The Florida State University

DigiNole Commons
Electronic Theses, Treatises and Dissertations The Graduate School

4-8-2012

Examining The Effects Of Family Relationships On Mental And Physical Health: Testing The Biobehavioral Family Model With An Adult Primary Care Sample
Sarah Beth Woods
The Florida State University

Follow this and additional works at: http://diginole.lib.fsu.edu/etd Recommended Citation

Woods, Sarah Beth, "Examining The Effects Of Family Relationships On Mental And Physical Health: Testing The Biobehavioral Family Model With An Adult Primary Care Sample" (2012). Electronic Theses, Treatises and Dissertations. Paper 5277.

This Dissertation - Open Access is brought to you for free and open access by the The Graduate School at DigiNole Commons. It has been accepted for inclusion in Electronic Theses, Treatises and Dissertations by an authorized administrator of DigiNole Commons. For more information, please contact lib-ir@fsu.edu.

THE FLORIDA STATE UNIVERSITY COLLEGE OF HUMAN SCIENCES

EXAMINING THE EFFECTS OF FAMILY RELATIONSHIPS ON MENTAL AND PHYSICAL HEALTH: TESTING THE BIOBEHAVIORAL FAMILY MODEL WITH AN ADULT PRIMARY CARE SAMPLE

By SARAH B. WOODS

A dissertation submitted to the Department of Family and Child Sciences in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Summer Semester, 2012

Sarah Woods defended this dissertation on March 28, 2012. The members of the supervisory committee were:

Wayne Denton Professor Directing Dissertation

Robert Glueckauf University Representative

Lenore McWey Committee Member

Ann Mullis Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the dissertation has been approved in accordance with university requirements.

ii

ACKNOWLEDGEMENTS An enormous thank you to my dissertation committee, without whom I may not have had faith that this project was worthwhile or accomplishable. Thank you, especially, to Dr. Wayne Denton. I am incredibly grateful for your gentle reassurances that this dissertation was only one small test of my abilities. Your confidence in my aspirations has been encouraging and sustaining; I am very grateful for your enthusiasm in my ideas and your recognition of my passion for medical family therapy. Thank you also to Dr. Lenore McWey, whose endless support, guidance, and mentoring has had immeasurable influence on my personal and professional lives. Thank you to Dr. Ann Mullis for your consistent reminders of the (important) life that exists after graduate school. Lastly, I am very grateful for the perspectives and encouragement of Dr. Robert Glueckauf, who was always willing to talk through my ideas and build them into more meaningful and significant contributions to the field. I am also forever grateful for the love and support of my family throughout this doctoral process; through my frustrations and successes, my family has been unendingly giving. My parents deserve enormous credit for my ability to start, and finish, this degree. They consistently listened and offered thoughts that gave me perspective, strength, and resilience. Most importantly, I am thankful for my husband, Jesse, who constantly believed in my ability to succeed, and made my dreams his dreams. He gave me courage and humor when I was inconsolable, and praise when I was proud. Without my family, this process would not have been meaningful and I am ever appreciative of them and their love.

iii

TABLE OF CONTENTS List of Tables..vi List of Figuresvii List of Abbreviationsviii Abstract...ix 1. INTRODUCTION...1 1.1 The Biobehavioral Family Model..1 1.2 Statement of the Problem...3 1.3 Study Purpose....4 1.4 Hypotheses.....5 2. LITERATURE REVIEW7 2.1 Families and Health.......7 2.2 The BBFM: An Organizing Framework....8 2.3 Summary..15 3. METHOD..16 3.1 Sample..16 3.2 Measures..19 3.3 Analyses...23 4. RESULTS..25 4.1 Exploratory Data Analysis...25 4.2 Construct Associations26 4.3 Stepwise Regressions...29 4.4 Model Testing..30 4.5 Bootstrapping Analyses...34 4.6 Conclusion...36 5. DISCUSSION37 5.1 Summary of Hypotheses..37 5.2 Model 1: Family Functioning..37 5.3 Model 2: Romantic Relationship Satisfaction.38 5.4 Limitations and Future Research.39 5.5 Implications for Clinical Practice44

iv

5.6 Conclusion...46 6. APPENDIX A: IRB Approval...54 7. APPENDIX B: Informed Consent Letter..55 8. APPENDIX C: Measures/Assessment...59 7. REFERENCES..70 8. BIOGRAPHICAL SKETCH.83

LIST OF TABLES 1 FEC Variables, Biobehavioral Reactivity Variables, and Disease Activity Variables: Descriptive Statistics (N = 125).26 Bivariate Pearson Correlations of Study Variables...27

vi

LIST OF FIGURES 1 2 3 The Biobehavioral Family Model (Wood, 1993)6 Participation flow diagram.17 The initial model built in AMOS to test Model 1, with family functioning (GFS/FAD scores) as the endogenous variable, depression (QIDS) as the mediator, and disease activity as a latent outcome variable (n = 125) *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.48 Model 1c (n = 125), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold49 Model 1d, with biobehavioral reactivity as a mediator composed of anxiety (OASIS Scale Score) and depression (QIDS Scale Score) (n = 125), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold...50 The initial model built in AMOS to test Model 2, with romantic relationship satisfaction (QMI scores) as the endogenous variable, depression (QIDS) as the mediator, and disease activity as a latent outcome variable (n = 76) *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.51 Model 2b (n = 76), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold52 Model 2c, with biobehavioral reactivity as a mediator composed of anxiety (OASIS Scale Score) and depression (QIDS Scale Score) (n = 76), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold...53

vii

LIST OF ABBREVIATIONS 1. CDC - Centers for Disease Control and Prevention 2. BBFM - Biobehavioral Family Model 3. FEC Family Emotional Climate 4. NHLBI - National Heart, Lung, and Blood Institute 5. ACTH - Adrenocorticotropic Hormone 6. GFS/FAD General Functioning Subscale of the Family Assessment Device 7. QMI Quality of Marriage Index 8. DAS Dyadic Adjustment Scale 9. PC Perceived Criticism 10. NIAAA - National Institute on Alcohol Abuse and Alcoholism 11. AUDIT Alcohol Use Disorders Identification Test 12. QIDS - Quick Inventory of Depressive Symptomatology Self Report 13. OASIS - Overall Anxiety Severity and Impairment Scale 14. BMI Body Mass Index 15. RMSEA - Root Mean Square Error of Approximation 16. CFI Comparative Fit Index 17. CI Confidence Interval 18. MFT Marriage and Family Therapist

viii

ABSTRACT Family and romantic relationships have been linked to both mental and physical health outcomes. Previous research has lacked attention on precise pathways by which these associations occur and continue to use predominately White, middle-class, nuclear families as the basis of study. The Biobehavioral Family Model (BBFM) is a biopsychosocial approach to health that integrates family emotional climate, biobehavioral reactivity (emotion dysregulation), and physical health outcomes into a comprehensive model. The present study was conducted to examine the ability of the BBFM to explain connections between family processes and health for primarily uninsured, low-income adult primary care patients. Patient participants (ages 18-65 years) self-reported their family functioning, romantic relationship satisfaction, anxiety, depression, alcohol use, illness symptoms, and physical well-being (n = 125). Data were also collected from patient medical charts. Separate models using family functioning (Model 1) and romantic relationship satisfaction (Model 2) as measures of family emotional climate were tested using path analyses and bootstrapping. Results demonstrated support for the BBFM in explaining health quality for this sample. Applying the BBFM to diverse primary care patients demonstrates pathways by which family processes affect the mental and physical health of these individuals. Recommendations for future research and clinical implications are discussed.

ix

CHAPTER 1 INTRODUCTION The World Health Organization (2006) defines health as a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity, asserting also that enjoyment of the highest attainable standard of health is one of the fundamental rights of every human being (p. 1). In the U.S., the Centers for Disease Control and Prevention (CDC) explains that well-being integrates mental health (mind) and physical health (body) resulting in more holistic approaches to disease prevention and health promotion (2011b, p. 1). Unfortunately, chronic illness and risk factors for poor physical health outcomes continue to be a concern in the United States. In 2008, almost 1 in 2 adults age 18 or older had 1 or more of the following chronic illnesses: diabetes, asthma, cancer, arthritis, cardiovascular disease, or chronic obstructive pulmonary disease (U.S. Department of Health and Human Services, 2011) and, more recently, 1 in 10 adults in the U.S. self-assessed their health to be fair or poor (National Center for Health Statistics, 2011). This is additionally concerning given the growth in health care costs (Agency for Healthcare Research and Quality, 2002); between 1998 and 2008, total personal health care expenditures almost doubled, growing to approximately $2 trillion, although almost 18% of the U.S. adult population under 65 years continues to be uninsured (National Center for Health Statistics, 2011). The mental and physical health of American adults are not unrelated to experiences of family and romantic relationships. Close relationships can both buffer and potentiate risk factors related to health, and there is now much evidence linking family and romantic relationships to health outcomes as researchers increasingly recognize the importance of understanding how social factors influence health (Carr & Springer, 2010). In addition, there is an increased focus in research tying relational variables to health outcomes on pathways by which these effects occur. Nevertheless, tests of the connections between social functioning and physical health continue to use White, middle-class, nuclear families as the basis of study, with a lack of attention on precise pathways and specific indicators of health (Carr & Springer, 2010; Wood, 2005). The Biobehavioral Family Model The Biobehavioral Family Model (BBFM) is a biopsychosocial approach (Engel, 1977) to health that has successfully integrated family functioning, psychological factors, and physical

health outcomes into one, comprehensive model (Wood, 1993). The BBFM connects principles of general systems theory (von Bertalanffy, 1969) with Minuchins (1974) structural family therapy model to explain the influence of psychosocial factors on biological processes and disease activity (Wood & Miller, 2002). The model theorizes the mutual influence of social, emotional, and physical factors on aspects of illness and purports that family/social relationships are a critical domain of functioning that can serve as protective or act to worsen health outcomes (Wood & Miller, 2002). Additionally, the BBFM addresses the limitation in the literature of a lack of organizing models that focus on factors relevant to explaining familypsychobiological pathways important in understanding disease (Wood et al., 2006, p. 1495). The BBFM theoretically applies to a broad developmental range of individuals and is able to address the processes affecting [the health of] any family member (adult or child) (Wood, Klebba, & Miller, 2000, p. 322). The authors justify this by individuals involvement with family regardless of age and stage of development, especially given the likely dependence on family members if ill (Wood et al., 2000; Wood & Miller, 2005). However, to date, studies of the BBFM have examined psychosocial processes affecting illness in children, specifically for children with asthma (Wood et al., 2000; Wood et al., 2006; Wood et al., 2007; Wood et al., 2008). Although this is logical given the stress-related nature of pediatric asthma (Miller & Wood, 2003), the BBFM is applicable to both psychologically manifested illnesses and diseases of primarily organic origin, as well as the illness experiences of adult family members (Wood & Miller, 2002). Lastly, because of its foundation in general systems theory, the model maintains neutrality with an unbiased framework of how families adapt; therefore, it is well-suited to the appreciation of cultural, racial, class, and gender factors as they relate to adaptive and maladaptive family process (Wood & Miller, 2002, p. 59-60). Constructs. The BBFM theorizes that family emotional climate (FEC) and biobehavioral reactivity (emotion regulation or dysregulation) are processes that influence one another and are either protective of or worsen physiological processes, including disease activity/severity and chronic illness (Wood, 1993; Wood et al., 2008). FEC describes the intensity and positivity or negativity of emotional processes within the family (Wood et al., 2008), while biobehavioral reactivity is the way in which an individual family member responds to emotional stimuli and is their degree of emotion regulation or dysregulation (Wood et al., 2008). Biobehavioral reactivity is typically measured as anxiety or depression (e.g., Wood et al., 2007; Wood et al., 2008).

The BBFM theorizes that the effects of the FEC on physical health are through emotional influences on individual family members; therefore, it predicts (a) a direct relationship between FEC and biobehavioral reactivity, (b) a direct relationship between biobehavioral reactivity and disease activity, and (c) a nonsignificant pathway between FEC and disease activity (Figure 1). These hypotheses have been confirmed in tests of the model with child asthma outcomes (Wood et al., 2008). A further description of these constructs and pathways is provided in Chapter 2. Statement of the Problem Although evidence repetitively demonstrates associations between familial and romantic relationships and either mental or physical health outcomes, the body of research supporting these relationships fails to simultaneously integrate effects of relational functioning on both mental and physical health and lacks a focus on precise pathways (Carr & Springer, 2010). Although researchers point to potential mediators or moderators of the relationship between close relationships and health, there is a lack of consensus on which processes need investigation and consistent recommendations that future research explore these associations (Proulx & Snyder, 2009). Only in establishing pathways, mechanisms, and directions of effect by which biopsychosocial variables influence physical well-being in a comprehensive model can we fully understand how to intervene in these factors and prevent negative outcomes. It is critical to integrate social, psychological, and biological indicators of disease (National Research Council, 2001). In addition, although research on relationships and heath has increased over the last decade, most studies of adult health focus on the protective effects of marriage and not on family relationships more broadly or non-traditional romantic relationships (Carr & Springer, 2010, p. 748). Neglecting to study the effects of relationships with family members on individual health outcomes ignores one of the most powerful influences on physical health and well-being; not all social relationships are created equal (Weihs, Fisher, & Baird, 2002). Although we increasingly understand how stress affects health, there is less literature examining the effects of family stress (Wood & Miller, 2005) and there has only recently been a growth in working to understand the impact of stressful close relationships and family relationship quality on physical health, especially for adults (Proulx & Snyder, 2009; Wickrama et al., 2001). Another concern with research to date is its reliance on models that privilege the White middle-class nuclear heterosexual family as the norm (Carr & Springer, 2010, p. 743). Given

mixed findings regarding the effects of race (Kaplan & Kronick, 2006; Springer & Mouzon, 2011), gender (Gardner & Oswald, 2004; Johnson, Backlund, Sorlie, & Loveless, 2000), and family structure and stability (Sbarra & Nietert, 2009; Wienke & Hill, 2009) on health and chronic illness, this is an important limitation. Differences in physical health outcomes also occur due in part to differing economic resources (Goldman, 2001; Rogers, Hummer, & Nam, 2000); research focusing only on middle-class individuals and families does not help us to understand processes related to health outcomes for families at lower socioeconomic levels. Lastly, along with a lack of organizing, comprehensive models and a need to focus on precise pathways of effect, there is a need for research on adult health to focus on specific health outcomes and to link family processes to high-prevalence conditions and known risk factors of disease (Carr & Springer, 2010). Carr and Springer (2010), in their recent decade review of research on families and health, encouraged researchers to move beyond broad measures of physical health (e.g., all-cause mortality, self-rated health) and investigate family factors linked to actual health outcomes (p. 756). Research that provides evidence for these pathways and specific outcomes would indicate key information needed to translate our understanding of families and health into intervention and practice. Study Purpose The purpose of this study is to address the shortcomings of the literature described above by (a) using a model integrating the effects of close relationships on both mental and physical health, (b) testing precise pathways in this model, including indirect mediation effects, (c) focusing on both family and romantic relationships as predictors of physical health outcomes, (d) using both self-rated health measures and specific indicators of health including current medical diagnoses as dependent variables, and (e) testing the biopsychosocial model with a diverse sample. The present study is a test of the BBFM with a sample of adult primary care patients, in order to attend to the limitations of previous research and expand our understanding of the effects of close relationship functioning on physical health. Although the BBFM has a broad developmental application and can theoretically be used to explain illnesses primarily psychological or biological in origin (Wood et al., 2000; Wood & Miller, 2002; Wood & Miller, 2005), it has not yet been tested with adult family members, nor with a broad array of chronic conditions important to adult health. Additionally, the use of an adult primary care sample at a

clinic serving a diverse, underserved patient population serves the studys goals of expanding the BBFM and addressing current limitations in the literature. A purpose of testing the models applicability to this population is to highlight factors that are protective of or detrimental to patient health; evidence of these linkages has several implications for practice, including the development of targeted, early interventions and tailored treatment plans. Overall, the use of the BBFM in this study provides organization and structure to the hypotheses and a sound theoretical foundation on which to test pathways linking relational functioning, emotion dysregulation, and physical health. Hypotheses The purpose of this study is to further test the BBFM with a sample of adult primary care patients, as the model should be especially helpful in understanding links between close relationship functioning and physical health outcomes for adults. This study attempted to closely replicate previous tests of the BBFMs ability to predict health quality (e.g., Wood et al., 2008) by using similar variables that are developmentally appropriate for adult patients. Because this study is a new approach to using the BBFM, it is therefore exploratory. As a result, rather than preliminarily delineating the relationships expected between specific measures, relationships between constructs in the model are broadly hypothesized and exploratory data analysis is used to first examine the data and potentially significant relationships. Consequently, both family functioning and romantic relationship functioning are used as independent variables representative of FEC. This is not unlike previous tests of the BBFM, which use both traditional measures of family climate and parents relationship quality/conflict as endogenous predictors (e.g., Wood et al., 2008). These predictors will be tested in two separate models. Multiple indicators of biobehavioral reactivity, or emotion dysregulation, are also used, including measures of depression, anxiety, and alcohol abuse; again, this is not unlike previous research using the BBFM (e.g., Wood et al., 2006; Wood et al., 2007). Lastly, both self-report and objective indicators of physical health are used, similar to prior BBFM research (e.g., Wood et al., 2007), but unique to families and health research more broadly (Carr & Springer, 2010). In sum, the following are hypothesized for both models: (1) A direct pathway between FEC and biobehavioral reactivity/emotion dysregulation; (2) A direct pathway between biobehavioral reactivity/emotion dysregulation and disease activity; and

(3) A nonsignificant pathway between FEC and disease activity (Figure 1).

FEC

Disease Activity

Biobehavioral Reactivity

Figure 1. The Biobehavioral Family Model (Wood, 1993)

CHAPTER 2 LITERATURE REVIEW Families and Health There is increasingly consistent evidence linking close relationships to health outcomes and well-being (Carr & Springer, 2010). Familial relationships, including romantic relationships, can both buffer and potentiate risk factors related to health (Wood & Miller, 2002). This is in part because these relationships are more emotionally intense than most other social relationships and because they are longitudinal, continuing over time (Weihs, Fisher, & Baird, 2002). Stable, safe, and supportive close relationships help individual members coping with illness to regulate emotional distress that is due in whole or part to the chronic disease (Weihs et al., 2002). In contrast, conflictual and negative family relationships can interfere with emotion regulation (Fiscella, Franks, & Shields, 1997), while resulting physiological changes due to emotion dysregulation can influence the development of disease (Kiecolt-Glaser et al., 1997; McEwen, 1998). Overall, however, empirical evidence substantiating psychological mechanisms that link social support and health indicators (e.g., cardiovascular function) is lacking (Uchino, Cacioppo, & Kiecolt-Glaser, 1996). In addition, there is a lack of research connecting the findings that negative family relationships affect emotion regulation and that this reactivity influences disease activity (Weihs et al., 2002). This is in part because mental health and physical health outcomes are often studied separately. Mental and physical health connections. Global reports of the contribution of either mental or physical health to mortality and disability underestimate the complexity of the interaction between the two, despite repetitive evidence that mental health concerns are related to and interact with physical health conditions (Prince et al., 2007). In a nationally representative sample of adults who completed the World Health Organization Disability Assessment Schedule, 53% reported 1 or more mental or physical conditions and these individuals reported an average 32 more role-disability days in the past year than matched controls; much of these effects could be accounted for by the effects of comorbid conditions (Merikangas et al., 2007). Of individuals with a serious mental illness, 74% have been given a diagnosis of 1 or more chronic health conditions, while 50% have a diagnosis of two or more. The number of health conditions an individual has is significantly related to annual costs of health care treatment (Jones et al., 2004).

