USF Libraries
USF Digital Collections

Depressive symptomatology, patient-provider communication, and patient satisfaction

MISSING IMAGE

Material Information

Title:
Depressive symptomatology, patient-provider communication, and patient satisfaction a multilevel analysis
Physical Description:
Book
Language:
English
Creator:
Novosel, Lorraine Marie
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Depression
Difficult patient
Patient-provider relationship
RIAS
DDPRQ
Dissertations, Academic -- Nursing -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Depression can be a profoundly disabling and costly disorder and is a major public health concern. Despite the efficacy of treatment options, it is often unrecognized, under-diagnosed, and inadequately treated in primary care settings. Research on patient-provider communication supports the connection among the quality of the patient-provider interaction, patient behavior, and health outcomes. The purpose of this study was to systematically examine the impact of patients' depressive symptoms on the patient-provider relationship, patient-provider communication, and patient satisfaction with the primary care office visit. One hundred twenty three patient-provider encounters were audiotaped and coded using the Roter Interaction Analysis System (RIAS).^ ^^A2 x 2 x 2 within-subjects factorial model provided the analytic framework for examining eight verbal communication behaviors categorized by speaker (patient or provider), type of utterance (question or information giving), and content of utterance (medical or psychosocial talk). Hierarchical linear modeling was used to analyze the two-level nested structure of the data. Results indicated that depression is associated with, but does not predict, increased provider-perceived difficulty in the patient-provider relationship. There was no significant change in either patient or provider communication behavior in relation to the severity of patients' depressive symptoms. Significantly more provider medical information was given during encounters with "difficult" patients and this behavior had a consistent negative effect on patient satisfaction.^ ^Patient-provider communication, by itself, does not appear to be a source of depressed patients' oft-reported dissatisfaction with medical care. Additional research is needed to further understand the core processes and structures of primary care practice in relation to the diagnosis and management of depression, their effect on patient outcomes, and to uncover opportunities for enhancing the effectiveness of depression care in primary care.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Lorraine Marie Novosel.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 180 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001920968
oclc - 190791030
usfldc doi - E14-SFE0001866
usfldc handle - e14.1866
System ID:
SFS0026184:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Depressive Symptomatology, Patient-Provide r Communication, and Patient Satisfaction: A Multilevel Analysis by Lorraine Marie Novosel A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Nursing University of South Florida Co-Major Professor: Judith F. Karshmer, Ph.D. Co-Major Professor: Jason W. Beckstead, Ph.D. Mary S. Webb, Ph.D. Eric M. Eisenberg, Ph.D. Date of Approval: April 4, 2007 Keywords: depression, difficult patient, pa tient-provider relationship, RIAS, DDPRQ Copyright 2007, Lorraine M. Novosel

PAGE 2

This dissertation is dedicated to my parents, Mary and Bill, who have unfailingly supported me in every possible way in all of my academic and life pursuits. Their unwavering faith in me motivated me to persist in completing this dissertation when my spirit waned. Thank you for instilling in me the confidence to believe in myself.

PAGE 3

Acknowledgments First and foremost, I gratefully acknowledge the participating physicians and nurse practitioners, and their office staffs, for graciously allowing me access to their patients as potential research participants. I am most indebted to the patients and providers for their time and willingness to share personal information in the spirit of improving the delivery of health care in this most challenging era. This project was funded by a grant from the American Academy of Nurse Practitioners Foundation. I thank Judith S. Dempster, DNSc, FNP, FAANP for sponsoring the Dempster Doctoral Dissertation Research Grant. I sincerely thank Susan M. Larson, M.S., Senior Research Associate, Johns Hopkins Bloomberg School of Public Health for her time, assistance, and patience teaching me the nuances of the RIAS coding system and for her consulting expertise throughout data analysis. I would like to thank my committee members for their support, comments, and suggestions on this project. In particular, I would like to thank Dr. Judith Karshmer for her time and editorial reviews of this manuscript. I also extend heartfelt appreciation to Dr. Jason Beckstead for his time, patience, and assistance with data analysis. You inspired and challenged me to embrace statistics and have been a trusted mentor throughout my doctoral program.

PAGE 4

Table of Contents List of Tables v List of Figures vi Abstract viii Introduction 1 Statement of the Problem 4 Background 6 Review of Literature 6 Depression Scope of the Problem 6 Diagnosis and Treatment 8 Depression and the Primary Care Practice 9 Barriers to Diagnosis and Treatment 11 The Difficult Patient 14 Patient-Provider Relationships 15 Patient-Provider Communication 17 Patient Satisfaction 22 Communication and Patient Satisfaction 24 Patient Health Status and Satisfaction 27 Depression and Satisfaction 27 Significance of the Study 29 Specific Aims 30 Theoretical Framework 30 Definitions 34 Depressive Symptomatology 34 Primary Care 34 Provider-Perceived Difficulty 34 Provider Medical Questioning 34 Patient Medical Questioning 35 Provider Medical Information Giving 35 Patient Medical Information Giving 35 Provider Psychosocial Questioning 36 Patient Psychosocial Questioning 36 Provider Psychosocial Information Giving 36 Patient Psychosocial Information Giving 36 i

PAGE 5

Patient Satisfaction 37 Hypotheses 37 Hypothesis 1 37 Hypothesis 2 37 Hypothesis 3 37 Hypothesis 4 38 Hypothesis 5 38 Hypothesis 6 38 Hypothesis 7 38 Hypothesis 8 38 Method 39 Pilot Study 39 The Study 40 Human Subjects Protection 40 Subjects 40 Sample Criteria 40 Sample Size Estimate 41 Measures 42 Provider Demographic Questionnaire 42 Patient Demographic Questionnaire 42 Beck Depression Inventory-Second Edition 43 Difficult Doctor-Patient Relationship Questionnaire 43 Roter Interaction Analysis System 44 Patient Satisfaction Survey 45 Procedure 45 Recruitment 45 Informed Consent Process 46 Data Collection 47 Raw Data Management 49 Analytic Plan 50 Results 52 Subjects 52 Provider Subjects 52 Patient Subjects 56 Data Management 59 Reliability Analysis of Measures 63 Preliminary Analysis 68 Hypothesis 1 72 Hypothesis 2 72 Hypothesis 3 72 Hypothesis 4 73 Hypothesis 5 73 ii

PAGE 6

Hypothesis 6 74 Hypothesis 7 74 Hypothesis 8 74 Hierarchical Linear Modeling 77 Unconstrained (Null) Model 79 Hypothesis 1 79 Hypothesis 2 83 Hypothesis 3 86 Hypothesis 4 89 Hypothesis 5 91 Hypothesis 6 93 Hypothesis 7 95 Exploratory Analysis 98 Discussion 106 Primary Aims and Hypotheses 106 Depressive Symptoms and the Difficult Patient 106 Depressive Symptoms and Patient-Provider Communication 109 Depressive Symptoms and Patient Satisfaction 111 Difficulty and Patient-Provider Communication 113 Patient-Provider Communication and Patient Satisfaction 115 Multilevel Modeling 117 Significance to Clinical Practice 119 Limitations 122 Direction for Future Research 124 Conclusions 126 References 128 Appendixes 145 Appendix A: IRB Approval Letters 146 Appendix B: Site Approval Letter 150 Appendix C: Provider Demographic Questionnaire 151 Appendix D: Patient De mographic Questionnaire 152 Appendix E: BDI-II Invoice (Authorization) 154 Appendix F: BDI-II Information 155 Appendix G: DDPRQ Authorization Letter 156 Appendix H: DDPRQ-10 Questionnaire 157 Appendix I: RIAS Agreement Letter 158 Appendix J: RIAS Information 159 Appendix K: Patient Satisfac tion Survey Authorization Letter 160 Appendix L: Patient Satisfaction Survey 161 Appendix M: Provider Recruitment Letter 164 Appendix N: Flyer 165 iii

PAGE 7

Appendix O: Provider Consent Forms 166 Appendix P: Patient Consent Forms 170 Appendix Q: HIPAA Authorization Form 174 Appendix R: Patient Notifica tion of Withdrawal from Study 176 Appendix S: Provider Notification of Patient Withdrawal from Study 177 Appendix T: Patient Notification of BDI-II Results 178 Appendix U: Provider Notification of BDI-II Results 179 About the Author End Page iv

PAGE 8

List of Tables Table 1 Characteristics of Provider Subjects By Provider Type ( N =11) 53 Table 2 Provider Ratings of Various Aspects of Depression Care Delivery ( N =11) 55 Table 3 Characteristics of Patient Subjects (N =123) 57 Table 4 Descriptive Statistics of Study Variables ( N =115) 62 Table 5 Reliability Analyses of Scales ( N =123) 64 Table 6 Reliability Analysis of Patient Satisfaction Survey Subscales ( N =123) 65 Table 7 Inter-rater Reliability Estimates of Individual and Composite RIAS Coded Variables (Medical) Using Pearsons r ( n=12) 66 Table 8 Inter-rater Reliability Estimates of Individual and Composite RIAS Coded Variables (Psy chosocial) Using Pearsons r ( n =12) 67 Table 9 Intercorrelations Between Communication Variables, Patient Depressive Symptoms, Provider-Perceived Difficulty, and Patient Satisfaction ( N =115) 69 Table 10 Summary of OLS Regression Analysis Predicting ProviderPerceived Difficulty and Pa tient-Provider Communication ( N =115) 70 Table 11 Summary of OLS Regression Analysis Predicting Patient Satisfaction ( N =115) 71 Table 12 Hierarchically Modeled Estimates of Fixed ( ) and Random ( ) Effects (Slopes) ( N =115) 78 Table 13 Additional RIAS Variables Included in Exploratory Analysis 100 Table 14 Inter-Coder Reliability Estimat es of RIAS Coded Global Affect 101 Variables Using Pearsons r ( n=12) v

PAGE 9

List of Figures Figure 1 Model Depicting the Hypothesized Relationships Between Patient Depressive Symptoms, Provider Perception of Difficulty in the Patient-Provider Relationship, Patient-Provider Communication, and Patient Satisfaction 51 Figure 2 Distributions of Provider DDPRQ Scores Adjusted for Patient Depressive Symptoms 81 Figure 3 Slopes of Provider-Perceived Di fficulty as a Function of Patient Depressive Symptoms for Each Provider 82 Figure 4 Distributions of Patient Satisfaction Scores Adjusted for Patient Depressive Symptoms, Provider Medical Questioning, and Provider Medical Information Giving 84 Figure 5 Slopes of Patient Satisfaction as a Function of Patient Depressive Symptoms for Each Provider Adjusted for Provider Medical Questioning and Provide r Medical Information Giving 85 Figure 6 Distributions of Provider Medical Information Giving Adjusted for Patient Depressive Symptoms and Provider-Perceived Difficulty 87 Figure 7 Slopes of Provider Medical In formation Giving as a Function of Patient Depressive Symptoms for Each Provider Adjusted for Provider-Perceived Difficulty 88 Figure 8 Slopes of Provider Psychosocial Questioning as a Function of Patient Depressive Symptoms for Each Provider Adjusted for Provider-Perceived Difficulty 90 Figure 9 Slopes of Provider Medical Information Giving as a Function of Provider-Perceived Difficulty for Each Provider Adjusted for Patient Depressive Symptoms 92 Figure 10 Slopes of Provider Psychosocial Questioning as a Function of Provider-Perceived Difficulty for Each Provider Adjusted for Patient Depressive Symptoms 94 vi

PAGE 10

Figure 11 Slopes of Patient Satisfaction as a Function of Provider Medical Questioning for Each Provider Adjusted for Patient Depressive Symptoms and Provider Me dical Information Giving 96 Figure 12 Slopes of Patient Satisfaction as a Function of Provider Medical Information Giving for Each Provider Adjusted for Patient Depressive Symptoms and Provider Medical Questioning 97 vii

PAGE 11

Depressive Symptomatology, Patient-Provide r Communication, and Patient Satisfaction: A Multilevel Analysis Lorraine Marie Novosel ABSTRACT Depression can be a profoundly disabling and costly disorder and is a major public health concern. Despite the efficacy of treatment options, it is often unrecognized, under-diagnosed, and inadequately treated in primary care settings. Research on patientprovider communication supports the connection among the quality of the patientprovider interaction, patient behavior, and health outcomes. The purpose of this study was to systematically examine the impact of patients depressive symptoms on the patient-provider relationship, patient-provider communication, and patient satisfaction with the primary care office visit. One hundred twenty three patient-provider encounters were audiotaped and coded using the Roter Interaction Analysis System (RIAS). A 2 2 2 within-subjects factorial model provided the analytic framework for examining eight verbal communication behaviors categorized by speaker (patient or provider), type of utterance (question or information giving), and content of utterance (medical or psychosocial talk). Hierarchical linear modeling was used to analyze the two-level nested structure of the data. Results indicated that depression is associated with, but does not predict, increased provider-perceived difficulty in the patient-provider relationship. There was no viii

PAGE 12

significant change in either patient or provider communication behavior in relation to the severity of patients depressive symptoms. Significantly more provider medical information was given during encounters with d ifficult patients and this behavior had a consistent negative effect on patient satisfaction. Patient-provider communication, by itself, does not appear to be a source of depressed patients oft-reported dissatisfaction with medical care. Additional research is needed to further understand the core processes and structures of primary care practice in relation to the diagnosis and management of depression, their effect on patient outcomes, and to uncover opportunities for enhancing the effectiveness of depression care in primary care. ix

PAGE 13

1 Introduction One out of every four adults in the United States experiences a diagnosable mental disorder each year (Kessler, Chiu, Demler, Merikangas, & Walters, 2005). Based on 2004 U.S. Census Bureau estimates this figure translates to approximately 57.7 million adults (U.S. Census Bureau, 2005). The true burden of these disorders may be even greater, as reports from the Global Burden of Disease Study indicate the prevalence of mental illness is heavily underestimated (Murray & Lopez, 1997a). Researchers acknowledge that many individuals remain reluctant to discuss their mental health history because of the stigma associated with mental illness. Stigma, a widespread, powerful, and often overwhelming barrier to diagnosis and treatment, prevents many individuals from even acknowledging their own mental health issues, much less disclosing their concerns with health care professionals. It is estimated that nearly half of all Americans who experience a severe mental illness do not seek any treatment at all (Department of Health and Human Services [DHHS], 1999). Mental illness refers collectively to all diagnosable mental disorders (DHHS, 1999). Characterized by alterations in thinki ng, mood, and/or behavior, mental disorders are associated with varied levels of emotional distress, impaired functioning, and a wide range of medical and social consequences. They can be highly disabling, ranking second only to cardiovascular conditions as a leading cause of worldwide disability as defined by the World Health Organization (Murray & Lopez, 1997a). Mental illness affects persons

PAGE 14

2 of all ages and individuals from all racial, ethnic, religious, educational, and socioeconomic groups. Left untreated, these disorders can be profoundly disabling and costly. The economic burden in the United States is estimated to be over $100 billion annually. (National Alliance on Mental Illness [NAMI], 2006). Depression is one of the most common and costly mental disorders. The total burden of depression in the U.S., including di rect medical, suicide-related mortality, and workplace costs was estimated at $83 billion in 2000 (Greenberg et al., 2003). An illness of substantial public health and economic significance, depression is a leading cause of disability in the United States and is projected to be the leading cause of disability burden by 2020 (Murray & Lopez, 1997b). Most individuals who seek help for depressive symptoms are evaluated in primary care settings, yet the majority of depressed patients are not diagnosed. Individuals experiencing subthreshold depression are even more likely to go unrecognized and undiagnosed. Of those who are diagnosed, most receive inadequate treatment even after adjusting for demographic, social/behavioral, depression severity, and economic factors (Kessler, Berglund, et al., 2003). Over the past 20 years, these facts have stimulated a multitude of studies aimed at improving provider detection and management of depression. Although many interventions have been shown to improve diagnosis, treatment, and patient outcomes, these improvements are generally short-lived as primary care providers typically return to baseline practice styles within a few months following the intervention (Lin et al., 1997). Healthy People 2010 (DHHS, 2000) designated mental health as one of the leading health indicators that will be used to measure the health status of the nation at the

PAGE 15

3 end of this decade. Modern treatments for mental disorders are highly effective, with a variety of treatment options available for most disorders. Despite the efficacy of treatment options, researchers report the failure of primary care providers to recognize, diagnose, and effectively treat depression and/or refer patients to mental health specialists (Spitzer et al., 1994). From a public health perspective, the consequences of untreated or sub-optimally treated depression are far-reaching and costly. From a personal perspective, the associated disability is immeasurable and often quite profound. Continuity of care is a main attribute of primary care and believed to enhance patient-provider relations and facilitate the disclosure of emotional distress. However, it has been reported that the continuity of care provided by the primary care providerpatient relationship is not sufficient to promote the discussion, disclosure, and detection of psychosocial issues (Wissow et al., 2002). This finding necessitates closer examination since primary care is the main portal of entry into treatment for mental disorders for the vast majority of individuals. The role of the primary care provider demands expertise in communication as numerous studies of patient-provider communication support the connection among the quality of the patient-provider interaction, patient behavior, and health outcomes. Although many barriers to the diagnosis and treatment of depression have been identified, the contributory affects of both patient and provider on the patient-provider relationship and the medical office encounter require further exploration. In order to enhance an understanding of the poor performance of primary care clinicians in the delivery of depression care services, research efforts must first identify all factors that

PAGE 16

4 impede care. As gatekeepers, primary care providers hold the key to improving the mental health of the nations citizens. Statement of the Problem Depression can be a profoundly disabling and costly disorder and is a major public health concern. Despite a multitude of psychometrically sound screening and case-finding instruments, enhanced educational efforts, communication skills training, effective treatment options, and clinical practice guidelines, depression often remains unrecognized, under-diagnosed, and inadequately treated by primary care clinicians. Primary care providers struggle with multiple competing demands in the delivery of care. Yet, diagnosing and treating mental health disorders and providing personal support are part of the responsibilities and tasks that define primary care. The role of interpersonal communication in the patient-provider relationship has been extensively studied and research findings consistently show that it is central to satisfactory patient-provider relationships and better patient health outcomes. These outcomes include patient satisfaction, patient adherence to treatment recommendations, functional status, symptom resolution, and measures of physiologic status (Brown, Stewart, & Ryan, 2003). However, there is little empirical evidence about how patients depressive symptoms impact the patient-provider relationship or patient-provider communication during the primary care office visit. Although it is widely documented that depressed patients are generally less satisfied with medical care than individuals who are not depressed, there is no extant research that has identified the communication

PAGE 17

5 behaviors that are important to this vulnerable group. The purpose of this study was to systematically examine the influence of patients depressive symptoms on the patientprovider relationship, patient-provider communication, and patient satisfaction with the primary care office visit.

PAGE 18

Background Review of Literature Depression Scope of the Problem Mental disorders are prevalent in society. About half of all Americans will meet criteria for a DSM-IV disorder (American Ps ychiatric Association, 1994) at some point during their lifetime: one out of five will experience a mood disorder. Mood disorders include major depression (unipolar depression) dysthymia, and bipolar disorder (manicdepression). Major depression is the most common mental disorder in the United States, affecting approximately 14.8 million adults each year (National Institute of Mental Health [NIMH], 2006). Although symptoms can develop at any age, the median age of onset is 32, affecting adults during their most productive years of life (Kessler, Berglund, et al., 2005). Seemingly more prevalent in women than in men (Kessler, Berglund, et al., 2003), the clinical presentation of depression varies among individuals, each of whom experience their own distinct combination of cognitive, affective, somatic and vegetative symptoms. Depression is associated with significant disability and functional impairment (Lin et al., 2000; McQuaid, Stein, Laffaye, & McCahill, 1999; Wu, Parkerson, & Doraiswamy, 2002). The negative physical impact is comparable to or greater than illnesses such as low back pain, arthritis, diabetes mellitus, and heart disease (Schonfeld et al., 1997). The disabling effects are even more pronounced when depression exists, as 6

PAGE 19

it frequently does, with co-morbid physical or psychiatric illness (McQuaid et al.; RoyByrne, 1996; Stein, Cox, Afifi, Belik, & Saree n, 2006; Wu et al., 2002). It is associated with increased costs and utilization of health care services (Callahan et al., 2002; Greenberg & Birnbaum, 2005; Roy-Byrne & Katon, 1997; Shvartzman et al., 2005; Stein et al., 2006), increased hospital length-of-stays (Pearson et al., 1999), non-compliance with prescribed medical regimens (DiMatteo, Lepper, & Croghan, 2000), work absenteeism (Stein et al.) reduced worker productivity (Druss, Schlesinger, & Allen, 2001) and suicide (Harris & Barraclough, 1997) Katon (2003) found depression doubled the medical costs of co-existing medical/physical illness even after controlling the severity of those illnesses. There is also an increasing body of literature suggesting depressive symptoms and major depression may be associated with increased morbidity and mortality from such illnesses as heart disease (Musselman, Evans, & Nemeroff, 1998), diabetes (DeGroot, Anderson, Freedland, Clouse, & Lustman, 2001), and osteoporosis (Robbins, Hirsch, Whitmer, Cauley, & Harris, 2001). Substantially more individuals suffer milder, but clinically significant depressive symptoms that do not meet DSM-IV criteria for major depression (Brody et al., 1998). Various terms have been used to describe these milder symptoms including subthreshold, subsyndromal, and minor depression. Despite the terminology subthreshold depression is also associated with significant disability, functional impairment, co-morbidity (Lyness, King, Cox, Yoediono, & Caine, 1999), increased costs and utilization of health care services (Pearson et al., 1999), and work absenteeism (Beck & Koenig, 1996). Although watchful waiting may be just as effective as antidepressants for treatment of milder 7

PAGE 20

symptoms (Kessler, Zhao, Blazer, & Swartz, 1997), many individuals with mild depression develop more serious depressive symptoms (Kessler, Merikangas, et al., 2003) and warrant careful monitoring. The economic burden of depression is staggering. Using data from the National Comorbidity Survey Replication (NCS-R), Greenberg et al. (2003) estimated the total economic burden of depression in the United States at $83 billion annually. This figure incorporated direct medical (31%), suicide-related mortality (7%), and workplace costs (62%). Unipolar depression, dysthymia, and bipolar disorders were included in their analysis. They also considered the cost of presenteeism (reduced worker productivity) in their assessment. These factors help to explain why their estimated annual cost is significantly greater than the $44 billion frequently cited in the literature. Diagnosis and Treatment A variety of efficacious treatment opti ons including pharmacological agents and psychotherapy, alone or in combination, are available to reduce depressive symptoms and assist patients to recovery. Yet, the literature is replete with reports indicating that depression is frequently unrecognized, largely under-diagnosed and inadequately treated. Dwight-Johnson, Sherbourne, Liao, and Wells (2000) noted that the majority of individuals experiencing depressive symptoms desire help. Unutzer et al. (2003) found that individuals are willing to pay up to 9% of their total household income for treatment of depression. Yet, only 20-25% of subjects reported receiving any type of mental health treatment (DHHS, 2000; Kessler, Demler, et al., 2005). The majority of depressed individuals prefer mental health counseli ng, although desire for treatment and patient 8

PAGE 21

preferences for type of treatment vary by ethnicity, gender, age, income, severity of symptoms, knowledge of treatment options (Brown & Schulberg, 1998; Cooper et al. 2003; Dwight-Johnson et al.; Unutzer et al.), and spirituality (Cooper, Brown, Vu, Ford, & Powe, 2001). Previous experience with antidepressants especially influences older adults who not only fear dependence to drugs, but very often resist acknowledging depression as a true medical illness (Givens et al., 2006). Depression and the Primary Care Practice Primary care clinicians address the majority of problems that individuals bring to the health care setting. Although some patient problems require referral, consultation, or hospitalization, the majority of problems are treated and resolved at the primary care level. Hence, primary care is the main portal of entry into treatment for mental disorders and the place where many individuals prefer to receive mental health services (Brody, Khaliq, & Thompson, 1997). The conventional primary care model conceptualizes primary providers as gatekeepers, who are able to identify individuals with mental disorders, initiate and manage treatment, and refer appropriately to mental health specialists. The majority of individuals feel that it is at least somewhat important that their primary care provider attend to their mental health needs (Brody et al.), report being comfortable discussing psychological problem s (Spitzer et al., 1994), and welcome the opportunity to address psychosocial concerns during the course of their medical visit (Bertakis, Roter, & Putnam, 1991). Seventy-five percent of individuals who seek help for depressive symptoms are initially evaluated in primary care settings (Goldman, Nielsen, & Champion, 1999). Yet, 9

PAGE 22

primary care providers fail to recognize or diagnose depression in up to 70% of patients (Callahan et al., 2002; Docherty, 1997; Nuyen et al., 2005). They often defer treating those they do identify, may simply monitor progress in others, and seldom refer patients to mental health specialists (Watts et al., 2002). Individuals experiencing subthreshold depression are even more likely to go unrecognized, undiagnosed, and untreated in primary care (Solberg, Korsen, Oxman, Fisc her, & Bartels, 1999; Von Korff & Simon, 1996). Rost et al. (1998) found nearly one-third of primary care patients with major depression remained undetected up to one year despite ongoing office visits. Approximately 50% of those individuals admitted to suicidal ideation during that time. Greater disability and severity of depressive symptoms increases the chance of being diagnosed (Borowsky et al., 2000; Klinkman, Schwenk, & Coyne, 1997; McQuaid et al., 1999; Simon, Goldberg, Tiemens, & Ustun, 1999). Yet, greater severity of symptoms and diagnosis does not result in improved patient outcomes (Klinkman et al., 1997). Of those who are diagnosed, most receive inadequate treatment even after adjusting for depression severity and demographic, social, behavioral, and economic factors (Kessler, Berglund, et al., 2003). Da ta from the National Ambulatory Medical Care Surveys (NAMCS) and the Medical Outcomes Study indicate primary care providers are less likely to recognize or diagnose depression in males, the elderly, Medicaid recipients, and African-American patients (Borowsky et al., 2000; Harman, Schulberg, Mullsant, & Reynolds, 2001). Diagnosing and treating depression in the primary care setting can be effective. A significant public health problem, the U.S. Preventive Services Task Force (USPSTF) 10

