xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 002001440
007 cr mnu|||uuuuu
008 090501s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002650
Morgan, Mary Ann.
Cancer patients with pain :
b examination of the role of the spouse/partner relationship in mediating quality of life outcomes for the couple
h [electronic resource] /
by Mary Ann Morgan.
[Tampa, Fla] :
University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 96 pages.
Dissertation (Ph.D.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
ABSTRACT: A diagnosis of cancer, regardless of type or site, raises much fear and loss of control for patients and their spouses. While being married is associated with lower mortality from a wide range of illnesses, including cancer, the quality of marital interactions and the relationship is the stronger predictor of health outcomes, rather than marital status. When people are faced with their greatest life challenges, they attach great importance to the behavior of their intimate partner, with trust being a key component of relationship quality, thus lending stability, and emotional and practical support. The purpose of this study was to examine vulnerable cancer patients with pain and their partners. The quality of the patient/partner relationships and the partners' coping styles were evaluated as mediators in a Structural Equation Model (SEM) latent path analysis with the outcome measures of quality of life for the individuals.Stress, coping and outcome theory guided the questions for the study. Much of the previous dyadic research that has been done on couples when one member has a diagnosis of cancer has centered on disease specific populations. This study was the first to examine the couples' relationships from a symptom defining population. Therefore, a broad range of ages, patients with different cancer diagnoses and both genders in the roles of patients and caregivers was the population studied. Coping was explored as a dyadic process that includes transactional appraisal of stressors that mediates the subsequent effects on quality of life outcomes. Multivariate analysis was used to determine covariates to be included in the SEM based on a review of the literature. There was no evidence of coping as a mediator. The participants in this study had good quality relationships, and this did serve as a positive mediator on the outcomes for the patient.Recommendations for future research and nursing practice, including the use of a single item global assessment of relationship quality, that nurses can incorporate in their practice is discussed.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Susan C. McMillan, Ph.D.
t USF Electronic Theses and Dissertations.
Cancer Patients with Pain: Examination of the Role of the Spouse/Partner Relationship In Mediating Quality of Life Outcomes for the Couple by Mary Ann Morgan A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Nursing University of South Florida Major Professor: Susan McMillan, A.R.N.P., Ph.D. Kristine Donovan, Ph.D. Janine Overcash, A.R.N.P., Ph.D. Brent Small, Ph.D. Date of Approval: November 10, 2008 Keywords: gender, coping, distress, interference, dyadic analysis Copyright 2008, Mary Ann Morgan
To my husband, friend, and life partner Curt You have always been there for me through all of our adventures Always believing in me and encouraging me to be my bestMode ling for me that att itude is everythingI wouldnt change a minute of our life togeth erI look forward to rest of our lives
Acknowledgements Like many of my patients, my fa mily is my passion. I thank my husband for his patience and encouragement through many hours of late nights and early mornings. I thank my children for their support, assurance and humor along the way. I thank and acknowledge my dissertation committee members with sincere appreciation and gratitude. I thank Dr. Susan McMillan, A.R.N.P., Ph.D. for mentoring me and her unwavering confidence that anything is possible if you be lieve and persist. She always travels the paths with her students and it is a lifelong commitment. I thank Dr. Kristine Donovan, M.B.A ., Ph.D. for also mentoring me and patiently explaining the finer details of putting scienc e and theory together. She has shared her enthusiasm for the importance of dyadic resear ch in quality of life issues for cancer patients. I thank Dr. Janine Overcash, A.R.N. P., Ph.D. for her mentoring and dedication of time to help me be a better writer and thinker. She is an enthusiastic and noble professor. I thank Dr. Brent Small, Ph.D. for his patience and guidance. Through his gentle manner, I have learned statistics can be mastered and enjoyed. Like the other committee members, he understands the importan ce of understanding through experience. I thank Dr. Jason Beckstead for his passion for statistics, patience and time over the last few years. He is the LISREL enthusiast and expert. I also thank Richard Gross, M. D ., another mentor, colleague, and friend for his encouragement, support and flexibility with maintaining and enhancing my clinical practice.
i Table of Contents List of Tables iv List of Figures v Abstract vi Chapter One: Introduction 1 Statement of the Problem 6 Aims and Questions 8 Aim 1 8 Question 1 8 Question 2 8 Question 3 8 Question 4 8 Question 5 8 Aim 2 8 Question 6 8 Question 7 8 Question 8 9 Definition of Terms 9 Vulnerability 9 Financial Concerns 9 Pain 9 Distress 10 Interference 10 Symptom Burden 10 Patient/Partner Relationship 10 Partner Coping Style 11 Health Related Quality of Life for Patient and Partner 12 Significance to Nursing 12 Chapter Two: Review of the Literature 13 Conceptual Framework 13 Stressors and Other Covariates 15 Vulnerability 15 Symptom Burden 17 Interference 18 Gender 19 Mediators 21
ii Patient/Partner Relationship 21 Dyadic Adjustment/Communication 23 Coping Styles 26 Outcomes 29 Health Related Quality of Life for Patient and Partner 29 Summary 31 Chapter Three: Methods 33 Sample 33 Instruments 34 Dem ographic Questionnaire: (Patient/Partner) 34 Medical Characteristics of the Patient 34 Brief Pain Inventory (BPI) 34 Memorial Symptom Assessment Scale (MSAS) 35 Caregiver Demands Scale (CDS) 36 Dyadic Adjustment Scale (DAS) 36 Social Problem-Solving Inventory Revised (SPSI-R) 37 The Profile of Mood States (POMS) 37 Medical Outcomes Study Short Form Health Survey (SF-36) 38 Institutional Review Board (IRB) 40 Procedures 40 Data Analysis 40 Chapter Four: Results 46 Profile of Sample 46 Covariates 49 Stressors 52 Mediators 54 Structural Equation Model 54 Goodness of Fit Indices 60 Full Measurement Model 61 Non-Mediated Model 62 Fully Mediated Model 63 Partially Mediated Model 64 Gender and Role 67 Outcomes 69 Quality of Life for Patient and Partner 69 Chapter Five: Discussion 72 Findings 72 Covariates 72 Mediators 74 Outcomes 77 Limitations 78 Recommendations and Conclusions 80
iii References 82 About the Author End Page
iv List of Tables Table 1 Measures Completed by Each Member of the Dyad 39 Table 2 Demographics for Patient/Partner Dyads 47 Table 3 Prevalence of Cancer Types 48 Table 4 Current Status of Cancer Therapy Treatment: Frequency and Percent of Patients 49 Table 5 Regression Model for Covari ates with Patient Dyadic Adjustment Scale (DAS), POMS, General and Mental Health Scores from the SF-36 50 Table 6 Regression Model for Covari ates with Partner Dyadic Adjustment Adjustment Scal e (DAS), POMS, General a nd Mental Health Scores from the SF-36 51 Table 7 Ranges, Means and Standard Deviations for Worst Pain, Pain Intensity Pain Distress, Interference, and Symptom Burden 53 Table 8 Number and Percent of Most Frequently Reported Symptoms 54 Table 9 Means, Standard Deviati ons and Ranges for Indicator Variables Used In Study (N = 177) 57 Table 10 Bivariate Co rrelations for Variables Included in SEM Model 58 Table 11 Fit Indices for Measurement Pain Models 67 Table 12 Summary of Standardized Total Effects Among Latent Variables Partially Mediated Model 67 Table 13 Comparison of Role by Gender Me ans for General and Mental Health, Dyadic Adjustment and POMS 68 Table 14 Main Effects for Role by Gender for POMS, Dyadic Adjustment, General a nd Mental Health of Patients and Partners 70 Table 15 Comparison of SF-36 Scores fo r Patient/Partner to Same Age Norms 71
v List of Figures Figure 1 Conceptual Model 14 Figure 2 Structural Equation Statistical Model 44 Figure 3 Full Measurement Model 62 Figure 4 Non-Mediated Model 63 Figure 5 Fully Mediated Model 64 Figure 6 Partially Mediated Model 66
vi Cancer Patients with Pain: Examination of the Role of the Spouse/Partner Relationship In Me diating Quality of Life Outcomes for the Couple Mary Ann Morgan ABSTRACT A diagnosis of cancer, regardless of t ype or site, raises much fear and loss of control for patients and their spouses. While being ma rried is associated with lower mortality from a wide range of illnesse s, including cancer, the quality of marital interactions and the relationship is the stronger predictor of health outcomes, rather than marital status. When people are faced with their greatest life challenges, they attach great importance to the behavior of their intimate partner, with trust being a key com ponent of relationship quality, thus lending stability, a nd emotional and practical sup port. The purpose of this study was to examine vulnerable cancer patients with pain and their partners. The quality of the patient/partner relationships a nd the partners coping st yles were evaluated as mediators in a Structural Equation Model (SEM) latent path analysis with the outcome measures of quality of life for the individua ls. Stress, coping and outcome theory guided the questions for the study. Much of the pr evious dyadic research that has been done on couples when one member has a diagnosis of cancer has centered on disease specific populations. This study was the first to ex amine the couples relationships from a symptom defining population. Therefore, a broa d range of ages, pati ents with different cancer diagnoses and both genders in the roles of patients and caregivers was the
vii population studied. Coping was explored as a dyadic process that includes transactional appraisal of stressors that me diates the subsequent effect s on quality of life outcomes. Multivariate analysis was used to determine covariates to be included in the SEM based on a review of the literature. There was no evidence of coping as a mediator. The participants in this study ha d good quality relationships, and this did serve as a positive mediator on the outcomes for the patient. R ecommendations for future research and nursing practice, including th e use of a single item global assessment of relationship quality, that nurses can incorporate in their practice is discussed.
1 Chapter One Introduction It is estimated that men will have a one in two chance and women will have a one in three chance of lifetime cancer risk, excluding squamous a nd basal cell skin malignancy (American Cancer Society, [A CS] 2008). Receiving a diagnosis of cancer has a profound impact not only on patients, but also families. A spouse/partner is particularly affected because of fear and uncertainty about treatm ent, impaired functioning, pain, and financial threats in treatment costs and lost earnings and ultimately quality of life. More importantly, cancer carries the expectation or possibility of a terminal prognosis (Halford, Scott, & Smythe, 2000), and research has shown that pati ents and partners psychological adjustments are interrelated (Baider, Koch, Esacson & Kaplan De-Nour, 1998). A diagnosis of cancer invades the h ealth of the patient and affects the life style of not only the individual, but also the partner. Social support has been examined and has been found to have a positive effect on psyc hological adjustments (Schnoll, Harlow, Stolback & Brandt, 1998). The relationship between the part ner and cancer patient is a complex dyad that differs from the general so cial support structure of family and friends (Gale et al., 2001; Manne, Taylor, Dougherty & Kemeny, 1997; B odenmann, 2005). The reactions and coping styles of the partner and the cancer patient also are important (Cutrona, Russell & Gardner, 2005; Manne & Schnoll, 2001; Pistrang & Barker, 2005).
2 Issues of partner trust and s upport are vital to the cancer diagnosis and treatment process (Cutrona; Banthia et al., 2003). The coupl es relationship should be considered by clinicians and researchers because of its strong influence on patient outcomes (Couper et al., 2006). A diagnosis of cancer, regardless of t ype or site, raises much fear and loss of control for patients and partners (Maliski, Heile mann & McCorkle, 2002; Morse & Fife, 1998). Pain has been identified to be one of the most prevalent symptoms for cancer patients (Gordon, 2005; Modonesi, et al., 2005; Stromgre n, et al., 2006; Vallerand, 1997; Walsh & Ribicki, 2006). If pain is perceived as a symptom of th e cancer or progressive disease the affective dimension and the meaning of pain carries even more significance for the patient (Smith, Gracely & Safer, 1998). Symptom burden, which includes physical symptoms and interference with life, increases with stage of cancer (Modonesi et al., 2005). Symptoms cannot be addressed from a purely biomedical perspective. Th ey include physical intensity and emotional distress or affective components (McMill an, Tofthagen, & Morgan, 2008; Portenoy, Thaler, Kornblith, Lepore, Friedlander-Klar, C oyle, et al., 1994). Similarly, these issues are important for partners, because they are dealing with the psychological impact of a cancer diagnosis, and their quality of life will also be affected by the patients pain and symptom burden (Ezer et al., 2006). Gender also has been identified as an issue affecting adjustment and reaction to hardships in that females, whether patient or partner, report more distress (Ben-Zur, Gilbar & Lev, 2001; Gilbar & Zusman, 2007; Hagedoorn Sanderman, Bolks, Tuinstra and Coyne, 2008). Spouses/partners are at increased risk for both physical and
3 psychological disturbances, and this will vary depending on the quality of the marital relationship, the nature and t ype of illness, functional impairment, changes in family functioning, including finances and role changes, age, gender and coping style of the spouse/partner; all of these are important co nsiderations in dyadic or couple studies (Burman & Margolin, 1992; Westman, Ke inan, Roziner & Benyamini, 2008). The definition of the dyad is couples in a comm itted relationship who can be heterosexual or same sex partners. Henceforth in this pape r, the dyad will consist of the couple and the spouse/partner will be referred to as the pa rtner. This study invol ves only heterosexual couples, as does the populations in this review of the literature. Lazarus and Folkman (1984) address negative stress as a stimulus that causes a reaction. They address three types of negative stress stimuli: (1) a cataclysmic event which affects large numbers of people, such as a hurricane; (2) disast rous events can also occur to only one person or a few person, such as a life threatening illness; and (3) daily hassles. The second stressor can be just as powerful and distressi ng to the individuals involved as a major disaster. The level of threat or vulne rability determines the coping response. When fear is involved, emo tional coping style can impact cognitive functioning and positive problem-solving abilit ies. However, high emotional and intact cognitive functioning can both be present in individuals, even under the most difficult circumstances. Problem-focused coping involves not only efforts focused on a situation with alternative actions and objectives, it also incl udes strategies that are directed inward. These strategies may include shifting leve ls of aspiration, reducing ego involvement,
4 exploring alternative channels of gratificat ion and developing new skills and procedures (Lazarus & Folkman, 1984). Emotion-focused coping can include reappraisal, avoidan ce or seeking outside emotional support. Denial can be part of this type of coping and can be helpful or harmful if deceptive rather than temporaril y protective. Confusion and misunderstanding may be the result if this strategy is predominant (Lazarus & Folkman, 1984). The construct of dyadic coping i nvolves both partners, usually in a marital or committed relationship, and it involves the co mmunication between stress signals of one partner and subsequent coping re actions of the other partner. This coping can include everyday communication, interp ersonal conflict, joint pr oblem solving, giving and receiving emotional support and de aling with life stressors as a couple rather than as two individuals (Bodenmann, Pi het, & Kayser, 2006). Dyadic coping has both positive a nd negative natures. A pa rtners adjustment to the patients diagnosis of cancer is a critical component of how the patient adjusts to the disease (Manne, et al., 1997; Northo use, Templin, Mellon & George, 2000). Psychological distress is predicted by dyadi c adjustment and coping styles following cancer diagnosis for both patients and partners (Banthia et al., 2003; Ko et al., 2005) and if one partner becomes distressed the other is likely to follow (Gilbar & Zusman, 2007; Bishop et al., 2007). One partners awareness of poor health of the ot her partner may lead to crossover of distress in the caregiving partner through em pathy. In the context of health outcomes, this increased risk for the caregiving partner may result in poorer health as a consequence of the increased burden (Westm an et al., 2008). Partners may be at as great or greater risk for distress and compro mised health as patients, because their own
5 lives are forced into new and unpredictabl e situations (Baider, Walach, Perry & Denour,1998; Hinnen, Hagedoorn, Sa nderman, & Ranchor, 2007). Positive supportive coping occurs when partners assist or help the other partner via a variety of mechanisms including: words of wise counsel, communi cating belief in the other partners capabilities, and expressions of solidarity (Bodenman, 2005). Support also may include self-sacrifice, such as taking on more tasks or household duties that are normally divided (Cutrona et al., 2005). Empathy or an emotional connection has been regarded as particularly important for cance r patients (Carlson, Otte nbreit, St. Pierre & Bultz, 2001). Positive coping results in a feeling of mutual trust, reliability, commitment, and the perception of a promise th at despite difficult circumstances, support is available. Efforts that s upport one partner also help to reduce the supporting partners own stress and sustain the relatio nship as well (Bodenman, 2005). Negative forms of dyadic coping can affect the process of a cancer diagnosis and treatment (Schmaling & Sher, 2000). Negative forms of coping include hostile, ambivalent and superficial coping, in that, help is provided, but in a way that is perceived as being sarcastic, unwilling or insincere (Bodenmann, 2005). Characteristics of negative support or lack of support include: physical avoidance of the patient, avoidance of open communication with the patient, engaging in fo rced cheerfulness, or minimization of the illness and its consequences that result in pa tients feeling rejected or abandoned (Hinnen et al., 2007; Manne et al., 1997) Patients who use avoidance coping techniques tend to report greater distress and poorer adjustment to cancer di agnosis and treatment (Ben-Zur et al., 2001). Partners patterns of coping affect their level of distress, such that those who use avoidant or impulsive styles are more di stressed (Couper et al., 2006). Banthia et al.