Recent research outlining the physiological effects of relationships on health, although groundbreaking (Carr & Springer, 2010), continues to separately document effects on mental and physical health (e.g., Seeman, 2001). Mental and physical health issues in combination are complex (Prince et al., 2007), disabling (Merikangas et al., 2007), and costly (Egede, Zheng, & Simpson, 2002; Goetzel, Hawkins, Ozminkowski, & Wang, 2003; Jones et al., 2004). It is necessary to proceed with research on the effects of family relationships on health using a biopsychosocial model (Engel, 1977), such as the BBFM. The BBFM: An Organizing Framework The BBFM (Wood, 1993) is a biopsychosocial approach to understanding the mutual influences of family relationships and psychological processes on physical health outcomes. The model serves as the guiding framework for this study and maintains several theoretical assumptions based in general systems theory (von Bertalanffy, 1969) and the psychosomatic family model (Minuchin, Rosman, & Baker, 1978). Theoretical assumptions. The BBFM builds upon several of the foundational premises of Minuchins (1974) structural family model, also known as the psychosomatic family model when used to explain the reciprocal effects of family relationships and illness (Minuchin et al., 1978). The first is that the family is a system; the BBFM maintains a general systems framework (von Bertalanffy, 1969). The second is that individual functioning and relational patterns are interactive and mutually impact each other. The third is that interpersonal patterns interact with individual biobehavioral processes and that some of these are related to health and illness (Wood & Miller, 2002, p. 60). The authors of the BBFM propose that three areas of functioning, which coexist and mutually influence one another, need to be considered when understanding physical health: the biological, psychological, and social (Wood & Miller, 2002). The model explains that family relationship patterns influence the psychological and biological processes of individuals within the family (Wood, 1993; Wood et al., 2000), and uses psychobiologic mediators to connect family emotional climate and disease activity (Miller & Wood, 2003; Wood et al., 2006; Wood et al., 2007; Wood et al., 2008). Constructs. The BBFM was originally an attempt to address the limitations of Minuchins psychosomatic family model, which describes family relational processes and their association with child illness (Minuchin et al., 1975). Wood and others (1989) sought to define the limitations of the model, to reformulate it to include child biobehavioral reactivity (Wood,

1993), and to add parent-child attachment security as a mediating or moderating factor (Wood, Klebba, & Miller, 2000). In sum, the BBFM models the interdependence of family relationships, emotional processes, and physiological changes, describing the contributions of family emotional climate (FEC) and biobehavioral reactivity to disease severity (Wood, 1993; Wood & Miller, 2002). Family emotional climate. FEC is used in the BBFM to describe the intensity and positivity or negativity of emotional processes within the family (Wood et al., 2008). More specifically, the model posits that processes including proximity, relationship quality, and interpersonal responsivity are part of the familys emotional climate and interact with individual family member psychological and emotional processes. Proximity, within the FEC, is defined as sharing space, private information, and emotional experiences, while relationship quality is defined as interactions including mutual support, understanding, and respectful disagreement; responsivity is described as how family members react to one another (Wood & Miller, 2002). The BBFM specifically hypothesizes a direct relationship between FEC and biobehavioral reactivity, or more typically, a direct relationship between a negative FEC (characterized by hostility and criticism) and emotion dysregulation (typically measured as anxiety or depression) (e.g., Wood et al., 2007). Research using the BBFM has tested the effects of FEC on illness in several ways. Wood et al. (2000) measured both childrens perceptions of their parents conflict and childrens beliefs they had caused interparental conflict (triangulation/self-blame), to represent a negative FEC. These authors found that childrens self-blame was significantly related to their feelings of hopelessness and vagal activation. Alternatively, Wood et al. (2006) used parent report of emotional expressiveness in the family as their measure of FEC, as have others (Wood et al., 2007). Findings indicate a significant pathway between a negative balance of FEC and child symptoms of depression and anxiety, although pathways at times appear to differ for mothers and fathers relationships with children (Wood et al., 2006). Additional tests of the BBFM have used child reports of parental psychological aggression, child reports of exposure to violence and hostility in the home (Woods & McWey, 2011), and observational coding of family interactions (Wood et al., 2008) to represent FEC. All find similar relationships between a negative FEC and child experiences of emotion dysregulation, including anxiety and depression.

Family emotional climate is a highly relevant construct for adult family members (Carr & Springer, 2010; Weihs et al., 2002), and especially for understanding the well-being of adult primary care patients. Families are often the primary source of social support for adults (Fiscella et al., 1997), and family stress is often linked to a lower quality of life and higher healthcare utilization for primary care patients (Parkerson, Broadhead, & Tse, 1995). Additionally, higher perceived family criticism predicts an increase in primary care visits for both psychosocial and biomedical reasons (Fiscella et al., 1997). Family functioning is evident from the first primary care visit; whether family members are supportive of each other is critical to understanding patient coping and well-being (McDaniel, Campbell, Hepworth, & Lorenz, 2004). Lastly, in a sample of family practice patients surveyed about their families and relationship quality, the majority believed physicians should ask about family conflict (96%) and that their physicians could be helpful in intervening and providing information or referrals (93%); two-thirds of this same sample explained their physician had never asked about family conflict (Burge, Schneider, Ivy, & Catala, 2005). Just as previous researchers using the BBFM include both family closeness and parental romantic relationship quality in their conceptualization of FEC (Wood et al., 2008), the present study includes measures of both family functioning and romantic relationship satisfaction as independent variables. In fact, family is typically defined as two ore more closely, intimately connected persons who have strong emotional bonds, a history, and a future as a group (Gilliss, Highly, Roberts, & Martinson, 1989; Weihs et al., 2002; Wood & Miller, 2005). Therefore, measuring family emotional climate does not preclude assessing for both the emotional and supportive nature of an individuals romantic relationships, as well as their family relationships more broadly defined (e.g., family-of-origin, family-of-procreation, spiritual family, etc.). Biobehavioral reactivity. Biobehavioral reactivity is the portion of the BBFM that connects family process to disease-related physiological processes (Wood & Miller, 2002). Biobehavioral reactivity is the way in which a family member responds to emotional stimuli (Wood et al., 2008); it is the ability of an individual to regulate their emotions and is measured as emotion dysregulation, or, as symptoms of anxiety and depression (e.g., Wood et al., 2007; Wood et al., 2008). Emotions reflect an individual family members adaptations to changes in environmental stress and are self-regulatory responses that help to coordinate adjustments and behaviors (Thayer & Lane, 2000). Disorders of emotion regulation, including anxiety and

10

depression (Kovacs et al., 2006; Wood et al., 2008), represent an individuals inability to adapt and assume emotions appropriate for the demands of the environment (Friedman & Thayer, 1998). Emotion dysregulation is accompanied by physiological dysregulation and transmits (or escalates) the effect of stress and emotional challenge to disease processes by way of psychophysiological pathways (Wood et al., 2008, p. 23). This is a highly relevant construct in understanding the health and well-being of individual adult family members and primary care patients. Prevalence statistics of the rate of anxiety and depression in primary care vary, although all point to concerning, high rates (e.g., Ansseau et al., 2004). Uebelacker, Smith, Lewis, Sasaki, and Miller (2009) explain that depression is one of the most common presenting problems in primary care clinics, while others assert that anxiety and depression are the 2 most common mental health problems seen in primary care (Kroenke, Spitzer, Williams, Monahan, & Lowe, 2007). Disorders of emotion regulation are also often comorbid with psychosomatic medical conditions; for example, there is evidence that pain and depression are comorbid for 1 in 5 female outpatients (Poleshuck, Giles, & Tu, 2006). Interestingly, primary care patients with more severe medical illnesses are more likely to be recognized as having depression or anxiety by their providers (Robbins, Kirmayer, Cathebras, Yaffe, & Dworkind, 1994). Recent estimates suggest 40% of primary care patients receive mental health care solely from primary care providers (Uebelacker, Wang, Berglund, & Kessler, 2006). Additionally concerning is that they may not be receiving adequate treatment within these primary care visits (Wang et al., 2005) and would likely receive better follow-up care from a mental health specialist (Simon, VonKorff, Rutter, & Wagner, 2001). Depression screening significantly increases the duration of a primary care appointment (Schmitt, Miller, Harrison, & Touchet, 2010) and the recognition of depression and anxiety by primary care physicians is greatly affected by physician characteristics, including sensitivity to nonverbal emotional expressions and willingness to formulate a psychiatric diagnosis (Robbins et al., 1994). Biobehavioral reactivity, connected with physiological dysregulation, involves several biological systems, including the hypothalamic-pituitary-adrenal axis and the autonomic nervous system; these arousal processes then influence the development and course of physical diseases (Wood et al., 2008; Wood & Miller, 2005). Our emotional interpretations of stimuli, including close relationships, affect the neuroendocrine activity of the body, which results in emotional and

11

behavioral responses, including experiences of fear, anxiety, and other emotional states (Seeman, 2001). Recent research on close relationships and health is only beginning to link how relationships literally get under [the] skin (Seeman, 2001, p. 196) of adults, focusing on endocrine, metabolic, immune, and sympathetic nervous systems (Carr & Springer, 2010). Disease activity as the dependent variable. The BBFM predicts physical health outcomes as exogenous variables, although theoretically the model hypothesizes that families are reciprocally affected by a family members illness (Wood, 1993). Because of the focus of previous research using the BBFM on pediatric asthma, measures of disease activity have focused mainly on health outcomes related to asthma severity. For example, Wood et al. (2006) used National Heart, Lung, and Blood Institute (NHLBI; 1997) criteria to diagnose asthma and type the severity level of the illness for children. The authors used measures of pulmonary function, frequency of daytime and nighttime symptoms, and both a research nurse and asthma specialist to diagnose the disease. Later tests used similar NHLBI criteria (e.g., Wood et al., 2007; Wood et al., 2008). In contrast, other research has measured heart rate fluctuations and respiratory sinus arrhythmia to measure vagal activation, used as the dependent variable in testing the BBFM (Wood et al., 2000). Families and health research focusing on health outcomes for adult patients has focused on broad measures of physical health, including all-cause mortality and participant self-rated health (Carr & Springer, 2010). For example, there is repetitive evidence of the protective effect of marriage for general measures of health, including number of illnesses reported (Lorenz, Wickrama, Conger, & Elder, 2006) and self-rated health (Williams & Umberson, 2004), while marital strain leads to a decline in self-rated health (Umberson, Williams, Powers, Liu, & Needham, 2006). In general, there is less attention in families and health research on diseases of adulthood (Weihs et al., 2002). For example, of the three diseases with the highest costs to the U.S. healthcare system cardiovascular disease, COPD and asthma, and non-insulin dependent diabetes the latter two have been the subject of very little family-focused intervention research (Weihs et al., 2002, p. 16). This is despite the fact that family researchers have concluded that the influence of families on physical health is equally as powerful as traditional medical, biological risk factors (Campbell, 2003). Pathways. Two significant pathways are predicted in the BBFM: a significant relationship between FEC and biobehavioral reactivity, and a significant relationship between

12

biobehavioral reactivity and disease activity (Wood, 1993). Although previous research provides evidence of these pathways separately, tests of the BBFM are the first to tie these links together (e.g., Wood et al., 2007). Pathway between FEC and biobehavioral reactivity. The connection between FEC and emotion regulation is a pertinent pathway for understanding adult functioning and health. Family relationships, including family expressed emotion, are stronger predictors of mental health than social relationships more broadly defined (Franks, Campbell, & Shields, 1992). There is also evidence of a relationship between perceived criticism and expressed emotion in ones family and depression and anxiety (Shields, Franks, Harp, Campbell, & McDaniel, 1994), a relationship between family cohesiveness and psychosocial risk factors for pregnant women (Balcazar, Krull, & Peterson, 2001), and evidence that supportive family processes affecting an individuals work experiences decrease the odds of problem drinking (Grzywacz & Marks, 2000). Lastly, research suggests that adults exposed to family risk factors (e.g., conflict, aggression, lack of support) as a child have disruptions in their ability to regulate their emotions and adapt to stress; this interference in emotion regulation leads to later mental and physical health problems in adulthood (Taylor, Lerner, Sage, Lehman, & Seeman, 2004). In adulthood, marital distress and negative behaviors of ones partner are associated with poor mental health (Hawkins & Booth, 2005) and, although it has previously been hypothesized that emotion regulation (as part of psychological processes) serves as a mediator between marital quality (as part of marital functioning) and physiological outcomes (Burman & Margolin, 1992; KiecoltGlaser & Newton, 2001), this comprehensive view of close relationships and health has yet to be tested (Robles & Kiecolt-Glaser, 2003). The pathway between FEC and biobehavioral reactivity is critical in the BBFM, as it links family processes to individual emotional responses to stress (Wood & Miller, 2005). There is increasing evidence demonstrating changes in neuroendocrine activity resulting from close relationships (Robles & Kiecolt-Glaser, 2003; Seeman, 2001). Kiecolt-Glaser et al. (1997) demonstrated that marital conflict in older adults is significantly related to endocrine changes; for wives, negative behavior during conflict and marital satisfaction accounted for 16% to 21% of the variance in changes in cortisol, adrenocorticotropic hormone (ACTH), and norepinephrine. This is also the case for newlyweds: hostile behaviors during conflict resulted in changes in epinephrine, norepinephrine, and ACTH for newly-married couples (Malarkey, Kiecolt-Glaser,

13

Pearl, & Glaser, 1994). Moreover, more hostile couples demonstrated persistently higher rates of all three endocrine markers during conflict and for 15 minutes afterwards (Malarkey et al., 1994); elevated levels of epinephrine that lasted throughout the entire day (following the studyinitiated conflict) predicted couples who eventually divorced compared to those who did not (Robles & Kiecolt-Glaser, 2003). Recommendations for future research connecting relational processes and physical health through psychophysiological pathways include (a) use of objective health outcome assessments, (b) expanding samples to include more than healthy, happy participants, and (c) use of measures that assess both distress and support in close relationships (Robles & Kiecolt-Glaser, 2003). Pathway between biobehavioral reactivity and disease activity. As described above, the comorbidity of mental and physical health conditions is considerable. Four out of 10 patients with a chronic illness have also had a recent or concurrent mental health disorder (Katon & Sullivan, 1990). Disorders of emotion regulation (mood and anxiety) are linked to physical illnesses through biological pathways, both in the BBFM and in prior research (Cohen & Rodriguez, 1995). Emotion dysregulation that includes the continual, persistent, excessive activation of endocrine systems, including the sympathetic-adrenal medullary system as an example, are likely to result in a medical condition (Cohen & Rodriguez, 1995). Constant activation of these physiological processes is implicated in diseases such as coronary heart disease (Manuck, Marsland, Kaplan, & Williams, 1995), hypertension (Krantz & Manuck, 1984), and susceptibility to infectious diseases (Cohen & Rodriguez, 1995; Robles & KiecoltGlaser, 2003). In summary, the bodys regulatory systems connect emotional experience to the physiologic stress response (Weihs et al., 2002, p. 9). Changes that occur throughout the bodys systems, including neurochemical and endocrine systems, influence the development of disease (Kiecolt-Glaser et al., 1997). This is reflected in the notion of allostatic load, which is the cost of chronic exposure to fluctuating or heightened neural or neuroendocrine response resulting from repeated or chronic environmental challenge that an individual reacts to as being particularly stressful (McEwen & Stellar, 1993, p. 2093). Chronic stress from close relationships leads to sustained endocrine changes and emotion dysregulation, with long-term effects on physical health and increased vulnerability to developing illness (McEwen & Stellar, 1993). Evidence substantiates the connection between allostatic load and emotion regulation:

14

increased allostatic load, measured using biomarker indices, is associated with increased, acute depressive symptoms (Juster et al., 2011). Disease outcomes for adult patients resulting from allostatic load include coronary heart disease, obesity, diabetes, hypertension, ulcers, and asthma, among others (McEwen, 1998). Reduction in allostatic load is related to later, lower odds of mortality for older adults (Karlamanga, Singer, & Seeman, 2006). Summary The goal of this study is to test the significance of pathways between (a) FEC and biobehavioral reactivity, and (b) biobehavioral reactivity and disease activity with an adult primary care sample. Despite evidence existing for the relevance of each of these three constructs for adult family members, and despite increasing families and health research on each of these two pathways, comprehensive models integrating these constructs and pathways have yet to be tested (Robles & Kiecolt-Glaser, 2003). Additionally, although the BBFM has broad developmental applications (Wood & Miller, 2005), it has yet to be tested with adults or for health outcomes other than disease activity related to pediatric asthma. This study explores the relationships between (a) FEC, measured for both family relationships and romantic relationships, (b) biobehavioral reactivity, measured as symptoms of depression, anxiety, and alcohol abuse, and (c) disease activity, measured using both self-report and objective medical record data. The present test of the BBFM expands both literature on this biopsychosocial model and the families and health literature more broadly. Further, this study meets recent recommendations in this area to use (a) models that do not only privilege White, middle-class nuclear families, (b) specific health outcomes, including self-report and objective medical data, and (c) precise pathways linking family processes to health outcomes (Carr & Springer, 2010).

15

CHAPTER 3 METHOD Sample The sample included 125 adult patients receiving medical care at an urban primary care clinic in Leon County, Florida. The clinic provides comprehensive primary care services to primarily uninsured or underinsured adults and offers mental health care on-site to its patients. The clinic also offers specialty services including optometry, gynecology, and cardiology. This clinic operates on a sliding fee scale with the majority of patients paying a five-dollar copay for their care. Participants for the present study included primary care patients who presented for a previously scheduled appointment or for an appointment during open walk-in hours. Patients presented for a variety of reasons, including for follow-up care, chronic disease management (e.g., diabetes, hypertension, etc.), psychiatric concerns, and routine eye care. Recruitment. Approval for this study was obtained from the Human Subjects Committee at The Florida State University. Participants were recruited from the primary care site using approved flyers posted in the clinic that explained the process of providing their contact information to clinic staff. In addition, this author and two research team members positioned in clinic waiting rooms provided patients the ability to approach the researchers to participate and provided initial, basic information about the study to seemingly interested patients. Prospective participants were first asked if (a) they were between the ages of 18 and 65 years, (b) were current clinic patients, and (c) were comfortable reading and speaking English. Adults 18- to 65years-old were included because persons under the age of 18 are considered children/minors and adults over 65 are considered elderly, both of which are protected, potentially vulnerable populations (Office of Human Subjects Research, 2006). Participants that met these inclusion criteria were then given further details of the study, provided a thorough consent form with additional information, and encouraged to ask questions about the study. They were explicitly told that the survey was confidential, that their medical providers at the clinic would not have access to the information included, and that their decision to participate would not affect their medical care. Patients that consented to participate completed our paper-and-pencil assessment. They were informed through the consent process that their answers would be reviewed following completion of the survey for any concerns. In addition, patients were able to complete the

16

measures in the waiting room, the hallway of the clinic, their exam room, or in rooms used by the researcher for reviewing surveys following completion. All participants completed the survey confidentially, without the presence of a friend or family member; several participants required the researcher to read the measures aloud due to either vision or literacy concerns. Of those approached to participate (N = 232), 59.5% agreed to participate (n = 138). Several of these patients were interrupted during the consent process in order to meet with their medical provider. In addition, 6 patients began the survey but withdrew before completion for reasons of transportation, prior commitments, and/or discomfort with the content of the survey. Those who expressed feeling disconcerted by the survey (n = 2) were successfully referred to a mental health provider within the clinic. Finally, 2 participants completed the survey but later admitted they were not current patients, and 1 participant took the survey with them and did not return it following their appointment. A total of 125 patients completed the survey (Figure 2).