PAGE 23

recommends that adults be screened for depression in clinical practices that have systems in place to assure accurate diagnosis, effective treatment, and follow-up (USPSTF, 2002). Numerous screening instruments are available to assist primary care providers in detecting depressive symptoms in their patients. In busy clinical practice settings, asking patients one or two specific questions about depressed mood and anhedonia may be as effective as longer screening instruments (Kroenke, Spitzer, & Williams, 2003; Williams et al., 1999). Patients who screen positive require further diagnostic questioning to establish an appropriate diagnosis, taking into consideration gender, culture, and the presence of co-morbid somatic symptoms (Kerr & Kerr, 2001). Barriers to Diagnosis and Treatment Numerous barriers to the diagnosis and management of depression in primary care practice have been identified and detailed in the literature. Despite these barriers, many patients do attempt to communicate with their primary care providers. Sleath and Rubin (2002) found that patients were more likely to initiate discussions on depression than their providers were likely to inquire. Female gender, higher education, poor selfreported health status (Sleath & Rubin), greater psychological distress (Callahan et al., 1998) and patient-provider familiarity (Robinson & Roter, 1999) appear to facilitate disclosure of symptoms. Continuity of care is considered to be a main attribute of primary care and is widely believed to enhance patient-provider relations and disclosure of emotional distress. However, continuity appears to vary in its importance to different subsets of patients (Nutting, Goodwin, Flocke, Zyzanski, & Stange, 2003) and may have some untoward affects on patient care. Roter (2000) posits there may actually be some 11

PAGE 24

negative aspects of continuity, such as presumptuousness of familiarity, labeling, and simply the need for a fresh perspective (p. 13). This may at least partially explain the findings of Wissow et al. (2002) who found that the continuity of care associated with the patient-primary care provider relationship was insufficient to promote the discussion, disclosure, and detection of psychological issues. Patient presentation and chief complaint can impede a providers ability to recognize and diagnose depression as the primary reason for a depressed patient visit (Flocke, Frank, & Wenger, 2001). The high prevalence of co-morbid medical and psychiatric disorders is a well-established phenomenon in depressive illness and can mask symptoms and complicate the diagnostic reasoning process. In a secondary analysis of a depression intervention trial, K eeley et al. (2004) found nearly two-thirds of depressed patients presented exclusively with physical symptoms. Somatization is frequently the way that depression presents in primary care in addition to complaints of pain, fatigue, and sleep problems (Aragones, Labad, Pinol, Lucena, & Alonso, 2005; Wittchen, Lieb, Wunderlich, & Schuster, 1999). There are conflicting reports on the impact of somatization and the diagnosis of depression. Barkow et al. (2004) found the presence of somatic symptoms increased the likelihood that depressive symptoms would be detected, but Aragones et al. reported a link between somatization and under-detection of the underlying psychiatric process. Providers beliefs and attitudes about depression, knowledge, experience, availability of community resources, and personal level of comfort addressing patients emotional concerns are additional barriers that affect depression care. Yet, general 12

PAGE 25

practitioners appear to view barriers to providing effective treatment of depression as being more allied to external issues, particularly patient-centered factors (Nutting et al., 2002) and health system or service provision barriers (Telford, Hutchinson, Jones, Rix, & Howe, 2002) rather than internal factors such as their own knowledge and skills. In a study using focus group and survey methods, Solberg et al. (1999) found primary care physicians cite lack of time as the biggest barrier to the delivery of psychosocial care. Time management is paramount for primary care providers in todays health care environment. Klinkman (1997) presents a comprehensive discussion of the multiple competing demands that primary care providers face in the delivery of psychosocial care and partially explain the appearance of inadequate depression care. Direct observation reveals primary care visits average only 10 to 19 minutes of direct doctor-patient contact time (Flocke et al., 2001; Stange et al., 1998) although patients bring with them an average of 6 problems to which they need attention (Williams et al., 1999). A linear trend exists between the number of patient problems and length of the encounter and in the vast majority of visits more than one problem is discussed (Flocke et al.). There is also evidence that the discussion of behavioral or emotional concerns, especially when raised during visits not originally identified as a psychological consultation, significantly increases the length of the office visit (Callahan et al., 1998; Cooper, Valleley, Polaha, Begeny, & Evans, 2006; Flocke et al.). Sleath and Rubin (2002) observed talk about depression occurred in only 25% of depressed patient visits adding support to the conclusion of Rost et al. (2000) that the a ttention patients with depression get during a 13

PAGE 26

given medical visit is less associated with the severity of the patients depressive symptoms than with the number or recency of other problems. The Difficult Patient Typically 10% to 20% of patients provoke a level of physician distress and frustration that transcends the expected and accepted level of difficulty (Hahn, Thompson, Wills, Stern, & Budner, 1994; Hahn et al., 1996). The difficulty is not a property of the patient per se, but is an attribute of the patient-provider relationship as perceived by the provider. This frustration is often expressed through the labeling of such patients with derogatory names. Although the labeling is not a formal diagnosis, it represents a subjective and negative categorization that once applied, is likely to be inscribed in a providers psyche. Frequently referred to as heartskink (ODowd, 1988), problem (Drossman, 1978), or difficult patients in the literature, use of these and similar monikers is widespread in clinical medicine. Groves (1978) seminal article on the hateful patient first legitimized physicians feelings of frustration and dislike toward certain patients. While it is clear that physicians are responsible for the problem patient designation, McGaghie and Whitenack (1982) carefully noted that they are also responsible for the medical actions that are associated with the labeling process. It is not clear how such feelings manifest during the patient-provider encounter or if providers predictably respond with dislike to specific subsets of patients. Certain features are common to difficult patients. A strong association with mental disorders, especially depressive and anxiety disorders is well documented (Hahn, 2001; Jackson & Kroenke, 2001; Kroenke, Jackson, & Chamberlin, 1997). Other 14

PAGE 27

characteristics include an increased number of patient-reported physical symptoms and somatization (Hahn), multiple non-specific complaints (Steinmetz & Tabenkin, 2001), abrasive behavioral style (Schwenk, Ma rquez, Lefever, & Cohen, 1989), greater functional impairment, higher utilization of health care services, more unmet expectations, lower levels of satisfaction with care (Hahn et al., 1996), more telephone calls to office, and more referral requests to specialists (Linn et al., 1991). Provider characteristics also play a role in defining the difficult patient. A poor provider psychosocial attitude (Jackson & Kroenke, 1999), less practice experience (Steinmetz & Tabenkin), and increased medical uncertainty in the care of the patient (Schwenk et al.) have been found to underlay physicians perceptions of difficult patients. Using quantitative and qualitative methods Mas Garriga et al. (2003) found physicians believe that the skills and strategies they have to help them manage such patients are limited. Patient-Provider Relationships Patients and providers constantly influence each other during the medical encounter. Patient ratings of the patient-provider relationship are increasingly being recognized by health care organizations as a measure of quality of care. Hall, Horgan, Stein, & Roter (2002) studied patients and physicians in established relationships and found patients accurately estimated their physicians liking of them. They also found correlations between patients liking of their physician and better self-reported health status, affective state following the encounter, better ratings of physician behavior, and increased visit satisfaction. 15

PAGE 28

There is evidence that the quality of the patient-provider relationship affects patient care. Barsky, Wyshak, Latham, and Klerman (1991) found physicians were less aware of symptoms of depression and anxiety in frustrating patients than in other patients. The authors suggested that a bad patient-provider relationship might dull a doctors sensitivity to patients symptoms. There is also some evidence from the patient perspective that the quality of the relationship affects depression care. OMalley, Forrest, and Miranda (2003) reported that in their sample of low-income African American women, those who rated their providers as having more respect for them were more likely to report being asked about and treated fo r depression. In a cross-sectional analysis of 18-month data Meredith, Orlando, Humphrey, Camp, & Sherbourne (2001) found an association between higher patient ratings of the interpersonal relationship and receipt of quality depression care (guideline-concordant care). Although much has been written about the nature of difficulties in the patientprovider relationship, the literature is mostly descriptive. Researchers grappled with ways to operationalize and measure providers perceptions of the patient-provider relationship until a valid and reliable instrument became available in recent years. Since then, a review of the literature reveals that although researchers have been able to quantify the degree of provider-perceived difficulty in the patient-provider relationship, the data have primarily been used in efforts to determine patient correlates of difficulty and, to a small extent, examine patient outcomes. Of particular interest is the impact of a providers negative feelings for his/her patient on the patient-provider encounter. Bensing and Dronkers (1992) analyzed 103 videotaped real-life general practice 16

PAGE 29

consultations. The authors found providers asked fewer questions, especially psychosocial questions, in encounters in which the physician appeared to be irritated, anxious, or nervous. There is no known research that has measured providers perceived difficulty with patients and then observed how that difficulty influences the ensuing medical encounter. It is possible that a dislik e for, or discomfort with, a specific patient (or subset of patients) has an impact on the dynamics of and challenges inherent in the delivery of depression care in the primary care environment. Patient-Provider Communication Many factors contribute to the evolution of the patient-provider relationship. However, interpersonal communication is the central element and is fundamental to both establishing and preserving the partnership. Roter (2000) describes the medical dialog as the fundamental vehicle through which the paradigmatic battle of perspectives is waged and the therapeutic relationship is defined (p. 5). There is much more to communication than just spoken words. Brown, Nelson, Bronkesh, and Wood (1993) write patients often judge the quality of communication not only by words but by a handshake, eye contact, and the white spaces when no words are spoken but an emotional or personal connection is made (p. 256). Difficulties in the patient-provider relationship are often a consequence of a breakdown in communication between patient and provider. Levinson, Stiles, Inui, & Engle (1993) found that physicians often blame patients for the problem. In their study exploring physicians perceived communication difficulties in frustrating patient visits, Levinson et al. found physicians difficulties clustered around six primary problem areas: (1) lack of trust/agreement, (2) too many [patient] problems, (3) feeling 17

PAGE 30

distressed, (4) lack of understanding, (5) lack of adherence, and (6) demanding/controlling patient. Much has been written about the role, content, and structure of interpersonal communication during the patient-provider relationship. Ong, de Haes, Hoos, & Lammes (1995) identified three different purposes of communication: (1) creating a good interpersonal relationship, (2) exchanging information, and (3) making treatment-related decisions. They also discussed specific communicative behaviors that are displayed during medical consultations including (1) instrumental (cure oriented) vs. affective (care oriented), (2) verbal vs. non-verbal, (3) privacy, (4) high vs. low controlling behavior, and (5) medical vs. everyday language vocabularies. Hall, Roter, & Katz (1988) collapsed nearly 250 different elements of communication into five categories including (1) information giving, (2) question asking, (3) partnership-building, (4) rapport-building, and (5) socioemotional talk. Stewart (1995) and Beck, Daughtridge, and Sloane (2002) provide comprehensive reviews on physician-patient communication and health outcomes. Thompson and Parrott (2002) outlined advances that have been made toward the development of a theory of health communication (p. 708-709). Communication patterns during the medical encounter have also received much attention in the literature. Physician communication styles have been described and defined as affiliation and control (Buller & Buller, 1987), mutuality, paternalism, consumerism, and default (Emanuel & Emanuel, 1992), patient-centeredness, directing vs. sharing, and affiliation vs. dominance (Williams, Weinman, & Dale, 1998), and person-focused, biopsychosocial, biomedical, and high physician control (Flocke, Miller, 18

PAGE 31

& Crabtree, 2002). Roter et al. (1997) found five distinct patterns emerged in their analysis of 537 audiotaped primary care encounters and classified them according to the frequency of both physician and patient questioning and information giving on biomedical and psychosocial issues. They categorized the patterns as narrowly biomedical, expanded biomedical, biopsychosocial, psychosocial, and consumerist. Two-thirds of the visits were characterized as being physician-dominated and narrowly focused on biomedical concerns. The asymmetrical nature of the patient-provider relationship has been well documented. Accordingly, the medical encounter contains elements of power and control. Thompson and Parrott (2002) note, Control processes are seen as particularly important within the health care context because of the traditional paternalistic relationship between physicians and patients (p. 692). Control in relationships is expressed by sending messages that define, direct, or dominate. Techniques that can be used include confrontation, question authority, make assertions, disconfirm, change topics, initiate or terminate interactions (OHair, 2003), interruption (Realini, Kalet, & Sparling, 1995), rejecting topics of conversation (Suchman, Markakis, Beckman, & Frankel, 1997), and dominating talk time (Roter et al., 1997). Physicians can exercise their power to control the content and course of the patient encounter in many ways. Often, that control is exerted by the use of questions (Beckman & Frankel, 1984). Physicians beliefs about psychosocial care can also influence their communication patterns. Those with more positive attitudes have been found to use more statements of emotion, fewer closed-ended questions, and have more 19

PAGE 32

psychosocial discussions with patients (Levinson & Roter, 1995). Yet, Roter, Lipkin, Stewart, and Stiles (1988) reported that medical encounters that entailed a lot of psychosocial talk and little biomedical talk, although most satisfying to patients, appeared to be the most anxiety-producing pattern of communication for physicians. Their analysis adds support to the earlier findings of Dungal (1978) who found physicians reported more anxiety and frustration, less interest and comfort, and a less satisfactory note when they were dealing with psychosocial problems. Patients seldom verbalize their emotions directly and spontaneously, tending to offer clues instead. However, there is evidence that physicians interviewing styles do not change according to patient characteristics such as emotional distress (Del Piccolo, Mazzi, Saltini, & Zimmerman, 2002; Deveugele, Derese, & DeMaeseneer, 2002). In fact, Suchman et al. (1997) found physicians a llowed both indirect and direct expressions of emotion to pass without acknowledgement, opting instead to return to the preceding topic, usually the diagnostic exploration of symptoms. Although there is no empirical evidence that patient ordering of concerns is related to medical importance or severity, Beckman and Frankel (1984) found patients were able to complete their opening statement of concern only 23% of the time before being interrupted by their physician within 18 seconds, on average. Physicians gain control of the dialog nearly 80% of the time after they interrupt patients (Realini et al., 1995). These data present a disturbing portrayal of some of the dynamics involved in patient care and highlight significant challenges patients face when seeking treatment for depression in primary care. 20

PAGE 33

The effects of patient characteristics on the patient-provider interaction, such as age, gender, ethnicity/race, culture, socioeconomic status, education, appearance, disability, and health status are well documented in the literature. In a secondary analysis of over 1,300 patient encounters, Hall, Roter, Milburn, and Daltroy (1996) found physicians conveyed negativity in their behaviors with sicker or more emotionally distressed patients and were less likely to engage in social conversation. Patients, in turn, provided more biomedical and psychosocial information, engaged in more emotionally concerned talk, asked more psychosocial questions, and used more submissive voice tone. The authors concluded that the lack of social conversation could reflect the realities of sicker patients medical care and/or psychological avoidance. Much of the research that has examined depression and patient-provider communication has focused on the dynamics of the exchange in relation to the physicians recognition and diagnosis of depression, provider communication skills training, and patient outcomes following various specialized, intensive, or collaborative care depression management programs. Callahan et al. (1996) observed less chatting occurred during depressed patient office visits. There is also some evidence relating aspects of communication during the encounter with the receipt of guideline-concordant depression care and improvement of patients depressive symptoms (Clever et al., 2006). Little is known, however, about discourse during the medical encounter as it relates to the depressed patient primary care provider dyad. Historically, researchers have focused on providers behaviors towards patients. In recent years there has been a shift in emphasis to the patients experience and individuality. Hall and Visser (2000) noted that 21

PAGE 34

studying providers has always implied studying patients. Their behaviors are intertwinedsuch that the behavior and outcomes of each [depends] on the attitudes, expectancies, skills, and behavior of the other (p. 115). In order to fully understand the dynamics of and challenges inherent in the diagnosis and treatment of depression in primary care, it is clear that research must focus on the contributory affects of both patients and providers on the dyadic exchange during the encounter. Previous findings indicate th at primary care providers are uncomfortable dealing with patients psychosocial problems and that physicians exercise power in the relationship to control the content and course of dialog during consultations. Although physicians most likely use various techniques to control the depressed patient visit agenda, it is reasonable to speculate that their actions serve to direct the discourse toward a more comfortable biomedically focused pattern of communication. Patient Satisfaction Patient satisfaction is a legitimate measure of health care quality because of its significant associations with other patient outcomes (Brown et al., 2003). Patients expectations and requests are central to most theories of patient satisfaction and understanding and responding to patients needs is an inherent goal of medicine. Yet, adoption of a universal definition is difficu lt. Brown et al.s (1993) definition of satisfaction as what each patient says it is (p. 9) illustrates that individuals form their own unique opinions of care according to his or her particular needs and experiences. Although patient satisfaction is arguably the most researched outcome of the patientprovider interaction, salient differences in methodologies often confound the ability to 22

PAGE 35

interpret and generalize findings. A cursory scan of the literature is all that is needed to recognize many significant differences in the way patient satisfaction and its various dimensions have been operationalized. Furthe rmore, comparative studies using different satisfaction instruments, even when measuring similar aspects of care, have been found to yield unreliable and inconsistent results (Cohen, Forbes, & Garraway, 1996; Ross, Steward, & Sinacore, 1995). There are many dimensions of patient satisfaction. Ware, Snyder, Wright, and Davies (1983) taxonomy included (1) interpersonal manner, (2) technical quality, (3) accessibility/convenience, (4) finances, (5) efficacy/outcomes, (6) continuity, (7) physical environment, and (8) availability. In a meta-analysis of the patient satisfaction literature Hall and Dornan (1988) categorized aspects of satisfaction with (1) access, (2) cost, (3) overall quality, (4) humaneness, (5) competence, (6) amount of information supplied by the provider, (7) bureaucratic arrangements, (8) physical facilities, (9) providers attention to psychosocial problems of the patient, (10) continuity of care, (11) outcome of care, and (12) overall satisfaction. The authors found that aspects of satisfaction related to the provider, including overall quality, humaneness, and competence ranked the highest. The position score of satisfaction with attention to psychosocial problems ranked the lowest. They acknowledged the various aspects of satisfaction were measured with very uneven frequencies and expressed concern that, considering the prevalence of unresolved psychosocial problems and their impact on health care utilization, only 3% of studies measured that particular aspect of care. 23

PAGE 36

Many determinants of patient satisfaction with care have been documented. Of these, physician communication, especially when measured immediately after the encounter (Jackson et al., 2001), is widely believed to have the most direct impact on patients assessments of the quality of care they receive. Yet, patients and physicians do not agree on the relative importance of effective communication (Laine et al., 1996). Buller and Buller (1987) reported nearly 75% of the variance in evaluations of medical care was associated with patient ratings of physicians communication. Thompson and Parrott (2002) cautiously noted that, at times, the study of patient satisfaction has reflected an assumed relationship between patient-provider communication and satisfaction since there has been no direct test of the relationship. Ong, Visser, Lammes, & de Haes (2000) recorded patient visits and found cancer patients satisfaction was most clearly predicted by the affective quality of the consultation. Interpretation of their findings is somewhat limited, however, since the affective quality of the interaction was not determined by the patients but by third-party coders. Overall patient satisfaction ratings of varied aspects of medical care are generally favorable. Notwithstanding, the literature is replete with evidence of widespread patient dissatisfaction with what is perhaps the most important quality of care indicator: provider communication during the medical encounter. Any dissatisfaction is a significant issue with which to contend. Communication and Patient Satisfaction Different surveys performed in the primary care setting suggest tentative conclusions on the impact of specific communicative behaviors on patient satisfaction. 24

PAGE 37

Williams, Weinman, et al. (1998) provided a comprehensive discussion organized according to the communication behavior categories of (1) information provision by the doctor and/or patient, (2) information-seeking behaviors of doctors and patients, (3) the doctor-patient relationship and expression of negative or positive affect by doctor and patient, and (4) the communication style of the doctor. Brown et al. (2003) reported patient satisfaction positively associated with medical encounters in which patients perceived their provider as caring, understanding, and competent, and with visits characterized by a balanced inquiry into biomedical, psychological, and sociological concerns. They also noted continuity in the patient-provider relationship and provider awareness of patients expectations to be key factors impacting satisfaction. Several studies provide insight to the impact of communication patterns on patient satisfaction. Marvel, Doherty, and Weiner (1998) and Bertakis et al. (1991) found patients reported the greatest satisfaction in visits that encouraged them to talk about psychosocial issues. In a large multi-site collaborative study of primary care visits Roter et al. (1997) also found patient satisfaction with the psychosocial pattern of communication was significantly higher than any other pattern. This pattern was characterized by a balanced amount of physician biomedical and psychosocial talk while patients devoted more than twice their talk to psychosocial (55%) compared to biomedical (25%) topics. Biomedical pattern s received the lowest ratings. Flocke, et al. (2002) observed over 2,800 primary care visits and found physicians with a personfocused practice style rated highest on patient satisfaction. They described these physicians as being personable and friendly, receptive to the patients agenda, willing to 25

PAGE 38

negotiate with patients, and more focused on the patient than the disease. Satisfaction ratings were lowest for physicians who dominated the encounter and disregarded the patients agenda. Buller and Buller (1987) found affiliative styles of communication were related positively to patient satisfaction, whereas dominant/active styles had a negative relationship. Other studies also reported an inverse relationship between provider dominance during the encounter and patient satisfaction (Cecil & Killeen, 1997; Di Blasi, Harkness, Ernst, Georgiou, & Kleijnen, 2001; Flocke et al., 2002; Roter et al., 1997). Communication and satisfaction have been examined in relation to a number of variables including previous patient-provider interaction experience and number of prior encounters (Buller & Buller, 1987; Sixma, Spreeuwenberg, & van der Pasch, 1998), patient expectations (Jackson, Chamberlin, & Kroenke, 2001), gender (Roter, Geller, Bernhardt, Larson, & Doksum, 1999), age (Callahan et al., 2000), social class (Hall, Roter, et al., 1988), cultural barriers (Hornberger, Itakura, & Wilson, 1997), education and income (Weiss, 1988), length of patient visit (Gross, Zyzanski, Borawski, Cebul, & Stange, 1998), physical disabilities (Kroll, Beatty, & Bingham, 2003), patient selfefficacy (Zandbelt, Smets, Oort, Godfried, & de Haes, 2004), patient involvement in decision making (Clever et al., 2006) physician communication skills training (Bredart, Bouleuc, & Dolbeault, 2005), quality of care (Chang et al., 2006; Flocke et al., 2002), intention to discontinue care (Federman et al., 2001) and malpractice claims (Beckman, Markakis, Suchman, & Frankel, 1994). 26

PAGE 39

Patient Health Status and Satisfaction Certain factors may predispose some individuals toward satisfaction or dissatisfaction with care, including overall lif e satisfaction, confidence in the medical community, locus of control, and patient health status. Patients who have, or perceive themselves as having, worse physical or mental health are generally less satisfied with their medical care than patients who report better health. At the present time there is no clear understanding of the underpinnings or direction of this association. Patient health status appears to impact communication during the medical encounter, and there is some evidence that it might, in turn, influence patient satisfaction. In a secondary analysis of data, Hall, Milburn, Roter, and Daltroy (1998) explored two explanatory models and found support for a direct hypothesis that poor health reduced satisfaction directly. Causal modeling also supported a physician mediation hypothesis and found that physicians reaction to sicker patients supported an indirect explanation for patient dissatisfaction in the form of curtailed social conversation. Hall, Epstein, DeCiantis, and McNeil (1993), Hall, Milburn, and Epstein ( 1993), and Like and Zyzanski (1987) have all reported a correlation between physicians personal feelings about patients and patient satisfaction. Depression and Satisfaction It is well documented that individuals with depression are significantly less satisfied with their medical care. One hypothesis is that depressed patients are less satisfied with care due to the adverse effects of depression on mood and cognition. However, little is known about the impact of depression on the patient-provider 27

PAGE 40

encounter. Furthermore, there is limited research to identify the communication behaviors that are important to this vulnerable group. There is general evidence that health care providers interpersonal skills are a very important aspect of care to depressed patients (Cooper et al., 2000). Some studies have found patient-reported communication problems associated with depression (Druss et al., 2001; Lerman et al., 1993; van Os et al., 2005), consistent with a plethora of reports indicating widespread dissatisfaction with provider communication. Clever et al. (2006) investigated the impact of patient involvement in decision-making (IDM) on depression care. They found higher patient IDM ratings associated with significantly greater satisfaction. In a longitudinal study Orlando and Meredith (2002) found depressed patients who reported higher satisfaction with care were more likely to receive higher quality depression care at six months compared to those who were less satisfied. Few studies have examined depressed patients satisfaction with medical care in relation to other outcomes. Solberg, Fischer, Rush, and Wei (2003) found dissatisfied depressed patients at the index visit continued to be dissatisfied with care and had poorer outcomes during a three-month follow-up period. Dissatisfied depressed patients have also been found to likely have persistent depressive symptoms and reduced productivity two years following initial care (Druss et al., 2001). The significant association between depression and (dis)satisfaction suggests that further research is needed to identify the sources of satisfaction for such patients. A thorough understanding of what depressed patients find important in their care is 28

PAGE 41

necessary in order to improve depression care, patient satisfaction, and its many associated outcomes. Significance of the Study Depression is one of the most common mental disorders and a leading cause of disability in the U.S. Left untreated, the personal and societal consequences are pervasive and costly. Despite a multitude of screening instruments and highly effective treatment options, depression is often undiagnosed and inadequately treated by primary care providers. The goal of this study was to develop a better understanding of factors that may impact the diagnosis and effective management of depression in the primary care setting. The purpose of this study was to systematically examine the influence of patients depressive symptoms on the patient-provider relationship, patient-provider communication, and patient satisfaction with the primary care office visit. It examined the contributory effects of both patients and providers on the dyadic exchange during the medical encounter. It was hypothesized that some communication behaviors during the medical encounter might actually promote patient dissatisfaction and ultimately discourage patients from discussing their mental health issues with their providers. Such a condition, if the case, would function as a barrier to the diagnosis and effective management of depression in the primary care setting. It would be a significant set-back to the current charge to encourage patients to discuss their emotional needs with their primary care provider as part of the overall effort to effectively diagnose and treat depression in primary care. 29

