M., Stephens, M. A. P., Druley, J. A., & Greene, K. A. (2006). Effects of spousal
control and support on older adults’ recovery from knee surgery. Journal of Family
Psychology, 20, 302-310.
Franks, M. M., Stephens, M. A. P., Rook, K. S., Franklin, B. A., Keteyian, S. J., & Artinian, N.
T. (2006). Spouses’ provision of health-related support and control to patients
participating in cardiac rehabilitation. Journal of Family Psychology, 20, 311-318.
Franks, M. M., Wendrof, C. A., Gonzalez, R., & Ketterer, M. (2004). Aid and influence: Health-
promoting exchanges of older married partners. Journal of Social and Personal
Relationships, 21, 431-445, doi:10.1177/0265407504044839
Gallant, M. P., Spitze, G. D., & Prohaska, T. R. (2007). Help or hindrance? How family and
friends influence chronic illness self-management among older adults. Research on
Aging, 29, 375-409, doi:10.1177/0164027507303169
Garcia, R. L., Kenny, D. A., & Ledermann, T. (2012). Moderation in the actor-partner
interdependence model. Unpublished paper, University of Connecticut.
Gleason, M. E. J., Iida, M., Shrout, P. E., & Bolger, N. (2008). Receiving support as a mixed
blessing: Evidence for dual effects of support on psychological outcomes. Journal of
Personality and Social Psychology, 94, 824-838, doi:10.1037/0022-3514.94.5.824
Guilford, R., & Bengston, V. (1979). Measuring marital satisfaction in three generations:
Positive and negative dimension. Journal of Marriage and the Family, 41, 387-398.
'%!
Hagedoorn, M., Dagan, M., Puterman, E., Hoff, C., Jeroen Meijerink, W. J. H., DeLongis, A., &
Sanderman, R. (2011). Relationship satisfaction in couples confronted with colorectal
cancer: The interplay of past and current spousal support. Journal of Behavioral
Medicine, 34, 288-297, doi:10.1007/s10865-010-9311-7.
Helgeson, V. S., Novak, S. A., Lepore, S. J., & Eton, D. T. (2004). Spouse social control efforts:
Relations to health behavior and well-being among men with prostate cancer. Journal of
Social and Personal Relationships, 21, 53-68.
Heavey, C. L., Larson, B. M., Zumtobel, D. C., & Christensen, A. (1996). The communication
patterns questionnaire: The reliability and validity of a constructive communication
subscale. Journal of Marriage and Family, 58, 796-800.
Hendrick, S. S. (1988). A generic measure of relationship satisfaction. Journal of Marriage and
the Family, 50, 93-98.
Heshbacher, P. K., Downing, R. W., & Stephansky, P. (1978). Assessment of psychiatric illness
severity by family physicians. Social Science and Medicine, 12, 45-52.
Hough, R. L., Landsverk, J. A., Stone, J. D., & Jacobson, G. F. (1982). Comparison of
psychiatric screening questionnaires from primary care patients. Final report for NIMH
Contract No. 278-81-0036.
Hunt, S. A., Baker, D. W., Chin, M. H., Cinquegrani, M. P.,

Feldman, A. M., Francis, G. S., et al.
(2001). ACC/AHA

guidelines for the evaluation and management of chronic heart

failure
in the adult: Executive summary (Committee to Revise the 1995 Guidelines

for the
Evaluation and Management of Heart Failure). Journal of the American College of