Approached (n=232) Interrupted during consent process (n=4) Withdrew due to prior commitments (n=3) Withdrew due to discomfort with survey content (n=2) Withdrew due to transportation issues (n=1)

Agreed to Participate (n=138)

Completed Assessment (n=128)

Excluded, not a patient (n=2) Excluded, did not return the survey (n=1)

Final Sample (n=125)

Figure 2. Participation flow diagram.

17

Demographic characteristics. Patients (N = 125) included 34% males (n = 43) and 65% females (n = 81); one participant did not report. The average age was 46 years (SD = 12.2), and the majority identified as Black or African American (59%), with other races identified as White (34%), American Indian or Alaska Native (2%), and Asian (1%); several participants identified as biracial (4%) and 1 participant did not report their race. In addition, 6% of participants identified as Hispanic/Latino ethnicity; 12 participants did not report their ethnicity. Regarding insurance status, 90% of the present sample was uninsured. Of the remainder, several had Medicaid coverage (n = 9) while others had Medicare (n = 2) and one participant had Universal Health Care as the insurance plan noted in their medical chart. The majority of patient participants reported being currently in a romantic relationship (62%, n = 76). In addition, 33% reported their current marital status as divorced, 31% reported never having married, 19% reported they were currently married and living together, 8% reported being currently married but separated, and 7% reported their current marital status as widowed. Participants also reported household income: the majority (66%) reported a household income of less than $10,000, while 14% reported $10,000-$19,999, 12% reported $20,000$39,999, 2% reported $40,000-$59,999, and 3% reported $60,000-$79,999 (3 participants did not report). Participants most often reported having graduated from high school or having obtained a GED (35%), while 27% reported not having graduated from high school and 19% reported some college; the remainder (18.4%) reported a college degree (Associates, Bachelors, Masters, or Professional degrees). Lastly, participants reported their current employment status: the vast majority (71%) reported no current employment, while 19% reported part-time and 10% reported full-time employment. Because the present sample included the full range of adults from ages 18- to 65-yearsold, it is important to consider the broad developmental range of these persons. The participants span early to middle adulthood (Hewstone, Fincham, & Foster, 2005) and may be experiencing change that is maturational, normative, or in response to predictable events and contextual changes (Franz, 1997). According to the family life cycle perspective (Carter & McGoldrick, 1999), adults in the present sample may range in life cycle stages, including preparing to leave home, joining through marriage, having young children, adolescents, or launching young adults, or may be learning to support their own aging parents (Nichols & Schwartz, 2004). Thankfully, the vast majority of the present sample was able to identify persons with whom they have close

18

relationships and consider to be family: 124 participants completed at least 75% of the family functioning measure, and were actively encouraged to respond to the items thinking of whoever they consider to be family. Measures The study questionnaire was completed using paper-and-pencil. All measures were offered in English. Family emotional climate. Three measures were used to assess family emotional climate; one measure represents participants perceptions of their family relationship quality and two are self-report measures of romantic relationship quality. The two romantic relationship measures were only completed by participants who indicated they are currently married or in a romantic relationship. General Functioning Family Assessment Device. Participants completed the General Functioning Subscale of the Family Assessment Device (GFS/FAD; Epstein, Baldwin, & Bishop, 1983). The GFS/FAD assesses relational functioning in families and has good reliability, demonstrated using Cronbachs alpha and spit-half correlations (Byles, Byrne, Boyle, & Oxford, 1988). The measure is valid, as demonstrated through correlations with other family variables, and is recommended as a global assessment of family functioning (Byles et al., 1988). The measure includes 12 items rated on a 4-point Likert scale (Strongly Agree to Strongly Disagree); example items include, We feel accepted for who we are, and, We can express feelings to each other. Scores on the GFS/FAD range from 1.00 (healthy family functioning) to 4.00 (unhealthy family functioning); the higher the score, the more problematic the family member perceives the familys overall functioning (Ryan, Epstein, Keitner, & Miller, 2005). A score of 2.00 or greater indicates problematic family functioning. Quality of Marriage Index. Participants completed the Quality of Marriage Index (QMI; Norton, 1983), a 6-item questionnaire assessing romantic relationship quality. The QMI has demonstrated high internal consistency coefficients ( = .96) for husbands and wives (Karney, Bradbury, Fincham, & Sullivan, 1994) and convergent validity is supported by correlations with the Dyadic Adjustment Scale (DAS; Spanier, 1976) and the Kansas Marital Satisfaction Scale (Schumm et al., 1986). Items 1 through 5 on the scale use a 7-point Likert scale ranging from 1 (very strong disagreement) to 7 (very strong agreement); item 6 uses a 10-point Likert scale ranging from 1 (very unhappy) to 10 (perfectly happy). Items are summed to produce a score

19

ranging from 6 to 45; higher scores are indicative of greater relationship satisfaction. A cutoff score of 29 or less equates to a score of 97 or below on the DAS and indicates relationship distress (Heyman, Sayers, & Bellack, 1994). Perceived Criticism. Perceived criticism was measured with two items using 10-point Likert scales (Hooley & Teasdale, 1989). Those two items ask, How critical is your partner of you?, (PC-Partner) and, How critical are you of your partner? (PC-Self) and are anchored by not at all critical and very critical indeed. Perception of ones romantic partner as critical is strongly linked to relapses in symptoms of depression (Hooley & Teasdale, 1989), substance abuse (Fals-Stewart, OFarrell, & Hooley, 2001) and symptoms of anxiety disorders (Chambless & Steketee, 1999). The two-item measure has demonstrated discriminant validity and moderate convergent validity among depressed and dysthymic patients (Riso, Klein, Anderson, Ouimette, & Lizardi, 1996). There is presently no clinical cutoff for this measure. Biobehavioral reactivity. Four measures were used to assess participants responses to emotional stimuli (Wood et al., 2008) and the degree of their emotional regulation or dysregulation (Wood & Miller, 2002) by assessing alcohol use disorders, symptoms of depression, and anxiety severity. NIAAA Screen. The National Institute on Alcohol Abuse and Alcoholism (NIAAA) endorses the use of a one question to screen patients for alcohol consumption (NIAAA, 2005) This screen has been found to correctly predict whether patients meet either NIAAAs criteria for at-risk drinking or DSM-IV criteria for an alcohol use disorder (Taj, Devera-Sales, & Vinson, 1998). The question asks, On any single occasion during the past 3 months, have you had more than 5 drinks containing alcohol? and participants answered either yes or no (Taj et al., 1998). Alcohol Use Disorders Identification Test (AUDIT). Alcohol use was also assessed using the AUDIT (Saunders, Aasland, Babor, de la Fuente, & Grant, 1993), which includes 10 items that measure whether a patient has problems with alcohol but who may not be dependent. Example items include, How often do you have six or more drinks on one occasion? and, How often during the last year have you had a feeling of guilt or remorse after drinking? A cutoff score of 8 or more on this measure indicates harmful or hazardous drinking; the measure is equally appropriate for both males and females (Allen, Litten, Fertig, & Babor, 1997). The validity of the AUDIT is well-established by both relationships to other self-report measures and

20

to biochemical measures of excessive drinking (Allen et al., 1997); internal consistency has been repetitively demonstrated to be high (e.g., = .94, Skipsey, Burleson, & Kranzler, 1997), including for primary care patients ( = .77, Schmidt, Barry, & Fleming, 1995). Quick Inventory of Depressive Symptomatology Self Report. The 16-item Quick Inventory of Depressive Symptomatology Self Report (QIDS; Rush et al., 2003) was used to assess the severity of depressive symptoms within the nine diagnostic symptom domains that characterize a major depressive episode in the DSM-IV (American Psychiatric Association, 2000). The measure demonstrates good internal consistency ( = .86, Rush et al., 2003; = .85, Trivedi et al., 2004) and content, criterion, and construct validity (Rush et al., 2003; Trivedi et al., 2004). Each item asks participants to describe how often each symptom of depression occurred in the past 7 days; an example asks respondents to rate feeling sad and includes response choices ranging from I do not feel sad to I feel sad nearly all the time. The exception is the QIDS item asking participants to describe whether they lost or gained weight; this item asks patients to reflect on the last 14 days. Scores on the QIDS range from 0 to 27; a cutoff score of 7 or higher indicates depression (Rush et al., 2003). More specifically, scores ranging from 7 to 10 indicate mild depression, scores from 11 to 15 indicate moderate depression, scores from 16 to 20 indicate severe depression, and scores from 21 to 27 indicate very severe depression (Health Technology Systems, 2012; Rush et al., 2003). Overall Anxiety Severity and Impairment Scale. Participants completed the 5-item Overall Anxiety Severity and Impairment Scale (OASIS), which addresses intensity and frequency of anxiety, interference with work or school, and interference with personal relationships (Norman, Cissell, Means-Christensen, & Stein, 2006). The measure was developed to capture anxiety severity and impairment in a brief manner and items are based on the DSM-IV guidelines of severity and impairment (American Psychiatric Association, 2000). Each question has a 5-point response scale, ranging from 0 to 4, with descriptors reflective of the aspect being assessed. For example, one item asks, In the past week, how often have you felt anxious?; participants can answer on a range of 0 (0 = NO ANXIETY in the past week) to 4 (4 = CONSTANT ANXIETY. Felt anxious all of the time and never really relaxed). The OASIS demonstrates adequate reliability ( = .80) and excellent convergent validity with other established anxiety measures (Norman et al., 2006). Scores on the OASIS range from 0 to 20;

21

scores greater than or equal to 8 identifies clinically significant anxiety (Campbell-Sills et al., 2009). Disease activity. To assess the impact of family and romantic relationships and symptoms of emotion dysregulation on the physical health of these patient participants, we used both self-report data and indicators collected from patient electronic medical records. Health Review. To assess specific illness symptoms from the participants perspectives, we used the 21-item Health Review (Epidemiology Data Center, 2011; Jenkins, Kreger, Rose, & Hurst, 1980; Rose, Jenkins, & Hurst, 1978). The measure is a checklist of symptoms primarily related to infectious disease and continuing health problems and asks participants to answer yes or no to whether, in the past month, they have experienced any of the symptoms. The measure focuses on specific, well-operationalized symptom clusters (Kiecolt-Glaser & Newton, 2001, p. 480) and is consistently related to physicians diagnoses (Jenkins et al., 1980; Kiecolt-Glaser, Dura, Speicher, Trask, & Glaser, 1991; Orts et al., 1995), supporting its validity. The measure demonstrates good interrater reliability for individual symptoms when administered as an interview and excellent agreement between raters who applied International Classification of Diseases, Ninth Revision (ICD-9) criteria to patients self-reports; test-retest reliability is also high (Kiecolt-Glaser et al., 1991). The items used for this administration were retrieved from the baseline battery of the REACH II study at the University Center for Social and Urban Research (Epidemiology Data Center, 2011). Scores on the Health Review range from 0 to 21; higher scores indicate more illness symptoms. RAND-36. Patient perceptions of their general health were assessed using four subscales of the RAND 36-item Health Survey (Hays, Sherbourne, & Mazel, 1995), including the Physical Functioning (10 items), Role Limitations Due to Physical Health (4 items), Pain (2 items), and General Health (5) subscales. The measure has been demonstrated to be reliable and valid, is widely used for similar purposes (Hays & Morales, 2001), and has high internal consistency (VanderZee, Sanderman, Heyink, & Haes, 1996). The items of this scale are identical to the more widely used SF-36 (Ware & Sherbourne, 1992), with the exception of a simpler, more straightforward scoring method (Rand Health, 2010). All subscales for the RAND-36 are scored so that a high score demonstrates a more favorable health state (Hays et al., 1995). There are presently no clinical cutoffs for this measure.

22

Medical chart data. Health data were also gathered directly from participants electronic medical records. Information collected included the participants most recent Body Mass Index (BMI; calculated from a persons height and weight), which provides an indicator of unhealthy weight that may potentially lead to health problems (CDC, 2011a). Each patients BMI was also categorized into the CDCs four weight statuses: underweight (BMI of below 18.5), normal (18.5 to 24.9), overweight (25.0 to 29.9) and obese (BMI of 30 or higher) (CDC, 2011a). Additional information collected included the number of primary care appointments the patient had within the past year, number of medical diagnoses (received in the past year and chronic), and the number of mental health diagnoses (given by NHS mental health providers and primary care providers). All medical diagnoses were verified using the ICD-9, published by the World Health Organization (CDC, 2009); diagnoses representing ICD-9 supplementary v-codes were excluded. The number of mental health diagnoses for the present study excluded diagnoses of depression and anxiety to avoid artificially inflating associations between measures of depression and anxiety included as mediators and the disease activity outcomes. Participant health insurance status and type were also gathered to characterize the sample. Analyses Exploratory data analysis. Due to the exploratory nature of this study, exploratory data analysis was first used to better understand the data collected and to test for significant relationships among constructs prior to testing full models. An exploratory approach is recommended to more effectively use data to test hypotheses by first learning about the data, distributions of variables, and relationships between variables (Hartwig & Dearing, 1979; Smith & Prentice, 1993). Therefore, the first step in these analyses was to create boxplots of each variable to examine the data for outliers. Second, associations between constructs included in the BBFM models were tested using correlation analyses. Third, relationships between (a) FEC variables and biobehavioral reactivity variables, and (b) biobehavioral reactivity variables and physical health outcomes were tested using stepwise regressions. Variables that demonstrated significant relationships were then tested further using path analyses. Model testing. To test the BBFM models proposed in the hypotheses above (one model using family functioning as the independent variable, a second using romantic relationship satisfaction as indicative of FEC), path analyses were used. Variables with significant relationships highlighted in the exploratory data analyses were used to build the models we

23

tested. Kline (2011) suggests a minimum sample size of 100 for structural equation modeling. He also suggests, along with Jackson (2003), that for statistical power to equal .80 in path analysis using maximum likelihood, the ratio of sample size to parameters in the model should be at least 10:1. Therefore, acceptable statistical power is achieved in this study, with a sample size of 125. Standardized path analysis coefficients were estimated in AMOS (Arbuckle, 1997) using direct maximum likelihood estimation. Model fit statistics are reported for each model. Consistent with Byrnes (2010) recommendations, assessment of model fit was based on three criteria: 2 likelihood ratio statistics, comparative fit index (CFI) and root mean square error of approximation (RMSEA). The goal is to achieve a nonsignificant chi-square (2) value, indicating that the variances and covariances of [the] hypothesized modeldid not differ significantly from those in the data set (Bikos & Kocheleva, 2012, p. 11). Additionally, the goal is to have a CFI value greater than .95 and an RMSEA of less than .05. Regarding missing data, maximum likelihood estimation in AMOS ensures the full sample is used for model testing. Maximum likelihood creates estimates of data missing for individual participants. In this approach, missing values are not imputed, but all observed information is used to produce the maximum likelihood estimation of parameters (Acock, 2005, p. 1018). Therefore, cases are not deleted if data is incomplete; all cases are entered into the maximum likelihood estimation for path analysis (Kline, 2011). Maximum likelihood is standard in structural equation modeling and use of a different estimation method requires justification (Hoyle, 2000). In addition, bootstrapping methodology, a nonparametric resampling procedure (Preacher & Hayes, 2008), was used to confirm the results of the path analyses. As demonstrated in previous research (e.g., MacDonnell, Naar-King, Murphy, Parsons, & Harper, 2010), bootstrapping is appropriate for use with smaller sample sizes because it increases the power of statistical results (Shrout & Bolger, 2002) and involves repeatedly sampling from the actual data to generate an empirical approximation of the sampling distribution and confidence intervals for the indirect effect (Preacher & Hayes, 2008). Bootstrapping generates confidence intervals for the size of the indirect path; a statistically significant mediation effect is indicated if the values between the upper and lower confidence limits in the confidence interval (CI) do not include zero. Bias-corrected and accelerated intervals were examined, as recommended by Efron (1987) as an improvement on traditional CI and bootstrapping methods.

24

CHAPTER 4 RESULTS As outlined above, analyses will proceed in five steps: (1) examination of the data for outliers and scores outside expected scale ranges, (2) correlation analyses, (3) stepwise regressions, (4) model testing using path analyses, and (5) bootstrapping tests of mediation. Exploratory Data Analysis Family emotional climate independent variables. The range of scores for the QMI fell within expected (6 to 45), as did the range of scores for the perceived criticism items (both 1 to 10) and the scores for the GFS/FAD subscale (1.00 to 4.00). An examination of boxplots for each scale demonstrated one outlier for the GFS/FAD subscale: one participant scored a 4.00, which, although it falls within the acceptable scale score range, fell far from the mean of 2.09. Biobehavioral reactivity mediators. The range of scores for the AUDIT fell within the expected range (0 to 23), as did the range of scores for the OASIS (0 to 20) and the QIDS (0 to 25). An examination of the boxplots for each scale demonstrated two outliers for the AUDIT: two participants scored a 23, which falls within the expected score range, although they fall far from the mean of 3.77. Disease activity dependent variables. The range of scores for the Health Review fell within the expected range (0 to 20), as did all four RAND-36 subscales (0 to 100 for each). The boxplot for the Health Review scale revealed two outliers: one participant scored a 17 while a second scored a 20. While these two scores fall in the acceptable scale score range, they fell far from the mean of 4.89. None of the RAND-36 boxplots demonstrated outlier values. Because all values for each scale fell within the expected range, all cases will be used in the remainder of the statistical analyses. Study variables. A description of all study variables is provided in Table 1. As described above, the GFS/FAD, QMI, AUDIT, QIDS, and OASIS all have precise clinical cutoff scores; the BMI has clinical cutoff scores defined by the CDC for underweight, overweight and obese status (CDC, 2011a). In the present sample, the mean scores for the GFS/FAD, QIDS, and BMI measures were all at or above the recommended clinical cutoff scores. In other words, the average individual in this sample was obese, had problematic family functioning, and reported clinical levels of depression.

25

Table 1. FEC Variables, Biobehavioral Reactivity Variables, and Disease Activity Variables: Descriptive Statistics (N = 125) Variables GFS/FAD QMI PC-Partner PC-Self NIAAA AUDIT QIDS OASIS Health Review RAND-36 Subscales Physical Functioning Role Limitations Pain General Health BMI 60.33 49.80 52.98 50.96 31.87* 32.35 43.36 31.79 26.51 8.40 2.64 1.90 .333 123 123 125 124 116 116 116 116
a

M 2.09* 31.95 5.43 5.68 .32 3.77 9.36* 5.44 4.89

SD .67 10.00 2.92 2.92 .469 5.07 5.76 5.62 3.97

n 109 66 77 76 124 119 124 120 115

Number of 4.47 appointments Number of medical 1.91 diagnoses Number of mental .10 health diagnoses a NIAAA Screen: 0 = no, 1 = yes. *Mean scores represent scores at or above clinical cutoff.