PAGE 42

Specific Aims The specific aims of this study were to: 1. Determine the influence of patient depressive symptoms on (a) provider perception of difficulty in the patient-provider relationship, (b) patientprovider communication, and (c) patient satisfaction. 2. Determine the influence of provider perception of difficulty in the patientprovider relationship on patient-provider communication during the medical encounter. 3. Explore the potential mediating effect of provider-perceived difficulty in the patient-provider relationship on the association between patient depressive symptoms and patient-provider communication during the medical encounter. 4. Explore the potential mediating effect of patient-provider communication on the association between patient depressive symptoms and patient satisfaction. 5. Identify patient and provider verbal communication behaviors that promote patient satisfaction with the medical encounter. Theoretical Framework Emersons social exchange theory of power-dependence relations was used as a guide to design this study (Emerson, 1962, 1972). The theory focuses attention on the characteristics of the social relationship, whether person-to-person, person-to-group, or group-to-group. Central to the theory is the concept of power explicitly defined as a property of the social relation and not an attribute of a person or group (called actors). 30

PAGE 43

As social relations entail ties of mutual dependence between parties, power (P) is the function of dependence (D) of one actor on another. The degree of dependence of one actor upon another actor is directly proportional to the value the actor places on the resources (or valued behavior) mediated by the other actor and inversely proportional to the availability of those resources (or behaviors) from other sources outside of the relationship. Emerson views power as potential: although it is a component of all interactions, it may or may not be exercised in a given situation. Power will be empirically manifest only if actor A makes some demand, and only if this demand runs counter to actor Bs desires (Emerson, 1962, p. 33). Understanding his conceptualizations of exchange relationship and transaction are also important. An exchange relationship exists between two parties when they exchange at least occasionally and could be expected to do so, at least occasionally, into the foreseeable future. A transaction is a mutually reinforcing or rewarding instance of exchange that can be initiated by either person in the social relation (Emerson, 1972). Another principal concept in Emersons theory is the idea that exchange relations can be balanced or imbalanced based on the level of dependency each party has on the other. Unbalanced relations are unstable and encourage the use of power. Exercising power opens the door to cost reduction and balancing operations. To Emerson, the costs involved are anchored in modifiable attitudes and values and cost reduction is a process involving change in values (personal, social, economic) which reduces the pains incurred in meeting the demands of a powerful other (Emerson, 1962, p. 35) Social exchanges can also be negotiated or reciprocal. In negotiated exchange the terms of the 31

PAGE 44

exchange are negotiated and agreed upon and each party knows what he or she will receive from the transaction. In reciprocal exchanges however, the terms of the exchange are not negotiated and each party chooses behaviors that have consequences for their interaction partners without knowing what the other person will choose to do with them (Smith-Lovin, 2001). Recognizing the reciprocity of social relations, Emerson represents the power-dependence relation as the following pair of equations: Pab =Dba (the power of A over B is equal to, and based upon, the dependence of B upon A) and Pba=Dab (the power of B over A is equal to, and based upon, the dependence of A upon B) (Emerson, 1962, p. 33) According to Emerson, actor A, paired with actor B in a social relation, conduct an instance of exchange called a transaction. A initiates behavior X1 during the transaction that evokes stimulus feedback Y from B. In reciprocity, B delivers help to A if and only if X1 is rewarding to B. If As behavior X1 does not evoke a rewarding behavior from B, then through the conditioned reinforcement of feedback Y, As behavior will either (a) change to form X2 which is reciprocated, or (b) the A:B relation will extinguish. While this stimulus consequence defines the temporal boundary of that particular transaction, as a feedback system, it regulates the next transaction. Therefore, each transaction is linked to a history and a future for the actors (Emerson, 1972). By virtue of their roles, the traditional patient-provider relationship is defined by an imbalance of power. Patients automatically enter into the relationship in a dependent position since they enter into it out of need. Patients require the provider in order to gain access to resources and services to fulfill their needs. Rendering important services or 32

PAGE 45

providing valuable benefits in return confers [the provider] a claim to superior status. Reciprocating denies this claim but [patient] failure to reciprocate validates the claim, thus amplifying the power imbalance in the patient-provider relationship (Blau, 2002). Applying Emersons social exchange theory to the patient-provider interaction in general, and to this study in particular, it was hypothesized that the patient (actor A) paired with the provider (actor B) in a social relation meet in an instance of exchange called a transaction (the medical encounter). The patient initiates behavior X1 (expression of depressive symptoms) during the transaction that evokes stimulus feedback Y (expression of provider difficulty in the patient-provider relationship). If the transaction is to be a mutually rewarding and reinforcing exchange and behavior X1 (patient expression of depressive symptoms) is not rewarding to the provider, theory would predict that the provider will deliver help to the patient if and only if the patient behavior is rewarding to the provider. If the patient behavior X1 (expression of depressive symptoms) does not evoke a rewarding behavior from the provider, then through the conditioned reinforcement of feedback Y (expression of provider difficulty in the patient-provider relationship), the patient behavior will either (a) change to a form X2 which is reciprocated by the provider, or (b) the patient-provider relation will ultimately extinguish. Either way, the patient is likely to be dissatisfied with the transaction (medical encounter) if their behavior is not reinforced and rewarded by the provider and their needs are not met. Since each transaction is linked to a history and a future for the actors, the depressed patient enters into the next patient-provider encounter dissatisfied 33

PAGE 46

and labeled by the provider as difficult, and the patient-provider relationship is strained. Definitions The following definitions are specifically applied to concepts used in this study. Depressive Symptomatology The presence and severity of depressive symptoms reported by patients and measured by the Beck Depression Inventory-II (BDI-II). Primary Care The provision of integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context of family and community (Institute of Medicine, 1996). Provider-Perceived Difficulty The presence and intensity of difficulty in the patient-provider relationship as perceived by the provider and measured by the Difficult Doctor-Patient Relationship Questionnaire-10 Item (DDPRQ-10). Provider Medical Questioning The proportion of provider talk during the encounter comprised of medical questioning and measured by the Roter Interaction Analysis System (RIAS). It is a composite number derived by totaling the frequencies of (1) closed-ended provider questions directed to the patient about (a) medical condition, and (b) therapeutic regimen 34

PAGE 47

and (2) open-ended provider questions directed to the patient about (a) medical condition, and (b) therapeutic regimen divided by the total number of provider utterances during the encounter. Patient Medical Questioning The proportion of patient talk during the encounter comprised of medical questioning and measured by the RIAS. It is a composite number derived by totaling the frequencies of (1) all patient questions directed to the provider about (a) medical condition and (b) therapeutic regimen divided by the total number of patient utterances during the encounter. Provider Medical Information Giving The proportion of provider talk during the encounter comprised of medical information giving and measured by the RIAS. It is a composite number derived by totaling the frequencies of provider information giving directed to the patient about (a) medical condition, (b) therapeutic regimen, and (c) counseling medical/therapeutic divided by the total number of provider utterances during the encounter. Patient Medical Information Giving The proportion of patient talk during the encounter comprised of medical information giving and measured by the RIAS. It is a composite number derived by totaling the frequencies of patient information giving directed to the provider about (a) medical condition and (b) therapeutic regimen divided by the total number of patient utterances during the encounter. 35

PAGE 48

Provider Psychosocial Questioning The proportion of provider talk during the encounter comprised of psychosocial questioning and measured by the RIAS. It is a composite number derived by totaling the frequencies of (1) closed-ended provider questions directed to the patient about (a) lifestyle, and (b) psychosocial-feelings and (2) open-ended provider questions directed to the patient about (a) lifestyle, and (b) psychosocial-feelings divided by the total number of provider utterances during the encounter. Patient Psychosocial Questioning The proportion of patient talk during the encounter comprised of psychosocial questioning and measured by the RIAS. It is a composite number derived by totaling the frequencies of (1) all patient questions directed to the provider about (a) lifestyle, and (b) psychosocial-feelings divided by the total number of patient utterances during the encounter. Provider Psychosocial Information Giving The proportion of provider talk during the encounter comprised of psychosocial information giving and measured by the RIAS. It is a composite number derived by totaling the frequencies of provider information giving directed to the patient about (1) gives information-lifestyle, and (2) counsels-lifestyle/psychosocial divided by the total number of provider utterances during the encounter. Patient Psychosocial Information Giving The proportion of patient talk during the encounter comprised of psychosocial information giving and measured by the RIAS. It is a composite number derived by 36

PAGE 49

totaling the frequencies of patient information giving directed to the provider about (1) gives information-lifestyle, and (2) gives information-psychosocial divided by the total number of patient utterances during the encounter. Patient Satisfaction The degree of fulfillment of wishes, expectations, or needs with, or the pleasure derived from, the medical encounter as reported by patients and measured by the patient satisfaction survey. Hypotheses This study examined the influence of patient depressive symptomatology on the patient-provider relationship, patient-provider communication, and patient satisfaction with the primary care office visit by testing the following hypotheses: Hypothesis 1 As the severity of patient depressive symptoms increases, provider perception of difficulty in the patient provider relationship increases. Hypothesis 2 As the severity of patient depressive symptoms increases, patient satisfaction decreases. Hypothesis 3 As the severity of patient depressive symptoms increases, provider and patient medical questioning and information giving increases. 37

PAGE 50

Hypothesis 4 As the severity of patient depressive symptoms increases, provider and patient psychosocial questioning and information giving decreases. Hypothesis 5 As provider perception of difficulty in the patient-provider relationship increases, provider and patient medical questioning and information giving increases. Hypothesis 6 As provider perception of difficulty in the patient-provider relationship increases, provider and patient psychosocial questioning and information giving decreases. Hypothesis 7 As provider and patient medical questioning and information giving increases, patient satisfaction decreases. Hypothesis 8 As provider and patient psychosocial questioning and information giving increases, patient satisfaction increases. The following chapter describes the methods used in testing these hypotheses. 38

PAGE 51

Method Pilot Study A pilot study was conducted to determine the feasibility of the present study. The goals of the pilot study were to (1) evaluate the availability and accessibility of required resources to conduct the study and maintain HIPAA compliance at various practice sites, (2) evaluate the appropriateness, readability, and completion times of the survey questionnaires, (3) evaluate the logistics of audio-recording, and (4) examine the relationships between patient depressive symptoms, difficulty in the patient-provider relationship, patient-provider communication, and patient satisfaction. Four primary care providers from three practice sites in southwest Florida and 26 adult patients participated. Completion of the pilot study provided the essential data to confirm that the required resources at each practice site were available and accessible, the survey content appropriate, the process of data collection practical, and that the relationships of interest could be explored in the present study. The data were subjected to descriptive, correlation, and regression analyses. Ninety-two percent of patient subjects felt it was very or somewhat important their primary care provider address their emotional health needs. Findings were consistent with previous research indicating that depressed patients tend to be less satisfied with medical care and that the presence of mental disorders in patients is associated with increased provider-perceived difficulty in the patient-provider relationship. Although the findings lacked statistical significance, 39

PAGE 52

patient and provider questioning and provider information giving on psychosocial issues was associated with increased patient satisfaction, whereas both patient and provider questioning and information giving on medical issues was associated with lower levels of satisfaction. Patient information giving on psychosocial issues also associated with less patient satisfaction although interpretation of the data was limited. Scale reliabilities were acceptable with alpha coefficients ranging from 0.62 to 0.90. Results of the pilot study provided the essential findings to support continuing with the present research. The Study Human Subjects Protection The present study was reviewed and approved by the University of South Florida Institutional Review Board protocol #103075. Documents reflecting the approval of the initial and continuing review applications are located in Appendix A. Subjects Sample criteria. Provider subjects were recruited from multiple medical practice sites that are all part of a large multi-physician multi-specialty practice group in southwest Florida (Appendix B). All provider participants in this study met the following criteria: 1. English speaking; 2. Adult age 21 or over; 3. Medical or advanced practice nursing clinical practice in the primary care area of Internal Medicine, Family Practice, or Geriatrics; 40

PAGE 53

4. Voluntarily agreed to participate. Patient subjects were recruited from the same physician practice group as the provider subjects. All patient participants met the following criteria: 1. English speaking; 2. Adult age 21 or over; 3. Presented to the medical office for a medical encounter with their primary care provider; 4. History of at least one previous encounter with the provider (otherwise referred to as an established patient); 5. Not experiencing an emergent medical need; 6. Did not appear to be in a grossly agitated state, or exhibit gross verbal or physical behavior; 7. No evidence of impaired or questionable cognitive ability; 8. Voluntarily agreed to participate. Sample size estimate. The pilot study data provided information on the direction and magnitude of the associations between patients depressive symptoms, provider perception of difficulty in the patient-provider relationship, patient-provider communication, and patient satisfaction. A series of simple and multiple regression equations were used to analyze the pilot study data. The multiple regression of patient satisfaction onto 9 predictor variables, assuming power of .80, = .05, and based on an estimated effect size of R2 = .40, indicated that N = 60 was needed for this study. Using the adjusted R2 = .067 for this same regression equation and again assuming power of .80 41

PAGE 54

and = .05 indicated N = 114 would be sufficien t to detect statistical significance. Tabachnick and Fidell (2001) offer rules of thumb when determining required sample size. They suggest N 50 + 8m (where m is the number of independent variables) for testing the multiple correlation and N 104 + m for testing individual predictors, assuming a medium-size relationship between the independent variables and the dependent variable, = .05 and = .20 (p. 117). Their suggestions indicated that N 122 would be required for stability for testing the multiple regression and N 113 for testing individual predictors. Consideri ng these four pieces of information, projected costs, pilot study experience, and anticipating a 5% loss of data (due to audio recording failure, etc.), the original protocol sought to enroll 120 patient subjects and 12 provider subjects into this study. Measures Provider demographic questionnaire. This 11-item self-report survey documented routine demographic and practice information and self-assessed ratings of the providers knowledge, ability to diagnose, and ability to treat depression. It asked the provider to rate the impact of several frequently cited barriers to the diagnosis and treatment of depression by primary care providers (Appendix C). Patient demographic questionnaire. This 14-item self-report survey documented routine demographic information and containe d 5 brief questions relating to self-assessed health status, opinion on treatment of emotiona l health needs in primary care, and opinion on the effect of length of office visit on th eir perception of care received (Appendix D). 42

PAGE 55

Beck Depression InventorySecond Edition (BDI-II). The Beck Depression Inventory (BDI) is one of the most widely used instruments for evaluating depression. Like the BDI, the Beck Depression Inve ntorySecond Edition (BDI-II) is a 21-item selfreport measure designed to assess the presence and intensity of depressive symptoms in adults and adolescents aged 13 years and older (Beck, Steer, & Brown, 1996). Each item is a list of four statements arranged in increasing severity about a particular symptom of depression over the preceding two weeks. The BDI-II has a possible range of 0-63 and is scored so that a higher value indicates a greater degree of depressive symptomatology. Reliability of the BDI-II (coefficient = 0.92) is higher than the original BDI (coefficient = 0.86), and is in alignment with DSM-IV criteria (American Psychiatric Association, 1994). The BDI-II can be self-administered or read aloud by the examiner for individuals with reading difficulties or sensory deficits. The cut score threshold followed the BDI-II manual scoring guidelines to maximize sensitivity (Beck et al., 1996) (Appendixes E and F). The Difficult Doctor-Patient Relationship Questionnaire (DDPRQ-10). The Difficult Doctor-Patient Relationship Questionnaire-10 (DDPRQ-10) is a 10-item selfreport instrument designed to detect and measure difficulties in the doctor-patient relationship as perceived by the physician. The medical provider completes it after seeing the patient for the medical encounter. The 10-item DDPRQ was developed for use in the PRIME-MD 1000 Study (Spitzer et al., 1994). Each item is measured on a sixpoint Likert scale and the score is the sum of all 10 items after the values of reversed items have been corrected for direction. The DDPRQ-10 has a possible range of 10-60. 43

PAGE 56

Scoring for this study was reversed from the original DDPRQ-10 and was scored so that a higher value indicated a greater level of provider-perceived difficulty in the doctorpatient relationship. It has high face validity, and construct validity has established a strong association with the presence of mental disorders, multiple unexplained somatic symptoms, and abrasive personality style with physician-experienced difficulty in the doctor-patient relationship (Hahn, n.d.). The DDPRQ-10 captures 96% of the variance of the original 30-item instrument (DDPRQ-30) and had a Cronbachs = 0.88 in the original sample (Hahn, 2000) (Appendixes G and H). The Roter Interaction Analysis System (RIAS). The RIAS is a method of coding verbal patient-provider interaction during th e medical visit. Derived from social exchange theories related to interpersonal influence, problem solving, and reciprocity, the RIAS is the most widely used system of medical interaction assessment and has been used in over 125 communication studies in a multitude of health care settings (Roter, n.d.). Coding is done directly from audio or video recordings using RIAS software, eliminating the need to transcribe the audio record to written form. Coding is tailored to dyadic exchange specific to the medical encounter. The RIAS is applied to the smallest unit of expression to which a meaningful code can be assigned, defined as an utterance. All patient and provider dialogue is coded into mutually exclusive and exhaustive categories that reflect the content and context of the routine dialogue between patients and providers during the medical exchange. The RIAS is highly reliable when applied by trained coders and has proved to be clinically meaningful in a number of empirical studies (Sandvik et al., 2002). It has consistently shown an average of 0.85 inter-rater 44

PAGE 57

reliability for both patient and provider categories and clusters. It has demonstrated content and discriminant validity (Roter & Larson, 2002). Roter and colleagues provided the expert coding for the audio recordings collected in this study (Appendixes I and J). Patient satisfaction survey. The patient satisfaction survey is a 36-item self-report instrument that measured one global satisfaction item and five distinct and reliable aspects of patient satisfaction specific to the immediate medical encounter. It was adapted from the work of Bertakis et al. (1991). Each item is measured on a five-point Likert scale and the score is the sum of all 36 items after the values of reversed items have been corrected for direction. The survey has a possible range of 36-180 and was scored so that a higher value indicated greater satisfaction with the medical encounter. The five subscales revealed by factor analysis are: task-directed skill ( = 0.90), interpersonal skill ( = 0.87), attentiveness ( = 0.81), partnership ( = 0.76), and emotional support ( = 0.71). The average inter-scale correlation is 0.51 (range 0.290.65) (Bertakis et al., 1991). (Appendixes K and L). Procedure Recruitment. A recruitment letter describing the study was mailed to all eligible primary care providers within the provider practice group (Appendix M). A telephone call was placed to each providers office one week after the mailing of the letter seeking to arrange a meeting with the provider to discuss the study in detail. If the provider was interested in hearing more about the study, a telephone or personal meeting was scheduled. During the meeting, the investigator screened the provider for eligibility, 45

PAGE 58

presented all information contained in the informed consent documents, answered any questions, and asked for their voluntary participation. The intake receptionist at each medical office was asked to distribute a flyer (Appendix N) introducing the study to each adult patient who presented to the office on the day(s) of data collection as they signed in for their office visit. The flyer informed potential patient recruits that the details of the study would be provided to them in privacy once they were escorted back to the examination room. Once the office nurse/medical assistant brought each patient to the private examination room and finished their tasks, she/he then asked each individual if they were interested in learning more about the study. If the patient expressed interest in hearing more about the study, the nurse/medical assistant relayed this information to the investigator. The investigator then entered the examination room, screened each potential patient recruit for eligibility, presented all information contained in the informed consent and HIPAA documents, answered any questions, and asked for their voluntary participation. Informed consent process. Each provider, who was screened for eligibility and expressed interest in participating in the study, was given a printed copy of the informed consent and consent for audio taping documents to read and sign (Appendix O). The consent documents explained the purpose and plan of the study, benefits and risks of participation, confidentiality of records, voluntary participation, and contact person with telephone number to call with any additional questions. Once informed consent was secured, each provider subject completed the provider demographic questionnaire. Each 46

PAGE 59

provider was then given a copy of his/her signed consent documents and a mutually agreed upon date(s) for data collection was established. Each patient, who was screened for eligibility and expressed interest in participating in the study was given a printed copy of the informed consent, consent for audiotaping, and HIPAA authorization docum ents to read and sign (Appendixes P and Q). The informed consent documents explained the purpose and plan of the study, benefits and risks of participation, payment for participation, confidentiality of records, voluntary participation, and contact person with telephone number to call with any additional questions. Once informed consent was secured, each patient subject was given a copy of his/her signed documents. Voluntary consent was also required from all third party individuals who were going to be present in the examination room during the patient-provider encounter. If a third party individual declined voluntary participation in the study, the corresponding patient subject was excluded from participation. The informed consent process for third party individuals mirrored the process used for patient subjects. All participating third party individuals signed the same informed consent and consent for audiotaping documents as patients. Each was then given a copy of his/her signed documents. Data collection. Patient participants were asked to complete the patient demographic questionnaire and BDI-II while waiting for the provider to enter the examination room for the medical encounter. In the event that the subject was unable to complete the instruments prior to the time the provider entered the examination room, they were asked to complete them immediately following the patient-provider interaction 47

PAGE 60

and prior to administration of the post-encounter satisfaction survey. Subjects were informed that audio recording was about to begin and the digital voice recorder was turned on. The investigator then left the examination room to photocopy the signed consent documents. The investigator re-entered the examination room, collected the completed surveys from the subject, provided them with a copy of his/her signed consent documents, and then left the examination room. The BDI-II was scored immediately. Study protocol dictated that patient subjects who admitted to hopelessness and/or suicidal ideation (as measured by items 2 and/or 9 of the BDI-II) with a rating of 2 or 3 were to be withdrawn from further participation (Appendix R) and the provider notified promptly (Appendix S). Otherwise, providers were blinded to the patients BDI-II results, data collection proceeded and the patient-provider encounter was audio recorded using a digital voice recorder. When it was determined that the patient-provider encounter was finished, the digital voice recorder was turned off. In consultation with the nurse/medical assistant, the feasibility of having the patient subject remain in the examination room to complete the post-encounter questionnaire was determined. If time constraints prohibited the subject from remaining in the examination room, they were escorted to another area within the office to complete the post-encounter satisfaction survey. The subject was instructed to complete the questionnaire and place it in the envelope provided, seal the envelope, and give the envelope to the receptionist upon check out. Each subject was given written notification of his or her BDI-II results in an envelope (Appendix T). 48

PAGE 61

When it was evident the patient-provider encounter was complete and the provider exited the examination room, the provider was given the provider perception questionnaire (DDPRQ-10) with instructions to complete as soon as possible that day, place in the envelope provided, seal the envelope, and return the envelope directly to the investigator. The provider was given written notification of each patient subjects BDI-II results (Appendix U) after they returned the completed DDPRQ-10. The investigator remained at the office site until all questionnaires were returned and a debriefing session was held with the provider. Raw Data Management All survey responses, complete and incomplete, were entered into an SPSS program database (SPSS 10.0 for Windows, SPSS Inc.) using a personal computer with password protection to secure confidentiality. Once the data were entered into the database, the completed surveys were stored in a locked file cabinet. The audio recordings of the patient-provider encounters, collected by digital voice recorder, were transferred to the computer, converted to .wav file format using Sony Digital Voice Editor software (version 2.27, Sony Electronics, Inc.) and saved to CD-R. The voice files on the digital voice recorder were erased once the records were converted and successfully stored. The voice files (on CD), were sent in batches of 30 records to Dr. Roter at The Johns Hopkins University Bloomberg School of Public Health for RIAS coding throughout the data collection period. The CDs were sent via USPS priority mail with delivery and signature confirmation. When Dr. Roter and her research staff 49

PAGE 62

completed coding of the audio files, she forwarded the RIAS data as SPSS data files and returned all CDs via USPS. Analytic Plan Statistical analyses to test the proposed hypotheses were guided by Figure 1. A 2 2 2 within-subjects factorial model provided the analytic framework for examining eight verbal communication behaviors categorized by speaker (patient or provider), type of utterance (question or information giving), and content of utterance (medical or psychosocial talk). The data had a 2-level nested structure; the patientprovider encounters were nested within provi ders. The hierarchical structure of the hypothesized model revealed the importance of analyzing the data within context in order to elucidate relationships at more than one level. Thus, the data violated the assumption of independence that underlies traditional regression analysis. Multilevel modeling relaxes the independence assumption and allows for correlated error structures (Luke, 2004). This study was designed to examine how patient and provider characteristics influence the patient-provider encounter, providing theoretical justification for a multilevel approach to analysis. Because the patient-provider encounters were not independent, were clustered by providers, and were likely to exhibit correlated errors, multilevel modeling was also statistically justified. Results are presented in the following chapter. 50

PAGE 63

provider provider patient patient question infom e d p s ym e d p s yquestion info Patient Depressive Symptoms Patient Satisfaction Provider Difficulty in Relationship Figure 1. Model depicting the hypothesized relationships between patient depressive symptoms, provider perception of difficulty in the patient-provider relationship, patientprovider communication, and patient satisfaction. 51

PAGE 64

Results Subjects Provider Subjects Recruitment letters were mailed to all potential provider subjects in the multiphysician practice group whose medical or advanced practice nursing clinical practices were in the primary care areas of internal medicine, family practice, or geriatrics ( n=20). Forty-five percent of the providers did not return telephone messages after the mailing. Eleven providers including six physicians and five advanced registered nurse practitioners (ARNPs) expressed an initial interest in hearing the details of the study during a follow-up telephone inquiry. All 11 providers met eligibility criteria and voluntarily consented to participate. Detailed provider demographic information is presented in Table 1. The seven males and four females ranged in age from 31 to 62 years with a mean age of 43. All providers had been in practice for at least one year. Eighty-one percent reported having over six years of professional practice experience. They treated an average of 21 patients per day and 50 to 150 patients per week. Physicians and male providers reported more daily (24 and 22 vs. 19 and 18) and weekly (118 and 111 vs. 76 and 76) patient encounters than the ARNPs and female providers. Collectively they estimated that they considered 15% of their patient population as difficult (range 5 to 40%). 52