6 (2003) reported that dyadic strength modera ted the effects of avoidant coping and intrusive thinking (considered maladaptive) on mood disturbance. However, members of stronger dyads reported less distress, despite maladaptive coping behaviors compared to dyads that were in dysfunctional relationships. With the mapping of the human genome, research is producing insights and advances into the causes and cures for can cer. It is hoped that with increased understandings of mechanisms of disease, we will be able to use targeted therapies to individualize care to be most effective with the least t oxicities (Feetham, Thomson, & Himshaw, 2005; Tranin, Masny, & Jenkins, 2003 ). This will have implications for patient/partner dyads and families because genetic information affects generations (Tranin, Masny & Jenkins, 2003). Spouses of women who have not yet been diagnosed with breast cancer but who are BRCA1 or BRC A2 mutation carriers have lower quality of life than husbands of wives who do not car ry these mutations (Mireskandari et al., 2006). Even though there is an emphasis on gene tic research, the result ant quality of life, including familial implications, symptoms, treatment effects and side-effects of the disease with its cancer-related symptom di stress, including pai n, continue to be challenging and the focus of needed research to be addressed according to the National Institutes of Health State-of -the Science Panel (NIH, 2002). Statement of the Problem Being married is associated with lower mortality from a wide range of illnesses, including cancer, but the quality of marital interactions and relationship is the stronger predictor of health outcomes, rather than the marital status (Ren, 1997; Halford et al., 2000; Schmaling & Sher, 2000). Mutual support that is e xperienced by the partners
7 predicts well-being for patients and partners (H alford). Demonstrations of the partners support are critical during times of stress, wh en people feel most vulnerable (Cutrona et al., 2005; Hinnen et al., 2007; Manne, Ostr off, Sherman et al., 2004). Northouse and colleagues (2000) reported, in a study of col on cancer patients, that patients who had higher marital satisfaction ha d lower hopelessness and fewer adjustment problems. Most of the studies examining c ouples who have been impacted by cancer include breast and prostate patients (e.g. Banthia, et al., 2003; BenZur et al, 2001; Bultz, Speca, Brasher, Geggie & Page, 2000) with a few in the gastrointestinal, colon, hematopioetic stem-cell transplant and lung cancer patient populations (e.g. Ba dr & Taylor, 2006; Northouse et al., 2000; Porter, Keefe, Hurw itz & Faber, 2005). A meta-analysis of distress studies of couples coping with cancer by Hagedoorn and colleagues (2008) considered these cancer specific studies a majo r limitation in generalization of results. No studies that have report ed the quality of the marita l or dyadic relationship and adjustment have included cancer patients sy mptom burden, symptom distress, or pain. No studies included a wide range of cancer di agnoses. With so many genetic implications and familial concerns, the quality and impact of the partner relationship takes on even greater importance in caring for cancer patients. The purpose of this dyadic study wa s to examine the effects of pain and symptom burden, including physical and emotional compone nts, on physical and mental health of the cancer patients and partners. Seeing a l oved one endure pain and other distressing symptoms has implications for distress and qua lity of life for both members of the dyad. The role of the dyadic relationship, the partne rs resourcefulness or coping style, and
8 quality of life outcomes of distress, physical and mental health for the patient and partner were analyzed. Aims and Questions Aim 1 To examine patients and their part ners when challenged by stressors of pain and symptom burden and its effect on their QOL. Question 1 Are covariates such as age, gender, financial concerns and length of the marriage related to the relatio nship quality or quality of life outcomes? Question 2. Is pain or symptom burden related to the dyadic committed relationship? Question 3. Does pain or symptom burden impact quality of life for the patient and partner in terms of emotional di stress and overall mental and physical health? Question 4 Does the quality of the relationship mediate the outcomes of emotional distress and overall mental and physical health of the individuals? Question 5. Are the patients and partners distress related to each other and are there any differences for gender? Aim 2 To examine the coping style of the partner when challenged by the patients pain or symptom burden. Question 6. Does pain or symptom burden have a relationship to the coping style of the partner? Question 7 Does the coping style of the partners impact their own or the patients distress and mental and physical health?
9 Question 8 Does the partners coping style mediate the QOL outcomes for the patient and partner? Definition of Terms For the purpose of this study, the following terms were defined. Latent constructs and indicators used in the structural equation model that was recommended for dyadic study analysis (Kenny, Kashy & Cook, 2006) were included. Vulnerability Vulnerability is the latent construc t that is defined as the readiness of a person to react to a situation or stressor. This is clos ely linked with commitment in that, the more a person is committed or cares, the more vulnerable he or she is to a particular threat (Lazarus & Folkman, 1984). For this study, pain and symptom distress, interference, and symptom burden from having cancer are the stressors or triggers that make the dyad vulnerable. Financial Concerns A diagnosis of cancer includes added stressors and role a lterations that may include financial impacts due to illness treatment, loss of employment and insurance issues (Pasacreta, Barg, Nuamah & McCorkle, 2000). This stressor is measured with a single indicator constructe d of four items. Pain Pain is a subjective and multidimensional experience that is determined by patients self-report. It should include intensity (severity of physical sensation), timing, quality, distress (emotional), and interference with daily functioning (Shin, Kim, Kim, Chee &
10 Im, 2007; Vallerand, 1997; McMillan & Sm all, 2002; Portenoy, Thaler, Kornblith, Lepore, Reiedlander-Klar & Kiyasu, 1994). Distress Distress is the degree or quantity of physical or mental upset, worry or suffering that is experienced or perceived. This in cludes the subjective meanings of the disease/treatment for the individual or dyad (Rhodes, McDaniel & Matthews, 1998). Interference The ability of patients to continue to participate in activities that are meaningful to them affects quality of life. Pain and dist ress can interfere with mood, walking or other physical activity, work, social relations, activity and sleep (Vallerand, Templin, Haenau & Riley-Doucet, 2007). Symptom Burden The prevalence of symptoms, frequency, severity (intensity ) and distress resulting from the disease of cancer or its treatmen t is termed symptom burden (Gapstur, 2007). Patient-Partner/Relationship Patients and their partners subject ive evaluations of the qua lity of their bond is the patient/partner relationship. Marital interaction or adjustment is the way members of the dyad communicate, behave, engage in activi ties together, express affection and the degree of satisfaction with the state of thei r marriage or partnership (Manne, Alfieri, Taylor, & Dougherty, 1999; Spanier, 1976). Fo r this study, the partic ipants are assumed to be in heterosexual, committed, intimate relationships.
11 Partner Coping Style This construct includes the part ners style of problem-sol ving, or coping, based on the situation of having a part ner with cancer who has pain or distress from symptoms. Individuals use many styles of dealing with life stressors, includi ng problem or emotion focused coping. Positive problem-solving a nd emotion focused coping or negative coping, such as avoidance and/or impulsive styles were examined in this study. Positive coping is associated with greater psychological adjustme nt and less mood disturbance while negative coping is associated with greater distress and adjustment (Ben-Zur et al., 2001; Romero, Lindsay, Dalton, Nelson & Fr iedman, 2007). This was limited to measurement from the partner only with the theoretical supposition that transaction or crossover of coping for the dyad occurs. Othe r dyadic studies have used the only the partners coping style as a mediator of distress in patient s and partners (Ko, et al., 2005). The construct of Dyadic Coping has specific instruments developed to measure dyadic participants involved in an intimate committed relationship which involves both partners and where the transmission of stress in one partner signals coping reactions of the other partner or a genuine experience of joint coping with the emphasis on the dyad rather than the individual (Revenson, Ka yser & Bodenmann, 2005). Other researchers have used data from the COPE scale collect ed from both patient and partner and used a dyadic coping model by comparing emotiona l and problem focused means of the individuals and calculated abso lute differences to assess dyadic coping (Ben-Zur, 2001). Because this was a secondary analysis, this dyadic coping measure was not available, but believed to be operant in the indirect m easure of the partners coping style.
12 Health Related Quality of Life for Patient and Partner A multidimensional assessment of health is necessary to achieve an understanding of the impact of disease on health-relate d quality of life (McHorney, Ware, & Raczek, 1993). The physical and mental he alth and distress of the patient and partner are at risk when a patient is vulnerable from a diagnosis of cancer and has pain or symptom burden. This construct included outcome measures for patient and partner for this study. Significance to Nursing Cancer impacts not only the individual, but the entire family. The role of the partner relationship has been identifie d as the most important one for many patients. Couples should be incorporated in research and in planning nursing assessments and interventions as the dyad adjusts to treatment and life st yle changes resulting from a diagnosis of cancer. Partners have as much or more distress than patients. Pain has both physi cal intensity and emotional dist ress dimensions. Social support can facilitate emotional healing. The signifi cance of the partner dyad in the arena of communication, distress, and adjustment has been examined in breast and prostate patients because of the intimacy issues involved with these cancers. There have been some studies in gastrointestinal and col on cancer patient partner dyads. Most of the research has been disease specific. No studi es have been reported that examine cancer patients with pain or symptom burden and the role of the partner relationship. This study may help provide knowledge of the importance of the dynamics of this role in outcomes and support nursing assessments and inte rventions that include partners.
13 Chapter Two Review of the Literature This chapter incl udes a discussion of the conceptual framework that guided the study and a review of the lite rature. The conceptual framework that was presented is based on the theory of individual stress, appraisal and coping by Lazarus and Folkman (1984) and then incorporates the theory of dyadic coping. Dyadic coping as developed by Revenson, Kayser, and Bodenmann, (2005) is an expansion of Lazarus and Folkmans original theory and includes or focuses on patients and their partners, as dyads, who are in an intimat e committed relationship. The review of the literature focu sed on the key concepts of cancer, distress, pain, dyads and partner relationships. The follo wing data bases were searched: Pub Med, CINAHL, JSTOR and related sources. The research studies are discussed under the topics of stressors and covari ates, mediators, and outcomes. Vulnerability due to pain and symptom burden, functional impairment and interference, financial impacts and role adjustments are discussed. Covariates are gender and age. Mediator s that are discussed are partner relationships, dyadic adjustmen t/communication, and copi ng styles. Outcomes include quality of life as physical and mental health and emotional distress for the patient and partner. A summary then concludes this chapter. Conceptual Framework The Transactional Model of St ress and Coping by Lazarus and Folkman (1984) was the framework that guided this study. This m odel evaluated the pro cesses of coping with
14 stressful events. Stressful experiences consisted of person-environment transactions. Primary appraisal consisted of the persons appraisal of the significance, vulnerability or potential harm that resulted from the stre ssor. A diagnosis of cancer, pain intensity, distress and its subsequent in terference with life activities for the dyad, and symptom burden were stressors for this study. Age, ge nder, length of the relationship and financial impacts were covariates with these stressors Secondary appraisal was the evaluation of the controllability of the stre ssor and the individuals avai lable resources that included social or cultural support. This lead to the initiation of copi ng or the strategies that were used to manage or deal with the stressor in an attempt to eliminate or minimize negative outcomes. These coping strategies consis ted of problem focused or emotion focused approaches. Outcomes that were measured included physical, mental and emotional health indicators. Figure 1. Conceptual Model S Mediator Partner Resourcefulness or Coping Dyadic Coping Assumed Mediator Partner Relationship Dyadic Adjustment Covariates Age Gender Length of Relationship Financial Concerns Outcomes Patient and Partner Quality of Life Patients Physical Mental and Emotional Health Partners Physical Mental and Emotional Health Stressors Vulnerability Pain Intensity (Worst) Distress from Pain Pain Interference Symptom Burden
15 The theory of Dyadic Coping (Rev enson, Kayser et al., 2005) is an expansion of individual stress and coping by Lazarus and Folkman (1984). Two mechanisms of dyadic coping have been postulated. The firs t sees individuals us ing coping strategies that are problem or emotion focused. Th e individuals examine the congruency and discrepancy of their partners strategies and the association with their marital or relationship satisfaction and pe rsonal well-being. The second approach views coping as a genuine dyadic phenomenon in that there is a systemic-transactional or crossover of coping between the dyad. This crossover is termed dyadic coping. Both mechanisms see dyadic coping as a process in which coping re actions of one partner takes into account stress signals of the other partner with resu lts impacting both indi viduals (Bodenmann et al., 2006). St ressors and Other Covariates Vulnerability For purposes of this study, patients and th eir partners are vulnerable from having a diagnosis of cancer, pain a nd symptom burden. Vulnerability (strength of commitment, caring or readiness to react to a stressor) in fluences the flow of events. A qualitative study using semi-structured inte rviews in focused groups of prostate cancer patients and their wives elicited four themes when inte rviewing dyads together, and patients and caregivers separately, with the same themes emerging from each group indicating congruency of the couples understanding of th e impact of the cance r diagnosis (Harden et al., 2002). Similar themes were elicited in a focused qualitative study for partners of patients with colon cancer and a stoma (Persson, Severinsson, & Hellstrom, 2003). The combined themes of these two studies are summarized as follows: (1) enduring
16 uncertainty that included str uggling with choices, interrupt ion of life and an emotional roller coaster (2) livi ng with treatment effects or the al tered body (3) coping with change that included closer family relationships and le ss socialization, shifting roles, anger at the unexpected intrusion of cancer, control issu es with communication and (4) needing help that involved professional cari ng or lack of feeling cared fo r by busy health practitioners, and lack of tailored information for their s ituation (Harden et al., 2002; Persson et al., 2003 ). In addition, Shands, Lewis, Sinshe imer and Cochrane (2006) also identified worries about the children as a core concer n for women with early stage breast cancer and their husbands. Carlson et al. (2001) assessed emotional distress in prostate and breast cancer patients using the Profile of Mood Stat es (POMS). The POMS was completed by patients and then completed by spouses as if they were the patients Partners of breast cancer patients had less accurate understanding of their wives distress than wives of prostate cancer patients. Fema le partners showed more c ongruence with their husbands answers than male partners had of women s breast cancer experiences. Investigators questioned whether gender, age and length of ti me dyads were together may have been a factor in the disparate congr uencies. Male patients were older and the dyads had, on average, been married much longer. This coul d be attributed to gender, but also may be influenced by length of time the couples were together. Partners did not complete the POMS to assess their own distress. Hagedoorn et al. (2008) also wa rn in dyadic studies men with prostate cancer are generally older than women w ith breast cancer. This makes for guarded generalizations, and age is an important consideration.