Construct Associations Associations between variables were initially tested using bivariate, two-tailed Pearson correlations (Table 2). These were conducted to examine the potential for scales to be used in later model testing (i.e., if demonstrated as significant in preliminary correlation tests). The correlations were also examined to check for multicollinearity; low multicollinearity is a statistical assumption of path analysis (Kline, 2011). Two correlations were greater than .70, the 26

Table 2. Bivariate Pearson Correlations of Study Variables

Measures 1. GFS/ FAD 2. QMI 3. PC-S 4. PC-P 5. NIAAA 6. AUDIT 7. QIDS 8. OASIS 9. HR-21 10. R36PF 11. R36RL 12. R36PA 13. R36GH 14. BMI 1 1.00 -.27* .13 .27* .04 .10 .28** .12 .23* -.27** -.18 -.10 -.10 -.16 1.00 -.08 .01 .08 .02 -.31* -.24 -.15 -.03 -.04 -.02 .15 .10 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

1.00 .69** .01 -.03 -.05 -.06 -.04 -.01 .01 .00 .10 .05 1.00 .01 .08 .01 .01 -.04 -.08 .023 .08 .03 -.09 1.00 .63** .07 .11 .03 .02 -.04 -.01 .01 -.07 1.00 .14 .07 .01 -.01 -.01 .05 -.07 -.18 1.00 .70** .40** -.23* -.39** -.38** -.44** -.13 1.00 .37** -.27** -.46** -.33* -.32** -.10 1.00 -.29** -.45** -.44** -.43** .16 1.00 .66** .43** .41** -.07 1.00 .54** .48** -.08 1.00 .48** .01 1.00 .03 1.00

15. BMI-.10 -.02 .08 -.05 -.04 -.12 -.12 -.04 .03 -.03 -.03 .05 .03 .79** 1.00 CDC 16. APPTS -.12 -.13 -.16 -.11 -.07 -.07 -.03 -.11 -.07 -.07 .02 -.13 -.11 .05 .14 1.00 17. MED-.10 .02 -.12 -.18 -.07 .01 -.16 -.04 .015 -.14 -.12 -.11 -.10 .08 .21* .33** 1.00 DX 18. MH-.06 .18 -.00 -.02 .07 -.11 -.16 -.16 .02 .04 .02 .10 -.10 -.06 .02 .13 1.00 .24* DX Note. Significant correlations are in boldface. GFS/FAD = General Functioning Subscale of the Family Assessment Device; QMI = Quality of Marriage Index; PC-S = Perceived criticism-self item; PC-P = Perceived criticism-partner item; NIAAA = NIAAA Screen; AUDIT = Alcohol Use Disorders Identification Test; QIDS = Quick Inventory of Depressive Symptomatology Self Report; OASIS = Overall Anxiety Severity and Impairment Scale; HR-21 = Health Review; R36PF = RAND-36 Physical Functioning subscale; R36-RL = RAND-36 Role Limitations subscale; R36-PA = RAND-36 Pain subscale; R36-GH = RAND-36 General Health subscale; BMI = Body Mass Index; BMI-CDC = Body Mass Index weight status as determined by the CDC (CDC, 2011a); APPTS = Number of appointments in the past year; MED-DX = Number of medical diagnoses; MH-DX = Number of mental health diagnoses. *p<.05, ** p<.01 (2-tailed)

27

recommended cutoff for testing multicollinearity (Tate, 1998). The first, a correlation of .70 between QIDS and OASIS scores, was examined by calculating the variables VIF and tolerance statistics. The collinearity statistics indicated the two variables are not replications of one another: the VIF was 2.09 (less than the recommended 10) and the tolerance was .48 (greater than the recommended 0.1) (UCLA Academic Technology Services, 2012). The second, a correlation between BMI and BMI-CDC of .79, was also examined by calculating the variables VIF and tolerance statistics. The VIF was 2.71 and the tolerance was .37, indicating the two do not require further investigation. However, it is logical that BMI and BMI-CDC are closely related, given that BMI-CDC is calculated using BMI. If the use of these variables together in model testing is indicated by stepwise regression analyses, their relationship will be closely analyzed. As predicted, several of the family emotional climate variables were significantly correlated with the biobehavioral reactivity variables. The GFS/FAD was significantly correlated with depression, as measured by the QIDS (r = .28, p = .003), as was the QMI (r = -.31, p = .011). In addition, two of the biobehavioral reactivity variables, depression and anxiety, were significantly associated with several disease activity outcome variables. More specifically, depression was significantly related to Health Review scores (r = .40, p = .000) and all four RAND-36 subscales. Anxiety, as measured by the OASIS, was also significantly related to Health Review scores (r = .37, p = .000) and all four RAND-36 subscales. These results indicate that as family functioning becomes more problematic, depression worsens; as marital satisfaction increases, depression decreases. Also, as depression increases, so do the number of illness symptoms reported by patients. As anxiety becomes more severe, the number of health symptoms reported by patients increases. Lastly, as depression and anxiety symptoms increase, patients self-reports of their physical health as reported on all four subscales of the RAND-36 worsened. Additional results include a significant correlation between family functioning and Health Review scores (r = .23, p = .018), as well as between family functioning and the RAND-36 Physical Functioning subscale scores. In addition, scores on the PC-Self item were significantly related to patients number of mental health diagnoses. Although correlations do not indicate a significant pathway, they are important to highlight as these associations may affect our mediation results.

28

Stepwise Regressions As outlined above, we tested relationships between (a) FEC variables and biobehavioral reactivity variables, and (b) biobehavioral reactivity variables and physical health outcomes using stepwise regressions. The goal, as stated above, is to conduct path analyses for two models: one with family functioning as the independent variable (Model 1) and one with romantic relationship satisfaction as the independent variable (Model 2). Therefore, preliminary stepwise regressions were conducted accordingly in order to explore which relationships make the most sense to test in AMOS. Romantic relationship functioning. First, a stepwise regression was conducted using QMI and the perceived criticism items as predictors and depression scores as the outcome variable. Of those three, only QMI scores were a significant predictor of depression (t = -2.55, p = .013). We used a similar process to test QMI and perceived criticism as predictors of anxiety as measured by the OASIS; none of the variables were indicated as significant. This was also true when we used AUDIT scores as the dependent variable. A logistic regression was used to test QMI, PC-Partner, and PC-Self scores as predictors of NIAAA screen answers, given the dependent variables dichotomous nature; none of these relationships were significant. Family functioning. Because only one scale was used to assess family functioning (GFS/FAD), we conducted a linear regression to test the relationship between family functioning scores and depression. GFS/FAD scores were a significant predictor of QIDS scores (F = 9.04, p = .003). This indicates that as family functioning became more problematic, self-reported depression symptoms worsened. A similar process was used to test family functioning as a predictor of anxiety; this relationship was not significant (F = 1.53, p = .22). GFS/FAD scores were also not a significant predictor of alcohol use, as measured by the AUDIT (F = 1.02, p = .316). A logistic regression was used to test the NIAAA screen as a dependent variable; GFS/FAD was not a significant predictor (p = .672). Biobehavioral reactivity predicting disease activity. First, a stepwise regression was used with all potential biobehavioral reactivity variables (QIDS, OASIS, AUDIT, and NIAAA screen scores) as independent variables and Health Review scores as the dependent variable. Only depression scores were highlighted as significant (t = 4.26, p = .000). This indicates that as the number and severity of self-reported depression symptoms increase, patients self-reported illness symptoms also increase. Similar analyses were used with all four RAND-36 subscales.

29

Depression was the sole significant predictor of Physical Functioning subscale scores (t = -3.13, p = .002), Pain subscale scores (t = -4.13, p = .000), and General Health subscale scores (t = 5.07, p = .000). Both depression (t = -2.22, p = .029) and anxiety scores (t = -2.19, p = .031) were significant predictors of RAND-36 Role Limitations subscale scores. These results indicate that as self-reported symptoms and severity of depression (and anxiety) increase, physical health worsens (as indicated by decreasing RAND-36 subscale scores). Stepwise regressions were also used to test medical chart data as dependent variables. Again using QIDS, OASIS, AUDIT, and NIAAA screen scores as independent variables, we tested the number of primary care appointments in the past year, number of medical diagnoses, number of mental health diagnoses, BMI, and BMI weight status as determined by CDC guidelines as outcomes. Unfortunately, there were no significant relationships between our biobehavioral reactivity variables and these medical chart indicators of disease activity. Overall, significant relationships were found between marital satisfaction and depression; family functioning and depression; depression and illness symptoms; and both depression and anxiety and physical health. Because these findings were generally replicated in both the correlation analyses and stepwise regressions, this is how model testing will begin for both models. Therefore, initially for Model 1 (family functioning), anxiety, alcohol abuse, and medical chart disease activity outcome variables will not be tested. The same is true for the Model 2 (romantic relationship satisfaction). Anxiety will then be added as a potential additional mediating variable in both models. Model Testing Both models of the BBFM were tested with path analyses, and standardized path analysis coefficients were estimated in AMOS using direct maximum likelihood estimation to handle missing data. A model generating approach (Byrne, 2010; Joreskog, 1993) was used when fitting the structural model. More specifically, testing began with an a priori specified model and, because it may be incorrectly specified given the exploratory nature of the study, the goal was to identify significant paths not included at first and to trim nonsignificant paths. First, the initial, hypothesized model was used to ensure the test was theoretically meaningful and parsimonious (Bikos & Kocheleva, 2012; Byrne, 2010; Kline, 2011). To modify each model and find the models with the best fit to the data and theory, a model trimming approach (Kenny, 2011; Kline, 2011) was used, which involves eliminating nonsignificant pathways one at a time on the basis

30

of empirical considerations (Kline, 2011), or, examining the regression coefficients (Anderson, Parmenter, & Mok, 2002). Trimming models during the process of model testing by eliminating variables in order to improve model fit is common practice (e.g., Anderson et al., 2002; DeGarmo & Martinez, 2006; Driver & Gottman, 2004; Gallagher, Ting, & Palmer, 2008; SidoraArcoleo, Feldman, & Spray, 2012) and ensures the best-fitting, most parsimonious model. Model 1: Family functioning. The initial model (Figure 3) used the entire study sample (n = 125) and tested the relationships between family functioning (GFS/FAD), depression (QIDS), and disease activity. Disease activity was a latent, unobserved variable; Health Review scores and all four RAND-36 subscales were used as observed outcome variables loading onto disease activity based on preliminary findings of significant relationships derived from the exploratory analyses. Goodness-of-fit indices for the model indicate the first model does not fit the data (2 = 33.93, p = .001, CFI = .910, RMSEA = .114). Standardized regression weights are reported in Figure 3; these represent the path coefficients for the model. To maintain the test of the BBFM, pathways between family functioning and depression, depression and disease activity, or family functioning and disease activity were not eliminated. Instead, focus was on the RAND-36 measure, which included several subscales. In order to attempt a parsimonious, yet meaningful model, the RAND-36 subscale with the weakest path coefficient was trimmed: RAND-36 General Health scores. The second model (model 1a) is identical to the initial model, but with RAND-36 General Health scores (and the attached pathway) removed. A slight improvement in goodnessof-fit was observed, where the 2 decreased and the CFI improved a small amount (2 = 22.82, p = .004, CFI = .917, RMSEA = .122). The next weakest path coefficient was found between disease activity and RAND-36 Pain subscale scores; therefore, this variable and its pathway were trimmed next. The third model (model 1b) is identical to the first model, but with RAND-36 General Health and RAND-36 Pain subscale scores (and attached pathways) trimmed. Again, a slight improvement in goodness-of-fit was observed (2 = 14.25, p = .007, CFI = .919, RMSEA = .144). Next, RAND-36 Physical Functioning subscale scores were trimmed as a variable, due to its paths weaker standardized regression coefficient. The fourth model (model 1c; Figure 4) is again identical to the initial model, but with RAND-36 General Health, Pain, and Physical Functioning subscale scores (and attached

31

pathways) removed from the model test. The improvement in goodness-of-fit was drastic and demonstrates acceptable agreement between the data and the model tested (2 = .324, p = .569, CFI = 1.000, RMSEA = .000). This model provides support of the theoretical model: pathways between family functioning and depression, and depression and disease activity were significant, while the pathway between family functioning and disease activity was nonsignificant. As discussed above, because anxiety demonstrated significant associations with disease activity outcomes in preliminary analyses, we added this into the model to test anxiety as an additional mediator in the BBFM. The last model (model 1c) was used as the base and a biobehavioral reactivity unobserved variable was created with anxiety and depression loading as observed variables (model 1d; Figure 5). This addition did not substantially affect goodnessof-fit statistics; this test demonstrated acceptable agreement between the data and the model tested (2 = 4.135, p = .247, CFI = .992, RMSEA = .055). In addition, the magnitude of the pathway coefficient between biobehavioral reactivity and anxiety was substantial. Of note is that this model altered the level of significance of the pathway between family functioning and biobehavioral reactivity (p < .05), as compared to in the previous model (model 1c; p <.01). However, the t-ratio for this pathway (2.38) still represents a significant effect of family functioning on biobehavioral reactivity. Therefore, this model is judged to be the best-fitting model for the data, is theoretically meaningful, and is the final model of our test of the BBFM using family functioning as the endogenous variable. Model 2: Romantic relationship satisfaction. The initial model (Figure 6) used the portion of the study sample that reported they were in a current romantic relationship (n = 76) to prevent maximum likelihood estimations conducted in AMOS from fabricating QMI scores for participants who are not married or partnered. Associations between romantic relationship satisfaction (QMI), depression (QIDS), and disease activity were tested. Similar to the model testing above using family functioning as the endogenous variable, disease activity was a latent, unobserved variable. Health Review scores and all four RAND-36 subscales were used as observed outcome variables loading onto disease activity based on preliminary findings of significant relationships derived from the exploratory analyses. Goodness-of-fit indices for the model indicate the first model fits the data, but that there is room for improvement (2 = 21.232, p = .068, CFI = .938, RMSEA = .092). Pathway coefficients are presented in Figure 6. As before, pathways between romantic relationship satisfaction and depression, depression and disease

32

activity, or romantic relationship satisfaction and disease activity were not eliminated in the model building and trimming process. Instead, the RAND-36 subscale with the weakest path coefficient was trimmed: RAND-36 Pain scores. The second romantic relationships model (model 2a) is identical to the initial model, but with RAND-36 Pain scores (and the attached pathway) removed. Unfortunately, the goodnessof-fit statistics worsened: the 2 became significant, the CFI decreased, and the RMSEA increased (2 = 19.657, p = .12, CFI = .890, RMSEA = .139). Therefore, Pain subscale scores were re-entered into the model. The next weakest path coefficient was between disease activity and RAND-36 Physical Functioning subscale scores; therefore, this variable and its pathway were trimmed. This third model (model 2b; Figure 7) demonstrated an enormous improvement in fit. The 2 became nonsignificant, the CFI was above our cutoff of .95, and the RMSEA was below .05 (2 = 6.126, p = .633, CFI = 1.000, RMSEA = .000). As with Model 1 (family functioning), anxiety was added into Model 2 to test these scores as an additional mediator in the BBFM. The last model (model 2b) was used as the basis of the test and a biobehavioral reactivity unobserved variable was created with anxiety and depression loading as observed variables (model 2c; Figure 8). This addition did not substantially affect goodness-of-fit statistics; this test demonstrated acceptable agreement between the data and the model tested (2 = 11.309, p = .503, CFI = 1.000, RMSEA = .000). In addition, the magnitude of the pathway coefficient between biobehavioral reactivity and anxiety was large. Therefore, this model is judged to be the best-fitting model for the data, and is the final model of this test of the BBFM using romantic relationship satisfaction as the endogenous variable. In summary, for both Model 1 and Model 2, pathways between FEC and biobehavioral reactivity were significant, as were pathways between biobehavioral reactivity and disease activity. Pathways between FEC and disease activity were not significant. For the final Model 1, this means that, as family functioning became more problematic (GFS/FAD scores increased), depression and anxiety scores worsened (QIDS and OASIS scores increased). It also indicates that, as patients experienced more biobehavioral reactivity, their disease activity increased; in other words, the number of illness symptoms they reported (Health Review scores) increased and their physical well-being worsened (i.e., patients experienced more role limitations due to their physical health, or, RAND-36 Role Limitations Due to Physical Health subscale scores decreased). For the final Model 2, our results indicate that, as marital satisfaction increases,

33

depression and anxiety improve (QIDS and OASIS scores decreased). In addition, as patients experienced more biobehavioral reactivity, their disease activity increased. More specifically, the number of illness symptoms patients reported (Health Review scores) increased and their physical well-being worsened (i.e., patients experienced more role limitations due to their physical health and more pain, or, RAND-36 Role Limitations Due to Physical Health and RAND-36 Pain subscale scores decreased). Bootstrapping Analyses Model 1: Family functioning. The next goal was to confirm significant relationships found in our final Model 1 tested in AMOS (model 1d; Figure 5). Depression significantly mediated the effects of family functioning on disease activity, as measured by Health Review scores, with a 95% bias-corrected and accelerated CI [.13, 1.07]. Depression also significantly mediated the effects of family functioning on RAND-36 Role Limitations scores with a 95% bias-corrected and accelerated CI [-12.72, -2.16]. Multiple mediator models model 1. Because the final model tested in AMOS included anxiety as an additional mediator, both depression and anxiety were subsequently tested in a multiple mediator model. Calculating a multiple mediator model using bootstrapping highlights the effects of all mediators as a set and calculates the difference between total and direct effects (the total indirect effect through all mediators), as well as demonstrating which specific mediators significantly contribute to the indirect effect (Preacher & Hayes, 2008). Using family functioning as the independent variable, QIDS and OASIS scores were found to significantly mediate the effects of family functioning on Health Review scores as a set (CI [.02, 1.04]) but were not significant independently (i.e., both confidence intervals included zero), indicating both mediators contribute to the indirect effects in the model. This finding is substantiated through examination of the output delineating normal theory tests of indirect effects; the indirect effects of family functioning on Health Review scores through both mediators (depression and anxiety) nears significance (p = .06), but neither mediator produces a significant indirect effect on its own (QIDS, p = .27; OASIS, p = .27). Similar results were found when the RAND-36 Role Limitations subscale was used as the dependent variable. Total indirect effects were significant, with a 95% bias-corrected and accelerated CI [-14.05, -.17]; depression (CI [-11.35, .12]) and anxiety (CI [-9.13, .52]) were

34

nonsignificant independently. Normal theory tests of the indirect effects demonstrate that the total mediation effects near significance (p = .05), which contributes to our bootstrapping results. Model 2: Romantic relationship satisfaction. The next goal was to confirm significant relationships found in our final Model 2 tested in AMOS (model 2c; Figure 8). Depression did not significantly mediate the effects of romantic relationship satisfaction on disease activity, as measured by Health Review scores, with a 95% bias-corrected and accelerated CI [-.09, .00], although it neared significance. Depression significantly mediated the effects of romantic relationship satisfaction on RAND-36 Role Limitations scores with a 95% bias-corrected and accelerated CI [.09, 1.16]. Depression also significantly mediated the effects of romantic relationship satisfaction on RAND-36 Pain subscale scores (CI [.05, .71]). Multiple mediator models model 2. Because the final model tested in AMOS included anxiety as an additional mediator for Model 2, depression and anxiety were next tested in a multiple mediator model in the bootstrapping analyses. First, the total indirect effects of depression and anxiety as mediators were nonsignificant; as a set, depression and anxiety did not significantly mediate the effects of romantic relationship satisfaction on Health Review scores with a 95% bias-corrected and accelerated CI [-.11, .00]), although this CI neared significance. Neither mediator was significant independently, although anxiety produced a CI [-.14, .00], which neared significance; depression produced a CI [-.06, .04]. Similar results were found when the RAND-36 Role Limitations subscale was used as the dependent variable. Total indirect effects were nonsignificant, with a 95% bias-corrected and accelerated CI [-.01, 1.18]; depression (CI [-.02, 1.14]) and anxiety (CI [-.07, .92]) were also nonsignificant independently. Normal theory tests of the indirect effects demonstrate that the total mediation effects near significance (p = .05), which contributes to our bootstrapping results. We additionally tested the indirect effects of romantic relationship satisfaction on RAND-36 Pain subscale scores, through depression and anxiety. The total indirect effects were significant, with a 95% bias-corrected and accelerated CI [.03, .73]. Neither depression (CI [-.05, .76] nor anxiety (CI [-.11, .65]) produced significant results independently. As a whole, the bootstrapping analyses produce mixed results in comparison to our model testing completed in AMOS. Some of these differences can be attributed to the use of latent variables in AMOS; our bootstrapping analyses tested anxiety and depression as mediators for singular dependent variables, whereas AMOS tested the model as a whole, using multiple

35

physical health variables loading onto disease activity. These bootstrapping analyses suggest that Health Review scores may not be a strong piece of the overall model, and also indicate that anxiety may be contributing to an overall mediation effect but likely does not mediate the effects of FEC on disease activity on its own. This conclusion may have been predicted by our preliminary analyses, in that anxiety was not significantly correlated with GFS/FAD or QMI scores (Table 2), nor were our FEC variables significant predictors of OASIS scores in the stepwise regressions. Conclusion FEC, as tested separately in models using family functioning and romantic relationship satisfaction as endogenous variables, is significantly related to biobehavioral reactivity (depression and anxiety), which is in turn related to disease activity. As hypothesized, FEC is not directly related to disease activity. Although anxiety was added as an additional mediator in the models tested in AMOS without negatively affecting goodness-of-fit statistics, its use was not further substantiated through bootstrapping analyses. Although correlations and one stepwise regression finding suggested anxiety was significantly associated with measures of physical health, it is likely the lack of a relationship between FEC and anxiety that produces these mixed results.