PAGE 65

Table 1 Characteristics of Provider Subjects By Provider Type (N=11) Variable MD (n=6) ARNP (n=5) %* Gender Female 0 4 36 Male 6 1 64 Age Mean 38.2 48.4 Range 31-44 33-62 Race American Indian/Alaskan Native 0 0 0 Asian 0 0 0 Black/African American 0 0 0 Native Hawaiian/Pacific Islander 0 0 0 White 6 5 100 Ethnicity Hispanic or Latino 2 0 18 Not Hispanic or Latino 4 5 81 Type of Clinical Practice Internal Medicine 1 1 18 Family Practice 3 4 63 Geriatrics 2 0 18 Number of Years in Practice Less than one year 0 0 0 1-5 years 1 1 18 6-10 years 3 1 36 11-20 years 2 1 27 Over 20 years 0 2 18 53

PAGE 66

Table 1 (Continued) Average Number of Patients Seen Per Day 23.7 18.9 Per Week 117.5 75.5 Note. Values are rounded and may not total 100%. 54

PAGE 67

The providers rated the impact of various internal (self) and external factors on their ability to diagnose and treat depression in their primary care patients (Table 2). They ranked time constraints and multiple competing demands as having the greatest negative impact on their ability to provide depression care to their patients. Patient reluctance to go to mental health specialists was ranked second. Table 2 Provider Ratings of Various Aspects of Depression Care Delivery (N=11) Variable M Range Self-Ratings* Knowledge of depression 8.1 7-10 Ability to diagnose depression 7.7 6-10 Ability to treat depression 7.5 5-10 Influence of Outside Factors** Time constraints and multiple competing demands 9.0 6-10 Patients dont provide enough information 6.0 3-9 Lack of availability/access to mental health specialists 7.7 5-10 Patient reluctance to go to mental health specialists 8.6 6-10 Note. Based on scale of 0-10 with rating of 10 indicating high proficiency. ** Based on scale of 0-10 with rating of 10 indicating great influence. 55

PAGE 68

Patient Subjects One hundred and fifty-two potential patient subjects expressed an interest in hearing details of the study and were subsequently approached directly by the investigator to discuss possible participation. Twenty-nine patients either declined to participate or failed to meet the eligibility criteria for enrollment. The remaining 123 patients met inclusion criteria and voluntarily consented to participate. All patient subjects continued to meet inclusion criteria based upon review of their BDI-II responses and 123 patient-provider encounters were observed. The number of patient encounters per provider ranged from 7-16 with a mean of 11.2. No patient or provider subjects withdrew or were withdrawn from the study during or after data collection. Detailed patient demographic information is presented in Table 3. The patient subjects were mostly Caucasian (98%) older adults (mean = 63 years, median = 67 years). Seventy-five percent were 50 years of age or older and one-fourth were over the age of 76. Over 40% listed their occupation as professional and reported a college degree or graduate degree as their highest level of education. Slightly more than half of the subjects were retired. According to BDI-II scoring criteria 77% percent reported experiencing minimal depressive symptoms (0-13), 13% mild depression (14-19), 7% moderate depression (20-28), and 3% extrem e depression (29-63). The majority of patients (88%) felt that (in general) it was very or somewhat important that their primary care provider addressed their emotional health needs. Nearly all (96%) indicated that the amount of time they spend with their primary care primary care provider was at least a somewhat important factor when evaluating quality of care. 56

PAGE 69

Table 3 Characteristics of Patient Subjects (N=123) Variable Female (n=71) Male (n=52) %* Age Mean 63.1 62.2 Range 21-90 21-89 Race American Indian/Alaska Native 1 0 0 Asian 1 0 0 Black/African American 0 2 2 Native Hawaiian/Pacific Islander 0 0 0 White 69 50 98 Ethnicity Hispanic or Latino 3 4 6 Not Hispanic or Latino 60 43 84 Occupation Professional 23 32 45 Technical 3 6 7 Clerical 17 2 15 Laborer 1 6 6 Housewife 21 1 18 Other 5 5 8 Retired Yes 42 26 55 No 25 26 41 Highest Level of Education Grade School 1 1 2 High School 31 19 41 Technical School 12 7 15 College Degree 22 17 32 Graduate Degree 5 7 10 57

PAGE 70

Table 3 (Continued) Marital Status Married 35 32 54 Single 7 7 11 Divorced 9 7 13 Widowed 20 4 20 Note. Values are rounded and may not total 100%. 58

PAGE 71

Data Management The data were entered into an SPSS program database. All variables were first subjected to univariate descriptive analysis with graphic representations and were examined for accuracy of data entry, out-of-range values, and plausible means and standard deviations. Suspicious values in data entry were compared to the original data and corrected. Means and standard deviations were reasonable. Preliminary BDI-II, DDPRQ, and satisfaction scores were tabulated for each case by summing individual scale item responses. The audio recordings of two patient encounters were inaudible and unusable for analysis and those cases were deleted. Eight composite communication variables categorized by speaker (provider or patient), type of utterance (question or information giving), and content of utterance (medical or psychosocial talk) were computed from the RIAS data file. Suspicious values in data entry were compared to the original audio records and corrected. The BDI-II, DDPRQ, and satisfaction scale items were assessed for multicollinearity and singularity by examining bivariate correlation matrices. The correlations ranged from r = .02 to .70 (BDI-II), r = .09 to .75 (DDPRQ), and r = .00 to .79 (satisfaction). The data were examined next to assess the amount and distribution of missing values. The DDPRQ and RIAS variable data files were complete. Ten of the 21 items on the BDI-II had missing data. Nine of those items were missing one or two values (0.8 1.6%). The tenth item had a total of eight missing values (6.5%). Six subjects did not answer any items on the satisfaction survey and those six cases were 59

PAGE 72

deleted, leaving 115 cases for further analysis. Thirty-two of the 36 items on the satisfaction survey had from 0.8 to 5.2% missing values. Non-randomly missing values can affect the generalizability of results (Tabachnick & Fidell, 2001). Therefore, the amount and pattern of missing data were examined. Independent sample t -tests (two-tailed, 95% C.I.) were used to determine if the patient subjects who failed to respond to at least one of the 21 BDI-II items ( n = 15) differed from subjects who responded to all items ( n = 100). No significant differences were found between the two groups on any of the demographic variables, practice site, sequence of participation, raw BDI-II or DDPRQ scores. Similar t-tests were also used to determine if the patient subjects who failed to respond to at least one of the 36 satisfaction items ( n = 24) differed from subjects who responded to all items ( n = 91). No significant differences were found between the two groups on any of the demographic variables, practice site, sequence of participation, raw DDPRQ or satisfaction scores. Because the deletion of all cases with missing values would have resulted in a substantial loss of subjects, various options were explored to determine the best method for handling the missing data. Mean item values for the BDI-II and satisfaction scales were computed for each case based upon the number of scale items the subject answered. Total BDI-II and satisfaction scores were then tabulated by multiplying the case mean values by the total number of items in each scale. This method standardized the survey scores and enabled all cases to be retained for further analysis. 60

PAGE 73

To improve interpretability of the data, each communication variable (previously defined as a proportion of total utterances) was adjusted to reflect the percentage of either all patient talk or all provider talk. The 11 primary study variables (BDI-II, DDPRQ, provider medical questioning, patient medical questioning, provider medical information giving, patient medical information giving, provider psychosocial questioning, patient psychosocial questioning, provider psychosocial information giving, patient psychosocial information giving, and patient satisfaction) were assessed for univariate and multivariate normality through various SPSS programs. Descriptive statistics are presented in Table 4. The BDI-II, DDPRQ, patient medical questioning, provider psychosocial questioning, patient psychosocial questioning, provider psychosocial information giving, and patient psychosocial information giving distributions exhibited moderate positive skewness (2.6 t 8.02, df = 114, p < .01). Patient satisfaction was moderately negatively skewed ( t = 4.21, df = 114, p < .01). BDI-II, DDPRQ, patient medical questioning, provider psychosocial questioning, patient psychosocial questioning, and provider psychosocial information giving distributions were leptokurtic ( t > 2.6, df = 114, p < .01). Four cases with univariate outliers were identified (> 4.0 SDs above the mean). One additional case was identified through Mahalanobis distance as a multivariate outlier with p < .001. The cases were examined to verify that they were properly part of the intended sample population as skewed distributions can create problems insofar as they violate the assumption of normality that underlies many other statistical tests. Therefore, 61

PAGE 74

Table 4 Descriptive Statistics of Study Variables (N=115) Variable M SE Mdn SD BDI-II 9.18 0.72 7.35 7.75 DDPRQ 18.46 0.71 17.00 7.62 SAT 159.40 1.79 167.00 19.17 DRMEDQ* 9.97 0.50 9.41 5.39 PTMEDQ* 2.62 0.23 2.22 2.50 DRMEDINFO 33.41 1.23 33.16 13.16 PTMEDINFO* 36.61 1.27 35.90 13.60 DRPSYQ* 2.33 0.23 1.79 2.47 PTPSYQ* 0.55 0.08 0.00 0.87 DRPSYINFO* 5.20 0.51 3.51 5.51 PTPSYINFO* 16.81 1.20 14.11 12.83 Note. Variables are reported as percentage of either all provider talk or all patient talk. BDI-II = Beck Depression Inventory-II. DDPRQ = Difficult Doctor-Patient Relationship Questionnaire-10. SAT = patient satisfac tion survey. DRMEDQ = provider medical questioning. PTMEDQ = patient medical questioning. DRMEDINFO = provider medical information giving. PTMEDINFO = patient medical information giving. DRPSYQ = provider psychosocial questioning. PTPSYQ = patient psychosocial questioning. DRPSYINFO = provider psychosocial information giving. PTPSYINFO = patient psychosocial information giving. 62

PAGE 75

options for dealing with the non-normality of the variables were explored. Deletion of the outlier cases would reduce statistical power. Retaining the cases would increase the risk of Type I and Type II errors and would limit the generalizability of the results. Transformation of variables when they are skewed to the same moderate extent hinders interpretation of the results, especially when the scale(s) is meaningful or widely used (Tabachnick & Fidell, 2001). Recognizing the potential impact that the outlier cases could have on the interpretation of results, all subsequent analyses of the data were conducted in parallel sets. The data were examined both with ( N =115) and without the outlier cases (N =110) and the results were compared. There were no significant differences in results. Therefore, to maintain adequate statistical power and study integrity, these five cases were left in the final analysis. For clarity, the results presented in this chapter refer to the final sample ( N =115) unless specified otherwise. Reliability Analysis of Measures The BDI-II, DDPRQ, and patient satisfaction scales were evaluated for internal consistency. Scale and subscale reliabilities were acceptable with alpha coefficients ranging from 0.89 to 0.95 (Tables 5 and 6). Analyses were conducted on the raw data and then repeated after missing values were replaced by case mean substitution. Ten percent of the audio records ( n=12) were randomly selected and coded by two RIAS experts to assess for errors or bias in coding. Inter-rater reliability estimates were computed for the individual and composite RIAS coded communication variables using Pearsons r Correlation coefficients for the eight composite variables were all positive and ranged from 0.72 to 0.96 (Tables 7 and 8). 63

PAGE 76

Table 5 Reliability Analyses of Scales (N=123) Scale Number of n# M SD Items in Scale BDI-II 21 100 .3444 .1584 .9131 .9032 DDPRQ-10 10 115 .4660 .1673 .8972 .8972 SAT 36 91 .3611 .1868 .9532 .9483 Note. # Listwise deletion. Cronbachs alpha after missing items replaced using case mean substitution ( N =115). BDI-II = Beck Depression Inventory-II. DDPRQ = Difficult Patient-Provider Relationship Questionnaire10. SAT = patient satisfaction survey. 64

PAGE 77

Table 6 Reliability Analysis of Patient Satisfaction Survey Subscales (N=123) Scale Number of n# M SD Items in Scale Task-directed skill 10 105 .4828 .1694 .9032 .9006 Interpersonal skill 8 104 .2962 .1788 .7710 .7762 Attentiveness 5 107 .3290 .1140 .7103 .7070 Partnership 9 100 .3957 .1772 .8549 .8497 Emotional support 3 106 .3283 .2557 .5945 .6090 Note. # Listwise deletion. Cronbachs alpha after missing items replaced using case mean substitution ( N =115). 65

PAGE 78

Table 7 Inter-rater Reliability Estimates of Individual and Composite RIAS Coded Variables (Medical) Using Pearsons r (n=12) Direction Variable Variable Description r of Speech Code D Pa cmedd closed question medical .91 ctherd closed question therapeutic .92 omedd open question medical .70 otherd open question therapeutic .11** medqued composite: question asking medical .84$ D Pa imedd gives information medical .88 itherad gives information therapeutic .93 cnlmdd counsels medical/therapeutic .81 infomedd composite: information-giving medical .87 P Db qmedp all questions medical .63* qtherp all questions therapeutic .81 medquep composite: question asking medical .81$ P Db imedp gives information medical .99 itherp gives information therapeutic .89 infomedp composite: information-giving medical .94 Note. a provider talk directed to patient. b patient talk directed to provider. mean 1.0-2.0. ** mean < 1.0. $ composite value includes only variables with mean > 2.0. 66

PAGE 79

Table 8 Inter-rater Reliability Estimates of Individual and Composite RIAS Coded Variables (Psychosocial) Using Pearsons r (n=12) Direction Variable Variable Description r of Speech Code D Pa clsd closed question lifestyle .96 cpsd closed question psychosocial .87** olsd open question lifestyle .32** opsd open question psychosocial .73** psyqued composite: question asking psychosocial .96$ D Pa ilsd gives information lifestyle .72 cnllsd counsels lifestyle/psychosocial .35* infopsyd composite: information-giving psychosocial .72$ P Db qlsp all questions lifestyle .41** qpsp all questions psychosocial .67** psyquep composite: question asking psychosocial $ P Db ilsp gives information lifestyle .94 ipsp gives information psychosocial .89 infopsyp composite: information-giving psychosocial .92$ Note. a provider talk directed to patient. b patient talk directed to provider. mean 1.0-2.0. ** mean < 1.0. $ composite value includes only variables with mean 2.0. 67

PAGE 80

Preliminary Analysis The data had a 2-level nested structure with patient-provider encounters (level-1) nested within providers (level-2). According to protocol the data were to be analyzed using hierarchical linear modeling multiple regression. Before the advanced modeling techniques were applied, the data were first subjected to bivariate correlation and standard multiple regression analysis (OLS) to determine if the multilevel relationships proposed in the original hypotheses were present. An alpha level of .05 was used for all statistical tests. The zero-order correlations among the variables are presented in Table 9. Tables 10 and 11 provide a summary of the regression analyses predicting (1) provider perception of difficulty in the patient-provider relationship, (2) patient-provider communication, and (3) patient satisfaction. 68

PAGE 81

Table 9 Intercorrelations Between Communication Variables, Patient Depressive Symptoms, Provider-Perceived Difficulty, and Patient Satisfaction (N=115) Variable v1 v2 v3 v4 v5 v6 v7 v8 BDI-II DDPRQ SAT v1 1.00 v2 -.09 1.00 v3 .04 .52** 1.00 v4 .47** -.04 .18* 1.00 v5 -.19* -.19* -.42** -.42** 1.00 v6 -.27** .14 .03 -.29** .08 1.00 v7 -.39** -.03 -.10 -.43** .42** .35** 1.00 v8 -.32** -.21* -.44** -.56** .70** .20* .38** 1.00 BDI-II -.00 -.12 -.21* -.09 .23* -.07 -.04 .12 1.00 DDPRQ .11 .11 .21* .11 -.13 -.08 -.13 -.13 .31 1.00 SAT -.19* -.11 -.30 -.00 .12 -.05 .17 .09 -.32 -.41 1.00 Note. v1 = provider medical questioning. v2 = patient medical questioning. v3 = provider medical information giving. v4 = patient medical information giving. v5 = provider psychosocial questioning. v6 = patient psychosocial questioning. v7 = provider psychosocial information giving. v8 = patient psychosocial information giving. BDI-II = Beck Depression Inventory-II. DDPRQ = Difficult Doctor-Patient Relationship Questionnaire 10. SAT = patient satisfaction. *p .05. **p .005. p .001. 69

PAGE 82

Table 10 Summary of OLS Regression Analysis Predicting Provider-Perceived Difficulty and Patient-Provider Communication (N=115) Predictor Outcome B SE B BDI-II DDPRQ 0.302** 0.088 0.307** Provider medical questioning -0.025 0.069 -0.035 Patient medical questioning -0.056 0.031 -0.174 Provider medical information -0.516** 0.157 -0.304** Patient medical information -0.235 0.172 -0.134 Provider psychosocial questioning 0.095* 0.030 0.297* Patient psychosocial questioning -0.005 0.011 -0.048 Provider psychosocial information -0.001 0.070 -0.001 Patient psychosocial information 0.289 0.161 0.174 DDPRQ Provider medical questioning 0.083 0.070 0.117 Patient medical questioning 0.054 0.032 0.165 Provider medical information 0.524** 0.160 0.303** Patient medical information 0.263 0.175 0.147 Provider psychosocial questioning -0.073* 0.031 -0.224* Patient psychosocial questioning -0.007 0.011 -0.062 Provider psychosocial information -0.097 0.071 -0.134 Patient psychosocial information -0.303 0.164 -0.180 Note. BDI-II = Beck Depression Inventory-II. DDPRQ = Difficult Doctor-Patient Relationship Questionnaire-10. p < .05. ** p .001. 70

PAGE 83

Table 11 Summary of OLS Regression Analysis Predicting Patient Satisfaction (N=115) Predictor B SE B BDI-II -0.979** 0.209 -0.396** Provider medical questioning -0.831* 0.337 -0.234* Patient medical questioning 0.526 0.741 0.069 Provider medical information -0.635** 0.153 -0.436** Patient medical information 0.211 0.153 0.150 Provider psychosocial questioning 0.727 0.940 0.094 Patient psychosocial questioning -2.650 1.960 -0.120 Provider psychosocial information 0.467 0.351 0.134 Patient psychosocial information -0.182 0.191 -0.122 Note. p < .05. ** p .001. 71

PAGE 84

Hypothesis 1 As the severity of patient depressive symptoms increases, provider perception of difficulty in the patient-provider relationship increases. The correlation and regression coefficients revealed a significant positive relationship ( r = .31, p = .001) ( B = 0.302, p = .001), indicating that as the severity of patient depressive symptoms increased, provider perception of difficulty in the patient-provider relationship increased. The results supported Hypothesis 1. Hypothesis 2 As the severity of patient depressive symptoms increases, patient satisfaction decreases. The correlation and regression coefficients demonstrated a significant inverse relationship ( r = -.32, p = .001) ( B = -0.979, p = .000), indicating that as the severity of patient depressive symptoms increased, patient satisfaction decreased. The results supported Hypothesis 2. Hypothesis 3 As the severity of patient depressive symptoms increases, provider and patient medical questioning and information giving increases. This hypothesis was not supported. Counter to the predicted direction of the relationships, as the severity of patient depressive symptoms increased, provider medical questioning, patient medical questioning, and patient medical informati on giving decreased although these findings lacked statistical significance. However, as the severity of patient depressive symptoms increased, there was a significant decrease in provider medical information giving 72

PAGE 85

( r = -.21, p = .024) ( B = -0.516, p = .001). This finding disconfirmed one of the relationships predicted in Hypothesis 3. Hypothesis 4 As the severity of patient depressive symptoms increases, provider and patient psychosocial questioning and information giving decreases. This hypothesis was not supported. Although lacking statistical significance, as the severity of patient depressive symptoms increased, patient psychosocial questioning and provider psychosocial information giving decreased, and, counter to the predicted direction of the relationship, patient psychosocial information giving increased. However, as the severity of patient depressive symptoms increased, there was a significant increase in provider psychosocial questioning ( r = .23, p = .014) ( B = 0.095, p = .002). This finding disconfirmed one of the relationships predicted in Hypothesis 4. Hypothesis 5 As provider perception of difficulty in the patient-provider relationship increases, provider and patient medical questioning and information giving increases. As predicted, the correlation and regression coeffi cients were all positive although the results supported only one of the hypothesized relationships. As provider perception of difficulty in the patient-provider relationship increased, provider medical questioning, patient medical questioning, and patient medical information giving increased although these results lacked significance. However, as provider perception of difficulty in the patient-provider relationship increased, provider medical information giving increased 73

PAGE 86

significantly ( r = .21, p = .024) ( B = 0 .524, p = .001). The results partially supported Hypothesis 5. Hypothesis 6 As provider perception of difficulty in the patient-provider relationship increases, provider and patient psychosocial questioning and information giving decreases. As predicted, the coefficients all demonstrated inverse relationships, indicating that as provider perception of difficulty in the patie nt-provider relationship increased, provider and patient psychosocial questioning and information giving decreased. Although the bivariate correlation between provider difficulty and provider psychosocial questioning lacked statistical significance ( r = -.13, p = .156), the regression coefficient was significant ( B = -0.073, p = .019) The results partially supported Hypothesis 6. Hypothesis 7 As provider and patient medical questioning and information giving increases, patient satisfaction decreases. The correlation and regression coefficients were inconsistent with respect to the hypothesized relationships between patient medical questioning and information giving and satisfaction. However, as provider medical questioning ( r = -.19, p = .045) ( B = -0.831, p = .015) and provider medical information giving increased ( r = -.30, p = .001) ( B = -0.635, p = .000), there was a corresponding significant decrease in patient satisfaction. The results partially supported Hypothesis 7. Hypothesis 8 As provider and patient psychosocial questioning and information giving increases, patient satisfaction increases. These relationships did not reach a level of 74

PAGE 87

significance although the results indicated that as provider psychosocial questioning and provider psychosocial information giving increased, patient satisfaction increased. In addition, and counter to what was predicted, as patient psychosocial questioning increased, patient satisfaction decreased. The correlation and regression coefficients were inconsistent with respect to the hypothesized relationship between patient psychosocial information giving and patient satisfaction. The results failed to support Hypothesis 8. In addition to testing the proposed hypotheses, this study was designed to also explore (1) the potential mediating effect of provider-perceived difficulty in the patientprovider relationship on the association between patients depressive symptoms and patient-provider communication during the medical encounter and (2) the potential mediating effect of patient-provider communi cation on the association between patients depressive symptoms and patient satisfaction with the medical encounter. In general, a given variable may be said to function as a mediator to the extent that it accounts for the relation between the predictor and the criterion (Baron & Kenny, 1986). To establish mediation, it must be demonstrated that (1) the predictor variable is correlated with the outcome, (2) the predictor variable is correlated with the mediator, and (3) the mediator affects the outcome variable when controlling for the predictor variable in establishing the effect of the mediator on the outcome. The correlation analyses offered preliminary support to consider that (1) provider-perceived difficulty in the patient-provider relationship might function as a mediator between patients depressive symptoms (predictor variable) and (a) provider medical information giving and (b) provider 75

PAGE 88

psychosocial questioning (outcome variables) and (2) provider medical information giving might mediate the relationship between patients depressive symptoms (predictor variable) and patient satisfaction (outcome variable). The data were next examined to assess if the give and take dialog between patients and providers was associated with patient satisfaction. Four interaction terms specific to the patient-provider dyadic exchange were computed by centering each of the main effects variables and computing the product of these values: (1) provider medical questioning patient medical information giving, (2) patient medical questioning provider medical information giving, (3) provider psychosocial questioning patient psychosocial information giving, and (4) patient psychosocial questioning provider psychosocial information giving. Centering the main effects so that each variable had a mean of zero prior to forming the interaction terms reduced potential problems of multicollinearity (Tabachnick & Fidell, 2001). The interaction terms were examined using SPSS correlation and OLS regression. There were no significant correlations between any of the interaction terms and patient satisfaction. Forward entry regression was employed to determine if the addition of the interaction terms improved prediction of patient satisfaction beyond that afforded by BDI-II scores and the eight original communication variables. Addition of the interaction terms to the initial model did not significantly improve R2. Thus, given that the interaction terms were not included in the originally hypothesized model and failed to enhance the prediction of patient satisfaction, they were dropped from subsequent analysis. 76

PAGE 89

In summary, bivariate correlation and standard regression analysis of the data provided preliminary support for Hypothesis 1 and Hypothesis 2 and partial support for Hypothesis 5, Hypothesis 6, and Hypothesis 7. Hypothesis 3 and Hypothesis 4 were partially disconfirmed. The results failed to demonstrate a need for further examination of the remainder of the originally hypothesized relationships. Hierarchical Linear Modeling The simple correlation model demonstrated seven significant predictive relationships among the study variables. However, since traditional OLS regression analysis assumes independent observations, the preliminary results were interpreted with caution. To control for the nested structure of the data, the significant relationships demonstrated in the preliminary OLS regression analysis were subsequently analyzed in Hierarchical Linear Modeling (HLM version 6.02, Scientific Software International, Inc.) using restricted maximum likelihood estimation and an alpha level of .05. This section reviews the multilevel modeling approach used and presents results of these analyses. Level-1 predictors were group mean-centered. The results of the hierarchical analysis are presented in Table 12. 77