17 Cancer may change the way spouse s communicate with each other and can affect marital satisfaction and stability with added stressors including role adjustments and financial impacts. Partner and patient dyadic studies have not shown definitive relationships between age, stag e of disease and adjustment to the diagnosis of cancer (Northouse et al., 2000: Northouse, Templin & Mood, 2001). Most couple studies have been with homogeneous populations and have not considered these as variables. Some researchers assessing individual patients, not dyads, have found older patients have more distress, particularly if they also had seri ous medical problems and physical impairments, while younger patients had more financial strains and role adaptations, particularly if there was a loss of income due to illne ss (Mor, Allen & Malin, 1994; Vinokur, Threatt, Caplan & Zimmerman,1989). Schnoll and colleag ues (1998) found that the relationships of age and stage of disease to the patient s psychological adjustment, in a non-dyadic study, was mediated through the in dividuals own style of coping. Symptom Burden The symptoms that cancer patients experience have been evaluated as the persons perception and response to th e occurrence and the amount of distress or degree of emotional upset that the symptoms produ ce (Rhodes & Watson, 1987). This symptom burden is multidimensional and impacts quali ty of life (Portenoy, Thaler, Kornblith, Lepore, Friedlander-Klar & Kiyasu, 1994). Treatment relate d side effects that affect physical and mental well-being of prostate cancer patients include fatigue, incontinence, and sexual dysfunction (Hawes et al., 2006; Kno ll, Burkert, Rosemeier, Roigas & Gralla, 2007). Various treatment side effects can l ead patients to rely more heavily on their
18 spouses or partners. This can challenge the caregivers ability to adapt to an unpredictable and complex illness journey (Hawes et al., 2000). Interference Cancer is often accompanied by multiple problems, including pain, and as the disease progresses, with physical deteriorat ion. This frequently leads to social restrictions for patients and their partners and higher distress than for those who are newly diagnosed or in first remission (Manne et al., 1999; Morse & Fife, 1998). Female patients who had high physical impairment and partner active engagement (positive coping) interactions or mu tual constructive communicati on have higher relationship satisfaction despite the additio nal burden (Hagedoorn, Kuifer et al., 2000; Manne et al., 2006). Female patients reported more negative feelings when partner behavior was rated as overprotective ( protective buffering) (Hagedoorn). The relationship quality and communication, not physical impairment, seem to be key to relationship satisfaction and subsequent distress. It is po ssible that physical impairment re sults in a greater sensitivity to the reactions of others, or it may in crease the quality of communication between partners (Manne). For t hose patients and spouses who had poorer quality relationships, caregiving can be perceived as burdensome (Kim & Carver, 2007). Manne et al. ( 1999) reported in a study that incl uded 111 male and 108 female patients with advanced stage breast and gastrointestinal cancers that wives had higher levels of interference in or restriction on their personal and social activities than husbands. For participants with high dyadi c satisfaction, the quality of the relationship moderated the association between functional impairment, interference in partner activities, partner negative mood and patient perception of partner negative behaviors
19 toward them as impairment worsened. For participants with less satisfying relationships, interference was associated w ith partner distress and pati ents perception of critical attitudes toward them. Negativ e partner responses in close re lationships have a stronger impact on patient psychological distress and well-being than positive responses for patients with disease-related functional im pairment (Manne et al., 1997). Symptom burden should be considered particularly in lower quality marital relationships. Gender Associations between perceived partner support and well-being differ across genders. Most studies have examined dist ress of female patient/partner and male patient/partner at a group level (Manne, 1998; Hagedoorn, B uunk et al., 2000; Northouse et al., 1998). Because there have been no clear patterns identified, Hagadoorn et al., (2008) conducted a meta-analysis examining di stress in dyads coping with cancer. This led to the conclusion that indi viduals levels of distress were determined by gender. Women reported more distress than men regard less of their role. The authors found that there was a modest distress with in couples that provided furt her support to the idea that dyads react as an emotional system rather than as individuals. Because male partners of women who have a diagnosis of cancer have re ported less distress than female partners (Baider et al., 1998; Gilbar & Zusman, 2007; Hagedoorn, Buunk et al., 2000; Hagedoorn, Kuijer et al., 2000; Kornblith et al., 1994; So loway et al., 2004), some researchers have proposed that men are more reluctant to acknowledge threatening experiences and respond to distress using repression and distan cing strategies (Lutzky & Knight, 1994). Male patients were more distressed than h ealthy controls; thus men were able to acknowledge they had distress (Hagedoorn, B uunk et al., 2000). Therefore, Hagedoorn,
20 Buunk et al. (2000) argued that womens grea ter distress regardless of role, was due to their being more relationship oriented than men, thus were more influenced by their partners condition than male counterparts. Women in North American cultures focus more attention on relationships than men do (Acitelli & Badr, 2005) and tend to be caregivers even when they are ill (Revenson, Ab raido-Lanza et al., 2005). Men tend to be less likely than women to seek outside support and are more apt to rely on their partners for assistance (Banthia, et al., 2003; Knoll et al., 2007) thus, not surpri singly, regardless of whether women were the partner or patient th ey reported more distress and assumption of nurturing roles (Baider et al., 1998; Couper et al, 2007; Northouse, et al., 2000; Soloway, Soloway, Kim & Kava, 2004). Baider and Bengel (2001) questi on the meaning of reported findings of women having more distre ss than men and suggest the possibility of male bias implying this may be indicative of womens introspective and self-reflective nature, rather than emotional overreaction and instability. Northouse et al. (2000) examined ge nder and role differences in distress for patients with colorectal cancer in a longitudinal study at three time periods, one-week post diagnosis, 60 days (T2) and one year post su rgery (T3). Male partners had higher levels of distress than male patients at baseline, but at 60 days there were no differences in distress, indicating that male partners have most distress at diagnosis. They found low intracouple correlations for the total sample (r = .09 to r =.28), male-patient and femalepatient dyads over time, but female-patient dyads were more similar and had higher depression at T2 and T3. When examining disclosure of t houghts and feelings of ga strointestinal cancer patients to their partners, Port er et al. (2005) reported they were underpowered to detect
21 any gender differences. Female patients re ported more cancer-related communication problems than their partners but it was unclear if this difference was due to gender or role (Kornblith et al., 2006). Female patients w ho had emotional support from their partners had a greater sense of well-being while ma le patients did not show association of emotional support affecting their psychologica l outcome. Both males and females had significant correlations with nega tive behaviors from spouses leading to a lower sense of well-being and greater distress, indicating that negative partner responses were not moderated by gender (Manne et al., 1997). Mediators Patient/Partner Relationship The quality of marita l relationships has been used as a predictor in the future health and well-being of individuals with little emphasis placed on the development of illness, such as cancer, after the re lationship has been establishe d (Ren, 1997). In a summary of an analysis of the association between marital rela tionships and health problems, Burman and Margolin (1992) recommend that data suggests that part ners should be considered important elements in health outcomes, and it is the relationship qua lity, not status that should be addressed. Adaptation to stressful events is facilitated by adequate social support, and in particular if there is a strong marital relationship built on trust. Trust is critical to quality and stability in intimate re lationships and this is linked to mental and physical health outcomes (Baider et al., 1998; Cutrona et al., 2005). Gale et al. (2001), reported a study among women being evaluated for breast cancer, th at the greatest distress was in women who were in low-qua lity relationships compared to women without partners or women in high qualit y marital relationships. Patient/partner
22 relationships that demonstrate high marital qua lity frequently attribute negative behaviors to stress, and this can counteract any destruc tive impact. In contrast, when trust is low, undesirable behaviors are attrib uted to enduring characteristics of the partner, and negative interactions may occur at high rates of frequency and intens ity (Cutrona et al, 2005; Manne et al., 1997; Manne et al., 1999). Bultz et al. (2000) reported a randomized controlled trial of a brief psychoeducational support group for partners of stage I and II breast cancer patients. The men discussed coping strategies, challenges, feelings about their anxieties and fears, communication with partners about cancer and sexuality, how to provide emotional support, body image issues, genetic risks for daughters and the dying process. They were able to ask the physician questions they had not previously asked (or had been afraid to ask in the presence of their partners). Ther e was deterioration in dyadic satisfaction and level of support for the contro l group while the intervention group remained stable. They concluded that through more open communication, the partners in the intervention group may have been able to avert some of the cancer-stress related challenges for patients. As part of a larg er intervention study for prostate can cer patients and their partners (Hawes et al., 2006), four specific types of problems were id entified by partners to be worked on in individual problem solving sessio ns. The issues that were selected as the first and second priorities included: responsibil ity of financial matters if the patient dies, continuing with womens hobbies, and balance of womens illness with patients illness. Then, priorities in decreasing order were patient issues, which addresse d: patients lack of action or sloppiness, patients temper or verb al abuse, and patients smoking, exercise or
23 eating habits. They linked these to patie nts depression, dysfunctional communication and lack of demonstrating affection. Northouse, Templin, Mood and Ober st (1998) evaluated ma rital satisfaction in a longitudinal study examining malignant breast cancer versus benign br east disease effects on dyads. The groups were formed based on thei r diagnosis of benign vs. malignant masses. Both groups reported fairly high levels of dyadic satisfaction and family functioning at diagnosis. There were greater decreases for the couples in the malignant group, secondary to the greater number of illness-related demands experienced by the breast cancer patients and partners, more negative moods and poorer quality of dyadic relationships. However, for the dyads in th e malignant group who had the highest scores on dyadic satisfaction at the beginning of th e study, scores were maintained at each assessment time. A group of colon cancer patients with higher dyadic satisfac tion were found to have lower hopelessness and indirectly had less emot ional distress and role problems with adjustment. Partners relationship satisfac tion had a direct effect on emotional distress (Northouse et al., 2000). Dyadic Adjustment/Communication Although it is generally recogn ized that couples communication about cancerrelated issues and concerns plays a valuable role in both partners adaptation to cancer, this has received little empirical attention (Manne et al., 2006). Northouse and PeterGolden (1993) identified three concerns as uni versal to partners of patients with cancer: (1) dealing with the fear and threat associated with a cancer diagnosis, (2) helping their partners cope with the emotional cons equences of cancer and (3) managing the
24 adjustments in daily life caused by the disease. Couples of ten feel lost and confused about how to respond to their partners dist ress (Pistrang & Barker, 2005). Hilton (1994) found two basic communication patterns when dealing with emotional problems for patients with early stage breast cancer: (1) c ouples who shared the vi ew that talking was important talked openly and (2) couples who di d not share this view had more difficulties with their communication. Recipr ocal self-disclosure and hum or between partners in a study of early stage breast cancer patients led to less distress during videotaped discussions of cancer-related i ssues (Manne, Ostroff, Sherman et al., 2004). Kornblith et al., (2006) reported that as can cer-related communication problems increased for patients and partners, their dyadic relationshi p worsened and distress increased. Manne et al. (2006) used th e communication patterns questionnaire (CPQ) and adapted it to be cancer-specific. This cons ists of three subscales that classify communication strategies: constructive communication, mutual avoidance communication and demand-withdrawal communication (occurs when one partner pressures the other partner to talk about th e issue while the other partner withdraws or becomes passive or defensive). Constructive communication was associated with higher relationship satisfaction and lower di stress for patient and partner. Mutual avoidance communication was associated with greater distre ss for patient and partner, although the dyadic relationship satisfaction remained high. Demand-withdrawal communication considered maladaptive, was consistent with higher distress for both patient and partner and lower dyadic satisfaction. If one person in the dyad wants to ta lk about cancer, and the other does not, this mismatch causes the greatest problems with adjustment (Kornblith et al., 2006). Pr eexisting spousal communication problems do not disappear
25 when the couple is facing a life-threatening illness and may even exacerbate a crisis (Hawes et al., 2006). Manne et al. (2005) reported a randomized c ontrolled couple-focused group intervention for women with early stage breast cancer stage I to III. The intervention was designed to support communication exchange s and coping skills. Couples who were in the intervention group ha d lower levels of depressive symptoms, and women who initially had unsupportive partners and greater physical impa irment in the intervention group had the most benefit, reporting lowe r depression scores over 6 months when compared to the control group. Partners are often the comm unication conduit between patients and health professionals and key advocates during diagnosis and treatment (Davison, Goldenberg, Gleave & Degner, 2003; Harden et al., 2002). Soloway et al. (2004) examined prostate cancer couples issues of sexual function a nd psychological distress and reported that partners had higher levels of psychological distress. The wives were reticent about addressing sexual issues because they belie ved that by not openly communicating their distress, they were protecting their partners anxiety about their prostate cancer. This is an example of protective buffering Similar findings were report ed by Ezer et al., (2006). Erectile dysfunction is the most co mmon long-term side effect of prostate cancer treatment and can have implications for the pa rtner relationship (Sol oway et al., 2004). This can lead to emotional distancing and pr otective behaviors attempting to protect each others dignity. Partners rate d patients lower in ability to gain erections and sexual performance than patients rated themselves but despite difficulties with erectile dysfunction identified in the study, if dyads continued to communicate during diagnosis,
26 treatment and recovery on critical sexual issu es, the dyad was able to adjust to the sexual function outcomes of treatment. Open communication between couples may lessen distress. Researchers/clinicians do and shoul d advocate that dyads need to communicate during diagnosis, treatment and the recove ry process for the best outcomes (Badr & Taylor, 2006; Bultz et al., 2000; Maliski et al., 2002; Soloway et al., 2004). Examining patients with br east, prostate, colon, stomach, melanoma and nonHodgkins lymphoma, partners have been found to be as distressed as patients, particularly with spousal wives reporting more distress than husbands. (Baider et al., 1998; Gilbar & Zusman, 2007; Hagedoorn, B uunk, Kuijer, Wobbes & Sanderman, 2000; Kornblith, Herr, Ofman, Scher & Holland, 1994; Soloway et al., 2004). Interestingly, when comparing dyads in which both partners had a diagnosis of cancer, there was no significant difference in psyc hologic distress than when ju st one partner had cancer (Baider, Walach, Perry & De-Nour, 1998). Coping Styles Coping styles have been classifi ed in many ways, but primarily two categories are identified as overarching, problem-focused (positive) and emotion-focused (negative), with other patterns designated as subcategor ies (Ben-Zur et al., 2001; Gilbar & Zusman, 2007; Lazarus & Folkman, 1984). Problem-focus ed coping involves actions that are responses that are directed at an external event. Th is includes planning, seeking instrumental social support, suppressi on of competing activities, and positive reinterpretation. Emotion-focused coping re sponses are attempts to control the individuals emotional reactions or internal state. This includes ventilation, intrusive thoughts and avoidance strategies (Banthia et al., 2003). Avoi dant coping is denial or
27 shunning the stressor. Problem focused coping ha s been more effective at preservation of social functioning and sense of well-being (B en-Zur et al., 2001). Coping is an ongoing process, not just an isol ated event (Gilbar & Zusma n, 2007). In assessing dyadic relationships, coping strategies that have been identified are active engagement and protective buffering. Active engagement involves the patient dyad in a discussion, exploring thoughts, emotions, and initiating positive efforts at problem solving. This is seen as positive or supportive coping (Bodenmann, 2005; Hi nnen et al., 2007). Protective buffering involves hiding ones concerns, denying worries, and yielding to the partner to avoid disagreements. This type of coping is intended to reinforce or strengthen the efforts, psychological, physical and social functioning of the other partner or to increase dyadic satisfaction (Bodenmann, 2005) but it can also increase the users own distress. Some authors see this strategy as a st rength (Bodenmann, 2005; Revenson, AbraidoLanza, Majerovitz & Jordan, 2005) while ot hers see it as generally unsupportive or maladaptive (Hinnen). Protective buffering and active engagement are unrelated constructs and should be evalua ted independently. More ofte n the older patient dyad uses protective buffering whereas younger or more highly educated patient dyads use active engagement Distress is also strongl y and positively related to protective buffering and less distress was associated with more active engagement (Hinnen). Avoidance coping is significantly associated with mood di sturbance while positive coping shows no relationship to mood (Romero et al., 2007). Romero et al. found that husbands and wives perceptions of the impact of a diagnosis of breast cancer on them as a unit was incongruent for those who used avoidant coping, thus leading to higher levels of distress. Because partners can be as distressed as patients, they may hold back
28 from disclosing their own thoughts and feelings, especially when there is fear of disease progression, financial concerns or their own emotions (Baider et al., 1998; Porter et al., 2005). Regardless whether patient or partner, holding back from disclosure led to more distress. Dyads who had more disclosure between them also had better communication with health care providers (Porter) empha sizing the benefits of dyadic cohesion. Women breast cancer patients reporte d more distress than their husbands although there was a close association between patient/partners di stress (Ben-Zur et al., 2001; Gilbar & Zusman, 2007). When both spouses us ed emotion-focused strategies, there was greater distress and poorer functioning (Ben-Zur). Ko et al. (2005) assessed partne rs problem solving skills as positive problem solving and dysfunctional problem solving. Positive problem solving included positive problem orientation and rational proble m-solving measures, while dysfunctional problem-solving included a negative orientation, carelessness/impulsiveness and avoidance style. They examined partner dist ress as measured by the POMS as a mediator on patient distress, also measured by the PO MS. Dysfunctional problem solving had a negative effect on both patient and partner. Constructive problem-solving indicated less patient distress. Posttraumatic growth is described as positive changes an i ndividual may experience after a traumatic event, a nd it is relatively common among adults with cancer, particularly for younger patients (Manne, Os troff, Winkel et al., 2004). Thornton and Perez (2006) found that men treated for pros tate cancer and their partners reported modest and similar levels of posttraumatic growth, similar to Maliski et al. (2004). This growth was accounted for by coping strategies that used information-seeking, positive
29 reframing and emotional support. Cognitive and emotional processes demonstrated as caring feelings contributed to increases in patient psychological gr owth, and patients had more growth when partners had expressed more caring sentiments and communication (Manne, Ostroff, Winkel et al., 2004). Outcomes Health Related Quality of Life for Patient and Partner Couper et al. (2004) in a review of the literature of psyc hological outcomes for men with prostrate cancer found very few studies that sought data directly from partners. The POMS has been used as an outcome and medi ator in multiple couple studies; some with the partners as if were the pa tients (Romero et al., 2007) a nd some with both patients and partners (Bultz et al., 2000; Carlson, et al., 2001; Hawes et al., 2006; Ko et al., 2005; Soloway et al., 2004). Many of the studies that involve cancer patients a nd partners examine psychological outcomes that include depression as a measure of distress, such as the CES-D (Davison et al., 2003; Tuinstra et al., 2005; Hagedoorn, Kuijer, Buunk et al., 2000). Hagedoorn and Kuijer used it for patients alone along with the Rand SF-36 Physical limitations scale looking at partner support. Hagedoorn, B uunk et al. (2000) in another study used the CES-D for both patients and partners, along with a Visual Analog Scale of Quality of Life on a scale of 0 (worst imaginable life) to 10 (best imaginable life). Female patients, female partners and male patients reported more psychological distress and lower quality of life when compared to healthy controls. In contrast, male partners were similar to healthy controls in distress and QOL.