36

CHAPTER 5 DISCUSSION This study was conducted to test the Biobehavioral Family Model with a sample of adult primary care patients in order to further understanding of the effects of close relationships on physical health. In addition, the goals of this study included using this model to integrate the effects of family emotional climate on both mental and physical health, to test precise pathways including indirect mediation effects, and to focus on both family and romantic relationships as predictors of health. Lastly, the purpose of the study was to address limitations in the families and health literature by using both self-report measures and specific, objective indicators of physical health, as well as testing the BBFMs proposed pathways in a diverse, underserved sample. Although the BBFM has a broad developmental application (Wood et al., 2000; Wood & Miller, 2005) it had not yet been tested with adult patients. In addition, although the BBFM is theoretically applicable to illnesses of either psychological or organic origin (Wood & Miller, 2002), the model had not been tested with chronic conditions critical to adult health. In testing the BBFMs ability to explain health outcomes for adult primary care patients, findings substantiated the three hypotheses outlined in Chapter 1 and the use of the BBFM with adult family members. Multiple implications for future research and clinical practice are highlighted by the results of this study. Summary of Hypotheses This study tested two versions of the BBFM with underserved adult primary care patients. To test family emotional climate (FEC), we used both a measure of family functioning to test the model with our entire sample (Model 1) and a measure of romantic relationship satisfaction with a subsample of patients who self-identified as being in a close romantic relationship (Model 2). For both Model 1 and Model 2, hypotheses included: (1) a direct pathway between FEC and biobehavioral reactivity, (2) a direct pathway between biobehavioral reactivity and disease activity, and (3) a nonsignificant pathway between FEC and disease activity (Figure 1). These relationships outline the mediation relationship of the BBFM: emotion dysregulation is the variable through which (problematic) family functioning affects physical health. Model 1: Family Functioning

37

All three hypotheses were supported for Model 1. Preliminary analyses, including tests of variable associations with Pearson correlations and stepwise regressions, indicated promise that the GFS/FAD was a significant predictor of biobehavioral reactivity (depression), and that both depression and anxiety were significant predictors of self-report measures of physical health (Health Review and RAND-36 subscales). Using a model generating approach (Byrne, 2010), nonsignificant pathways were trimmed in order to find the model that best fit the data and theory (Anderson et al., 2002; Kenny, 2011; Kline, 2011). The final model (Figure 5) demonstrated acceptable agreement between the data and the model tested and produced a significant relationship between FEC (family functioning) and biobehavioral reactivity (measured by both depression and anxiety), as well as a significant relationship between biobehavioral reactivity and disease activity (as measured by self-reported illness symptoms with the Health Review and role limitations with the RAND-36 Role Limitations Due to Physical Health subscale). In addition, a nonsignificant relationship was found between FEC and disease activity. These results indicate that the BBFM, using family functioning as the predictor variable, is able to explain physical health outcomes for adult primary care patients. Self-reports of family functioning, including the ability to turn to family members for support, family decision-making, and the ability to talk to family members about sadness, affects primary care patients mental health. In other words, the more problematic patients viewed their family functioning to be, the more likely they were to report anxiety and depression. Adults emotional dysregulation in turn predicted their disease activity: increases in the symptoms and severity of depression and anxiety resulted in worse physical health outcomes, including more self-perceived illness symptoms and role limitations. Role limitations included, for example, difficulty in performing work and other activities. These findings demonstrate that the BBFM is valuable in predicting the effects of family functioning on the health of adult primary care patients. Model 2: Romantic Relationship Satisfaction All three hypotheses were supported for Model 2. As with Model 1, preliminary analyses indicated that the QMI was a significant predictor of biobehavioral reactivity (depression), and that both depression and anxiety were significant predictors of self-report measures of physical health (Health Review and RAND-36 subscales). Significant pathways were found through testing Model 2 in AMOS (Figure 8), including a significant relationship between FEC (romantic relationship satisfaction) and biobehavioral reactivity (as measured by both depression and

38

anxiety), and a significant relationship between biobehavioral reactivity and disease activity (including self-reported illness symptoms measured with the Health Review, role limitations measured with the RAND-36 Role Limitations Due to Physical Health subscale, pain measured with the RAND-36 Pain subscale, and overall health measured with the RAND-36 General Health subscale). In addition, a nonsignificant relationship was found between FEC and disease activity. The findings for Model 2 confirm that the BBFM, using romantic relationship satisfaction as the endogenous variable, is able to explain physical health outcomes for adult primary care patients. Self-reports of romantic relationship satisfaction, including perceptions of having a stable or strong relationship, feeling happy with ones partner, and feeling like part of a team, affect primary care patients mental health. The more satisfied patients were in their romantic relationships, the less likely they were to report anxiety and depression. Adults emotional dysregulation in turn predicted their disease activity: increases in symptoms and severity of depression and anxiety resulted in worse physical health outcomes, including more self-perceived illness symptoms, role limitations, pain and worse overall health. Overall, these findings demonstrate that the BBFM is valuable in predicting the effects of romantic relationship processes on the health of adult primary care patients. Limitations and Future Research This study addressed several shortcomings of the current literature, including using a model that integrated the effects of both family and romantic relationships on mental and physical health. However, an additional goal of this study was to use both self-rated health measures and objective indicators of health collected from patient medical charts. Despite finding evidence supporting the use of the BBFM to explain the associations between family emotional climate, biobehavioral reactivity, and disease activity, preliminary analyses demonstrated a lack of significant associations between data collected from medical charts and other study variables. The number of primary care appointments in the past year, number of medical diagnoses, number of mental health diagnoses, BMI, and BMI weight status were all variables tested in the correlation and stepwise regression analyses. Unfortunately, none of these data were associated with FEC or biobehavioral reactivity variables. The only exception was an association found between the PC-Self item and the number of mental health diagnoses found in

39

patients medical charts. Because the perceived criticism measures were not associated with any other variables, they were excluded from further analyses. There are several possible reasons why these medical chart indicators may not have been significantly related to the other study measures. One is that, although it was a goal of this study to include primary care patients with a wide array of chronic conditions and to expand the use of the BBFM in this way, the proposed indicators of physical health may not have been specific enough to determine disease activity. For example, BMI was only significantly related to BMI weight status. BMI may not be a specific enough indicator to capture the physical health of patients with a large variety of medical conditions. A second possible reason for the lack of significant findings for the objective health measures may be medical chart error. There is evidence that outpatient electronic medical records may lack accuracy for a variety of reasons (Wagner & Hogan, 1996; Weeks, Corbett, & Stream, 2010), including multiple interactions between humans and technological systems that impact data entry (Azar et al., 2012). It may have been that the data collected from patient medical charts was incomplete and therefore not representative of patients actual prescription lists, diagnoses, or current height/weight. Previous tests of the BBFM with an asthmatic pediatric sample were able to find significant associations between child depression, anxiety, and disease severity through use of objective measures of pulmonary functioning conducted for the study by medical providers (e.g., Wood et al., 2007). Future tests of the BBFM with adult patients should continue to focus on specific indicators of health (Carr & Springer, 2010), although narrowing patient samples to those with specific illnesses or measuring health functioning directly may be more likely to yield significant findings. An additional limitation of the current literature includes infrequent tests of precise pathways linking family processes and health (Carr & Springer, 2010); use of the BBFM in this study is an attempt to ameliorate this. Although tests of the model and its pathways in AMOS yielded meaningful, significant findings, the attempt to confirm these findings using bootstrapping tests of mediation produced mixed results. There are several reasons this may have occurred, including differences in how bootstrapping mediation tests differ from path analyses in AMOS. The largest difference is that bootstrapping is able to test multiple mediator models, but is only suitable for testing one independent variable and one dependent variable (Preacher & Hayes, 2008). Therefore, because the models built in AMOS used multiple indicators of disease

40

activity, the multiple mediator models tested using bootstrapping analyses were required to differ. This ultimately produced some differing results. A second possibility for the variation in results includes that anxiety, included in the latent variable of biobehavioral reactivity in both Model 1 and Model 2, is likely not a perfect mediator. Although anxiety was significantly associated with physical health indicators in the correlation analyses, it was not consistently identified as a significant predictor of disease activity in the stepwise regressions. Therefore, although it was included in the final models constructed in AMOS because its inclusion did not adversely affect the model fit statistics, it produced significant pathways, and because it increased the meaning of the resulting model, it may have been highlighted as a weak mediator through the bootstrapping analyses. Therefore, future research should include further tests of anxiety as a measure of biobehavioral reactivity for adult family members. In addition, although this was an exploratory study providing the initial test of the BBFM with adult patients, alcohol abuse was thought to be a potential mediator of the association between family emotional climate and disease activity. Unfortunately, the two measures of alcohol use did not produce significant findings. It may be that alcohol use is not truly representative of biobehavioral reactivity and emotion dysregulation in this primary care population, despite repetitive evidence in the literature that emotion regulation is a primary causal factor of alcohol abuse (Berking et al., 2011). Nonsignificant findings may have also occurred due to participants discomfort in disclosing alcohol abuse. Primary care providers are often concerned that patients do not respond honestly to surveys of alcohol use (Beich, Gannik, Malterud, 2002) and patients willingness to volunteer risky health behaviors such as alcohol use is related to their comfort and their perception of the relevance of sharing the information to their health care (Gerbert et al., 1999). In regards to the present data, only 17% of patient participants fell at or above the AUDIT cutoff for concerning alcohol abuse. However, 32% admitted to problematic drinking on the NIAAA screen. This indicates that, when asked only one item, one in three participants disclosed having 5 or more alcoholic drinks during at least one occasion in the past 3 months. Therefore, it may not be an unwillingness to disclose alcohol use that resulted in nonsignificant findings. Interestingly, despite this studys findings indicating concerning levels of alcohol use in the present sample, only one patient had documented alcohol use problems in their medical chart.

41

Future research should not only test alcohol use as a measure of emotion dysregulation in using the BBFM with adult patients, but should also examine the impact of having behavioral health providers screen for health-risk behaviors that patients may otherwise be uncomfortable disclosing to their medical providers. Collaborative care and integrated mental health services provide a likely opportunity for screening patients, identifying adults with risky alcohol use, and immediately intervening to provide counseling that reduces alcohol use and improves health (Whitlock, Polen, Green, Orleans, & Klein, 2004). Along with nonsignificant findings for medical chart indicators of health and alcohol use, a measure of perceived criticism was included as a potential measure of FEC; neither the PCPartner nor the PC-Self item were significantly associated with measures of biobehavioral reactivity. Although the present study found support for all three hypotheses with both models and included a focus on both family and romantic relationships, the inclusion of a perceived criticism measure did not explain associations between FEC and health. Research has long substantiated perceived criticism as a predictor of mental health: responses to the question, How critical is your spouse of you? has been demonstrated to be the single strongest predictor of depression relapse for married patients (Hooley & Teasdale, 1989). One reason for the lack of significant associations between perceived criticism and biobehavioral reactivity in this sample may be the studys small sample size. Although the entire sample consisted of 125 primary care patients, only 76 were currently in a romantic relationship. This may also be reflective of the broad developmental range of the study sample, which ranged from 18- to 65-years-old: not all participants may have been pursuing close romantic relationships in their current life cycle stage (Carter & McGoldrick, 1999). A second possible cause of the nonsignificant findings for perceived criticism may be literacy concerns with the present sample. These two items produced the most difficulty for patients, who during review of their survey sometimes explained they believed critical meant important and had therefore answered on the higher end of the scale indicating the significance of their partner in their lives. It is likely that many of these mistakes occurred, were not caught by the author, and therefore affected the resulting data and analyses. It is critical that future research pilot test assessments within their community-based setting to better ascertain patient ability and comfort with the measures. Although several of the patients in this study had completed a college degree, greater than one in four had not graduated from high school and may

42

have had challenges reading or comprehending the items. The sample accessed for the present test of the BBFM is a strength of this study, however, further precautions should be taken in future research to ensure ease of study participation for primary care patients. Lastly, it is a limitation of this study that the data were collected at one point in time. Although the present discussion suggests family emotional climate affects disease activity through biobehavioral reactivity, it may be that physical health affects family functioning. The found associations may occur in the opposite direction suggested presently; theoretically, the BBFM would support reciprocal influences in both directions. Therefore, future research should test the application of the BBFM longitudinally. Ascertaining how family emotional climate has causal, long-term effects on mental and physical health has the potential to highlight how interventions can be best targeted for adult primary care patients. This could have far-reaching effects: health promotion and the prevention of chronic illness is critical in primary care (Pandhi et al., 2011) and integrating findings about patient family functioning into these settings may enhance the chance of flagging patients at later risk of depression, anxiety, and disease activity. In addition to these areas of future study suggested by the present investigations limitations, several other implications for future research are highlighted. An example is that future research using the BBFM should clarify how tests of the model advance our theoretical understanding of the underlying processes occurring for these individuals and their families. The author of the model underlines that no single study has tested the BBFM in its entirety (Wood, 2012), although a recent investigation tested all levels of the model (Lim, Wood, Miller, & Simmens, 2011). The complexity of the model, and its multiple potential applications to specific illness types, family functioning variables, and individual psychological variables (Wood, 2012) requires that future research continue to flesh out the processes underlying the models constructs. This is critical to improve our theoretical understanding of the conceptual mechanisms of the BBFM. A final area of future research suggested by the present study is a more thorough consideration of sample background factors and how they are associated with the models variables and pathways. More specifically, additional research is necessary to better understand how demographic characteristics (e.g., socioeconomic status, health insurance status, race/ethnicity, age, gender) of specific populations and individual patients affect the processes of the BBFM and found associations. The present sample was a multi-need, primarily uninsured

43

sample that presented, on average, as depressed with problematic family functioning. In addition, the patients who participated in this study reported on average almost 5 illness symptoms (20% reported 9 or more), had a 32 for their BMI, and had attended 5 primary care appointments in the past year. Findings with medically underserved patients using the BBFM may highlight additional needs of this population and would have several health care policy implications. Further understanding of how the BBFM explains health for diverse populations, especially medically underserved patients and their families, is critical. Implications for Clinical Practice The results of this study have several potential implications for clinical practice. The found relationship between FEC and disease activity, as mediated by biobehavioral reactivity, provides a prospective opportunity for clinical intervention and improved health outcomes. The BBFM is based in several theoretical assumptions of Minuchins structural family therapy and psychosomatic family model (Minuchin et al., 1978). This may mean that interventions targeted to adult primary care patients reporting problematic family functioning and related mental and physical health issues should include family therapy. Structural family therapy, specifically, would focus on assessing the patients familys interactions and restructuring family patterns in order to actively create changes in the family system (Goldberg & Goldberg, 2008). Although the statistical results from the present study are not causal and further research to determine the longitudinal nature of the effects of family on adult health is necessary, it may be that intervening therapeutically in patients family functioning benefits their mental and physical health. It has been repetitively demonstrated that negative family relationships have a stronger influence on health than positive relationships (Campbell, McDaniel, & Cole-Kelly, 2005). Examining the family dimension of physical health and offering therapy as necessary is considered critical to health care and symptom management (e.g., Lewandowski, Morris, Draucker, & Risko, 2007). It is especially important that therapeutic approaches be based in theory (Lewandoski et al., 2007). Structural family therapy is likely to provide benefits to primary care patients experiencing common conditions (e.g., pain) and can be tailored to the specific family patterns highlighted in patients completed assessments (Kerns & Otis, 2003). In fact, Minuchins work has served as the basis for several family treatments developed for improved patient health (e.g., Fiese, 2005; Loeb & le Grange, 2009). Demonstrating efficacy of family treatments for adult patients is indicated as a next step from the results of the present

44

study. There have been very few family or romantic relationship intervention studies for adults experiencing illness (Campbell, 2003), especially studies of family therapy. Suggesting family interventions for improved physical health outcomes is not new: Campbell (2003) suggests that understanding [biological, behavioral, and psychophysiological] pathways or mechanisms is helpful in developing and testing family interventions and choosing appropriate mediating variables and outcome measuresIntervention studies that target [these] pathways and measure appropriate mediating variables will help us to better understand the processes by which families influence health (p. 266). While there are several types of interventions that may be useful for adult primary care patients and their families, they all fall within medical family therapy (McDaniel, Hepworth, & Doherty, 1992) and can be categorized as family education and support, family psychoeducation, or family therapy; the latter two are almost always provided by family therapists (Campbell, 2003). A recent review of couple and family interventions for health problems suggests overall support for including family members in treatment. However, the studies reviewed were overwhelmingly randomized clinical trials of family support interventions and family psychoeducation; the only studies testing family therapy were focused for pediatric populations (Shields, Finley, Chawla, & Meadors, 2012). A purpose of this study was to test the ability of the BBFM to explain how family processes affect physical health for a diverse, underserved sample of adults. In addition, it was a goal to test the models hypothesized pathways with a sample of primary care patients presenting with a broad array of reported symptoms and documented medical conditions. The finding that both family functioning and romantic relationship satisfaction relate to disease activity through patient levels of depression and anxiety, and that these results were produced using the present sample, indicates that marriage and family therapists (MFTs) have an important, substantiated, and potentially unfulfilled role in treating primary care patients. MFTs have the ability to provide the type of care (i.e., family relational interventions) that is indicated by the BBFMs pathways (Wood et al., 2008) and should be included in health care teams in order to offer a critical systemic understanding of patient functioning (Campbell, 2003). MFTs specifically targeting issues highlighted through completion of the GFS/FAD (e.g., problematic family decision-making) and QMI (e.g., a couples ability to work together as a team) have the potential to thereby improve anxiety, depression, and physical health. Multilevel interventions, targeting variables tested in this study, may have the most potential to succeed