PAGE 90

Table 12 Hierarchically Modeled Estimates of Fixed () and Random () Effects (Slopes) (N=115) Predictor Outcome BDI-II DDPRQ 0.215422 0.04464 Provider medical information giving 0.366190 0.11127 Provider psychosocial questioning 0.076276 0.02404** Patient satisfaction -1.096203* 1.14247** DDPRQ Provider medical information giving 0.504637* 0.09080 Provider psychosocial questioni ng -0.050093 0.00200 Provider medical questioning Patient satisfaction -0.193769 0.20231 Provider medical information giving Patient satisfaction -0.488496** 0.00516 Note. Corresponds to the gamma ( ) coefficients and corresponds to the mu ( ) estimates in the multilevel modeling equations reported in th e text. BDI-II = Beck Depression Inventory-II. DDPRQ = Difficult Doctor-Patient Relationship Questionnaire-10. p < .05. ** p .005. 78

PAGE 91

Unconstrained Model (Null) To measure the proportion of variance in patient satisfaction accounted for by providers (level-2 units), the intraclass correlation coefficient (ICC) was estimated. The unconstrained model, with no level-1 or level-2 predictors was specified by the following equation: SATij = 00 + 0 j + rij. In this equation, SATij = the estimated satisfaction score for a particular patient within a particular provider, 00 = the grand mean of patient satisfaction across all patient subjects, 0 j = variability in patient satisfaction between providers, and rij = variability in patient satisfaction between patients within providers (random error). The ICC was .26, indicating that 26% of the variance in patient satisfaction was accounted for by providers. This substantial ICC demonstrated a main effect of provider on patient satisfaction in the study sample and provided empirical justification for a multilevel approach to analysis. Hypothesis 1 As the severity of patient depressive symptoms increases, provider perception of difficulty in the patient-provider relationship increases. This hypothesis was tested by regressing DDPRQ scores onto BDI-II scores. The mixed-effects model was specified by the following equation: DDPRQij = 00 + 10 (BDI-II)ij + 0 j + 1 j (BDI-II)ij + rij. In this equation, 00 = the grand mean of DDPRQ scores across all subjects, 10(BDI-II)ij = the effect (slope) of BDI-II scores for the level-2 unit (provider) and 1 j(BDI-II)ij = between provider variability in the slopes. A non-significant, positive relationship was 79

PAGE 92

observed ( B = 0.215422, SE B = 0.123324, p = 0.111). In contrast to the OLS regression analysis findings, the hypothesis that patient depressive symptoms predict providerperceived difficulty in the patient-provider relationship was not supported. Modeling revealed significant variability between providers in terms of their DDPRQ scores ( 2 = 20.71774, df = 10, p = .023). A graphical summary of the distributions of scores for the 11 providers is displayed in Figure 2. There was no significant difference in slopes between providers (2 = 12.14437, df = 10, p = .275) (Figure 3). A large random effects variance component for level-1 (49.17215) demonstrated substantial un-modeled variability in DDPRQ scores (random error). 80

PAGE 93

9.00 19.68 30.35 41.03 51.70DDPRQ 0 13.00 Figure 2. Distributions of provider DDPRQ scores adjusted for patient depressive symptoms. 81

PAGE 94

-9.18 0.32 9.82 19.32 28.82 14.34 19.66 24.98 30.30 35.61BDI-II ScoresDDPRQ Scores Figure 3. Slopes of provider-perceived difficulty as a function of patient depressive symptoms for each provider. 82

PAGE 95

Hypothesis 2 As the severity of patient depressive symptoms increases, patient satisfaction decreases. This hypothesis was tested by regressing patient satisfaction onto BDI-II scores (see also Hypothesis 7). The mixed-effects model was specified by the following equation: SATij = 00 + 10 (BDI-II)ij + 20 (DRMEDQ)ij + 30 (DRMEDINFO)ij + 0 j + 1 j (BDI-II)ij + 2 j (DRMEDQ)ij + 3 j (DRMEDINFO)ij + rij. In this equation, 20 (DRMEDQ)ij = the effect of provider medical questioning, 30 (DRMEDINFO)ij = the effect of provider medical information giving, 2 j (DRMEDQ)ij = variability between providers in the effect of provider medical questioning, and 3 j (DRMEDINFO)ij = variability between providers in the effect of provider medical information giving. A significant, inverse relationship between the severity of patient depressive symptoms and patient satisfaction was observed ( B = -1.096203, SE B = 0.394580, p = .020). The hypothesis that the severity of patient depressive symptoms predicts patient satisfaction was supported and confirmed the findings of the OLS regression analysis. There was significant variability between providers both in terms of their patients satisfaction ( 2 = 66.97756, df = 10, p = .000) (Figure 4) and slope of BDI-II (2= 25.49043, df = 10, p = .005) (Figure 5). A large random effects variance component for level-1 (185.76746) demonstrated substantial unmodeled variability in patient satisfaction (random error). 83

PAGE 96

86.40 114.30 142.20 170.10 198.00Patient Satisfaction 0 13.00 Figure 4. Distributions of patient satisfaction scores adjusted for patient depressive symptoms, provider medical questioning, and provider medical information giving. 84

PAGE 97

0 7.50 15.00 22.50 30.00 100.00 115.00 130.00 145.00 160.00BDI-II ScoresPatient Satisfaction Figure 5. Slopes of patient satisfaction as a function of patient depressive symptoms for each provider adjusted for provider medical questioning and provider medical information giving. 85

PAGE 98

Hypothesis 3 As the severity of patient depressive symptoms increases, provider and patient medical questioning and information giving increases. Preliminary analysis significantly disconfirmed the hypothesized relationship between patient depressive symptoms and provider medical information giving. This relationship was further examined by regressing provider medical information gi ving onto BDI-II scores (see also Hypothesis 5). The mixed-effects model was specified by the following equation: DRMEDINFOij = 00 + 10 (BDI-II)ij + 20 (DDPRQ)ij + 0 j + 1 j (BDI-II)ij + 2 j (DDPRQ)ij + rij. In this equation, DRMEDINFOij = provider medical information giving, 20 (DDPRQ)ij = the effect of provider-perceived difficulty, and 2 j (DDPRQ)ij = variability between providers in the effect of provider-perceived difficulty in the patient-provider relationship. A non-significant, inverse relationship was observed ( B = -0.366190, SE B = 0.180610, p = .070). In contrast to the OLS re gression analysis findings, the hypothesis that patient depressive symptoms predict provider medical information giving during the encounter was not supported. There was significant variability among providers in terms of medical information giving during the encounter ( 2 = 65.89477, df = 10, p = .000). A graphical summary is displayed in Figure 6. There was no significant difference in slopes between providers ( 2 = 14.67536, df = 10, p = .144) (Figure 7). A large random effects variance component for level-1 (102.87816) demonstrated substantial un-modeled variability in provider medical information giving (random error). 86

PAGE 99

0 13.00 2.25 20.48 38.71 56.94 75.17 DMEDINFO Figure 6. Distributions of provider medical information giving adjusted for patient depressive symptoms and provider-perceived difficulty. 87

PAGE 100

-9.18 0.32 9.82 19.32 28.82 30.27 37.93 45.58 53.24 60.89BDI-II ScoresProvider MED Information Figure 7. Slopes of provider medical information giving as a function of patient depressive symptoms for each provider ad justed for provider-perceived difficulty. 88

PAGE 101

Hypothesis 4 As the severity of patient depressive symptoms increases, provider and patient psychosocial questioning and information giving decreases. Preliminary analysis failed to support this hypothesis. However, the results significantly disconfirmed the relationship between patient depressive symptoms and provider psychosocial questioning. This relationship was further examined by regressing provider psychosocial questioning onto BDI-II scores (see also Hypothesis 6). The mixed-effects model was specified by the following equation: DRPSYQij = 00 + 10 (BDI-II)ij + 20 (DDPRQ)ij + 0 j + 1 j (BDI-II)ij + 2 j (DDPRQ)ij + rij. In this equation, DRPSYQij = provider psychosocial questioning. A non-significant, positive relationship was observed ( B = 0.076276, SE B = 0.059205, p = .202). In contrast to the OLS regression analysis fi ndings, the hypothesis that patient depressive symptoms predict provider psychosocial questioning was not supported. However, the slope of depressive symptoms, as demonstrated in Figure 8, varied significantly among providers ( 2 = 39.12708, df = 10 p = .000). A large random effects variance component for level-1 (4.44438) demonstrated substantial un-modeled variability in provider psychosocial questioning (random error). 89

PAGE 102

0 2.50 5.00 7.50 10.00Provider PSY Questioning 5.82 13.32 20.82 28.32BDI-II Scores Figure 8. Slopes of provider psychosocial questioning as a function of patient depressive symptoms for each provider adjusted for provider-perceived difficulty. 90

PAGE 103

Hypothesis 5 As provider perception of difficulty in the patient-provider relationship increases, provider and patient medical questioning and information giving increases. Preliminary analysis supported the relationship between provider perception of difficulty in the patient-provider relationship and provider medi cal information giving. This relationship was examined by regressing provider medical information giving onto DDPRQ scores (see also Hypothesis 3). The mixed-effects model was specified by the following equation: DRMEDINFOij = 00 + 10 (BDI-II)ij + 20 (DDPRQ)ij + 0 j + 1 j (BDI-II)ij + 2 j (DDPRQ)ij + rij. In this equation, DRMEDINFOij = provider medical information giving. A significant, positive relationship was observed ( B =0 .504637, SE B = 0.171022, p = .015) (Figure 9) and confirmed the findings of the OLS regre ssion analysis. The hypothesis that providerperceived difficulty in the relationship predicts provider medical information giving was supported. There was little variation in slopes across providers ( 2 = 16.11970, df = 10, p = .096). 91

PAGE 104

25.00 30.00 35.00 40.00 45.00Provider MED Information 10.29 16.54 22.79 29.04DDPRQ Scores Figure 9. Slopes of provider medical information giving as a function of providerperceived difficulty for each provider adjusted for patient depressive symptoms. 92

PAGE 105

Hypothesis 6 As provider perception of difficulty in the patient-provider relationship increases, provider and patient psychosocial questioning and information giving decreases. Preliminary analysis supported the relationship between provider perception of difficulty in the patient-provider relationship and provider psychosocial questioning. This relationship was examined by regressing pr ovider psychosocial questioning onto DDPRQ scores (see also Hypothesis 4). The mixed-effects model was specified by the following equation: DRPSYQij = 00 + 10 (BDI-II)ij + 20 (DDPRQ)ij + 0 j + 1 j (BDI-II)ij + 2 j (DDPRQ)ij + rij. A non-significant, inverse relationship was observed ( B = -0.050093, SE B = 0.033465, p = .165). In contrast to the OLS regression analysis findings, the hypothesis that provider-perceived difficulty predicts provider psychosocial questioning was not supported. There was no statistically significant difference between providers in terms of amount of psychosocial questioning ( 2 = 7.52070, df = 10, p .500) or slope of provider-perceived difficulty ( 2 = 9.24527, df = 10, p > .500) (Figure 10). 93

PAGE 106

-8.46 0.79 10.04 19.29 28.54 1.27 2.83 4.39 5.95 7.51DDPRQ ScoresProvider PSY Questioning Figure 10. Slopes of provider psychosocial questioning as a function of providerperceived difficulty for each provider adjusted for patient depressive symptoms. 94

PAGE 107

Hypothesis 7 As provider and patient medical questioning and information giving increases, patient satisfaction decreases. Preliminary analysis supported the relationships between provider medical questioning and provider information giving and patient satisfaction. These relationships were examined by regressing patient satisfaction onto provider medical questioning and provider medical info rmation giving (see also Hypothesis 2). The mixed effects model was specified by the following equation: SATij = 00 + 10 (BDI-II)ij + 20 (DRMEDQ)ij + 30 (DRMEDINFO)ij + 0 j + 1 j (BDI-II)ij + 2 j (DRMEDQ)ij + 3 j (DRMEDINFO)ij + rij. A non-significant, inverse relationship was observed between provider medical questioning and patient satisfaction ( B = -0.193769, SE B = 0.307920, p = .543). In contrast to the OLS regression analysis findings, the hypothesis that provider medical questioning predicts patient satisfaction was not supported. The slope of provider medical questioning did not significantly vary across providers ( 2 = 5.77424, df = 10, p > .500) (Figure 11). A significant, inverse relationship was observed between provider medical information giving and patient satisfaction (B = -0.488496, SE B = 0.128600, p = .004) (Figure 12) and confirmed the findings of the OLS regression analysis. The slope of provider medical information giving was consistent across providers ( 2 = 6.38846, df = 10, p > .500). The hypothesis that provider medical information giving predicts patient satisfaction was supported. 95

PAGE 108

-9.24 -2.65 3.93 10.52 17.11 74.32 94.94 115.57 136.19 156.82Provider MED QuestioningPatient Satisfaction Figure 11. Slopes of patient satisfaction as a function of provider medical questioning for each provider adjusted for patient depressive symptoms and provider medical information giving. 96

PAGE 109

0 7.50 15.00 22.50 30.00 70.00 97.50 125.00 152.50 180.00Provider MED InformationPatient Satisfaction Figure 12. Slopes of patient satisfaction as a function of provider medical information giving for each provider adjusted for patient depressive symptoms and provider medical questioning. 97

PAGE 110

The adequacy of the multilevel modeling at predicting patient satisfaction from BDI-II scores and provider medical information giving was further examined to see how closely the model fit the data. This was tested by comparing the deviance value of the model to the deviance value of the unconstrained (null) model. Deviance, a measure of lack of fit between the data and the model, is distributed as a chi-square statistic with degrees of freedom equal to the difference in the number of parameters estimated in each model (Luke, 2004). The difference between the two deviances was x2 (5, N = 115) = 32.38389, p = .000. Adding BDI-II and provider medical information giving as predictors of patient satisfaction provided a significantly better fit to the data than the unconstrained model. Multilevel modeling did not demonstrate the bivariate relationships that would have been required in order to establish a mediation effect of provider-perceived difficulty in the patient-provider relationship, provider medical information giving, or provider psychosocial questioning. Exploratory Analysis Hierarchical modeling indicated that the severity of patient depressive symptoms and the amount of provider medical inform ation giving during the medical encounter predict patient satisfaction. However, the analysis failed to support the majority of the hypothesized relationships between the communication variables and patient satisfaction. A large random effects variance component for level-1 (185.76746) demonstrated substantial un-modeled variability in patient satisfaction between patients within providers (random error). In an attempt to further elucidate the underpinnings of patient 98

PAGE 111

satisfaction with the medical encounter an exploratory analysis was conducted. The RIAS database provided an abundant opportunity for further examination. Twenty-two additional provider and 19 patient communication variables in the RIAS database were available for review (Table 13). Prior to analysis, the variables were normalized so that the values reflected the proportion of either all provider talk or all patient talk during the encounter and were then adjusted to percentage values to improve interpretability of the data. In addition to the utterance by utterance RIAS coding, the coders also rated the affect of the patient-provider dialogue (i.e., the emotional context) (Table 13). The coders did not directly link these ratings to particular events or utterances during the exchange, but assigned ratings based on their overall affective impressions of the speakers. Ratings were assigned for both provider and patient speech. Inter-rater reliability estimates were computed using Pearsons r. The correlation coefficients were all positive and ranged from .83 to 1.0 (Table 14). The variables were first examined by a series of SPSS correlations and OLS regressions. Significant bivariate correlations were observed between patient satisfaction and the provider variables of approval-direct ( r = .28, p = .002), back-channels (e.g., indicators of sustained interest, attentive listening, or encouragement) ( r = .20, p = .034), friendliness/warmth (r = .25, p = .007), sympathetic/empathetic (r = .21, p = .027), and respectfulness (r = .19, p = .04), and the patient variables of anger/irritation ( r = -.29, p = .002), interest/attentiveness ( r = .21, p = .023), friendliness/warmth ( r = .28, p =.003), responsiveness/engagement ( r = .26, p = .006), and sympathetic/empathetic ( r = .18, p = .054). The variables were modeled in a forward regression analysis to see if they 99

PAGE 112

Table 13 Additional RIAS Variables Included in Exploratory Analysis Communication Variables Provider Patient personal remarks personal remarks laughs, tells jokes laughs, tells jokes approval-direct approval-direct compliment-general compliment-general disagreement, criticism-direct disagreement, criticism-direct disagreement, criticism-general disagreement, criticism-general empathy/legitimation statements empathy/legitimation statements concern, worry concern, worry reassures, optimism reassures, optimism partnership statements gives information-other self-disclosure shows agreement, understanding gives information-other paraphrase, checks for understanding shows agreement, understanding transitions back-channels gives orientation, instructions paraphrase, checks for understanding all questions-other transitions asks for service gives orientation, instructions asks for reassurance open question-other asks for understanding asks for opinion bid for repetition asks for permission asks for reassurance bid for repetition Global Affect Variables Provider Patient anger/irritation anger/irritation anxiety/nervousness anxiety/nervousness dominance/assertiveness depression/sadness interest/attentiveness emotional distress/upset friendliness/warmth dominance/assertiveness responsiveness/engagement interest/attentiveness sympathetic/empathetic friendliness/warmth hurried/rushed responsiveness/engagement respectfulness sympathetic/empathetic respectfulness 100

PAGE 113

Table 14 Inter-Coder Reliability Estimates for RIAS Coded Global Affect Variables Using Pearsons r (n=12) Coder agreements/ Variable possible agreements r Provider Affect anger/irritation 12/12 1.0 anxiety/nervousness 12/12 1.0 dominance/assertiveness 12/12 1.0 interest/attentiveness 12/12 1.0 friendliness/warmth 11/12 .92 responsiveness/engagement 12/12 1.0 sympathetic/empathetic 11/12 .92 hurried/rushed 11/12 .92 respectfulness 10/12 .83 Patient Affect anger/irritation 12/12 1.0 anxiety/nervousness 12/12 1.0 depression/sadness 12/12 1.0 emotional distress/upset 12/12 1.0 dominance/assertiveness 12/12 1.0 interest/attentiveness 12/12 1.0 friendliness/warmth 10/12 .83 responsiveness/engagement 12/12 1.0 sympathetic/empathetic 12/12 1.0 respectfulness 12/12 1.0 101

PAGE 114

improved the prediction of satisfaction above and beyond BDI-II scores and the original eight communication variables. Significant re gression coefficients were observed for the provider variables of dominance/assertiveness ( B = 8.062, p = .005), and shows approvaldirect ( B = 3.209, p = .000), and the patient variables of friendliness/warmth ( B = 7.000, p = .002), depression/sadness ( B = -9.818, p = .016), and respectfulness ( B = 5.692, p = .048). Adding these variables also significantly improved R2 of the model ( R2 change = .020, p = .048). To evaluate the robustness of these relationships while controlling for the nested structure of the data, the variables were then modeled in HLM to determine if they improved the prediction of patient satisfaction scores beyond that afforded by BDI-II scores and provider medical information giving. A significant, positive relationship was observed between patient friendliness/warmth and patient satisfaction ( B = 6.290095, SE B = 2.397012, p = .026). This result indicated that the affective quality of a patients speech, specifically the expression of friendliness/warmth during the medical encounter is a predictor of patient satisfaction. The adequacy of the multilevel modeling at predicting patient satisfaction from BDI-II scores, provider medical information giving, and patient friendliness/warmth was further examined to see how closely the model fit the data. This was tested by comparing model deviance values. The difference between the two deviances was 2 (4, N = 115) = 16.82947, p = .002. Adding patient friendliness/warmth to BDI-II scores and provider medical information giving as predictors of patient satisfaction provided a significantly better fit to the data. 102

PAGE 115

Characteristics of the patient-provider encounter were examined. A third party (associated with the patient) participated during 17% of the encounters (n=19). In 12%, a second provider (provider in-training) participated ( n=14). The provider left the examination room at least once during 46% of visits ( n=53) and a staff member interrupted the patient-provider encounter in 15% (n= 17). The relationships between these four variables and patient satisfaction were examined using SPSS correlation and standard regression analysis. The presence of a second provider during the examination significantly correlated with patient satisfaction ( r = -.19, p = .046). There were no other significant bivariate correlations. Forward entry regression was used to determine if the addition of these variables improved the variance of patient satisfaction beyond that afforded by the eight original communication variables and BDI-II scores. None of the variables were found to be significant predictors of patient satisfaction, nor improve R2 in the presence of the previously modeled variables using OLS regression. The data were also explored in an attempt to elucidate the underpinnings of provider-perceived difficulty in the relationship since multilevel modeling failed to support the predicted relationship between patient depressive symptoms and provider difficulty. No significant bivariate correla tions were observed between any of the encounter characteristic variables of third pa rty, second provider, provider left room, or staff interruption with DDPRQ scores. Forwar d entry regression was used to determine if any of these variables predicted DDPRQ scores above and beyond that afforded by BDIII scores alone. None of the variables were found to be significant predictors of provider103

PAGE 116

perceived difficulty in the patient-provider relationship, nor improve R2 in the presence of BDI-II scores using OLS regression. The global affect ratings were re-examined to see if the emotional context of the encounter predicted DDPRQ scores. Significant correlations were demonstrated between DDPRQ scores and provider friendliness/warmth (r = -.19, p = .046), patient emotional distress/upset ( r = .22, p = .018), patient friendliness/warmth (r = -.23, p = .012), and patient sympathetic/empathetic (r = -.34, p = .000). Forward entry regression was used to determine if the any of the provider or patient ratings predicted DDPRQ scores above and beyond that afforded by BDI-II scores alone. The patient global affect variable of sympathetic/apathetic was the only variable that demonstrated a significant relationship with DDPRQ scores in the presence of BDI-II scores ( B = .707, p = .000) and also significantly improved R2 of the model ( R2 change = .125, p = .000). The robustness of this relationship was then tested in HLM, adjusting for the nested structure of the data. A significant, inverse relationship was observed ( B = -8.525196, SE B = 2.138115, p = .000). This result indicated that the affective quality of a patients speech, specifically the expression of sympathy/empathy that is directed toward the provider is a predictor of provider perception of difficulty in the patient -provider relationship. The adequacy of the multilevel modeling at predicting DDPRQ scores from patient sympathy/empathy was examined by comparing the deviance value of the model to the unconstrained (null) model. The difference between the two deviances was x2 (2, N = 115) = 20.79954, p = .000. Adding patient sympathy/empathy as a predictor of DDPRQ provided a significantly better fit to the data than the unconstrained model. 104

PAGE 117

These findings are discussed in the next chapter. 105

PAGE 118

Discussion Primary Aims and Hypotheses Depressive Symptoms and theDifficult Patient A specific aim of this study was to determine if a patients depressive symptomatology impacts the patient-provider relationship by increasing the likelihood of being perceived as a difficult or frustrating patient by the provider. The findings indicate that the prevalence and severity of depressive symptoms in the patient sample was similar to what has been reported elsewhere. This confirms that substantially more individuals suffer milder, yet possibly significant, depressive symptoms that do not meet DSM-IV criteria for major depressive disord er (American Psychiatric Association, 1994). Subthreshold depression, like major depression, is associated with significant disability, functional impairment, comorbidity, increased costs and utilization of health care services, and work absenteeism. These individuals are unlikely to seek treatment for their depression. Yet the evidence indicates that many individuals with mild symptoms are likely to develop more serious depressive symptoms. It is critical that primary care providers recognize, diagnose, and effectively treat depression in their patients in order to address these problems before significant impairment results. The DDPRQ was scored and analyzed as a continuous variable. It can also be used as a dichotomous variable with a cut-point of 30 (Hahn et al., 1996). If used in that manner, providers would have classified 9% of the patient sample as difficult. This is 106

PAGE 119

slightly less than what would have been anticipated, as other studies using the DDPRQ typically report that 10% to 20% of patients provoke a level of physician distress and frustration that transcends the expected and accepted level of difficulty (Hahn et al., 1994; Hahn et al., 1996; Jackson & Kroenke, 2001). There are several possible explanations. Provider selection, their interests in and attitudes toward depression care, and a social desirability response set might have biased DDPRQ ratings in this sample. In addition, a patient self-selection bias could have influenced providers responses. Since previous research has indicated that patients can accurately estimate their physicians liking of them (Hall et al., 2002), patients who had particularly satisfying established relationships with their providers might have been more inclined to participate. Although it is possible that the data underestimate the prevalence of provider perception of difficulty in primary care patients, there is no indication that this sample differed from those of previous studies. Consistent with previous research, the data demonstrated a significant positive correlation between depressive symptoms and provider-perceived difficulty. However, depressive symptomatology, whether measured by BDI-II scores or implied by the affective tonal quality of patient speech (i.e., sounding depressed, sad, or emotionally distressed), did not predict provider-perceived difficulty in this patient and provider sample after controlling for other predictors. This finding counters the main premise of this study and merits examination. The association between mental disorders, especially depressive and anxiety disorders is well documented (Hahn et al ., 1994; Hahn et al., 1996; Hahn, 2001; Jackson, 107

PAGE 120

2005; Jackson & Kroenke, 2001; Kroenke et al., 1997). Furthermore, many of the behavioral characteristics associated with and displayed by depressed patients mirror those that are frequently ascribed to difficult patients. However, two distinct methodological features of this study strengthen the current finding. First, in contrast to several of the aforementioned studies the providers in this study were blinded to (and therefore presumably not influenced by) patie nts depression screening results during the encounter. Second, this study sample was limited to established patient-provider pairs. New or first time patients who had yet to begin a provider relationship were excluded from participation. Contrary to a priori speculation this study suggests that providers do not respond in a predictable way to their depressed patients. This nonsignificant finding should not be interpreted as though the null is unequivocally true however. The data demonstrated an effect size of r = 0.31 that is consistent with the conventional medium effect size of a population r of .30 (Cohen, 1977, p. 60). Hierarchical modeling revealed significant variability between providers in terms of DDPRQ scores and substantial unexplained differences in scores between patients within providers (random error). This indicates that either the effect of depressive symptoms on provider-perceived di fficulty, or the study sample size, would have to be larger to have confidence in detecting it. It is also possible that the outlier cases impacted the multilevel regression coefficient and solution since extreme cases can lead to Type I and Type II errors (Tabachnick & Fidell, 2001). This explanation is partially supported by the observed multilevel regression coefficient p value of 0.07 (95% C.I. = 0.18235 0.67071) when the outlier cases were excluded from analysis compared 108