30 A study by Wagner, Bigatti and Stor niolo (2006) compared qua lity of life (QOL) for husbands who had wives with breast cancer to husbands of healthy wives. Husbands of wives with breast cancer scored lower on gene ral health, vitality, role-emotional, and mental health subscales of the Medical Ou tcomes Study (MOS-SF36) than those with healthy wives. Higher QOL for the partners of patients with breast cancer was associated with less caregiver burden, lower use of emotion focuse d coping and higher social support. Stage of disease and time since diagnos is were not related to QOL in husbands. Age and education were not different for th e husbands of wives with breast cancer and healthy wives groups. Examining post-traumatic growth in prostate cancer survivors and their partners and its subsequent affect on qua lity of life (QOL), Thornton and Perez (2006) reported modest degree in post-traumatic growth in both patients and partners with similar means. They used the Rand Health Survey (Rand-36) and found no relationshi p in post-traumatic growth to QOL for patients or partners. The Rand-36 (1992) is another version of the MOS-SF-36 (Ware, 1976) with identical subscales. Some longitudinal studies have examined breast cancer patients and colorectal cancer patients and their spouses to addr ess the correspondence or transmission of distress between patient and spouse couples administering ques tionnaires to both individuals in the dyad using the BSI and PAIS or CES-D. Results indicated that if both partners are in a high-high di stress group, they are likely to remain there one year later (Northouse et al., 1998; Tu instra et al., 2004). A study that included gastrointestinal patie nts/partners used the shortened version of the Impact of Events Scale (IES) and the Functional Assessment of Cancer Therapy
31 (FACT) as outcome measures to examine whether disclosure of concerns between patients and partners affected th eir quality of life (Porter, et al., 2005). They reported that high levels of holding back and less levels of disclosure were associated with increased distress for both pati ent and partner. Another outcome measure in coupl e studies that has been used was the Mental Health Inventory (MHI) used by Manne, et al. (1997). It has positive well-being and psychological distress subscales. Data wa s collected from patients in married relationships. No data was collected direc tly from spouses. They found associations between spouse support and psychological outc omes in patients, particularly that negative responses have a stronger impact on well-being. Galbraith, Pedro, Jaffe and Allen (2008) reported a descriptive study that compared patients/partners ranking of their quality of life with their general physical and mental health (SF-36) and marriage quality (DAS). They related these for both members of the dyad at four time intervals over an eighteen month period. Their findings showed that patients and partners quality of life was affected by the patients treatment, particularly for the emotional role at each time sequence. They concluded that partners feel the effects of the patients experiences empathetically as they go through treatment. Summary Chapter II has esta blished that the lives of both pati ents and their spousal/partners are affected by a diagnosis of cancer, and the qua lity of the partner re lationship can be the most significant social and emotional support for patients. Partners can be just as distressed or more distressed than patients. The dyad can be vulnerable from the disease and its treatment or side effects of either. For purposes of this study, patients and their
32 partners are vulnerable or ha ve stressors from having a diagnosis of cancer, which includes pain and symptom burden. Vulnerability (defined as how strong the commitment, caring or readiness to react to a stressor) influences the flow of events. Ideally, dyadic coping or the recognition of stre ss in one partner initi ates coping reactions in the other partner to lessen distress in both and preserve the relationship. Patient/Partner relationship sa tisfaction and styles of problem-solving by the partner are examined as mediators in th e health quality of life for both the patient and partner. The quality of the marital relationshi p has been shown to infl uence the adjustment of both patient and spouse. Age, gender, financ ial impacts, and functional limitations are variables that have been identified to infl uence the outcomes or quality of life as measured by distress for the dyad. General and mental or emotional health outcomes for both the patient and spouse/partn er have not been examined. Most of the studies that have examin ed patient/partner relationships with regard to a cancer diagnosis have been with disease sp ecific populations, name ly prostate, breast, colon and a few other cancers. Because th is study was a seconda ry analysis of an interventional pain study for caregivers, many varieties of cancer diagnoses, and partners and patients roles from both genders are in cluded. Distress or mood states for both patient and spouse were compared and quality of life for the dyad for both the patient and partners mental and general health were examined.
33 Chapter Three Methods This chapter presents the sample, instruments used for data collection, Institutional Review Board exemption, procedures, and data analysis plan. The purpose of this crosssectional study was to examine the stressors of vulnerability from a cancer diagnosis, pain and symptom burden on the mental and ph ysical well-being of the patients and their partners. The role of the partner relationship and dyadic coping (as a crossover of the partners coping style), were c onsidered mediators for the ou tcome of quality of life for the patients and partners general physical, mental or emotional health. This was a secondary analysis of data from a larger National Institutes of Health (NIH) funded interventional study (5R01N R008270) and a standard dyadic design (Kenny, 2006) in that each person is linked to one, a nd only one, other person in the study. Sample The sample consisted of patients and their partners accrued at a National Cancer Institute (NCI) designated comprehensive cancer center in the southeas tern United States. For this secondary analysis, the patients had to be in a committed partner relationship. Patients had to have a diagnosis of cancer in any stage, and a pain intensity level of at least 3 on a 0 to 10 scale. Patients had to be at least 18 years of age, have at least a sixthgrade education, and have no documented neurol ogic or psychiatric disorders that would interfere with self-report. Patients were excluded if they did not have a partner, or if they had psychiatric problems or were unable to read or understand English. Only
34 heterosexual partners were included in this study since the Dyadic Adjustment Scale was normed with heterosexual couples, and, like ma ny of the studies including partners, there was low participation of same sex couples. Instruments A group of instruments was used to assess aspects of the conceptual framework. The demographics include covariates of age a nd gender. Instruments that assess stressors are: the Brief Pain Inventor y (BPI), Memorial Symptom Assessment Scale (MSAS), and Caregivers Demand Scale (CDS). Mediators are included as the Dyadic Adjustment Scale (DAS) and the Social Problem-Solving Inventory-Revised (SPSI-R). Instruments that measured the outcomes are: Profile of Mood States (POMS) and the Medical Outcomes Study Short Form Health Survey (MOS SF-36). Some instruments were completed by the dyad and some by indivi dual members of the dyad. (Table 1). Demographic Questionnaire: (patient) and (partner) Demographic variables were assesse d by self-report. Variable s reported in this study were: age, race, ethnicity, gender, educati on, occupation, religion, income and length of marriage/relationship. Medical Characteristics of the Patient The medical characteristics were obtained from the chart audit that was conducted from the electronic medical record. Type of cancer was utilized for this study. Brief Pain Inventory The purpose of the Brief Pain Inventory (BPI) was to assess pain in cancer patients using self-report. It measures pain at its worst, its least, average and current level. It includes a checklist of adjectiv es to describe pain. The inst rument is scored on a 0 to 10
35 scale with general interpretation as follows: 1-3 (mild pain); 4-6 (moderate pain); and 710 (severe pain). Additional information was collected on the role pain has on interference with daily functioni ng for the patient with a range of 0-70. This is a subscale that asks patients to rate how much pain interferes with seven functional areas using a series of 0-10 point scales. Evidence of validity was presented by Serlin, Mendoza, Nakamura, Edwards, and Cleeland, (1995). They reported significant correlations of the Interference Subscale with mood disturbance items from the (Prof ile of Mood States) POMS. Serlin et al. (1995) reported relia bility as Cronbach alphas ranging from 0.860.91. Test-retest reliability was strongest for pa in at its worst (r=0.93). Pain at its worse and pain interference subscale to tal were indicators used for st ressors or vulnerability. Memorial Symptom Assessment Scale Symptom intensity and frequency is different from symptom distress (McMillan & Small, 2002; McClement, Woodgate, & Degner, 1997; Portenoy, Thaler, Kornblith, Lepore, Friedlander-Klar, Coyle, et al., 1994). The Memorial Symptom Assessment Scale (MSAS) is a self-report measure that as sesses a group of diverse symptoms that are common in cancer patients by differentiating frequency, intensity and distress. Distress items are scored from 0 (no distress) to 3 (wor st distress). Patients score distress for each of the 25 symptoms that they endorse as being present. A summation of all the items or total score gives the Global Distress Scor e (GDS). The revised version (Portenoy, Thaler, Kornblith, Lepore, Friedlander-Klar, Kiyasi, 1994) was used for this study. Factor analysis confirmed two factors that di stinguish three groups of symptoms as Psychological, High Prevalence Physical Symptoms and Low Prevalence Physical Symptoms. Reliability coefficients have been reported as indicating strong internal
36 consistency for the subscales (alphas = 0.830.92). Individual sympto ms also provide a detailed description that incl udes severity, frequency and dist ress that can be particularly relevant in some circumstan ces (Portenoy, Thaler, Kornblith, Lepore, Friedlander-Klar, Kiyasi, 1994). The Global Distress Score (GDS ) was used as an indicator of symptom burden. Pain severity and pain distress were in dicators of vulnerability from the specific symptom pain in the SEM model for this study. Caregiver Demands Scale The Caregiver Demands Scale (CDS ) (Stetz, 1987) is a 46 item self-report measure of demands that caregivers may experience along two dimensions (physical caregiving and role alterations). Four questions under the subscale Financial Alterations will be utilized for this study. The spouse/partner firs t answers yes or no. If yes, then they rate the demand on a 5 point Likert-type scale (1= not at all to 5 extremely stressful). The total score is summed, with higher scores representing greater perceived demands. Internal consistency reliability (C ronbachs alpha) has been reported as 0.87. This instrument has been used in cancer populations (Pasacreta, Barg, Nuamah & McCorkle, 2000). Dyadic Adjustment Scale The Dyadic Adjustment Scale (DAS) is a thirty-two item scale that is designed to assess marital satisfaction for couples married or in a committed relationship (Spanier, 1976). This instrument measures dyadic c onsensus, dyadic satisfaction, affectional expression and dyadic cohesion that direc tly assess general communication between couples and multiple items in which comm unication patterns are undercurrent. Higher scores indicate greater marita l satisfaction. Alpha coeffici ents have been reported from
37 0.92 to 0.95 (Northouse et al., 1998). Construct and criterion validity and reliability were reported by Spanier (1976). Patient DAS and Partner DAS scores were used as indicators for both SEM models. Social Problem-Solving Inventory Revised The Social Problem-Solving Inventory-Revised (SPSI-R) (DZurilla, Chang, Nottingham & Faccini, 1998) is a measure that consists of 52 items that use a five-point Likert-type scale that assesses problem-solvi ng skills. There are five subscales: (1) Positive Problem Orientation; (2) Negative Problem Orientation; (3) Rational Problem Solving; (4) Impulsivity/Carele ssness Style; and (5) Avoidanc e Style. A total score is calculated to reflect overall problem-solving ability with high er scores indicating better problem-solving ability. Construct validity was reported (r = 0.33 to 0.75) with subscales from other problem-solving measures. Internal consistency using alpha coefficients have ranged from 0.69 to 0.96. This instrument has been used with cancer patients. The Positive Problem Orientation, Impulsiveness/C arelessness and Avoidance subscales were used as indicators in the SEM model. The Profile of Mood States The Profile of Mood States (POM S) (McNair, 1992) is a 65 item Likert-type scale that evaluates six affective states: (1)T ension-Anxiety; (2) Depression-Dejection; (3)Anger-Hostility; (4) Vigor-Activity; and (6 ) Confusion-Bewilderment. Internal consistency has been reported as 0.87 to 0.95 within these subscales. Test-retest reliability ranged from 0.65 to 0.74 over a th ree-week interval. A global measure of affective state, or total mood disturbance sc ore can be calculated by summation of the six subscale scores. High scores except for th e Vigor-Activity subscal e indicate worse mood
38 with lower scores indicative of better mood. This global sc ore was used for the patient and partner as a qual ity of life outcome. Medical Outcomes Study Short Form Health Survey The Medical Outcomes Study ShortForm Health Survey (MOS SF-36) (Ware et al., 1993) is a thirty-six item self-report multidim ensional rating scale that measures eight health-related concepts: (1) physical functioning; (2) role limitati ons due to physical health problems; (3) bodily pa in; (4) general health; (5) vitality (energy/fatigue); (6) social functioning; (7) role limitations due to emotional problems; (8) mental health (psychological distress and psychological wellbeing). The items use Likert-type scales with five or six points and a few with two or three points. Construct validity correlations range from r = 0.52 to 0.78 for subscales with other quality of life measures. Test-retest reliability coefficients for the subscales ra nge from 0.68 to 0.93. Factor analytic studies identified two summary scores: the Physical Health Component Score and the Mental Health Component Score. In the original proposal, the Physical Health Component Score and the Mental Health Component Score were to be used as indicators for Quality of Life for the patient and partner. Due to difficulties with the lambda loadings in the Structural Equation Model, the General and Mental Hea lth subscales were substituted to give a broad overview of the couples well-being.
39 Table 1 Measures Completed by Each Member of the Dyad Variables Measures Patient Partner Covariates (Demographics) Demographi c Questionnaire x x Age Demographic Questionnaire x x Gender Demographic Questionnaire x x Income Demographic Questionnaire x x Length of Relationship Demographic Questionnaire x x Stressors (Vulnerability) Worst Pain BPI x Distress MSAS x Interference BPI (i nterference subscale) x Financial Concerns CDS (financial concerns subscale) x Potential Mediators Relationship Quality DAS x x Partners Resourcefulness SPSI-R x Outcomes (QOL) Emotional Distress POMS x x Physical Health SF-36 x x Mental Health SF-36 x x
40 Institutional Review Board The parent interventional study received approval by the Protocol Review and Monitoring Committee of the H. Lee Moffitt Ca ncer Center and received approval from the University of South Florida Institu tional Review Board (IRB Number 100675). No further patient/partner contact nor chart review was made, thus no harm was done to participants. No new IRB approval was needed for this study. Procedures Patients were invi ted to participate in the study while attending regularly scheduled outpatient appointments at the cancer cente r. The study was explained to both the patients and their partners, questions were answered and written consent was obtained. Questionnaires were then completed individu ally by each participant and given to the research team to be entered into the data bank. For this study, the data set was obtained from the Principal Investigator. The data was cleaned. Five same sex partners were eliminated from the study. Four dyads had large sections of missing data. These were also deleted. The remaining dyads had less than 95% missing data characterized as missing completely at random. Mean substitutions were made for the missing values (Tabachnick & Fidell, 2007). Data Analysis Demographic data, type of cancer and stage of disease are re ported using descriptive statistics, including frequencies and descriptive statistics of means, ranges and standard deviations (SD). Aim 1, to examine a pa tient and their partner when challenged by stressors of pain and symptom burden and its affect on their QOL are addressed as follows. Question 1: Are the covariates such as age, gender, financial concerns and length
41 of the marriage related to the relationship qua lity or quality of life outcomes? This was answered by multiple regressions analyses. The covariates were entered with the DAS as the dependent variable for the patient. Th e partners DAS score was entered first, then age, gender, financial concerns and length of marriage were entered. This same procedure was then be used with the Partne rs DAS score as the de pendent variable and the covariates with the Patient DAS score entered first. These same multiple regression analyses of covariates was conducted for the dyads for the POMS, General Health and Mental Health scores which are the outcom e measures. The patient and partner both need to be considered in each regression analysis because they are considered an interdependent model. Patients and their partners have the same exposure within a household, thus are not indepe ndent of one another (Tabac hnick & Fidell, 2007). The significant covariates were th en addressed in the SEM mode l and subsequent analysis. Questions two through four of AIM 1 were answered by the SEM model. Correlations, means, and standard deviations for the indicator variables were completed using Statistical Package for the Social Sc iences (SPSS) (version 16.0) and are reported. These results were examined for significance. Raw data sets of the indicator vari ables were then entered into LISREL with Prelis. This was then used in a series of structur al equation models to assess if the partner relationship or partner resourcefulness mediated the out comes of distress and QOL. Latent variable path analysis uses the eight matrices of LISREL combined to consider measurement of the variables and their struct ural relations (Kellowa y, 1998). The fit of the measurement model is conducted first and provides a baseline for the fit of the full latent model. A full measurement model was created by loading the appropriate
42 indicators on the latent constructs with all constructs treated as endogenous. The indicators for each construct n eeds to correlate fairly highly with each other to assure assessment of the same construct. The Positive Problem Orientation and the Impulsiveness/Carelessness and Avoidance scales were correlated, but negatively. The negative indicators were reversed scored by multiplying by -1 so that the correlations were in the same direction. Therefore, nega tive coping was interpreted by low scores on these two scales. Both variables of positive problem solving and negative coping were still present. The same procedure was done for the POMS, so that the quality of life outcome measures were in the same direc tion. The LISREL program was then able to provide calculations with few iterations and good fit indices. This full measurement model provides correlations for all the constructs and is the best model that can possibly be obtained with the data (Kelloway). Once that was established, further models were structured to examine partially mediated, fully mediated and non-mediated pathways. The structural equation model (SEM) computed significance for the direct and indirect pathways and calculated goodness of fit indices for the mediation models. Kenny, Kashy and Cook (2006) recommend multi-level modeling or structural equation modeling when conducting dyadic analysis. Question 2: Is pain or symptom burden related to the dyadic committed relationship? This was a direct pathway from the stressors to Quality of Partner Relationship. Question 3: Does pa in or symptom burden impact Quality of Life for the patient and partner in terms of emotional distress and overall mental and general health? This was also a direct pathway from the st ressors to the patients QOL outcomes and a direct pathway to the partners QOL. Question 4: Does the Quality of the Relationship mediate the outcomes of emotional distress and overall mental and
43 general health of the individuals? This is answered thr ough the pathways set for fully mediated, non-mediated and partially mediated models. Aim 2 seeks to examine the c oping style of the partner when challenged by the patients pain or symptom burden and the eff ect of the coping style on the outcomes for the patient and partner. Thes e are the questions for Aim 2. Question 6: Does pain or symptom burden have a relationshi p to the coping style of the partner? This is a direct pathway from the stressors to Partners copi ng. Question 7: Does the coping style of the partner affect their own or their partners dist ress, mental and genera l health? This is a direct pathway from the stressors to the QOL for each member of the dyad. Question 8: Does the partners coping style mediate the QOL outcomes for the patient and partner? This theory assumes the partners characteristics (e.g. coping st yle) affect his or her own score on an outcome variable (e .g. distress, general health, ment al health) and also affects the patients outcome scores on the same variables. Questions 6 through 8 were also included as constructs with indi cators in the SEM full measurement model and were included in the pathways for direct effects, non-mediated, partially mediated and fully mediated models as above. (See Figure 2. The Statistical Structural Equation Model). Several fit indices are reported to justify the fit of the mode ls. These are discussed in the results section under the Goodness of Fit Indices section with the discussion of each of the latent path models.