Cardiology, 38, 2101-2113.
'&!
Johnson, N. J., Backlund, E., Sorlie, P. D., & Loveless, C. A. (2000). Marital status and
mortality: The national longitudinal mortality study. Annals of Epidemiology, 10, 224-
238.
Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability:
A review of theory, method, and research. Psychological Bulletin, 118, 3-34.
Kiecolt-Glaser, J. K., & Newton, T. L. (2001). Marriage and health: His and hers. Psychological
Bulletin, 27, 472-503.
Kenny, D. A., & Cook, W. (1999). Partner effects in relationship research: Conceptual issues,
analytic difficulties, and illustrations. , 433-448.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic Data Analysis. New York: Guilford
Press.
Kenny, D. A., & Kashy, D. A. (2010). Dyadic data analysis using multilevel modeling. In J. Hox
& J. K. Roberts (Eds.), The handbook of multilevel analysis, pp. 335-370. London: Taylor
Francis.
Kenny, D. A., West, T. V., & Garcia, R. (2012). Dyadic analysis using mulitlevel modeling:
Data Analysis Training Institute of Connecticut. Storrs, Connecticut: University of
Connecticut, Department of Psychology.
Khan, C. M. (2010). Spousal support and control targeting exercise in older adults with
diabetes: Roles of patients’ emotional responses and gender. Retrieved from
Dissertations and Theses database. (AAT 3429231)
Kiecolt-Glaser, J. K., & Newton, T. L. (2001). Marriage and health: His and hers. Psychological
Bulletin, 27, 472-503.
Kleiboer, A. M., Kuijer, R. G., Hox, J. J., Schreurs, K. M. G., & Bensing, J. M. (2006).
''!
Receiving and providing support in couples dealing with multiple sclerosis: A diary study
using an equity perspective. Personal Relationships, 13, 485-501.
Knapp, S. J., & Lott, B. (2010). Forming the central framework for a science of marital quality:
An interpretive alternative to marital satisfaction as a proxy for marital quality. Journal of
Family Theory & Review, 2, 316-333.
Krull, J. L. & MacKinnon, D. P. (2001). Multilevel modeling of individual and group
level mediated effects. Multivariate Behavioral Research, 36, 249-277.
Lang, F. R., & Carstensen, L. L. (2002). Time counts: Future time perspective, goals, and social
relationships. Psychology and Aging, 17, 125-139.
Laurenceau, J. P., & Bolger, N. (2005). Using diary methods to study marital and family
processes. Journal of Family Psychology, 19, 86-97, doi:10.1037/0893-3200.19.1.86
Lavela, S. L., & Ather, N. (2010). Psychological health in older adult spousal caregivers of older
adults. Chronic Illness, 6, 67-80.
Lewis, M. A., & Rook, K. S. (1999). Social control in personal relationships: Impact of health
behaviors and psychological distress. Health Psychology, 18, 63-71.
Liang, J., Krause, N. M., & Bennett, J. M. (2001). Social exchange and well-being: Is giving
better than receiving. Psychology and Aging, 16, 511-523.
Lloyd-Jones, D., Adams, R., Carnethon, M., De Simone, G., Ferguson, T. B., Flegal, K., et al.
(2009). Heart disease and stroke statistics 2009 update: A report from the American Heart
Association Statistics Committee and Stroke Statistics Subcommittee. Circulation, 119,
480-486.
Luke, D. A. (2004). Multilevel Modeling. Newbury Park: Sage Publications.
'(!
MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annual Review of
Psychology, 58, 593-614.
Maisel, N. C., & Gable, S. L. (2009). The paradox of received social support: The importance of
responsiveness. Psychological Science, 20, 928-932.
Manzoli, L., Villari, P., Pirone, G. M., & Boccia, A. (2007). Marital status and mortality in the
elderly: A systematic review and meta-analysis. Social Science & Medicine, 64, 77-94.
McCall, D. (1995). Epidemiology, etiology, and natural history. In D. McCall & S. H.
Rahimtoola (Eds.). Heart Failure (pp. 1-13). New York: Chapman & Hall.
McHorney, C., Ware, J. E., Jr., & Raczek, A. E. (1993). The MOS 36-item short-form health
survey (SF-26): II. Psychometric and clinical tests of validity in measuring physical and
mental health constructs. Medical Care, 31, 247-263.
Miller, W. R., & Rollnick, S. (1991). Motivational interviewing: Preparing people to
change addictive behavior. New York: Guilford Press.
Nesselroade, J. R. (1991). The warp and woof of the developmental fabric. In R. Downs, L.
Liben, & D. S. Palermo (Eds.), Visions of aesthetics, the environment, and development:
The legacy of Joachim F. Wohlwill (pp. 213-240). Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
New York Heart Association Criteria Committee. (1994). Nomenclature and criteria for
diagnosis of diseases of the heart and great vessels, 9
th
Ed. Boston: Little, Brown & Co.
Orbuch, T. L., House, J. S., Mero, R. P., & Webster, P. S. (1996). Marital quality over the life
course. Social Psychology Quarterly, 59, 162-171.
')!
Pinquart, M. (2003). Loneliness in married, widowed, divorced, and never-married older adults.
Journal of Social and Personal Relationships, 20, 31-53.
Pinquart, M., & Sörensen, S. (2003). Differences between caregivers and noncaregivers in
psychological health and physical health: A meta-analysis. Psychology and Aging, 18, 250-
267.
Pinquart, M., & Sörensen, S. (2007). Correlates of physical health of informal caregivers: A
meta-analysis. Journal of Gerontology: Psychological Sciences, 62, 126-137.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data
analysis methods. Thousand Oaks, CA: Sage.
Revenson, T. A. (1994). Social support and marital coping with chronic illness. Annals of
Behavioral Medicine, 16, 122-130.
Revenson, T. A. (2003). Scenes from a marriage: Examining support, coping, and gender within
the context of chronic illness. In J. Suls & K. A. Wallston (Eds.), Social psychological
foundations of health and illness (pp. 530-559). Malden, MA: Blackwell Publishing.
Rohrbaugh, M. J., Mehl, M. R., Shoham, V., Reilly, E. S., & Ewy, G. A. (2008). Journal of
Consulting and Clinical Psychology, 76, 781-789, doi:10.1037/a0013238
Rohrbaugh, M. J., Shoham, V., Cleary, A. A., Berman, J. S., & Ewy, G. A. (2009). Health
consequences of partner distress in couples coping with heart failure. Heart & Lung, 38,
298-305.
Rohrbaugh, M. J., Shoham, V., & Coyne, J. C. (2006). Effect of marital quality on eight-year
survival of patients with heart failure. The American Journal of Cardiology, 98, 1069-
1072.
(+!
Rohrbaugh, M. J., Shoham, V., Coyne, J. C., Cranford, J. A., Sonnega, J. S., & Nicklas, J. M.
(2004). Beyond the “self” in self-efficacy: Spouse confidence predicts patient survival
following heart failure. Journal of Family Psychology, 18, 184-193, doi:10.1037/0893-
3200.18.1.184
Rohrbaugh, M. J., Shoham, V., Cranford, J. A., Niklas, J. M., Sonnega, J. S., & Coyne, J. C.
(2002). Couples coping with congestive heart failure: Role and gender differences in
psychological distress. Journal of Family Psychology, 16, 3-13.
Setoguchi, S., & Stevenson, L.W. (2009). Hospitalizations in patients with heart failure: Who
and why. Journal of the American College of Cardiology, 54, 1703-1705,
doi:10.1016/j.jacc.2009.08.015
Shrout, P. E., Herman, C. M., & Bolger, N. (2006). The costs and benefits of practical and
emotional support on adjustment: A daily diary study of couples experiencing acute
stress. Personal Relationships, 13, 115-134.
Siegler, I. C., Bosworth, H. B., & Poon, L. W. (2003). Disease, health, and aging. In R. Lerner &
A. Easterbrooks (Eds.), Handbook of psychology: Developmental psychology (Vol. 6, pp.
423-442). New York: Wiley.
Snijders, T. A. B., & Bosker, R. J. (2012). Multilevel Analysis: An introduction to basic and
advanced multilevel modeling. London: Sage.
Stephens, M. A. P., Rook, K. S., Franks, M. M., Khan, C., & Iida, M. (2010). Spouses use of
social control to improve diabetic patients’ dietary adherence. Families, Systems, &
Health, 28, 199-208.
Stephens, M. A. P., Fekete, E. M., Franks, M. M., Rook, K. S., Druley, J. A., & Greene, K.
(2009). Spouses’ use of pressure and persuasion to promote osteoarthritis patients’
(*!
medical adherence after orthopedic surgery. Health Psychology, 28, 48-55,
doi:10.1037/a0012385
Story, T. N., Berg, C. A., Smith, T. W., Beveridge, R., Henry, N. J. M., & Pearce, G. (2007).
Age, marital satisfaction, and optimism as predictors of positive sentiment override in
middle-aged and older married couples. Psychology and Aging, 22, 719-727.
Trief, P. M., Morin, P. C., Izquierdo, R., Teresi, J., Starren, J., Shea, S., Weinstock, R. S. (2006).
Marital quality and diabetes outcomes: The IDEATel project. (2006). Families, Systems,
& Health, 24, 318-331, doi:10.1037/1091-7527.24.3.318
Trief, P. M., Ploutz-Snyder, R., Britton, K. D., & Weinstock, R. S. (2004). The relationship
between marital quality and adherence to the diabetes care regimen. Annals of Behavioral
Medicine, 27, 148-154, doi:10.1207/s15324796abm2703_2
Tucker, J. S. (2002). Health-related social control within older adults’ relationships. Journal of
Gerontology: Psychological Sciences, 57B, 387-395.
Tucker, J. S., & Anders, S. L. (2001). Social control and health behaviors in marriage. Journal of
Applied Social Psychology, 31, 467-485.
Tucker, J. S., & Mueller, J. S. (2000). Spouses’ social control of health behaviors: Use and
effectiveness of specific strategies. Personality and Social Psychology Bulletin, 26, 1120-
1130.
Tucker, J. S., Orlando, M., Elliott, M. N., Klein, D. J. (2006). Affective and behavioral responses
to health-related social control. Health Psychology, 25, 715-722.
Umberson, D. (1992). Gender, marital status and the social control of health behavior. Social
Science Medicine, 14, 907-917.
("!
Umberson, D. (1987). Family status and health behaviors: Social control as a dimension of social
integration. Journal of Health and Social Behavior, 28, 306-319.
Unger, D. G., Jacobs, S. B., & Cannon, C. Social support and marital satisfaction among couples
coping with chronic constructive airway disease. Journal of Social and Personal
Relationships, 13, 123-142, doi:10.1177/0265407596131007
Ware, J. E. (2000). SF-36 Health survey update. Spine, 25, 3130-3139.
Weishaus, S., & Field, D. (1988). A half century of marriage: Continuity or change? Journal of
Marriage and the Family, 50, 763-774.
West, T. V. (2012). Repeated measures with dyads. In J. A. Simpson & L. Campbell (Ed.), The
Oxford handbook of close relationships. New York: Oxford University Press.
Whitton, S. W., & Whisman, M. A. (2010). Relationship satisfaction instability and depression.
Journal of Family Psychology, 24, 791-794, doi:10.1037/a0021734
Williamson, G. M., & Shaffer, D. R. (2001). Relationship quality and potentially harmful
behaviors by spousal caregivers: How we were then, how we are now. Psychology and
Aging, 16, 217-226.
(#!
Footnote

1
Despite having the published data to support the exclusive use of spouse psychological
distress at baseline as a covariate in the current analyses, models were tested with patient distress
as well as spouse distress to ensure accuracy of results. Inclusion of patients’ baseline
psychological distress did not alter the pattern or significance of results, so it was dropped from
the analyses.