45

(e.g., Wood et al., 2008). MFTs also have the ability to focus on adaptive family processes and preventative interventions to buffer the effects of relational stress on patients health, as well as considering multiple definitions of family in the treatment process (Wood, 2005). The goal of medical family therapy is to enhance both patient agency and communion (McDaniel et al., 1992), while simultaneously using knowledge of systems theory to actively collaborate with primary care providers in caring for patients and their families. There are immense health concerns for primary care patients, especially less educated, low-income patients (Cwikel, Zilber, Feinson, & Lerner, 2007) served in urban primary care practices (Olfson et al., 2000). These patients tend to be sicker and account for a large portion of health care costs (Sutherland, Fisher, & Skinner, 2009). In the present sample alone, the average number of medical conditions was 2, while 20% of patient participants had 4 or more documented in their medical chart. In addition, few underserved, low-socioeconomic status patients are likely to seek out (Howard et al., 1996) or receive mental health care (Gallo et al., 1995; Olfson et al., 2000). Primary care physicians report difficulty in obtaining outpatient mental health services for their patients, made more difficult by severe access problems for lowincome uninsured patients (Cunningham, 2009). Recent changes in mental health parity legislature do not assist with the shortage of mental health providers, especially those willing or able to accept primary care patients with Medicaid coverage (Cunningham, 2009). Therefore, given the severity of health problems underserved primary care patients present with, identifying treatable issues that may worsen health, such as family functioning, romantic relationship satisfaction, depression, or anxiety, is of importance. Assessing patients level of risk for these factors is critical. Surveys, such as those administered in the present study, can be used routinely in primary care and can highlight avenues for further appraisal and intervention. MFTs have the unique ability to systemically intervene, using family and relational strategies, with primary care patients and their family members. Multilevel interventions that target relational distress, emotion dysregulation, and symptom management are critical. MFTs understanding of systems theory also contributes to their ability to effectively integrate into primary care practices and to recognize the importance of practicing collaboration with medical providers to improve patient care and health outcomes. Conclusion

46

There is much evidence linking family and romantic relationships to health outcomes as researchers increasingly recognize the importance of these associations (Carr & Springer, 2010). Despite this increased focus, tests of the connections between social functioning and physical health continue to use White, middle-class, nuclear families as the basis of study, with a lack of attention on precise pathways and specific indicators of health (Carr & Springer, 2010; Wood, 2005). The Biobehavioral Family Model is a biopsychosocial approach (Engel, 1977) to health that has successfully integrated family functioning, psychological factors, and physical health outcomes into one, comprehensive model (Wood, 1993). The model has been tested most often with asthmatic pediatric samples. This study addresses limitations in the literature while testing the BBFM with adult primary care patients and further elucidates the paths through which family emotional climate affects disease activity. Through the use of model-testing and mediation analyses, the results of this study indicate that both family functioning and romantic relationship satisfaction affect physical health through patient experiences of depression and anxiety. Additional research is necessary to test objective indicators of physical health and whether the BBFMs pathways are significant longitudinally. Interventions can be tailored to fit the relational needs of individual patients in primary care and may be especially effective when delivered by a collaborative marriage and family therapist.

47

RAND-36 Physical Functioning Subscale

RAND-36 Role Limitations Subscale

RAND-36 General Health Subscale

Health Review Score .57***

-.70*** -.83*** -.65*** RAND-36 Pain Subscale Disease Activity -.68***

GFS/FAD

.11

.28** QIDS Scale Score

.48***

2 (13) = 33.931, p = .001, RMSEA = .114, CFI = .910

Figure 3. The initial model built in AMOS to test Model 1, with family functioning (GFS/FAD scores) as the endogenous variable, depression (QIDS) as the mediator, and disease activity as a latent outcome variable (n = 125) *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.

48

RAND-36 Role Limitations Subscale

Health Review Score

-.66**

.68***

GFS/FAD .17

Disease Activity

.28** .54*** QIDS Scale Score

2 (1) = .324, p = .569, RMSEA = .000, CFI = 1.000

Figure 4. Model 1c (n = 125), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.

49

RAND-36 Role Limitations Subscale

Health Review Score

-.66**

.68***

GFS/FAD .17

Disease Activity

.28** .54*** Biobehavioral Reactivity

OASIS Scale Score

.81***

.86***

QIDS Scale Score

2 (1) = 4.135, p = .247, RMSEA = .055, CFI = .992

Figure 5. Model 1d, with biobehavioral reactivity as a mediator composed of anxiety (OASIS Scale Score) and depression (QIDS Scale Score) (n = 125), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.

50

RAND-36 Physical Functioning Subscale

RAND-36 Role Limitations Subscale

RAND-36 General Health Subscale

Health Review Score .55***

-.68*** -.81*** -.73*** RAND-36 Pain Subscale Disease Activity -.65***

QMI

.11

-.31** QIDS Scale Score

.54***

2 (13) = 21.232, p = .068, RMSEA = .092, CFI = .938

Figure 6. The initial model built in AMOS to test Model 2, with romantic relationship satisfaction (QMI scores) as the endogenous variable, depression (QIDS) as the mediator, and disease activity as a latent outcome variable (n = 76) *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.

51

RAND-36 Role Limitations Subscale

RAND-36 General Health Subscale

Health Review Score .60***

-.76*** -.73*** RAND-36 Pain Subscale Disease Activity .10

-.65***

QMI

-.31**

.61*** QIDS Scale Score

2 (8) = 6.126, p = .633, RMSEA = .000, CFI = 1.000

Figure 7. Model 2b (n = 76), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.

52

RAND-36 General Health Subscale RAND-36 Pain Subscale -.64***

RAND-36 Role Limitations Subscale

Health Review Score

-.76*** -.73*** .61***

QMI .15

Disease Activity

-.34 ** .71*** Biobehavioral Reactivity

OASIS Scale Score

.87***

.88***

QIDS Scale Score

2 (12) = 11.309, p = .503, RMSEA = .000, CFI = 1.000

Figure 8. Model 2c, with biobehavioral reactivity as a mediator composed of anxiety (OASIS Scale Score) and depression (QIDS Scale Score) (n = 76), *p<.05, **p<.01, ***p<.001. Significant paths are indicated in bold.

53

APPENDIX A
Office of the Vice President For Research Human Subjects Committee Tallahassee, Florida 32306-2742 (850) 644-8673, FAX (850) 644-4392 APPROVAL MEMORANDUM Date: 5/9/2011 To: Wayne Denton [wdenton@fsu.edu] Address: 225 Sandels Building Tallahassee, FL 32306-1491 Dept.: FAMILY & CHILD SCIENCE From: Thomas L. Jacobson, Chair Re: Use of Human Subjects in Research Assessing Need for Mental Health and Addiction Treatment Services in a Primary Care Medical Clinic The application that you submitted to this office in regard to the use of human subjects in the research proposal referenced above has been reviewed by the Human Subjects Committee at its meeting on 02/09/2011. Your project was approved by the Committee. The Human Subjects Committee has not evaluated your proposal for scientific merit, except to weigh the risk to the human participants and the aspects of the proposal related to potential risk and benefit. This approval does not replace any departmental or other approvals, which may be required. If you submitted a proposed consent form with your application, the approved stamped consent form is attached to this approval notice. Only the stamped version of the consent form may be used in recruiting research subjects. If the project has not been completed by 2/8/2012 you must request a renewal of approval for continuation of the project. As a courtesy, a renewal notice will be sent to you prior to your expiration date; however, it is your responsibility as the Principal Investigator to timely request renewal of your approval from the Committee. You are advised that any change in protocol for this project must be reviewed and approved by the Committee prior to implementation of the proposed change in the protocol. A protocol change/amendment form is required to be submitted for approval by the Committee. In addition, federal regulations require that the Principal Investigator promptly report, in writing any unanticipated problems or adverse events involving risks to research subjects or others. By copy of this memorandum, the Chair of your department and/or your major professor is reminded that he/she is responsible for being informed concerning research projects involving human subjects in the department, and should review protocols as often as needed to insure that the project is being conducted in compliance with our institution and with DHHS regulations. This institution has an Assurance on file with the Office for Human Research Protection. The Assurance Number is FWA00000168/IRB number IRB00000446. Cc: Kay Pasley, Chair, HSC No. 2011.5568

54

APPENDIX B

55

56

57

58

APPENDIX C
Q-1. Are you currently married or in a romantic relationship? Please circle one: YES NO

If you are NOT currently married or in a romantic relationship, SKIP TO PAGE 2!

If you answered YES, please continue:

Q-2. Please rate how much you agree or disagree with the following statements. (Circle the number of your answer)
VERY STRONG DISAGREEMENT a. We have a good marriage (relationship)..1 b. My relationship with my partner is very stable....1 c. Our marriage (relationship) is strong....1 d. My relationship with my partner makes me happy..........................1 e. I really feel like part of a team with my partner...1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5 VERY STRONG AGREEMENT 6 6 6 6 6 7 7 7 7 7

f. One the scale below, indicate the point which best describes the degree of happiness, everything considered, in your marriage (relationship). The middle point, happy, represents the degree of happiness which most people get from marriage (relationships). The scale gradually increases on the right side for those few who experience extreme joy in marriage (a relationship) and decreases on the left side for those who are extremely unhappy. VERY UNHAPPY 1 2 HAPPY 5 PERFECTLY HAPPY 10

Q-3. Please choose the response that best represents your feelings. (Circle the number of your answer) NOT AT ALL CRITICAL a. How critical is your partner of you?.1 2 b. How critical are you of your partner... 1 2 VERY CRITICAL 9 10 9 10

3 3

4 4

5 5

6 6

7 7

8 8

59

PAGE 2
Q-4. On any single occasion during the past 3 months, have you had more than 5 drinks containing alcohol? Please circle your answer: YES NO

Q-5. Please checkmark the answer that is correct for you: a. How often do you have a drink containing alcohol?
NEVER MONTHLY OR LESS TWO TO FOUR TIMES A MONTH TWO TO THREE TIMES PER WEEK FOUR OR MORE TIMES PER WEEK b. How many drinks containing alcohol do you have on a typical day when you are drinking? 1 or 2 3 or 4 5 or 6 7 or 9 10 or more None I do not consume alcohol

Q-6. Please circle the answer that is correct for you:

LESS THAN MONTHLY 2 2 TO 3 TIMES PER WEEK 4 4 OR MORE TIMES PER WEEK 5

NEVER a. How often do you have six or more drinks on one occasion?.................................................................1 b. How often during the last year have you found that you were not able to stop drinking once you had started?......................................................................... 1 c. How often during the last year have you needed a first drink in the morning to get yourself going after a heavy drinking session?.................................................................. 1 d. How often during the last year have you had a feeling of guilt or remorse after drinking?......................................... 1 e. How often during the last year have you been unable to remember what happened the night before because you had been drinking?................................................................. 1

MONTHLY 3

60

Q-7. Please circle the answer that is correct for you:

NO a. Have you or someone else been injured as a result of your drinking?......................................................................................1 b. Has a relative or friend, or a doctor or other health worker, been concerned about your drinking or suggested you cut down?.................................................................................................1 YES, BUT NOT IN THE LAST YEAR 2 YES, DURING THE LAST YEAR 3

Q-8. This part of the questionnaire contains a number of statements about families. Please read each statement carefully and decide how well it describes your family. You should answer according to how you see your family. For each statement there are four possible responses. (Circle the number of your answer)
STRONGLY AGREE a. Planning family activities is difficult because we misunderstand each other..1 b. In times of crisis we can turn to each other for support,.1 c. We cannot talk to each other about the sadness we feel1 d. Individuals are accepted for who they are..1 e. We avoid discussing our fears and concerns1 f. We can express feelings to each other1 g. There are lots of bad feelings in the family....1 h. We feel accepted for who we are1 i. Making decisions is a problem for our family..1 j. We are able to make decisions about how to solve problems.1 k. We dont get along well together.....1 l. We confide in each other1 AGREE 2 2 2 2 2 2 2 2 2 2 2 2 DISAGREE 3 3 3 3 3 3 3 3 3 3 3 3 STRONGLY DISAGREE 4 4 4 4 4 4 4 4 4 4 4 4

Q-9. The following items ask about anxiety and fear. For each item, CHECKMARK the answer that best describes your experience over the PAST WEEK.
a. In the past week, how often have you felt anxious? 0 = NO ANIXETY in the past week 1 = INFREQUENT ANXIETY. Felt anxious a few times. 2 = OCCASSIONAL ANXIETY. Felt anxious as much of the time as not. It was hard to relax. 3 = FREQUENT ANXIETY. Felt anxious most of the time. It was very difficult to relax. 4 = CONSTANT ANXIETY. Felt anxious all of the time and never really relaxed.

61

b. In the past week, when you have felt anxious, how intense or severe was your anxiety? 0 = LITTLE OR NONE: Anxiety was absent or barely noticeable. 1 = MILD: Anxiety was at a low level. It was possible to relax when I tried. Physical symptoms were only slightly uncomfortable. 2 = MODERATE: Anxiety was distressing at times. It was hard to relax or concentrate, but I could do it if I tried. Physical symptoms were uncomfortable. 3 = SEVERE: Anxiety was intense much of the time. It was very difficult to relax or focus on anything else. Physical symptoms were extremely uncomfortable. 4 = EXTREME: Anxiety was overwhelming. It was impossible to relax at all. Physical symptoms were unbearable.

c. In the past week, how often did you avoid situations, places, objects, or activities because of anxiety or fear? 0 = NONE: I do not avoid places, situations, activities, or things because of fear 1 = INFREQUENT: I avoid something once in a while, but will usually face the situation or confront the object. My lifestyle is not affected. 2 = OCCASIONAL: I have some fear of certain situations, places, or objects, but it is still manageable. My lifestyle has only changed in minor ways. I always or almost always avoid the things I fear when Im alone, but can handle them in someone comes with me. 3 = FREQUENT: I have considerable fear and really try to avoid the things that frighten me. I have made significant changes in my lifestyle to avoid the object, situation, activity, or place. 4 = ALL THE TIME: Avoiding objects, situations, activities, or places has taken over my life. My lifestyle has been extensively affected and I no longer do things that I used to enjoy.

d. In the past week, how much did your anxiety interfere with your ability to do the things you needed to do at work, at school, or at home? 0 = NONE: No interference at work/home/school from anxiety. 1 = MILD: My anxiety has caused some interference at work/home/school. Things are more difficult, but everything that needs to be done is still getting done. 2 = MODERATE: My anxiety definitely intereferes with tasks. Most things are still getting done, but few things are being done as well as in the past. 3 = SEVERE: My anxiety has really changed my ability to get things done. Some tasks are still being done, but many things are not. My performance has definitely suffered. 4 = EXTREME: My anxiety has become incapacitating. I am unable to complete tasks and have had to leave school, have quit or been fired from my job, or have been unable to complete tasks at home and have faced consequences like bill collectors, eviction, etc.

e. In the past week, how much has anxiety interfered with your social life and relationships? 0 = NONE: My anxiety doesnt affect my relationships. 1 = MILD: My anxiety slightly interferes with my relationships. Some of my friendships and other relationships have suffered, but, overall, my social life is still fulfilling. 2 = MODERATE: I have experienced some interference with my social life, but I still have a few close relationships. I dont spend as much time with others as in the past, but I still socialize sometimes. 3 = SEVERE: My friendships and other relationships have suffered a lot because of anxiety. I do not enjoy social activities. I socialize very little. 4 = EXTREME: My anxiety has completely disrupted my social activities. All of my relationships have suffered or ended. My life is extremely strained.

62

Q-10. Please CHECKMARK the one response to each item that is the most appropriate to how you have been feeling over the PAST 7 DAYS.
a. Falling asleep: 0 = I never took longer than 30 minutes to fall asleep. 1 = I took at least 30 minutes to fall asleep, less than half the time (3 days or less out of the past 7 days). 2 = I took at least 30 minutes to fall asleep, more than half the time (4 days or more out of the past 7 days). 3 = I took more than 60 minutes to fall asleep, more than half the time (4 days or more out of the past 7 days). b. Sleep during the night: 0 = I didnt wake up at night. 1 = I had a restless, light sleep, briefly waking up a few times each night. 2 = I woke up at least once a night, but I got back to sleep easily. 3 = I woke up more than once a night and stayed awake for 20 minutes or more, more than half the time (4 days or more out of the past 7 days). c. Waking up too early: 0 = Most of the time, I woke up no more than 30 minutes before my scheduled time. 1 = More than half the time (4 days or more out of the past 7 days), I woke up more than 30 minutes before my scheduled time. 2 = I almost always woke up at least one hour or so before my scheduled time, but I got back to sleep eventually. 3 = I woke up at least one hour before my scheduled time, and couldnt get back to sleep. d. Sleeping too much: 0 = I slept no longer than 7-8 hours/night, without napping during the day. 1 = I slept no longer than 10 hours in a 24-hour period including naps. 2 = I slept no longer than 12 hours in a 24-hour period including naps. 3 = I slept longer than 12 hours in a 24-hour period including naps. e. Feeling sad: 0 = I do not feel sad. 1 = I feel sad less than half the time. 2 = I feel sad more than half the time. 3 = I feel sad nearly all the time. PLEASE CHECKMARK THE ONE RESPONSE TO EACH ITEM THAT IS MOST APPROPRIATE TO HOW YOU HAVE BEEN FEELING OVER THE PAST 7 DAYS. Please complete either 11 or 12 (not both) Q-11. Decreased appetite: 0 = There was no change in my usual appetite. 1 = I ate somewhat less often or smaller amounts of food than usual. 2 = I ate much less than usual and only by forcing myself to eat. 3 = I rarely ate within a 24-hour period, and only by really forcing myself to eat or when others persuaded me to eat. Q-12. Increased appetite: 0 = There was no change in my usual appetite. 1 = I felt a need to eat more frequently than usual. 2 = I regularly ate more often and/or greater amounts of food than usual. 3 = I felt driven to overeat both at mealtime and between meals.

63

Please complete either 13 or 14 (not both) Q-13. Decreased weight (within the last 14 days): 0 = My weight has not changed 1 = I feel as if Ive had a slight weight loss. 2 = Ive lost 2 pounds or more. 3 = Ive lost 5 pounds or more. Q-14. Increased weight (within the last 14 days): 0 = My weight has not changed. 1 = I feel as if Ive had a slight weight gain. 2 = Ive gained 2 pounds or more. 3 = Ive gained 5 pounds or more.

Q-15. Please CHECKMARK the one response to each item that is the most appropriate to how you have been feeling over the PAST 7 DAYS.
a. Concentration/decision-making: 0 = There was no change in my usual ability to concentrate or make decisions. 1 = I occasionally felt indecisive or found that my attention wandered. 2 = Most of the time, I found it hard to focus or to make decisions. 3 = I couldnt concentrate well enough to read or I couldnt make even minor decisions.

b. Perception of myself: 0 = I saw myself as equally worthwhile and deserving as other people. 1 = I put the blame on myself more than usual. 2 = For the most part, I believed that I caused problems for others. 3 = I thought almost constantly about major and minor defects in myself.

c. Thoughts of my own death or suicide: 0 = I didnt think of suicide or death. 1 = I felt that life was empty or wondered if it was worth living. 2 = I thought of suicide or death several times for several minutes over the past 7 days. 3 = I thought of suicide or death several times a day in some detail, or I made specific plans for suicide or actually tried to take my life.

d. General interest: 0 = There was no change from usual in how interested I was in other people or activities. 1 = I noticed that I was less interested in other people or activities. 2 = I found I had interest in only one or two of the activities I used to do. 3 = I had virtually no interest in the activities I used to do. e. Energy level: 0 = There was no change in my usual level of energy. 1 = I got tired more easily than usual. 2 = I had to make a big effort to start or finish my usual daily activities (for example: shopping, homework, cooking, or going to work). 3 = I really couldnt carry out most of my usual daily activities because I just didnt have the energy.