44 Partner Resourceful ness/Coping Quality of Partner Relationshi p Patient QOL General Mental Emotional Health Vulnerability Financial Concerns Partner QOL General Mental Emotional Health Figure 2. Structural Equation Statistical Model. Aim 1, Question 5: Are the patient s and partners distress related to each other and are there any differences for gender? This was answered by a multivariate repeated measures design that examined the dyad as th e subject in a two by two design of role by gender. The POMS for the patient and the PO MS for the partner as the within subject factor and patient gender as the between subj ect factor was conducted. This was repeated for the POMS of the partner and POMS of the patient as the within subject factor and the partners gender as the between subject factor Because patients and their partners are not independent of one another and the dyad is considered as the subject in this multivariate analysis the patients score was time 1 and the partners score was time 2, then vice versa (Tabachnick & Fidell, 2007).
45 Kelloway (1998) recommends a sa mple size of at least 200 in order to have sufficient power, but a ratio of sample size to parameters should be between 5:1 and 10:1. There was sufficient power for the models with all participants included, but insufficient power for the gender analysis when trying to divide the models for partner coping by gender. Because gender could not be addressed in SEM, the instruments completed by one member of the dyad were examined for differences in gender by ANOVA and the same multivariate time series analysis used to detect role and gender differences in the POMS explained in the analysis for Aim 1, Question 5 was used for the variables completed by both members of the dyad.
46 Chapter Four Results Chapter Four presents the study re sults. Results are organized as follows: (1) the sample demographics, types of disease, descriptions of pain and symptoms; (2) description of the results for the covariat es for the study; (3) comparison of role by gender for the distress measure of the PO MS; (4) results of the SEM models for mediation outcomes and (5) the results of the quality of life outcome measures for the patient and partner. Profile of Sample The sample consisted of 354 part icipants, 177 dyads of heterosexual couples in intimate, committed relationships. There were 102 male patients/partners and 75 female patients/partners. The individuals in the c ouples were similar to each other in age, education and socioeconomic level. The couples were in relationships for a mean of 28 years and a range of 10 months to 60 years. The mean age of ca regivers 55 years was similar to and patients mean age was 57 years. There was a wide range of ages of the participants with the youngest being 22 years and the oldest 81 years. The mean income was $40,000 to $60,000, with 30% of the sample greater than $60,000. Fifty percent of the sample were classified as professionals or administrators/managers. Ninety percent of the sample was White, with only 2% Bl ack and 6% Hispanic. They were highly educated for a general population with the mean education level of 14 years for both partners and patients (Table 2).
47 Table 2 Demographics for Patient/Partner Dyads Variable Patients Partners Age in years Mean 57.49 55.50 Standard Deviation 12.33 12.42 Range 24.78 to 80.75 22.46 to 79.04 Family income $10,00119,999 4% $20,000 39,999 30.7% $40,000 59,999 16.0% $60,000 100,000 29.3% >$100,000 5.3% Prefer not to answer 14.7% Length of Relationship in years Mean 28.16 Standard Deviation 14.73 Range <1 to 60 The patients were representative of a wide spectrum of different primary cancer diagnosis. Lung cancer was the most prev alent, followed by head and neck cancer patients, then breast cancer (Table 3).
48 Table 3 Prevalence of Cancer Types Type of Cancer Fr equency Percent Lung 45 26.2 Head and Neck 22 12.8 Breast 18 10.5 Colorectal 14 8.1 Gynecological 13 7.1 Leukemia/Lymphoma 8 5.0 Pancreatic 8 5.0 Sarcoma 8 5.0 Unknown Primary 7 4.1 Prostate 7 4.1 Bladder/Renal 4 2.2 Carcinoid 3 1.7 Melanoma 2 1.2 Multiple Myeloma 2 1.2 Gastric/Esophageal 2 1.2 Testicular 1 .6 Data for staging of cancers was diffi cult to obtain from the chart audits and was not considered reliable; however, the current status of cancer therapy treatment was included in the chart audit from the original study. Fo rty-five percent of the participants were considered stable whether they were receivi ng chemotherapy or radiation or just at the
49 center for surveillance. Thirty-seven percent of the patients with pain in the study were receiving treatment but continued to have progressive disease. Ten percent of the patients were considered to be in remission, but continued to have symptoms and pain from the cancer treatment, surgery, chem otherapy or radiation (Table 4). Table 4. Current Status of Cancer Therapy Treatm ent: Frequency and Percent of Patients Treatment Status Frequency Percent Under treatment, progressing 66 37.2 Under treatment, stable 61 34.4 No treatment, stable 20 11.2 Remission 17 9.6 Missing 7 4.1 No treatment, progressing 6 3.5 Covariates Question 1, Aim 1 asks if the covariates age, gender, le ngth of relationship, and financial concerns affect the quality of the rela tionship or the quali ty of life for the patient/partner. The covariates were entered into a regression model separately in stepwise regression and together as predicators for patient/partner for the variables Dyadic Adjustment Scale (DAS), POMS, General Health, and Mental Health. The results were the same regardless of entry in the equati on so only the model with all the variables entered simultaneously is shown. This proced ure was done to see if the covariates should be included in the SEM model. These results are reported in Table 5 and Table 6.
50 Table 5 Regression Model for Covariates with Patie nt Dyadic Adjustment Scale (DAS), POMS, General, and Mental Health Scores from the SF-36 Models R 2 Standardized Model 1 Patient DAS Dependent Variable .335 Partner DAS .525 .559* Patient age -.074 -.061 Length of relationship .060 .059 Financial Concerns -.174 -.078 Gender -2.111 -.070 Model 2 Patient POMS Dependent Variable .177 Partner POMS .242 .234* Patient age -.045 -.018 Length of relationship .045 .021 Financial Concerns 1.167 .253* Gender 6.845 .110 Model 3 Patient General Health .113 Dependent Variable Partner General Health .236 .225* Patient age .106 .073 Length of relationship .012 .010 Financial Concerns -.447 -.167* Gender -.595 -.016 Model 4 Patient Mental Health .066 Dependent Variable Partner Mental Health .196 .187* Patient age -.059 -.037 Length of relationship .065 .048 Financial Concerns -.366 -.122 Gender -3.580 -.089 *p < .05
51 Table 6 Regression Model for Covariates with Partner Dyadic Adjustment Scale (DAS), POMS, General, and Mental Health Scores from the SF-36 Models R 2 Standardized Model 1 Partner DAS Dependent Variable .325 Patient DAS .598 .561* Partner age .045 .035 Length of relationship -037 -035 Financial Concerns -.103 -.043 Gender -.453 -.014 Model 2 Partner POMS Dependent Variable .228 Patient POMS .218 .225* Partner age .298 .125 Length of relationship -.191 -.095 Financial Concerns 1.590 .358* Gender 4.530 .076 Model 3 Partner General Health .115 Dependent Variable Patient General Health .214 .225* Partner age -.055 -.040 Length of relationship -.086 -.074 Financial Concerns -.574 -.225* Gender -.174 -.005 Model 4 Partner Mental Health .138 Dependent Variable Patient Mental Health .163 .170 Partner age .069 .045 Length of relationship -.072 -.056 Financial Concerns -.693 -.243* Gender -6.871 -.179* *p < .05
52 Because the patient and partner are not independent, the partners or patients corresponding variables were also entered into the equation as predictors and these were significant in each model as expected. None of these variables covaried significantly with the Dyadic Adjustment Scale. Fina ncial Concerns did have a significant negative effect with the patients and partners quality of life scor es. The partners gender was also a significant negative covariate for part ners mental health outcomes with women having lower mean scores than men in subsequent analyses. Because Financial Concerns was significant for the couples quality of lif e, it was included in the SEM model as a single indicator latent construct and was considered a st ressor. The SEM model was also run for the partners coping style separate ly for gender, but was not powered enough when the sample was split to give an adequate fit to be evaluated as a covariate in the mediation models. Because gender was consid ered a significant variable for the study, further analysis of gender differences was completed using MANOVA for variables completed by both members of the dyad and ANOVA for those measures completed only by one of the individuals. This is di scussed further in the section headed Gender and Role. Stressors To be eligible to participate in the study the patients had to have had a worst pain score of at least a thre e on a scale of 0 to 10 at the time of consent. Fifty-five percent of the sample characterized their pain as conti nuous and forty-five percent described it as occasional. A summary of pain, including in terference, intensity, distress and symptom burden indicators means, ranges and standard deviations are provided. (Table 7). Two
53 of the patients had scores of 2 on the BPI and one had 1. They did not meet the eligibility criteria at the time of admission, but were retained in the study. Table 7 Ranges, Means and Standard Deviations for Worst Pain, Pain Intensity, Pain Distress, Interference, and Symptom Burden Measure N = 177 Possible Range Mean Standard Deviation Worst Pain (BPI) 0 to 10 7.89 2.007 Pain Intensity (MSAS) 0 to 3 2.14 .838 Pain Distress (MSAS) 0 to 3 1.94 1.034 Interference (BPI) 0 to 70 37.667 17.618 Global Distress Score (MSAS) 0 to 60 16.57 11.95 In addition to pain, the patien ts also completed the MSAS that assessed the presence, intensity and distress of 25 symptoms This is the measure used to define the latent construct Symptom Burden. Table 8 shows the frequency of the symptoms and percentage of patients affect ed for the most commonly reported symptoms. Pain is the primary symptom for these patients and should have been present in all 177, because it was the inclusion criteria to pa rticipate in the stu dy. Seven of the patients did not select pain as a symptom, although they had at least a score of 3 on the BPI. This measure did ask about pain over the last week. Symptom burden was used in the SEM model, but was highly correlated (r = 0.81) w ith vulnerability from pain. It did not add to the understanding of vulnerability. Therefore, it was not used in the SEM analysis so that a more parsimonious model would be explored.
54 Table 8 Number and Percent of Most Frequently Reported Symptoms Symptom Frequency Percent of patients affected Pain 167 94 Muscle weakness 112 63 Sleep problems 105 59 Emotional upset 101 57 Drowsiness 101 57 Constipation 95 54 Nausea 83 47 Change in appetite 81 46 Numbess 78 44 Problems with concentration 76 43 Lightedness/Dizzyness 71 40 Mediators The Structural Equation Model Aim 1, questions 2 through 4 addre ss the role of mediation of pain and financial concerns by the marital relationship, Aim 2 que stions 6 through 8 address the role of mediation by the coping styles of the partne rs on the quality of life outcomes for the patients and partners. The st ructural equation model addresses both of these questions. The means, standard deviations and correla tions for the variables were calculated in SPSS and are reported in tables 8 and 9. For this study, the raw data set containing the
55 fifteen variables were imported from SPSS into the LISREL program. The physical and mental health component scores (standard ized scores from the SF-36) although not included in the model are provided for comparis ons to other studies. These standardized scores were replaced by the general health and mental health subscales for better fit in the structural model, without losing the outcome measures that were sought. SEM is a statistical procedure de signed to evaluate latent constructs that have measured indicators in a multivariate model using covariance matrices. An original covariance matrix from the data set is replic ated through a series of iterations to give values to the indicators and specified pathways in a reprodu ced covariance matrix. Error variance is also measured in the structural equation model (Ratner, Bottorff, & Johnson, 1998). LISREL output reports each parameters effect size and the standard error of estimate for that parameter. The ratio of the pa rameter effect size to it s standard error is a reported as a t-test. Because the sample sizes are presumed to be large (at least > 164), the t values are interpreted using critical values for the z test or standardized scores. Any value + 1.96 was considered to be significant at the 95% confidence level (Kelloway, 1998). Each construct was made up of multiple indicators and in this study, some were completed by the patient, some by the partner, and some by the dyad. The exogenous variables are no t caused by other variab les. The endogenous variables are mediating variable s or outcome variables. For this study, cancer pain with three indicators and financial concerns with one indicator are both the exogenous variables. The middle endogenous or mediators are relationship quality with two indicators and coping style of the partner with three indicators. The outcomes or endogenous variables are the patient quality of life with three indi cators and partner
56 quality of life with the same three indicator s. The full measurement model and the latent pathway models are discussed: the fully mediated model, non-mediated model and the partially mediated model. A summary of the goodness of fit indices for each model is reported after discussion of each individual mode l. Ranges, means, standard deviations and correlations for the indicator vari ables are in Table 9 and Table 10. The Impulsive/Carelessness, Avoida nce and POMS are not reversed scored in these tables. Higher scores mean higher values of negative coping and distress. Twelve percent of the partners scored below 100 on the Dyadic Adjustment scale, with 10% of the patients falling below 100, i ndicative of poorer quality relationships. Most of the couples who scored lower did not have scores that corresponded with their partners. Couples who are at risk for divorce have scores in the 70s and most married couples have a mean score of 114.8 (Spanier, 1976). Three partners and one patient had scores below 80. Ten patients and seven partners had scores greater th an or equal to 145, indicating very high quality marriages or relationships. Two caregivers had scores on their physical health less th an the patients corresponding sc ores. Twelve partners and twenty patients had scores greater than 70 on the POMS, indicating greater distress. Twenty-seven patients had scores less than 0 on the POMS and thirty partners had scores of less than 0 on the POMS. These were very low distress scores. The patients Physical Health Component Scores were two standard deviations below the mean on the SF-36. The patients and partners Mental Health Co mponent Scores and the partners Physical Health Component Scores were within one standard deviation of the mean. When examining scores that fell well below the mean s for all the indicator variables, there were no clear patterns to explain the high and low ranges of the variables.