2
Given the lack of findings of moderation using the composite measure of marital
quality, moderation analyses were conducted again using the marital satisfaction measure (RAS)
to assess whether the lack of findings were due to problems with the composite measure. The
marital satisfaction component of the composite measure was tested because it is most closely
associated with the daily rating of marital closeness and its variability, which was shown to
moderate some of the tested associations. However, no change in the pattern of results was found
when using marital satisfaction as a moderator in the analyses.
($!
Table 1. Descriptive Summary of Patients

Characteristics Mean (SD) or n (%)
Gender (male) 43 (72%)
Age (years) 67 (11.68)
Ethnicity (white) 51 (85%)
Education (college graduate) 41 (43%)
Length of marriage (years) 34.76 (16.69)
Left ventricular ejection fraction (%) 29.1 (8.7)
NYHA Class III/IV 20 (33%)
Time since diagnosis (years) 4.8 (5.1)
History of hypertension 15 (25%)
History of diabetes 19 (32%)
History of myocardial infarction 25 (42%)
Emergency room visit in previous 6 months 25 (42%)
Note. SD = standard deviation; NYHA = New York Heart Association

85
!
Table 2. Within-Partner Pearson Correlations for Study Variables
Note. Correlations for patients’ report of variables are reported below the diagonal; correlations for spouses’ report of variables are reported above the diagonal;
correlations for the same measure between patients and spouses are reported on the diagonal ; BL – Baseline; FU – Follow-up; Pt – Patient; Pos – Positive; Neg –
Negative; CHF symptom measure is scaled so that higher numbers indicate more symptoms; Physical and mental health measures are scaled so that higher
numbers indicate better health.
*p < .05; **p < .01; ***p < .001

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13
1. Illness Severity 1 -.20 -.16 -.01 .24 -.27* -.04 -.01 .13 .00 .28* .21 -.02
2. Illness Duration -.20 1 .23 .16 .12 -.05 .16 -.04 -.08 -.08 .07 -.03 -.27*
3. Age -.26* .35** .90*** .59*** -.10 .23 -.11 .27* -.16 .06 -.03 -.04 .01
4. Marital Duration -.01 .16 .57*** 1 .01 .02 -.19 -.08 -.12 -.07 .00 -.01 -.08
5. BL Psychological Distress .51*** .01 -.25 .00 .40** -.27* .19 .14 .29* .22 .37** -.22 -.45**
6. BL Marital Quality -.13 .04 .04 -.11 -.39** .58*** -.29* .20 -.28* .12 -.26 -.11 .11
7. Variability Marital Closeness .17 .01 .01 .01 .01 -.08 .28* -.13 .42** -.07 .07 -.13 -.17
8. Pos Spousal Influence -.12 -.10 .17 -.13 -.09 .27* -.04 .44*** .11 .29* .23 -.08 -.13
9. Neg Spousal Influence -.10 .05 .00 -.16 .16 .02 .10 -.23 .33* -.13 .09 .01 -.27*
10. FU Pt Compliance -.06 -.11 .12 .12 -.13 .13 -.02 .15 -.29* .66*** -.20 .09 .04
11. FU Pt CHF Symptoms .40** .10 -.26* -.01 .59*** -.31* .22 -.09 .14 -.27* .73*** -.12 -.22
12. FU General Physical Health -.47*** -.07 .24 .23 -.59*** .16 -.27* .05 -.19 .22 -.61*** .01 .61***
13. FU General Mental Health -.49*** -.13 .22 .21 -.62*** .24 -.27* .04 -.31* .19 -.67*** .78*** .22
!"#
Table 3. Between-partner correlations for study variables

Note. BL – Baseline; FU – Follow-up; Pt – Patient; Sp – Spouse; Pos – Positive; Neg – Negative; CHF symptom measure is scaled so that higher numbers
indicate more symptoms; Physical and mental health measures are scaled so that higher numbers indicate better health.
*p < .05; **p < .01; ***p < .001

Variable 1-Sp 2-Sp 3-Sp 4-Sp 5-Sp 6-Sp 7-Sp 8-Sp 9-Sp 10-Sp
1. Age-Pt
2. BL Psychological Distress-Pt -.21
3. BL Marital Quality-Pt .05 -.08
4. Variability Marital Closeness-Pt -.08 .34** -.16
5. Pos Spousal Influence-Pt .16 .07 .32* -.10
6. Neg Spousal Influence-Pt -.01 .10 .13 .17 .05
7. FU Pt Compliance-Pt -.09 .18 .08 -.22 .09 -.22
8. FU Pt CHF Symptoms-Pt -.17 .38** -.18 .04 .08 .04 -.10
9. FU General Physical Health-Pt -.18 -.50*** .18 -.20 -.11 -.06 .05 -.46***
10. FU General Mental Health-Pt -.17 -.48*** .23 -.19 -.06 -.08 .01 -.47*** .04
Variable 1-Pt 2-Pt 3-Pt 4-Pt 5-Pt 6-Pt 7-Pt 8-Pt 9-Pt 10-Pt
1. Age-Sp
2. BL Psychological Distress-Sp -.10
3. BL Marital Quality-Sp .18 -.31*
4. Variability Marital Closeness-Sp -.06 .13 -.23
5. Pos Spousal Influence-Sp .24 .02 .22 -.32*
6. Neg Spousal Influence-Sp -.21 .04 -.02 .08 -.04
7. FU Pt Compliance-Sp .06 .03 .23 -.01 .45** .00
8. FU Pt CHF Symptoms-Sp -.06 .43*** -.36** .15 -.13 .05 -.22
9. FU General Physical Health-Sp .00 .02 .01 -.25 .11 -.13 .04 -.13
10. FU General Mental Health-Sp .03 -.20 .06 -.19 .18 -.17 -.06 -.21 .04
87
!
Table 4. Comparison between patients and spouses on demographic and study variables
Characteristics Patients
Mean (SD)
Spouses
Mean (SD)
Paired t P value
Age (years) 67.0 (11.68) 65.58 (10.73) 2.70* .031
BL Psychological Distress (HSCL) 43.10 (11.11) 39.40 (9.88) 2.47* .016
BL Marital Quality 0.01 (0.81) 0.01 (0.79) -.016 .987
Variability in Marital Closeness 1.01 (0.67) 1.08 (0.73) -.613 .543
Positive Spousal Influence
17.98 (5.29) 16.11 (5.32) 2.61* .011
Negative Spousal Influence
15.29 (4.63) 13.04 (4.43) 3.12** .003
FU Patient Compliance
3.75 (0.58) 3.68 (0.78) 1.00 .322
FU Patient CHF Symptoms
1.41 (0.31) 1.44 (0.37) -1.03 .306
FU General Physical Health
49.21 (24.22) 64.62 (25.51) -3.10** .003
FU General Mental Health
67.01 (21.13) 74.34 (19.55) -1.81 .077
Daily Spousal Influence
4.76 (2.59) 4.50 (2.79) -1.76 .080
Ratio Pos:Neg Spousal Influence
1.32 (0.71) 1.39 (0.60) -.239 .812
Daily Marital Closeness
7.48 (2.16) 7.97 (1.94) -1.70 .096
Note. SD = standard deviation; BL = baseline measure; FU = follow-up measure; HSCL = Hopkins Symptom
Checklist. Variables below the line are used for the analysis of Aim 4. Daily spouse influence and daily marital
closeness are from the daily diary data; means reflect the sample mean of each participants’ mean of the 14 days of
measurement and are only used for descriptive purposes.
!!"
Table 5. APIM of the effect of spousal influence attempts on treatment compliance at follow-up
and with moderation by baseline marital quality (Aims 1 and 3).