64

f. Feeling more sluggish than usual: 0 = I thought, spoke, and moved at my usual pace. 1 = I found that my thinking was more sluggish than usual or my voice sounded dull or flat. 2 = It took me several seconds to respond to most questions and I was sure my thinking was more sluggish than usual. 3 = I was often unable to respond to questions without forcing myself.

g. Feeling restless (agitated, not relaxed, fidgety): 0 = I didnt feel restless. 1 = I was often fidgety, wringing my hands, or needed to change my sitting position. 2 = I had sudden urges to move about and was quite restless. 3 = At times, I was unable to stay seated and needed to pace around.

Q-16. In the PAST MONTH, have you experienced any of the following symptoms? (Circle the number of your answer)
NO a.Temperature of 100 degrees F (37.7C) or more...1 b. Headache lasting more than 1 hour...1 c. Skin rash or hives..1 d. Painful, irritated, or burning eyes1 e. Ear ache or ear infection..1 f. Toothache1 g. Sore throat.1 h. Sneezing, stuffy, or runny nose..1 i. Dry cough....1 j. Coughing up substances other than saliva, or thin phlegm.1 k. Wheezing (from chest).... 1 l. Unusual shortness of breath.....1 m. Unplanned weight loss.1 n. Nausea and/or vomiting....1 o. Stomach pain or abdominal cramps1 p. Heartburn.1 q. Chest pain other than heartburn..1 r. Rapid or pounding heart.1 s. Diarrhea1 t. Bloody or black stools..1 u. Discomfort from hemorrhoids1 YES 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

65

Q-17. The following items are about activities you might do during a typical day. Does your health now limit you in these activities? If so, how much? (Circle the number of your answer)
YES, LIMITED A LOT a. Vigrouous activities, such as running, lifting heavy objects, participating in strenuous sports..............................................................1 b. Moderate activities, such as moving a table, pushing a vacuum cleaner, bowling, or playing golf.1 c. Lifting or carrying groceries.1 d. Climbing several flights of stairs.1 e. Climbing one flight of stairs.1 f. Bending, kneeling, or stooping.1 g. Walking more than a mile.1 h. Walking several blocks..1 i. Walking one block...1 j. Bathing or dressing myself........1 YES, LIMITED A LITTLE 2 NO, NOT LIMITED AT ALL 3

2 2 2 2 2 2 2 2 2

3 3 3 3 3 3 3 3 3

Q-18. During the PAST 4 WEEKS, have you had any of the following problems with your work or other regular daily activities as a result of your physical health? (Circle the number of your answer)
YES a. Cut down the amount of time you spent on work or other activities...1 b. Accomplished less than you would like...1 c. Were limited in the kind of work or other activities.1 d. Had difficulty performing the work or other activities (for example, it took extra effort)..1 NO 2 2 2

Q-19. Please CHECKMARK one response to each item.

a. How much bodily pain have you had during the past 4 weeks? None Very mild Mild Moderate Severe Very severe b. During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)? Not at all A little bit Moderately Quite a bit Extremely

66

c. In general, would you say your health is: Excellent Very good Good Fair Poor

Q-20. How TRUE or FALSE is each of the following statements for you. (Circle the number of your answer)
DEFINITELY TRUE a. I seem to get sick a little easier than other people..................................................................1 b. I am as healthy as anybody I know..1 c. I expect my health to get worse.1 d. My health is excellent..1 MOSTLY TRUE 2 2 2 2 DONT KNOW 3 3 3 3 MOSTLY FALSE 4 4 4 4 DEFINITELY FALSE 5 5 5 5

Q-21. As a child or adolescent, did you ever live away from your biological parent? Please circle your answer: YES NO
a. If YES, who did you live with? Checkmark all that apply: Foster parent Grandparent Other relative Friend/Non-relative

Q-22. Are you currently providing care as a traditional, kin, or other type of foster caregiver? Please circle your answer: YES NO
a. If YES, which type of foster care are you providing? Checkmark all that apply: Traditional Kin Other

Q-23. Would you be interested in meeting with a couple counselor as part of a future research project?
Yes No Not applicable I am not currently in a relationship

Q-24. Do you believe your partner would be interested in meeting with a couple counselor as part of a future research project?
Yes No Not applicable I am not currently in a relationship

67

This section asks questions that describe some general characteristics about you. This information helps us understand general characteristics of the people who have completed the survey. Please CHECKMARK one response to each item.

1. Birth Date:___/___/____

2. Sex:

Female

Male

3. Number of times you have been married: 0 2 4 1 3 5 or more

4. Number of times you have been divorced: 0 2 4 1 3 5 or more

5. Number of times you have been widowed: 0 2 4 1 3 5 or more

6. Please indicate your current MARITAL STATUS: Married and living together Divorced Never married Married but separated (not currently living together) Widowed

7. If you are NOT MARRIED but are CURRENTLY IN A RELATIONSHIP, please indicate which description best describes your relationship: Casually dating (I date other people as well) Living with my partner less than one year NOT APPLICABLE I am married Dating one person but do not live together Living with my partner more than one year NOT APPLICABLE I am NOT CURRENTLY in a relationship

8. What was your approximate household income last year? Under $10,000 $20,000 - $39,999 $60,000 - $79,999 $10,000 - $19,999 $40,000 - $59,999 $80,000 or above

68

9. What is the HIGHEST level of education you have completed? Did not graduate high school and did not obtain GED High school graduate or obtained GED Some college no degree Associates degree Bachelors degree Masters degree Professional degree Doctoral degree

10. Are you currently employed? Yes, full-time Yes, part-time No

11. What is your ethnicity? Hispanic or Latino Not Hispanic or Latino 12. What is your race? (You may check MORE than one) American Indian or Alaska Native Asian Black or African-American Native Hawaiian or Other Pacific Islander White

Thank you for completing our survey. Please let a research assistant know you are finished and we will be right with you!

69

REFERENCES Acock, A. C. (2005). Working with missing values. Journal of Marriage and Family, 67, 10121028. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders fourth edition text revision. Washington DC: American Psychiatric Association. Agency for Healthcare Research and Quality. (2002). Health care costs (AHRQ Publication No. 02-P033). Retrieved from http://www.ahrq.gov/news/costsfact.pdf Allen, J. P., Litten, R. Z., Fertig, J. B., & Babor, T. (1997). A review of research on the Alcohol Use Disorders Identification Test (AUDIT). Alcoholism: Clinical and Experimental Research, 21, 613-619. Anderson, M. I., Parmenter, T. R., & Mok, M. (2002). The relationship between neurobehavioural problems of severe traumatic brain injury (TBI), family functioning and the psychological well-being of the spouse/caregiver: Path model analysis. Brain Injury, 16, 743-757. Ansseau, M., Dierick, M., Buntinkx, F., Cnockaert, P., De Smedt, J., Van Den Haute, M., & Vander Mijnsbrugge, D. (2004). High prevalence of mental disorders in primary care. Journal of Affective Disorders, 78, 49-55. Arbuckle, J. L. (1997). AMOS users guide (version 4.0). Chicago, IL: SmallWaters. Azar, K. M., Moreno, M. R., Wong, E. C., Shin, J. J., Soto, C., & Palaniappan, L. P. (2012). Accuracy of data entry of patient race/ethnicity/ancestry and preferred spoken language in an ambulatory care setting. Health Research and Education Trust, 47, 228-240. Balcazar, H., Krull, J. L., & Peterson, G. (2001). Acculturation and family functioning are related to health risks among pregnant Mexican American women. Behavioral Medicine, 27, 62-70. Beich, A., Gannik, D., & Malterud, K. (2002). Screening and brief intervention for excessive alcohol use: Qualitative interview study of the experiences of general practitioners. BMJ, 325, 870-874. Berking, M., Margraf, M., Ebert, D., Wupperman, P., Hofmann, S. G., & Junghanns, K. (2011). Deficits in emotion-regulation skills predict alcohol use during and after cognitivebehavioral therapy for alcohol dependence. Journal of Consulting and Clinical Psychology, 79, 307-318. Bikos, L. H., & Kocheleva, J. (2012). Life role salience dimensions and mental health outcomes among female expatriate spouses in Turkey. Journal of Career Development. Advance online publication. doi: 10.1177/0894845311427906

70

Burge, S. K., Schneider, D., Ivy, L., Catala, S. (2005). Patients advice to physicians about intervening in family conflict. Annals of Family Medicine, 3, 248-254. Burman, B., & Margolin, G. (1992). Analysis of the association between marital relationships and health problems: An interactional perspective. Psychological Bulletin, 112, 39-63. Byles, J., Byrne, C., Boyle, M. H., & Oxford, O. R. (1988). Ontario Child Health Study: Reliability and validity of the General Functioning Scale of the McMaster Family Assessment Device. Family Process, 27, 97-104. Byrne, B. M. (2010). Structural equation modeling with AMOS: Basic concepts, applications, and programming (2nd ed.). New York, NY: Routledge. Campbell, T. L. (2003). The effectiveness of family interventions for physical disorders. Journal of Marital and Family Therapy, 29, 263-281. Campbell, T., McDaniel, S., & Cole-Kelly, K. (2005). Family issues in health care. In R. B. Taylor (Ed.), Taylors diagnostic and therapeutic challenges: A handbook (pp. 16-33). New York, NY: Springer. Campbell-Sills, L., Norman, S. B., Craske, M. G., Sullivan, G., Lang, A. J., Stein, M. B. (2009). Validation of a brief measure of anxiety-related severity and impairment: The Overall Anxiety Severity and Impairment Scale (OASIS). Journal of Affective Disorders, 112, 92-101. Carr, D., & Springer, K. W. (2010). Advances in families and health research in the 21st century. Journal of Marriage and Family, 72, 743-761. Carter, E. A., & McGoldrick, M. (1999). The expanded family life cycle: Individual, family, and social perspectives (3rd ed.). Boston, MA: Allyn and Bacon. Centers for Disease Control and Prevention (2009). International Classification of Diseases, Ninth Revision (ICD-9). Retrieved from http://www.cdc.gov/nchs/icd/icd9.htm Centers for Disease Control and Prevention (2011a). Healthy weight its not a diet, its a lifestyle! Retrieved from http://www.cdc.gov/healthyweight/assessing/bmi/ Centers for Disease Control and Prevention (2011b). Well-being concepts. Retrieved from http://www.cdc.gov/hrqol/wellbeing.htm Chambless, D. L., & Steketee, G. (1999). Expressed emotion and behavior therapy outcome: A prospective study with obsessive-compulsive and agoraphobic outpatients. Journal of Consulting and Clinical Psychology, 67, 658-665. Cohen, S., & Rodriguez, M. S. (1995). Pathways linking affective disturbances and physical disorders. Retrieved from Carnegie Mellon University website: http://repository.cmu.edu/

71

psychology/266 Cwikel, J., Zilber, N., Feinson, M., & Lerner, Y. (2007). Prevalence and risk factors of threshold and sub-threshold psychiatric disorders in primary care. Social Psychiatry and Psychiatric Epidemiology, 43, 184-191. Cunningham, P. J. (2009). Beyond parity: Primary care physicians perspectives on access to mental health care. Health Affairs, 28, w490-w501. DeGarmo, D. S., & Martinez Jr., C. R. (2006). A culturally informed model of academic wellbeing for Latino youth: The importance of discriminatory experiences and social support. Family Relations, 55, 267-278. Driver, J. L., & Gottman, J. M. (2004). Daily marital interactions and positive affect during marital conflict among newlywed couples. Family Process, 43, 301-314. Efron, B. (1987). Better bootstrap confidence intervals. Journal of the American Statistical Association, 82, 171-185. Egede, L. E., Zheng, D., & Simpson, K. (2002). Comorbid depression is associated with increased health care use and expenditures in individuals with diabetes. Diabetes Care, 25, 464-470. Engel, G. L. (1977). The need for a new medical model: A challenge for biomedicine. Science, 196, 129-136. Epidemiology Data Center (2011). Welcome to REACH II. Retrieved from http://www.edc.gsph.pitt.edu/reach2/ Epstein, N. B., Baldwin, L. M., & Bishop, D. S. (1983). The McMaster Family Assessment Device. Journal of Marital and Family Therapy, 9, 171-180. Fals-Stewart, W., OFarrell, T. J., & Hooley, J. M. (2001). Relapse among married or cohabiting substance-abusing patients: The role of perceived criticism. Behavior Therapy, 32, 787801. Fiese, B. H. (2005). Introduction to the special issue: Time for family-based interventions in pediatric psychology? Journal of Pediatric Psychology, 30, 629-630. Fiscella, K., Franks, P., Shields, C. G. (1997). Perceived family criticism and primary care utilization: Psychosocial and biomedical pathways. Family Process, 36, 25-41. Franks, P., Campbell, T. L., & Shields, C. G. (1992). Social relationships and health: The relative roles of family functioning and social support. Social Science and Medicine, 34, 779-788. Franz, C. E. (1997). Stability and change in the transition to midlife: A longitudinal study of

72

midlife adults. In M. E. Lachman & J. B. James (Eds.), Multiple paths of midlife development (pp. 45-66). Chicago, IL: The University of Chicago Press. Friedman, B. H., & Thayer, J. F. (1998). Autonomic balance revisited: Panic anxiety and heart rate variability. Journal of Psychosomatic Research, 44, 133-151. Gallagher, D., Ting, L., & Palmer, A. (2008). A journey into the unknown; taking the fear out of structural equation modeling with AMOS for the first-time user. The Marketing Review, 8, 255-275. Gallo, J. J., Marino, S., Ford, D., & Anthony, J. C. (1995). Filters on the pathway to mental health care, II: Sociodemographic factors. Psychological Medicine, 25, 1149-1160. Gardner, J., & Oswald, A. (2004). How is mortality affected by money, marriage, and stress? Journal of Health Economics, 23, 1181-1207. Gerbert, B., Bronstone, A., Pantilat, S., McPhee, S., Allerton, M., & Moe, J. (1999). When asked, patients tell: Disclosure of sensitive health-risk behaviors. Medical Care, 37, 104111. Gilliss, C. L., Highly, B. L., Roberts, B. M., & Martinson, I. M. (1989). Toward a science of family nursing. Menlo Park, CA: Addison-Wesley. Goetzel, R. Z., Hawkins, J., Ozminkowski, R. J., & Wang, S. (2003). The health and productivity cost burden of the top 10 physical and mental health conditions affecting six large U.S. employers in 1999. Journal of Occupational and Environmental Medicine, 45, 5-14. Goldberg, H., & Goldberg, I. (2008). The structural model. In (H. Goldberg & I. Goldberg (Eds.), Family therapy: An overview, Seventh edition. Belmont, CA: Thompson Higher Education. Goldman, N. (2001). Social inequalities in health: Disentangling the underlying mechanisms. Annals of the New York Academy of Sciences, 954, 118-139. Grzywacz, J. G., & Marks, N. F. (2000). Family, work, work-family spillover, and problem drinking during midlife. Journal of Marriage and Family, 62, 336-348. Hartwig, F., & Dearing, B. E. (1979). Exploratory data analysis. Thousand Oaks, CA: Sage Publications, Inc. Hawkins, D. & Booth, A. (2005). Unhappily ever after: Effects of long-term, low-quality marriages on well-being. Social Forces, 84, 445-465. Hays, R. D., & Morales, L. S. (2001). The RAND-36 measure of health-related quality of life. Annals of Medicine, 33, 350-357.

73

Hays, R. D., Sherbourne, C. D., & Mazel, R. M. (1993). The RAND 36-Item Health Survey 1.0. Health Economics, 2, 217-227. Health Technology Systems. (2012). Inventory of Depressive Symptomatology (IDS) & Quick Inventory of Depressive Symptomatology (QIDS). Retrieved from http://www.idsqids.org/idsqids.pdf Hewstone, M., Fincham, F. D., & Foster, J. (2005). Psychology. Hoboken, NJ: Wiley-Blackwell. Heyman, R. E., Sayers, S. L., & Bellack, A. S. (1994). Global marital satisfaction versus marital adjustment: An empirical comparison of three measures. Journal of Family Psychology, 8, 432-446. Hooley, J. M., & Teasdale, J. D. (1989). Predictors of relapse in unipolar depressives: Expressed emotion, marital distress, and perceived criticism. Journal of Abnormal Psychology, 98, 229-235. Howard, K. I., Corniolle, T. A., Lyons, J. S., Vessey, J. T., Lueger, R. J., & Saunders, S. M. (1996). Patterns of mental health service utilization. Archives of General Psychiatry, 53, 696-703. Hoyle, R. H. (2000). Confirmatory factor analysis. In H. E. A. Tinsley & S. D. Brown (Eds.), Handbook of applied multivariate statistics and mathematical modeling (pp. 465-497). New York, NY: Academic Press. Jackson, D. L. (2003). Revisiting sample size and number of parameter estimates: Some support for the N:q hypothesis. Structural Equation Modeling: A Multidisciplinary Journal, 10, 128-141. Jenkins, C. D., Kraeger, B. E., Rose, R. M., & Hurst, M. W. (1980). Use of a monthly health review to ascertain illness and injuries. American Journal of Public Health, 70, 82-84. Johnson, N. J., Backlund, E., Sorlie, P. D., & Loveless, C. A. (2000). Marital status and mortality: The National Longitudinal Mortality Study. Annals of Epidemiology, 10, 224238. Jones, D. R., Macias, C., Barreira, P. J., Fisher, W. H., Hargreaves, W. A., & Harding, C. M. (2004). Prevalence, severity, and co-occurrence of chronic physical health problems of persons with serious mental illness. Psychiatric Services, 55, 1250-1257. Joreskog, K. G. (1993). Testing structural equation models. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 299-316). Newbury Park, CA: Sage. Juster, R.-P., Marin, M.-F., Sindi, S., Nair, N. P., Ng, Y. K., Pruessner, J. C., & Lupien, S. J. (2011). Allostatic load associations to acute, 3-year and 6-year prospective depressive symptoms in healthy older adults. Physiology & Behavior, 104, 360-364.