57 Table 9 Means, Standard Deviations and Ranges for Indicator Variables Used in Study (N=177) Variable Ranges Mean Standard Deviation Financial Concerns 0 to 20 5.70 6.67 Pain Distress 0 to 3 1.94 1.03 Pain at worst 1 to 10 7.89 2.01 Pain Interference 0 to 70 37.07 17.62 Patient Dyadic Adjustment 79 to 151 121.82 14.96 Partner Dyadic Adjustment 66 to 148 118.88 15.94 Positive Problem Solving 2 to 20 12.28 3.75 Impulsive/Carelessness 0 to 27 7.17 5.09 Avoidance 0 to 22 4.70 3.77 Patient POMS -37 to 117 29.44 30.74 Patient General Health 12 to 97 51.82 17.88 Patient Mental Health 4 to 100 67.95 19.93 Partner POMS -38 to 124 25.08 29.67 Partner General Health 17 to 97 69.69 17.00 Partner Mental Health 12 to 100 73.08 19.02 Patient Physical CS* 15.60 to 56.19 28.84 8.25 Patient Mental CS* 19.72 to 68.52 45.67 11.81 Partner Physical CS* 17.34 to 61.66 45.39 9.13 Partner Mental CS* 14.26 to 68.12 49.23 9.74 *Note: CS indicates Component Score
58 Table 10 Bivariate Correlations for Vari ables Included in SEM Model FC DIS WP IF PDAS CDAS PPO IC FC 1.000 DIS .214** 1.000 WP .224** .541** 1.000 IF .199** .619** .582** 1.000 PDAS -.145 -.165* -.187* -.173* 1.000 CDAS -.127 -.031 -.078 -.082 .568** 1.000 PPO -.150* .087 -.111 -.007 .268** .385** 1.000 IC -.019 -.015 .036 -.009 -.300** -.196 -.289** 1.000 AV .086 -.067 .015 -.060 -.232** -.228** -.440**.500** PPOM .394** .314** .347** .471** -.324** -.171* -.066 .132 CPOM .420** .186* .131 .112 -.410** -.334** -.210** .155* Note. FC = Financial Concerns; DIS = Distress; WP= Worst Pain; IF = Interference; PDAS = Patient dyadic adjustment scale; C DAS = Partner dyadic adjustment scale; PPO = Positive problem solving; IC = Im pulsive/Carelessness; AV = Avoidance; PPOM = Patient POMS; CP OM = Partner POMS; *Correlation significant at the 0.05 level (2 tailed) **Correlation significant at the 0.01 level (2 tailed)
59 Table 10 (Continued) FC DIS WP IF PDAS CDAS PPO IC PGEN -.248** -.217** -.283** -.342** .169* .153* .250** .030 PMEN -.175* -.360** -.315** -.472** .371** .229** .098 -.112 CGEN -.245** -.133 -.081 -.100 .269** .250** .260** .009 CMEN -.276** -.081 -.108 -.047 .276** .324** .356** -.089 AV PPOM CPOM PGEN PMEN CGEN CMEN AV 1.000 PPOM .044 1.000 CPOM .242** .333* 1.000 PGEN -.112 -.346** -.140 1.000 PMEN .004 -.716** -.253** .390** 1.000 CGEN -.123 -.219** -.423** .265** .177* 1.000 CMEN -.184* -.258** -.680** .153* .204** .435** 1.000 Note: FC = Financial Concerns; DIS = Distress; WP= Worst Pain; IF = Interference; PDAS = Patient dyadic adjustment scale; C DAS = Partner dyadic adjustment scale; PPO = Positive problem solving; IC = Im pulsive/Carelessness; AV = Avoidance; PPOM = Patient POMS; CPOM = Partner POMS; PGEN = Patient general health; PMEN = Patient mental health; CGEN = Part ner general health; CMEN = Partner mental health *Correlation significant at the 0.05 level (2 tailed) **Correlation significant at the 0.01 level (2 tailed)
60 The correlations should be examin ed prior to proceeding with the SEM even though theory guides the model. Financial concer ns did not correlate significantly with the Relationship quality variables, the DAS scores Vulnerability from pain correlated with the patients Dyadic Adjustment scores, but not the partners. Distress and the patients and partners POMS were significant. Correla tions from the stressors to Partner coping was weak r = .15 for Financial concerns. Th ere were no correlations from Vulnerability which was distress, intensity and interfer ence to coping. Ther e were significant correlations between the Quality of the rela tionship and the Partner coping styles. The outcome QOL variables for the patient a nd partner correlated significantly with relationship quality. Vulnerabil ity has a relationship to the patients quality of life, but only distress for the partners was significan t with pain. Financial concerns have relationships with both patie nts and partners QOL. Goodness of Fit Indices Models should have three aspects of fit that include a theoretical basis, empirical evidence and parsimony (Ratner, 1998). Th e model was conceptu alized based on the review of the literatu re, and theory guided its construction. The goodness of fit indices are reported for each model for empirical evidence using the following goodness of fit criteria: the chi-square to degrees of fr eedom ratio, the root mean squared error of approximation (RMSEA), the comparative fit index (CFI) and the standardized root mean squared residual (SRMR). The chi square ra tio to degrees of freedom should not be significant, indicating that the implied covariance is not different from the observed data set, but this varies with sample size and is not a definitive test (Ratner, 1998). The RMSEA is based on an analysis of the resi duals and should be le ss than 0.10 for a good
61 fit and values of < 0.05 for a very good fit. Th e SRMR is the standard ized square root of the mean of the squared discrepancies be tween the implied and observed covariance matrices and have 0 as a lower bound and 1 as upper bound with values <.05 considered a good fit (Kelloway, 1998). The CF I is a relative measure of fit that compares the null model to the estimated model. The CFI should be at least .95 and if it is less than .90, it is a poor fit. (Kenny, 2006). (Table 11). Full Measurement Model In the full measurement model each indicator is loaded on its corresponding construct and no pathways are set between constr ucts. Each latent construct is correlated with every other latent construct. This determines if the indicators are appropriately measuring the constructs and if the model fit is suitable to proceed for further evaluation. LISREL works better with multiple indicators than with single indicators (Diamantopoulos & Siguaw, 2000). The parameters for the full measurement model were acceptable for a good fitting model (Figur e 3). This model has the best fitting indices that were possible with this data set. Statistical significance is designated for the correlations and indicators that were set to be freely estimated in the program. At least one indicator for each construct is set to 1 for a starting value, thus significance is not given for that indicator. This allowed for the subsequent mode ls with path analysis to evaluate for mediation of the marital relationship and coping style of the partner. Fit indices are provided for each model to assess for acceptance or rejecti on of the theoretical constructs. For the fully mediated, non-mediated and partially mediated models, Financial Concerns and Vulnerability from Pain are the exogenous c onstructs and Relationship quality, Partners
62 coping style, and Quality of Life for patient and partner are the endogenous constructs. Direct and indirect pathwa ys were set in LISREL to evaluate for mediation. Figure 3. Full Measurement Model Non-Mediated Model The non-mediated model had the tw o exogenous constructs of Financial Concerns and Pain to each of the endogenous variables as direct pathways. Relationship quality and Partner coping style had no pathways set to mediate the QOL outcomes. Financial concerns had significant direct effects on patient and partner QOL outcomes. Vulnerability from pain had significant eff ects on Relationship quality and Patient QOL. There were no significant eff ects on Partner coping or Part ner QOL. The fit indices for
63 this model, though using less degrees of freed om had unacceptable fit to explain the data, thus mediation was present. (See Figure 4). For clarity of the diagrams, the i ndicators are not include d. Their significance and values are essentially the same as indicated on the full measurement model. The direct and indirect pathways with the standa rdized coefficients are as illustrated. Financial Concerns Relationship Quality Vulnerability from Pain .04 Partner Coping St y le Partner QOL Patient QOL -.07 -.17* -.60* -.41* -.11 -.24* -.12 Figure 4. Non-Mediated Model Note: Direct Effect indicated by Fully Mediated Model The fully mediated model had dire ct pathways set from the exogenous variables, Financial concerns and Vulnerability from Pain to the mediators, Relationship quality and Partner coping style. Then indirect pathways were set from the mediators to the outcome variables PQOL and CQOL. There were no dir ect pathways set to the outcome variables.
64 This model also did not have good fit indices, so exclusive mediati on did not explain the relationships. (See Figure 5). There were significant direct pathways from both Financial concerns to Relationship quality and then to both patient and partner QOL. Vulnerability also had direct significant pathways through Relationship quality to the QOL outcomes for the couple. Coping did not mediate either exogenous variable. Financial Concerns -.24* Relationship Quality Vulnerability from Pain Partner Coping St y le .08 -.13 -.31* Partner QOL -.12* .60* .18* .61* Patient QOL Figure 5. Fully Mediated Model Partially Mediated Model The partially mediated model added direct pathways in addition to the mediation pathways. This model gave the best explan ation for the relationships of the variables with acceptable fit indices. (See Figure 6) Partial mediation was the best model empirically and is discussed in detail. Th ere were significant direct pathways from Financial concerns to both PQOL and CQOL. There was a negative pathway to
65 Relationship quality, but it was not significant, so Relationship quality did not mediate the direct effects of Financial concerns. The direct pathway from Pain to patients QOL was significant, however, there was no effect on the partners QOL directly. Pa in had a significant direct effect on the Relationship quality, and an i ndirect effect on PQOL through Relationship quality, with a reduction of the direct effect meeting the criteria for partial mediation. In order for mediation to be present, th ere are three conditions: 1. va riations in the independent variable have significant pathways to the presumed mediator which was the Relationship quality 2) Variations in the mediator have si gnificant pathways to the dependent variable or patient QOL and 3) the direct pathway fr om the exogenous variab le to the dependent variable, in this case, patient quality of life should become 0 for full mediation or decrease for partial mediation. (Baron & Kenny, 1986). There was a decrease on the direct effect from -.51 to -.58 and this wa s a significant total effect. Since there was no direct effect of Pain on the partner, th ere was no mediation for the partner. There were no sign ificant pathways from Pain or Financial concerns to Partner coping, thus it was not a mediat or for patient or partner quality of life. There was a significant direct relationship from Partners coping style to the partners QOL, but not for the patients QOL, thus indicating that th e partners coping styl e affected only their own QOL. There was a significant correlati on for the patients general health and positive problem solving that did not manifest in the SEM model. Changing Partner coping to an exogenous variable would eliminate pathways from Financial concerns to Relationship quality, Pain to coping, and c oping to patient QOL. It would be a more parsimonious model, but then coping would not be a mediator which was the theory
66 guiding the questions in Aim 2. The standardized total e ffects of the endogenous and exogenous variables are summarized in Table 12. Six percent of th e variance in this model examining Financial concerns and Pain as stressors was explained by the positive effect s of the quality of the relationship in contributing to the patients QOL, 1% was explained from the coping by the partners having a positive effect on their own QOL, 52% of the variance was explained by the effects of the stressors on the patients qual ity of life and 42% of the explanation was due to the effects on the part ners quality of life. Financial Concerns Relationship Quality Vulnerability from Pain Partner Coping St y le Partner QOL .34* -.36* -.08 ..44* .16* -.12 -.51* -.12 -.19* .04 -.13 .13 Patient QOL Figure 6. Partially Mediated Model Direct Pathways are indicated by Indirect Pathways are indicated by
67 Table 11 Fit Indices for Measurement Pain Models Model 2 df RMSEA CFI SRMR Full Measurement Model 135.24 76 0.06 .96 .05 NonMediated Model 209.78 82 0.10 ..90 .12 Fully Mediated Model 223.33 82 0.10 .90 .12 Partially Mediated Model 159.13 78 0.07 .95 .09 Table 12 Summary of Standardized Total Effects Among Latent Variables Partially-Mediated Model Exogenous on Endogenous on Endogenous Endogenous Financial Vulnerability Relationship Partner Patient Partner Concerns Pain Quality Coping QOL QOL Relationship -.13 -.19 __ __ __ __ Quality Partner Coping -.12 -.07 __ __ __ __ Patient QOL -.17* -.58* .34* -.03 __ __ Partner QOL -.43* -.08 .44* .16* __ __ *p < .05 Gender and Role Aim 1, Question 5 asked if the patient and partner results on the POMS were related and if there was a difference for gender. Th e general linear model wa s used to conduct a
68 MANOVA as a time series analysis using the dyad as the unit of analysis (the first time was the patient and the second time was the pa rtner). The partner/patient dyad or role was the within subjects factor and partner ge nder was the between subj ects factor. This also gave the mirror image when run for pa tient gender since thes e were heterosexual couples, and there was a male/female for each partner relationship. There were no significant differences for role or gend er on the POMS, although significance was approached at p= 0.058 with female pa tients having higher means (Table 13). Table 13 Comparison of Role by Gender Means for General and Mental Health, Dyadic Adjustment and POMS Role Gender Patient Partner Differences Mental Health Male 68.82 77.12 Role* Female 66.77 70.12 Partner Gender by Role* General Health Male 52.26 69.96 Role*, no Gender Female 51.23 69.50 Dyadic Adjustment Male 122.78 118.45 Role*, no Gender Female 120.51 119.19 POMS Male 26.93 23.41 No Role, no Gender Female 32.85 26.30 p < .05 Since gender could not be eval uated by the SEM model, subsequent evaluations were done with the same MANOVA time series evaluations looking for role or gender differences for the following variables: Dyad ic Adjustment, General Health, and Mental
69 Health. As expected, there were significan t differences for the pa tients general health compared to the partners, but no differences for gender. There were differences by role for marital adjustment, with male and female partners having statisti cally less satisfaction than patients. There was an interaction for dyad by gender only for mental health p =.015 with female partners more negatively affecte d. (Table 14). There were no differences between patients by gender nor between partners by gender. All of the eight subscales fo r the SF-36 were also examined with MANOVA for gender differences, with only mental health having significant results for gender. The variables that were completed individually by one member of the dyad were also examined for gender differenc es by ANOVA: Finance, Distre ss from Pain, Intensity of Pain, Interference from Pain, Positive problem solving, Impulsiveness/carelessness and Avoidance. No significant differences for gender were found on any of the variables. Outcomes Quality of Life for Patient and Partner The distress levels as measured by the POMS for the patient and partner were not significantly different from each other. The patients scored less on the subscales of the SF-36 compared to same age norms in every ca tegory (Table 15). The partners had more bodily pain than the norm (higher scores in dicate no pain) when self-reporting on the SF36, lesser scores on physical role, vitality, social functioning, emotional role and mental health, however, it is unknown if these are stat istical differences. Pain measures in the SEM model were completed by the patient. Th ere was no direct effect for pain on the Partners Quality of Life in the SEM model that was completed by the patients, although there was a correlation of distress from the pa tients pain and the PO MS for the partner.
70 Table 14 Main Effects for Role and Gender for PO MS, Dyadic Adjustment, General and Mental Health of Patients and Partners Source SS df MS F P Within Subjects POMS Role 2190.268 1 2190.268 3.635 .058 Gender*Dyad 1678.268 1 1678.286 2.786 .097 Dyadic Adjustment Role 690.186 1 690.186 6.645 .011* Gender*Dyad 51.564 1 51.564 .496 .482 General health Role 27973.230 1 27973.230 124.478 .000* Gender*Dyad 47.863 1 47.863 .213 .645 Mental Health Role 2928.348 1 2928.348 9.968 .002* Gender*Dyad 1770.924 1 1770.924 6.028 .015* Between Subjects POMS Gender 198.581 1 198.581 .162 .687 Dyadic Adjustment Gender 195.867 1 198.867 .522 .471 General Health Gender 6.978 1 6.978 0.18 .893 Mental Health Gender 529.965 1 529.965 1.161 .283 *p <.05
71 Table 15 Comparison of SF-36 Scores for Pa tient/Partner to Same Age Norms Mean PF RP BP GH VT SF RE MH Norm 76.24 73.66 67.51 64.62 60.37 81.37 80.26 75.01 Patient 34.86 15.25 38.62 51.82 31.53 48.52 50.85 67.95 Partner 77.60 71.61 55.63 69.69 56.04 79.52 72.69 69.69 (Physical Functioning (PF), Role Physical (RP) Bodily Pain (BP), General Health (GH), Vitality (VT), Social Functioning (SF), Role Emotional (RE), and Mental Health (MH).)