No Moderation
Moderation by
Actor Marital
Quality
Moderation by
Partner Marital
Quality
APIM Parameters Estimate p Estimate p Estimate p
Covariates
Illness Severity ! Pt Comply -.068 .494 -.073 .503 -.039 .712
Spouse Distress ! Pt Comply .017+ .052 .016 .083 .020* .036
Illness Severity ! Sp Comply .021 .875 -.001 .996 -.006 .772
Spouse Distress ! Sp Comply .022 .071 .019 .083 .026* .046
Actor effects
Pt Pos Infl ! Pt Comply .033 .060 .033 .095 .032 .096
Pt Neg Infl ! Pt Comply -.039* .038 -.038+ .052 -.041* .038
Sp Pos Infl ! Sp Comply -.002 .923 .002 .932 -.006 .813
Sp Neg Infl ! Sp Comply -.043 .120 -.039 .214 -.047 .111
Partner effects
Pt Pos Infl ! Sp Comply .065** .007 .059* .028 .059* .028
Pt Neg Infl ! Sp Comply -.003 .911 -.006 .820 -.006 .820
Sp Pos Infl ! Pt Comply -.013 .447 -.012 .516 -.019 .308
Sp Neg Infl ! Pt Comply -.030 .121 -.029 .121 -.029 .121
Main effects of Moderator Variable
Pt Marital Quality ! Pt Comply -- -- -.023 .788 -- --
Sp Marital Quality !Sp Comply -- -- -.087 .519 -- --
Interaction effects
Pt Marital Quality x Pt Pos Infl ! Pt Comply -- -- .010 .623 -- --
Pt Marital Quality x Pt Neg Infl ! Pt Comply -- -- -.012 .553 -- --
Pt Marital Quality x Sp Pos Infl ! Pt Comply -- -- -.005 .780 -- --
Pt Marital Quality x Sp Neg Infl ! Pt Comply -- -- .003 .902 -- --
Sp Marital Quality x Sp Pos Infl ! Sp Comply -- -- .009 .788 -- --
Sp Marital Quality x Sp Neg Infl ! Sp Comply -- -- -.015 .601 -- --
Sp Marital Quality x Pt Pos Infl ! Sp Comply -- -- .014 .674 -- --
Sp Marital Quality x Pt Neg Infl ! Sp Comply -- -- .014 .611 -- --
Main effects of Moderator Variable
Pt Marital Quality ! Sp Comply -- -- -- -- .195 .148
Sp Marital Quality !Pt Comply -- -- -- -- .134 .175
Interaction effects
Pt Marital Quality x Pt Pos Infl ! Sp Comply -- -- -- -- .002 .578
Pt Marital Quality x Pt Neg Infl ! Sp Comply -- -- -- -- .000 .998
Pt Marital Quality x Sp Pos Infl ! Sp Comply -- -- -- -- .014 .578
Pt Marital Quality x Sp Neg Infl ! Sp Comply -- -- -- -- .000 .989
Sp Marital Quality x Sp Pos Infl ! Pt Comply -- -- -- -- -.012 .537
Sp Marital Quality x Sp Neg Infl ! Pt Comply -- -- -- -- .008 .680
Sp Marital Quality x Pt Pos Infl ! Pt Comply -- -- -- -- .002 .937
Sp Marital Quality x Pt Neg Infl ! Pt Comply -- -- -- -- -.010 .592
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos Infl
= positive influence attempts; Neg Infl = negative influence attempts; Comply = patient treatment compliance at follow-up.
* p < .05; ** p < .01; + p < .06.
!"#
Table 6. APIM of the effect of spousal influence attempts on treatment compliance at follow-up
and with moderation by closeness variability (Aim 3 continued).

Moderation by Actor
Closeness Variability
Moderation by Partner
Closeness Variability
APIM Parameters Estimate p Estimate p
Covariates
Illness Severity ! Pt Comply -.011 .913 -.077 .436
Spouse Distress ! Pt Comply .019* .048 .025* .010
Pt Marital Quality ! Pt Comply -.002 .988 -.022 .860
Sp Marital Quality ! Pt Comply .133 .314 .074 .586
Illness Severity ! Sp Comply .081 .561 .024 .873
Spouse Distress ! Sp Comply .028* .049 .024 .090
Sp Marital Quality ! Sp Comply -.047 .808 -.110 .568
Pt Marital Quality ! Sp Comply .268 .122 .212 .213
Actor effects
Pt Pos Infl ! Pt Comply .022 .225 .028 .123
Pt Neg Infl ! Pt Comply -.063* .033 -.044* .027
Sp Pos Infl ! Sp Comply -.004 .887 -.005 .870
Sp Neg Infl ! Sp Comply -.091** .008 -.050 .111
Partner effects
Pt Pos Infl ! Sp Comply .049* .048 .065* .018
Pt Neg Infl ! Sp Comply .015 .577 .032 .232
Sp Pos Infl ! Pt Comply -.021 .328 -.023 .238
Sp Neg Infl ! Pt Comply -.029 .121 -.019 .352
Main effects of Moderator Variable
Pt Closeness Variability ! Pt Comply .164 .248 -- --
Sp Closeness Variability !Sp Comply .475* .012 -- --
Interaction effects
Pt Closeness Variability x Pt Pos Infl ! Pt Comply -.055* .018 -- --
Pt Closeness Variability x Pt Neg Infl ! Pt Comply .013 .654 -- --
Pt Closeness Variability x Sp Pos Infl ! Pt Comply .015 .599 -- --
Pt Closeness Variability x Sp Neg Infl ! Pt Comply .031 .266 -- --
Sp Closeness Variability x Sp Pos Infl ! Sp Comply .023 .467 -- --
Sp Closeness Variability x Sp Neg Infl ! Sp Comply .042 .211 -- --
Sp Closeness Variability x Pt Pos Infl ! Sp Comply -.045 .202 -- --
Sp Closeness Variability x Pt Neg Infl ! Sp Comply -.039 .210 -- --
Main effects of Moderator Variable
Pt Closeness Variability ! Sp Comply -- -- -.033 .885
Sp Closeness Variability !Pt Comply -- -- -.162 .129
Interaction effects
Pt Closeness Variability x Sp Pos Infl ! Sp Comply -- -- .021 .616
Pt Closeness Variability x Sp Neg Infl ! Sp Comply -- -- .016 .755
Pt Closeness Variability x Pt Pos Infl ! Sp Comply -- -- .024 .489
Pt Closeness Variability x Pt Neg Infl ! Sp Comply -- -- -.028 .574
Sp Closeness Variability x Pt Pos Infl ! Pt Comply -- -- -.034 .101
Sp Closeness Variability x Pt Neg Infl ! Pt Comply -- -- -.001 .957
Sp Closeness Variability x Sp Pos Infl ! Pt Comply -- -- .011 .605
Pt Closeness Variability x Sp Pos Infl ! Sp Comply -- -- .027 .197
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos
Infl = positive influence attempts; Neg Infl = negative influence attempts; Comply = patient treatment compliance at follow-up. * p <
.05; ** p < .01; + p < .06.
!"#
Table 7. APIM of the effect of treatment compliance at follow-up on CHF symptom level at
follow-up (Aim 1 continued).
APIM Parameters Estimate (p)
Covariates
Illness Severity ! Pt CHF Symptoms .125** .004
Spouse Distress ! Pt CHF Symptoms .013** .001
Illness Severity ! Sp CHF Symptoms .095 .092
Spouse Distress ! Sp CHF Symptoms .016** .002
Actor effects
Pt Compliance ! Pt CHF Symptoms -.185* .020
Sp Compliance ! Sp CHF Symptoms -.087 .265
Partner effects
Pt Compliance ! Sp CHF Symptoms -.101 .331
Sp Compliance ! Pt CHF Symptoms .019 .742
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of
the variable; Pos Infl = positive influence attempts; Neg Infl = negative influence attempts.
* p < .05; ** p < .01

!"#
Table 8. APIM of the effect of spousal influence attempts on general physical health at follow-up
and with moderation by baseline marital quality (Aims 2 and 3).