PAGE 121

with p = 0.11 (95% C.I. = 0.31097 0.82193) when retained. Neither confidence level contains zero. Therefore, it is possible that providers may have a tendency to respond to their depressed patients in a predictable way that could not be detected using the conventional alpha value (.05) in this study. The relationship between depressive symptoms and provider-perceived difficulty is likely confounded by other unmeasured variables. Exploratory analysis indicates that the affective quality of patient speech, specifically the expression of sympathy or empathy, significantly lowers the likelihood of a patient being considered as a difficult patient. It is interesting to note that the affective quality of provider speech, especially expressions of anger/irritation or anxiety/ nervousness that presumably would reflect provider frustration with a patient, neither correlate with nor appear to influence provider ratings of difficulty in the relationship. Future research efforts should explore for possible interaction effects between patients depressive symptoms and patient and provider affective behaviors on provider per ception of difficulty. Difficulty may also arise from a unique combination of particular patient and provider characteristics that would be difficult to capture in any study. Depressive Symptoms and Patient-Provider Communication Depressive symptomatology significantly correlates with provider medical information giving and provider psychosocial questioning yet it does not appear to have a significant influence on the verbal content of the encounter. It was hypothesized that patient and provider communication behaviors, specifically medical and psychosocial talk, could be predicted based on the severity of patients depressive symptoms. This 109

PAGE 122

hypothesis was not supported. None of the patient or provider communication variables were predicted by patients depressive symptoms. Results were consistent whether the depressive symptoms were measured by BDI-II scores or implied by the affective tonal quality of patient speech. This finding is consonant with past research that demonstrated providers interviewing styles do not change according to patient characteristics such as emotional distress (Del Piccolo et al., 2002; Deveugele et al., 2002; Sleath & Rubin, 2002) and converges with a proposed link between health status and provider behavior (Hall et al., 1996). Although this finding lacks statistical significance it is clinically meaningful. The majority of patients reported that (in general) it was at least somewhat important that their primary care provider attend to their mental health needs and previous research has shown that patients welcome the opportunity to discuss their psychosocial concerns during the course of the medical visit (Bertakis et al., 1991). Since the majority of individuals with depression are initially evaluated in primary care settings and many rely on a primary care provider for all of their health needs, it is troubling to observe that communication behaviors dont necessarily change in response to patients depressive symptoms. This finding may reflect some of the dynamics involved in the delivery of depression care. Whether such behavior contributes to the under-diagnosis and suboptimal treatment of depression in primary care cannot be answered by this study. Several caveats bear mentioning. First, it is not known exactly how many patients in this study actually desired or had a pre-visit expectation to discuss their depressive symptoms during the observed visit. Less than 3% of patients who responded to the 110

PAGE 123

question listed depression as their major reason for the appointment on the pre-visit questionnaire. Second, this study provided only a snapshot glimpse into the black box of the medical encounter. It did not examin e patients medical or psychosocial history, co-morbid health problems, medication use, recency or content of last office visit, or length of the patient-provider relationship. Continuity of care is a main attribute of primary care and widely believed to enhance disclosure of symptoms. Perhaps one advantage to familiarity is the ability to read between the lines. For example, in a well-established relationship one brief patient statement (e.g., I feel lousy, not good, same thing, or even you know ) may be all that needs to be said in order for a provider to understand his/her patients individual situation. Hence, it is possible th at verbal coding, which is based on frequency counts, may not adequately reflect the communication process between depressed patients and their providers. Depressive Symptoms and Patient Satisfaction It is well documented that individuals with depression are significantly less satisfied with their medical care. This study hypothesized that depressive symptomatology would have a negative impact on patient satisfaction with the medical encounter. The data support this hypothesis and indicate that the more depressed a patient is, the less satisfied he/she is with their care. This finding is not only statistically significant but substantively meaningful as well. Patient satisfaction is a legitimate measure of health care quality and is important not only as a financial investment for 111

PAGE 124

health care organizations but also because of its significant associations with other patient outcomes including adherence to treatment, physiologic status, and symptom resolution. Hierarchical modeling revealed significant unexplained variability between providers both in terms of their patients satisfaction ratings as well as the effect depressive symptoms had on their patients satisfaction. However, this study cannot untangle nor explain the source(s) of depresse d patients dissatisfaction. Part of it likely stems from the adverse nature of depression itself (i.e., altered mood and cognition). Whether depressive symptoms bias patient assessments of satisfaction, or if these patients receive lower quality of care, as suggested by Orlando and Meredith (2002), cannot be answered by these findings. Further interpretation is also limited by the fact that patients expectations of the office encounter were not measured. A lack of unmet expectations has been shown to be a powerful predictor of patient satisfaction at various time points (Jackson et al., 2001). Therefore, it is possible that depressed patients in this study were less satisfied than their non-depressed counterparts because their pre-visit expectations may not have been met. A qualitative approach to inquiry, specifically narrative analysis, which focuses on the recounted or observed stories of participants experiences (storytelling) would give voice to depressed patients and help to elucidate the underpinnings of their (dis)satisfaction with medical care. Patient-provider communication does not appear to be a source of depressed patients (dis)satisfaction as the data failed to demonstrate a communication mediation hypothesis between depressive symptoms and pa tient satisfaction. Two distinct features of this study enhance this finding. First, pa tient satisfaction was measured immediately 112

PAGE 125

after the observed visit. Although satisfaction may have been confounded by additional uncontrolled factors (e.g., waiting time, friendliness of the office staff, convenience of appointment, non-verbal communication of the provider, etc.), ratings were not biased by the passage of time (i.e., patient recollection of events). Second, the dyadic exchange during the encounter was examined by direct observation and therefore did not depend on subjects recollection or interpretation of events. Difficulty and Patient-Provider Communication The second aim of this study was to determine if provider perception of difficulty in the relationship influences verbal communication, specifically, medical or psychosocial talk, during the encounter. As hypothesized, the results indicate that provider-perceived difficulty predicts at least part of the verbal exchange: provider medical information giving. The more difficult or frustrating a patient is, the more medical information they receive during the visit. Providers were strikingly consistent in this behavior. Whether providers give more information in response to patient questions or requests for information, or because of a patients clinical status, cannot be answered. It could be that providers, when dealing with difficult patients, fall back on their professional responsibility to provide medical information and, as a trade-off, subsequently withdraw as social participants during the encounter. Provider perception of difficulty did not predict patient or pr ovider psychosocial questioning or information giving, patient medical questioning or in formation giving, or provider medical questioning. 113

PAGE 126

Bensing and Dronkers (1992) found providers asked fewer questions, especially psychosocial questions, in encounters in which the physician appeared to be irritated, anxious, or nervous. It is interesting to note two particular trends in the current data despite weak bivariate correlations ( r = -.13 to .21). As difficulty scores increased, there was an across the board increase in patient and provider medical talk and a corresponding decrease in all psychosocial talk. Use of medical and psychosocial talk may indicate subtle ways in which providers tend to dominate consultations. Although interpretation of this observation is limited, this could possibly suggest that providers, when dealing with difficult patients, may attempt to cont rol the content and direction of the encounter by using medical jargon. If this is the case, this may be in response to increased uncertainty in the care of difficult patients (Schwenk et al., 1989). It may also be one method providers employ, either consciously or unconsciously, in an attempt to avoid discussing other topics. The function of provider medical info rmation giving during encounters with difficult patients is unclear. To determine if it is an attempt by providers to control the content and direction of the encounter, this could be further examined using a qualitative approach to inquiry, namely discourse analys is. Discourse analysis focuses on how talk within medical (or other) encounters functions to change, establish, or maintain social power relationships (Roter & McNeilis, 2003). Alternatively, the audiotapes of the visits could be coded using a relational control method approach, rooted in relational communication theory. This method examines the exchange of paired sequential messages over time as the basic unit of analysis. Hence, each message is treated as both 114

PAGE 127

a response to the preceding message and a stimulus for the message that follows (Rogers & Farace, 1975). Patient-Provider Communication and Patient Satisfaction This study examined the association between medical and psychosocial talk and patient satisfaction immediately after the offi ce visit. Only one communication behavior had a consistent significant negative effect on patient satisfaction: provider medical information giving during the encounter. The reason for this cannot be explained. However, an overload of information, especially if given too quickly (e.g., due to time constraints) or at a level that the patient does not understand, may result in the patient feeling confused and dissatisfied. Openness to the patients agenda and willingness to negotiate options may facilitate good communication and convey an understanding of patient preferences and values regarding health. Patients in this study may have felt that they could not ask questions or that their provider did not listen to what they tried to say. Alternatively, this finding may be a function of unmet patient expectations. Once again, a qualitative approach to inquiry, specifically narrative analysis, would give voice to patients and help to elucidate the underpinnings of their (dis)satisfaction with provider medical information giving. Patient satisf action was not predicted from patient or provider psychosocial questioning or informa tion giving, patient medical questioning or information giving, or from provider medical questioning. The data suggest an indirect link betw een provider perception of difficulty and patient (dis)satisfaction in addition to the direct link that has been reported previously (Hahn et al. 1996). The more difficult or frustrating a patient is, the more medical 115

PAGE 128

information they receive during the visit. However, this study indicates that the more medical information given to patients during the encounter, the less satisfied they are with their care. This offers one possible explanation as to why difficult patients report lower levels of satisfaction, since dissatisfaction has been shown to be associated with a high-control style of provider communicati on (e.g., domination of the encounter and/or disregard of the patients agenda) (Buller & Buller, 1987; Cecil & Killeen, 1997; Di Blasi et al., 2001; Flocke et al., 2002; Roter et al., 1997). Future research should explore the potential mediating effect of provider medi cal information giving on the association between provider perception of difficulty and patient satisfaction. Provider communication, especially when measured immediately after the encounter, is widely believed to have the most direct impact on patient satisfaction whereas satisfaction ratings 2 weeks and 3 months after the consultation are related to medical outcome, such as health status (Jackson et al., 2001). However, satisfaction ratings may reflect something other than the verbal communication and affective behaviors examined in this study. With the exception of provider medical information giving ( r = -.30, p .001), all of the other communication variables demonstrated a weak (or no) bivariate association with satisfaction ( r = -.05 to .17, p .05) vis a vis generally high patient satisfaction ratings. This suggests that patient-provider communication, especially psychosocial discussion, may not be necessary for patients to be content with the office visit. If this is the case, then this observation challenges Roter et al. (1988) who found patients were most satisfied with medical encounters that entailed a lot of psychosocial talk. High levels of patient satisfaction are a common finding in 116

PAGE 129

satisfaction research and may reflect a ceiling effect of various satisfaction instruments. Patients may also have a tendency to give higher ratings because they are uncomfortable being critical of the person who is responsible for their health care. Yet, it has been posited that satisfaction may more often reflect attitudes (e.g., they are doing the best they can, or well, its not really their job to do Williams, Coyle, and Healy, 1998, p. 1358 ) and should not be taken to indicate that patients have had or are having good experiences in relation to a particular service. Multilevel Modeling The study unequivocally underscores the importance of context (grouping of variables) when examining phenomenon of interest to health researchers. Hierarchical linear modeling is a multivariate technique that was theoretically, statistically, and empirically justified for use in this study. Traditional statistical techniques including ordinary least squares regression analysis are not suitable for clustered data with correlated errors, largely because they violate the assumption of independence. Unfortunately, these methods are often inappropriately used in nursing research. In standard regression analysis, the resulting standard errors are smaller than they should be and thereby increase the chance of committing a Type I error (Luke, 2004). This risk is poignantly demonstrated in this study, as ma ny of the significant OLS findings were not replicated once the nested structure of the data was taken into account. Multilevel modeling provided an improved, honest, and substantially different estimate of the 117

PAGE 130

associations between patient depressive symptoms, provider perception of difficulty, patient-provider communication, and patient satisfaction. Single level statistical techniques can provide a lot of precision over inferences. However, they limit the ability to evaluate contextual effects. One of the benefits of multilevel modeling is its ability to partition the variance and covariance components among levels (i.e., decomposing the covariation among a set of encounter-level variables e.g. patient satisfaction into within provider and between provider components) (Raudenbush & Bryk, 2002). In this study multilevel modeling identified significant differences between providers (e.g., distributions of DDPRQ scores, effect (slope) of BDI-II scores on patient satisfaction) and within providers (e.g., DDPRQ and patient satisfaction scores) that would not have been detected in a single level analysis. Furthermore, modeling allowed for an improved estimation of effects by estimating separate regression equations for each individual provider. This study was not sufficiently powered to examine the effect of level-2 variables (e.g., provider gender, age, ethnicity, years in professional practice, etc.) on level-1 data (e.g., DDPRQ scores, patient-provider communication, or patient satisfaction). A significantly larger number of groups (providers) would be required to test level-2 estimates. Hierarchical modeling has one additional significant advantage over traditional statistical methods: its ability to evaluate cross-level effects. However, this study was not sufficiently powered to examine for cross-level interactions. Such analyses would be based on the total number of level-2 units (i.e., providers) rather than on the total sample size. Eleven providers (groups) participated in this study. Kreft and De Leeuw (1998) 118

PAGE 131

discuss studies indicating that at least 30 groups and 30 observations within each group may be needed to detect cross-level interactions. They conclude that sufficient power can be obtained when groups are not too small, and the number of groups is larger than 20 (p. 126). Future research should examine for second-level effects as discussed above and for cross-level effects, such as between individual patient characteristics (e.g., gender, age, race, ethnicity, etc.) and provider characteristics (e.g., gender, age, type of provider, years in professional practice, attitude toward depression, etc.) on provider perception of difficulty, patient-provider communication, and patient satisfaction in a larger sample of patients and providers. Significance to Clinical Practice Consistent with previous research, this study suggests that primary care providers do not recognize patients cues and are not responsive to patients depressive symptoms during the medical encounter. Unequivocally, primary care providers are faced with multiple competing demands in the delivery of depression care. Patients and providers alike face the dilemma of deciding what to discuss during visits, particularly given the many issues that might arise in a primary care visit and the increasing pressure placed on providers to see more patients in less time. It is important for providers to understand patients goals and expectations for the visit in order to clarify their own role during the encounter. Without a mutual understanding of goals and expectations, the relationship may come to what has been defined as a dysfunctional standstill or relationship default (Roter & McNeilis, 2003). 119

PAGE 132

One way for providers to understand their patients goals and expectations for the visit is to use a pre-visit patient questionnaire. This may, however, entail a trade-off between conflicting provider goals: trying to improve their response to patient concerns while not significantly increasing the length (and subsequent cost) of the office visit (Hornberger, Thom, & MaCurdy, 1997). Regardless of the time constraints and financial barriers that create complexity and restrictions within care delivery, the role of the primary care provider is to provide comprehensive patient care. Effective interpersonal communication is crucial to elicit patient concerns. It is also fundamental to establishing and preserving the patient-provider relationship. Primary care providers can be uncomfortable discussing depression with their patients and there is evidence that providers can exhibit increased anxiety during medical encounters that entail a lot of psychosocial talk (Roter et al., 1988). Balint groups can help providers to care for their depressed and/or difficult patients. Developed by psychoanalyst Michael Balint, these programs are designed to help general care providers deal with the psychological aspects of their patients problems by incorporating psychological techniques into general practice (Balint, 1964). The focal point of the seminars is the quality of the patient-provider relationship and the providers emotional response to his/her patient. Now with an international presence, the groups demonstrate how patients and providers constantly influence each other during the medical encounter. Balint groups help providers gain insight on how their emotional responses to patients impact the course and content of the medical visit, patient outcomes, future patient encounters and health care management, as well as patient and provider levels of 120

PAGE 133

satisfaction. Primary care providers should also take advantage of continuing education offerings, patient-provider communication skills seminars and intensive training programs when available. Any significant, enduring improvement in the overall treatment of depression by primary care providers will clearly require additional substantive changes in the structure of care delivery above and beyond improving providers communication skills and receptiveness to depressed patients needs. Individual clinicians and practice groups should explore local options for integrating mental health services into primary health care, as discussed in the U.S. Surgeon Generals report on The Integration of Mental Health Services and Primary Health Care (DHHS, 2001). For example, one option is to have a mental health specialist on-site in the primary care providers office to deliver care and to bridge primary and specialty mental health care. Another option may be to utilize case managers to screen patients for depression and to monitor their adherence to prescribed therapy and response to treatment. This study also has important implications for educators. Graduating primary clinicians often feel ill equipped to deal with the complexities of mental illness in the course of routine practice. Primary care providers report little formal training in the diagnosis and treatment of depression and sparse guidance about what level of severity can be treated effectively in primary care vs. specialty mental health care (DHHS, 2001). There are also few training programs that em phasize the integration of mental health services and primary care (DHHS, 2001). Although many graduate nursing and medical school programs have adopted mental health coursework and communication skills 121

PAGE 134

training into their primary care curricula, there are currently few incentives for educational institutions to step beyond existing training programs. Educators should work in concert with professional associations to develop a set of common core competencies specific to the provision of mental health services for all primary care clinicians. Limitations These findings must be interpreted in light of several limitations. A possible selection bias of providers and patients limits the generalizability of these findings to other provider and patient groups. Providers who responded to the initial recruitment letter and agreed to participate may have differed from the providers who chose not to participate. There was limited variation among subjects and a lack of racial and ethnic diversity among patients. All providers were Caucasian and saw patients in a fee-forservice setting in a single urban community. The data collection and coding methodology impose some limitations. Although the BDI-II is the most widely used instrument for detecting depression symptoms, its reliability is affected by patients interpretation of its emotional terms and their conception of depression (Kerr & Kerr, 2001). Although each subject was assured of the confidentiality of their survey responses and grip-seal return envelopes were distributed with each BDI-II, DDPRQ, and satisfaction surve y, the possibility of a social desirability response set bias might have influenced subjects responses in a desire to present a 122

PAGE 135

favorable image. Patients and providers could have modified their normal interview behaviors because of the audio recording procedure (Hawthorne effect). This study observed both the instrumental (cure oriented) and affective (care oriented) verbal communication behaviors that were displayed during the encounter. However, non-verbal behaviors, which are inhe rently a part of every social interaction and can often speak louder than words, were not observed. Although the dyadic exchange during the encounter was measured by direct observation, verbal coding (RIAS) is a reliable representation of actual medical practice; one cannot read into a patients head or know what a providers intentions were when asking certain questions or making particular statements. Verbal coding, which is based on frequency counts, may not adequately reflect the total communication process. The restricted range in sampling of cases limits interpretation of findings. The majority of patients reported experiencing minimal depressive symptoms, were highly satisfied with their visit, and were not perceived by their providers as very difficult patients. Furthermore, five of the eight ve rbal communication variable distributions were moderately skewed. Therefore, the sample data may not accurately reflect population estimates. The relationships among these variables could be much stronger than shown here. The study design did not allow for the evaluation of verbal communication behaviors in relation to how established a patient was (i.e., length of the patientprovider relationship or number of previous encounters), recency of last office visit, whether the provider recognized a patients depressive symptomatology, if there was a 123

PAGE 136

past history or current diagnosis of depressi on in the patient record, if the patient was currently being treated for depression, or the presence of co-morbid psychiatric disorders. Lastly, conclusions drawn from this study are limited by its cross-sectional design, which precludes any inference of causation. Direction for Future Research This study supports large body of research indicating that depressed individuals are less satisfied with their medical care. The proposed model did not, however, explain the process or source(s) of depressed patients dissatisfaction with care. Patient-provider communication, by itself, does not appear to be a contributing factor. Thus, an important priority for future research is to identify sources of satisfaction that are important to this vulnerable group. Methodological triangulation (e.g., using observation, patient interviews, and/or review of medical records) would help to elucidate this problem. In addition, prospective, longitudinal studies that observe patient-provider behaviors over time would help to discern the temporal relationship between these two variables. With the growing body of research demonstrating that patient satisfaction is related to improved patient outcomes, it is imperative that health care organizations and policy makers look beyond the financial incentives related to patient satisfaction and support research efforts to understand and enhance depressed patients care experiences to improve their quality of life. Incorporating a pre-visit interview or questionnaire designed to assess patients expectations in future research will help to illuminate a potential link between health 124

PAGE 137

status, provider behavior, and satisfaction w ith care. Likewise, including historical information such patient medical and mental health history, medication use, and length of the patient-provider relationship in future analyses will enhance our understanding of the process of depressed patient care and may add insight to the patient and provider behaviors observed in this study. This study did not fully examine the verbal or affective behaviors of third parties (such as second providers or accompanying family members) during the encounter. Although their presence did not appear to influence patient satisfaction, their impact on provider perception of difficulty or content of the consultation is not known. Thus, a potential area for future study includes exploring the contributions of other participants during the clinical encounter. Additionally, it would be interesting to examine the dynamics of the office visit by considering patient and third party as one unit (i.e., patient), and provider and second provider as one unit (i.e., provider). A qualitative analysis of the data would to help disentangle the complexities and subtleties of patient-provider communication during the medical encounter. By more closely examining the verbal dyadic exchange, it would be possible to better understand the relationships of the patients and providers in this study. Therefore, it is important to listen to the audiotapes of the patient encounters in order to get a better idea exactly what was communicated during these visits as well as what that information might mean to each party. Communication patterns could be examined and providers behaviors in response to patients psychosocial talk, most notably indicators of depression and direct expressions of depression, could be explored. The dialogue could also be examined for 125

PAGE 138

various indicators of control during the exchange and would help to characterize the information gained when providers interrupt patients (or patients interrupt providers). In addition, this could help to define the communication strategies that move discourse during the clinical encounter toward (or away from) the traditional medical model approach to patient care. Finally, an important next step is further analysis of this data to explore for possible interaction effects among the level-1 variables (e.g., What if the patient is depressed and angry? What if the patient is deemed a difficult patient and is assertive, or anxious, or attentive? Does provider medical information giving have the same impact on patient satisfaction if there is also a lot of provider reassurance, or laughing, or personal talk during the visit?). The abundant RIAS database provides vast opportunity for further examination. Conclusions This study demonstrated that depression is associated with increased providerperceived difficulty in the patient-provider relationship. Providers do not appear, however, to respond in a predictable way to their depressed patients nor significantly change their interviewing style according to the severity of depressive symptoms. There is significantly more provider medical information given during encounters with difficult patients and this behavior has a negative influence on patient satisfaction. In addition, findings indicate that patient-provi der communication, by itself, does not appear to be a source of depressed patients oft-reported dissatisfaction with medical care. The 126

PAGE 139

sources of and processes contributing to depressed patients satisfaction remain poorly understood. Primary care providers are in a position to take a leadership role in improving the quality of life of their patients with depression. Additional analyses of data from this study and others will be needed to further understand the core processes and structures of primary care practice in relation to the diagnosis and management of depression, their effect on patient outcomes, and to uncover opportunities for enhancing the effectiveness of depression care in primary care. 127

PAGE 140

References Aragones, E., Labad, A., Pinol, J.L., Lucena, C. & Alonso, Y. (2005). Somatized depression in primary care attenders. Journal of Psychosomatic Research, 58, 145-151. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Balint, M. (1964). The doctor, his patient, and the illness (Revised ed.). Madison, WI: International Universities Press, Inc. Barkow, K., Heun, R., Ustun, T.B., Berger, M ., Bermejo, I., & Gaebel, W., et al. (2004). Identification of somatic and anxiety symptoms which contribute to the detection of depression in primary health care Abstract European Psychiatry: The Journal of the Association of European Psychiatrists, 19, 250-257. Baron, R.M. & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. Barsky, A.J., Wyshak, G., Latham, K.S., & Klerman, G.L. (1991). Hypochondriacal patients, their physicians, and their medical care. Journal of General Internal Medicine, 6, 413-419. Beck, R.S., Daughtridge, R., & Sloane, P.D. (2002). Physician-patient communication in the primary care office: a systematic review. Journal of the American Board of Family Practice, 15, 25-38. Beck, D.A. & Koenig, H.G. (1996). Minor depression: a review of the literature. International Journal of Psychiatry in Medicine, 26, 177-209. Beck, A.T., Steer, R.A., & Brown, G.K. (1996). Beck Depression Inventory manual (2nd ed.). San Antonio, TX: The Psychological Corporation. Beckman, H.B. & Frankel, R.M. (1984). The effect of physician behavior on the collection of data. Annals of Internal Medicine, 101, 692-696. 128

PAGE 141

Beckman, H.B., Markakis, K.M., Suchman, A.L., & Frankel, R.M. (1994). The doctorpatient relationship and malpractice. Lessons from plaintiff depositions. Archives of Internal Medicine, 154, 1365-1370. Bensing, J.M. & Dronkers, J. (1992). Instrumental and affective aspects of physician behavior. Medical Care, 30, 283-298. Bertakis, K.D., Roter, D., & Putnam, S.M. (1991). The relationship of physician medical interview style to patient satisfaction. Journal of Family Practice, 32, 175-181. Blau, P.M. (2002). Theory of social exchange. In J. Berger & M. Zelditch, Jr. (Eds.), New Direction in Contemporary Sociological Theory (pp. 346-349). Lanham, MD: Rowan Littlefield Publishers. Borowsky, S.J., Rubenstein, L.V., Meredith, L.S., Camp, P., Jackson-Triche, M., & Wells, K.B. (2000). Who is at risk of nondetection of mental health problems in primary care? Journal of General Internal Medicine, 15, 381-388. Bredart, A., Bouleuc, C., & Dolbeault, S. (2005). Doctor-patient communication and satisfaction with care in oncology. Current Opinion in Oncology, 17, 351-354. Brody, D.S., Hahn, S.R., Spitzer, R.L., Kroenke K., Linzer, M., deGruy, F.V., et al. (1998). Identifying patients with depression in the primary care setting. A more efficient method. Archives of Internal Medicine, 158, 2469-2475. Brody, D.S., Khaliq, A.A., & Thompson, T.L. (1997). Patients perspectives on the management of emotional distress in primary care settings. Journal of General Internal Medicine, 12, 403-406. Brown, S.W., Nelson, A., Bronkesh, S.J., & Wood, S.D. (1993). Patient satisfaction pays: Quality service for practice success. Gaithersburg MD: Aspen Publishers. Brown, C. & Schulberg, H.C. (1998). Diagnosis and treatment of depression in primary medical care practice: the application of research findings to clinical practice. Journal of Clinical Psychology, 54, 303-314. Brown, J.B., Stewart, M., & Ryan, B.L. (2003). Outcomes of patient-provider interaction. In T.L. Thompson, A.M. Dorsey, K.I. Miller, & R. Parrott (Eds.), Handbook of Health Communication (pp. 141-161). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Buller, M.K. & Buller, D.B. (1987). Physicians communication style and patient satisfaction. Journal of Health and Social Behavior, 28, 375-388. 129