72 Chapter Five Discussion The focus of this study was to exam ine cancer patients with pain and the role of partner relationships and partner coping styl e mediating the quality of life outcomes of the patient and partner. Chapter five discusses the findings, study limitations, conclusions, implications for nursing practice and education, and offers recommendations for future research. Findings Covariates The strength of the study was that it included a wide range of cancer diagnoses, and both genders were represented as caregivers and patients. Fifty-five percent of the patients in the study were stable, whether under treatment, not under treatment or in remission. Prostate cancer with the highe st number of new diagnoses per year (ACS, 2005), had only 4% of patients participate, possibly due to early intervention and cure, they did not require pain management, thus were not eligible for this study. The remaining 45% of the sample had progressive disease. This would imply a profound risk for a decrease in quality of life for half the patients and partners. This study consisted of mostly white participants, and this lim its ability to generalize to other ethic groups. The couples had sim ilar education levels and age. Kenny (2006) says this is typical compositional effects that are to be expected with married couples in that couples tend to be simila r in education level, age, a nd socioeconomic status. Aim 1,
73 question 1, examined covariates to be used in the model. Age and length of the relationship were not found to be significant covariates influencing these participants. Financial concerns were a major concern for pa rtners and also affected the patients. The significant standardized B coefficient for patients was -.13 and -.36 for partners. This measure was completed by the partner and this ma y have influenced the results. Partners probably feel more pressure financially with the uncertainty of having a loved one with cancer, additional caregiving responsibilities and role adjustments. They may even provide some protective buffering to patient s, shielding patients in assuming more financial burden. It is also possible that patients had changed their priorities, putting financial concerns lower on th e list. Even though the pati ents did not complete this measure, it did show an effect on thei r quality of life thr ough this model. Gender was also a covariate that was identified. The orig inal proposal stated the intent to run the same model separately for male partners and female partners with corresponding patients. Due to LISREL requi rements of large sample sizes, the SEM model was not suited to examine differences in gender. Gender differences were not found in the results of the PO MS, although the results approa ched (significance p = .058) with the female patients having higher distress scores. Ther e were significantly negative scores for female partners on the mental health subscale compared to male partners. This finding was consistent with the earlier covari ate analysis done prior to designing the SEM model, lending further support for the results. Mental health and the Profile of Mood States are similar measures of mood and emo tional distress. The POMS covers a wide range of moods or emotions and gives insight into distre ss: anger-hostility, depression/dejection, fatigue, v itality, confusion, and tension anxiety. The Mental health
74 subscale consists of only five items that give an overview of an introspective outlook on life: happy, peaceful, nervous, sad, and discouraged. Women partners were more discouraged and saddened than their male count erparts. Many of th e studies that have been done were in younger women with breast cancer and their spouses, and older men with prostate cancer and their spouses. Although there were in conclusive results demonstrating higher distress or impaired mental health well being, these results lend support to being more similar to other inve stigators findings of females having higher distress whether patient or partner with a more balanced study of gender and role mix in a variety of cancer diagnoses and different stag es (Baider et al., 1998; couper et al., 2007; Northouse, et al., 2000; Soloway, et al., 2004. This continues to be a variable that needs to be examined as proposed by Hagedoorn et al. (2008). Mediators Aim 1, questions 2 through 4 addre ssed the relationship quali ty as a mediator of financial concerns and vulnerability. Financial concerns did have an effect directly on both the patient and partner, but interestingly, it did not have a significant effect on the relationship quality. This pertinent negative finding is supportive in that even though this is a significant stressor for the couple, it di d not significantly affect their relationship. Financial concerns have been cited to be a major problem for marriages and a cause for divorce (Karney, Story & Bradbury, 2005). Vulnerability from pain had a signi ficant negative effect on th e patients quality of life. There was no direct effect on the partne rs quality of life. The relationship quality did mediate the patients quality of life positiv ely, despite the stress. The relationship had a direct positive effect on Quality of Life of both members of the couples and this is in
75 spite of disease progression and pain for 50% of the dyads. Because both members of the dyad were together when they were appr oached, there was open communication and cohesion even during the enrollment in the study. Several members commented that the questionnaires once completed, actually gave them stimulus for discussion. Even though partners had less satisfaction with the rela tionship, perhaps this was influenced by changes the illness precipitated in their own lives, the patient s continued to feel support. The patient and partner mean scores were slightly higher on the DAS (mean 119 to 123) compared to scores reported in previous studies of cancer couples (mean 108 to 120) (Northouse, 1998; Manne, 1999; Solomon, 2004) Spanier (1976), who developed the Dyadic Adjustment Scale, found that mean scores for divorced persons was 70.7 and married was 114.8. Examination of the individual s with low scores in this present study provided no clear patterns with respect to pain score, age, education, disease, stage of treatment, or coping styles. With all the st ress these cancer patients and partners had, for most, their relationships seemed to remain strong and supportive, a positive influence on quality of life for most couples. Partners did have more bodily pa in when self-reporting on the SF-36 compared to the norm for their age groups. The responsibilities of caregiving could have been more physically demanding than a general population would feel. They could be lifting more, carrying more, or just the am ount of increased workload could have induced muscle strain and discomfort. Or, perhaps, there was some crossover of pain through empathy that was not captured in the SEM model. Aim 2, questions 1 through 3 addr essed the mediation by th e partners coping style when a couple is challenged with a patient having a diagnosis of cancer and pain. The
76 couple was also vulnerable from the financ ial concerns. There were no significant pathways to coping from the stressors, t hus, no evidence of mediation by coping. There was no evidence of a crossover or transmittal of dyadic coping from the SEM model. The partners coping scores were generally more positive than negative and did affect their own quality of life positively. The results show there was no direct effect of the partners coping on the patients quality of life directly. Th ere were significant positive and negative correlations between coping st yles and the relationship quality. The two theories of dyadic coping address that there is a crossover of coping as a systemictransaction, or the a lternative hypothes is that individuals examine their partners strategies and then in itiate their own responses to pres erve relationship satisfaction in dealing with stressors, depending on the re lationship commitment. Transaction was not demonstrated by this model. Coping was related to the quality of the relationship, so when the partner perceived stress, perhaps they acted through the relationship to reassure the patient and have a positive effect on both members quality of life. The coping instrument was completed only by the partner and addressed items as I rather than we, so this may have influenced the outcome. However, financial c oncerns did reflect the patients QOL even though it was completed by the partner. Not having a dyadic coping measure or at least coping measures from both patient and partner may have influenced these results. Intuitively, one would think that a partners coping style would affect the patients outcomes, including pain management strategies that frequently require input from another. In retrospect, the correlati ons for the stressors and coping were minimal with only financial concerns weakly correlated. Coping probably could have been determined as a non-mediator when examining the initial correlations.
77 These partners had higher means on positive coping (mean 12.27 SD 3.74) compared to other studies (mean 13.22 SD 3.57). They had lower scores on negative coping indicated by the Negative Problem s Solving subscale (mean 8.08 SD 6.36) compared to (mean 10.14 SD 6.55), Impulsi ve-carelessness subs cales (mean 7.17 SD 5.09) compared to earlier research (mean 8.86 SD 6.85) and Avoidance subscale (mean 4.70 SD 3.77) compared to earlier re sults (6.23 SD 4.97) (Ko, 2005). It may also be possible that individuals who have more positive coping strategies have better quality relationships. One of the limitations identified earlier in the study was that these participants used more positive copi ng styles by the nature of self-selection for the study. This may also have influenced the evaluation of coping style as a mediator. If the partners had higher negative coping skills they may have had a negative effect on the relationship and indir ectly on the patient. Outcomes The distress levels as measured by the POMS for the patient and partner were not significantly different from each other, but are higher for the patient than scores reported in previous studies of co uples (Banthia, 2003; Ko, 2005). The patient scores on the POMS were (mean 29.89 SD 30.74) compared to other studies (mean 17.15 SD 33.22) while the partners scores were slightly lower (mean 25.08 SD 29.67) compared to other studies (mean 26.70 SD 37.04) (Ko, 2005). These differences suggest patients distress was affected by the symptom pain. Distress, general health, and mental health indicators exhi bited a decrease in their quality of life from pain and cancer, particularly for the pa tient, but there was a positive effect from the relationship. The POMS, and the SF-36 as measures of quality of life
78 show that cancer has impacted patients health in all the s ubscales, although the patients mental health was not as different from the norm despite great physical and role impairment. Having a partner that has cancer di d show a slight general decline in most of the SF-36 subscales, but the partners own gene ral health mean was actually higher than the norm. This may be a relative perception fo r the partners self-evaluations when their frame of reference may have been influen ced by close proximity to an ill person. Perhaps, since the marital relationships were generally strong and these were positive copers, they found meaning and satisfaction in caring for the patients, despite the greater burden. Limitations This study was a secondary anal ysis (Glass, 1976) of dya ds using baseline data from patients and their partners from a la rge intervention study supported by the National Institutes of Health (5R01NR 008270). This an alysis was limited to the existing data that had already been collected. Th e participants used self-report and some of the couples may have completed their questionnaire with th eir partners present. The use of selfreport instruments is a known limitati on bias in many research studies. Dyadic coping instruments were not collected from the dyad, only the partner and addressed the individual, not the couple as a unit. This construct was theoretically inferred as a crossover from the self-report of the partners problem-solving style, but not supported in this study, a problem identified in the original propos al, but explored in theory. The original proposal included th e use of the Physical and Mental Component Scores in the SEM model rather than the Gene ral and Mental health subscales that were
79 utilized in the study. In a f actor analysis completed by Bucholz, Krol, Rist, Nieuwkerk & Schippers (2008), there was overlap between role-physical and role-emotional on the separate component scores. Perhaps these co rrelations between fact ors contributed to poorer lambda loadings in the SEM model that attempted to use the component scores. Couples who agreed to participate in this study had distre ss and pain that they actively sought to lessen by enrolling is this study. This was evidenced by the distress measured by the POMS scores for patients. By choosing to enroll in this study there was a bias toward individuals who use positive c oping strategies to lessen distress. The study may also have been biased because by agreeing to participate in the study, the participants had fairly good marital or dya dic relationships since they were hoping to improve their partners pain. Because the patients had a diagnosis of cancer and many had progressive disease, comparison of patient s and partner s phys ical and mental or emotional health outcomes were expected to differ but were still appropriate variables to study. This was a cross-sectional study that examined only one episode in time where the processes that were examined are constantly ongoing and changing. The sample included limited diversity due to eligibility criteria a nd requirements of bei ng able to read and write English. A pain score of at least a 3 on a scal e of 0 to 10 was part of the inclusion criteria. One patient was accrued in the study who had a score of 1 and two participants had a score of 2 at the time of enrollment. Although these patients should not have been eligible according to strict criteria, they were retained. These dyads were enrolled earlier in the study when the patients were recruite d as outpatients in the Pain and Palliative Care Clinic. These three patients all wished to eliminate or decrease their pain scores to
80 lower levels, so even though they did not i ndicate higher pain scores, their lives were affected by pain, thus they were included in the study. Recommendations and Conclusions Health care providers need to be aw are of the both patients and partners when caring for patients if they are in committed relationships. Shamley and Cross (1982) performed a factor analysis of the Dyadic Adjustme nt Scale and have recommended a shortened version that consists of 6 items. Even more applicable to clinical practice, they report that using the global single it em that asks for degrees of happiness with the relationship ranked on a Likert-type scale is sufficient for quick screening (Sharpley & Cross). In this sample of strong marital relationships, the relationship quality had a positive effect on both the patients and partners QOL and mediated the stressor, cancer pain. It is possible that patients who are single or in low quality marital relationships may have even greater negative impacts on their quality of life, distress, general and mental health and their partners influence may impact them negatively. Further research is needed comparing single patients with dyads. Future studies need to continue to explore dyadic research despite the difficulties with enrollment, and large sample sizes required for data analysis. People do not live in isolation and are greatly influenced by their social envi ronments. In particular, when examining concepts that are paramount to cancer patients, such as love, uncertainty, meaning of life, body image, and changing role status, the influence of intimate partners and close friends has great potential to influe nce quality of life outcomes. There should be relationship studies that include a greater proportion of the different racial and cultural concerns that were lacking in this study.
81 Some constructs for this study were completed by the patient, some by the partner and some by both patient and partner. SEM wa s still the appropriate choice of analysis (D.A. Kenney, personal communication, Octobe r 3, 2008). Despite the limitation of a coping measure by only one member of the dya d, the findings of positive influences of a good quality, committed, intimate relationship an d the quality of life comparisons for patients and partners was a worthwhile st udy. These outcomes comparing the patient, partner and same age norms was also informativ e, though not a statistical analysis used in this study. For future research, dyadic coping and dyadic relationships should be explored in quality of life outcome research. With the advent of personalized health care, that includes genetic testing and targeted therapies, there are even more implications for potential quality of life issues and decisions confronting committed partners and their children.
82 References Acitelli, L. & Badr, H. (2005). My illness or our illness? Attending to the relationship when one partner is ill. In Couples coping with stress: Emerging perspectives on dyadic coping (pp. 121-136.). Washington, D. C.: American Psychological Association. American Cancer Society (2008). Cancer statistics. Retrieved June 13, 2008 from: http://www.cancer.org/docroot/PRO/c ontent/PRO_1_1_Cancer_Statistics_2008 American Cancer Society (2005). Cancer f acts and figures. Retrieved October 5, 2008 from: http://www.cancer.org/downloads/STT/CAFF2005f4PWSecured.pdf Badr, H. & Taylor, C. (2006). Social cons traints and spousal co mmunication in lung cancer. Psycho-Oncology, 15 673-683. Baider, L., Koch, U., Esacson, R. & De-Nour A. (1998). Prospective study of cancer patients and their spouses: The weakness of marital strength. Psycho-Oncology, 7, 49-56. Baider, L., Walach, N., Perry, S., & De-Nour A. (1998). Cancer in married couples: Higher or lower distress. Journal of Psychosomatic Research, 45 (3), 239-248. Baider, L. & Bengel, J. (2001). Cancer and the spouse: Gender-rel ated differences in dealing with health care and illness. Critical Reviews in Oncology/Hematology, 40(2), 115-123.
83 Banthia, R., Malcarne, V., Varni, J., Ko, C., Sadler, G. & Greenbergs, H. (2003). The effects of dyadic strength and coping st yles on psychological distress in couples faced with prostate cancer. Journal of Behavioral Medicine, 26 (1), 31-50. Baron, R. & Kenny, D. (1986). The moderator-m ediator variable distinction in social psychological research: Con ceptual, strategic, and stat istical considerations. Journal of Personality and Social Psychology, 51 1173-1182. Ben-Zur, H., Gilbar, O & Lev, S. (2001). Coping with breast cancer. Psychosomatic Medicine, 63 32-29. Bishhop, M., Beaumont, J., Hahn, E., Cella, D., Andrykowski, M., Brady, M. et al. (2007). Late effects of cancer and hema topoietic stem-cell transplantation on spouses or partners compared with su rvivors and survivor-matched controls. Journal of Clinical Oncology, 25 (11), 1403-1411. Bodenmann, G. (2005). Dyadic coping and its significance for marital functioning. In T. Revenson, Kayser, K. & Bodenmann, G. (Ed.), Couples coping with stress: Emerging perspectives on dyadic coping (pp. 33-49.). Washington, D. C.: American Psychological Association. Bodenmann, G., Pihet, S. & Kayser, K. ( 2006). The relationship between dyadic coping and marital quality: A 2year longitudinal study. Journal of Family Psychology, 20(3), 485-493. Buchholz, A., Krol, A., Rist, F., Nieuwkerk, P. & Schippers, Gerard. (2008). An assessment of factorial structure and hea lth-related quality of life in problem drug users using the short form 36 health survery. Quality of Life Research, 17, 10211029.
84 Bultz, B., Speca, M., Brasher, P., Geggie, P. & Page, S. (2000). A randomized controlled trial of a brief psychoeducational support gr oup for partners of early stage breast cancer patients. Psycho-Oncology, 9 303-313. Burman, B. & Margolin, G. (1992). Analys is of the association between marital relationships and health problems: An interactional perspective. Psychological Bulletin, 112 (1), 39-63. Carlson, L., Ottenbreit, N., St. Pierre, M. & Bultz, B. (2001). Partner understanding of the breast and prostate cancer experience. Cancer Nursing, 24(3), 231-239. Cleeland, C., Mendoza, T., Wang, X., Chou, C., Harle, M., Morrissey, M., & Engstrom, M. (2000). Asse ssing symptom distress in cancer patients: the M. D. Anderson Symptom Inventory. C ancer, 89( 7), 1634 1646. Cook, W. & Kenny, D. (2005). The actor-partn er interdependence model: A model of bidirectional effects in developmental studies. International Journal of Behavioral Development, 29 101-109. Couper, J., Bloch, S., Love, A., Macvean, M., Guchnesne, G. & Kissane, D. (2006). Psychosocial adjustment of female partne rs of men with prostate cancer: A review of the literature. Psycho-Oncology, 15 937-953. Cutrona, C., Russell, D., & Gardner, L. ( 2005). The relationship enhancement model of social support. In Couples coping with stress: Emerging perspectives on dyadic coping. (pp. 73-96.). Washington, D.C.: Amer ican Psychological Association. Davison, B., Goldenberg, S., Gleave, M. & Degner, L. (2003). Provision of individualized information to men and their partners to facilitate treatment decision making in prostate cancer. Oncology Nursing Forum, 30 (1), 2003.
85 Diamantopoulos, A. & Siguaw, J. (2000). Introducing LISREL London: Sage Publications. D'Zurilla, T., Chang, E., Nottingham, E. & Faccini, L. (1998). Social problem-solving deficits and hopelessness, de pression, and suicidal risk in college students and psychiatric inpatients. Journal of Clinical Psychology, 54 (8), 1091-1107. Ezer, H., Ricard, N., Bouchard, L., Souhami, L., Saad, F., Aprikian, A., et al. (2006). Adaptation of wives of pros tate cancer following diagnosis and 3 months after treatment: A test of the family adaptation theory. International Journal of Nursing Studies, 43(7), 827-838. Feetham, S., Thomson, E. & Himshaw, A. (2005). Nursing leadership in genomics for health and society. journal of Nursing Scholarship, 37 (2), 102-110. Gale, L., Bennett, P., Tallon, D., Brooks, E., Munnoch, K. & Schreiber-Kounine, C. et al. (2001). Quality of partner relationship and emotional responses to a health threat. Psychology Health and Medicine, 6 (4), 373-386. Gapstur, R. (2007). Symptom burden: A con cept analysis and implications for oncology nurses. Oncology Nursing Forum, 34 (3), 673-680. Galbraith, M., Pedro, L., Jaffe, A. & Alle n, T. (2008). Describing health-related outcomes for couples experien cing prostate cancer: Differences and similarities. Oncology Nursing Forum, 35 (5), 794-801. Gilbar, O. & Zusman, A. (2007). The correlation between coping strategies, doctorpatient/spouse relationships and psyc hological distress among women cancer patients and their spouses. Pshycho-Oncology, 16 1010-1018.