No Moderation
Moderation by
Actor Marital
Quality
Moderation by
Partner Marital
Quality
APIM Parameters Estimate p Estimate p Estimate p
Covariates
Illness Severity ! Pt Physical Health -7.85+ .052 -10.54* .021 -12.16** .001
Spouse Distress ! Pt Physical Health -.736* .011 -.832** .006 -.893** .003
Illness Severity ! Sp Physical Health 10.00 .094 7.79 .185 8.25 .112
Spouse Distress ! Sp Physical Health -.733 .108 -.815 .076 -.763 .097
Actor effects
Pt Pos Infl ! Pt Physical Health .382 .522 .477 .488 .143 .833
Pt Neg Infl ! Pt Physical Health -.764 .243 -.783 .228 -.694 .334
Sp Pos Infl ! Sp Physical Health -.370 .659 -.096 .912 -.236 .811
Sp Neg Infl ! Sp Physical Health .693 .498 .721 .516 .619 .581
Partner effects
Pt Pos Infl ! Sp Physical Health .786 .358 .477 .600 .668 .527
Pt Neg Infl ! Sp Physical Health -.618 .557 -.785 .463 -.577 .618
Sp Pos Infl ! Pt Physical Health -.766 .165 -1.01 .087 -.687 .245
Sp Neg Infl ! Pt Physical Health .772 .238 .609 .363 .619 .420
Main effects of Moderator Variable
Pt Marital Quality ! Pt Physical Health -- -- -1.45 .701 -- --
Sp Marital Quality !Sp Physical Health -- -- -4.27 .453 -- --
Interaction effects
Pt Marital Quality x Pt Pos Infl ! Pt Physical Health -- -- .841 .305 -- --
Pt Marital Quality x Pt Neg Infl ! Pt Physical Health -- -- -1.46 .076 -- --
Pt Marital Quality x Sp Pos Infl ! Pt Physical Health -- -- -.996 .174 -- --
Pt Marital Quality x Sp Neg Infl ! Pt Physical Health -- -- -.312 .704 -- --
Sp Marital Quality x Sp Pos Infl ! Sp Physical Health -- -- 1.23 .252 -- --
Sp Marital Quality x Sp Neg Infl ! Sp Physical Health -- -- -.685 .546 -- --
Sp Marital Quality x Pt Pos Infl ! Sp Physical Health -- -- 1.23 .331 -- --
Sp Marital Quality x Pt Neg Infl ! Sp Physical Health -- -- -1.10 .315 -- --
Main effects of Moderator Variable
Pt Marital Quality ! Sp Physical Health -- -- -- -- -.192 .975
Sp Marital Quality !Pt Physical Health -- -- -- -- -2.36 .568
Interaction effects
Pt Marital Quality x Pt Pos Infl ! Sp Physical Health -- -- -- -- .417 .708
Pt Marital Quality x Pt Neg Infl ! Sp Physical Health -- -- -- -- -.353 .780
Pt Marital Quality x Sp Pos Infl ! Sp Physical Health -- -- -- -- .039 .975
Pt Marital Quality x Sp Neg Infl ! Sp Physical Health -- -- -- -- -.118 .927
Sp Marital Quality x Sp Pos Infl ! Pt Physical Health -- -- -- -- .606 .521
Sp Marital Quality x Sp Neg Infl ! Pt Physical Health -- -- -- -- -.170 .833
Sp Marital Quality x Pt Pos Infl ! Pt Physical Health -- -- -- -- -.907 .255
Sp Marital Quality x Pt Neg Infl ! Pt Physical Health -- -- -- -- -.723 .393
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos Infl
= positive influence attempts; Neg Infl = negative influence attempts; * p < .05; ** p < .01; + p < .06.
!"#
Table 9. APIM of the effect of spousal influence attempts on general physical health at follow-up
and with moderation by closeness variability (Aim 3 continued).

Moderation by Actor
Closeness Variability
Moderation by Partner
Closeness Variability
APIM Parameters Estimate p Estimate p
Covariates
Illness Severity ! Pt Physical Health -11.94** .002 -10.19* .030
Spouse Distress ! Pt Physical Health -.887** .009 -.743* .029
Pt Marital Quality ! Pt Physical Health .954 .820 -.826 .855
Sp Marital Quality ! Pt Physical Health 1.09 .819 .083 .987
Illness Severity ! Sp Physical Health 6.80 .175 6.98 .224
Spouse Distress ! Sp Physical Health -1.06+ .054 -.652 .167
Sp Marital Quality ! Sp Physical Health -11.47 .110 -10.13 .093
Pt Marital Quality ! Sp Physical Health 4.46 .481 6.00 .264
Actor effects
Pt Pos Infl ! Pt Physical Health .537 .426 .500 .465
Pt Neg Infl ! Pt Physical Health -.880 .272 -1.10 .186
Sp Pos Infl ! Sp Physical Health .257 .815 -.268 .761
Sp Neg Infl ! Sp Physical Health 1.09 .421 .957 .359
Partner effects
Pt Pos Infl ! Sp Physical Health .707 .467 1.17 .169
Pt Neg Infl ! Sp Physical Health -.248 .840 -.402 .696
Sp Pos Infl ! Pt Physical Health -.863 .189 -.800 .271
Sp Neg Infl ! Pt Physical Health .789 .338 1.21 .141
Main effects of Moderator Variable
Pt Closeness Variability ! Pt Physical Health -5.48 .415 -- --
Sp Closeness Variability !Sp Physical Health -9.71 .275 -- --
Interaction effects
Pt Closeness Variability x Pt Pos Infl ! Pt Physical Health .560 .578 -- --
Pt Closeness Variability x Pt Neg Infl ! Pt Physical Health -1.02 .376 -- --
Pt Closeness Variability x Sp Pos Infl ! Pt Physical Health -1.05 .397 -- --
Pt Closeness Variability x Sp Neg Infl ! Pt Physical Health .800 .521 -- --
Sp Closeness Variability x Sp Pos Infl ! Sp Physical Health -1.02 .505 -- --
Sp Closeness Variability x Sp Neg Infl ! Sp Physical Health -.694 .679 -- --
Sp Closeness Variability x Pt Pos Infl ! Sp Physical Health -.209 .872 -- --
Sp Closeness Variability x Pt Neg Infl ! Sp Physical Health .223 .893 -- --
Main effects of Moderator Variable
Pt Closeness Variability ! Sp Physical Health -- -- -4.73 .315
Sp Closeness Variability !Pt Physical Health -- -- -15.18 .126
Interaction effects
Pt Closeness Variability x Sp Pos Infl ! Sp Physical Health -- -- 3.64* .041
Pt Closeness Variability x Sp Neg Infl ! Sp Physical Health -- -- -1.34 .523
Pt Closeness Variability x Pt Pos Infl ! Sp Physical Health -- -- -.724 .577
Pt Closeness Variability x Pt Neg Infl ! Sp Physical Health -- -- 3.02 .136
Sp Closeness Variability x Pt Pos Infl ! Pt Physical Health -- -- -.244 .787
Sp Closeness Variability x Pt Neg Infl ! Pt Physical Health -- -- -.140 .877
Sp Closeness Variability x Sp Pos Infl ! Pt Physical Health -- -- .382 .715
Pt Closeness Variability x Sp Pos Infl ! Sp Physical Health -- -- .815 .421
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos
Infl = positive influence attempts; Neg Infl = negative influence attempts. * p < .05; ** p < .01; + p < .06.
!"#

Table 10. APIM of the effect of spousal influence attempts on general mental health at follow-up
and with moderation by baseline marital quality (Aims 2 and 3).