74

Karlamanga, A. S., Singer, B. H., & Seeman, T. E. (2006). Reduction in allostatic load in older adults is associated with lower all-cause mortality risk: MacArthur Studies of Successful Aging. Psychosomatic Medicine, 68, 500-507. Karney, B. R., Bradbury, T. N., Fincham, F. D., & Sullivan, K. T. (1994). The role of negative affectivity in the association between attributions and marital satisfaction. Journal of Personality and Social Psychology, 66, 413-424. Kaplan, R. M., & Kronick, R. G. (2006). Marital status and longevity in the United States population. Journal of Epidemiology and Community Health, 60, 760-765. Katon, W. J., & Sullivan, M. D. (1990). Depression and chronic medical illness. Journal of Clinical Psychiatry, 51, 3-11. Kenny, D. A. (2011). Path analysis. Retrieved from http://davidakenny.net/cm/pathanal.htm Kerns, R. D., & Otis, J. D. (2003). Family therapy for persons experiencing pain: Evidence for its effectiveness. Seminars in Pain Medicine, 1, 79-89. Kiecolt-Glaser, J. K., Dura, J. R., Speicher, C. E., Trask, J., & Glaser, R. (1991). Spousal caregivers of dementia victims: Longitudinal changes in immunity and health. Psychosomatic Medicine, 53, 345-362. Kiecolt-Glaser, J. K., Glaser, R., Cacioppo, J. T., MacCallum, R. C., Snydersmith, M., Kim, C., & Malarkey, W. B. (1997). Marital conflict in older adults: Endocrinological and immunological correlates. Psychosomatic Medicine, 59, 339-349. Kiecolt-Glaser, J. K., Newton, T. L. (2001). Marriage and health: His and hers. Psychological Bulletin, 4, 472-503. Kline, R. B. (2011). Principles and practice of structural equation modeling (3rd ed.). New York, NY: The Guilford Press. Kovacs, M., Sherrill, J., George, C. J., Pollock, M., Tumuluru, R. V., & Ho, V. (2006). Contextual emotion-regulation therapy for childhood depression: Description and pilot testing of a new intervention. Journal of the American Academy of Child and Adolescent Psychiatry, 45, 892-903. Krantz, D. S., & Manuck, S. B. (1984). Acute psychophysiologic reactivity and risk of cardiovascular disease: A review and methodologic critique. Psychological Bulletin, 96, 435-464. Kroenke, K., Spitzer, R. L., Williams, J. B., Monahan, P. O., & Lowe, B. (2007). Anxiety disorders in primary care: Prevalence, impairment, comorbidity, and detection. Annals of Internal Medicine, 146, 317-325. Lewandowski, W., Morris, R., Draucker, C. B., & Risko, J. (2007). Chronic pain and the family:

75

Theory-driven treatment approaches. Issues in Mental Health Nursing, 28, 1019-1044. Lim, J., Wood, B. L., Miller, B. D., & Simmens, S. J. (2011). Effects of paternal and maternal depressive symptoms on child internalizing symptoms and asthma disease activity: Mediation by interparental negativity and parenting. Journal of Family Psychology, 25, 137-146. Loeb, K., & le Grange, D. (2009). Family-based treatment for adolescent eating disorders: Current status, new applications and future directions. International Journal of Child and Adolescent Health, 2, 243-254. Malarkey, W., Kiecolt-Glaser, J. K., Pearl, D., & Glaser, R. (1994). Hostile behavior during marital conflict alters pituitary and adrenal hormones. Psychosomatic Medicine, 56, 4151. Manuck, S. B., Marsland, A. L., Kaplan, J. R., & Williams, J. K. (1995). The pathogenicity of behavior and its neuroendocrine mediation: An example form coronary artery disease. Psychosomatic Medicine, 57, 275-283. McDaniel, S. H., Campbell, T. L., Hepworth, J., & Lorenz, A. (2004). Family-oriented primary care: A manual for medical providers. New York, NY: Springer-Verlag New York LLC. McDaniel, S. H., Hepworth, J., & Doherty, W. J. (1992). Medical family therapy. New York, NY: Basic Books. McEwen, B. S. (1998). Stress, adaptation, and disease: Allostasis and allostatic load. Annals of the New York Academy of Sciences, 840, 33-44. McEwen, B. S., & Stellar, E. (1993). Stress and the individual: Mechanisms leading to disease. Archives of Internal Medicine, 153, 2093-2101. Merikangas, K. R., Ames, M., Cui, L., Stang, P. E., Ustun, T. B., Von Korff, M., & Kessler, R. C. (2007). The impact of comorbidity of mental and physical conditions on role disability in the US adult household population. Archives of General Psychiatry, 64, 1180-1188. Miller, B. D., & Wood, B. L. (2003). Emotions and family factors in childhood asthma: Psychobiologic mechanisms and pathways of effect. Advanced Psychosomatic Medicine, 24, 131-160. Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press. Minuchin, S., Baker, L., Rosman, B. L., Libman, R., Milman, L., & Todd, T. C. (1975). A conceptual model of psychosomatic illness in children: Family organization and family therapy. Archives of General Psychiatry, 32, 1031-1038.

76

Minuchin, S., Rosman, B. L., & Baker, L. (1978). Psychosomatic families: Anorexia nervosa in context. Cambridge, MA: Harvard University Press. National Center for Health Statistics. (2011). Health, United States, 2010 with special feature on death and dying. Retrieved from http://www.cdc.gov/nchs/hus.htm National Heart, Lung, and Blood Institute. (1997). National Asthma Education and Prevention Program expert panel report 2: Guidelines for the diagnosis and management of asthma (NIH publication no. 97-4051). Bethesda, MD: National Heart, Lung, and Blood Institute. National Research Council. (2001). New horizons in health: An integrative approach. Washington, DC: National Academy Press. Nichols, M. P., & Schwartz, R. C. (2004). Family therapy concepts and methods (6th ed.). Boston, MA: Pearson Education, Inc. NIAAA. (2005). Screening for alcohol use and alcohol-related problems. Alcohol Alert, 65, 1-7. Norman, S. B., Cissell, S. H., Means-Christensen, A. J., & Stein, M. B. (2006). Development and validation of an Overall Anxiety Severity and Impairment Scale (OASIS). Depression and Anxiety, 23, 245-249. Norton, R. (1983). Measuring marital quality: A critical look at the dependent variable. Journal of Marriage and the Family, 45, 141-151. Office of Human Subjects Research. (2006). Sheet 8: Answers to questions frequently asked of NIHs Office of Human Subjects research (OHSR). Retrieved from: http://ohsr.od.nih. gov/info/pdf/InfoSheet8.pdf Olfson, M., Shea, S., Feder, A., Fuentes, M., Nomura, Y., Gameroff, M., & Weissman, M. M. (2000). Prevalence of anxiety, depression, and substance use disorders in an urban general medicine practice. Archives of Family Medicine, 9, 876-883. Orts, K., Sheridan, J. F., Robinson-Whelen, S., Glaser, R., Malarkey, W. B., & Kiecolt-Glaser, J. K. (1995). The reliability and validity of a structured interview for the assessment of infectious illness. Journal of Behavioral Medicine, 18, 517-530. Pandhi, N., DeVoe, J. E., Schumacher, J. R., Bartels, C., Thorpe, C. T., Thorpe, J. M., & Smith, M. A. (2011). Preventive service gains from first contact access in the primary care home. Journal of the American Board of Family Medicine, 24, 351-359. Parkerson, G. R., Broadhead, W. E., & Tse, C.-K. (1995). Health status and severity of illness as predictors of outcomes in primary care. Medical Care, 33, 53-66. Poleshuck, E. L., Giles, D. E., & Tu, X. (2006). Pain and depressive symptoms among

77

financially disadvantaged womens health patients. Journal of Womens Health, 15, 182193. Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879-891. Prince, M., Patel, V., Saxena, S., Maj, M., Maselko, J., Phillips, M. R., & Rahman, A. (2007). No health without mental health. Lancet, 370, 859-877. Proulx, C. M., & Snyder, L. A. (2009). Families and health: An empirical resource guide for researchers and practitioners. Family Relations, 58, 489-504. Rand Health (2010). Medical Outcomes Study: 36-item short form survey scoring instructions. Retrieved from http://www.rand.org/health/surveys_tools/mos/ mos_core_36item_scoring.html Riso, L. P., Klein, D. N., Anderson, R. L., Ouimette, P. C., & Lizardi, H. (1996). Convergent and discriminant validity of perceived criticism from spouses and family members. Behavior Therapy, 27, 129-137. Robbins, J. M., Kirmayer, L. J., Cathebras, P., Yaffe, M. J., & Dworkind, M. (1994). Physician characteristics and the recognition of depression and anxiety in primary care. Medical Care, 32, 795-812. Robles, T. F., & Kiecolt-Glaser, J. K. (2003). The physiology of marriage: Pathways to health. Physiology & Behavior, 79, 409-416. Rogers, R. G., Hummer, R. A., & Nam, C. B. (2000). Living and dying in the USA: Behavioral, health, and social differentials of adult mortality. San Diego, CA: Academic Press. Rose, R. M., Jenkins, D., & Hurst, M. W. (1978). Health change in air traffic controllers: A prospective study - I. Background and description. Psychosomatic Medicine, 40, 142165. Rush, A. J., Trivedi, M. H., Ibrahim, H. M., Carmody, T. J., Arnow, B., Klein, D. N., Keller, M. B. (2003). The 16-item Quick Inventory of Depressive Symptomatology (QIDS), Clinician Rating (QIDS-C), and Self-Report (QIDS-SR): A psychometric evaluation inpatients with chronic major depression. Biological Psychiatry, 54, 573-583. Ryan, C. E., Epstein, N. B., Keitner, G. I., Miller, I. W., & Bishop, D. S. (2005). Evaluating and treating families: The McMaster Approach. New York, NY: Routledge. Saunders, J. B., Aasland, O. G., Babor, T. F., de la Fuente, J. R., & Grant, M. (1993). Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption II. Addiction, 88, 791-804.

78

Sbarra, D. A., & Nietert, P. J. (2009). Divorce and death: Forty years of the Charleston heart study. Psychological Science, 20, 107-113. Schmidt, A., Barry, K. L., & Fleming, M. F. (1995). Detection of problem drinkers: The Alcohol Use Disorders Identification Test (AUDIT). Southern Medical Journal, 88, 52-59. Schmitt, M. R., Miller, M. J., Harrison, D. L., & Touchet, B. K. (2010). Relationship of depression screening and physician office visit duration in a national sample. Psychiatric Services, 61, 1126-1131. Schumm, W. R., Paff-Bergen, L. A., Hatch, R. C., Obiorah, F. C., Copeland, J. M., Meens, L. D., & Bugaighis, M. A. (1986). Concurrent and discriminant validity of the Kansas Marital Satisfaction Scale. Journal of Marriage and the Family, 48, 381-387. Seeman, T. E. (2001). How do others get under out skin: Social relationships and health. In C. D. Ryff & B. H. Singer (Eds.), Emotion, social relationships, and health (pp. 189-209). New York, NY: Oxford University Press. Shields, C. G., Finley, M. A., Chawla, N., & Meadors, P. (2012). Couple and family interventions in health problems. Journal of Marital & Family Therapy, 38, 265-280. Shields, C. G., Franks, P., Harp, J. J., Campbell, T. L., & McDaniel, S. H. (1995). Family emotional involvement and criticism scale (FEICS): II. Reliability and validity studies, Family Systems Medicine, 12, 361-377. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422-445. Sidora-Arcoleo, K., Feldman, J. M., & Spray, A. (2012). A multi-factorial model for examining racial and ethnic disparities in acute asthma visits by children. Annals of Behavioral Medicine, 43, 15-28. Simon, G. E., VonKorff, M., Rutter, C., & Wagner, E. (2000). Randomized trial of monitoring, feedback, and management of care by telephone to improve treatment of depression in primary care. BMJ, 320, 550-554. Skipsey, K., Burleson, J. A., & Kranzler, H. R. (1997). Utility of the AUDIT for identification of hazardous or harmful drinking in drug-dependent patients. Drug and Alcohol Dependence, 1997, 157-163. Smith, A. F., & Prentice, D. A. (1993). Exploratory data analysis. In G. Keren & C. Lewis (Eds.), A handbook for data analysis in the behavioral sciences (p. 349-390). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

79

Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of marriage and other dyads. Journal of Marriage and the Family, 38, 15-28. Springer, K. W., & Mouzon, D. M. (2011). Macho men and preventive health care: Implications for older men in different social classes. Journal of Health & Social Behavior, 52, 212-227. Sutherland, J. M., Fisher, E. S., & Skinner, J. S. (2009). Getting past denial The high cost of health care in the United States. New England Journal of Medicine, 361, 1227-1230. Taj, N., Devera-Sales, A., & Vinson, D. C. (1998). Screening for problem drinking: Does a single question work? Journal of Family Practice, 46, 328-335 Tate, R. L. (1998). An introduction to modeling outcomes in the behavioral and social sciences (2nd Ed.). Edina, MN: Burgess International Group. Taylor, S. E., Lerner, J. S., Sage, R. M., Lehman, B. J., & Seeman, T. E. (2004). Early environment, emotions, responses to stress, and health. Journal of Personality, 72, 13651393. Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61, 201-216. Trivedi, M. H., Rush, A. J., Ibrahim, H. M., Carmody, T. J., Biggs, M. M., Suppes, T., Kashner, T. M. (2004). The Inventory of Depressive Symptomatology, Clinician Rating (IDS-C) and Self-Report (IDS-SR), and the Quick Inventory of Depressive Symptomatology, Clinician Rating (QIDS-C) and Self-Report (QIDS-SR) in public sector patients with mood disorders: A psychometric evaluation. Psychological Medicine, 34, 73-82. U.S. Department of Health and Human Services. (2011). General health status. Retrieved from http://www.healthypeople.gov/2020/about/GenHealthAbout.aspx Uchino, B. N., Cacioppo, J. T., & Kiecolt-Glaser, J. K. (1996). The relationship between social support and psychological processes: A review with emphasis on underlying mechanisms and implications for health. Psychological Bulletin, 119, 488-531. UCLA Academic Technology Services. (2012) Stata web books, regression with stata: Chapter 2 regression diagnostics. Retrieved from http://www.ats.ucla.edu/stat/stata/webbooks/reg/ chapter2/statareg2.htm Uebelacker, L. A., Smith, M., Lewis, A. W., Saski, R., & Miller, I. W. (2009). Treatment of depression in a low-income primary care setting with co-located mental health care. Families, Systems, & Health, 27, 161-171. Uebelacker, L. A., Wang, P. S., Berglund, P., & Kessler, R. C. (2006). Clinical differences

80

among patients treated for mental health problems in general medical and specialty mental health settings in the National Comorbidity Survey Replication. General Hospital Psychiatry, 28, 387-395. Umberson, D., Williams, K., Powers, D. A., Liu, H., & Needham, B. (2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47, 1-16. VanderZee, K. I., Sanderman, R., Heyink, J. W., & Haes, H. D. (1996). Psychometric qualities of the RAND 36-Item Health Survey 1.0: A multidimensional measure of general health status. International Journal of Behavioral Medicine, 3, 104-122. von Bertalanffy, L. (1969). General systems theory: Foundations, development, applications. New York, NY: George Braziller. Wagner, M. M., & Hogan, W. R. (1996). The accuracy of medication data in an outpatient electronic medical record. Journal of the American Medical Informatics Association, 3, 234-244. Wang, P. S., Lang, M., Olfson, M., Pincus, H. A., Wells, K. B., Kessler, R. C. (2005). Twelvemonth use of mental health services in the U.S.: Results from the National Comorbidity Survey Replication (NCS-R). Archives of General Psychiatry, 62, 629-640. Ware, J. E., & Sherbourne, C. D., (1992). The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item selection. Medical Care, 30, 473-483. Weeks, D. L., Corbett, C. F., & Stream, G. (2010). Beliefs of ambulatory care physicians about accuracy of patient medication records and technology-enhanced solutions to improve accuracy. Journal for Healthcare Quality, 32, 12-21. Weihs, K., Fisher, L., & Baird, M. (2002). Families, health, and behavior: A section of the commissioned report by the Committee on Health and Behavior: Research, Practice, and Policy Division of Neuroscience and Behavioral Health and Division of Health Promotion and Disease Prevention Institute of Medicine, National Academy of Sciences. Families, Systems, & Health, 20, 7-46. Whitlock, E. P., Polen, M. R., Green, C. A., Orleans, T., & Klein, J. (2004). Behavioral counseling interventions in primary care to reduce risky/harmful alcohol use by adults: A summary of the evidence for the U.S. Preventive Services Task Force. Annals of Internal Medicine, 140, 557-568. Wickrama, K. A. S., Lorenz, F. O., Wallace, L. E., Peiris, L., Conger, R. D., & Elder, G. H. (2001). Family influence on physical health during the middle years: The case of onset of hypertension. Journal of Marriage and Family, 63, 527-539. Wienke, C., & Hill, G. J. (2009). Does the marriage benefit extend to partners in gay and

81

lesbian relationships? Evidence from a random sample of sexually active adults. Journal of Family Issues, 30, 259-289. Williams, K., & Umberson, D. (2004). Marital status, marital transitions, and health: A gendered life course perspective. Journal of Health and Social Behavior, 45, 81-98. Wood, B. L. (1993). Beyond the psychosomatic family: A Biobehavioral Family Model of pediatric illness. Family Process, 32, 261-278. Wood, B. L. (2005). Commentary: Is it time for family-based interventions in pediatric psychology? Journal of Pediatric Psychology, 30, 694-697. Wood, B. L. (2012). Biopsychosocial. In L. LAbate (Ed.), Paradigms in theory construction (pp. 169-186). New York, NY: Springer. Wood, B. L., Klebba, K. B., & Miller, B. D. (2000). Evolving the Biobehavioral Family Model: The fit of attachment. Family Process, 39, 319-344. Wood, B. L., & Miller, B. D. (2002). A biopsychosocial approach to child health. In F. W. Kaslow & J. L. Lebow (Eds.), Comprehensive handbook of psychotherapy, Vol. 4 (p. 5980). Hoboken, NJ: John Wiley & Sons. Wood, B. L., & Miller, B. D. (2005). Families, health and illness: The search for mechanisms within a systems paradigm. In W. M. Pinsof & J. Lebow (Eds.), Family psychology: The art of the science (pp. 493-520). New York, NY: Oxford Press. Wood, B. L., Lim, J., Miller, B. D., Cheah, P. A., Simmens, S. ... Ballow, M. (2007). Family emotional climate, depression, emotional triggering of asthma, and disease severity in pediatric asthma: Examination of pathways of effect. Journal of Pediatric Psychology, 32, 542-551. Wood, B. L., Lim, J., Miller, B. D., Cheah, P. A., Zwetsch, T. ... Simmens, S. (2008). Testing the Biobehavioral Family Model in pediatric asthma: Pathways of effect. Family Process, 47, 21-40. Wood, B. L., Miller, B. D., Lim, J., Lillis, K., Ballow, M., Stern, T., & Simmens, S. (2006). Family relational factors in pediatric depression and asthma: Pathways of effect. Journal of American Academy of Child and Adolescent Psychiatry, 45, 1494-1502. Woods, S. B., & McWey, L. M. (2011). A biopsychosocial approach to asthma in adolescents encountering child protective services. Journal of Pediatric Psychology. Advance online publication. doi: 10.1093/jpepsy/jsr104 World Health Organization. (2006). Constitution of the WHO. New York, NY: WHO.

82

BIOGRAPHICAL SKETCH Sarah Woods is originally from Syracuse, New York. She earned a Bachelors degree in Psychology from the University of Rochester in 2005 and completed her Masters in Marriage and Family Therapy in 2007, also at the University of Rochester, in the School of Medicine and Dentistry. Sarah practiced as a clinician full-time for several years following receipt of her Masters degree and earned her license to practice in the state of New York in 2009. Sarah entered the doctoral program in Marriage and Family Therapy at The Florida State University in August, 2009 and advanced to candidacy status in November, 2011. While at FSU, she has published several articles in the area of families, health, and clinical practice, has been awarded several scholarships, and was elected to the College of Human Sciences Glenn Society, Kappa Omicron Nu, and Phi Kappa Phi. Sarah has also served the profession as a member of the Executive Board of the Tallahassee Association for Marriage and Family Therapy throughout her time in Tallahassee. Her program of research focuses on medical family therapy and family processes that affect physical health. In addition, her interest is in developing collaborative, evidence-based interventions for use in primary care that target patients relational functioning.

83