86 Glass, G. (1976). Primary, secondary, and meta-analysis of research. Educational researcher, 5 (10) 3-8. Gordon, D. (2005). American Pain Society r ecommendations for improving the quality of acute and cancer pain management: Am erican Society Quality of Care Task Force. Archives of Internal Medicine, 165 ( 14), 1574-1580. Hagedoorn, M., Buunk, B., Kuijer, R., Wobbes, T., & Sanderman, R. (2000). Couples dealing with cancer: Role and gender di fferences regarding psychological distress and quality of life. Pshycho-Oncology, 9 232-242. Hagedoorn, M., Kuijer, R., Buunk, B., DeN ong, M., Wobbes, T. & Sanderman, R. (2000). Marital satisfaction in patients w ith cancer: Does support from intimate partners benefit those who need it most? Health Psychology, 19 (3), 274-282. Hagedoorn, M., Sanderman, R., Bolks, H., Tuinis tra, J. & Coyne, J. (2008). Disress in couples coping with cancer: A meta-analy sis and critical review of role and gender effects. Psychological Bulletin, 134(1), 1-30. Halford, W., Scott, J. & Smythe, J. (2000). Couples and coping with cancer: Helping each other through the night. In K. S. Schmaling, T. (Ed.), The psychology of couples and illness: Theory, research and practice. (pp. 135-170.). Washington, D.C.: American Psychological Association. Harden, J., Schafenacker, A., Northouse, L ., Mood, D., Smith, D., Pienta, K. et al. (2002). Couples' experience with prostate cancer: Focus Group Research. Oncology Nursing Forum, 29 701-709.
87 Hawes, S., Malcaren, V., Ko, C., Sadler, G., Banthia, R., Sherman et al. (2006). Identifying problems faced by spouses and partners of patients with prostate cancer. Oncology Nursing Forum, 33 (4), 807-814. Hilton, B. (1994). Family communication patte rns in coping with early breast cancer. Western Journal of Nursing Research, 16(4), 366-391. Hinnen, C., Hagedoorn, M., Sanderman, R. & Ra nchor, A. (2007). The role of distress, neuroticism and time since diagnosis in e xplaining support behaviors in partners of women with breast cancer: Re sults of a longitudinal analysis. PsychoOncology, 16, 913-919. Karney, B., Story, L. & Bradbury, T. (2005). Marriages in Context: Interactions between chronic and acute stress among newlywed s. In T. Revenson, Kayser, K. & Bodenmann, G. (Ed.), Couples coping with stress : Emerging perspectives on dyadic coping (pp. 33-49.). Washington, D. C.: American Psychological Association. Kelloway, K. (1998). Using lisrel for structural equati on modeling: A researcher's guide. Thousands Oaks, California: Sage Publications. Kenny, D. & Cook, W. (1999). Partner effects in relationship research: conceptual issues, analytic difficulties, and illustrations. Personal Relationships, 6 433-448. Kim, Y. & Carver, C. (2007). Frequency a nd difficulty in caregiving among spouses of individuals with cancer: Effects of adult attachment and gender. PshychoOncology, 16, 714-723.
88 Knoll, N., Burkert, S., Rosemeier, H., Roig as, J. & Gralla, O. (2007). Predictors of spouses' provided support for patients receiving laparoscopic radical prostatectomy peri-surgery. Psycho-Oncology, 16 312-319. Ko, C., Malcarne, V., Varni, J., Roesch, S., Banthia, R., Greenbergs, H. et al., (2005). Problem-solving and distress in prosta te cancer patients and their spousal caregivers. Support Cancer Care., 13 367-374. Kornblith, A., Herr, H., Ofman, U., Scher, H. & Holland, J. (1994). Quality of life of patients with prostate cancer and their spouses. The value of a data base in clinical care. Cancer, 73(11), 2791-2802. Kornblith, A., Regan, M., Youngmee, K., Greer, G., Parker, B., Bennett, S. et al., (2006). Cancer-related communication between female patients and male partners scale: A pilot study. Psycho-Oncology, 15 780-794. Lazarus, R. & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Lutsky, S. & Knight., B. (1994). Expl aining gender differences in caregiver disstress: The roles of emotional attentiveness and coping styles. Psychology and Aging, 9(4), 513-519. McClement, S., Woodgate, R., Degner, L. (1997). Symptom distress in adult patients with cancer. Cancer Nursing, 20 (4), 236-243. McHorney, C., Ware, J. & Raczek, A. (1993). The MOS 36-item short-form health survey (SF-36): II. Psychometric and clinical te sts of validity in measuring physical and mental health constructs. Medical Care, 31 (3), 247-263.
89 McMillan, S., Tofthagen, C., & Morgan, M. (2008). Relationships among pain, sleep disturbances and depressive symptoms in outpatients from a comprehensive cancer center. Oncology Nursing Forum, July, 35(4) 603 611. McMillan, S. & Small, B. ( 2002). Symptom distress and quality of life in patients with cancer and newly admitted to hospice home care. Oncology Nursing Forum, 29 (10), 1421-1426. McNair, D., Lorr, M. & Droppleman, L. (1992). Profile of mood states manual. San Diego, California: EdITS. Maliski, S., Heilemann, M. & McCorkle, R. (2002). From "death sentence" to "good cancer": Couples' transformation of a prostate cancer diagnosis. Nursing Research, 51(6), 391-397. Manne, S., Alfieri, T., Taylor, K. & Doughert y, J. (1999). Spousal negative responses to cancer patients: The role of social restriction, s pouse mood, and relationship satisfaction. Journal of Consulting and Clinical Psychology, 67 (3), 352-361. Manne, S., Taylor, K., Dougherty, J. & Ke meny, N. (1997). Supportive and negative responses in the partner relationship: Their association with psychological adjustment among individuals with cancer. Journal of Behavioral Medicine, 20(2), 101-125. Manne, S., Ostroff, J., Sherman, M., Heyman, R., Ross, S. & Fox, K. (2004). Couples' support-related communication, psychol ogical distress, and relationship satisfaction among women with early stage breast cancer. Journal of Consulting and Clinical Psychology, 72 (4), 660-670.
90 Manne, S., Ostroff, J., Winkel, G., Gold stein, L., Fox, K. & Grana, G. (2004). Posttraumatic growth after breast cancer: Patient, partner, and couple perspectives. Psychosomatic Medicine, 66 442-442-454. Manne, S., Ostroff, J., Winkel, G., Fox, K., Grana, G., Miller, E. et al. (2005). Couplefocused group intervention for wome n with early st age breast cancer. Journal of Consulting and Clinical Psychology, 73 (4), 634-646. Manne, S., Ostroff, J., Norton, T., Fox, K., Goldstein, L. & Grana, G. (2006). Cancerrelated relationship communication in c ouples coping with early stage breast cancer. Psycho-Oncology, 15 234-247. Manne, S. & Schnoll, R.. (2001). Measuri ng supportive and unsupportive responses during cancer treatment: A factor analytic assessment of the pa rtner responses to cancer inventory. Journal of Behavioral Medicine, 24 (4), 297-321. Mireskandari, S. Meiser, B., Sherman, K., Warn er, B., Andrews, L. & Tucker, T. (2006). Evaluation of the needs and concerns of partners of women at high risk of developing breast/ovarian cancer. Psycho-Oncology, 15(2) 96-108. Modonesi, C., Scarpi, D., Maltone M., Derni, S., Fabbri, L., Martini, F., Sansoni, E., & Amadori, D. (2005). Impact of palliative care unit admission on symptom control evaluated by the edmonton symptom assessment system. Journal of Pain and Symptom Management, 30 (4), 367-373. Mor, V., Allen, S. & Malin, M. (1994). The psychosocial impact of cancer on older versus younger patients and their families. Cancer, 74(7), 2118-2127. Morse, S. & Fife, B. (1998). Coping with a part ner's cancer: Adjustment at four stages of the illness trajectory. Oncology Nursing Forum, 25 (4), 751-759.
91 NIH. (2002). NIH Stateofthescience statem ent on symptom management in cancer: Pain depression, and fatigue. Paper presented at the National Institutes of Health (NIH) State-of-the-Scien ce Conference on Symptom Management in Cancer: Pain, Depression, and Fatigue. Northouse, L. & Peter-Golden. H. (1993). Cancer and the family: Strategies to assist spouses. Seminars in Oncology Nursing, 9 (2), 74-82. Northouse, L., Templin, T., Mood, D. & Oberst M. (1998). Couples' adjustment to breast cancer and bening breast dis ease: A longitudinal study. Psycho-Oncology, 7 3748. Northouse, L., Mood, D., Templin, T., Mellon, S. & George, T. (2000). Couples' patterns of adjustment to colon cancer. Social Science & Medicine, 50 271-284. Northouse, L., Templin, T. & Mood, D. (2001) Couples' adjustment to breast disease during the first year following diagnosis. Journal of Behavioral Medicine, 24 (2), 115-136. Pasacreta, J., Barg, F., Nuamah,I. & McCork le, R. (2000). Particip ant characteristics before and 4 months after attendance at a family caregiver cancer education program. Cancer Nursing, 23(4), 295-303. Persson, E., Severinsson, E. & Hellstrom, A. (2004). Spouses' perceptions of and reactions to living with a partner who has undergone surgery for rectal cancer resulting in a stoma. Cancer Nursing, 27(1), 85-90.
92 Pistrang, N. & Barker., C. (2005) How partners talk in times of stress: A process analysis approach. In T. Revenson, Kayser, K. & Bodenmann, G. (Ed.), Couples coping with stress: Emergi ng perspectives on dyadic coping (pp. 97-119.). Washington, D.C.: American Psychological Association. Portenoy, R., Thaler, H., Kornblith, A., Lepore, J, Friedlander-Klar, Coyle, T., Smart Curley, T., Kemeny, N., Norton, L., Hoskins, W ., and Scher. (1994). Symptom prevalen ce, characteristics and distress in a cancer population. Quality of Life Research, 3 183-189. Portenoy, R., Thaler, H., Kornblith, A., Lepor e, J, Friedlander-Klar, H., Kiyasu, F. (1994). The Memorial Symptom Assessment Scale: An Instrument for the Evaluation of Symptom Prevalence, Ch aracteristics and Distress. European Journal of Cancer, 30A 1326-1336. Porter, L., Keefe, F., Hurwitz, H. & Faber, M. (2005). Disclosure between patients with gastrointestinal cancer and their spouses. Psycho-Oncology, 14 1030-1042. Rand Corporation (2008). Medical outcomes study: 36 item short-form survey. Retrieved September 19, 2008 from http://www.rand.org/he alth/surveys_too ls/mos/mos_core_36item.html Ratner, P. Bottorff, J. & Johnson, J. (2008) Structural equation modeling. In Advanced Design in Nursing Research (pp 178 208). Thousand Oaks: Sage Publications. Ren, X. (1997). Marital status and quality of relationships: The impact of health perception. Social Science Medicine, 44(2), 241-249.
93 Revenson, T., Abraido-Lanza, A., Majerovitz S. & Jordan, C. (2005). Couples coping with chronic illness: What's gender got to do with it? In Couples coping with stress: Emerging perspectives on dyadic coping. (pp. 137-156.). Washington, D. C.: American Psychological Association. Revenson, T., Kayser, K. & Bodenmann, G. (2005). Couples coping with stress: Emerging perspectives on dyadic coping Washington, D. C.: American Psychological Association. Rhodes, V., & Watson, P. (1987). Symptom Dist ress: The Concept Past and Present. Seminars in Oncology Nursing, 3 242-247. Rhodes, V., McDaniel, R. & Matthews, C. (1998). Hospice patie nts and nurses perceptions of self-care defic its based on symptom experience. Cancer Nursing, 21(5), 312-19. Romero, C., Lindsay, J., Dalton, W., Nels on, D. & Friedman, L. (2007). Husbands' perceptions of wives' adjustment to br east cancer: The impact on wives' mood. Pshycho-Oncology. Schmaling, K. & Sher, T. (2000). The psychology of couples and illness: Theory, research and practice. Washington, D.C.: American Psychological Association. Schnoll, R., Harlow, L., Stolbach, L. & Bra ndt, U. (1998). A structural model of the relationships among stage of disease, ag e, coping, and psychological adjustment in women with breast cancer. Pshycho-Oncology, 7 69-77. Serlin, RC, Mendoza, TR, Nakamura, Y, Ed wards, KF, Cleeland, CS. (1995). When is cancer pain mild, moderate or severe? Gr ading pain severity by its interference with function. Pain, 61, 277-284.
94 Shands, M., Lewis, F., Sinsheimer, J. & Coch rane, B. (2006). Core concerns of couples living with early stage breast cancer. Pshycho-Oncology, 15, 1055-1064. Shin, H., Kim, K., Kim, Y., Chee, W., & Im, E. (2007). A comparison of two pain measures for asian american cancer patients. Western Journal of Nursing Research 1-15. Smith, W., Gracely, R., & Safer, M. (1998). The meaning of pain: cancer patients' rating and recall of pain intensity and affect. Pain, 78(2), 123-129. Soloway, C., Soloway, M., Kim, S. & Kava, B. (2005). Sexual, psycho logical and dyadic qualities of the prostate cancer 'couple'. British Journal International, 95 780785. Spanier, G. (1976). Measuring dyadic adjustme nt: New scales for assessing the quality of marriage and similar dyads. Journal of Marriage and Family, 38(1), 15-28. Stetz, K. (1987). Caregiving demands during advanced cancer: Spouse's needs. Cancer Nursing, 10, 260-268. Stromgren, A. S., Sjogren, P., Goldschmidt, D., Petersen, M. A., Pedersen, L., & Groenvold, M. (2006). Symptom priority and course of symptomatology in specialized palliative care. Journal of Pain and Symptom Management, 31 (3), 199-206. Tabachnick, B. & Fidell, L. (2007). Using multivariate statistics, fifth edition Boston, MA: Pearson. Thornton, A. & Perez, M. (2004). Posttraumatic growth in prostate cancer survivors and their partners. Psycho-Oncology, 15, 285-296.
95 Tishelman, C., Taube, A., & Sachs, L. (1991). Self-reported symptom distress in cancer patients: Reflections of di sease, illness or sickness? Social Science Medicine, 33(11), 1229-1240. Tranin, A., Masny, A. & Jenkins, J. (2003). Genetics in oncology practice Pittsburgh, PA: Oncology Nursing Society. Tuinstra, J., Hagedoorn, M., Sonderen, E., Ranchor, A., Van den Bos, G. Nijboer, C. et al. (2004). Psychological distress in couples dealing with colorectal cancer: Gender and role differences and intracouple correspondence. British Journal of Health Psychology, 9 Vallerand, A., Templin, T., Hasenu, S., & Riley-Doucet, C. (2007). Factors that affect functional status in patients with cancer-related pain. 1-17. Vallerand, A. (1997). Measurem ent issues in the comprehe nsive assessment of cancer pain. Seminars in Oncology Nursing, 13 (1), 16-24. Vinokur, A., Threatt, B., Caplan, R. & Zimm erman, B. (1989). Physical and psychosocial functioning and adjustment to breast cancer. Cancer, 63, 394-405. Wagner, C., Bigatti, S., & Storniolo, A. (2006). Quality of life of husbands of women with breast cancer. Psycho-Oncology, 15 109-120. Walsh, D., Rybicki, L. (2006). Symptom clustering in advanced cancer. Support Care Cancer, 14, 831-836. Ware, J., Kosinski, M. & Keller, S. (1993). SF-36 Physical and mental health summary scales: A user's manual. Boston, Massachusetts: The Health Institute.
96 Wells, N., Murphy, B., Wujcik, D., & Johns on, R. (2003). Pain-related distress and interference with daily life of ambul atory patients with cancer with pain. Oncology Nursing Forum, 30 (6), 977-986. Westman, M., Keinan, G., Roziner, I. & Be njamini, Y. (2008). The crossover of perceived health between spouses. J ournal of Occupational Health Psychology, 13(2), 168-180.
About the Author Mary Ann Morgan received a Bach elor of Science of Nursing Degree from the University of Maryland, Baltimore, Maryland in 1974, a Masters of Nursing from Louisiana State University Medical Center, New Orleans, Louisiana in 1988, and a PostMasters Family Nurse Practitioner Certificate from the University of South Florida, Tampa, Florida in 2002. She has worked as a staff nurse, clinical manager, research assistant on an RO1, NIH funded study and is currently a nurse practitioner at Moffitt Cancer Center. She is an active member of the Oncology Nursing Society (ONS) and is currently the treasurer for the local chapter. She has had abstracts presented at the Southern Nursing Research Society, 4 th Biennial Cancer Survivorship Conference, and the World Congress of International Union Against Cancer. She has manuscripts that have been accepted for publication in the Oncology Nursing Forum and Clinical Journal of Nursing Oncology