No Moderation
Moderation by
Actor Marital
Quality
Moderation by
Partner Marital
Quality
APIM Parameters Estimate p Estimate p Estimate p
Covariates
Illness Severity ! Pt Mental Health -5.37 .106 -7.20 .069 -4.87 .174
Spouse Distress ! Pt Mental Health -.604 .012 -.615* .019 -.574* .025
Illness Severity ! Sp Mental Health 5.06 .227 4.56 .299 5.15 .287
Spouse Distress ! Sp Mental Health -.752* .022 -.839* .017 -.815* .021
Actor effects
Pt Pos Infl ! Pt Mental Health .195 .693 .281 .641 .252 .660
Pt Neg Infl ! Pt Mental Health -1.34* .016 -1.38* .018 -1.25* .043
Sp Pos Infl ! Sp Mental Health -.438 .461 -.303 .642 -.475 .493
Sp Neg Infl ! Sp Mental Health -.605 .403 -.320 .701 -.477 .543
Partner effects
Pt Pos Infl ! Sp Mental Health .866 .155 .638 .345 1.10 .141
Pt Neg Infl ! Sp Mental Health -.265 .721 -.402 .616 -.359 .658
Sp Pos Infl ! Pt Mental Health -.229 .613 -.355 .489 -.251 .611
Sp Neg Infl ! Pt Mental Health .813 .135 .687 .244 .661 .307
Main effects of Moderator Variable
Pt Marital Quality ! Pt Mental Health -- -- 1.53 .645 -- --
Sp Marital Quality !Sp Mental Health -- -- -2.59 .543 -- --
Interaction effects
Pt Marital Quality x Pt Pos Infl ! Pt Mental Health -- -- -.289 .687 -- --
Pt Marital Quality x Pt Neg Infl ! Pt Mental Health -- -- -.289 .684 -- --
Pt Marital Quality x Sp Pos Infl ! Pt Mental Health -- -- .040 .950 -- --
Pt Marital Quality x Sp Neg Infl ! Pt Mental Health -- -- -.274 .705 -- --
Sp Marital Quality x Sp Pos Infl ! Sp Mental Health -- -- .202 .801 -- --
Sp Marital Quality x Sp Neg Infl ! Sp Mental Health -- -- -.190 .824 -- --
Sp Marital Quality x Pt Pos Infl ! Sp Mental Health -- -- .914 .338 -- --
Sp Marital Quality x Pt Neg Infl ! Sp Mental Health -- -- .224 .784 -- --
Main effects of Moderator Variable
Pt Marital Quality ! Sp Mental Health -- -- -- -- -1.81 .673
Sp Marital Quality !Pt Mental Health -- -- -- -- .146 .967
Interaction effects
Pt Marital Quality x Pt Pos Infl ! Sp Mental Health -- -- -- -- -.406 .603
Pt Marital Quality x Pt Neg Infl ! Sp Mental Health -- -- -- -- .477 .591
Pt Marital Quality x Sp Pos Infl ! Sp Mental Health -- -- -- -- -.276 .749
Pt Marital Quality x Sp Neg Infl ! Sp Mental Health -- -- -- -- .262 .773
Sp Marital Quality x Sp Pos Infl ! Pt Mental Health -- -- -- -- -.252 .752
Sp Marital Quality x Sp Neg Infl ! Pt Mental Health -- -- -- -- -.097 .887
Sp Marital Quality x Pt Pos Infl ! Pt Mental Health -- -- -- -- -.390 .560
Sp Marital Quality x Pt Neg Infl ! Pt Mental Health -- -- -- -- -.008 .991
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos Infl
= positive influence attempts; Neg Infl = negative influence attempts; * p < .05; ** p < .01; + p < .06.
!"#
Table 11. APIM of the effect of spousal influence attempts on general mental health at follow-up
and with moderation by closeness variability (Aim 3 continued).

Moderation by Actor
Closeness Variability
Moderation by Partner
Closeness Variability
APIM Parameters Estimate p Estimate p
Covariates
Illness Severity ! Pt Mental Health -7.52* .036 -9.18** .006
Spouse Distress ! Pt Mental Health -.710* .013 -.527* .027
Pt Marital Quality ! Pt Mental Health 3.29 .327 -1.33 .674
Sp Marital Quality ! Pt Mental Health -1.24 .740 2.30 .514
Illness Severity ! Sp Mental Health 6.69 .155 3.14 .497
Spouse Distress ! Sp Mental Health -.784+ .052 -.844* .031
Sp Marital Quality ! Sp Mental Health -.788 .864 -.244 .955
Pt Marital Quality ! Sp Mental Health -3.21 .529 -3.04 .527
Actor effects
Pt Pos Infl ! Pt Mental Health .409 .442 .759 .117
Pt Neg Infl ! Pt Mental Health -1.51* .020 -2.18** .000
Sp Pos Infl ! Sp Mental Health -.048 .951 -.165 .818
Sp Neg Infl ! Sp Mental Health -.631 .513 -.551 .514
Partner effects
Pt Pos Infl ! Sp Mental Health .685 .322 1.07 .124
Pt Neg Infl ! Sp Mental Health .224 .798 -.332 .691
Sp Pos Infl ! Pt Mental Health -.018 .971 -.776 .129
Sp Neg Infl ! Pt Mental Health 1.06 .106 .662 .307
Main effects of Moderator Variable
Pt Closeness Variability ! Pt Mental Health -.417 .937 -- --
Sp Closeness Variability !Sp Mental Health .572 .927 -- --
Interaction effects
Pt Closeness Variability x Pt Pos Infl ! Pt Mental Health -.062 .938 -- --
Pt Closeness Variability x Pt Neg Infl ! Pt Mental Health .345 .704 -- --
Pt Closeness Variability x Sp Pos Infl ! Pt Mental Health 1.22 .218 -- --
Pt Closeness Variability x Sp Neg Infl ! Pt Mental Health -.072 .941 -- --
Sp Closeness Variability x Sp Pos Infl ! Sp Mental Health -.359 .742 -- --
Sp Closeness Variability x Sp Neg Infl ! Sp Mental Health -.103 .931 -- --
Sp Closeness Variability x Pt Pos Infl ! Sp Mental Health -.074 .936 -- --
Sp Closeness Variability x Pt Neg Infl ! Sp Mental Health -1.25 .294 -- --
Main effects of Moderator Variable
Pt Closeness Variability ! Sp Mental Health -- -- .436 .956
Sp Closeness Variability !Pt Mental Health -- -- -4.29 .195
Interaction effects
Pt Closeness Variability x Sp Pos Infl ! Sp Mental Health -- -- 1.50 .289
Pt Closeness Variability x Sp Neg Infl ! Sp Mental Health -- -- -.728 .669
Pt Closeness Variability x Pt Pos Infl ! Sp Mental Health -- -- -.292 .781
Pt Closeness Variability x Pt Neg Infl ! Sp Mental Health -- -- .015 .992
Sp Closeness Variability x Pt Pos Infl ! Pt Mental Health -- -- -.335 .596
Sp Closeness Variability x Pt Neg Infl ! Pt Mental Health -- -- .428 .502
Sp Closeness Variability x Sp Pos Infl ! Pt Mental Health -- -- 1.94* .011
Pt Closeness Variability x Sp Pos Infl ! Sp Mental Health -- -- .462 .513
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos
Infl = positive influence attempts; Neg Infl = negative influence attempts. * p < .05; ** p < .01; + p < .06.
!"#
Table 12. APIM of the effect of daily influence attempts on next day marital closeness and with
moderation by ratio of positive to negative influence (Aim 4).