PAGE 142

Callahan, E.J., Bertakis, K.D., Azari, R., Robbins, J.A., Helms, L.J., & Chang, D.W. (2000). The influence of patient age on primary care resident physician-patient interaction. Journal of the American Geriatric Society, 48, 30-35. Callahan, E.J., Bertakis, K.D., Azari, R., Robbi ns, J.A., Helms, L.J., & Leigh, J.P. (2002). Association of higher costs with symptoms and diagnosis of depression. Journal of Family Practice, 51, 540-544. Callahan, E.J., Bertakis, K.D., Azari, R., Robbins, J., Helms, L.J., & Miller, J. (1996). The influence of depression on physician-patient interaction in primary care. Family Medicine, 28, 346-351. Callahan, E.J., Jaen, C.R., Crabtree, B.F., Zyzanski, S.J., Goodwin, M.A., & Stange, K.C. (1998). The impact of recent emotional distress and diagnosis of depression or anxiety on the physician-patient encounter in family practice. Journal of Family Practice, 46, 410-418. Cecil, D.W., & Killeen, I. (1997). Control, compliance, and satisfaction in the family practice encounter. Family Medicine, 29, 653-657. Chang, J.T., Hays, R.D., Shekelle, P.G., MacLean, C.H., Solomon, D.H., Reuben, D.B., et al. (2006). Patients global ratings of their health care are not associated with the technical quality of their care. Annals of Internal Medicine, 144, 665-672. Clever, S.L., Ford, D.E., Rubenstein, L.V., Rost, K.M., Meredith, L.S., Sherbourne, C.D., et al. (2006). Primary care patients involvement in decision-making is associated with improvement in depression. Medical Care, 44, 398-405. Cohen, J. (1977). Statistical power analysis fo r the behavioral sciences (Rev. ed.). New York: Academic Press. Cohen, G., Forbes, J., & Garraway, M. (1996). Can different patient satisfaction survey methods yield consistent results? Comparison of three surveys. BMJ, 313, 841844. Cooper, L.A., Brown, C., Vu, H.T., Ford, D.E., & Powe, N.R. (2001). How important is intrinsic spirituality in depression care? A comparison of white and AfricanAmerican primary care patients. Journal of General Internal Medicine, 16, 634638. Cooper, L.A., Brown, C., Vu, H.T., Palenchar, D.R., Gonzalez, J.J., Ford, D.E., et al. (2000). Primary care patients opinions regarding the importance of various aspects of care for depression. General Hospital Psychiatry, 22, 163-173. 130

PAGE 143

Cooper, L.A., Gonzales, J.J., Gallo, J.J., Rost, K.M., Meredith, L.S., & Rubenstein, L.V. (2003). The acceptability of treatment for depression among African-American, Hispanic, and white primary care patients. Medical Care, 41, 479-489. Cooper, S., Valleley, R.J., Polaha, J., Begeny, J., & Evans, J.H. (2006). Running out of time: physician management of behavioral health concerns in rural pediatric primary care. Pediatrics, 118 132-138. DeGroot, M., Anderson, R. Freedland, K.E., Clouse, R.E., & Lustman, P.J. (2001). Association of depression and diabetes complications: a meta-analysis. Psychosomatic Medicine, 63, 619-630. Del Piccolo, L., Mazzi, M., Saltini, A., & Zimmerman, C. (2002). Inter and intra individual variations in physicians verbal behaviour during primary care consultations. Social Science & Medicine, 55, 1871-1885. DiMatteo, M.R., Lepper, H.S., & Croghan, T.W. (2000). Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence. Archives of Internal Medicine, 160 21012107. Department of Health and Human Services (1999). Mental health: A report of the Surgeon General. Rockville, MD: US Department of Health and Human Services, SAMHSA, CMHS, NIH, NIMH. Report available at http://www.surgeongeneral.gov/library/mentalhealth/home.html Department of Health and Human Services (2001). Report of a Surgeon Generals working meeting on the integration of mental health services and primary health care. Atlanta, GA. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Office of the Surgeon General. Report available at http://www.surgeongeneral.gov/library/mentalhealthservices/mentalhealthservices .pdf Department of Health and Human Services (2000). Healthy People 2010: Understanding and improving health (2nd ed.). Washington, DC: US Government Printing Office. Deveugele, M., Derese, A., & De Maeseneer, J. (2002). Is GP-patient communication related to their perceptions of illness severity, coping, and social support? Social Science & Medicine, 55, 1245-1253. Di Blasi, Z., Harkness, E., Ernst, E., Geor giou, A., & Kleijnen, J. (2001). Influence of context effects on health outcomes: a systematic review. Lancet, 357, 757-762. 131

PAGE 144

Docherty, J.P. (1997). Barriers to the diagnosis of depression in primary care. Journal of Clinical Psychiatry, 58(Suppl. 1), 5-10. Drossman, D.A. (1978). The problem patient: evaluation and care of medical patients with psychosocial disturbances. Annals of Internal Medicine, 88, 366-372. Druss, B.G., Schlesinger, M., & Allen, H.M. (2001). Depressive symptoms, satisfaction with health care, and 2-year work outcomes in an employed population. American Journal of Psychiatry, 158, 731-734. Dungal, L. (1978). Physicians responses to patients: a study of factors involved in the office interview. Journal of Family Practice, 6, 1065-1073. du Pre A. (2000). Communicating about health: Current issues and perspectives. Mountain View, CA: The McGraw-Hill Companies. Dwight-Johnson, M., Sherbourne, C.D., Liao, D., & Wells, K.B. (2000). Treatment preferences among depressed primary care patients. Journal of General Internal Medicine, 15, 527-534. Emanuel, E.J., & Emanuel, L.L. (1992). Four models of the physician-patient relationship. JAMA, 267, 2221-2226. Emerson, R.M. (1962). Power-dependence relations. American Sociological Review, 27 (1), 31-41. Emerson, R.M. (1972). Exchange theory, Pa rt I: A psychological basis for social exchange. Exchange theory, Part II: Exchange relations and network structures. In J. Berger, M. Zelditch Jr., & B. Anderson (Eds.). Sociological Theories in Progress, Volume Two (pp. 38-87). Boston: Houghton Mifflin Co. Federman, A.D., Cook, E.F., Phillips, R.S., Puopolo, A.L., Haas, J.S., Brennan, T.A., et al. (2001). Intention to discontinue care among primary care patients: influence of physician behavior and process of care. Journal of General Internal Medicine, 16, 668-674. Flocke, S.A., Frank, S.H., & Wenger, D.A. (2001). Addressing multiple problems in the family practice office visit. Journal of Family Practice, 50, 211-216. Flocke, S.A., Miller, W.I., & Crabtree, B.F. (2002). Relationships between physician practice style, patient satisfaction, and attributes of primary care. Journal of Family Practice, 51, 835-840. 132

PAGE 145

Givens, J.L., Datto, C.J., Ruckdeschel, K., Knott, K., Zubritsky, C., & Oslin, D.W. (2006). Older patients aversion to antidepressants: a qualitative study. Journal of General Internal Medicine, 21, 146-151. Goldman, L.S., Nielsen, N.H., & Champion, H.C., for the Council on Scientific Affairs, American Medical Assosication (1999). Awareness, diagnosis, and treatment of depression. Journal of General Internal Medicine, 14, 569-580. Greenberg, P.E. & Birnbaum, H.G. (2005). The economic burden of depression in the US: societal and patient perspectives. Expert Opinion on Pharmacotherapy, 6, 369-376. Greenberg, P.E., Kessler, R.C., Birnbaum H.G., Leong, S.A., Lowe, S.W., Berglund, P.A., et al. (2003). The economic burden of depression in the United States: how did it change between 1990 and 2000? Journal of Clinical Psychiatry, 64, 14651475. Gross, D.A., Zyzanski, S.D., Borawski, E.A ., Cebul, R.D., & Stange, K.C. (1998). Patient satisfaction with time spent with their physician. Journal of Family Practice, 47, 133-137. Groves, J.E. (1978). Taking care of the hateful patient. New England Journal of Medicine, 298, 883-887. Hahn, S.R. (n.d.). The Difficult Doctor-Patient Relationship Questionnaire. Manual available from author. Hahn, S.R. (2000). The difficult doctor patient relationship questionnaire DDPRQ. In M. Maruish (Ed.), Handbook of Psychological Assessment in Primary Care Settings. Mahwah, NJ: Lawrence Erlbaum Publishers. Hahn, S.R. (2001). Physical symptoms and physician-experienced difficulty in the physician-patient relationship. Annals of Internal Medicine, 134(9 Pt 2) 897-904. Hahn, S.R., Kroenke, K., Spitzer, R.L., Brody, D., Williams, J.B., Linzer, M., et al. (1996). The difficult patient: prevalence, psychopathology, and functional impairment. Journal of General Internal Medicine, 11, 1-8. Hahn, S.R., Thompson, K.S., Wills, T.A., Stern, V., & Budner, N.S. (1994). The difficult doctor-patient relationship: somatization, personality and psychopathology. Journal of Clinical Epidemiology, 47, 647-657. 133

PAGE 146

Hall, J.A. & Dornan, M.C. (1988). What patients like about their medical care and how often they are asked: a meta-analysis of the satisfaction literature. Social Science and Medicine, 27, 935-939. Hall, J.A., Epstein, A.M., DeCiantis, M.L., & McNeil, B.J. (1993). Physicians liking for their patients: more evidence in the role of affect in medical care. Health Psychology, 12, 140-146. Hall, J.A., Horgan, T.G., Stein, T.S., & Roter, D.L. (2002). Liking in the physicianpatient relationship. Patient Education and Counseling, 48, 69-77. Hall, J.A., Milburn, M.A., & Epstein, A.M. (1993). A causal model of health status and satisfaction with medical care. Medical Care, 31, 84-94. Hall, J.A., Milburn, M.A., Roter, D.L., & Daltroy, L.H. (1998). Why are sicker patients less satisfied with their medical care? Tests of two explanatory models. Health Psychology, 17, 70-75. Hall, J.A., Roter, D.L., & Katz, N.R. (1988). Meta-analysis of correlates of provider behavior in medical encounters. Medical Care, 26 657-675. Hall, J.A., Roter, D.L., Milburn, M.A., & Daltroy, L.H. (1996). Patients health as a predictor of physician and patient behavior in medical visits. A synthesis of four studies. Medical Care, 34, 1205-1218. Hall, J.A. & Visser, A. (2000). Health communication in the century of the patient. Patient Education and Counseling, 41, 115-116. Harman, J.S., Schulberg, H.C., Mulsant, B.H., & Reynolds, C.F. 3rd. (2001). The effect of patient and visit characteristics on diagnosis of depression in primary care Abstract Journal of Family Practice, 50, 1068. Harris, E.C. & Barraclough, B. (1997). Suicide as an outcome for mental disorders. A meta-analysis. British Journal of Psychiatry, 170, 205-228. Hornberger, J., Itakura, H., & Wilson, S.R. (1997). Bridging language and cultural barriers between physicians and patients [Abstract]. Public Health Reports, 112, 410-417. Hornberger, J., Thom, D., & MaCurdy, T. (1997). Effects of a self-administered previsit questionnaire to enhance awareness of patients concerns in primary care. Journal of General Internal Medicine, 12, 597-606. 134

PAGE 147

Institute of Medicine (1996). Primary care: Americas health in a new era. M.S. Donaldson, K.D. Yordy, K.N. Lohr, & N.A. Vanselow (Eds.). Washington, DC: National Academies Press. Jackson, J.L. (2005). Communication about symptoms in primary care: impact on patient outcomes. The Journal of Alternative and Complementary Medicine, 11 (Suppl. 1), 51-56. Jackson, J.L., Chamberlin, J., & Kroenke, K. (2001). Predictors of patient satisfaction. Social Science & Medicine, 52, 609-620. Jackson, J.L., & Kroenke, K. (1999). Difficult patient encounters in the ambulatory clinicclinical predictors and outcomes. Archives of Internal Medicine, 159, 1069-1075. Jackson, J.L., & Kroenke, K. (2001). The effect of unmet expectations among adults presenting with physical symptoms. Annals of Internal Medicine, 134(9 Pt 2) 889-897. Katon, W.J. (2003). Clinical and health services relationships between major depression, depressive symptoms, and general medical illness. Biological Psychiatry, 54, 216-226. Keeley, R.D., Smith, J.L., Nutting, P.A., Miriam Dickinson, L., Perry Dickinson, W., & Rost, K.M. (2004). Does a depression intervention result in improved outcomes for patients presenting with physical symptoms? Journal of General Internal Medicine, 19, 615-623. Kerr, L.K., & Kerr, L.D. (2001). Screening tools for depression in primary care: the effects of culture, gender, and somatic symptoms on the detection of depression. Western Journal of Medicine, 175, 349-352. Kessler, R.C., Berglund, P., Demler, O., Jin, R., Koretz, D., Merikangas, K., et al. (2003). The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA, 289, 3095-3105. Kessler, R.C., Berglund, P., Demler, O., Jin, R., Merikangas, K.R., & Walters, E.E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62, 593-602. 135

PAGE 148

Kessler, R.C., Chiu, W.T., Demler, O., Merikangas, K.R., & Walters, E.E. (2005). Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62, 617-627. Kessler, R.C., Demler, O., Frank, R.G., Olfs on, M., Pincus, H.A., Walters, E.E., et al. (2005). Prevalence and treatment of mental disorders, 1990-2003. New England Journal of Medicine, 352, 2515-2523. Kessler, R.C., Merikangas, K.R., Berglund, P., Eaton, W.W., Koretz, D.S., & Walters, E.E. (2003). Mild disorders should not be eliminated from the DSM-V. Archives of General Psychiatry, 60, 1117-1122. Kessler, R.C., Zhao, S., Blazer, D.G., & Swartz, M. (1997). Prevalence, correlates, and course of minor depression and major depression in the national comorbidity survey. Journal of Affective Disorders, 45,19-30. Klinkman, M.S. (1997). Competing demands in psychosocial care. A model for the identification and treatment of depressive disorders in primary care. General Hospital Psychiatry, 19, 98-111. Klinkman, M.S., Schwenk, T.L., & Coyne, J.C. (1997). Depression in primary care more like asthma than appendicitis: the Michigan Depression Project Abstract Canadian Journal of Psychiatry, 42, 966-973. Kreft, I., & de Leeuw, J. (1998). Introducing multilevel modeling. Thousand Oaks, CA: Sage Publications Inc. Kroenke, K., Jackson, J.L., & Chamberlin, J. (1997). Depressive and anxiety disorders in patients presenting with physical complaints: clinical predictors and outcome. American Journal of Medicine, 103, 339-347. Kroenke, K., Spitzer, R.L., & Williams, J.B. (2003). The Patient Health Questionnaire-2: validity of a two-item depression screener. Medical Care, 41, 1284-1292. Kroenke, K., Taylor-Vaisey, A., Dietrich, A.J ., & Oxman, T.E. (2000). Interventions to improve provider diagnosis and treatment of mental disorders in primary care. A critical review of the literature. Psychosomatics, 41, 39-52. Kroll, T., Beatty, P.W., & Bingham, S. (2003). Primary care satisfaction among adults with physical disabilities: the role of patient-provider communication. Managed Care Quarterly, 11, 11-19. 136

PAGE 149

Laine, C., Davidoff, F., Lewis, C.E., Ne lson, E.C., Nelson, E., Kessler, R.C., et al. (1996). Important elements of outpatient care: a comparison of patients and physicians opinions. Annals of Internal Medicine, 125, 640-645. Lerman, C., Daly, M., Walsh, W.P., Resch, N., Seay, J., Barsevick, A., et al. (1993). Communication between patients with breast cancer and health care providers. Determinants and implications. Cancer, 72, 2612-2620. Levinson, W., & Roter, D. (1995). Physicians psychosocial beliefs correlate with their patient communication skills. Journal of General Internal Medicine, 10, 375-379. Levinson, W., Stiles, W.B., Inui, T.S., & Engle, R. (1993). Physician frustration in communicating with patients. Medical Care, 31, 285-295. Like, R., & Zyzanski, S.J. (1987). Patient satisfaction with the clinical encounter: social psychological determinants. Social Science and Medicine, 24, 351-357. Lin, E.H., Katon, W.J., Simon, G.E., Von Korff, M., Bush, T.M., Rutter, C.M., et al. (1997). Achieving guidelines for the treatment of depression in primary care: is physician education enough? Medical Care, 35, 831-842. Lin, E.H., Katon, W., Von Korff, M., Bush, T., Lipscomb, P., Russo, J., et al. (1991). Frustrating patients: physician and patient perspectives among distressed high users of medical services. Journal of General Internal Medicine, 6 241-246. Lin, E.H., Von Korff, M., Russo, J., Katon, W., Simon, G.E., Unutzer, J., et al. (2000). Can depression treatment in primary care reduce disability? A stepped care approach. Archives of Family Medicine, 9, 1052-1058. Luke, D.A. (2004). Multilevel modeling. Thousand Oaks, CA: Sage Publications. Lyness, J.M., King, D.A., Cox, C., Yoediono, Z., & Caine, E.D. (1999). The importance of subsyndromal depression in older primary care patients: prevalence and associated functional disability. Journal of the American Geriatrics Society, 47, 647-652. Marvel, M.K., Doherty, W.J., & Weiner, E. (1998). Medical interviewing by exemplary family physicians. Journal of Family Practice, 47, 343-348. Mas Garriga, X., Cruz Domenech, J.M., Fananas Lanau, N., Allue Buil, A., Zamora Casas, I., & Vinas Vidal R. (2003). Difficult patients in primary care: a quantitative and qualitative study [Abstract]. Atencion Primaria, 31, 214-219. 137

PAGE 150

McGaghie, W.C., & Whitenack, D.C. (1982). A scale for measurement of the problem patient labeling process. Journal of Nervous & Mental Disease, 170, 598-604. McQuaid, J.R., Stein, M.B., Laffaye, C., & McCahill, M.E. (1999). Depression in a primary care clinic: the prevalence and impact of an unrecognized disorder. Journal of Affective Disorders, 55, 1-10. Meredith, L.S., Orlando, M., Humphrey, N., Camp, P., & Sherbourne, C.D. (2001). Are better ratings of the patient-provider relationship associated with higher quality care for depression? Medical Care, 39, 349-360. Murray, C.J., & Lopez, A.D. (1997b). Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study. Lancet, 349, 1498-1504. Murray, C.J., & Lopez, A.D. (1997a). Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet, 349, 1436-1442. Musselman, D.L., Evans, D.L., & Nemeroff, C.B. (1998). The relationship of depression to cardiovascular disease. Epidemiology, biology, and treatment. Archives of General Psychiatry, 55, 580-592. National Alliance on Mental Illness (NAMI) (2006). About mental illness. Available from NAMI web site, http://www.nami.org/Content/NavigationM enu/Inform_Yourself/About_Mental_I llness/About_Mental_Illness.htm National Institute of Mental Health (NIMH) (2006). The numbers count: Mental disorders in America. Available from NIMH website, http://www.nimh.nih.gov/publicat/numbers.cfm Nutting, P.A., Goodwin, M.A., Flocke, S.A., Zyzanski, S.J., & Stange, K.C. (2003). Continuity of primary care: to whom does it matter and when? Annals of Family Medicine, 1, 149-155. Nutting, P.A., Rost, K., Dickinson, M., Werner, J.J., Dickinson, P., Smith, J.L., et al. (2002). Barriers to initiating depression treatment in primary care practice. Journal of General Internal Medicine, 17, 103-111. Nuyen, J., Volkers, A.C., Verhaak, P.F., Schellevis, F.G., Groenewegen, P.P., & Van den Bos, G.A. (2005). Accuracy of diagnosing depression in primary care: the impact of chronic somatic and psychiatric co-morbidity. Psychological Medicine, 35, 1185-1195. 138

PAGE 151

ODowd, T.C. (1988). Five years of heartsink patients in general practice. BMJ, 297, 528-530. OHair, D. (2003). Research traditions in provi der-consumer interaction: implications for cancer care. Patient Education and Counseling, 50, 5-8. OMalley, A.S., Forrest, C.B., & Miranda, J. (2003). Primary care attributes and care for depression among low-income African American Women. American Journal of Public Health, 93, 1328-1334. Ong, L.M., de Haes, J.C., Hoos, A.M., & Lammes, F.B. (1995). Doctor-patient communication: a review of the literature. Social Science & Medicine, 40, 903918. Ong, L.M., Visser, M.R., Lammes, F.B., & de Haes, J.C. (2000). Doctor-patient communication and cancer patients quality of life and satisfaction. Patient Education and Counseling, 41, 145-156. Orlando, M. & Meredith, L.S. (2002). Understanding the causal relationship between patient-reported interpersonal and technical quality of care for depression. Medical Care, 40, 696-704. Pearson, S.D., Katzelnick, D.J., Simon, G.E., Manning, W.G., Helstad, C.P., & Henk, H.J. (1999). Depression among high utilizers of medical care. Journal of General Internal Medicine, 14, 461-468. Raudenbush, S.W., & Bryk, A.S. (2002). Hierarch ical linear models. Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage Publications. Realini, T., Kalet, A., & Sparling, J. (1995). Interruption in the medical interaction. Archives of Family Medicine, 4, 1028-1033. Robbins, J., Hirsch, C., Whitmer, R., Cauley, J., & Harris, T. (2001). The association of bone mineral density and depression in an older population. Journal of the American Geriatrics Society, 49, 732-736. Robinson, J.W., & Roter, D.L. (1999). Psychosocial problem disclosure by primary care patients. Social Science & Medicine, 48, 1353-1362. Rogers, L.E., & Farace, R.V. (1975). Analysis of relational communication in dyads: New measurement procedures. Human Communication Research, 3, 222-239. Ross, C.K., Steward, C.A., & Sinacore, J.M. (1995). A comparative study of seven measures of patient satisfaction. Medical Care, 33, 392-406. 139

PAGE 152

Rost, K., Nutting, P., Smith, J., Coyne, J.C., Cooper-Patrick, L., & Rubenstein, L. (2000). The role of competing demands in the treatment provided primary care patients with major depression. Archives of Family Medicine, 9, 150-154. Rost, K., Zhang, M., Fortney, J., Smith, J., Coyne, J., & Smith, G.R. Jr. (1998). Persistently poor outcomes of undetected major depression in primary care. General Hospital Psychiatry, 20, 12-20. Roter, D. (n.d.). Roter Interaction Process Analysis. Retrieved February 1, 2007, from http://www.rias.org/articles Roter, D.L. (2000). The enduring and evolving nature of the patient-physician relationship. Patient Education and Counseling, 39, 5-15. Roter, D.L., Geller, G., Bernhardt, B.A., Larson, S.M., & Doksum, T. (1999). Effects of obstetrician gender on communication and patient satisfaction. Obstetrics and Gynecology, 93, 635-641. Roter, D. & Larson, S. (2002). The Roter interaction analysis system (RIAS): Utility and flexibility for analysis of medical interactions. Patient Education and Counseling, 46, 243-251. Roter, D.L., Lipkin, M., Stewart, M., & Stiles, W. and the Collaborative Study Group of the Task Force on Medical Interviewing (1988). Five communication patterns in primary care encounters [Abstract] Clinical Research 36, 749A. Roter, D.L., & McNeilis, K.S. (2003). The nature of the therapeutic relationship and the assessment of its discourse in routine medical visits. In T.L. Thompson, A.M. Dorsey, K.I. Miller, & R. Parrott (Eds.), Handbook of Health Communication (pp. 121-140). Mahwah, NJ: Lawrence Erlbaum Associates. Roter, D.L., Stewart, M., Putnam, S.M., Lipkin, M. Jr., Stiles, W., & Inui, T.S. (1997). Communication patterns of primary care physicians. JAMA, 277, 350-356. Roy-Byrne, P.P. (1996). Generalized anxiety and mixed anxiety-depression: association with disability and health care utilization. Journal of Clinical Psychiatry, 57 (Suppl. 7), 86-91. Roy-Byrne, P.P., & Katon, W. (1997). Generali zed anxiety disorder in primary care: the precursor/modifier pathway to increased health care utilization. Journal of Clinical Psychiatry, 58 (Suppl. 3), 34-38. 140