No Moderation
Moderation by
Actor Ratio
Pos:Neg Infl
Moderation by
Partner Ratio
Pos:Neg Infl
APIM Parameters Estimate p Estimate p Estimate p
Covariates
Pt Marital Closeness Day t ! Pt Marital Closeness Day t + 1
.223** .000 .192** .001 .201 .001
Sp Marital Closeness Day t ! Sp Marital Closeness Day t + 1
.151** .007 .120* .035 .132 .022
Actor effects

Pt Infl Day t! Pt Marital Closeness Day t + 1
.121* .019 .134* .013 .157 .007
Sp Infl Day t ! Sp Marital Closeness Day t + 1
.063 .238 .062 .272 .042 .484
Partner effects

Pt Infl Day t ! Sp Marital Closeness Day t + 1
-.035 .345 -.058 .142 -.031 .443
Sp Infl Day t ! Pt Marital Closeness Day t + 1
.042 .279 .066 .112 .075 .094
Main effects of Moderator Variable

Pt Ratio Pos/Neg Infl ! Pt Marital Closeness Day t + 1
-- -- -.042 .959 -- --
Sp Ratio Pos/Neg Infl ! Sp Marital Closeness Day t + 1
-- -- 1.01** .005 -- --
Interaction effects

Pt Ratio Pos/Neg Infl x Pt Infl ! Pt Marital Closeness Day t + 1
-- -- .332* .029 -- --
Pt Ratio Pos/Neg Infl x Sp Infl ! Pt Marital Closeness Day t + 1
-- -- -.077 .612 -- --
Sp Ratio Pos/Neg Infl x Sp Infl ! Sp Marital Closeness Day t + 1
-- -- -.013 .866 -- --
Sp Ratio Pos/Neg Infl x Pt Infl ! Sp Marital Closeness Day t + 1
-- -- .038 .531 -- --
Main effects of Moderator Variable

Pt Ratio Pos/Neg Infl ! Sp Marital Closeness Day t + 1
-- -- -- -- .899 .254
Sp Ratio Pos/Neg Infl ! Pt Marital Closeness Day t + 1
-- -- -- -- -.449 .333
Interaction effects

Pt Ratio Pos/Neg Infl x Sp Infl ! Sp Marital Closeness Day t + 1
-- -- -- -- .154 .386
Pt Ratio Pos/Neg Infl x Pt Infl ! Sp Marital Closeness Day t + 1
-- -- -- -- -.037 .755
Sp Ratio Pos/Neg Infl x Sp Infl ! Pt Marital Closeness Day t + 1
-- -- -- -- .044 .567
Sp Ratio Pos/Neg Infl x Pt Infl ! Pt Mental Health Day t + 1 -- -- -- -- -.072 .281
Note. Estimates are unstandardized regression coefficients. Pt = patient report of the variable; Sp = spouse report of the variable; Pos Infl
= positive influence attempts; Neg Infl = negative influence attempts; * p < .05; ** p < .01.

!"#
Figure 1. The Actor-Partner Interdependence Model (APIM).

Note: X = data for person A, Time 1; X’ = data for person B, Time 1; Y = data for person A, Time 2; Y’ = data for
person B, Time 2; U = residual portion of person A’s Time 2 score; U’ = residual portion of person B’s Time 2
score. Single-headed arrows indicate predictive path. Double-headed arrows indicate correlated variables. Paths
labeled as a indicate actor effects and paths labeled as p indicate partner effects.

!"#
Figure 2. Conceptual APIM model of the indirect effect of spousal influence attempts on CHF
symptoms via patient compliance.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.
!"#
Figure 3. Conceptual APIM model of the effect of spousal influence attempts on patient and
spouse general physical health.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.
!!"
Figure 4. Conceptual APIM model of the effect of spousal influence attempts on patient and
spouse mental health.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.
!""#
Figure 5. Conceptual APIM model with moderation of the indirect effect of spousal influence
attempts on CHF symptoms via patient compliance by marital quality and closeness variability.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.

!"!#
Figure 6. Conceptual APIM model with moderation of the effect of spousal influence attempts
on patient and spouse general physical health by marital quality and closeness variability.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.

!"#$
Figure 7. Conceptual APIM model with moderation of the effect of spousal influence attempts
on patient and spouse mental health by marital quality and closeness variability.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.
!"#$
Figure 8. Conceptual model of repeated measures APIM of the effect of providing and receiving
influence attempts on next day marital closeness.

!"#$
Figure 9. Conceptual model of repeated measures APIM with moderation of the effect of
providing and receiving influence attempts on next day marital closeness by relative positivity to
negativity of the influence.

!"#$
Figure 10. APIM of spousal influence attempts on patient compliance at follow-up.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.
a = path moderated by patient closeness variability in subsequent moderation analyses.
* p < .05; ** p < .01.
!"#$
Figure 11. Moderation of the effect of patient report of positive influence on patient report of
follow-up compliance by patient closeness variability.

Note: Pt = patient report of the variable; CloseVar = closeness variability.
1
1.5
2
2.5
3
3.5
4
4.5
5
Low Positive Influence-Pt High Positive Influence-Pt
Pt Follow-Up Compliance
Low Pt CloseVar
High Pt CloseVar
!"#$
Figure 12. APIM of patient compliance on CHF symptoms.

Note: T2 = time 2 variables measured at six-month follow-up.
* p < .05.

!"#$

Figure 13. APIM of the effect of positive and negative influence attempts on patient and spouse
general physical health at follow-up.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up. a = path is moderated by patient closeness variability in subsequent moderation analyses.
!"#$
Figure 14. Moderation of the effect of spouse report of positive influence on spouse general
physical health at follow-up by patient closeness variability.

Note: Pt = patient report of the variable; Sp = spouse report of the variable; CloseVar = closeness variability.
0
10
20
30
40
50
60
70
80
90
100
110
Low Positive Influence-Sp High Positive Influence-Sp
Sp Physical Health at Follow-up
Low Pt CloseVar
High Pt CloseVar
!!"#
Figure 15. APIM of the effect of positive and negative influence attempts on patient and spouse
mental health at follow-up.

Note: T1 = time 1 variables measured during baseline assessment; T2 = time 2 variables measured at six-month
follow-up.
b = path is moderated by spouse closeness variability in subsequent moderation analyses.
* p < .05; ** p < .01.
!!!"
Figure 16. Moderation of the effect of spouse report of positive influence on patient mental
health at follow-up by spouse closeness variability.

Note: Pt = patient report of the variable; Sp = spouse report of the variable; CloseVar = closeness variability.

0
10
20
30
40
50
60
70
80
90
100
Low Positive Influence-Sp High Positive Influence-Sp
Pt Mental Health at Follow-up
Low Sp CloseVar
High Sp CloseVar
!!"#
Figure 17. Repeated measures APIM of the effect of patient report of daily influence received on
patient report of next day marital closeness.

Note: a = path is moderated by patient ratio positivity to negativity of influence attempts in subsequent moderation
analyses.
* p < .05.

!!"#
Figure 18. Moderation of the effect of patient report of daily influence on patient report of daily
marital closeness by the ratio of positivity to negativity of the influence received. .