PAGE 153

Sandvik, M., Eide, H., Lind, M., Graugaard, P.K., Torper, J., & Finset, A. (2002). Analyzing medical dialogues: strength and weakness of Roters interaction analysis system (RIAS). Patient Education and Counseling, 46, 235-241. Schwenk, T.L., Marquez, J.T., Lefever, R.D., & Cohen, M. (1989). Physician and patient determinants of difficult physician-patient relationships. The Journal of Family Practice, 28, 59-63. Shvartzman, P., Weiner, Z., Vardy, D., Friger, M., Sherf, M., & Biderman, A. (2005). Health services utilization by depressive patients identified by the MINI questionnaire in a primary care setting. Scandinavian Journal of Primary Health Care, 23, 18-25. Schonfeld, W.H., Verboncoeur, C.J., Fifer, S.K., Lipschutz, R.C., Lubeck, D.P., & Buesching, D.P. (1997). The functioning and well-being of patients with unrecognized anxiety disorders and major depressive disorder. Journal of Affective Disorders, 43, 105-119. Simon, G.E., Goldberg, D., Tiemens, B.G., & Ustun, T.B. (1999). Outcomes of recognized and unrecognized depression in an international primary care study. General Hospital Psychiatry, 21, 97-105. Sixma, H.J., Spreeuwenberg, P.M., & van der Pasch, M.A. (1998). Patient satisfaction with the general practitioner: A two-level analysis. Medical Care, 36, 212-229. Sleath, B., & Rubin, R.H. (2002). Gende r, ethnicity, and physician-patient communication about depression and anxiety in primary care. Patient Education and Counseling, 48, 243-252. Smith-Lovin, L. (2001). Social psychology. In J.R. Blau (Ed.), The Blackwell Companion to Sociology (pp. 407-420). Malden, MA: Blackwell Publishers, Inc. Solberg, L.I., Fischer, L.R., Rush, W.A., & Wei, F. (2003). When depression is the diagnosis, what happens to patients and are they satisfied? American Journal of Managed Care, 9, 131-140. Solberg, L.I., Korsen, N., Oxman, T.E., Fischer, L.R., & Bartels, S. (1999). The need for a system in the care of depression. Journal of Family Practice, 48, 973-979. Spitzer, R.L., Williams, J.B., Kroenke, K., Li nzer, M., deGruy, F.V., Hahn, S.R., et al. (1994). Utility of a new procedure for diagnosing mental disorders in primary care: The PRIME-MD 1000 study. JAMA, 272, 1749-1756. 141

PAGE 154

Stange, K.C., Zyzanski, S.J., Jan, C.R., Callahan, E.J., Kelley, R.B., Gillanders, W.R., et al. (1998). Illuminating the black box. A description of 4,454 patient visits to 138 family physicians. The Journal of Family Practice, 46, 377-389. Stein, M.B., Cox, B.J., Afifi, T.O., Belik, S.L., & Sareen, J. (2006). Does co-morbid depressive illness magnify the impact of chronic physical illness? A populationbased perspective. Psychological Medicine, 36, 587-596. Steinmetz, D., & Tabenkin, H. (2001). The d ifficult patient as perceived by family physicians. Family Practice, 18, 495-500. Stewart, M.A. (1995). Effective physician-patient communication and health outcomes: a review. Canadian Medical Association Journal, 152, 1423-1433. Suchman, A.L., Markakis, K., Beckman, H.B., & Frankel, R. (1997). A model of empathic communication in the medical interview. JAMA, 277, 678-682. Tabachnick, B.G., & Fidell, L.S. (2001). Using multivariate statistics (4th ed.). Boston: Allyn & Bacon. Telford, R., Hutchinson, A., Jones, R., Rix, S ., & Howe, A. (2002). Obstacles to effective treatment of depression: a general practice perspective. Family Practice, 19, 4552. Thompson, T.L., & Parrott, R. (2002). Interpersonal communication and health care. In M.L. Knapp & J.A. Daly (Eds.), Handbook of Interpersonal Communication (3rd ed.) (pp. 680-725) Thousand Oaks, CA: Sage. Unutzer, J., Katon, W.J., Russo, J., Sim on, G., von Korff, M., & Lin, E. (2003). Willingness to pay for depression treatment in primary care. Psychiatric Services, 54, 340-345. U.S. Census Bureau. (2005, June). Annual estimates of the population by selected age groups and sex for the United States: April 1, 2000 to July 1, 2004 (NCEST2004-02). Available from U.S. Census Bureau web site, http://www.census.gov U.S. Preventive Services Task Force (2002). Screening for depression. Available from Agency for Healthcare Research and Quality (AHRQ) website, http://www.ahrq.gov/clinic/uspstf/uspsdepr.htm 142

PAGE 155

van Os, T.W., van den Brink, R.H., Tiemens, B.G., Jenner, J.A., van der Meer, K., & Ormel, J. (2005). Communicative skills of general practitioners augment the effectiveness of guideline-based depression treatment. Journal of Affective Disorders, 84, 43-51. Von Korff, M., & Simon, G. (1996). The prevalence and impact of psychological disorders in primary care: HMO research needed to improve care. HMO Practice 10, 150-155. Ware, J.E. Jr., Snyder, M.K., Wright, W. R., & Davies, A.R. (1983). Defining and measuring patient satisfaction with medical care. Evaluation and Program Planning, 6, 247-263. Watts, S.C., Bhutani, G.E., Stout, I.H., Ducker, G.M., Cleator, P.J., McGarry, J., et al. (2002). Mental health in older adult recipients of primary care services: is depression the key issue? Identification, treatment, and the general practitioner. International Journal of Geriatric Psychiatry, 17, 427-437. Weiss, G.L. (1988). Patient satisfaction with primary care: evaluation of sociodemographic and predispositional factors. Medical Care, 26, 383-392. Williams, B., Coyle, J., & Healy, D. (1998). The meaning of patient satisfaction: an explanation of high reported levels. Social Science and Medicine, 47, 1351-1359. Williams, J.W., Mulrow, C.D., Kroenke, K., Dhanda, R., Badgett, R.G., Ormori, D., et al. (1999). Case-finding for depression in primary care: a randomized trial. American Journal of Medicine, 106, 36-43. Williams, S., Weinman, J., & Dale, J. (1998). Doctor-patient communication and patient satisfaction: a review. Family Practice, 15, 480-492. Wissow, L.S., Roter, D., Larson, S.M., Wang, M.C., Hwang, W.T., Johnson, R., et al (2002). Mechanisms behind the failure of residents longitudinal primary care to promote disclosure and discussion of psychosocial issues. Archives of Pediatrics & Adolescent Medicine, 156, 685-692. Wittchen, H.U., Lieb, R., Wunderlich, U., & Schuster, P. (1999). Comorbidity in primary care: presentation and consequences. Journal of Clinical Psychiatry, 60 (Suppl. 7), 29-36. Wu, L.R., Parkerson, G.R. Jr., & Doraiswamy, P.M. (2002). Health perception, pain, and disability as correlates of anxiety and depression symptoms in primary care patients. Journal of the American Board of Family Practice, 15, 183-190. 143

PAGE 156

Zandbelt, L.C., Smets, E., Oort, F.J., Godfried, M.H., de Haes, H.C. (2004). Satisfaction with the outpatient encounter: a comparison of patients and physicians views. Journal of General Internal Medicine, 19, 1088-1095. 144

PAGE 157

Appendixes 145

PAGE 158

Appendix A: IRB Approval Letters 146

PAGE 159

Appendix A (Continued) 147

PAGE 160

Appendix A (Continued) 148

PAGE 161

Appendix A (Continued) 149

PAGE 162

Appendix B: Site Approval Letter 150

PAGE 163

Appendix C: Provider Demographic Questionnaire Provider Questionnaire 1. What is your gender? Female Male 2. What is your age? _______ years 3. What is your race? American Indian or Alaskan Native Native Hawaiian or Pacific Islander Asian White [ Black or African American 4. What is your ethnicity? Hispanic or Latino Not Hispanic or Latino 5. What is your title? MD DO ARNP Other __________________ 6. How would you primarily define your clinical practice? Internal Medicine Geriatrics Family Practice Other ______________________ 7. How many years have you been in practice? Less than one year 6-10 years Over 20 years 1-5 years 11-20 years 8. On average, how many patients do you see in a typical day? _______ a week? ______ 9. What percentage of your patient population would you consider difficult? _______% 10. On a scale of 1-to-10 (10 being most profic ient) please rate yourself on the following: Your overall personal knowledge about depression? _________ Knowledge and ability to diagnose depression in your primary care patients? ________ Knowledge and ability to treat depression in your primary care patients? _________ 11. On a scale of 1-to-10 (10 being a great deal) to wh at extent do you feel the following negatively affect the care your patients receive for depression: Time constraints and multiple competing demands on providers _________ Patients dont provide you with enough information _________ Lack of availability or access to qua lified mental health specialists _________ Patient reluctance to go to mental health specialists _________ Other _____________________________________________________________ 151

PAGE 164

Appendix D: Patient Demographic Questionnaire Patient Questionnaire ___________________________________________________________________________ For each question, check the one best answer in the box provided. 1. Is this the first time you are seeing this doctor or nurse practitioner? [ ] Yes If you answered yes, please STOP [ ] No You are not eligible to participate in this study. 2. What is your gender? [ ] Female [ ] Male 3. What is your age? __________ years 4. What is your race? [ ] American Indian or Alaska Native [ ] Native Hawaiian or Pacific Islander [ ] Asian [ ] White [ ] Black or African American 5. What is your ethnicity? [ ] Hispanic or Latino [ ] Not Hispanic or Latino 6. What is/was your lifes major occupation? [ ] Professional [ ] Laborer ] Technical [ ] Housewife [ ] Clerical [ ] Other __________________ 7. Are you retired? [ ] Yes [ ] No 8. What is your highest level of education? [ ] Grade School [ ] College Degree [ ] High School [ ] Graduate Degree [ ] Technical School 9. What is your current marital status? [ ] Married Divorced [ ] Single [ ] Widowed 10. Why are you seeing the doctor or nurse practitioner today? _________________________________________________________________________ 152

PAGE 165

Appendix D (Continued) 11. Did you ask for this appointment today or did your doctor tell you to come in today? I made this appointment My doctor wanted me to come in 12. Have you been told that you have a serious medical problem? [ ] No [ ] Yes, Please list your diagnosis: ______________________________________________________________ For the following questions, please circle the number that most closely reflects how you feel. 13. Do you think it is important for your doctor (or nurse practitioner) to address your emotional health needs? Very Somewhat Somewhat not Not at all Important Important Unsure Important Important 1 2 3 4 5 14. To what extent do you think the amount of time you spend with your doctor (or nurse practitioner) affects the quality of care you receive? Very Somewhat Somewhat not Not at all Important Important Unsure Important Important 1 2 3 4 5 153

PAGE 166

Appendix E: BDI-II Invoice (Authorization) 154

PAGE 167

Appendix F: BDI-II Information The Beck Depression Inventory II is a registered trademark of The Psychological Corporation, Harcourt Brace & Company, San An tonio, Texas. Purchasing information is available online at http://harcourtassessment.com 155

PAGE 168

Appendix G: DDPRQ Authorization Letter 156

PAGE 169

Appendix H: DDPRQ-10 Questionnaire PRACTITIONER PERCEPT ION QUESTIONNAIRE (DDPRQ-10) Directions: CIRCLE the most appropriate number and SEAL this form in the ENVELOPE provided. Please return the sealed envelope directly to the study investigator. A Great Not Deal At All 1. How much are you looking forward to this patients 1 2 3 4 5 6 next visit after seeing this patient today? 2. How frustrating do you find this patient? 1 2 3 4 5 6 3. How manipulative is this patient? 1 2 3 4 5 6 4. To what extent are you frustrated by this patients 1 2 3 4 5 6 vague complaints? 5. How self-destructive is this patient? 1 2 3 4 5 6 6. Do you find yourself secretly hoping that this patient 1 2 3 4 5 6 will not return? 7. How at ease did you feel when you were with this patient 1 2 3 4 5 6 today? 8. How time consuming is caring for this patient? 1 2 3 4 5 6 9. How enthusiastic do you feel about caring for this patient? 1 2 3 4 5 6 10. How difficult is it to communicate with this patient? 1 2 3 4 5 6 157

PAGE 170

Appendix I: RIAS Agreement Letter 158

PAGE 171

Appendix J: RIAS Information The Roter Interaction Analysis System (RIAS) is copyright property of Dr. Debra Roter, The Johns Hopkins University, Baltimore, Maryland. Information is available online at http://www.rias.org 159

PAGE 172

Appendix K: Patient Satisfaction Survey Authorization Letter 160

PAGE 173

Appendix L: Patient Satisfaction Survey Patient Post-Visit Survey The following statements are things people sometimes say about doctor vi sits (or nurse practitioner visits NP). This is just asking for your opinion. There are no right or wrong answers. Based on your visit TODAY please CIRCLE the answer that best reflects how you feel. When you are finished, please put this survey in the envelope SEAL IT and give it to the receptionist before you leave the office today Thank you. Strongly Strongly Agree Agree Unsure Disagree Disagree 1. This was a very satisfying visit. 1 2 3 4 5 2. My doctor/NP was very careful to check everything 1 2 3 4 5 when examining me. 3. My doctor/NP encouraged me to talk about my 1 2 3 4 5 worries. 4. My doctor/NP interrupted me. 1 2 3 4 5 5. My doctor/NP is competent and well trained. 1 2 3 4 5 6. My doctor/NP acted bored at times during my 1 2 3 4 5 visit today. 7. My doctor/NP MISSED important information 1 2 3 4 5 that I gave him/her. 8. My doctor/NP was NOT as thorough as he/she 1 2 3 4 5 should have been. 9. The medical problems I have had in the past 1 2 3 4 5 were IGNORED during my visit today. 10. My doctor/NP and I laughed and joked together 1 2 3 4 5 during my visit. 11. My doctor/NP asked if I understood the 1 2 3 4 5 information he/she gave me about my condition or treatment. 12. My doctor/NP has a good understanding of my 1 2 3 4 5 past health history. 161

PAGE 174

Appendix L (Continued) Strongly Strongly Ag ree Agree Unsure Disagree Disagree 13. I have great confidence in my doctor/NP. 1 2 3 4 5 14. Im very satisfied with the medical care 1 2 3 4 5 I received. 15. My doctor/NP really seemed to care about me 1 2 3 4 5 and my health problems. 16. My doctor/NP explained things in words I 1 2 3 4 5 could understand. 17. I depend on my doctor/NP in order to feel better 1 2 3 4 5 both physically and emotionally. 18. My doctor/NP told me exactly what he/she 1 2 3 4 5 planned to do next in my treatment. 19. My doctor/NP seemed annoyed today. 1 2 3 4 5 20. My doctor/NP told me all I wanted to know 1 2 3 4 5 about my condition and treatment. 21. My doctor/NP knew what health problems 1 2 3 4 5 I wanted to talk about today. 22. My doctor/NP encouraged me to tell him/her 1 2 3 4 5 everything I thought was important. 23. I count on my doctor/NP to set my mind at ease 1 2 3 4 5 when I am worried. 24. My doctor/NP had a complete underanding 1 2 3 4 5 st of the things that are wrong with me. 162

PAGE 175

Appendix L (Continued) Strongly Strongly Agree Agree Unsure Disagree Disagree 25. My doctor/NP seemed to be in a hurry. 1 2 3 4 5 26. My doctor/NP DID NOT explain my medical 1 2 3 4 5 problems to me. 27. I think all the health problems we discussed 1 2 3 4 5 today were important. 28. I have health problems which should have 1 2 3 4 5 been discussed today but were NOT. 29. My doctor/NP asked for my opinion when 1 2 3 4 5 trying to decide on the best way to treat my problem. 30. My doctor/NP seemed nervous today. 1 2 3 4 5 31. My doctor/NP clearly explained why I should 1 2 3 4 5 do the things he/she asked me to do. 32. My doctor/NP acted as though he/she were 1 2 3 4 5 doing me a favor by talking to me. 33. My doctor/NP tells me if he/she is worried 1 2 3 4 5 about my condition or how I am doing. 34. My doctor/NP answered all my questions. 1 2 3 4 5 35. My doctor/NP made me feel important. 1 2 3 4 5 36. My doctor/NP acted bossy and domineering 1 2 3 4 5 at times during my visit today. Thank you for your time. Please put this survey in the envelope SEAL IT and give it to the receptionist before you leave 163

PAGE 176

Appendix M: Provider Recruitment Letter Date Name Address [City], [State] zip Dear Name : I have been practicing with Anchor Health Centers as an ARNP for a little over five years now. In addition, I have been pursuing my Ph.D. at the University of South Florida for the past three years. I am now actively working on my doctoral dissertation research and am recruiting and inviting Anchor providers and patients to participate in my study. I am examining patients depressive symptoms, the patient-provider relationship, and patient-provider communi cation during the primary care encount er and am recruiting Family Practice, Internal Medicine, a nd Geriatric providers and their established patients over 21 years of age. I have designed the study so that yo ur time commitment would be minimal. I would ask you to (1) complete a one-page demogra phic questionnaire, (2) allow me to audiotape your office encounters with ten of your established patients, and (3) ask you to complete one 10-item questionnaire after seeing each patient that takes less than one minute to complete. This study has been has been approved by the University of South Florida Institutional Review Board. Several Anchor providers participated in a sim ilar pilot study I conducted last year and they reported that participation and data collection resulted in little, if any, interruption to their (or their staff members) daily routine. I would be on-site in your office for all data collection and would ask only that (1) your front offi ce staff hand a flyer to each patient upon check-in, and (2) your nurse/medical assistant ask each patient if they are interest ed in hearing more about the study once they are brought to the exam room. Your staff would have no further involvement. I would like an opportunity to meet with you at your convenience to discuss the study in detail and answer any questions you may have. I will phone your office early next week and will ask to speak with your office manager Name Successful completion of my dissertation is the final requirement I need to fulfill for my degree. Your participation and support of my research would be so greatly appreciated! Sincerely, Lorraine M. Novosel, PhD(c),ARNP,CS [address] [City], [State] [zip] [phone] [pager] 164

PAGE 177

Appendix N: Flyer 165

PAGE 178

Appendix O: Provider Consent Forms 166

PAGE 179

Appendix O (Continued) 167

PAGE 180

Appendix O (Continued) 168

PAGE 181

Appendix O (Continued) 169

PAGE 182

Appendix P: Patient Consent Forms 170

PAGE 183

Appendix P (Continued) 171

PAGE 184

Appendix P (Continued) 172

PAGE 185

Appendix P (Continued) 173

PAGE 186

Appendix Q: HIPAA Authorization Form ANCHOR HEALTH CENTERS REQUEST FOR AUTHORIZATION TO RELEASE PROTECTED HEALTH INFORMATION TO THIRD PARTIES By signing this authorization, I authorize Anchor Health Centers (AHC) to use and/or disclose certain protected health information (PHI) about me to, or for, the party, or parties, listed below. This authorization permits AHC to use or disclose the following individually identifiable health information to: The Johns Hopkins University, Bloomberg School of Public Health, Department of Health Policy and Management Person or entity to receive the information 624 North Broadway, Baltimore, Maryland 21205 Address Specifically describe the information to be released, such as date(s) of service, level of detail to be released, origin of information, purpose of the disclosure etc. using the space below. The audiotaped recording of todays office visit will be released to researchers at the above entity for the purpose of having the audiotapes coded. This authorization pertains only to the event of disclosure. If another disclosu re is required, I understand that a second authorization must be si gned. I understand that this authori zation will expire at the end of the research study. When my information is used or disclosed pursuant to this authorization, it may be subject to re-disclosure by the recipient and may no longer be protected by the federal HIPAA Privacy Rule. I have the right to revoke this authorization in writi ng except to the extent that AHC has acted in reliance upon this authorization. My written revocation must be submitted in writing using the reverse of this form. Signed by: _______________________________ ____________________________ Signature of Patient or Legal Guardian Relationship to Patient _______________________________ ____________________________ Print Patients Name Date _______________________________ Print Name of Legal Guardian 174

PAGE 187

Appendix Q (Continued) ANCHOR HEALTH CENTERS REQUEST FOR AUTHORIZATION TO RELEASE PROTECTED HEALTH INFORMATION TO THIRD PARTIES By signing this revocation, I am reversing the authori zation that I previously gave Anchor Health Centers (AHC) to use and/or disclose certain protected health information (PHI) about me to or for the party, or parties, listed on the front of this form. I understand that disclosures made prior to this revocati on were made with my authorization and that I must contact those entities who received this information directly if I do not want them to further use or disclose this information. Signed by: _______________________________ ____________________________ Signature of Patient or Legal Guardian Relationship to Patient _______________________________ ____________________________ Print Patients Name Date _______________________________ Print Name of Legal Guardian 175

PAGE 188

Appendix R: Patient Notification of Withdrawal from Study Date: ______________ Your answers to the questions you gave earlier show that you have symptoms that suggest you may be suffering severe depression. Because this is so important to your health and safety, you are no longer eligible to be a part of this study and are being withdrawn immediately. Your doctor or nurse practitioner is being notified of your results, and you will have your office visit with them as planned. Thank you for your interest in participating in this study. Score: _____________ 176

PAGE 189

Appendix S: Provider Notification of Patient Withdrawal from Study Date: ______________ Dear Dr/ARNP ________________________________: Your patient, __________________________________, was screened for depression today as part of the study examining patient-provider communication during the primary care visit. In the best interest of this patients health and safety, it is important to notify you that they responded affirmatively to: Item #2: Pessimism / Hopelessness Item #9: Suicidal Thoughts or Wishes The Beck Depression Inventory-II Manual indicates such individuals should be closely scrutinized for suicide potential (Beck, et al., 1996). The patient has been given oral and written notification of these findings. In the best interest of their health and safety, they are being withdrawn from the study at this time. The patient is aware that they are no longer eligible to participate in the study and that you are being notified of these results. _____________________________________ Signature of Investigator Beck, A.T., Steer, R.A., & Brown, G.K. (1996). Beck Depression Inventory manual (2nd ed.). San Antonio, TX: The Psychological Corporation. 177

PAGE 190

Appendix T: Patient Notification of BDI-II Results Date: ______________ Your answers to the questions you gave earlier indicate: ____ You have symptoms that may be due to mild depression. It is recommended that you discuss this or show this paper to your doctor or nurse practitioner. The office staff will be glad to schedule an appointment for you to discuss this in private. ____ You have symptoms that are commonly due to depression. It is highly recommended that you discuss this or show this paper to your doctor or nurse practitioner as soon as you can The office staff will be glad to schedule an appointment for you to discuss this in private. ____ You have symptoms that suggest you may be suffering severe depression. It is strongly recommended that you discuss this or show this paper to your doctor or nurse practitioner as soon as you can The office staff will be glad to schedule an appointment for you to discuss this in private. Because this is so important to your health and safety, your doctor or nurse practitioner will be notified of this finding. Score: _____________ 178

PAGE 191

Appendix U: Provider Notification of BDI-II Results Date: ______________ Dear Dr/ARNP ________________________________: Your patient, __________________________________, was screened for depression today as part of the study examining patient-provider communication during the primary care visit. In the best interest of this patients health and safety, it is important to notify you that their score was ________. According to published scoring guidelines, this result indicates: minimal depressive symptoms (0-13) mild depression (14-19) moderate depression (20-28) extreme depression (29-63) The patient has been given oral and written notification of the score and it has been recommended to them that they discuss their score with you as soon as possible. The patient is aware that this score is being shared with you in the best interest of their health and safety. _____________________________________ Signature of Investigator 179

PAGE 192

180 About the Author Lorraine Marie Novosel, MSN, ARNP, CS, has been in the nursing profession for more than 20 years. She maintains dual national certification as an adult and gerontological nurse practitioner through the ANCC and currently practices internal medicine in Naples, Florida. She holds an A.A.S. from Youngstown State University (1983), B.S. in Nursing from Pennsylvania St ate University (1992), and M.S.N. majoring as an adult nurse practitioner from Kent State University (1995) where she received the Linnea Henderson Award for Academic Excellence. She will receive her Ph.D. in Nursing from the University of South Florida in 2007. Ms. Novosel holds professional licensure in Florida, Ohio, and Pennsylvania and is a member of Phi Kappa Phi National Honor Society, Sigma Theta Tau International, American Academy of Nurse Practitioners, American Nurses Association, and the Florida Nurses Association. She is a published author and has presented her research at the state and national levels.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001920968
003 fts
005 20080114141656.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080114s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001866
035
(OCoLC)190791030
040
FHM
c FHM
049
FHMM
090
RT41 (ONLINE)
1 100
Novosel, Lorraine Marie.
0 245
Depressive symptomatology, patient-provider communication, and patient satisfaction :
b a multilevel analysis
h [electronic resource] /
by Lorraine Marie Novosel.
260
[Tampa, Fla] :
University of South Florida,
2007.
3 520
ABSTRACT: Depression can be a profoundly disabling and costly disorder and is a major public health concern. Despite the efficacy of treatment options, it is often unrecognized, under-diagnosed, and inadequately treated in primary care settings. Research on patient-provider communication supports the connection among the quality of the patient-provider interaction, patient behavior, and health outcomes. The purpose of this study was to systematically examine the impact of patients' depressive symptoms on the patient-provider relationship, patient-provider communication, and patient satisfaction with the primary care office visit. One hundred twenty three patient-provider encounters were audiotaped and coded using the Roter Interaction Analysis System (RIAS).^ ^^A2 x 2 x 2 within-subjects factorial model provided the analytic framework for examining eight verbal communication behaviors categorized by speaker (patient or provider), type of utterance (question or information giving), and content of utterance (medical or psychosocial talk). Hierarchical linear modeling was used to analyze the two-level nested structure of the data. Results indicated that depression is associated with, but does not predict, increased provider-perceived difficulty in the patient-provider relationship. There was no significant change in either patient or provider communication behavior in relation to the severity of patients' depressive symptoms. Significantly more provider medical information was given during encounters with "difficult" patients and this behavior had a consistent negative effect on patient satisfaction.^ ^Patient-provider communication, by itself, does not appear to be a source of depressed patients' oft-reported dissatisfaction with medical care. Additional research is needed to further understand the core processes and structures of primary care practice in relation to the diagnosis and management of depression, their effect on patient outcomes, and to uncover opportunities for enhancing the effectiveness of depression care in primary care.
502
Dissertation (Ph.D.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 180 pages.
Includes vita.
590
Adviser: Jason W. Beckstead, Ph.D.
653
Depression.
Difficult patient.
Patient-provider relationship.
RIAS.
DDPRQ.
690
Dissertations, Academic
z USF
x Nursing
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1866