Note: Pt = patient report of the variable; Pt Daily Infl = patient report of receipt of daily influence attempts; Low
ratio = ratio of positive to negative influence attempts is 1:2; High ratio = ratio of positive to negative influence
attempts is 2:1.
0
1
2
3
4
5
6
7
8
9
10
Low Pt Daily Infl High Pt Daily Infl
Pt Daily Marital Closeness
Low Pt ratio
High Pt ratio
!!"#
Appendix A
Hopkins Symptom Checklist – 25
Not at all A little Quite a bit Extremely
Suddenly scared for no reason 1 2 3 4
Feeling fearful 1 2 3 4
Faintness, dizziness, or weakness 1 2 3 4
Nervousness or shakiness inside 1 2 3 4
Heart pounding or racing 1 2 3 4
Trembling 1 2 3 4
Feeling tense or keyed up 1 2 3 4
Headaches 1 2 3 4
Spells of terror or panic 1 2 3 4
Feeling restless, can't sit still 1 2 3 4
Feeling low in energy, slowed down 1 2 3 4
Blaming yourself for things 1 2 3 4
Crying easily 1 2 3 4
Loss of sexual interest or pleasure 1 2 3 4
Poor appetite 1 2 3 4
Difficulty falling or staying asleep 1 2 3 4
Feeling hopeless about the future 1 2 3 4
Feeling blue 1 2 3 4
Feeling lonely 1 2 3 4
Feeling trapped or caught 1 2 3 4
Worrying too much about things 1 2 3 4
Feeling no interest in things 1 2 3 4
Thoughts of ending your life 1 2 3 4
Feeling everything is an effort 1 2 3 4
Feeling worthless 1 2 3 4
!!"#
Appendix B

Positive Health-Related Spousal Influence Attempts.
How often do you do the following:
not at all sometimes very often
1. Help him/her view taking medications as part of a
normal daily routine..................................................1 ...............2 ...............3 ...............4 ...............5
2. Praise him/her for taking his/her medications ..........1 ...............2 ...............3 ...............4 ...............5
3. Praise him/her for following a healthy diet ..............1 ...............2 ...............3 ...............4 ...............5
4. Praise him/her for getting enough exercise ..............1 ...............2 ...............3 ...............4 ...............5
5. Join him/her in maintaining a healthy diet ...............1 ...............2 ...............3 ...............4 ...............5
6. Exercise with him/her...............................................1 ...............2 ...............3 ...............4 ...............5

Negative Health-Related Spousal Influence Attempts.
How often do you do the following:
7. Argue, complain, or criticize him/her about
how s/he takes medication........................................1 ...............2 ...............3 ...............4 ...............5
8. Argue, complain, or criticize him/her about
his/her diet ................................................................1 ...............2 ...............3 ...............4 ...............5
9. Argue, complain, or criticize him/her about
exercising .................................................................1 ...............2 ...............3 ...............4 ...............5
10. Warning him/her to pay attention to symptoms?......1 ...............2 ...............3 ...............4 ...............5
11. Insisting that s/he feels worse than s/he will
admit? .......................................................................1 ...............2 ...............3 ...............4 ...............5
12. Urging him/her to report symptoms to medical
personnel ..................................................................1 ...............2 ...............3 ...............4 ...............5

!!"#
Appendix C
Compliance with Treatment Regimen

Patient Questions:
In the past month, to what extent have you followed your doctors’ advice to…..
rarely or never sometimes always
1. Take medications exactly as prescribed?..................1 ...............2 ...............3 ...............4 ...............5
2. Maintain a healthy diet? ...........................................1 ...............2 ...............3 ...............4 ...............5
3. Limit the amount of sodium (salt) in your diet?.......1 ...............2 ...............3 ...............4 ...............5
4. Weigh yourself regularly? ........................................1 ...............2 ...............3 ...............4 ...............5
5. Get enough exercise?................................................1 ...............2 ...............3 ...............4 ...............5
6. Pace yourself to avoid overexertion?........................1 ...............2 ...............3 ...............4 ...............5
7. Manage the stress in your life? .................................1 ...............2 ...............3 ...............4 ...............5

!!"#

Appendix D
Relationship Assessment Scale (RAS).
1. How well do you feel your partner meets your needs?
Low satisfaction 1 2 3 4 5 High satisfaction

2. In general, how satisfied are you with your relationship?
Low satisfaction 1 2 3 4 5 High satisfaction

3. How good do you feel your relationship is compared to most?
Very poor 1 2 3 4 5 Very good

4. How often do you wish you hadn’t gotten into this relationship?
Never 1 2 3 4 5 Always

5. To what extent does your relationship meet your original expectations?
Low satisfaction 1 2 3 4 5 High satisfaction

6. How much do you love your partner?
Very little 1 2 3 4 5 Very much

7. How many problems are in your relationship?
Very few 1 2 3 4 5 Very many

!!"#
Appendix E

Constructive Communication Scale (CCS)

Please rate each item on a scale from 1 = very unlikely to 7 = very likely.
WHEN SOME PROBLEM IN THE RELATIONSHIP ARISES:
1. Both members try to discuss the problem.
DURING A DISCUSSION OF A RELATIONSHIP PROBLEM:
2. Both members blame, accuse, and criticize each other.
3. Both members express their feelings to each other.
4. Both members threaten each other with negative consequences.
5. Both members suggest possible solutions and compromises.
6. Man calls woman names, swears at her, or attacks her character.
7. Woman calls man names, swears at him, or attacks his character. ....
!!"#

Appendix F

!"#$

!"!#
Appendix G
Daily Diary Questions
Patient Questions
1. How close or connected did you feel to your partner yesterday?
0 1 2 3 4 5 6 7 8 9 10
not at all extremely close and connected
2. To what extent did your partner try to influence you yesterday to keep a healthy diet, get the
right amount of exercise, or take medications as prescribed?
0 1 2 3 4 5 6 7 8 9 10
not at all very much
Spouse Questions
1. How close or connected did you feel to your partner yesterday?
0 1 2 3 4 5 6 7 8 9 10
not at all extremely close and connected
2. To what extent did you try to influence your partner yesterday to keep a healthy diet, get the
right amount of exercise, or take medications as prescribed?
0 1 2 3 4 5 6 7 8 9 10
not at all very much

Abstract (if available)

Abstract Coping with chronic illness is increasingly being viewed as a relational process rather than an individual-based phenomenon. Although spouses serve as each others' primary source of support and assistance, relatively little is known about how spousal involvement in illness management affects outcomes for patients and spouses. The aims of this project were to (1) identify the distinct effects of two types of health-related spousal influence attempts (positive and negative) on patient compliance and subsequent chronic heart failure (CHF) symptoms, (2) examine the potential broader consequences of these influence attempts for patients' and spouses' general physical and mental health, (3) investigate how the marital context (specifically marital quality and variability in closeness) influences the above associations, and (4) explore the day-to-day effects of providing and receiving influence attempts on feelings of marital closeness using a daily diary methodology. Analyses for Aims 1 through 3 were conducted using an actor-partner-interdependence model (APIM) with a short-term longitudinal dataset (n = 60 couples). Data revealed that negative spousal influence attempts at baseline were associated with lower treatment compliance for patients six months later and in turn greater CHF symptom levels, whereas positive influence attempts had the opposite effect when feelings of marital closeness were stable. Spouses exhibited a negative association between positive influence attempts and physical health, but no association between influence attempts and physical health was found for patients. Negative influence attempts were negatively associated with patients' mental health six months later

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The influences of anxiety, coping, and social support on physical functioning among heart failure patients

PDF

The effects of familism and cultural justification on the mental and physical health of family caregivers

PDF

The effects of wisdom-related personality traits on caregivers’ health: an application of the resilience model

PDF

Pregnancy in the time of COVID-19: effects on perinatal mental health, birth, and infant development

PDF

Heart, brain, and breath: studies on the neuromodulation of interoceptive systems

Asset Metadata

Creator Geren, Jennifer L. (author)

Core Title The effects of health-related spousal influence on couples coping with chronic heart failure: an application of the actor-partner interdependence model

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Publication Date 07/21/2013

Defense Date 04/02/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag actor partner models,chronic heart failure,couples coping,dyadic data,OAI-PMH Harvest,spousal influence

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Knight, Bob G. (committee chair), Baucom, Brian (committee member), Clark, Florence (committee member), Margolin, Gayla (committee member), Mather, Mara (committee member)

Creator Email jenniferleegeren@gmail.com,kellough@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-293220

Unique identifier UC11288090

Identifier etd-GerenJenni-1804.pdf (filename),usctheses-c3-293220 (legacy record id)

Legacy Identifier etd-GerenJenni-1804.pdf

Dmrecord 293220

Document Type Dissertation

Rights Geren, Jennifer L.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA