Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Psychometric assessment of the Multidimensional Quality of Life Questionnaire for Persons Infected with… Remple, Valencia Paulette 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-0223.pdf [ 6.91MB ]
Metadata
JSON: 831-1.0090275.json
JSON-LD: 831-1.0090275-ld.json
RDF/XML (Pretty): 831-1.0090275-rdf.xml
RDF/JSON: 831-1.0090275-rdf.json
Turtle: 831-1.0090275-turtle.txt
N-Triples: 831-1.0090275-rdf-ntriples.txt
Original Record: 831-1.0090275-source.json
Full Text
831-1.0090275-fulltext.txt
Citation
831-1.0090275.ris

Full Text

(  PSYCHOMETRIC ASSESSMENT OF THE MULTIDIMENSIONAL QUALITY OF LIFE QUESTIONNAIRE FOR PERSONS INFECTED WITH HIV OR WITH AIDS IN A COHORT OF HIV-INFECTED WOMEN by VALENCIA PAULETTE REMPLE B.S.N., The University of British Columbia, 1997 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN NURSING in THE FACULTY OF GRADUATE STUDIES (School of Nursing) We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA April 2002 © Valencia Paulette Remple, 2002  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  School of Nursing The University of British Columbia Vancouver, Canada Date  Abstract This methodological study was designed to assess the psychometric properties of a relatively new HIV-specific quality of life questionnaire, the Multidimensional Quality of Life Questionnaire for Persons Living with HIV or with AIDS (MQOL-HIV), in a cohort of Canadian, HIV-infected women. The study was prompted by an identified lack of information regarding the validity and reliability of HIV-specific quality of life (QOL) scales in populations of HIV-infected women. The specific analyses conducted were content validity, concurrent validity, internal consistency, and test-retest reliability. The study was conducted in the fall of 2001 at the Oak Tree Clinic, British Columbia's outpatient referral centre for women, children, and families living with HIV. Seven health care professionals and six patients participated in the content validity testing, which involved rating the validity of the 40 items and 10 domains of the MQOL-HIV. Eighty-five patients participated in the concurrent validity and internal consistency testing, which involved baseline completion of the MQOL-HIV and the Medical Outcomes Study - HIV Health Survey (MOSHIV) during a clinic visit. Forty-three patients participated in the test-retest reliability testing, which involved completing a second MQOL-HIV two weeks after baseline. Content validity findings indicated that most of the items and all of the domains of the scale were valid for HIV-infected women, but that revision and replacement of some of the items were needed. Several important domains and aspects of domains were identified as missing from the scale. Five domains, the Summary Index, and the overall scale demonstrated strong concurrent validity when correlated with similar domains contained in the more commonly used MOS-HIV. Most of the domains demonstrated strong internal consistency and test-retest reliability.  The findings were consistent with the literature, indicating that QOL scales require revalidation in new populations, and that dimensions and aspects of QOL may differ for HIVinfected women. Although the MQOL-HIV was developed with HIV-infected patients, some of whom were women, the need for subsequent validation in female populations was underscored. The content validity findings supported those found by others in that the QOL domains and aspects of salience may vary depending on gender and cultural context. This study represents one of the first of its kind to validate an HIV-specific QOL scale in women, and provides preliminary findings in terms of the unique dimensions of QOL for HIV-infected women.  Table of Contents Abstract List of Tables  ii viii  List of Figures  x  Acknowledgements  xi  CHAPTER 1 Introduction  1  Background and Problem Identification Statement of Purpose Research Questions Significance of the Study Organization of the Thesis CHAPTER 2 Review of the Literature  1 4 4 4 5  6  Conceptualization of Quality of Life 7 Historical Background 7 Uses of the Concept 8 Definitions of Quality of Life 10 Health-related definitions of quality of life 11 Achievement-related definitions of quality of life 13 Happiness-related definitions of quality of life 14 Satisfaction-related definitions of quality of life 14 Well-being-related definitions of quality of life 15 Utility-related definitions of quality of life 16 Dimensions of Quality of Life 16 Essential Attributes of Quality of Life 20 Quality of Life of Particular Populations 22 Quality of life in HIV-infected populations 22 Gender differences in QOL 25 Cultural considerations 28 Measuring Quality of Life 29 How Quality of Life is Operationalized and Measured 29 Types of Scales 31 How Scales are Developed and Tested 33 Development 33 Assessing an instrument's psychometric properties 37 The Multidimensional Quality of Life Questionnaire for Persons Infected with HIV or with AIDS (MQOL-HIV) 42 Description, Scoring, and Administration 42 Development 43 iv  Psychometric Properties Content validity Criterion-related validity Construct validity Reliability Responsiveness Critique of the Instrument Summary CHAPTER 3 Methods  44 44 44 46 46 47 48 49 51  Design 51 Context of the Study: The Oak Tree Clinic 52 Sampling Methods 52 Content Validity Testing 52 Concurrent Validity and Reliability Testing ...53 Patient Sample Recruitment Plan and Data Collection Procedures 54 Description of Methods for Each Type of Validity and Reliability Testing 54 Content Validity 54 Concurrent Validity 56 Test-Retest Reliability 56 Internal Consistency 57 Missing Data 57 Instrumentation 58 Multidimensional Quality of Life Questionnaire for Persons Infected with HIV or with AIDS (MQOL-HIV) 58 Medical Outcomes Study - HIV Health Survey (MOS-HIV) 58 Socio-demographic and Health Information Forms 61 Quality of Life Status Change Question 62 Protection of Human Rights 62 Summary 63 CHAPTER 4 Findings Characteristics of the Sample The Patients and Health-Care Professionals Who Comprised the Expert Panel and Work Groups The Patients Who Responded to the Questionnaires Results for Each Type of Validity and Reliability Testing Research Question #1: Content Validity CVI of the items CVI of the domains Research Question #2:Concurrent Validity Research Question #3: Internal Consistency Research Question #4: Test-Retest Reliability Summary  65 65 65 66 72 75 75 81 84 85 86 87  v  CHAPTER 5 Discussion The Sample The Expert Panel Members The Patients Who Completed the Questionnaires Discussion of the Major Findings Content Validity Content validity of the items Content validity of the domains and missing aspects and domains Concurrent Validity Internal Consistency Test-Retest Reliability Methodological Considerations Summary CHAPTER 6 Summary, Limitations, Conclusions, Implications, and Recommendations Study Summary Study Limitations Conclusions Implications for Nursing Practice, Research, and Theory Recommendations for Future Research  89 89 89 90 91 91 91 95 101 102 103 105 107 109 109 110 113 114 116  References  119  Appendix A Dimensions of QOL as Operationalized by Various Authors  136  Appendix B Socio-Demographic Questionnaire  137  Appendix C Rating Form for Health-Care Professional Members  140  Appendix D Patient Panel Scoring Sheet  145  Appendix E Chart Review Data Collection Form  148  Appendix F Informed Consent Form - Patient Version  149  Appendix G Informed Consent Form - Patient Work Group Version  152  vi  Appendix H Informed Consent Form - Health-Care Professional Version  List of Tables Table 1 Elements of the Ferrans' (1996) Conceptual Model of Quality of Life Table 2 Medical Outcomes Study Conceptual Framework Table 3 fflV-Specific QOL Instruments  19 20 24  Table 4 Example of Content Validity Matrix for Hypothetical QOL Scale  39  Table 5 MQOL-HIV Domain-Specific Criterion-Related Correlations for the MQOL-HIV Domains  45  Table 6 MQOL-HIV Test-Retest Reliability and Internal Consistency  47  Table 7 Dimensions, Subscales, and Number of Items in the MOS-HTV  59  Table 8 Demographic Characteristics of the Sample  67  Table 9 HIV Acquisition and Substance Use Characteristics of the Sample  70  Table 10 HIV and Clinical Characteristics of the Sample  71  Table 11 Distribution of the MQOL-HIV Domain Scores  73  Table 12 Distribution of the MQOL-HIV Summary Index and Total Instrument Scores  74  Table 13 Content Validity Index (CVI) of the MQOL-HIV Items as Rated by the Health-Care Professional Expert Panel (HCP) and Patient Work Groups (PWG)  76  Table 14 Content Validity Index (CVI) of 10 Domains of the MQOL-HIV, as Rated by the Health-Care Professional Expert Panel (HCP) and the Patient Work Groups (PWG)  81  Table 15 Domains and Aspects Identified as "Missing" from the MQOL-HIV by Health Care Professionals (HCP) and Patient Workgroup Participants  82  Table 16 Domain Concurrent Validity Analysis  85  Table 17 MQOL-HIV Reliability Analyses  86  ix  List of Figures Figure 1 Data Collection Sequence for Reliability and Validity Testing  51  Figure 2 Single-Item QOL Status Change Question  62  x  Acknowledgements I wish to thank the members of my thesis committee for their guidance and support throughout this project. To my chair, Dr. Ann Hilton, thank you for your guidance and consistent, constructive feedback. To Dr. Pamela Ratner, thank you for contributing such excellent suggestions. To the third member of my committee, Anne Wyness, thank you for providing your insightful perspective on this work. I also wish to thank my husband, James. Your tremendous patience and encouragement during this project were invaluable. I could not have completed this without your support. Lastly, acknowledgement goes to the staff and patients at the Oak Tree Clinic. To my employer and thesis consultant, Dr. David Burdge, thank you for your unwavering support and encouragement. To Evelyn Maan, I couldn't have completed this study without your tireless efforts. Finally, I wish to thank the women of the Oak Tree Clinic, who gave of their time and themselves for this project.  xi  CHAPTER 1: INTRODUCTION In the Spring of 2000, the multidisciplinary clinical and research team of the Oak Tree Clinic, the British Columbia outpatient referral centre for women, children, and families infected with the human immunodeficiency virus (HIV), identified the assessment of patients' quality of life (QOL) as an important priority for both research and clinical care. As the clinic research program coordinator, my interest in the concept of QOL was subsequently piqued, and I began to explore the literature in an effort to identify an appropriate measurement scale for our clinic population. It became rapidly apparent, however, that not only was an appropriate instrument not readily available, but there existed no agreement on the conceptual definition of QOL and its operationalization. As I delved further into the literature, I realized that measuring a complex concept such as QOL was not a simple matter of administering the most popular QOL questionnaire. This realization became the impetus for this project, as I sought a promising HIVspecific QOL scale that could be validated in our population of HIV-infected women. Background and Problem Identification The human immunodeficiency virus is a retrovirus that impairs the immune system, resulting in the eventual development of opportunistic infections and malignancies that characterize the acquired immune deficiency syndrome (AIDS) (Center for Disease Control, 1993). Recent treatment advances include the development of combination highly active antiretroviral therapy (HAART) that can now reduce the level of viral load, restore portions of the immune system, and prevent opportunistic infections (Li et al., 1998; Rosenberg et al., 1997). Over the past several years, dramatic improvements in HIV treatment have resulted in declines in hospitalization rates, morbidity, and mortality (Carpenter et al., 2000; Marschner et al., 1998; Palellaet al., 1998). For individuals living with HIV infection, it is not only the duration of their life that is of concern, but also the quality of their life (Leplege et al., 1997). Despite advances in treatment, 1  HIV disease continues to have multiple manifestations, many of which may compromise individuals' QOL, such as loss of appetite, nausea, vomiting, night sweats, exhaustion, weight loss, and depression (Cunningham et al., 1998). Franchi and Wenzel (1998) reviewed 20 published papers and concluded that HIV-related symptoms significantly and negatively affect QOL. As well, HAART is often accompanied by difficult side effects and toxicities as well as inconvenient and pill-intensive medication regimens that also degrade patients' QOL (Aversa, Kimberlin, & Segal, 1998). For HIV-infected individuals who are asymptomatic, it may be difficult to reconcile the aspirations of longevity and QOL, particularly when faced with committing to a lifelong treatment regimen that can provide longer survival, but may also make the quality of their lives worse. Over the past few decades, QOL has come to the fore as an important concept in the assessment and documentation of positive outcomes in health-care research and practice. Despite its prolific use, however, it remains a poorly defined concept, both conceptually and operationally (Gill & Feinstein, 1994; Wood-Dauphinee, 1999). Most authors view QOL as a multidimensional construct that is subjectively experienced, but agreement on the specific dimensions of QOL and how best to measure them has not yet been reached (Apolone & Mosconi, 1998; Ferrans, 1990a, 1992; Haas, 1999a, 1999b; Meeberg, 1993; Stewart, 1992; Wood-Dauphinee, 1999). QOL and its measurement have become increasingly popular with researchers involved with HIV-infected populations. Several HIV-specific QOL instruments have been developed since 1989 (de Boer, van Dam, & Sprangers, 1995), the most frequently employed being the Medical Outcomes Study - HIV Health Survey (MOS-HIV) (Wu et al., 1991). The MOS-HIV is considered a "generic" scale and therefore may not be sensitive to small differences or changes in QOL. Additionally, because of its "evolutionary" development (i.e., it evolved from a generic, general population QOL scale), the MOS-HIV may not cover all dimensions of QOL  that are salient to HIV-infected individuals. A relatively new instrument, the Multidimensional Quality of Life Questionnaire for Persons Living with HIV or with ADDS (MQOL-fflV) (Smith, Avis, Mayer, & Swislow, 1997), was developed by studying HIV-infected individuals themselves, some of whom were women, and therefore may contain the specific dimensions of QOL of relevance to HIV-infected populations. Although the HIV epidemic in developed countries was initially recognized predominantly among gay men, women now represent one of the largest growing groups of newly diagnosed cases (British Columbia Centre for Disease Control Society, 2000). Nonetheless, most literature describing the experiences of living with HIV continues to focus almost exclusively on men (van Servellen, Sarna, & Jablonski, 1998). Just as it is generally understood that QOL varies according to gender in general and other disease-specific populations (Brazier et al., 1992; Dibble, Padilla, Dodd, & Miaskowski, 1998; Hjermstad, Fayers, Bjordal, & Kaasa, 1998; Jenkinson, Coulter, & Wright, 1993; Kinney, Burfitt, Stullenbarger, Rees, & Debolt, 1996; Riedinger et al., 2001), recognition is growing that the experience of living with HIV and its impact on QOL may be very different for women than for men (Cederfjall, Langius-Eklof, Lidman, & Wredling, 2001; Sarna, van Servellen, Padilla, & Brecht, 1999; Sowell et al., 1997). It is therefore important that evidence of the validity and reliability of instruments used to assess QOL in HIV-infected women be established. The MQOL-HIV is a promising new instrument that, because of its inductive development, may tap QOL dimensions of relevance to HIV-infected populations. However, there is limited published evidence of its validity and reliability, and it has not been formally validated in HIV-infected women (personal communication, K. W. Smith, October 30, 2000). According to Kline (1979), any scale that is intended for use in a population for which it was not originally developed requires revalidation to allow proper interpretation of the results and for inferences to be made. Because HIV-infected women represent a unique population from those 3  groups traditionally used to develop HIV-specific QOL scales (i.e., predominantly gay men), revalidation of the MQOL-FflV is required to determine its relevance to and usefulness in this population. Statement of Purpose It was the purpose of this study to assess the validity and reliability of an HIV healthrelated QOL measure, the MQOL-HIV, in a cohort of HIV-infected, Canadian women. Specifically, this study assessed the content validity, concurrent validity, internal consistency, and test-retest reliability of the instrument. Research Questions The research questions that guided this study were: 1. What is the level of content validity of the MQOL-HIV in a cohort of HIV-infected women? 2. What is the level of concurrent validity of the MQOL-HIV when contrasted with the MOS-HIV in a cohort of HIV-infected women? 3. What is the internal consistency of the MQOL-HIV in a cohort of HIV-infected women? 4. What is the two-week test-retest reliability of the MQOL-HTV in a cohort of HIVinfected women? Significance of the Study As Kline (1979) states, "The use of any scale with groups for which it was not initially validated requires revalidation ... if one is to make the same inferences as one would with the original validation sample" (pp. 1699-1700). Therefore, the psychometric assessment of the MQOL-HIV in HIV-infected women is required before initiating its use within this population. Results from this study will allow researchers studying QOL in HIV-infected women to interpret study results using the MQOL-HIV scale. They may also provide some insight into the possible unique dimensionality of QOL in HIV-infected women. 4  Organization of the Thesis This thesis consists of six chapters. In Chapter One, the background and problem identification, statement of purpose, research questions, and significance of the research have been presented. In Chapter Two, a review of the selected literature relevant to the identified research problem is presented. In Chapter Three, the research methods are presented, including a description of the research design, study context, sampling methods, data collection and analysis methods, instrumentation, and protection of human rights considerations. In Chapter Four, the findings are presented, including a description of the sample and the findings for each of the research questions. In Chapter Five, a discussion of the findings is presented. The summary, limitations, conclusions, implications for nursing practice, research, and theory, and recommendations for future research are presented in Chapter Six.  5  C H A P T E R 2: R E V I E W O F T H E L I T E R A T U R E The following literature review focuses on conceptual and methodological topics of importance when assessing psychometric properties of a disease-specific Q O L instrument in HIV-infected women. The chapter is divided into three sections, including a review of the conceptualization of Q O L , a review of Q O L measurement, and a review and critique of the HTVspecific, Q O L questionnaire used in this study, the Multidimensional Quality of Life Questionnaire for Persons L i v i n g with H I V or with A I D S ( M Q O L - H I V ) . Before any discussion about measurement and instrumentation, a review of how the concept of Q O L is conceptualized in the literature is of foremost importance. A conceptual framework allows the researcher to put into perspective what, in fact, is being evaluated. What is quality oflifel  This is a question that receives much attention in the literature, with no consensus  reached. F o r this reason, a large portion of this review focuses on the conceptualization of Q O L . Because the population of interest in this study is women living with H I V , and because the female HIV-infected population in British Columbia is culturally diverse, what is currently known about Q O L of persons living with H I V and of women is also reviewed, including a brief discussion of cultural considerations. The second section of the literature review focuses on Q O L measurement. Although there are many topics that could be explored in relation to Q O L measurement, only those viewed as relevant to this project are discussed. The operationalization of Q O L is reviewed, as well as the types of scales that can be used to measure it. Because the current study proposes to assess a relatively new instrument, some understanding of the common methods used in instrument development, testing, and refinement is required to put this study in context. Specifically, methods used to assess an instrument's validity and reliability are reviewed.  6  The third section describes the study instrument, the MQOL-HIV, and provides a synthesis of the literature on its development, testing, and use. The chapter concludes by putting conceptualization and measurement issues into context as they relate to the study. Conceptualization of Quality of Life Historical Background Since ancient times, QOL has been described by such philosophers as Aristotle (trans. 1963), who referred to eudaimonia - the happiness of well-being, or the idea that having understanding is the best way to live one's life. Early Jews and Christians referred to QOL in terms of one's willingness to surrender to the will of God and to choose good over evil (Ecclesiastes 12:13-14, New International Version). In the past century, the term quality of life was first described by the economist Pigou in 1920 but did not surface again until a couple of decades later (Wood-Dauphinee, 1999). After World War U and the deprivations of the Great Depression in the 1930's, the concept of QOL came further to the forefront as society acknowledged social inequities and focused their attention on living a good life instead of simply surviving (Harrison, Juniper, & MitchellDiCenso, 1996; Wood-Dauphinee). As well, the broadening of the definition of health by the World Health Organization (WHO) in 1947 to include physical, emotional, and social wellbeing, raised the question of whether and how health should be measured (Wood-Dauphinee). The development of welfare states in Europe and North America stimulated literature on measures of social indicators in the 1970's (Wood-Dauphinee), a time in which the perspective of the patient was also increasing in importance in the realm of health-care and medicine (Harrison et al.). In surveys of Americans, performed by Cantril in the 1950's, and by Campbell and colleagues and Flanagan in the 1970's, social indicators, such as level of education, housing, and local crime rates, were said to reflect a nation's overall QOL (Allison, Locker, & Feine, 1997; Anderson & Burckhardt, 1999; Campbell, Converse, & Rodgers, 1976). However, these 7  objective indicators were soon recognized as comprising only a small proportion of an individual's QOL and subjective indicators were eventually given greater recognition (Haas, 1999a). Research in the area of QOL began in the sociology and psychology disciplines, moving from there into the health-care disciplines (Haas, 1999a). In 1977, "quality of life" became a key term of MEDLINE (Wood-Dauphinee, 1999). In the 1980's, "gap" theories of QOL emerged, stating that QOL is the difference between a person's expectations and present life circumstances (Haas); that is, the narrower the gap, the higher the quality of life (Caiman, 1987). The term health-related quality of life (HRQOL) came into use in the mid-1980's, suggesting that it is not goods such as jobs, housing, and schools that constitute health, rather those things that directly affect health (Wood-Dauphinee); this term continues to be used interchangeably with QOL. Over the past few decades, QOL has come to the fore as an important concept in the assessment of outcomes in health-care research and practice. Despite its increasing popularity and extensive use, there currently exists no consensus in the health disciplines on the definition of QOL, the dimensions or domains that comprise QOL, nor its essential attributes (Stewart, 1992; Wood-Dauphinee, 1999). The approach used in the remainder of this review is to identify how the concept of QOL is used, how QOL is defined, putative dimensions of QOL, and essential attributes of the concept. Uses of the Concept QOL is arguably one of the most versatile terms employed, being used in the media, politics, and advertising, and in research in the fields of sociology, medicine, nursing, psychology, economics, geography, social history, and philosophy (Farquhar, 1995). Understanding the context in which QOL is used is important in that it may influence the way that it is conceptualized. In the health science literature, some common ways in which QOL is used are as a measure of utility to evaluate resource allocation, a measure of value in directing 8  health policy, an outcome measure to assist in treatment decision making, an indicator of treatment efficacy, and as an indicator of quality of care. Descriptions of the types of definitions of QOL follow in the next section. As a measure to evaluate resource allocation, health economists refer to QOL when assessing the merits of alternative allocations of resources to enable consumers to maximize their quality of life; this is commonly referred to as utility (Grabowski & Hansen, 1996; Spiegelhalter et al., 1992). Utility is often described as a level of usefulness or ability to function normally and productively within a society (Spiegelhalter et al.). Closely tied to resource allocation evaluation, QOL is used as a value indicator that assists in directing health policy. Health policy makers deal with the question of whether the costs of medications or treatments are appropriate or too high; by comparing the total cost and benefit (i.e., total clinical, economic, and QOL value) of two or more treatments, the relative value of those treatments can be established (Spilker, 1996; Stewart, 1992). In the past two decades, QOL has been used as an outcome measure of efficacy in medical and pharmaceutical clinical trials, thereby assisting in treatment decision making (Spilker, 1996). Researchers acknowledge that although traditional biomedical measures may provide information about physiological responses to therapy, they do not reflect how patients feel and function as a result of treatment, nor how treatment affects individuals' perceptions of their benefit (Wiklund, Dimenas, & Wahl, 1990). Many claim that subjective measures of health-related QOL are valid and reliable and are responsive to important clinical changes that result from treatment (MacKenzie, Charlson, DiGioia, & Kelley, 1986; Wiklund et al.; Wilson & Cleary, 1995; Wood-Dauphinee, 1999). Because it is now the common view that the clinical decision-making process should be shared between health-care providers and patients, it follows that QOL assessment, as the representation of patients' aims, ends, and values, is indispensable for sound decision making (Brock, 1996). 9  Researchers and administrators employ QOL as a tool to evaluate quality of care (Wiklund et al., 1990; Wu & Cagney, 1996). The tradition of assessing health care with measures of morbidity and mortality based on laboratory, clinical, and radiological testing, and such measures as hospital-patient-days and readmission rates, is beginning to shift towards measures that incorporate patients' points of view (Wood-Dauphinee, 1999). QOL measures are used to evaluate the quality of care for particular conditions in specific populations and across groups; this information is subsequently used to identify variations in practice and areas where improvement is needed (Wu & Cagney). Finally, QOL is used as an outcome measure in intervention evaluation, a practice that is becoming more commonplace, particularly in nursing research (Harrison et al., 1996; Thorne & Jillings, 1996). Harrison et al. claim that the anticipated outcomes of nursing interventions often relate to the dimensions that make up QOL, and that the evaluation of such interventions necessitates the use of QOL measurement. Because QOL is used in a variety of contexts and for a variety of purposes, it follows that the concept means different things to different users. It is also understandable then, why there remains no consensus on a clear, conceptual definition of the term. The following section addresses the question, What is the meaning of QOL? Definitions of Quality of Life In the literature, it appears as though there are as many definitions of QOL as there are authors who use and write about it. One way of providing clarity when sifting through a large and diverse body of literature is to group definitions into categories. Although different authors categorize definitions of QOL in different ways, as I reviewed the literature on the definitions and conceptualizations of QOL, six "types" of definitions became apparent, those related to health, achievement, happiness, satisfaction, well-being, and utility. By far the most common  type, health-related definitions focus on those dimensions of QOL that pertain strictly to health 10  and functioning, excluding those dimensions that are claimed to be "non-health related." Definitions that focus on achievement address a person's fulfillment of personal goals or potential. Those definitions that relate to happiness and satisfaction hold that QOL is the level of happiness or satisfaction individuals have in important areas of their lives. Definitions that focus on well-being deal with a more global, overall perception spanning all areas of life. Finally, definitions based on notions of utility focus on quality of life as it pertains to a person's ability to function within society — the assessment of which often results in the computation of a "score," indicating the value of a life. Health-related definitions of quality of life. Health-related definitions of QOL are often referred to as describing health-related quality of life (HRQOL). These definitions are largely concerned with the physical and functional status of individuals in terms of deviations from what is considered "normal." Ferrans (1990b) terms these "normal life" definitions, in that they contain such notions as the ability to live a life of physical ability and functioning similar to what would be lived by "healthy" or "normal" individuals in similar situations or age categories. This is seen in Harrison et al.'s (1996) definition, which asserts that health-related QOL is an individual's perception of aspects of their life that they consider important in relation to their expectations of 'normal' living. Taking a different focus, Schipper, Clinch, and Olweny's (1996) definition, "the functional effect of an illness, and its consequent therapy, upon a patient, as perceived by the patient" (p. 16), highlights that some authors focus strictly on QOL as a concern about function. It is generally agreed that HRQOL is subjectively perceived and that it encompasses multiple, health-specific dimensions; often this entails the purposeful exclusion of "non-health" dimensions such as relationships, spirituality, housing, jobs, the neighbourhood, and socio-economic status (Apolone & Mosconi, 1998; Haas, 1999a; Wood-Dauphinee, 1999).  11  Among those who define HRQOL, many use the term interchangeably with QOL; one has only to perform a MEDLINE search for "QOL" to find that publications using QOL in the title either explicitly employ HRQOL in the body of the article or imply its use by measuring only those domains that relate directly to physical health and functioning. In an extensive review of the QOL literature, Gill and Feinstein (1994) found that none of the articles reviewed distinguished between QOL and HRQOL. In her concept analysis of QOL, Haas (1999b) identified two categories of HRQOL that are used interchangeably with QOL, namely functional status and health status. According to Apolone and Mosconi (1998), QOL and HRQOL are used synonymously to describe an evaluation of the effects that a disease, treatment, or more complex intervention has on a person's physical or emotional performance in everyday activities. The title of Spilker's (1996) second edition of Quality of Life and Pharmacoeconomics in Clinical Trials is incongruous with the opening statement "health-related quality of life is the subject of this book" (p. 1), although he acknowledges that "because of the intimate connection between social relationships and health ... and also between spirituality and health, all domains should be assessed when health-related quality of life is measured" (p. 1). Schipper et al. (1996), although clearly referring to HRQOL, state that '"quality of life' in clinical medicine represents the functional effect of an illness and its consequent therapy upon a patient, as perceived by the patient" (p. 16). Perhaps the crux of determining whether HRQOL is synonymous with QOL lies in one's definition of health. According to Wood-Dauphinee (1999), most health-related QOL definitions are based on the 1947 WHO definition that expanded the concept of health to include physical, emotional, and social well-being. More recent definitions encompass the notion that health is a state of both wellness and illness and hold that individuals with disease can perceive themselves as well. In defining HRQOL, Guyatt, Feeny, and Patrick (1993) state that health status, functional status, and quality of life are often used interchangeably to describe the same domain 12  of health. Because the notion that health is oppositional to disease is contrary to the view of many patients with advanced illness who perceive good health in the presence of disease (Connelly, Wofford, & Philbrick, 2000; Fryback, 1993; Kagawa-Singer, 1993), it would appear that viewing health, and subsequently F1RQOL as a synonymous measure of QOL, as something solely encompassing the physiological and psychological responses to disease and disability may result in the loss of the breadth, depth, and richness of human experience that constitutes true or full QOL. Achievement-related definitions of quality of life. Some definitions of QOL focus on the achievement of personal goals. Most authors conceptualize this as a relatively global concept, wherein individuals measure their current status against some desired standard. Falling into this type are the "gap theories" described in the 1980's, in which QOL is described as the difference between one's current life circumstances and the standard to which one compares oneself (Caiman, 1987; Haas, 1999b); thus the smaller the gap, the higher the QOL, and conversely, the less a person is able to realize his/her expectations and goals, the lower the QOL (Schipper et al., 1996). For example, Caiman describes QOL as: the difference, at a particular period of time, between the hopes and expectations of the individual and the individual's present experience. It is concerned with the difference between perceived goals and actual goals. It is an assessment of the potential for growth, (p. 46) As pointed out by Ferrans (1990b), the key feature of these definitions is that they focus on the individual's own goals, not those determined by others nor those that may necessarily apply to all people.  13  Happiness-related definitions of quality of life. Happiness is the theme of several authors' definitions of QOL. With the exception of McCall (1975), who refers to the happiness of a society when stating that QOL is "obtaining the necessary conditions for happiness throughout a society" (p. 299), most authors refer to an individual's own perception of happiness with their life. Aristotle (trans., 1963), who discusses different areas of one's life that provide happiness, and Oleson (1990), who considers QOL to be "the subjective perception of happiness or satisfaction with life in domains of importance to the individual" (p. 188), allude to different dimensions with which a person can be happy. Conversely, Dalkey and Rourke (cited in Farquhar, 1995), who simply refer to "a person's sense of ... his happiness or unhappiness" (p. 503), and McCall describe QOL as a more global experience. Although happiness is frequently included in QOL definitions, it should not be confused with QOL itself. Happiness is an emotion described by Campbell et al. (1976) as a state that is the product of the presence of positive feelings and the absence of negative feelings. They also point out that happiness is a short term, fleeting emotion, and therefore may not be the best indicator of QOL. Bradburn (cited in Oleson, 1990) also defines happiness as a short term emotion, referring to it as transitory and involving moods of gaiety or euphoria. Satisfaction-related definitions of quality of life. Often coupled with happiness is the concept of satisfaction, at times appearing to be used as an interchangeable concept (Dalkey & Rourke, in Farquhar, 1995; Oleson, 1990). Campbell et al. (1976) define satisfaction as a cognitive and judgemental experience that is maintained over time as a result of an evaluation of one's life situation. Most agree that satisfaction is subjectively determined by the individual, although Meeberg (1993) adds that others must agree that the individual's living conditions are satisfactory when she states that QOL is "a feeling of overall life satisfaction, as determined by the mentally alert individual whose life is being 14  evaluated. Other people ... must also agree that the individual's living conditions are not lifethreatening and are adequate in meeting that individual's basic needs" (p. 37) . 1  Implied in many definitions is the notion that individuals may identify unique areas or dimensions of their lives that require satisfaction; an example of this is Oleson (1990), who states that QOL is "the subjective perception of happiness or satisfaction with life in domains of importance to the individual" (p. 189). Another variation on the satisfaction theme is introduced by Campbell et al. (1976), who identify the satisfaction of needs as a component of QOL when they describe "an experience defined mainly in terms of satisfaction of needs [recognizing that] individual needs differ greatly from one person to another" (p. 9). Sartorius (1987) however, suggests that satisfaction may at times be at odds with QOL, in that satisfaction is defined as "the achievement of a goal or the sense of approaching it," while QOL is a "derivation of the levels of achievement of the various goals a person may have" (p. 19). Haas (1999b) also points out that satisfaction is not synonymous with QOL in that it fails to consider objective indicators of QOL. Well-being-related definitions of quality of life. Well-being is often used synonymously with satisfaction and QOL. In an analysis of well-being, Acton (in Stewart, 1992) noted that the term was often used interchangeably with QOL, although suggested that the two should be conceptually clarified to identify differences between the two concepts. The definitions proposed by Dalkey and Rourke (in Farquhar, 1995) and Naughton, Shumaker, Anderson, and Czajkowski (1996) view QOL as synonymous with well-being or satisfaction, whereas Ferrans (1990b) identifies well-being as a consequence of satisfaction within areas of one's life when she states that QOL is "a person's sense of well-being that stems from satisfaction or dissatisfaction with the areas of life that are important to him or her" (p. 15). On the other hand, Haas' (1999b) model of QOL identifies well-being as an indicator  1  of QOL. Well-being is generally viewed as being subjective, although as Haas points  Meeberg's (1998) definition implies that others have, in effect, "veto" power over an individual's conception of their QOL.  15  out, this attribute distinguishes it from QOL in that she claims that one of the critical attributes of QOL is that there can be an objective component when she states that QOL is: a multidimensional evaluation of an individual's current life circumstances in the context of the culture in which they live and the values they hold. QOL is primarily a subjective sense of well-being encompassing physical, psychological, social, and spiritual dimensions. In some circumstances, objective indicators may supplement or, in the case of individuals unable to subjectively perceive, serve as a proxy assessment of QOL. (p. 219) Utility-related definitions of quality of life. Utility-type definitions of QOL are largely employed within the domain of health economics and policy making. Utility most often, in a global sense, refers to the ability to lead a useful life. Ferrans (1992) refers to this as social utility, or the ability to lead a socially "useful" life. Stated explicitly by Burt et al. (cited in Caiman, 1987), who refer to the extent to which an individual feels he or she has the power to determine his individual well-being within society, is the notion that society contains norms against which an individual's QOL can be compared, and that varying degrees can be determined by society itself. In some cases, utility-related QOL is conceptualized as something that can be assigned a numerical value (i.e., identifying the value of a life), such as by Shaw (cited in Meeberg, 1993), who defines QOL as the following formula: "QOL = NE x (H + S) ... where QOL equals natural endowment (NE) times combination of home (H) and society (S)" (p. 34). Dimensions of Quality of Life Although most experts agree that QOL is a multidimensional concept (Schipper, Clinch, & Powell, 1990; Wood-Dauphinee, 1999), delineation of the distinct dimensions - or domains of QOL varies. Early researchers who performed large-scale surveys of the American population identified up to 17 domains of life satisfaction that constituted QOL (Campbell et al., 1976). 16  More recent conceptualizations incorporate fewer, broader dimensions that embrace a large variety of aspects of human life. While some authors conceptualize a single dimension of QOL (i.e., QOL as a global concept), others identify groups of broad domains, each containing specific aspects. Adding to the confusion is the fact that no two authors identify the broad domains nor the specific aspects in the same way. Furthermore, different authors use different labels for dimensions and aspects of QOL and do not provide definitions of such, making identification of consensual dimensions nearly impossible. As an example, King (1998) describes five generally agreed-upon dimensions: (1) physical, (2) psychological, (3) social, (4) somatic/disease- and treatment-related symptoms, and (5) spiritual. According to King, the physical dimension is that which most traditional measures resemble, namely measures of strength, energy, and physical function. The most commonly assessed psychological symptoms include anxiety, depression, and fear. Social well-being refers to how individuals function within relationships and within the context of work and community. The somatic dimension includes disease symptoms and side effects of therapy. And finally, the spiritual dimension refers to the perception that life has meaning and may encompass religious beliefs (Schipper et al., 1990). Appendix A further highlights the lack of consensus and the wide range of concepts used to label dimensions of QOL. It should be noted that dimensions conceptualized by different authors often do not share the same label nor definition. For example, some view the physical dimension as an indicator of function, others as physical health (though health may or may not be defined), and still others as the consequence of disease and treatment. The psychological domain is viewed by some as encompassing the cognitive, mental and emotional aspects of QOL, yet by others as a dimension distinct from these concepts. Also common, the social domain may encompass anything from social functioning to the concept of social support.  17  To further muddy the conceptual waters, there is lack of agreement regarding what constitutes a dimension versus an element of a dimension. Haas' (1999b) model of QOL includes the physical, psychological, social, and spiritual dimensions, and includes subjective well-being and obj ecti ye functional status as indicators of QOL. She considers the dimensions as general, although does not provide information regarding just what types of elements fall within these dimensions of QOL. Ferrans' (1996) model of QOL includes four general domains, health and functioning, psychological/spiritual, social and economic, and family, and further  identifies specific elements that comprise each dimension (see Table 1). The Medical Outcomes Study (MOS) conceptual framework conceptualizes all health measures as indicators of two underlying dimensions of health, namely physical and mental (Wu et al., 1991). Indicators fall within four categories: physical, mental, social functioning and well-being, and health perceptions (see Table 2). Additional issues related to the various conceptual frameworks of QOL are raised by Stewart (1992). She adds what she calls the "response dimension," meaning that part of the definition that pertains to what it is about the dimensions and elements that is of interest. For example, the response could be subjective satisfaction with the particular dimension (e.g., satisfaction with physical functioning or intimacy), or objective limitations in the dimension (e.g., limitations due to diminished cognitive or sexual functioning). As well, Stewart raises the question as to whether some of the dimensions and elements are components or determinants of QOL. For example, does one's economic status in part constitute QOL, or is it a determinant of psychological well-being or security? Similarly, diminished energy or vitality are common symptoms of disease or side effects of treatment, yet are often included in conceptualizations of QOL. According to Stewart, this leads to confusion in developing hypotheses as to what is an outcome and what is an intervening treatment or disease variable.  18  Table 1. Elements of the Ferrans' (1996) Conceptual Model of Quality of Life Health and Functioning Domain Usefulness to others Physical independence Ability to meet family responsibilities Own health Pain Energy (fatigue) Stress or worries Control over own life Leisure time activities Potential for a happy old age/retirement Ability to travel on vacations Potential for a long life Sex life Health care  Social and Economic Domain Standard of living Financial independence Home (house, apartment) Neighbourhood Job/Unemployment Friends Emotional support from others Education Family Domain Family happiness Children Relationship with spouse Family health  Psychological/Spiritual Domain Satisfaction with life Happiness in general Satisfaction with self Achievement of personal goals Peace of mind Personal appearance Faith in God  The above discussion and examples are only a small sampling of the diversity found in the literature regarding the constituents of QOL. They demonstrate that there is no agreement on the dimensions of QOL nor the elements found within. Although most conceptual frameworks include the very broad physical, psychological, and social dimensions (see Appendix A), all contain additional, varying combinations of elements purported to comprise QOL.  19  Table 2. Medical Outcomes Study Conceptual Framework (adapted from Stewart, 1992) Physical Health Physical Functioning or Well-being Physical functioning Pain Energy/fatigue Sleep problems Psychophysiologic symptoms  X X X X X  Mental Functioning and Well-being Cognitive functioning Psychological distress Psychological well-being Social Functioning and Well-being Social activity limitations due to health Role limitations due to physical health Role limitations due to emotional problems Health Perceptions Current health perceptions Health outlook  Mental Health  X X X X X X  X X  X X  X X  X X  Essential Attributes of Quality of Life Because there is no consensus on the conceptual definition of QOL, it is not surprising, then, that the critical attributes of the concept are also unclear. Nonetheless, some commonalities exist and therefore four common attributes are tentatively proposed: (a) it is multidimensional in nature, (b) it is subjectively experienced, (c) it can be measured, and (d) it is somehow related to the determination of satisfaction or well-being within the dimensions. Although a review of published definitions shows a split consensus regarding the comprehensiveness or multidimensionality of QOL, most published reviews and concept analyses of the conceptualization and definition of QOL claim that it is commonly referred to as a multidimensional concept (Apolone & Mosconi, 1998; Fava, 1990; Ferrans, 1990a, 1992; Haas, 1999a, 1999b; Meeberg, 1993; Stewart, 1992; Wood-Dauphinee, 1999). The exception is 20  definitions found within the utility category, which are unanimously global in scope. However, it should be noted that in some instances, though the definition itself may not refer to multidimensionality, other elements within the definition may imply it, depending on their use. For example, Parse (1994) defines QOL broadly as "not what those outside the life looking in think it is, but rather it is what the person there living the life says it is" (p. 17). It is implied in Parse's definition that the way in which the person living the life conceptualizes his or her QOL may indeed be multidimensional. With the exception of the utility definitions, there is general consensus that in most cases QOL is subjectively experienced and assessed by the individual. Nonetheless, this is not unanimously agreed upon, as pointed out by authors such as Haas (1999a, 1999b) and Meeberg (1993), who claim that others must either agree with the individual, or may act as proxies in the evaluation of another's QOL. It is implied by the terms used within definitions, such as perception, sense, feeling, and evaluation, that QOL can be expressed and thus evaluated and measured. Meeberg (1993) adds to this in her identification of one of the critical attributes as having the mental capacity to measure one's own life as satisfactory or otherwise. Haas (1999a) too identifies the assessment of QOL as being a critical attribute, agreeing with Allison et al. (1997) that the assessment is based on values and that it is dynamic and changing depending on the context in which it is measured. Implied in the assessment is that some form of standardized 'norms' exist to which a person aspires, be they established by the individual or by society. Finally, the assessment of QOL is related in some way to the determination of one's level of satisfaction or well-being with specific dimensions of one's life. Stewart (1992) refers to this as the 'response dimension,' in that how a person responds to dimensions of their life determines their assessment of their QOL. Oleson (1990) and Meeberg (1993) identify satisfaction with life  21  as a critical attribute of QOL, whereas Ferraris (1996) sees well-being as a result of satisfaction in areas of life that are viewed as important to the individual. Quality of Life of Particular Populations Although there is a wealth of literature on the QOL of specific populations, this section focuses on what is known about QOL in HIV-infected populations, gender differences in QOL, and cultural considerations. Quality of life in HIV-infected populations. The HIV-related QOL literature focuses mainly on the identification of dimensions of QOL unique to HIV, HIV-specific instrument development and testing, descriptive research (i.e., describing the QOL of HIV-infected cohorts), and the effect of treatments on QOL (i.e., QOL as an outcome measure in clinical drug trials). The dimensions of QOL presented in most publications are aimed at general populations, and may not provide a good "fit" in the HIV context, de Boer et al. (1995) suggest that the core set of general QOL dimensions in general populations may require supplementation with specific domains of concern for HIV-infected patients. For example, they assert that sexual functioning is seen as a relevant additional domain in HIV-infected patients because of its importance in the lives of many HIV-infected persons, especially gay men, and that many individuals contract the 2  disease through sexual contact and have to cope with the knowledge of being a carrier of a potentially fatal, sexually transmittable disease. Body image may also be a dimension or element of importance for HIV-infected individuals, especially for those with Kaposi's sarcoma (now rare), or with severe bodyweight loss and medication-induced lipodystrophy (de Boer et al.). Some of the additional dimensions of QOL that have been identified as relevant to HIV include disclosure worries and health distress (Holmes & Shea, 1998; Wu et al., 1991), financial worries  The assertion that sexual functioning is seen as particularly relevant to gay men appears to assume that it is not relevant to other populations, such as heterosexual men and gay or heterosexual women. I believe this to be a rather stereotypical view, and argue that sexual functioning - or more accurately, sexuality - is likely a relevant domain of QOL for all human populations.  2  22  (Holmes & Shea; Smith et al., 1997), sexual functioning or intimacy (de Boer et al.; Leplege et al, 1997; O'Brien & Pheifer, 1993; Smith et al.), transcendence or existential well-being (Cohen, Hassan, Lapointe, & Mount, 1996; de Boer et al.), and self concept and body image (de Boer et al.; O'Brien & Pheifer). Because the number of HIV-infected individuals who come from marginalized communities (i.e., including drug use and poverty) is growing, issues such as financial status, depression, and housing may be of increasing relevance (Coscia et al., 2001; McClure, Catz, Prejean, Brantley, & Jones, 1996). Because it is generally accepted that the QOL dimensions of importance to HIV-infected populations may differ from those of the general or other disease-specific groups, a great deal of literature is dedicated to reporting the development and refinement of HIV-specific QOL instruments. Several disease-specific instruments are now available, as shown in Table 3. It should be noted that most of these instruments were developed before the era of HAART and the chronification of HIV disease; therefore, much of the current literature focuses on assessing the validity and reliability of these instruments. Results of descriptive studies of the QOL of HIV-infected populations vary depending on the population studied, the instrument(s) used, and the point in the spectrum of infection of the participants. Most authors report very low QOL scores in various dimensions in patients with advanced HIV disease and AIDS (Lenderking, Testa, Katzenstein, & Hammer, 1997; Wachtel et al., 1992). As Watchel et al. observe, this is not surprising given the debilitating effects of AIDS, such as weight loss, repeated and protracted hospitalizations, and the fact that persons with HIV disease often contend with discrimination and prejudice in the areas of employment, housing, insurance, and medical care with obvious effects on role and social functions. These findings confirm those of Wu et al. (1991), who found that the impact of HIV-related symptoms was devastating in all measured dimensions of QOL. Similarly, in a review of longitudinal assessments of multi-dimensional health-related QOL in HIV-infected patients, de Boer et al. 23  Table 3.fflV-SpecificQOL Instruments Instrument  Source  Medical Outcomes Study - HIV Health Survey (MOS-HIV)  Wu etal., 1991  Multidimensional Quality of Life Questionnaire for Persons Infected with HIV or with AIDS (MQOL-HIV)  Smith et al., 1997  Functional Assessment of Human Immunodeficiency Virus Infection (FAHI) quality of life instrument  Cella, McCain, Peterman, Mo, & Wolen, 1996  HIV Overview of Problems-Evaluation System (HOPES)  Ganz, Coscarelli Schag, Kahn, Petersen, & Hirji, 1993  AIDS Health Assessment Questionnaire (AIDS-HAQ)  Lubeck & Fries, 1997  HIV/AIDS-Targeted Quality of Life (HAT-QoL)  Holmes & Shea, 1997  HIV-PARSE  Bozzette, Hays, Berry, & Kanouse, 1994  HIV Quality of Life - 31 (HIV-QL31)  Leplege et al., 1997  Living with HIV  Holzemer, Spicer, Skodol, Kemppainen, & Coleman, 1998  HIV-QoL  Cleary et al., 1993  (1995) found that over time HIV-infected patients reported poorer physical, role, and sexual functioning, although their psychological functioning improved or remained stable. Other studies of relatively healthy samples of HIV-infected individuals found higher levels of QOL in asymptomatic or early stage HIV disease (Lubeck & Fries, 1997; Ragsdale & Morrow, 1990; Smith et al., 1997). Although researchers used a variety of instruments and sample at different points in the spectrum of disease, the general consensus appears to be that QOL decreases as the severity of HIV illness increases.  24  Leplege and Berzon (1997) point out that since much of the treatment-related research in HIV is focused on the development of antiretroviral medications, the use of QOL scales in treatment trials is now routine and these instruments have become essential components of the clinical testing process for AIDS medications. Although the new HAART medications provide health benefits, many of the regimens have negative effects on patients' QOL because of toxicities, side effects, and the strict and pill-intensive daily, life-long routine (Nieuwkerk et al., 2000; Tsasis, 2000). It is becoming more generally recognized that QOL, and the trade-off between QOL and survival, are important considerations in the selection of medical therapies in HIV disease (Cohen, Revicki, Nabulsi, Sarocco, & Jiang, 1998). Gender differences in QOL. Although Hjermstad et al. (1998) claim that it is generally accepted that QOL varies by gender, there are few studies that describe the QOL of women specifically or that identify how and why these variations exist. Most QOL studies analyze men's and women's data together and do not consider gender as a separate factor or covariate. In the literature that does exist however, it appears as though not only do both healthy and diseased women perceive their QOL differently than men do, but they rate their QOL differently in different dimensions. As well, there is preliminary evidence to suggest that the QOL dimensions of salience may differ between women and men. Gender differences in QOL exist in both healthy and diseased populations. In a large, randomized study of the general Norwegian population, Hjermstad et al. (1998) found that overall, women scored lower than men in all QOL scales in all age categories except for cognitive function and social function. In these latter two dimensions, women over the age of 60 years scored higher than men of similar age. Similar findings have been published in populationreferenced data for the SF36, with women generally reporting lower QOL than men (Brazier et al., 1992; Jenkinson et al., 1993). These findings hold when disease-specific populations are 25  assessed. Several reviews describe gender differences regarding aspects of QOL in cardiac patients (Kinney et al., 1996), patients with genital herpes (Jadack, Hyde, & Keller, 1995), and patients with mental disorders (Linzer et al., 1996); in general, women report lower QOL than men in most studies. Regarding gender differences in specific dimensions of QOL, study findings vary. In a study of age and ejection fraction-matched men and women with heart failure, Riedinger et al. (2001) found that women obtained significantly poorer general life satisfaction, physical function, and social and general health scores than men. After controlling for degree of heart failure, women still had significantly poorer ratings for intermediate activities of daily living and social activity. In a study of 1,026 sarcoidosis patients, de Vries, van Heck, and Drent (1999) reported that women and men differed significantly in the broader domains of physical health and psychological health. Findings of gender differences in overall QOL and in specific dimensions of QOL are also found in HIV-infected populations. Comparing their findings with other, male-specific studies using the same instrument (HOPES), Sarna et al. (1999) found that HIV-infected women indicated more severe problems than men in every summary QOL scale. Wachtel et al. (1992), using the MOS-HIV, found that women with HIV infection reported lower overall QOL than men, as did Piette, Wachtel, Mor, and Mayer (1995). Similar findings were reported by McDonnell, Gielen, Wu, O'Campo and Faden (2000), in that HIV-infected women reported lower levels of well-being and functioning than in other, published male samples. Cederfjall et al. (2001) looked specifically at gender differences in QOL among 55 women and 134 men receiving outpatient HIV care; HIV-infected women were significantly younger than the men, had less advanced disease, and scored significantly lower on well-being, sense of coherence, and social support scales than the men.  26  There is limited, but compelling evidence that the actual QOL dimensions of relevance may vary according to gender. Dibble et al. (1998), in a study aimed at exploring gender differences and similarities in the dimensions of QOL, identified two main factors for women, psychosocial well-being and physical competence, and two different factors for men, vitality and  personal resources. They concluded that measurement of QOL requires gender-specific questions to accurately assess the dimensions of relevance to women and men. Although not studying dimensions specifically, Sowell et al. (1997) reported that they had to modify some of the questions used from a selection of generic and HIV-specific instruments to make them "female-appropriate." Although no detail was provided regarding why or how these modifications were made, the researchers clearly believed that the available instruments did not appropriately meet the needs of women. Perhaps the crux of why gender differences exist in reported QOL and the salient dimensions of QOL lies in the unique factors and stressors that exist for women. In HIVinfected populations both gender-specific stressors and demographical characteristics that influence QOL tend to vary between women and men. It has been suggested that factors influencing the QOL of HIV-infected women differ from those affecting men (Nussbaum & Sen, 1993). The interrelationships of poverty, caregiving roles, and social expectations may construct a unique context in which HIV-infected women live their lives. According to some authors, many of the stressors of HIV infection can be viewed as unique to women, such as the non-HTVrelated stressors of childbearing and rearing (van Servellen et al., 1998), and symptomatic gynecological conditions (Breibart et al. cited in van Servellen et al.; Melnick, Wertheimer, & Pinn, 1995). Semple et al. (1993), in a study of 31 women, identified stressors especially important to HIV-infected women, including gynecological problems, child and family role worries, marital/partner issues, occupational and economic issues, and social network problems. Although they recognize that some of these domains may also relate to men, some domains, such 27  as childb'family, economics, gynecological problems, and social network issues may be particularly important to women. Semple et al. reported that the women in their sample were more likely than men to have dependent children, be single parents, and have more difficulty disclosing HIV status in the context of heterosexual dating and to their children. Gynecological 3  problems were also seen as particularly important because they are robust prognostic indicators of rapid disease progression. Financial concerns, while prevalent in HIV-infected populations, may be worse for single mothers (Semple et al). Cultural considerations. In addition to differences in QOL according to gender, cultural diversity must be given consideration, particularly in the context of the culturally diverse HIV-infected female population in British Columbia. Despite there being no universally accepted definition of culture (Johnson, 1996), some researchers have attempted to define QOL in a culturally-sensitive way. In a book on cross-cultural QOL assessment, Shumaker and Berzon's (1995) definition of QOL purposefully includes only those domains believed to be universally viewed as essential to achieving desired goals, assuming, then, that regardless of culture, people will consider the domains directly relevant to their HRQOL. They define HRQOL as follows: Health-related quality of life refers to people's subjective evaluations of the influences of their current health status, health care, and health promoting activities on their ability to achieve and maintain a level of overall functioning that allows them to pursue valued life goals and that is reflected in their general well-being. The domains of functioning that are critical to HRQOL include: social, physical and cognitive functioning; mobility and self-care; and emotional well-being (p. 7).  It should not be inferred from the findings of Semple et al. that all HIV-infected women are heterosexual or that disclosure of HIV infection - or any other sexually transmitted infection - is only a concern in the context of heterosexual dating. Indeed, this is likely a concern in all human populations, regardless of sex or sexual orientation. 3  28  Guarnaccia (1996) claims that QOL itself is inherently a cultural concept, and that there exist three key problems to adapting instruments to culturally diverse populations: (a) researchers have an underdeveloped notion of culture and its impact on QOL, (b) in focusing on a particular ethnic population there is less attention given to intra- and intercultural diversity among study populations, and (c) the adaptation and translation of instruments are often inadequate. An awareness that cultural context can affect perceptions of QOL is essential. There is general agreement that in the context of QOL measurement there are systematic variations across cultures in terms of reporting style, illness experience, expectations of treatment, and acceptability of treatment (Cella, Wiklund, Shumaker, & Aaronson, 1995). In addition, withincountry differences with great cultural diversity (e.g., Canada) may be more dramatic than in culturally similar countries (Cella et al.). Furthermore, specific diseases can carry different impact across cultures, such as HIV infection (Berzon, Leplege, Lohr, Lenderking, & Wu, 1997). Measuring Quality of Life Although there are a myriad of topics that could be covered related to the measurement of QOL - indeed, many books have been written on the subject - for the purposes of this work, the focus is on several topics of particular importance to the study. These topics include how QOL is operationalized and measured, types of QOL scales currently available, and methods by which these scales are developed and tested. How Quality of Life is Operationalized and Measured To operationalize a concept for measurement, a clearly defined conceptual or theoretical framework is required. According to Shumaker, Ellis, and Naughton (1997), operationalization should respect the boundaries of the conceptual definition and capture the elements of it that are sensitive to change over time. However, as some authors point out, in the case of QOL, the usual scientific method of basing measurements on clearly defined concepts is often reversed, in that the concept is defined by the operations and instruments used to measure it (Albrecht et al., 29  1994; Guyatt, Jaeschke, Feeny, & Patrick, 1996; Wood-Dauphinee, 1999). Gill andFeinstein's oft-cited 1994 review found that only 15% of authors conceptually defined QOL, and more recent authors concur that this lack of conceptual definition still exists (Haberman & Bush, 1998; Wood-Dauphinee, 1999). The way in which QOL is operationalized depends on how it is defined and conceptualized. For example, those who define QOL as unidimensional (or global) operationalize QOL with a single item, such as a simple five-point rating scale or a 10 millimetre visual analogue scale (VAS) (Haberman & Bush, 1998). However, most view QOL as a multidimensional concept, and therefore operationalize QOL by identifying specific dimensions (i.e., those discussed previously) and specific items within each dimension. Because QOL is considered to be subjectively experienced, the most common way to measure it is through a questionnaire, completed either by the participant him or herself, or through an interview format, either in person or by telephone. Although self-administered questionnaires require minimal research resources, they may result in high rates of non-response or missing answers. In some cases, proxy or surrogate respondents are used if the participant cannot respond for him or herself, such as in the case of very young or ill respondents (Guyatt et al., 1996), but this issue is heavily debated in that one cannot be sure that a proxy's perceptions are the same as those of the patient (Guyatt et al.; Harrison et al., 1996; Wood-Dauphinee, 1999). Mode of administration can also affect how QOL is reported. For example, some claim that symptoms tend to be reported as improved in face-to-face interviews, perhaps because the patient wishes to please the interviewer (Wiklund et al., 1990). Results may differ depending on where the interview takes place (i.e., home vs. hospital), whether they are spontaneous reports vs. responses to active questioning, or derived from self-assessment questionnaires (Wiklund et al.). Still other studies have shown that interviews produce less missing data and are preferred by patients over self-administration (Cook et al., 1993; Weinberger, Oddone, Samsa, & 30  Landsman, 1996). Ultimately, the decision regarding choice of respondent and mode of administration must take into consideration the characteristics of the participant population and the resources available to the researcher. Types of Scales In general, there are two broad categories of scales, generic and disease-specific, and within each category, three types of scales used to measure QOL: unidimensional (or global) measures; single-scale, multidimensional measures; and the use of multiple, dimension-specific measures. Generic instruments measure QOL across a wide range of illnesses or conditions, and therefore provide a common database for comparing results. Some common examples of generic instruments include the Medical Outcome Study Short Form General Health Survey (SF36) (Ware, Brook, Davies, & Lohr, 1981), the Nottingham Health Profile (NPH) (Hunt et al., 1980), and the Sickness Impact Profile (SIP) (Bergner, Bobbitt, Carter, & Gilson, 1981). Many generic instruments provide investigators with an aggregate score and individual subscale scores, enabling not only the assessment of overall QOL but also providing a view of the various dimensions that make up QOL. Generic instruments may measure many dimensions of QOL and can provide a universal formula to evaluate impact of disease or treatment (Wiklund et al., 1990). The primary advantage of generic measures is that the effects of treatment can be compared in different studies for different medical conditions; however, they may be less sensitive than specific instruments and may not detect small but important changes in selected populations (Harrison et al., 1996; Wiklund et al.). Specific measures are designed to focus on the QOL of specific populations. Because they address QOL specific to particular conditions or treatments they may be more responsive to change (Wiklund et al., 1990). Instruments may be specific to a disease (such as asthma or HJV infection), to a population (such as the elderly or children), to a certain function (such as sleep or sexual function), or to a problem (such as pain) (Guyatt et al., 1996). Examples of specific 31  instruments are the Multidimensional Quality of Life - HIV questionnaire (MQOL-HTV) (Smith et al, 1997) the Child Health Questionnaire (CHQ) (Landgraf et al., 1998), and the Cancer Rehabilitation Evaluation System (CARES) (Schag, Ganz, & Heinrich, 1991). The main strength of specific measures is that they focus on the areas of greatest importance to patients and clinicians; however, the results cannot be compared across conditions or populations. Regarding types of scales, global scales generally consist of a single item that requires a rating of the individual's overall QOL. An example of this is a 100-mm VAS that may ask how an individual rates their overall QOL, where 0 = "worst imaginable QOL" and 100 = "best imaginable QOL" (Haberman & Bush, 1998). According to Haas (1999b), global measures are frequently an indicator of a person's satisfaction with life or their well-being; however, as pointed out earlier, well-being and satisfaction are not the same as QOL and therefore QOL should be distinguished from these concepts. Advantages of global measures are that they provide a quick, straight-forward method of assessing QOL and lend themselves to straightforward analysis (Stewart, 1992). However, they lack sensitivity and precision and may be unable to distinguish the effectiveness of different interventions (Holzemer & Wilson, 1995; Stewart). Farquhar (1995) notes too that global definitions are all-encompassing but tell us little about how QOL is to be operationalized. The second approach is to define QOL as multidimensional and operationally define it by including items of all its dimensions in one instrument (Farquhar, 1995; Ferrans, 1990a; Haas, 1999b). As Haas points out, problems with multidimensional scales include overlap of domains, variable emphasis on different domains, and lack of comprehensiveness. Additionally, she claims that scales can become problematic when domain scores are summed for an overall QOL score, in that different people may rate different domains in different ways, yet end up with similar overall scores, making comparisons difficult.  32  Finally, the multiple scale, or modular or battery approach (Stewart, 1992) consists of using a series of multiple instruments for measuring specific concepts, such as depression, functional status, self esteem, and so forth (Bowling, 1995). Problems with this approach include questionnaire overload, overlap, redundancy, the potential need for weighting of scores, and omission of certain dimensions of QOL (Haas, 1999b). As Bowling contends, studies that have used scales to measure a particular dimension, or sets of dimensions, of QOL have not conducted a "QOL study," per se, but rather a study of specific dimensions of QOL. How Scales are Developed and Tested The following section describes how scales are developed and tested in terms of their psychometric properties. Development. Scales are generally developed in two broad stages, development and refinement. The development phase involves either taking an existing, "standardized" instrument and modifying it to suit the population under study or using inductive methods to develop a scale from the ground up. One method of scale development involves using an existing scale as a base and modifying or evolving it to suit the needs of a specific population or study, Streiner and Norman (1995) state that "instruments rarely spring full grown from the brows of their developers ... they are usually based on what other people have deemed to be relevant, important, or discriminating" (p. 15). They claim that modifying existing scales saves time and effort, that previously developed items have usually gone through repeated processes of testing and therefore may be psychometrically sound and that there are only a limited number of ways to ask about a specific problem. It should be noted, however, that Streiner and Norman write about concepts used primarily in education and psychology, such as intelligence and personality, which perhaps have a longer history of theoretical development than QOL. It has been highlighted in the preceding 33  section that both the conceptualization and operationalization of QOL remain in their infancy, and no consensus has yet been reached on what QOL is, let alone how it is measured. Many authors have used the evolutionary method to develop QOL scales, modifying either generic or other disease-specific scales. Ferrans and Powers (1985; 1992) developed the generic Quality of Life Index and went on to modify it for specific disease states, such as cancer (Ferrans, 1990b). Regarding HIV-specific instruments, the HIV Overview of ProblemsEvaluations System (HOPES) (Schag, Ganz, Kahn, & Petersen, 1992) is a behaviour-based health-related QOL tool and a direct descendant of the Cancer Rehabilitation Evaluations System (CARES) (Schag et al., 1991). Perhaps the most commonly used HIV-specific scale, the Medical Outcomes Study HIV Health Survey (MOS-HIV) (Wu et al., 1991), is a questionnaire that was adapted from the Medical Outcomes Study Short Form 20 (Stewart, Hays, & Ware, 1988) by adding specific concepts relevant to HIV infection. Critics of the evolutionary method of scale development claim that it may be inadequate in identifying dimensions that are relevant to specific populations. Holmes and Shea (1998) highlight this when they state that: the processes by which modified instruments have been evolved for HIV/ ... (AIDS)-targeted use appear to have been driven mostly by investigators' beliefs about quality of life in HIV infection, and not to have included much or any input from HIV seropositive individuals. Thus, this method of evolution may have led to HIV/AIDS-targeted instruments that include dimensions of quality of life that are not relevant or important to the seropositive population, as well as not including dimensions that are. (p. 139) Inductive methods of scale development are considered by many to be most appropriate in the development phase of scales. One of the inductive methods used for scale development is identification of domains and theories through literature review. Streiner and Norman (1995) 34  recommend that researchers first identify whether an appropriate scale exists through extensive review of the literature, and that only in the absence of such a tool should a new one be developed. They also state that identifying dimensions and possibly items of a concept through thorough literature review should be seen as the first step in devising items in a scale. Lynn (1986) too identifies literature review as one of the steps in the developmental stage of scale construction. She claims that a thorough review of the literature to identify all dimensions and sub-dimensions of a concept will incorporate the ideas of a variety of experts, although noting that this is limited to those published. Another inductive method is that of incorporating the input of experts, both professional and members of the population of interest. Streiner and Norman (1995) state that "a point often overlooked in scale development is the fact that patients and potential research subjects are an excellent source of items" (p. 16). They assert that whereas clinicians may be the best observers of outward manifestations of a condition, only those who are experiencing it can report on its subjective elements. Streiner and Norman refer to two specific techniques, focus groups and key informant interviews. Focus groups involve facilitated small group discussions where members talk freely and spontaneously about themes considered important to the investigation. The tasks involve the generation of general themes and specific items within the themes. Key informant interviews involve in-depth interviews with a small number of individuals selected because of their unique knowledge, either because they possess the characteristics or disease of interest, or because of their professional expertise. Similar to focus groups, the objective of the interviews is to elicit information regarding the dimensions and items of the concept of interest. Several authors have employed inductive methods in developing QOL scales. In the first phase of their development of the HIV/AIDS-Targeted Quality of Life (HAT-QoL) instrument, Holmes and Shea (1997; 1998) performed in-depth interviews with 42 HIV-infected individuals, and using the nominal group technique (NGT), generated content for the questionnaire items. 35  This resulted in a 76-item draft that was subsequently refined in a second phase of development. Smith at al. (1997) also incorporated patient input from semi-structured interviews with 27 HIVinfected patients and six expert health-care providers. From these interviews, 12 major domains of the MQOL-HIV were identified, and the authors wrote a set of items aimed at eliciting both objective and subjective responses within each domain. After being reviewed by a panel of HIV/AIDS health-care providers, the resulting instrument draft consisted of 83 items. A third, and perhaps most seldom seen, method of inductive development is theory based. According to Vallerand, Breckenridge, and Hodgson (1998), the absence of formal theoretical models to guide the development of QOL measures has limited the use and generalizability of QOL assessment in clinical trials and population surveys. An example of this method is provided by Holzemer et al. (1998), who first explicated a theory before developing their Living With HIV Scale. Their first step of the generative phase was to discover a grounded theory of QOL in advanced HIV/AIDS that could offer a strong conceptual base on which to construct domains and items for an HIV-specific QOL assessment tool. Using the grounded theory approach introduced by Glaser and Strauss (1967), they performed in-depth interviews with persons living with HIV/AIDS, caregivers, and expert nurses in the field. Their grounded theory, "Salvaging Quality of Life," formed the basis for the second phase of their tool development, which involved identifying the domains and items for the scale. The second stage of scale development involves the refinement and testing of the instrument. The processes involved in this stage are twofold: reduction of the number of items to the minimum necessary and assessment of the instrument's psychometric properties (Streiner & Norman, 1995). Streiner and Norman describe various criteria to be used in determining which items are retained in a scale, most of which involve having participants from the target population (i.e., those with the characteristics or disease state the instrument is aimed at) complete and respond to the scale. Lynn (1986) describes a judgment-quantification stage, 36  where a panel of experts are polled to determine an item, subscale, and overall scale's relevance or content validity. Most scale developers describe a refinement stage of some kind. For example, Smith et al. (1997) administered the first draft of the MQOL-FHV to 121 HIV-infected volunteers in their refinement stage, the results of which resulted in the elimination of two domains, leaving 10 major domains that now comprise the tool. A description of how an instrument is assessed for its psychometric properties follows. Assessing an instrument's psychometric properties. Psychometric properties important in a QOL instrument include responsiveness, interpretability, validity, and reliability (Guyatt et al., 1996; Juniper, Guyatt, & Jaeschke, 1996). The responsiveness of an evaluative scale (those that measure within-subject change over time) refers to its ability to detect true change over time, and is traditionally measured statistically as effect size (Smith et al., 1997). The interpretability of a scale refers to its ability to convey meaning of its scores, that is, whether a score or change in score represents small, moderate, or large degrees of impairment or change (Guyatt et al.). The remainder of this section will focus on the two broad categories of interest in the context of this study, validity and reliability. The validity of an instrument refers to its ability to accurately measure what it is intended to measure (Harrison et al., 1996). There are three types of validity, namely content, criterionrelated (including concurrent and predictive), and construct (including convergent and  discriminant) validity. Content validity is defined as the extent to which items actually measure a content domain; criterion-related validity refers to the correlation of a scale with either some other measure or trait under study that is used and accepted in the field for the same concept or a predicted future event; and construct validity refers to the theoretical relationship of items to other items and variables (Harrison et al.; Streiner & Norman).  37  Hubley and Zumbo (1996) assert that the establishment of validity cannot be achieved through the reporting of "scanty, disconnected bits of evidence" (p. 214) obtained from a simple checklist of procedures, but rather that the concept of validity is "a broader, more integrative theory in which one amasses an evidential and consequential basis for the two main functions of test interpretation and test use" (p. 214). Streiner and Norman (1995) concur, stating that different types of validity are not separate attributes, but rather all address the same issue, namely the degree of confidence one can place in the inferences that can be drawn from scores on scales. Three types of validity — content, criterion-related, and construct - and how they are assessed, are described below. Content validity refers to issues surrounding item development and selection within a domain or subscale. Content validity is supported in a variety of ways, often during a scale's development phase when the literature is reviewed and synthesized and for when information regarding domains and items of relevance is elicited from expert and patient panels (Lynn, 1986; Smith et al., 1997; Streiner & Norman, 1995). According to Holmes and Shea (1998), if an instrument is developed predominantly on the basis of input from members of the target group, it is most likely to result in items that are relevant to its intended target population. One method of assessing content validity described by Streiner and Norman is referred to as a matrix method and can be used to assess a scale's overall content validity, or coverage. It involves identifying dimensions viewed as relevant to the concept being measured (based on previous research, theory, expert opinion and other sources) that form the column headings of the matrix. Each row on the matrix represents an individual item of the scale. Each item then is examined and a mark placed in the appropriate column(s) depending on what content area it assesses (see example in Table 4). Lynn (1986) describes a different method to establish content validity that can be used to evaluate the validity of individual items, subscales, and overall scales. This method involves the 38  Table 4. Example of Content Validity Matrix for Hypothetical QOL Scale Content Area  Item Physical 1 2 3 ...etc.  Psychological  Family  Spiritual  Economic ... etc.  X X  X X  X  assertion by a specific number of experts that the items, subscales, and overall scale are valid and relevant. Each item, subscale, and the overall scale are assigned a score between one to four, where T is not relevant and '4' is very relevant. The scores are then computed to determine the proportion of experts that rate the items as relevant (scores of either 3 or 4), or the proportion of items receiving a rating of 3 or 4 to determine the subscale and overall scale's content validity. Lynn points out that for those items or scales that experts rate 1 or 2, they should identify how the item should be revised or whether or not the item should be deleted. Experts should also identify any area(s) that they consider to be omitted from the scale (Lynn). Criterion-related validity is usually divided into two types: concurrent validity and predictive validity (Hubley & Zumbo, 1996; Streiner & Norman, 1995). To test for concurrent validity, a correlation coefficient is calculated for the scores on the new scale and a criterion measure, generally an existing, "standardized" scale for that same concept with established psychometric properties, or in the case of multidimensional scales, with various scales that measure the specific dimensions of interest (Streiner & Norman). Data for this type of analysis are collected by administering the two (or more) scales simultaneously. For example, Smith et al. (1997) used a total of 20 established criterion measures to assess the concurrent validity of the  39  10 dimensions of their MQOL-HIV. It is expected that if the new scale measures something similar to the standard scale, the correlation should be strong (Streiner & Norman). Predictive validity is concerned with the ability of a scale's scores to predict some future event or criterion (Streiner & Norman, 1995). This type of assessment requires investigators to wait until some future time to confirm or disconfirm the prediction. An example of this is the use of intelligence scores to predict success in college students' academic standing (Streiner & Norman). Construct validity refers to a wide range of approaches that are used to measure a 'hypothetical or latent construct,' rather than something that can be readily observed, such as temperature or weight (Streiner & Norman, 1995). QOL is this type of concept in that it is not observable but rather consists of several domains that are thought to comprise QOL. Several methods exist to assess construct validity, including convergent and discriminant validity, and factor analysis (Streiner & Norman). Extreme groups is a method in which two groups are administered the scale, only one of which is deemed to have the characteristic of interest. One group is expected to score significantly higher on the scale than the other. Convergent validity refers to how two scales purporting to measure the same construct are observed to be related to each other (Streiner & Norman, 1995). This is generally assessed by calculating correlation coefficients between measures; thus, it is expected that theoretically similar measures will produce high correlations. Opposite to convergent validity, discriminant validity refers to the notion that scales that do not purport to measure the same construct are not observed to be related to each other (Streiner & Norman, 1995). Factor analysis is a complex statistical method used to identify clusters of concepts or variables, often used in identifying the subscales or dimensions of a scale (Streiner & Norman, 1995). Data collected from participants completing the scale can be analyzed to determine if the  40  observed structure pattern of the concepts corresponds to the hypothesized one (Streiner & Norman). Reliability refers to the stability and homogeneity of a scale, including the internal consistency of items within a domain and the reproducibility of the scale (Streiner & Norman, 1995). Assessment of the reliability of a scale can be done in several ways, including determining the internal consistency of items within a subscale, inter-observer reliability, intraobserver reliability, and test-retest reliability (Streiner & Norman). Internal consistency of items refers to the correlation of items within the scale. For example, if four individual items in a questionnaire are supposed to measure physical function, they should correlate more strongly with each other, and weakly with other items that measure a different domain, such as family functioning. Data for calculating correlations are collected by administering the scale at a single time to a number of participants. Calculations that can be used to assess correlations include Cronbach's alpha, Kuder-Richardson, or split halves, all of which yield similar results (Streiner & Norman). A correlation of 0.70 or greater generally indicates acceptable internal consistency. Inter-observer reliability refers to the degree of agreement between different observers (Streiner & Norman, 1995). Kappa or Pearson's r are the tests most commonly used on data obtained from a set number of observers' scores of the same group of persons (Streiner & Norman). This assessment is not relevant in cases where scales are self-administered, such as self-reported QOL scales. Similar to inter-observer reliability, intra-observer reliability is relevant only in cases where observers are responsible for data collection, and refers to agreement between observations made by the same observer on more than one occasion (Streiner & Norman, 1995). Streiner and Norman describe a method to assess this form of reliability wherein videotapes of a group of patients are observed and scored on two different occasions by the same group of observers. The resulting coefficient measures variation that occurs within observers, as opposed to between them, as described above.  Test-retest reliability is concerned with the level of agreement in scores made by the same respondents on two separate occasions (Streiner & Norman, 1995). W i t h a relatively stable concept such as Q O L (i.e., it may not be as labile as other concepts, such as anxiety or mood), in the absence of incident, it is expected that a person's Q O L should remain relatively stable over a two-week period (Streiner & Norman). Thus, the most common way of assessing test-retest reliability is to have a group of participants complete the scale on two separate occasions, two weeks apart, and to calculate the correlation coefficient (Pearson product-moment correlation coefficient) for the scores of the items, subscales, and overall scale (Streiner & Norman). The Multidimensional Quality of Life Questionnaire for Persons Infected with H I V or with AIDS (MQOL-HIV) In the following section a description and synthesis of the literature is presented regarding the instrument of interest in this study, the M Q O L - H I V . Description, Scoring, and Administration The M Q O L - H I V is a multidimensional, 40-item instrument that purports to measure 10 dimensions of quality of life specific to persons with H I V ("Multidimensional Quality of L i f e , " 1997). Each domain is measured with four items. Separate scores for each domain are obtained by summing the frequency of scores (l=Never through 7=Always) for the four items in each domain. Certain items require reverse-scoring (7=Never through l=Always). Domain scores range from 4 to 28. In addition to Q O L scores for each domain, a single summary index for overall Q O L can be computed using the formula:  MQOL-HIV INDEX = 2 X Mental Health domain score + Physical Functioning domain score The possible range of scores for the overall Q O L index is 12 - 84, with higher scores representing greater Q O L . For example, i f a person scored the lowest values (i.e., 4) in both the Mental Health and Physical Functioning domains the overall Q O L index would be (2 x 4) + 4 = 12, and conversely, i f a person scored the highest value (i.e., 28) in both domains the index 42  would be (2 x 28) + 28 = 84. The authors claim that an effect size of .20 may indicate clinically important treatment differences or temporal changes. The authors also claim that the MQOLHIV can be administered by telephone, personal interview, or completed by the respondent, and that there are no significant mode effects (Smith et al., 1997). The instrument is rated at a grade eight (presumably American) reading level, and takes 10 minutes to complete (personal communication, K. W. Smith, October 30, 2000). The instrument is currently used in the United States, Spain, South Africa, and Switzerland, in both clinical studies and program evaluations (personal communication, H. Thurston, May 10, 2001). Four primary research papers have been published on its psychometric properties (Badia, Podzamczer, Garcia, Lopez-Lavid, & Consiglio, 1999; Badia, Podzamczer, Casado, Lopez-Lavid & Garcia, 2000; Osowiecki et al., 2000; Smith et al., 1997). With the exeption of the Osowiecki et al. paper, the studies have primarily been aimed at assessing the psychometric properties of the scale. Smith et al. published some of the data from the development of the instrument. Interestingly, Osowiecki et al. used the instrument in a cohort of women, but aside from reporting that scores from five of the subscales and the overall QOL score were similar to previously published scores, no further information regarding the use of the instrument in this particular population was provided. Development The instrument's development occurred over two phases, the first phase served to identify relevant domains and items, and the second to assess the psychometric properties of the instrument and refine the items ("Multidimensional Quality of Life," 1997). The first phase involved semi-structured interviews with 27 volunteers from an HIV/AIDS adult day program and a housing project for persons with advanced AIDS. Four of the participants were women. The open- and closed-ended questions of the interview asked about how being HIV-infected, or having AIDS affected different aspects of the participants' lives, and how a diagnosis of 43  HIV/AIDS affected these areas in detail. Additionally, an expert panel of six health-care providers who treat HIV/AIDS patients were interviewed. From these interviews, 12 major domains were identified and the authors wrote a set of items aimed at eliciting both objective and subjective responses in the 12 domains. After being reviewed by a panel of HIV/AEDS health care providers, the resulting instrument consisted of 83 items. Subsequent refinement of the instrument in the second phase resulted in a reduction to 10 domains, each with four items. Psychometric Properties Content validity. The method used to develop the instrument, as described above, provided some excellent preliminary support for the content validity of the MQOL-HIV in that the domains and items were identified and refined by an expert panel and persons living with HIV/AIDS. In the second phase of development, the instrument was administered to two samples. In one group, 121 HIVinfected volunteers identified through announcements and recommendations by clinicians at a hospital and community health centre (36 were women, 24 were African-American or Hispanic, and 36% reported an AIDS diagnosis) were asked to complete private interviews. Content validity was further assessed using a subset of 16 participants who sorted 83 index cards with items on them into related domains ("Multidimensional Quality of Life," 1997). Although the authors do not explicate the results of this test nor the characteristics of the 16 participants, they claim that this work resulted in the elimination of two of the domains, leaving 10 major domains which now comprise the tool ("Multidimensional Quality of Life"). The steps taken to develop the scale provide strong support for the content validity of the scale. Criterion-related validity. In the development of the instrument, the authors claimed to use a criterion-related approach to select the four items for each of the 10 domains ("Multidimensional Quality of Life," 1997). The results of this concurrent validity testing are shown in Table 5. The validity of 44  Table 5. MQOL-fflV Domain-Specific Criterion-Related Correlations for the MQOL-FHV Domains (adapted from "Multidimensional Quality of Life," 1997)  Domain  Criterion Measure(s)  Criterion Correlation*  Mental Health  Profile of Mood States (POMS) Depression - Rejection POMS Tension - Anxiety POMS Anger - Hostility RAND 36-Item Health Survey (RAND-36) (Hays, Sherbourne, & Mazel, 1993) Emotional Well-Being  Physical Functioning  Sickness Impact Profile (SIP) (Bergner et al., 1981) Mobility SIP Home Management RAND-36 Physical Functioning  .82  Physical Health  RAND-36 General Health  .72  Social Support  Duke-UNC Health Profile (Duke-UNC) (Parkerson etal., 1981) Functional Support Scale  .68  Social Functioning  SIP Social Interaction RAND-36 Social Functioning  .74  Cognitive Functioning  SIP Alertness Behavior Rochester - Boston Drilled Word (Borod, Goodglass, & Kaplan, 1980) Wechsler Adult Intelligence Scale (WAIS) Backward Digit Span Medical Outcomes Study - HIV Health Survey (MOS-HIV) (Wu et al., 1991) Cognitive Function POMS Confusion - Bewilderment  Partner Intimacy  Abbreviated Dyadic Adjustment Scale  .59  Sexual Functioning  Psychosocial Adjustment to Illness Scale (PAIS) Sexual Relationship Scale  .38  Medical Care  RAND Patient Satisfaction Questionnaire  .66  .81  .74  * Pearson correlation between criterion measure and M Q O L - H I V domain score (N=121 development sample subjects).  45  the MQOL-HIV index score was tested by correlating it with the HIV Symptom Index. The four items explaining the greatest amount of variation in the criterion measure were selected using stepwise regression analysis. No subsequent criterion-related validity testing has been published. Further evidence for concurrent validity is reported by Badia et al. (1999), who administered the MQOL-HIV to a cohort of 558 HIV-infected patients who were starting or switching antiretroviral therapy in Spain. They reported that seven of the 11 dimensions and summary 4  scores correlated significantly with the number of HIV symptoms experienced in the Spanish cohort of 558 HIV-infected patients. As well, the physical health, physical functioning, sexual functioning, social functioning, and medical care domains were found to discriminate between asymptomatic and symptomatic or ADDS patients (p<0.05). Construct validity. The developers claimed that convergent and discriminant validity were assessed during the instrument's development by examining whether items in a particular domain correlated higher with each other than with items in other domains ("Multidimensional Quality of Life," 1997). The developers do not provide any detail regarding the results of this testing, but claim that the results of these tests led them to eliminate two of the preliminary domains, resulting in the current 10 domains. Additionally, the index scores discriminated (effect size of 1.65) respondents who reported one or more bed days due to illness from those who had not been ill within the past two weeks (Smith et al.). Reliability. During the second phase of the instrument's development, a longitudinal cohort was constituted of 95 HIV-infected men from a community health centre ("Multidimensional Quality of Life," 1997). They completed the MQOL-HIV at baseline and two weeks and six months  This group used the Spanish language version of the MQOL-HIV. Although the instrument was presumably translated in the United States, no published information is available regarding this translation process nor is evidence of the translated versions validity provided. It is not known if the American Spanish version would be valid in Spain. 4  46  later. Reliability of the domains was tested by examining internal consistency and by using twoweek test-retest correlations. Internal consistency ranged from .72 to .86 for eight of the 10 subscales and .56 and .64 for Sexual Functioning and Mental Health, respectively (see Table 6). Table 6. MQOL-HIV Test-Retest Reliability and Internal Consistency (adapted from "Multidimensional Quality of Life," 1997) Domain  2-Week Test-Retest Correlation (Pearson's Correlation Coefficient)  Internal Consistency (Cronbach's Alpha)  Mental Health  .72  .64  Physical Health  .81  .78  Physical Functioning  .80  .76  Social Functioning  .73  .72  Social Support  .64  .86  Cognitive Functioning  .80  .86  Financial Status  .84  .80  Partner Intimacy  .81  .77  Sexual Functioning  .88  .56  Medical Care  .76  .74  Whereas the developers ("Multidimensional Quality of Life," 1997) reported a test-retest coefficient range of .64-.88, Badia at al. (1999), in the Spanish cohort, reported a range of 0.440.82. Responsiveness. Responsiveness of the instrument was assessed by both the instrument's developers and the Spanish group of investigators. In the sample followed over time, the correlation coefficient of the index scores and self-reports of whether their overall QOL had improved or deteriorated 47  was r=0.52 (p<0.0001) over a 5.5 month period (Smith et al., 1997). Although the Spanish (Badia et al., 2000) results showed that the MQOL-HIV did not detect change due to medical treatment over a three-month period on some of the specific subscales (i.e., medical care, social support, partner intimacy, and sexual functioning), they did not report results of the overall index scores. They did, however, acknowledge that the content of these subscales may not register changes in a three-month treatment period. Critique of the Instrument The primary strength of the MQOL-HIV lies in the method used for its development. According to Holmes and Shea (1998), if an instrument is developed predominantly on the basis of input from HIV-infected individuals, it is most likely to result in items that are relevant to its intended target group. If indeed QOL is conceptualized as something that is subjectively experienced by the target individual or population, then going to the target population to elicit information on what, from their perspective, constitutes their quality of life, is in keeping with the conceptual framework. Of benefit to this proposal is the incorporation of women in the group of HIV-infected individuals involved in the development of the MQOL-HIV, although the scale has not been subsequently validated in women (personal communication, K.W. Smith, October 30, 2000). Because the MQOL-HIV is a relatively new instrument in the field of disease-specific QOL instruments, there exists limited evidence of its reliability and validity. Other than during the development of the instrument, no subsequent assessment of its content validity has been published. Some concurrent validity testing has been performed, but the developers do not comment on the low coefficient found in the Sexual Functioning domain (see Table 5); nonetheless, they claim that the results indicate that the domain scores accurately reflect their intended QOL domains ("Multidimensional Quality of Life," 1997). Regarding reliability, two domains (Mental Health and Sexual Functioning) had a Cronbach's alpha less than 0.70 and one 48  domain (Social Support) had a test-retest coefficient of less than 0.70. These results raise the question of whether the domain scales in question can be considered valid if they do not have acceptable evidence of reliability. Of particular concern is the question of reliability of the Mental Health subscale, as it is used in computing the overall QOL index score. Furthermore, the characteristics of the two samples in which the psychometric testing was done (Badia et al., 1999; Badia et al., 2000; Smith et al., 1997) are somewhat homogeneous, limiting the tool's generalizability. With the exception of the participants included in the first phase of the instrument's development, Smith et al.'s samples consisted primarily of AfricanAmerican and White men, many of whom reported an AIDS diagnosis. According to Smith (personal communication, October 30, 2000), the instrument has not been validated in women or injection drug users (IDU). Although no information is given regarding the characteristics of the Spanish cohort, it is presumably homogeneous (i.e., Spanish) and the study excluded pregnant women and active IDUs. The Spanish cohort consisted of 25% women, but no information regarding gender differences is provided. As well, the subjects were recruited from hospitals, and therefore most were symptomatic or had AIDS. Summary As can be seen in the above review, QOL is a complex, elusive concept that has proven difficult to define. Although dissent in the literature remains, it may be argued that QOL's essential attributes are that: (a) it is multidimensional in nature, (b) it is subjectively experienced, (c) it can be measured, and (d) it is somehow related to the determination of satisfaction or wellbeing within its dimensions. It is abundantly clear that women and persons living with HIV represent unique populations that possess unique challenges and issues in the conceptualization and measurement of QOL. In many cases, it appears that women not only experience QOL differently than men, but also have differences in reporting styles. It is likely that QOL for women may consist of 49  different d i m e n s i o n s than m e n . A s w e l l , w i t h i n the c o n t e x t o f a c u l t u r a l l y d i v e r s e H I V - i n f e c t e d p o p u l a t i o n , the i m p a c t o f disease a n d the c o n c e p t o f Q O L m a y be p e r c e i v e d a n d d e f i n e d differently.  H I V disease, l i k e m a n y c h r o n i c diseases, presents its o w n t w i s t o n c l a r i f y i n g the  elements of Q O L . Measurement o f Q O L requires a clear operational definition upon w h i c h instrumentation can be based.  A g o o d i n s t r u m e n t is o n e that has f o l l o w e d a s y s t e m a t i c d e v e l o p m e n t p r o c e s s that  i n c l u d e s e l i c i t i n g i n p u t f r o m the target p o p u l a t i o n a n d r i g o r o u s testing o f its p s y c h o m e t r i c properties.  It h a s b e e n e s t a b l i s h e d t h a t p s y c h o m e t r i c t e s t i n g o f i n s t r u m e n t s m u s t o c c u r b e f o r e  t h e y are u s e d to p r o v i d e c o n f i d e n c e i n the i n t e r p r e t a t i o n o f results a n d i n f e r e n c e s  made.  T h e M Q O L - H I V is a p r o m i s i n g instrument because it incorporates a m u l t i d i m e n s i o n a l c o n c e p t u a l i z a t i o n o f Q O L . A l s o , it i n c o r p o r a t e s the s u b j e c t i v i t y o f Q O L , i n that m e m b e r s f r o m the target p o p u l a t i o n w e r e s t u d i e d to i d e n t i f y the d i m e n s i o n s o f l i f e o f interest o r o f i m p o r t a n c e to t h e m . P e r h a p s m o s t c o m p e l l i n g l y , it is v e r y e v i d e n t that b e f o r e it is a p p l i e d i n the s t u d y o f H I V - i n f e c t e d w o m e n , the M Q O L - H I V s r e l i a b i l i t y a n d v a l i d i t y m u s t be assessed w i t h i n this p o p u l a t i o n , g i v e n the issues p e r t a i n i n g to g e n d e r p r e s e n t e d here.  50  CHAPTER 3: METHODS This chapter is organized into six sections, including sections describing the overall study design, the context of the study, the sampling methods, the specific methods used for each type of testing, the instruments used, and the ethical considerations used in this study. Design This methodological study employed several approaches to assess the validity and reliability of the MQOL-HIV in a cohort of HIV-infected Canadian women. To test its content validity, a panel of expert health-care professionals and patients were asked to judge the relevance of the items, domains, and the overall scale. To test for concurrent validity, a correlational design was employed to determine the correlation between several of the MQOLHIV and MOS-HIV quality of life subscales. To test for internal consistency, Cronbach's alpha was determined for the total scale and subscales. To test for stability (test-retest reliability), a correlational design was used to determine the correlation of within-subject scores between two time periods, two weeks apart. Test-retest reliability was evaluated for the total scale and each subscale (see Figure 1). Figure 1.  Data Collection Sequence for Reliability and Validity Testing  Time 1  Time 2 (2 weeks)  MQOL-HIV  MQOL-HIV  MOS-HIV Socio-demographic questionnaire QOL status question  51  Context of the Study: The Oak Tree Clinic The study was conducted at the Oak Tree Clinic (OTC) in the Women's Health Centre of British Columbia. The OTC was established in 1994 in recognition of the special needs of women and families living with and affected by HIV (Burdge & Forbes, 1994). The multidisciplinary team at the clinic includes adult infectious disease specialists, general practitioners, paediatricians, pharmacists, outreach workers, obstetrician-gynaecologists, research nurses, clerks, a clinical nurse specialist, a dietician, a social worker, and volunteers who provide childcare. The team provides comprehensive care to 485 HIV-infected adults (73% women) and 75 children annually, with approximately 55 newly diagnosed referrals annually (Burdge, 2000). Approximately 80-90% of the patient population are actively taking antiretroviral therapy (ART), resulting in routine clinic visits every two or three months for follow up. The investigator was the OTC research program coordinator from August 1998, through September 2001. Sampling Methods Content Validity Testing To test for content validity, a panel of experts including health-care providers and patients were purposefully recruited. Purposive sampling is characterized by use of judgment of the investigator to obtain the maximum variation of perspective and situations (Holzemer et al., 1998). Health-care providers were asked to participate as experts, including the clinic physicians, outreach workers, the clinic social worker, and clinic pharmacists. Patients were also included as panel members and were selected on the basis of ethnicity, marital status, age, parental status, mode of HIV acquisition, history of EDU, antiretroviral therapy, and disease stage. Although it was not feasible to establish a panel in which all characteristics were equally  52  represented, purposive sampling ensured that certain characteristics common to the Oak Tree Clinic patient population were not over- or under-represented. In terms of the sample size for conducting the content validation, Lynn (1986) indicates that a minimum of five experts provides a sufficient level of control for chance agreement on item/dimension validity, however a minimum of three was considered acceptable; there was some concern that it would be difficult to locate content/domain experts who would agree to participate. To establish content validity beyond the 0.05 level of significance, a sample size of six would allow for one member to not endorse an item (Lynn); therefore, the number of patient participants recruited was six. Concurrent Validity and Reliability Testing Convenience sampling was used to obtain participants for testing the concurrent validity and reliability of the scale. The target population was HIV-infected women and the accessible population was women who attend the Oak Tree Clinic for HIV care. Inclusion criteria included confirmed HIV infection, ability to read English, willingness to participate, and clinic attendance for a routine appointment. Illiterate and non-English speaking patients were well known to clinic staff, and therefore were not approached. In conducting the concurrent validity and reliability work, sample size considerations included the following. To test the concurrent validity, a sample size of 84 was required, based on an alpha of 0.05, power of 0.8, and effect size of 0.3 (i.e., correlation of 0.30) (Cohen, 1988). To test for test-retest reliability and internal consistency at a set effect size of 0.7, alpha of 0.05, and power of 0.8, a sample size of 13 and 18, respectively, was desirable (Cohen). It should be noted that the sample size calculations were based on the assumption of random sampling, which was violated in this study. However, given the time and resource constraints of this study, a sample size of 84 was set.  53  Patient Sample Recruitment Plan and Data Collection Procedures Potential patient participants were approached by their clinic physician or nurse, who informed them of the study and obtained permission for the investigator to speak to them about the study. When permission was obtained, the investigator discussed the study in detail and obtained the consent of the participant. Consenting women were asked to complete the questionnaire packet at the same time as their clinic appointment and to return it to the investigator who checked it for completeness. The packet provided included a copy of the MQOL-HIV, a copy of the MOS-HIV, and a copy of the sociodemographic questionnaire (see Appendix B). Those participating in the test-retest component were given a second set of questionnaires in a packet that included an instruction sheet, a second copy of the MQOL-HIV, the single-item QOL question, and a self-addressed, postage-paid return envelope. A follow-up phone call was made if the questionnaire had not been returned by the third week. Description of Methods for Each Type of Validity and Reliability Testing To clarify the methods for each study component, the approach and analysis for the content validation is described first, followed by a description of the procedures for the concurrent validity testing, the test-retest testing, and the internal consistency reliability testing. Content Validity The method used to judge the content validity of the instrument was that described by Lynn (1986). Each member of the health-care professional expert panel was provided a packet, either in person (i.e., clinic staff) or by mail, including a copy of the MQOL-HIV and a written set of instructions (see Appendix C). The investigator additionally instructed each member in person how to complete the evaluation. The panel was instructed to rate each of the instrument items on a four-point scale, where '1' was not relevant,' and '4' was 'very relevant and succinct,' and to provide feedback on those items that were thought to require revision. The same process was followed for each of the 10 dimensions of the MQOL-HTV. Panel members 54  were instructed to indicate any omissions they noted in the instrument, either omitted items, dimensions, or both. The experts were asked to judge the instrument independently and to return the completed forms to the investigator, either in person or by mail (a self-addressed, stamped envelope was provided). The panel of patients was divided into two groups of three participants each. Each of these groups met for approximately three hours with the investigator as the facilitator, at a mutually agreed-upon time in the Clinic conference room (refreshments were provided). The purpose was for these patients to judge the content validity of each of the items and dimensions of the instrument. The meetings were tape-recorded. Notes were later taken from the audiotape. At the meeting, each participant was given a copy of the MQOL-HIV and a rating score sheet (see Appendix D). Facilitated by the investigator, the group was first asked to discuss their overall impressions of the scale, followed by a discussion of each item. The investigator guided the discussion by asking participants to give their impressions of the relevance of the items to women living with HIV. As well, the investigator set the pace of the discussion to ensure that all 40 items were covered in the allotted time. Following discussion of the items, the participants were asked to rate each item and dimension on the provided rating score sheet (see Appendix D). The meetings concluded with the identification of any omissions in the scale (e.g., any dimensions or items of relevance that the participants believed had not been included). To analyze the data, the index of content validity (CVI) was used to quantify the content validity, as described by Lynn (1986). The CVI for each item and dimension was determined by the frequency of experts who rated it as content valid (i.e., scores of 3 or 4), and the CVI for the entire instrument was the number of total items judged content valid. The frequency of experts whose endorsement is required to judge content validity depends on the total number of experts. To establish content validity beyond the .05 level of significance, if the total number of experts is five or less, all must agree. For a total number of six or more, the following frequencies of 55  endorsement were required, where the numerator signifies the number of experts endorsing the item or dimension as content valid, and the denominator signifies the total number of experts: 5/6, 6/7, 7/8, 7/9, or 8/10 (Lynn). The CVI data from the panels of health-care providers and patients were analyzed together and separately to identify any potential differences between the panels. Feedback from all participants was examined to identify areas of major concern and to determine whether there were general agreement and support for the scale. Concurrent Validity Consenting women were given one copy each of the MQOL-HIV, the MOS-HIV, the sociodemographic questionnaire, a pencil, a clipboard, and an envelope. They were asked to complete the instruments and to return the completed questionnaire packet in an envelope to the investigator. The investigator was present in the Clinic to check for completeness of the forms and to answer any questions the participants had. Data were analyzed using SPSS 9.0 statistical software (Norusis, 1988). Associations between the respective subscale scores of each instrument were computed using Pearson's correlation coefficient: mental health, physical functioning, social functioning and cognitive functioning were correlated on the two scales. The MQOL-HIV physical health subscale total was correlated with the General Health Perceptions subscale (MOS-HIV). Since an overall summary scale cannot be calculated for the MOS-HIV, it cannot be correlated with the QOL Summary Index of the MQOL-HIV. However, the MQOL-HIV Summary Index as well as the overall instrument gross sum score (i.e., adding all item scores together) were correlated with the MOS-HIV Quality of Life item. According to Streiner and Norman (1995), because any measurement has some associated error, one should expect that correlations between two measures purporting to measure the same attribute should fall in the midrange (i.e., above 0.3). Test-Retest Reliability All of the women who consented completed the MQOL-HIV at the baseline clinic 56  appointment and were given a second copy of the MQOL-FflV and an addressed, stamped envelope before leaving the clinic. The second questionnaire was numbered with each participant's unique identifier. The women were instructed to complete the second MQOL-HIV questionnaire and single-item QOL status question two weeks following the enrolment questionnaire and to mail it to the investigator. To minimize attrition, permission was obtained to follow up with a reminder phone call in two weeks time (there were only a few women at the OTC who requested that they not receive calls from OTC staff because of potential HIV-status disclosure concerns). Therefore, it was expected that most of the women would complete and return the questionnaire after the two-week time period. Mean scores between the baseline subscales and total scale score were correlated with those obtained two weeks later using Pearson's correlation coefficient. According to Streiner and Norman (1995), a correlation coefficient of greater than 0.5 to establish stability would be expected if adequate test-retest stability exists; however, for this study the coefficient was set at 0.7. Internal Consistency All data collected for the MQOL-HIV at the baseline were analyzed for internal consistency. Cronbach's alpha was computed for each subscale and the total scale. An r value of 0.70 generally indicates acceptable internal consistency (Streiner & Norman, 1995). Missing Data Missing data were handled for both instruments using the method recommended by their developers (Wu, 1996; personal communication, K. Smith, May 12, 2001). When at least one half of the individual items in a multi-item scale were completed, a respondent's average score across the completed items in a subscale was substituted for the missing item. For example, if three out of four items for a single QOL domain were answered, the missing one was substituted with the average score of the three completed items.  57  Instrumentation Four instruments were used in this study, including the MQOL-FJIV, the Medical Outcomes Study - HIV Health Survey (MOS-HIV), a sociodemographic questionnaire, and a single-item QOL status questionnaire. Multidimensional Quality of Life Questionnaire for Persons Infected with HIV or with AIDS (MQOL-HIV) The MQOL-HIV instrument has been previously described in the literature review. Medical Outcomes Study - HIV Health Survey (MOS-HIV) The MOS-HIV was one of the first disease-targeted QOL instruments available for the HIV-infected population (Wu et al., 1991; Wu, Revicki, Jacobson, & Matlitz, 1997). The instrument consists of 35 questions making up three broad dimensions - overall evaluation of health, functional status, and well-being - and 10 subscales of QOL: general health perceptions, physical function, role function, social function, cognitive function, pain, mental health, energy/fatigue, health distress, and quality of life (change in health/health transition are  considered separately) (see Table 7). The raw subscale scores are transformed using subscalespecific formulae (Wu, 1996), with possible transformed scores ranging from 0 - 100, where higher scores indicate better QOL. In addition to scores for each subscale, a physical health summary score and a mental health summary score on a 0-100 scale can be computed. According to the developers, it takes approximately five minutes to complete and can be administered as a self-, interview-, or telephone-administered instrument (Wu et al., 1997). The MOS-HIV has been completed by tens of thousands of HIV-infected patients in many countries of the world and is used on an ongoing basis in dozens of clinical trials in Canada, the United States, and Europe (Wu et al., 1997). The instrument is available in 14 languages. The development of the MOS-HIV provided some evidence of its content validity. The developers used the dimensions represented in the Medical Outcomes Study Short Form 20 58  Table 7. Dimensions, Subscales, and Number of Items in the MOS-HIV Dimensions of QOL  Subscales (number of items)  Overall evaluation of health  General health perceptions (5)  Functional status  Physical function (6) Role function (2) Social function (1) Cognitive function (4)  Well-being  Pain (1) Mental health (5) Energy/fatigue (4) Health distress (4) Quality of life (1)  Change in health  Health transition (1)  (SF20) (Stewart et al., 1988), adding concepts important to HIV/AIDS patients based on results obtained through reviewing the relevant literature and input from clinic patients, clinical trial participants, and HIV care providers (Wu et al., 1997). Expert panels of HIV-specialized clinicians and HIV-infected patients identified relevant content items (Wu et al., 1991). No details regarding the results of these assessments are available, nor are subsequent content validity data published. Evidence for validity was first reported by Wu et al. (1991) who found that scores for individual items correlated with scores for other items within the same, hypothesized subscale, and these correlations exceeded correlations with scores from other subscales. Burgess, Dayer, Catalan, Hawkins, and Gazzard (1993) assessed construct validity and found that most scales measuring aspects of physical health or functional performance showed statistically significant differences between groups of patients with different disease stages. In 1997, Wu et al. reported multi-trait analyses that provided evidence for convergent and discriminant construct validity.  59  Regarding concurrent validity, there is a large body of data showing the relationship between MOS-HIV scores and other indicators of health. MOS-HIV scores are moderately and significantly correlated with scores on numerous, standard health status instruments, symptom indices, performance measures, and clinical and examination findings, including the Sickness Impact Profile, Quality of Well-being Scale, cognitive function tests, and the Standard Gamble Utility (Wu et al., 1997). When comparing the MOS-HIV and the HIV-Related Quality of Life Questions (HJVQoL), Burgess et al. (1993) found that correlations between subscales on the two scales purporting to measure similar constructs were highly significant, ranging from coefficients of 0.59 to 0.66 (p<.001 for all). Correlations between the Mental Health Scale and the HAD Anxiety and Depression scales were -.74 (p<.001) and -.57 (p<.001), respectively. Correlations between Physical Functioning and Basic ADL were unacceptably low, likely due to a ceiling effect (most subjects were not ill). Regarding internal consistency reliability, data are available from studies in populations with asymptomatic infection (Wu et al., 1991), early AIDS-related complex (Wu et al., 1991), AIDS (Wu et al.,1991), within specific CD4 count ranges (Revicki, Wu, & Brown, 1995), and in samples with all stages of HIV disease (Burgess et al., 1993; Revicki et al.). In most cases, the Cronbach's alpha coefficients exceeded 0.70 for each subscale and the two summary scales. The exception is the Role Functioning scale, which initially produced a coefficient of 0.50 (Wu et al., 1991) but has subsequently produced correlation coefficients greater than 0.70 consistently (Wu et al., 1997). Regarding use of the MOS-HIV in women, McDonnell et al. (2000) reported "adequate" internal consistency in a cohort of 287 HIV-infected women, citing alphas ranging from 0.64 to 0.89. Revicki et al. (1995) reported similar results in a cohort of 180 African-American women.  60  The limitations of the MOS-HIV include its limited use in women and injection drug users and its potential ceiling effects (Wu et al., 1997). Although preliminary findings are purported to demonstrate reliability and validity comparable to that reported elsewhere, Wu et al. postulate that the overwhelming effects of IDU and the accompanying chaotic lifestyle may reduce responsiveness. Similar to other HIV-specific instruments, the MOS-HTV was developed and refined in an era prior to HAART, which has resulted in a now higher proportion of asymptomatic, "well" patients. This threatens to increase ceiling effects, because higher percentages of patients are likely to rate themselves at the highest possible score on some of the scales. Finally, although the MOS-HIV was developed for persons living with HIV, its developers (Wu et al., 1997) claim that it is a generic tool and therefore may not focus on some of the more specific problems and concerns of HIV-infected populations. Because the developers of the MQOL-HIV used data elicited from HIV-infected individuals to develop their instrument, it is possible that the dimensions and items represented on the scale tap the domains of salience to HIV-infected populations. As well, the MQOL-HIV is considered to be a disease-specific instrument, and therefore may be more sensitive than the MOS-HIV. Socio-demographic and Health Information Forms Relevant socio-demographic and health information was collected by means of a selfreport socio-demographic form and chart review (see Appendices B & E). Socio-demographic data collected included age, country of origin, race, religious preference, education level, occupational/financial status, marital/family status, and illicit drug use. Health information collected included duration of HIV infection, type and duration of antiretroviral therapy, comorbid medical conditions, occurrence of infections, relevant laboratory data, and number of symptoms/side effects. The socio-demographic and health data were intended to describe the sample used in this study and to allow the investigator to determine if the sample was 61  representative of the Oak Tree Clinic population. These specific variables were identified through review of the literature and discussion with the Medical co-Director of the Clinic as factors that may influence QOL in a population of HIV-infected women. Quality of Life Status Change Question Regarding the test-retest approach, it was possible that changes might have occurred during the two-week interval that influenced the participants' QOL. Smith (in "Multidimensinal Quality of Life," 1997) recommends including a question that assesses whether participants' overall QOL changed over the two-week period when performing test-retest assessments. For this reason, a single-item question was asked at the second testing period to identify any major changes in overall QOL (see Figure 2). The question was developed specifically for this study by the investigator.  Figure 2. Single-Item QOL Status Change Question  "Compared to two weeks ago, how would you rate your overall QUALITY OF LIFE now?' [check the box that best describes your answer] Much better Somewhat better now than two now than two weeks ago weeks ago • •  About the same as two weeks ago •  Somewhat worse now than two weeks ago  •  Much worse now than two weeks ago •  Protection of Human Rights Ethics and institutional approval was obtained from the UBC Behavioural Research Ethics Board and the Children's and Women's Health Centre of BC Research Review Corrimittee, respectively. All participants were approached in person by their clinic physician or nurse, who obtained permission to have the investigator introduce the study in more detail. The 62  investigator verbally explained the study to each potential participant. This discussion included details pertaining to the protection of confidentiality, management of data (i.e., the use of unique identifiers, not names or initials), plans for dissemination of results, and the voluntary nature of the study. Permission was obtained to review the participant's clinic medical record to obtain the needed health data. Once participants indicated that they understood the study and had no further questions, the investigator had them sign an informed consent form, a copy of which was given to the participants (see Appendices F & G). Participants retained the right to refuse to participate or to withdraw at any time without jeopardizing their ongoing medical care at the Oak Tree Clinic. Health-care professionals who participated in the content validity testing were approached in person by the investigator and written, informed consent was obtained (see Appendix H). To protect confidentiality, all consenting participants were assigned a unique numerical identifier. All study documents (i.e., questionnaires and scoring sheets) were numbered using these identifiers, and no names or initials appeared on any study paperwork. Electronic data were protected by password and did not contain any data that identified individual participants (i.e., names, initials). Hard copies of study data were stored in a locked filing cabinet in a locked, alarmed office. Audiotapes were destroyed following the study analysis. The investigator obtained copyright permission to use both the MQOL-HIV and the MOS-HIV in this project and to publish the results. Summary A description of the research design, study context, sampling methods, data collection and analysis methods, instrumentation, and protection of human rights considerations has been presented in this chapter. This methodological study employed several approaches to assess the content validity, concurrent validity, internal consistency and test-retest reliability of the MQOLHIV. Patients and health-care professionals were recruited from the Oak Tree Clinic using 63  purposive and convenience sampling methods. Data for content validity testing were collected during patient work groups and by having the health-care professionals complete the rating forms independently. Data for concurrent validity and reliability testing were collected by having patient participants complete the MQOL-HIV, MOS-HIV, and sociodemographic questionnaire at baseline, and a second copy of the MQOL-HJV two weeks later. Data were analyzed using the index of content validity to assess content validity, Pearson's correlation coefficient to assess concurrent validity, Cronbach's alpha to assess internal consistency, and Pearson's correlation coefficient to assess test-retest reliability. Ethics and institutional approval were obtained from the University of British Columbia Behavioural Research Ethics Review Board and the Children's and Women's Health Centre of BC Research Review Committee. All participants provided informed, written consent.  64  CHAPTER 4: FINDINGS In this chapter, the findings of the study are presented. A description of the sample is provided followed by a presentation of the results according to each research question. Characteristics of the Sample The sample consisted of (a) the patients and health-care professionals who comprised the content validity expert panel and patient work groups, and (b) the patients who completed the questionnaires. The Patients and Health-Care Professionals Who Comprised the Expert Panel and Work Groups Ten health-care professionals originally consented to participate in the expert panel, but only seven completed and returned the content validity rating packages, resulting in a response rate of 70%. The seven participants included three physicians, two pharmacists, one social worker, one dietician, and one outreach worker. Two were male and all but one of the professionals worked primarily in the Oak Tree Clinic. All six patients who were invited consented to participate in the work groups. The mean age of the group was 38 years. Two identified as Caucasian, one Japanese, one Indo-Canadian, one First Nations, one African, and all but one were born in Canada. Two were married, one of whom reported having a partner infected with HIV, three were single, and one was widowed, having lost her partner to AIDS. Five of the women had at least some college or university education. One of the women had never had children, one was married with children, three were single mothers, and one had lost her only child to AIDS. Of the women with living children, none reported having HIV-infected children. Four of the women reported acquiring HIV infection through heterosexual contact, and two reported sexual contact and dirty needle use as joint acquisition factors. The mean duration of HIV infection was 7.8 years. Five of the women were on antiretroviral medication (ART), with a mean duration of ART therapy of five years. One of the women was hepatitis C co-infected, and two reported lipodystrophy. 65  The Patients Who Responded to the Questionnaires In an effort to avoid selection bias, where possible, all women who presented for appointments over the data collection period were approached for participation by clinic staff. On some days, the staff were unable to approach all of the women who attended the clinic because some left prior to being approached. Almost all of the women who were approached agreed to participate, but approximately 10-12 left the clinic before completing the questionnaires and were therefore dropped from the study. The demographic, HIV acquisition, substance use, and clinical characteristics of the 85 participants are summarized in Tables 8 - 10, with comparison characteristics of the total Oak Tree Clinic (OTC) female population where available. The OTC database was under construction at the time of the study and did not contain complete data on all of the participant characteristics that were collected in this study; therefore, only those data that were available are included. In terms of demographic characteristics (see Table 8), the sample ranged from 19 to 60 years of age (M=36.5). About two thirds of the sample identified as Caucasian, one fifth identifying as First Nations or Aboriginal. Most women were born in Canada (85%) with the remainder born in Africa (6%), Western Europe (5%), Asia (4%) or other locations (1%). Almost one half of the women had either partial or complete high school preparation and almost 40% had some college or university education. Many of the women were unemployed and receiving welfare or disability payments (60%). A small number were employed part- or fulltime (26%). Almost two thirds reported a gross annual household income below $30,000. In terms of marital status, about one third did not have a partner at the time of the study. Of the two thirds who had a partner, almost one third reported that their partner was also HIVinfected; one half reported that their partner was not HIV-infected, and about 10% were not aware of their partner's HIV status. Two thirds of the women reported having children, and five 66  Table 8. Demographic Characteristics of the Sample (n=85) Characteristic  Frequency (%)  Mean age 36.5 years (SD=9.5) Range 19-60 years Ethnicity/Race Caucasian First Nations/Aboriginal Black Asian South Asian Hispanic Other or mixed  52 19 6 3 1 1 3  (61.2) (22.4) (7.1) (3.5) (1.2) (1.2) (3.5)  28 27 4 3 3 18  (32.9) (31.8) (4.7) (3.5) (3.5) (21.2)  2  (2.4)  Schooling (completed or partial) Elementary school High school College/university Missing  3 40 34 8  (3.5) (47.1) (40.0) (9.4)  Annual Gross Household Income <$10K $10-30K $31-50K >$50K Don't know Missing  24 32 9 14 1 5  (28.2) (37.6) (10.6) (16.5) (1.2) (5.9)  15 7 51 12  (17.7) (8.2) (60.0) (14.1)  Religion Protestant Catholic Atheist Native spirituality Buddhist All other (i.e., Jehovah's Witness Agnostic, Sikh, "other") Missing  Occupational Status Employed full-time Employed part-time Unemployed (on disability and/or welfare) Other or combination (i.e., part-time and/or welfare/disability)  67  Table 8 (continued). Demographic Characteristics of Sample (n=85) Marital Status' Single Married Divorced/separated Widowed Living with partner > lyr "Other" Missing  26 20 13 5 18 2 1  (30.6) (23.5) (15.3) (5.9) (21.2) (2.3) (1.2)  Partner HIV+ (n=59) Yes No Don't know "Not applicable"  18 30 6 5  (30.5) (50.8) (10.2) (8.5)  Any Children Yes No  59 26  (69.4) (30.6)  HIV+ Children (n=59) Yes No Don't know Missing  5 51 1 2  (8.5) (86.4) (1.7) (3.4)  Child Living Arrangements (n=59) With woman With father (not woman) With other family (not woman) Foster/adopted out Adult children (on their own) Other or combined answer Missing  23 3 4 6 10 11 2  (39.0) (5.1) (6.8) (10.2) (16.9) (18.6) (3.4)  Currently Pregnant Yes No  9 76  (10.6) (89.4)  2  3  Only one answer was allowed on the sociodemographic questionnaire; therefore, only those who identified as "single" did not have a partner. Gender of partner not specified. For unknown reasons, some women who previously reported having a partner answered "n/a" to this question. 1  2 3  68  of these (8.5%) had at least one HIV-infected child. Nine (10.6%) of the women were pregnant at the time of the study. In terms of how the women became infected with HIV (see Table 9), mode of acquisition was reported by most as sexual contact (57.6%), and the rest by sharing dirty needles (15.3%), combined sexual contact and dirty needle use (7.1%), sexual contact plus several other risks (2.4%), receipt of blood or blood product transfusion (2.4%), and unknown (11.8%). One quarter of the women reported street drug use within the six months preceding the study. Mode of drug use was primarily swallowing (81.8%) and injecting (63.6%), followed by snorting (27.3%), and smoking (18.2%). Nine (10.6%) of the women reported having been in detoxification or rehabilitation programs in the six months preceding the study, and eight (9.4%) were on methadone for treatment of heroine addiction at the time of the study. In terms of the HIV and clinical characteristics of the sample (see Table 10), most of the women were able to recall the year of their first positive HIV test, and the mean interval from then to the study time was 5.4 years. Fewer (69.4%) women were able to recall the actual year of HIV acquisition, for whom the mean duration of HIV infection at the time of the study was 8.2 years. Ten of the women in this sample were newly diagnosed with HIV infection in 2001. One half of the sample was taking antiretroviral medications at the time of the study, and most were taking other, non-ART medications, generally in multiple combinations. In terms of symptomology, most of the sample reported at least one symptom or side effect; the most common (i.e., those reported by over 25% of participants) are listed in Table 10.  69  Table 9. HIV Acquisition and Substance Use Characteristics of the Sample (n=85) Characteristic  Frequency (%)  Characteristics of Total Oak Tree Clinic Population  Mode of HIV Acquisition Sexual contact Dirty needle sharing Blood/product transfusion Combined sexual/dirty needle sharing "Other" or Combined Answer Don't know Missing  49 13 2 6 4 10 1  Sex with HIVinfected male 76%; dirty needle sharing 44%; blood products 2%; sex trade 10%; unknown 10%  (57.6) (15.3) (2.4) (7.1) (4.8) (11.8) (1.2)  !  Any Drug use in Past 6 Months Yes No Missing  22 62 1  (25.9) (72.9) (1.2)  31% of the total OTC population is currently not using street drugs  Mode of Drug Use (n=22) Injected Snorted Smoked Swallowed  14 6 4 18  (63.6) (27.3) (18.2) (81.8)  13% of the total OTC population is currently actively using injection drugs  2  1  3  Data not available  Detox/Rehab in Past Six Months Yes No Missing  9 75 1  (10.6) (88.2) (1.2)  Currently On Methadone Yes No  8 77  (9.4) (90.6)  Data not available  Multiple choices allowed; therefore, total does not add up to 100%. This represents only those patients that are known not to be currently using street drugs, and does not imply that the remaining 69% are. This represents patients who are known to be actively injecting drugs and is likely an underestimation of the true number. 1  2  3  70  Table 10. HIV and Clinical Characteristics of the Sample (n=85) Characteristic  Frequency (%)  Mean Years Since First HIV+ Test 5.4 yrs (SD=4.0) (n=73) Range 0 - 17 yrs Mean Years Since Known HIV Infection 8.2 yrs (SD=5.2) (n=59) Range 0-20 yrs Mean CD4 415 copies/ml (SD=218) Mean Viral Load 54,754 copies/ml (SD=121,396) Range undetectable - 643,000 copies/ml Currently Taking Antiretroviral Therapy (ART) Yes No  45 40  (52.9) (47.1)  Type of Non-ART Meds (n=64) Vitamins, herbals, or oral contraceptive only Multiple other medications  16 48  (25.0) (75.0)  Concurrent Medical Diagnosis  34  (40.0)  HCV co-infection  12  (14.1)  Lipodystrophy or lipid abnormalities  11  (12.9)  Any Symptoms/Side Effects (SE)  74  (87.1)  Most Common Concurrent Symptoms/SE Fatigue Headache Depression Nausea/vomiting Night sweats Rash/itch Diarrhea Anorexia General body pain Numbness/tingling Abdominal pain  49 43 39 30 28 28 27 27 27 23 22  (66.2) (58.1) (52.7) (40.5) (37.8) (37.8) (36.5) (36.5) (36.5) (31.1) (29.7)  Mean Duration on ART 39.6 months (SD=32.3) Range 1-149 months  71  In terms of representativeness, the sample was very representative of the total Oak Tree Clinic (OTC) population in terms of age, race, mode of HIV acquisition, injection drug use, symptomology, and virologic and immunologic status (i.e., CD4 and HIV viral load). Most of the Clinic population that was pregnant at the time of the study participated, and therefore the percentage of pregnant women in the sample was greater than that of the OTC population. In terms of street drug use, it should be noted that while 31% are documented in the OTC database as not currently using street drugs, this does not imply that the remaining 69% are; therefore, it is difficult to determine whether the sample is representative of the general population in terms of substance use. In terms of hepatitis C (HCV) co-infection, 14% of the sample reported being HCV-positive, while the OTC database indicates that over one half of the patients have laboratory-confirmed HCV-positive results. Results for Each Type of Validity and Reliability Testing Prior to presenting the results on the validity and reliability testing, descriptive information is presented for the MQOL-HIV domain scale scores. Descriptive data including the frequency distribution of the domain scores are presented in Table 11 and of the Summary Index Scores (SIS) and total instrument (i.e., sum of all 40 items) in Table 12. The tables also present the possible range of scores, the mean and standard deviations (SD), groupings of scores, and the percentage that each grouping represents of the total number of respondents for each domain, the SIS, and the total instrument. For each of the 10 domains a lower score indicates a lower perceived level of the specific domain, and for the SIS and total instrument, lower scores indicate lower overall perceived QOL. Hence, when a high percentage of respondents have low scores for a particular domain, it indicates that the respondents perceive a lower level of the specific domain. Likewise, when a high percentage of respondents have low scores for the SIS or total instrument, it indicates that the respondents perceive low levels of overall QOL.  72  Table 11. Distribution of the MQOL-HIV Domain Scores (n=85) Domain (possible range of scores)  Overall Mean (SD)  Score Groups (low score reflects lower QOL rating in that domain)  Mental Health (4-28)  17.8  4-9 10-15 16-21 22-28  (low)  4-9 10-15 16-21 22-28  (low)  4-9 10-15 16-21 22-28  (low)  Physical Health (4-28)  Physical Functioning (4-28)  20.6  20.1  (5.1)  (5.9)  (6.1)  (high)  (high)  (high)  Frequency (%)  7 16 41 21  (8.2) (18.9) (48.2) (24.7)  5 11 24 45  (5.9) (12.9) (28.3) (52.9)  5 15 26 39  (5.9) (17.6) (30.6) (45.9)  Social Functioning (4-28)  18.2  (6.1)  4-9 10-15 16-21  (low)  7 22 38 18  (8.2) (25.9) (33.0) (32.9)  Social Support (4-28)  20.8  (6.3)  4-9 10-15 16-21 22-28  (low)  2 16 20 47  (2.4) (20.0) (23.5) (54.1)  4-9 10-15 16-21 22-28  (low)  7 13 26 39  (8.2) (15.3) (30.6) (45.9)  4-9 10-15 16-21 22-28  (low)  8 19 19 39  (9.4) (22.4) (22.3) (45.9)  4-9 10-15 16-21 22-28  (low)  3 15 15 26  (5.1) (25.4) (25.4) (44.1)  Cognitive Functioning (4-28)  Financial Status (4-28)  Partner Intimacy (4-28) 1  20.2  19.3  19.9  (6.4)  (6.4)  (6.5)  (high)  (high)  (high)  (high)  73  Table 11 (continued). Distribution of MQOL-HIV Domain Scores (n=85) Sexual Functioning (4-28)  Medical Care (4-28)  17.2  20.9  (5.2)  (5.4)  4-9 10-15 16-21 22-28  (low)  4-9 10-15 16-21 22-28  (low)  (high)  (high)  6 20 35 24  (7.1) (28.2) (41.2) (23.5)  3 12 25 45  (3.5) (14.1) (29.5) (52.9)  For 59 women who reported having a partner.  Table 12. Distribution of the MQOL-HIV Summary Index and Total Instrument Scores Scale (possible range of scores) Overall Mean (SD) Summary Index Score (12-84)  Total MQOL-HIV (40-280) 1  55.7  (14.1)  Score Groups (low score reflects lower overall QOL) 12-23 24-35 36-47 48-59 60-71 72-84 40-63 64-87 88-111 112-135 136-159 160-183 184-207 208-231 232-255 256-280  (low)  (high) (low)  (high)  Frequency (%)  1 8 13 29 22 12  (1.2) (9.4) (15.3) (34.1) (25.9) (14.1)  0 0 0 2 9 10 16 6 13 3  (0.0) (0.0) (0.0) (3.4) (15.3) (16.9) (27.1) (10.2) (22.0) (5.1)  'For the 59 women who reported having a partner: the remaining 26 did not complete four of the items.  74  As can be seen in Table 11, the majority of respondents scored in the midrange of the possible domain scores in all domains except Social Support and Medical Care, in which case more than 50% of respondents were in the highest score category. Similarly, most respondents scored in the midrange of the possible SIS scores (Table 12). In terms of the distribution of the total instrument, none of the respondents scored in the lowest three score categories, and most of the respondents scored in the mid- to high-mid range of possible scores. The remainder of this section presents the analysis results according to the specific research questions. Research Question #1: Content Validity The first research question related to the content validity of the scale. The six patient participants were split into two groups of three that met on two separate occasions for the patient work groups. Each group session took about three hours. For both groups it took approximately one half hour for orientation to the background and purpose of the project. Some of the patients had trouble initially grasping the concept of content validity, but with careful explanation and group discussion, all participants eventually indicated understanding and participated actively in the process. The health-care professionals who participated in the content validation component also provided positive feedback and were supportive of the study. All claimed that the instructions provided were clear, and that the process was straightforward. Most reported that it took them about an hour to complete the evaluation. CVI of the items. Table 13 shows the Content Validity Index (CVI), as rated by the combined expert panel members of the health-care professional and patient work groups, for each of the 40 MQOL-FUV  75  Table 13. Content Validity Index (CVI) of the MQOL-FflV Items as Rated by the Health-Care Professional Expert Panel (HCP) and Patient Work Groups (PWG) Item (i.e., "How often in the past 2 weeks ...")  Mental Health Domain 1. You felt depressed 2. You felt anxious 3. You felt needed 4. You worried about things Physical Health Domain 5. You had health problems 6. You were too tired to do things you wanted to do 7. You were bothered by nausea 8. You were bothered by tingling sensations  Combined HCP HCP and (n=7) PWG (n=13)  Social Support Domain 17. You have received enough emotional support from people close to you 18. You had someone who could help you in an emergency 19. You had someone you could talk to about problems 20. You had someone who could give you useful advise Cognitive Functioning Domain 21. You were bothered by trouble remembering things 22. You were bothered by forgetting what you started to do 23. You have found that your mind wanders more than usual 24. You were bothered by a short attention span  6 5  12  7 7 5* 7  10 13 13 11  7 7 7 6  6 6 5  5* 5*  5 6  6 6  5  13 13 12 12  7 7 7 7  6 6 5 5  13  7  6  11 13 13  7 7 7  4*  13 13 13 13  7 7 7 7  6 6 6 5  13 12 9*  Physical Functioning Domain 9. You were able to do things around the house 10 10. You were able to get from place to place, such as shopping 11 and to the doctor 11. You were able to climb several flights of stairs without help 11 9* 12. You could perform strenuous sports like running or weight lifting Social Functioning Domain 13. You felt isolated 14. You have withdrawn from socializing 15. You lacked energy to socialize with friends 16. You spent quality time with friends  PWG (n=6)  4*  5  3*  3*  6 6  76  Table 13 (continued). Content Validity Index (CVI) MQOL-HIV Items as Rated by the HealthCare Professional Expert Panel (HCP) and Patient Work Groups (PWG). Financial Status Domain 25. You had to put off paying regular living expenses 26. You have not had enough money to pay for medication 27. You have not had enough money to pay for recreational activities 28. You were concerned about your financial future Partner Intimacy Domain 29. You felt satisfied with the activities you do with your spouse or partner 30. You felt satisfied with the amount of affection your spouse of partner expressed towards you 31. You were able to confide in your spouse or partner 32. You felt mistrust for your spouse or partner Sexual Functioning Domain 33. You felt satisfied with your sexual activity 34. You have wanted sex more often than you have had sex 35. You felt afraid of infecting someone through sexual contact 36. You felt afraid of sexual rejection Medical Care Domain 37. You wanted more information from your health care providers 38. You wanted more support from your doctors and other health providers 39. You doubted that your health providers are doing all they can for your health 40. You were able to get the medical care you needed Overall MQOL-HIV  12 12  7 6  5 6  11 13  7 7  4*  12  7  5  13  7  5  13 12  7 7  6 5  12 11 13 13  7 7 7 7  13  7  6  13  7  6  13  7  6  13  7  6  39/40  37/40  34/40  6  5 4*  6 6  * Proportion of respondents ranking item as '3' or '4' inadequate to achieve statistical significance at level of 0.05.  items and the overall instrument. All participants were asked to rate each item and each domain using a 1-4 scale where 1 was "not relevant," and 4 was "very relevant and succinct." The CVI of the items is the frequency count of participants that rated them as 3 or 4 (i.e., "relevant"), and the CVI of the total instrument is the number of items that achieved the level of agreement required to determine content validity at a significance level below 0.05. When there 77  were six raters, as in the case of the patient work group, an adequate item CVI required five raters to rate it as '3' or '4.' When there were seven raters, as in the case of the health-care professionals, then six were required to rate the items as '3' or '4.' When the groups were combined to total 13, then ten raters were required to consider the item adequate. When rated by the health-care professionals, three items did not receive the required proportion of agreement to be considered content valid, whereas the patient work groups identified six items as not valid. Of those that were rated as not valid, only one item, #3, did not reach an acceptable level as rated by both groups. When the ratings were combined between the health-care professionals and patient work groups, only item #3 did not achieve the required level of agreement to be considered content valid. In terms of specific comments on items rated not valid by the health-care professionals, comments regarding item #3 ("you felt needed") indicated that the wording was too ambiguous. Regarding items #9 ("you were able to do things around the house") and #10 ("you were able to get from place to place, such as shopping and to the doctor"), participants indicated that there should be a reference to baseline capabilities, and that the socioeconomic status and lifestyle of the population could have an impact on how it is answered. As one participant commented, in reference to item #9: ... need to keep in mind that the meaning inferred by this question will vary depending on one's socioeconomic status and may not be at all relevant to some women (i.e., women living on the street, shelters, or both). Also a low score may not be reflective of poor HIV specific health but of other factors (i.e., drug addiction). The patient participants provided more detail in terms of specific comments on items rated not valid. Regarding item #3, all six participants indicated that for women, being needed could be viewed as both positive and negative, although it was scored positively in the MQOL78  HIV. As one woman commented, she felt burdened by being needed by her clients and her young children (she is a single parent and works within the HIV-infected community), and feels a great sense of accomplishment that her adult child does not need her any more. One of the groups suggested that this item be revised to read "you felt appreciated" to make it more genderappropriate.  /  Item #5 ("you had health problems") was considered by most to be too vague, and as one woman put it, "ridiculous," because by definition HIV-infected individuals have health problems. They suggested that this item should be worded in such as way as to give a respondent an indication of what "health problems" might be so that women could separate it from the over-arching experience of living with HIV infection. Item #12 ("you could perform strenuous sports like running or weight lifting") was also labelled ridiculous by many of the women, in that strenuous sports, in particular weight-lifting, were not typical forms of exercise for HIV-infected women. All of the participants stated that they felt this item was "male." As well, two of the women were of First Nations and African backgrounds, and they stated that strenuous sports were not actively pursued by women from their cultural communities, as a rule. Revision suggestions included changing "sports" to "activities," and including examples such as running or lifting heavy objects. Although item #18 ("you had someone who could help you in an emergency") was rated by two of the six participants as not relevant, no comments or revision recommendations were given. Similar to item #12, participants indicated that in item #27 ("you have not had enough money to pay for recreational activities"), paying for recreational activities was rarely a high priority for HIV-infected women, particularly those who are street-involved or have substance abuse problems. It was suggested that "recreational activities" be replaced with "things you do for pleasure."  79  Finally, item #34 ("you have wanted sex more than you have had sex") was considered problematic by the women because they believed that it did not allow for the situation where a woman may be having more sex than she wanted (i.e., non-consensual sex). Many of the participants disclosed having been raped. One woman stated: I don't see anything here that even talks about the fact that we don't always have control, who we have sex with, when, you know, etcetera ... A lot of my clients [women in the Vancouver Downtown Eastside] are raped -1 mean, so have I. It happened because I'm a woman, not just because I have HIV. Additionally, the women felt the item placed too great an emphasis on the desire to participate in the sex act over other forms of intimacy. In addition to feedback on individual MQOL-HIV items, the patient participants provided comments about the validity and relevance of the instrument as a whole. They indicated that the number of items was acceptable, and that it was fairly straightforward to complete. Two of the women did not like the wording of one of the response options, namely "about as often as not," stating that they felt it was the same as "sometimes." One of the participants, who works exclusively within the HIV-infected Aboriginal/First Nations community, expressed serious concerns about the wording of many of the items, stating that negatively worded items, such as item #32 ("you felt mistrust for your spouse or partner"), would be received poorly in the Aboriginal/First Nations community. She stated: I hate to say it because I don't want to make First Nation people seem uneducated. The language in this is above my people .... My clients get bothered by answering these questions because they are so dysfunctional -1 mean, so many First Nations people come from dysfunctional upbringings -  80  and these negative things bring out the truth, but it hurts.... We are raised to never be negative, but to always see the positive. Two of the women with histories of substance abuse and street involvement, and one of the women who currently works with street-involved HIV-infected women, felt that much of the language in the instrument would require simplification and rewording to make it understandable and relevant to this population. For example, it was pointed out that being "able to do things around the house" (item #9) was irrelevant to the homeless, and that "shopping" (item #10) was the term used for shoplifting in Vancouver's Downtown Eastside. From the perspective of a recovering substance addict, one of the women pointed out that withdrawing from socializing (item #14) and isolating (item #13) are considered positive steps in recovery programs, whereas in the MQOL-HIV these are scored negatively. CVI of the domains. The CVI of the 10 domains are shown in Table 14. Similar to that of the items, the CVI Table 14. Content Validity Index (CVI) of 10 Domains of the MQOL-HIV, as Rated by the Health-Care Professional Expert Panel (HCP) and the Patient Work Groups (PWG) MQOL-HIV Domain Mental Health  CVI from CVI from combined HCP and PWG (n=13) HCP (n=7) 7 13  CVI from Patients (n=6) 6  Physical Health  12  6  6  Physical Functioning  11  6  5  Social Functioning  13  7  6  Social Support  13  7  6  Cognitive Function  13  7  6  Financial Status  11  6  6  Partner Intimacy  11  6  5  Sexual Function  12  7  5  Medical Care  13  7  6  10/10  10/10  10/10  Overall Domains of MQOL-HIV  81  of the domains is the frequency of participants that rated them as '3' or '4.' In terms of the numbers required to rate a domain as content valid, 5/6 patients, 6/7 health-care professionals, and 10/13 combined health-care professionals and patients were required to establish significance (Lynn, 1986). All 10 domains achieved the required level of agreement to be considered content valid by both the health-care professionals and patient members. Despite rating the 10 domains as content valid, most health-care professionals and all patient participants commented that there were several important quality of life domains of relevance to women that were missing in the instrument. Table 15 summarizes the missing domains, as identified by the health-care professionals and patient workgroup participants. Table 15. Domains and Aspects Identified as "Missing" from the MQOL-HIV by Health-Care Professionals (HCP) and Patient Workgroup Participants Domain/Aspect  Patients (n=6)  HCP (n=7)  Family/Child Health  X  X  Caregiver Role  X  X  Pregnancy  X  X  Spirituality  X  Medications & Pill Burden  X  Body Image & Self Esteem  X  Stigma & Disclosure  X  Physical Environment  X  X  The majority of comments and suggestions from the health-care professionals referred to family/child health, caregiver role, pregnancy, and physical environment. In terms of  family/child health and the caregiver role, one participant wrote: ... this [family and child health] is usually as important if not more important than the woman's actual health. This issue is the main topic for many [clinic]  82  visits. Focus on: (1) concern about health of children, especially if FJTV+, (2) who will take care of children if they [mother] die, (3) who will take care of [mothers'] parents if they [mother] die, and (4) happiness of children/family. In terms of pregnancy, one participant saw this as an issue that overlapped several other areas, listing "fears, social, stigma, financial, decisions about keeping pregnancy, fears about baby status, fears about [child] apprehension: dimensions - physical health; social function; social health...." Finally, in terms of physical environment, the outreach worker stated, "I think Physical Environment should be added as a dimension. Primarily housing but also neighbourhood. For example, questions related to safety, affordability, stability and proximity to services and supports." The patient participants identified a greater number of missing domains, including all those identified by the health-care professionals, but adding spirituality, medications and pill burden, self-esteem and body image, and stigma and disclosure. All the patients agreed that  women's perceptions regarding parenting, caregiving, and pregnancy were significant contributors to overall QOL. In terms of spirituality, one patient stated that spirituality "also involves fate and destiny. If things hadn't happened the way they've happened, my son wouldn't be my son." Another women explained to the group, " For my people [First Nations] there are four aspects to life, and without spirituality there is no balance, we are not happy, there can be no medicine wheel." Regarding medications and pill burden, one patient explained how this impacted her life, and how it overlapped with body image and disclosure issues: I haven't told people at my job that I have HIV ... how do I explain when we go out for lunch, or I'm out with clients, and they're all having their meal but I can't eat because I have to wait [for taking the medication] ... and then an hour later it's like, oh, ok, now I can eat and everybody is ready to leave the restaurant! Or, I really hate it when people look at me funny because I just have 83  a salad with no dressing or something ... they say "oh you're so skinny because you don't eat" and they don't know that I can't have the other food I want because it doesn't go with my meds and my cholesterol is already through the roof! [hypercholesterolemia is another characteristic of the lipodystrophy syndrome]. Also in terms of self esteem and body image, one patient shared her experiences of cutting herself off socially, refusing to leave her house for months, because she was ashamed of how she looked and upset by having strangers stare at her in public. She said her self-esteem, which used to be healthy, had suffered a great deal. Another woman described her low self-esteem as negatively impacting her ability to obtain what she needed from her health-care providers: I can't seem to ask for all the information I need from the doctors, I mean, I write all these things down before I come in, but then, it's like, I don't have the courage or self-esteem to ask. I'm getting better, though.... I think. Research Question #2: Concurrent Validity The second research question related to concurrent validity. The women completed both the MQOL-HIV and the MOS-HIV for the purposes of concurrent validity testing. Results based on Pearson's correlation coefficients between the two instruments are shown in Table 16. Correlations ranged between 0.55 (Social Functioning) and 0.81 (Cognitive Functioning). All correlation coefficients were significant at a level of p < 0.001 (2-tailed). According to Streiner and Norman (1995), because any measurement has some associated error, one expects that correlations between two measures purporting to measure the same attribute should fall in the midrange (i.e., above 0.3). In this case, all correlations fell within an acceptable range.  84  Research Question #3: Internal Consistency The third research question related to the internal consistency of the domains and the overall scale. Cronbach's alpha was calculated for each of the 10 domains and for the entire instrument. For the Partner Intimacy domain, the Cronbach's alpha was calculated only for the Table 16. Domain Concurrent Validity Analysis (n=85) MQOL-HIV Domain  MOS-HIV Domain  Summary Index Score  Quality of Life  0.60  Entire Instrument  Quality of Life  0.65  Cognitive Functioning  Cognitive Function  0.81  Mental Health  Mental Health  0.78  Physical Health  General Health Perceptions  0.56  Physical Functioning  Physical Function  0.62  Social Functioning  Social Function  0.55  1  Correlation (Pearson correlation coefficient)  1  For 59 women who indicated they had a partner (i.e., those without partners did not score items 29-32).  59 women who indicated they had a partner, as the remaining 26 did not score these four items. Internal consistency could not be calculated for the Summary Index because this is a score that is calculated using a formula, and does not consist of a specific number of items. The internal consistency results are summarized in Table 17. Cronbach alpha levels for the 10 domains ranged between 0.43 (Sexual Functioning) and 0.92 (Cognitive Functioning). The internal consistency of the entire instrument was 0.92. The domains Mental Health, Sexual Functioning, and Medical Care did not achieve the required alpha of 0.70 that indicates acceptable internal consistency, as per Streiner and Norman (1995).  85  Research Question #4: Test-Retest Reliability The fourth research question related to the instrument's stability. Forty-three (50.6%) of the women completed the two-week MQOL-HIV and the single-item QOL status question. Correlation coefficients for the MQOL-HIV Summary Index and total instrument raw scores (i.e., sum of all scores) and the 10 domain scores are summarized in Table 17. The test-retest correlation for the Partner Intimacy domain was calculated only for those 26 women who Table 17. MQOL-HIV Reliability Analyses Domain  Number of items  Internal Consistency (Cronbach's alpha)  1  2-week TestRetest Reliability 0.59  Summary Index Score  3  Entire Instrument  40  0.92  0.74  Mental Health  4  0.68  0.54  Physical Health  4  0.83  0.70  Physical Functioning  4  0.79  0.69  Social Functioning  4  0.86  0.61  Social Support  4  0.85  0.78  Cognitive Functioning  4  0.92  0.74  Financial Status  4  0.79  0.76  Partner Intimacy  4  0.87  0.84  Sexual Functioning  4  0.43  0.50  Medical Care  4  0.69  0.78  4  1 2 3 4 5 6  2  5  5  6  based on 85 baseline questionnaires. for 43 subjects who completed 2-wk questionnaires. All values significant at p < .001. Cronbach's alpha can not be calculated for the Summary Index Score as it represents the result of a formula. sum of all 40 item scores. for 26 subjects who both indicated they had a partner and completed the 2-wk MQOL-HIV. for 59 women who indicated they had a partner.  reported having a partner and completed the two-week questionnaire. All coefficients are statistically significant at p < .001. Although Streiner and Norman (1995) state that a correlation  86  coefficient of greater than 0.5 is expected to establish adequate test-retest stability exists, for this study the coefficient was set at 0.7. Therefore, although all domains correlated above 0.5, only the sum score of the total instrument, and the domains of Physical Health, Social Support, Cognitive Functioning, Financial Status, Partner Intimacy, and Medical Care achieved coefficients of 0.7 or greater. In terms of perceived QOL change over the two-week retest period (i.e., responses to the single item QOL status question), most (72.1%) indicated their overall QOL was about the same as two weeks prior, 20.9% indicated their overall QOL was somewhat or much better than two weeks prior, and 7.0% indicated that their overall QOL was somewhat or much worse than two weeks prior. When the change in Summary Index Score (SIS) from baseline to two weeks calculated by subtracting the baseline SIS from the two-week SIS - was correlated with the scores on the single item QOL status question, the Pearson's correlation coefficient was -.43 (p=0.003). A negative correlation is expected, since a higher score on the single item QOL status scale indicates worsening overall QOL. In terms of the women who did not complete the two-week questionnaire, many women who had agreed to complete the second MQOL-HIV did not return it despite a reminder phone call. When sociodemographic, HIV, and clinical characteristics of the responders and nonresponders were compared, there were no statistically significant differences except that responders were more likely to have children living at home than non-responders (p <.05). Summary The characteristics of the sample and findings for each of the research questions have been presented in this chapter. Seven health-care professionals and six patients participated in the content validity testing, and 85 patients participated in the concurrent validity and reliability testing. In terms of content validity, 37 and 34 of the 40 MQOL-HIV items were rated as valid by the health-care professionals and patient participants, respectively. All ten domains were 87  rated as valid by both groups, although several domains and aspects of QOL were identified as missing. In terms of concurrent validity, the correlation coefficients of the five domains, entire instrument, and Summary Index Score ranged from 0.55 to 0.81 when correlated with similar domains from the MOS-HIV. In terms of internal consistency, the Cronbach's alpha for the 10 domains ranged from 0.43 to 0.92. In terms of test-retest reliability, the correlation coefficients of the 10 domains, the entire instrument, and the Summary Index Score ranged from 0.50 to 0.84.  88  CHAPTER 5: DISCUSSION In this chapter a discussion is provided of the sample, the major findings, and the methodological considerations. The Sample The Expert Panel Members The 70% response rate for the health-care professionals who participated in the panel is considered good. According to Lynn (1986) it can be difficult to find even as many as three substantive experts when conducting content validity testing. Those who participated were representative of the types of professionals normally involved in the care of HIV-infected patients thereby capturing a wide range of experience and perspectives. Those who did not complete the packages were not asked why, but it is likely that particularly heavy work loads at the time of the study were an influential factor. In contrast to the health-care professionals, all women who were initially approached to participate in the work groups consented. The women were reasonably representative of the total sample and general Clinic population in terms of race, age, parenthood, and mode of HIV acquisition; however, they had been HIV-infected longer, had more education, and were more financially affluent than both the sample and the total OTC population. A possible explanation for this was the selection method used. The investigator purposively sampled the OTC population, focusing on inviting women who not only represented a range of social and life history backgrounds (i.e., parenthood, mode of HIV acquisition, known history of substance use or street living), but also who were known to be articulate, easily engaged, and proactive in either their own lives or in working with HIV-infected female groups. This sample therefore included those who perhaps had come to terms with their HJV disease; some had returned to school, embarked on successful careers, or provided volunteer work within HIV-infected communities. 89  If purposive sampling introduced bias, the study's external validity is compromised (Sackett, 1979). It is the opinion of the investigator, however, that a minimum level of comprehension and communication skills was required to participate in the complex task of ascertaining content validity. Despite this aspect of non-representativeness, because the participants had been patients at the OTC for many years, some of whom worked as peer counsellors and volunteers within HIV-infected women's groups, the work group participants were able to reflect on not only how the items and domains of the MQOL-HTV were valid for themselves, but also for the more general female HIV-infected population. The Patients Who Completed the Questionnaires The sample of women who completed the questionnaires was representative of the general OTC female population on most characteristics except in terms of antiretroviral therapy (ART), pregnancy, and hepatitis C co-infection. A smaller proportion of the sample was taking ART perhaps because almost a quarter of the women were newly diagnosed or pregnant. HIVinfected pregnant women typically attend the clinic every two weeks. There was therefore a greater likelihood of enrolling them than members of the general clinic population. Information about HCV status was obtained through self-report, whereas the OTC database contains laboratory-confirmed cases of HCV infection; therefore, it is not possible to determine if the prevalence of HCV infection in the sample was a true prevalence, and therefore representative of the overall OTC population. In addition, the sample was similar in terms of age, country of birth, education, household income, parenthood, and mode of HIV acquisition to a sample of HIVinfected women in BC surveyed by Kirkham and Lobb (1998). Although convenience sampling can introduce bias if those who participate are different from those who do not, and if differences influence how participants respond (Sackett, 1979), it is unlikely that the psychometric properties of the scale would vary significantly depending on whether participants were taking ART, HCV co-infected, or pregnant.  Discussion of the Major Findings Content Validity Overall, the MQOL-HIV achieved a fairly high level of content validity, as rated by both the patients and the health-care professionals. Thirty-four (85%) and 36 (90%) of the items were considered valid by the patients and professionals, respectively, and all of the domains were rated as valid by both groups. These results indicate that, for the most part, the QOL domains and the items chosen to measure them within the MQOL-HIV reflect what is salient in terms of QOL to HIV-infected women. Several of the items were considered problematic, however, and a number of domains were identified as missing from the instrument. A discussion of the specific problematic items and the missing domains and aspects of QOL is presented. Content validity of the items. Of the three items rated as not content valid by the health-care professionals and the six rated as not content valid by the patients, only item #3, "you felt needed," was rated as not valid by both groups. The reasons given for the ratings were different for the two groups. While the health-care professionals simply referred to this item as too "vague" or "broad," the patients indicated that feeling needed in the context of being a woman was not necessarily positive and perhaps was burdensome. This is contrary to how feeling needed is operationalized in the MQOL-HIV, in that high levels of feeling needed indicate a high level of mental health. The burdensome nature of being needed for women living with HIV is also found in the literature. Songwathana (2001) examined the impact of HIV infection on the social obligations and roles of Thai women and found that HIV infection and AIDS often increased the level of responsibility and caregiving roles of women. Bunting (2001) also reports that the tasks of caregiving in the context of families living with HIV become especially burdensome for women. Nussbaum and Sen (1993) suggested that the caregiving role and social gender expectations construct a unique context in the way women with HIV live their lives. Van Servellen et al.  (1998) and Semple et al. (1993) identified unique HIV-related stressors for women that echo some of the issues related to feeling needed raised in this study, namely those of child-rearing, child and family role worries, and marital/partner issues. Although Nussbaum and Sen and Semple et al. acknowledge that these concerns may relate to some men, as did the women in the patient work groups, they felt that they were of specific concern to women. In terms of the item about having "health problems," one half of the patients rated it as not valid, while all of the health-care professionals considered it valid. The women were unanimous in their opinion that, by definition, the fact that they were HIV-infected meant that they had health problems. Even those who referred to themselves as "well" or "healthy," meaning they were not currently suffering from any debilitating HIV symptoms or side effects from medications, stated that they had an over-arching health problem. The discrepancy between the health-care professional's and patient's ratings may be the result of a lack of clarity in the language used in the item. It is possible that the label "health problems" conjures different meanings to health-care professionals and patients. Within the medical community HIV disease is considered a chronic infection that is often asymptomatic. It is possible that health-care professionals consider patients to have health problems only in the context of symptom or side effect occurrence. In contrast, persons living with HIV, even in the absence of symptoms or side effects, may consider HIV infection to be an all-encompassing problem with their health. The item pertaining to the performance of strenuous sports such as running and weightlifting was rated valid by only one half of the patients, although most of the health-care professionals rated it as valid. The patients did not consider strenuous sports to be typical activities pursued by HIV-infected women, particularly running and weight lifting. As well, participation in strenuous sports of any kind was considered a foreign concept to the women of First Nations or African descent. These discrepancies highlight the notion that within-country 92  cultural differences can impact the perception and reporting of QOL (Cella et al., 1995). The choice of sporting activities as examples of physical functioning may have arisen from the predominantly Caucasian male sample used in the development of the instrument ("Multidimensional Quality of Life," 1997). The substitution of "activities" for "sports," with more general examples provided (e.g., running or lifting heavy objects), would likely make this item applicable to both HIV-infected men and women of any ethnic background. One item that generated considerable discussion and was rated not valid by two patients related to whether they wanted sex more often than they had sex. The discussion centred primarily on the discrepancy between sex that is wanted and not received, versus sex that is not wanted but received (i.e., non-consensual sex). Several of the participants disclosed a history of rape and sexual abuse, a common finding in studies of HIV-infected women (Bedimo, Kissinger, & Bessinger, 1997; Boisson & Rodrigues, 2002; Gielen et al., 2000; Kirkham & Lobb, 1998). Sexual abuse and violence is a part of the lives of many women, particularly those involved with injection drug use and street lifestyles (Freeman, Parillo, Collier, & Rusek, 2001) and has been reported as a significant determinant of HIV acquisition for women in particular (Garcia-Morena & Watts, 2000; Seals, 1996). Although violence and sexual abuse is clearly an issue for women living with HIV that could have an effect on such QOL domains as Sexual Functioning and Partner Intimacy, items addressing this issue are absent in QOL scales. Two items rated by the patients as valid were rated as not valid by the health-care professionals: "you could do things around the house" and "you could get from place to place, such as shopping or to the doctor." The health-care professionals expressed concerns because there was no baseline measure and the meaning of the questions might vary depending on the socioeconomic status of the respondents. The patients, however, did not share these concerns but suggested that the wording of both items required revision to make them more relevant to the real lives of HIV-infected women. They were concerned about the assumption that HIV-infected 93  women had a house and that they did things such as shopping and going to the doctor. Although all these patients had homes and did access HIV care, they felt that many HIV-infected women lived on the street and did not access care, nor "shop" in the traditional sense (i.e., "shopping" is the colloquial expression for shoplifting in the Vancouver Downtown Eastside). The socioeconomic disadvantages of HIV-infected women are well documented in the literature. Women living with HIV tend to be of lower socioeconomic status than men, are more often single parents, and may have greater financial stressors (Bunting, Bevier, & Baker, 1999; Lichtenstein, Laska, & Clair, 2002; Nussbaum & Sen, 1993; Semple et al., 1993). The wording of these items may not accurately reflect the situation of HIV-infected women because the MQOL-HIV was developed using a predominantly male sample recruited through HIV/AIDS daycare programs and clinics ("Multidimensional Quality of Life," 1997), and were likely to have had health insurance and housing. It makes sense that the item that pertained to not having enough money to pay for recreational activities was rated as not valid by the patients who felt that women living with HIV did not, in general, spend their money on recreational activities. All the health-care professionals, on the other hand, rated the item as valid. A possible explanation for this discrepancy may lie in differences in the interpretation of the term "recreation." Health-care professionals may consider recreation as something that persons from all sociodemographic strata may participate in, whereas HIV-infected women living in poverty may never consider paying for recreation. A possible explanation as to why this item is part of the Financial Status domain may be associated with the characteristics of the original instrument development sample. Participants were recruited through advertisements in HIV outpatient clinics, which implies that volunteers were likely not from lower socioeconomic populations (i.e., they had health insurance and access to HIV care), and therefore may have been more likely to have had enough money to pay for recreational activities and to have perceived this to be important. 94  Finally, the relevance of the item that pertained to having someone who could help in an emergency did not have unanimous support from the patients, but it is not clear why because no comments were provided. Although each item was brought up for discussion, not all generated discussion and some were passed over more quickly than others. Content validity of the domains and missing aspects and domains. All 10 domains of the MQOL-FJJV achieved the required agreement to be considered content valid. Given the general nature of the 10 domains, it is not surprising that they should be considered salient to the assessment of QOL for HIV-infected women; they appear to be general enough to be relevant to all populations, regardless of HIV status. A possible exception may be Medical Care, which may be of more relevance to diseased populations. Despite the domains rated as content valid, there were concerns related to domains and areas that were missing from the scale. Both the health-care professionals and patients identified two domains of QOL that may be of particular relevance to women, including family/child health and spirituality, and several areas that could be incorporated into existing domains, including caregiver role, pregnancy, physical environment, body image and self-esteem, medications and pill burden, and stigma and disclosure.  Family/child health was identified by health-care professionals and patients as an important, yet missing, domain of QOL. It is perhaps not surprising that the scale does not include this domain given the predominantly male sample used in the instrument's development. In the general population, it is widely understood that women are more likely than men to be single parents and this is perhaps more true for HIV-infected women than men (Metcalfe, Langstaff, Evans, Paterson, & Reid, 1998; Nussbaum & Sen, 1993; Semple et al., 1993). The notion of family and child health becomes particularly important in the context of women having a partner or child who is also living with HIV infection, as found by Songwathana (2001) and Bunting (2001). The notion that family/child health should be viewed as a separate domain of 95  QOL is also supported in the literature. In her development of a conceptual model of QOL, Ferrans (1996) identified family as one of the four primary domains of QOL. Spirituality was identified as an important missing domain of QOL by the patients alone. While they did not identify spirituality as being unique to women, they were unanimously adamant that this was needed in the MQOL-HIV. Two of the women who identified as atheist also agreed that spirituality was important, and that it did not necessarily speak to a faith in God or a Higher Power, but could also refer to a sense of fate. Spirituality as a dimension of QOL is well documented in the literature. In a review of QOL dimensions, King (1998) claims that spirituality is one of the five generally agreed-upon domains of QOL. Schipper et al. (1990) refer to spirituality as encompassing not only religious beliefs but also the perception that life has meaning. Paradoxically, as a domain, spirituality is absent in many, if not most, QOL instruments. Brady, Peterman, Fitchett, Mo, and Cella (1999) point out that spirituality may be particularly salient in the context of life-threatening illness. Their work with an ethnically diverse sample indicated that spirituality is associated with QOL to the same degree as is physical well-being, a dimension unquestioned in its importance to QOL. These findings were similar to those of Gioiella, Berkman, and Robinson (1998), who recommended that spirituality be included in the assessment of cancer patients' QOL. The female caregiver role was identified as a part of the family/child health domain that was missing by both the health-care professionals and patients. The patients also referred to the caregiver role within the context of their workplace. Two thirds of the patients worked in some form of supportive capacity with HIV-infected people. They expressed a very strong burden of caregiving as a result of this work, despite it being seen as positive within their lives. As one participant put it, she worked with HIV-infected First Nations women to facilitate the healing of the spirit, both that of her clients and of herself; she considered herself a caregiver not only to her two dependent children, but also to her clients. 96  Caregiver role stressors, closely tied to the burden of feeling needed, are also supported in the literature. Nussbaum and Sen (1993) and Semple et al. (1993) refer to the caregiving role, or caregiver worries, in the context of unique domains of QOL to women. These findings were confirmed by Songwathana (2001) and Bunting (2001). It should be noted, however, that this does not imply that HIV-infected men do not perform caregiving roles, simply that it may occur more often for women. Pregnancy was clearly identified by both the health-care professionals and patients as being unique to women, and intimately linked to several other domains, such as family/child health, physical health, social support, and social functioning. As many as 80% of HIV-infected women are of child-bearing age, and increasing numbers are becoming pregnant (Forsyth, Davis, Freudigman, Katz, & Zelterman, 2002). At the Oak Tree Clinic, approximately 25 women become pregnant each year. In this context, then, it is not surprising that pregnancy is viewed as an important QOL issue for HIV-infected women. Even pregnancies uncomplicated with HIV are associated not only with feelings of joy, love, a sense of fulfillment, and contentment, but also fear of the unknown and concerns about the future (Wiggins, 1979). Concepts such as identity, self-esteem, role theory, joy, touch and sensuality, frustration and conflict, anxiety, loss, and crisis have been used to describe the  psychosocial effect and experience of "normal" childbearing (Clark, Affonso, & Harris, 1979). It is not surprising, then, that adding to this the complicating physical effects of HIV infection, concerns regarding antiretroviral medication use during the pregnancy, the possibility of transmitting the virus to the infant, and uncertainties regarding ability to live long enough to raise the child may impact HIV-infected women's overall QOL. Body image and self-esteem were also high on the patients' list of concepts not addressed in the instrument. They spoke of having to adjust to dramatic physical changes they were experiencing as a consequence of their HIV disease and of the antiretroviral medications. Two 97  of the four women suffered from extreme lipodystrophy syndrome, a common occurrence for patients on long-term HAART characterized in women by dramatic changes in body fat distribution: wasting of the face, buttocks, and extremities, and fat gain in the breast, abdomen, and upper back ("buffalo hump") regions (Garg, 2000). As well, low self esteem seemed to be a barrier to requesting information from health-care providers. Self-esteem and body image are not likely to be QOL issues unique to women. There is evidence in the literature that HIV-infected men also are concerned with body image. Tate and George (2001), in a qualitative study of body image in HIV-infected men, found that most of the participants avoided social contact because of self-consciousness over their physical appearance. Wagner and Rabkin (1999) found that weight loss correlated strongly with higher levels of depression and lower overall QOL in a sample of HIV-infected men. In addition, there is an increase in cosmetic surgical interventions available for persons disfigured by HIV disease (Berger, 2001; personal communication, D. Peabody, HIV dietitian, July 2001). In terms of self-esteem, it appears to be closely linked to body image and there is evidence in the literature that it is related also to perceived stigma. In a psychometric analysis of the HIV Stigma Scale (Berger, Ferrans, & Lashley, 2001), personalized stigma and negative selfimage were two of the four factors identified. Collins, Wagner, and Walmsley (2000) conducted a qualitative survey of HIV-infected men and women with lipodystrophy and found that erosion of self image and self-esteem were dominant themes for this group, leading to social and sexual problems, forced HIV disclosure, and demoralization. Although perhaps not comprising a separate domain, the concepts of body image and self-esteem may be an important component of Mental Health, but may also be related to domains such as Social Functioning, Partner Intimacy, and Medical Care. Medications and pill burden were additional QOL concerns that the patients identified as missing but important. Of the six participants, five were on ART, and they discussed issues 98  about having their lives dictated by a strict pill schedule and the "rules" of administration (i.e., with or without food, two hours before or after food). All of the women stated that although they recognized the need to take the medication, and indeed they felt their general health had improved because of it, the psychological and daily adjustments required to take ART dramatically affected their overall QOL. This was also closely related to disclosure issues; several women with jobs shared how they tried to hide their pill-taking from colleagues. Although the MQOL-HTV encompasses the domain of Medical Care, it does not address the life-long commitment to what is often a profoundly difficult medication regimen. It should be noted here that the MQOL-HIV was developed in the early 1990's, prior to the era of highly active antiretroviral therapy (HAART) and multiple drug combination therapies. At that time, HIV infected persons were on mono or dual therapy, if anything, and it was not until 1996 when discoveries in HIV treatment began to make significant advances (Williams, 1997). Given this context, it is not surprising that items addressing medication and pill burden are not present in the current instrument. Nonetheless, in the current era of HIV/AIDS, it is evident that items measuring the impact of medication use should be included to fully tap the medical care component of living with HIV. Recent publications support the need to include the impact of HIV medication use when assessing QOL, in that many of the regimens have been found to have negative effects on patients' QOL (Nieuwkerk et al., 2000; Tsasis, 2000), and specific scales to measure the impact of medications are being developed (Aversa et al., 1998). In this study sample, approximately one half of the women were taking ART, but it is a relative certainty that all eventually will. Although it is clear that issues around the life-long commitment to HIV drug treatment can significantly impact QOL, this is not a domain unique to women. Another area that all the patients but only one health-care professional identified as missing was physical environment, in which they were referring primarily to housing and security. The women were unanimous in their perception that QOL was dramatically affected by 99  whether an individual had a safe, secure place to live. It is the unfortunate case of many HIVinfected women in Vancouver to not have stable housing, and in some cases, no housing at all (personal communication, C. Moody, Outreach Worker, Novemer 23, 2001). Although concerns regarding physical environment may not be unique to women, it is thought that HIV-infected women in general have more financial and housing concerns than men (Bunting et al., 1999; Nussbaum & Sen, 1993; Semple et al., 1993). Although perhaps not a specific domain of QOL, physical environment may be incorporated into the Financial Status domain. Finally, stigma and disclosure were identified by the women as significant issues of QOL absent in the MQOL-HIV. The women related these concepts not only to themselves, but also to their children and other members of their family. One single mother expressed her concerns about the need to protect her young children from potential stigma at school if it were known that their mother was infected with HIV. Although all members of the patient work groups had lived with HIV infection for many years, and several were spokespersons for HIV-positive groups, they all felt that stigma and disclosure issues were important in women's overall QOL throughout their lives, regardless of whether or how they had come to terms with their HIV disease. Although stigma and disclosure issues are likely of concern to HIV-infected men, there is evidence in the literature that this domain is particularly salient for women. This notion is supported by Semple et al. (1993), in that the women in their sample were more likely than men to have dependent children, be single parents, and have more difficulty disclosing HIV status in the context of heterosexual dating and to their children. Gaskins and Brown (1992) found that women were more likely than men to be reluctant to disclose their HIV status because they felt the need to protect their children. Similar findings were reported by Sowell et al. (1991) and Moneyham et al. (1996). Women in Metcalfe et al.'s study (1998) reported that they were more likely to be stigmatized than men, particularly if they were neither drug-users nor promiscuous, a 100  finding similar to that of Lichtenstein, Laska, and Clair (2002), who reported that women were subjected to greater stigmatization because of their association with "dirty sex" and moral threat to heterosexual communities, and because they lacked the social support networks of gay men. Women are also more likely than men to find themselves at risk of physical violence in the context of disclosing their HIV status to a would-be partner (Gielen, O'Campo, Faden, & Eke, 1997; Gielin et al., 2000). The MQOL-HIV original development sample did not identify stigma and disclosure as relevant QOL issues, possibly because the sample was predominantly male. However, it is clear that stigma and disclosure are issues of significance to HIV-infected women, and could be incorporated into domains such as Mental Health, Social Support, and Social Functioning. Concurrent Validity The MQOL-HIV demonstrated good concurrent validity in this study. The five domains, the total instrument score, and the Summary Index of the MQOL-HIV correlated moderately to strongly with the MOS-HIV domains. According to Streiner and Norman (1995), because any measurement has some associated error, one expects that correlations between two measures purporting to measure the same attribute should fall in the midrange (i.e., above 0.3). As all correlations fell between 0.55 and 0.81, one may conclude that the scale has good concurrent validity with the MOS-HIV. One of the assumptions of concurrent validity testing is that the instrument(s) chosen to correlate with the study instrument are valid in themselves (Streiner & Norman, 1995). Although its use in female populations has been limited, claims have been made that the MOSHTV has demonstrated acceptable reliability and validity in two studies (McDonnell et al., 2000; Revicki et al., 1995). As well, the MOS-HIV has been in use for over a decade, has been completed by tens of thousands of patients, and many studies have published evidence of its  101  overall validity and reliability. The MOS-HIV, therefore, is likely the best choice as a single, multidimensional QOL instrument for conducting concurrent validity testing. Although the five domains tested and the overall instrument demonstrated acceptable concurrent validity, the remaining five domains - namely Social Support, Financial Status, Partner Intimacy, Sexual Functioning, and Medical Care - were not tested in this study. It is therefore not known if they would perform as well as the five that were tested. To provide comprehensive evidence for the MQOL-HIV s concurrent validity, it would be necessary to correlate it with a battery of scales measuring the 10 domains, an endeavour that would require women to complete a large number of questionnaires at one sitting. This could prove logistically difficult, given the difficulties experienced by the women in this sample in completing three questionnaires. Internal Consistency Seven of the 10 MQOL-HIV domains achieved a Cronbach's alpha of 0.70 or greater, indicating that they were internally consistent. These findings indicate that the four items comprising each domain correlated strongly with each other, and therefore were likely measuring the same concept. The three that did not achieve the required level were Mental Health, Sexual Functioning, and Medical Care. These findings indicate that the four items within each domain were not correlated strongly with each other because they were not measuring a similar enough concept. In terms of the Mental Health domain, a possible explanation for this finding may be inferred from the results of the content validity testing. Both the health-care professionals and patient participants rated one of the items as not content valid, indicating that the item may not be relevant to the mental health of HIV-infected women. If one of the four items is not relevant to the domain it purports to measure, it is not surprising then, that the four items did not correlate strongly with each other. 102  The same explanation may be true for the weak correlations found within the Sexual Functioning domain. Indeed, although the patient participants only rated one of the items as not content valid, there was also a great deal of discussion regarding the wording and relevance of the other three items during the work groups. It seemed as though the women saw sexual functioning as important to the QOL of women, but felt that the emphasis on the functional aspect of sexuality was perhaps less pertinent to women than it may be for men. Consequently, if the sexual functioning items do not accurately tap the sexual functioning domain for women, one would expect that they would not correlate strongly with each other. The weak internal consistency for the Sexual Functioning domain was also a finding of the instrument's authors, who reported a Cronbach's alpha of .56 in their original sample ("Multidimensional Quality of Life," 1997). This evidence leads one to conclude that although the domain of Sexual Functioning - or perhaps more accurately, Sexuality - may be relevant to HIV-infected populations, the items chosen to measure it may not be appropriate. In terms of the internal consistency of the Medical Care domain, all of the items were rated as content valid by both the health-care professionals and patient participants. It should be noted that the Cronbach's alpha missed the 0.70 level by 0.01, suggesting that the possibility of acceptable internal consistency exists. Although all items may be content valid, it is possible that they measure several aspects of a broad concept of medical care that are different enough from each other so as to weaken the inter-item correlations. Test-Retest Reliability The total instrument score and six of the MQOL-HIV domains achieved the required correlation coefficient of 0.70 in the test-retest reliability analysis. If the test-retest reliability results would have been evaluated using the 0.50 cut-off suggested by Streiner and Norman (1995), all 10 domains and the Summary Index Score achieved this level. The level set for this  103  study, however, was 0.70, and therefore these findings indicate that the measurement of most of the domains and the entire instrument were stable over time. The Summary Index Score (SIS) and four of the MQOL-HIV domains, Mental Health, Social Functioning, Social Support, and Sexual Functioning, did not correlate at 0.70 or higher. A possible explanation for this finding for the Mental Health domain (correlation coefficient = 0.54) may be inferred from the results of the content validity and internal consistency testing. As discussed previously, both the health-care professionals and patient participants rated one of the items as not content valid, indicating that the item may not be measuring a unidimensional conceptualization of mental health in HIV-infected women. Additionally, the four items did not correlate strongly enough with each other to achieve an acceptable level of internal consistency. Therefore, if the four items purporting to measure mental health do not in fact do so, it is possible that the domain may have weak stability over time, as evidenced by weak test-retest correlation. A second possible explanation for poor test-retest reliability for Mental Health is that almost 30% (12) of the respondents indicated that their overall QOL had changed over the twoweek retest period, as indicated by the responses to the single-item QOL status question. One of the assumptions of test-retest reliability is that a correlation should be high if the construct being measured is relatively stable over the testing time interval. If changes in mental health were responsible for the perceived overall QOL change, this could explain the poor correlation. If this is the case, it is possible that the test-retest reliability of the Mental Health domain would perform better if only the scores of those who indicated no QOL change were correlated. In terms of Physical Functioning, the correlation coefficient of 0.69 was just below the 0.70 cut-off, suggesting that the domain may in fact be stable. Similar to that discussed above, it is possible that the domain is reliable, but that the scores of the 12 respondents who indicated  104  that their QOL changed introduced enough instability to weaken the correlation coefficient. The same may be the case for Social Functioning (correlation coefficient = 0.61). In terms of Sexual Functioning (correlation coefficient = 0.50), the possible explanations for the weak correlation between scores are similar to those for Mental Health, namely, that it is questionable whether some or all of the items are content valid. As well, the domain as a whole demonstrated the weakest internal consistency of all the domains. In light of these findings, it is perhaps not surprising that the domain demonstrates poor stability. Finally, the poor test-retest correlation of the SIS (correlation coefficient = 0.59) may be explained by the combination of the possible instability of the four domains plus the unacceptable internal consistency of three of the domains. Another explanation is presented by Streiner and Norman (1995). One of the assumptions behind test-retest reliability is that the construct being measured is in itself stable over a short interval of time, such as two weeks. It is possible, therefore, that the overall construct of QOL has some inherent instability. It is equally possible that some domains are more stable than others, in which case the results of this testing do not indicate poor stability, but rather legitimate change in QOL in those specific domains. Because perceived overall QOL change was measured in this study, this may not be the complete reason for the poor test-retest reliability of the SIS. Methodological Considerations A brief discussion of the methodological considerations in terms of data collection methods used for each of the research questions is presented. In terms of the data collection methods for content validity involving the health-care professionals, the content validity rating process was done in private by each respondent, guided by written instructions. It is possible that heavy workloads and a perception of not having enough time to devote to the task may have contributed to the fact that very little feedback in terms of item revision and domain additions was provided by members of the expert panel. 105  Based on the excellent feedback received during the patient workgroups, the amount of feedback received from the professional participants may have increased if the methods had included a short (i.e., half hour), tape-recorded interview with each health-care professional to elicit feedback and comments regarding item revisions and missing domains. In terms of data collection methods used in the two patient work groups, for both groups it took up to one half hour for them to reach an understanding of the notion of "content validity," as this was unfamiliar terminology and an unfamiliar process for all of them. However, without exception each participant was ultimately able to grasp the concept and provide insightful, reflective contributions to the discussion during the work groups. The group discussions resulted in large amounts of feedback regarding why items were or were not considered content valid, how items should be revised, and in particular, the domains of quality of life of relevance to women that were missing in the instrument. Each participant additionally expressed pleasure in being able to participate in the work groups, and several friendships appeared to have been forged, as evidenced by the exchanging of phone numbers. The work group process appeared to have been a positive learning experience for both the participants and the investigator. The methods used for collecting the data for concurrent validity testing involved approaching women when they presented at the clinic for regular appointments and having consenting patients complete the MQOL-HIV, MOS-HIV, and a sociodemographic questionnaire. Most women who were approached were receptive and willing to participate in the study, and the overall feedback from the participants was positive in regards to the importance of measuring QOL. Several of the women, however, experienced exhaustion when faced with completing three questionnaires at a sitting, particularly those who attended the clinic with small children, were feeling physically or emotionally unwell or fatigued (a common experience among this population), or were having concentration difficulties (i.e., those under the influence of street drugs or with cognitive involvement of their HIV disease). Three 106  questionnaires, then, may be too many for this population to complete at one time, and may have introduced bias by excluding those who could not complete the task. It was noted in the results that only one half of the 85 patient participants completed and returned the two-week retest MQOL-HIV. Many women who had agreed to complete the second MQOL-HIV did not return it despite a reminder phone call. Although 43 is well above the sample size calculation of 18 required to determine test-retest reliability, it is not clear how the almost 50% attrition rate may have contributed bias to the results. When comparing characteristics of respondents to non-respondents, the responders differed only in terms of being more likely to have children at home. Because it is not known why one half of the women were unable to complete and return the second questionnaire, it is difficult to identify alternative methods for future studies that may improve the response rate. Summary A discussion of the sample, the major findings, and the methodological considerations was presented in this chapter. The health-care professionals who participated in the content validity testing were representative of the types of professionals who normally provide care for HIV-infected patients. The patients who participated in the content validity testing were reasonably representative of the Clinic population but had been HIV-infected longer, had more education, and were more financially affluent. The patients who completed the questionnaires were representative of the Clinic population except in terms of antiretroviral therapy, pregnancy, and hepatitis C co-infection. In terms of content validity, the specific problematic items and missing domains and aspects of QOL of salience to HIV-infected women were discussed in relation to other findings in the literature. Although the MQOL-HIV demonstrated good concurrent validity, it was pointed out that only one half of the instrument was tested in this study, and that further, more comprehensive concurrent validity testing is required. Results from the internal consistency and 107  test-retest reliability testing were discussed in terms of possible explanations for why certain domains did not achieve required levels of reliability. Finally, a brief discussion of the methodological considerations in terms of data collection methods used for each of the research questions was presented.  108  CHAPTER 6: SUMMARY, LIMITATIONS, CONCLUSIONS, IMPLICATIONS, AND RECOMMENDATIONS In this chapter a brief summary of the study followed by a discussion of the study limitations, conclusions, implications for practice, research, and theory, and recommendations for future research are presented. Study Summary This study was conducted to assess the psychometric properties of a relatively new HIVspecific quality of life questionnaire, the Multidimensional Quality of Life Questionnaire for Persons Living with HIV or with AIDS (MQOL-HIV), in a cohort of Canadian, HIV-infected women. The literature review indicated that women and persons living with HIV possess unique challenges and thus present particular issues to the conceptualization and measurement of QOL. QOL measurement requires a clear operational definition upon which instrumentation can be based, and a good scale is one that has followed a systematic development process and possesses strong psychometric properties. A gap in the literature regarding validation of HIV-specific scales in female populations was identified. The MQOL-HIV was identified as a promising instrument because of its multidimensional operationalization of QOL, its basis on the subjectivity of QOL, and its development that included the use of inductive methods. Although the initial developmental sample included some women, the MQOL-HIV had not subsequently been validated in HIV-infected women. This study was therefore undertaken to address this knowledge gap. This methodological study applied several methods to assess the content validity, concurrent validity, internal consistency, and test-retest reliability of the MQOL-HTV, and was conducted with the staff and patients of the Oak Tree Clinic. Seven health-care professionals  109  and six patients participated in the content validity testing, and 85 patients participated in the concurrent validity and reliability testing. Content validity findings indicated that most of the items and all of the domains of the scale were valid for HIV-infected women, but that revision and replacement of some of the items was needed. Several important domains and aspects of domains were identified as missing from the scale. Five domains, the Summary Index, and the overall scale demonstrated strong concurrent validity when correlated with like domains contained in the more commonly used Medical Outcomes Study - FUV Health Survey (MOS-HIV). Most of the domains demonstrated strong internal consistency and test-retest reliability. The findings were consistent with those found in the literature, namely that QOL scales require revalidation in new populations, and that dimensions and aspects of QOL may differ for HIV-infected women. Although the MQOL-HIV was developed with HIV-infected patients, some of whom were women, the need for subsequent validation in female populations was underscored. The content validity findings support those found by others, such as Bunting (2001), Cella et al. (1995), Nussbaum and Sen (1993), Semple et al. (1993), Songwathana (2001), and Van Servellen et al. (1998), indicating that the QOL domains and aspects of salience may vary depending on gender and cultural context. This study represents one of the first of its kind to validate an HIV-specific QOL scale in women, and provides preliminary findings in terms of the unique dimensions of QOL for HIV-infected women. Study Limitations The limitations of this study relate to the sampling, mode of questionnaire administration, two-week attrition rate, and some of the statistical methods used. The sampling methods used in this study may have introduced bias. The purposive sampling method used to establish the content validity patient work groups systematically excluded women who, in the opinion of the investigator, were not likely to be able to perform the 110  study activities. Therefore, input regarding the content validity of the MQOL-FflV was not elicited from women who were currently addicted to street drugs, were incarcerated, were physically unwell, had cognitive involvement because of their HIV disease or substance use, or resided in the Vancouver Downtown Eastside. To ensure comprehensive coverage of the content validity of the instrument, it was important to engage participants who were 'up to' the task. It should be noted that although the six participants may not have been representative of the general Oak Tree Clinic female population in terms of several characteristics, all six were very familiar with the population and many worked with marginalized HIV-infected women's groups. The participants were able to provide feedback not only from their own perspective but also as a proxy for the women with whom they worked. The 85 women who completed the questionnaires consisted of a convenience sample. This may have introduced sampling bias in that women who were irregular Clinic attenders and who could read not English were excluded from the study. As well, several women were unable to complete the questionnaires, for various reasons, and were dropped from the analysis. Because it is possible that the validity and reliability of the MQOL-HIV may differ between the included and excluded women, and between the women who were able to complete the questionnaires and those who were not, the generalizability of the study results to the entire Clinic population may be limited. Generalizability of the study results to the greater female HIV-infected population may also be limited because the study was conducted in a single location. It is likely that the female HIV-infected population in Vancouver differs from those in other Canadian centres and other countries. Factors such as mode of HIV acquisition, cultural and racial backgrounds, socioeconomic status, government system, and health-care system could all contribute to how women in this study responded to the tests. Therefore, generalization of the results of this study must be done with caution.  Ill  The pencil-and-paper mode of administration introduced the possibility of missing answers and misunderstanding of questions. In an effort to reduce the number of missing answers, the investigator remained available to the participants when they completed the scales and checked all completed scales for completeness. As well, she availed herself to answer questions should there have been any misunderstanding. As most patients at the OTC spent two to four hours at the clinic, often waiting long periods in the waiting-room, most had time to complete the questionnaire. However, some of the completed questionnaires contained missing data, and although this was dealt with methodologically, it cannot be determined how this may have influenced the study results. The test-retest reliability results may have been biased by the high attrition rate. Although the responders and non-responders did not differ in demographic or HIV/clinical characteristics (with the exception of more responders having children living at home), there is not enough evidence to dismiss the possibility that the instrument's stability may have differed between the two groups. It should be noted, however, that this high attrition rate could also be an explanation for why one half of the domains did not achieve the required correlation coefficient to establish acceptable test-retest reliability. Finally, there are some limitations to the statistical testing conducted. Regarding the use of Cronbach's alpha, according to Streiner and Norman (1995), it is recognized that some scales, when tapping a heterogenous concept, will result in low internal consistency. This was seen in this study and a study by Smith et al. ("Multidimensional Quality of Life," 1997) in the Sexual Functioning domain. In such a case, the internal consistency may be increased by eliminating items that are not highly correlated with each other, but according to Streiner and Norman, this may result in an index that taps only one aspect of the concept and results in very low content  112  validity. Under such circumstances, Streiner and Norman recommend sacrificing internal 5  consistency for content validity. Alternatively, the item could be included but not tallied in the subscale score. Regarding the CVI method for assessing content validity, Lynn (1986) identifies two limitations: the possibility of chance agreement for the CVI and the dependence of the CVI on the number of categories used in the rating. Regarding the first limitation, this was controlled for by providing a total of 13 expert panel members (i.e., six patients and seven health-care professionals). The second limitation was addressed by the choice of a four-option rating scale. According to Lynn, this is preferable because it does not include the ambivalent middle rating occurring in odd numbered rating scales. Conclusions Despite the study limitations, this work represents one of the first of its kind in the evaluation of the psychometric properties of an HIV-specific QOL instrument in HIV-infected women and provides preliminary evidence for the validity and reliability of the MQOL-HIV in this unique population. Based on the ratings of six HIV-infected women and seven HIV health-care professionals, most of the items and all of the 10 domains appear to be content valid in the measurement of QOL in women living with HIV. Revision or replacement of the items rated by the patients and professionals as not valid, as well as minor language revisions throughout the instrument, would greatly strengthen the content validity of the scale in this population. Addition of domains that measure QOL in terms of family/child health and spirituality, and the incorporation of such concepts as caregiver role, pregnancy, medication issues, body image and self esteem, stigma and disclosure, and physical environment would further strengthen the content validity of the overall scale.  5  This could also imply the existence and conflation of two dimensions, as opposed to one complex dimension.  113  The results of the concurrent validity testing are promising, in that the five domains of Cognitive Functioning, Physical Health, Physical Functioning, Mental Health, and Social Functioning and the overall instrument correlated strongly with similar domains in the more commonly used MOS-HIV. If the MOS-HIV is a valid instrument for measuring QOL in HIVinfected women, as the authors claim (Wu et al., 1991, 1997), then one may conclude that the five MQOL-HIV domains and the overall instrument are also valid. In terms of reliability, the MQOL-HIV demonstrated strong internal consistency in all but three of the domains, and moderate to strong test-retest reliability in most of the domains and the overall instrument. The internal consistency and test-retest reliability of the instrument would likely be improved if the content validity of the measure were strengthened. Overall, the findings support the validity and reliability of the current version of the MQOL-HIV in female HIV-infected populations. With the recommended revisions, the measure would likely be a psychometrically powerful tool for measuring the QOL of HIV-infected women. Implications for Nursing Practice, Research, and Theory The findings have implications for nursing practice, research, and theory. First, the overall findings of this study have highlighted the need for nurses and other health-care professionals to critically appraise a scale prior to initiating its use. Although there is published evidence supporting the validity and reliability of the MQOL-HIV, its psychometric properties differed when tested in a different population than that for which it was originally developed. This supports the recommendations made by Kline (1979), who states that any scale intended for use in a population for which it was not originally developed requires revalidation to allow proper interpretation of the results and for sound inferences to be made. The findings, particularly the content validity results, may also guide nurses and other health-care professionals when interpreting results from the MQOL-HIV in HIV-infected women 114  or mixed groups of men and women. For example, interpreting responses to the item regarding "feeling needed" may differ depending on whether the respondents are male or female. If a group of women score highly on this item, it should not be assumed that this indicates a positive level of mental health. Another implication of the findings is that, although some revision is required to make the MQOL-HIV a psychometrically stronger scale for HIV-infected women, the scale in its current form may be used to measure QOL in this population. The content and concurrent validity results suggest that, for the most part, the scale is measuring the domains of QOL that it purports to measure in HIV-infected women. As well, the majority of the domains demonstrated good internal consistency and test-retest reliability, suggesting that the instrument is moderately to strongly stable in this population. These findings will allow nurses and other health-care professionals to use the MQOL-HIV to measure QOL in HIV-infected women, and to interpret the findings in a meaningful way. Finally, the study has conceptual and theoretical implications. The findings support the essential attributes of the QOL construct found in the literature, particularly its multidimensional, subjective nature, and highlight the notion that QOL dimensions of salience may vary from one population to the next. Both the health-care professional and patient work group members agreed that there were multiple QOL dimensions relevant to HIV-infected women. As well, the differences in responses between the professionals and patients in terms of MQOL-HIV items that were not valid, together with identification of missing QOL domains, highlight the notion that QOL is best reported by the individual experiencing it. For example, although both groups agreed that the item referring to feeling needed was not valid, the reasons for their ratings were very different; whereas the professionals believed it was too vague, the patients believed that feeling needed had entirely different meanings for women and men. This example supports the assertion made by many authors that QOL is best evaluated by the person experiencing it, such 115  as Parse (1994), 'who states that QOL is "not what those outside the life looking in think it is, but rather it is what the person there living the life says it is" (p. 17). In terms of the salient dimensions of QOL, findings from the content validity testing indicate that some dimensions of QOL that are relevant to HIV-infected women may not be relevant to HIV-infected men, or non HIV-infected women. It was clear from the information obtained from the patient participants that HIV-infected women represent a unique population for whom unique dimensions of QOL exist. For example, within the missing domain of Family/Child Health, pregnancy introduces QOL issues that are particularly important to women living with HIV. An explanation for the inability of researchers to reach consensus on what dimensions comprise the QOL construct may be suggested by these findings; QOL may consist of myriad dimensions, but only certain ones become prominent depending on the individual, population, or circumstances under which it is being evaluated. These findings will allow nurses and other theoretical researchers to view the construct of QOL with greater understanding, within the context of specific populations. Recommendations for Future Research The findings of this study suggest areas for future research. The major research areas include strengthening and re-testing the scale's content validity, construct validity, and reliability, and improving the generalizability of the scale. The findings of the content validity testing indicate that parts of the MQOL-HIV require revision to make it more relevant to female HIV-infected populations. Based on the CVI scores, several of the items require significant revision or replacement. As well, two domains and several areas of QOL that are salient to women are not covered in the scale and should be added. The scale could be revised using a similar approach to that described by the developers (MQOL-HIV, 1997) and by Streiner and Norman (1995), including focus groups and key informant interviews. Focus groups and interviews with HIV-infected women could be 116  conducted to further explore how the identified domains and issues relate to the QOL of women living with HIV. Based on this information, revised items or new items that address the identified missing QOL issues can be drafted. Choosing the items for the instrument could be conducted using the methods described by the developers. For example, all sample items can be written on index cards, and a group of HIV-infected women can be asked to sort them into domains. This enables the selection of appropriate items and domains to leave in the instrument. The resulting 'draft' of the revised questionnaire could then be subjected to repeated psychometric evaluation, such as that conducted in the current study. In terms of the construct validity of the MQOL-HIV, it was noted that the concurrent validity testing provided only preliminary evidence for concurrent validity. To generate more comprehensive evidence, comparative testing of all domains is required, preferably using a battery of standardized instruments for the correlation analyses. As well, more advanced analyses could add to the evidence of the instrument's construct validity, such as factor analysis and multidimensional scaling, both of which require much larger samples than that obtained for the current study (Streiner & Norman, 1995). Finally, the generalizability of the scale could be improved in several ways. First, random sampling of all HIV-infected female clinic patients could reduce the threat of sampling bias by not systematically excluding those who do not attend the clinic regularly. This could introduce logistical challenges, however, that may preclude this method of sampling. For example, if a random sample were obtained from the Oak Tree Clinic database, it may be impossible to locate some of the women identified due to transience, lack of stable housing, or lack of telephone access. As well, the MQOL-HIV would have to be administered in several ways, including in-person interview (for those who can not read) and employing the use of interpreters.  117  Perhaps a more feasible method to improve generalizability would be to initiate a multicentre study, enrolling participants from a number of clinics across Canada, and possibly other countries. This would enable generalization to a more heterogeneous female HIV-infected population. This approach would also increase the sample size, enabling more complex, in-depth validity testing methods. A final method to help improve the generalizability of the scale would involve translating the MQOL-HIV into the most common languages spoken by the target population. In the context of the Canadian HIV-infected population, translation into French and some of the more common immigrant languages, such as Punjabi, Vietnamese, and Mandarin, may allow for a more inclusive target population. Questionnaire translation is a complex and time-consuming process, however, requiring an understanding of the unique methodological issues related to word meaning, colloquial phrases, jargon, and idiomatic expressions (Hilton & Skrutkowski, 2002). Mathias, Fifer, and Patrick (1994) describe a basic method that requires having a professional translator translate the scale into a different language, followed by a second, independent translator translating it back into English. The scale then requires language and equivalence testing and further psychometric testing within the target population. In conclusion, this study has provided preliminary evidence in support of the validity and reliability of a promising new HIV-specific QOL scale, the MQOL-HIV. It is the hope of the investigator that the results of this study will stimulate not only further research regarding validation of this and other HIV-specific QOL measures for HIV-infected women, but also further research regarding the uniqueness of the QOL construct within this population.  118  References Albrecht, H., Helm, E. B., Plettenberg, A., Ernminger, C , Heise, W., Schwartlander, B., & Stellbrink, H. J. (1994). Kaposi's sarcoma in HIV infected women in Germany: More evidence for sexual transmission. A report of 10 cases and review of the literature. Genitourinary Medicine, 70. 394-398. Allison, P. J., Locker, D., & Feine, J. S. (1997). Quality of life: A dynamic construct. Social Science & Medicine, 45, 221-230. Anderson, K. L., & Burckhardt, C. S. (1999). Conceptualization and measurement of quality of life as an outcome variable for health care intervention and research. Journal of Advanced Nursing, 29, 298-306. Apolone, G., & Mosconi, P. (1998). Review of the concept of quality of life assessment and discussion of the present trend in clinical research. Nephrology, Dialysis, Transplantation, 13 Suppl. 1. 65-69. Aristotle. (1963). Ethics (J. Warrington, Trans.). New York: Dutton. Aversa, S., Kimberlin, C , & Segal, R. (1998). The Medication Attribution Scale: Perceived effects of antiretrovirals and quality of life. Quality of Life Research, 7, 205-214. Badia, X., Podzamczer, D., Garcia, M., Lopez-Lavid, C. C , & Consiglio, E. (1999). A randomized study comparing instruments for measuring health-related quality of life in HIVinfected patients. Spanish MOS-HIV and MQOL-HIV Validation Group. Medical Outcomes Study HIV Health Survey. AIDS, 13, 1727-1735. Badia, X., Podzamczer, D., Casado, A., Lopez-Lavid, C , & Garcia, M. (2000). Evaluating changes in health status in HIV-infected patients: Medical Outcomes Study-HfV and Multidimensional Quality of Life-HIV quality of life questionnaires. Spanish MOS-HIV and MQOL-HIV Validation Group. AIDS, 14, 1439-1447.  119  Bedimo, A. L., Kissinger, P., & Bessinger, R. (1997). History of sexual abuse among HIV-infected women. International Journal of STD and AIDS, 8. 332-335. Berger, B. E., Ferrans, C. E., & Lashley, F. R. (2001). Measuring stigma in people with HIV: Psychometric assessment of the HIV stigma scale. Research in Nursing and Health, 24, 518-529. Berger, D. S. (2001). New facial filling treatment for lipodystrophy. Positive Awareness, 12, 17-20. Bergner, M., Bobbitt, R. A., Carter, W. B., & Gilson, B. S. (1981). The Sickness Impact Profile: Development and final revision of a health status measure. Medical Care, 19, 787-805. Berzon, R. A., Leplege, A. P., Lohr, K. N., Lenderking, W. R., & Wu, A. W. (1997). Summary and recommendations for future research. Quality of Life Research, 6, 601-605. Boisson, E.V., & Rodrigues, L. C. (2002). Factors associated with HJV infection are not the same for all women. Journal of Epidemiology and Community Health, 56, 103-8. Bowling, A. (1995). What things are important in people's lives? A survey of the public's judgements to inform scales of health related quality of life. Social Science & Medicine, 41, 1447-1462. Bozzette, S. A., Hays, R. D., Berry, S. H., & Kanouse, D. E. (1994). A perceived health index for use in persons with advanced HJV disease: Derivation, reliability, and validity. Medical Care. 32,716-731. Brady, M. J., Peterman, A. H., Fitchett, G., Mo, M., & Cella, D. (1999). A case for including spirituality in quality of life measurement in oncology. Psychooncology, 8, 417-428. Brazier, J. E., Harper, R., Jones, N. M., O'Cathain, A., Thomas, K. J., Usherwood, T., & Westlake, L. (1992). Validating the SF-36 Health Survey Questionnaire: New outcome measure for primary care. British Medical Journal, 305, 160-164.  120  British Columbia Centre for Disease Control Society (BCCDC). (2000). fflV/AIDS update: Annual 2000. Vancouver, BC: British Columbia Centre for Disease Control, STD/AIDS Control. Brock, D. (1996). Quality of life measures in health care and medical ethics. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2nd ed.; pp. 497-510). Philadelphia: Lippincott-Raven. Bunting, S. M., Bevier, D. J., & Baker, S. K. (1999). Poor women living with HIV: Selfidentified needs. Journal of Community Health Nursing, 16, 41-52. Bunting, S. M. (2001). Sustaining the relationship: Women's caregiving in the context of HJV disease. Health Care for Women International, 22, 131-148. Burdge, D. (2000). Women and Family HIV Centre (Oak Tree Clinic) Program quality assurance annual report. April 1/99 - March 31/00. Unpublished manuscript. Vancouver, BC. Burdge, D. R., & Forbes, J. (1994). The Oak Tree Centre: Initiatives for HIV-infected women, children, youth and families. Leadership in Health Services, 3, 9-13. Burgess, A., Dayer, M., Catalan, J., Hawkins, D., & Gazzard, B. (1993). The reliability and validity of two HIV-specific health-related quality-of-life measures: A preliminary analysis. AIDS, 7, 1001-1008. Caiman, K. C. (1987). Definitions and dimensions of quality of life. In N. K. Aaronson & J. Beckmann (Eds.), The quality of life of cancer patients (pp. 1-9). New York: Raven Press. Campbell, A., Converse, P. E., & Rodgers, W. L. (1976). The quality of American life. New York: Russell Sage Foundation. Carpenter, C. C , Cooper, D. A., Fischl, M. A., Gatell, J. M., Gazzard, B. G., Hammer, S. M., Hirsch, M. S., Jacobsen, D. M., Katzenstein, D. A., Montaner, J. S., Richman, D. D., Saag, M. S., Schechter, M., Schooley, R. T., Thompson, M. A., Vella, S., Yeni, P. G., & Volberding, P.  121  A. (2000). Antiretroviral therapy in adults: Updated recommendations of the International AIDS Society-USA Panel. Journal of the American Medical Association, 283, 381-390. Cederfjall, C , Langius-Eklof, A., Lidman, K., & Wredling, R. (2001). Gender differences in perceived health-related quality of life among patients with FflV infection. AIDS Patient Care and STDs. 15. 31-39. Cella, D. F., McCain, N. L., Peterman, A. H., Mo, F., & Wolen, D. (1996). Development and validation of the Functional Assessment of Human Immunodeficiency Virus Infection (FAHI) Quality of Life Instrument. Quality of Life Research, 5. 450-463. Cella, D. F., Wilkund, I., Shumaker, S. A., & Aaronson, N. K. (1995). Integrating healthrelated quality of life into cross-national clinical trials. In S. A. Shumaker & R. Berzon (Eds.) The International Assessment of Health-Related Quality of Life: Theory, Translation, Measurement and Analysis (pp. 75-82). New York: Rapid Communications. Center for Disease Control. (1993). AIDS definition update by CDC. American Family Physician, 47, 265. Clark, A. L., Affonson, D. D., (with Harris, T. R.) (Eds.) (1979). Childbearing: a nursing perspective (2 ed.). Philadelphia, PA: F. A. Davis company. nd  Cleary, P. D., Fowler, F. J., Jr., Weissman, J., Massagli, M. P., Wilson, I., Seage, G. R., Ill, Gatsonis, C , & Epstein, A. (1993). Health-related quality of life in persons with acquired immune deficiency syndrome. Medical Care, 31, 569-580. Cohen, C , Revicki, D. A., Nabulsi, A., Sarocco, P. W., & Jiang, P. (1998). A randomized trial of the effect of ritonavir in maintaining quality of life in advanced HIV disease. Advanced HIV Disease Ritonavir Study Group. AIDS, 12, 1495-1502. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2 ed.). nd  Hillsdale, NJ: L. Erlbaum.  122  Cohen, S. R., Hassan, S. A., Lapointe, B. J., & Mount, B. M. (1996). Quality of life in HIV disease as measured by the McGill Quality of Life Questionnaire. AIDS, 10, 1421-1427. Collins, E., Wagner, C , & Walmsley, S. (2000). Psychosocial impact of the lipodystrophy syndrome in HIV infection. AIDS Reader, 10, 546-550. Connelly, J. E., Wofford, A. B., & Philbrick, J. T. (2000). Healthy patients who perceive poor health: Why are they worried sick? American Journal of Medical Science, 320, 36-42. Cook, D. J., Guyatt, G. H., Juniper, E., Griffith, L., Mcllroy, W., Willan, A., Jaeschke, R., & Epstein, R. (1993). Interviewer versus self-administered questionnaires in developing a disease-specific, health-related quality of life instrument for asthma. Journal of Clinical Epidemiology, 46, 529-534. Coscia, J. M., Christensen, B. K., Henry, R. R., Wallston, K., Radcliffe, J., & Rutstein, R. (2001). Effects of home environment, socioeconomic status, and health status on cognitive functioning in children with HIV-1 infection. Journal of Pediatric Psychology, 26, 321-329. Cunningham, W. E., Shapiro, M. F., Hays, R. D., Dixon, W. J., Visscher, B. R., George, W. L., Ettl, M. K., & Beck, C. K. (1998). Constitutional symptoms and health-related quality of life in patients with symptomatic HIV disease. American Journal of Medicine, 104, 129-136. de Boer, J. B., van Dam, F. S., & Sprangers, M. A. (1995). Health-related quality-of-life evaluation in HIV-infected patients. A review of the literature. Pharmacoeconomics, 8, 291-304. de Vries, J., van Heck, G. L., & Drent, M. (1999). Gender differences in sarcoidosis: Symptoms, quality of life, and medical consumption. Women & Health, 30, 99-114. Dibble, S. L., Padilla, G. V., Dodd, M. J., & Miaskowski, C. (1998). Gender differences in the dimensions of quality of life. Oncology Nursing Forum, 25, 577-583. Farquhar, M. (1995). Definitions of quality of life: A taxonomy. Journal of Advanced Nursing, 22, 502-508.  123  Fava, G. A. (1990). Methodological and conceptual issues in research on quality of life. Psychotherapy & Psvchosomatics, 54, 70-76. Ferrans, C. E. (1990a). Quality of life: Conceptual issues. Seminars in Oncology Nursing, 6, 248-254. Ferrans, C. E. (1990b). Development of a quality of life index for patients with cancer. Oncology Nursing Forum, 17, 15-19. Ferrans, C. E. (1992). Conceptualizations of quality of life in cardiovascular research. Progress in Cardiovascular Nursing, 7, 2-6. Ferrans, C. E. (1996). Development of a conceptual model of quality of life. Scholarly Inquiry for Nursing Practice, 10, 293-304. Ferrans, C. E., & Powers, M. J. (1985). Quality of Life Index: Development and psychometric properties. Advances in Nursing Science, 8(1), 15-24. Ferrans, C. E. & Powers, M. J. (1992). Psychometric assessment of the Quality of Life Index. Research in Nursing and Health, 15, 29-38. Forsyth, B. W., Davis, J. A., Freudigman, K. A., Katz, K. H., & Zelterman, D. (2002). Pregnancy and birth rates among HIV-infected women in the United States: The confounding effects of illicit drug use. AIDS. 16. 471-479. Franchi, D., & Wenzel, R. P. (1998). Measuring health-related quality of life among patients infected with human immunodeficiency virus. Clinical Infectious Diseases, 26, 20-26. Freeman, R. C , Parillo, K. M., Collier, K., & Rusek, R. W. (2001). Child and adolescent sexual abuse history in a sample of 1,490 women sexual partners of injection drug-using men. Women & Health, 34, 31-49. Fryback, P. B. (1993). Health for people with a terminal diagnosis. Nursing Science Quarterly, 6, 147-159.  124  Ganz, P. A., Coscarelli Schag, C. A., Kahn, B., Petersen, L., & Hirji, K. (1993). Describing the health-related quality of life impact of HIV infection: Findings from a study using the HIV Overview of Problems-Evaluation System (HOPES). Quality of Life Research. 2, 109119. Garcia-Moreno, C., & Watts, C. (2000). Violence against women: Its importance for HIV/AIDS. AIDS. 14, S253-S265. Garg, A. (2000). Lipodystrophies. American Journal of Medicine, 108, 143-152. Gaskins, S., & Brown, K. (1992). Psychosocial response among individuals with human immunodeficiency virus infection. Applied Nursing Research, 4, 111-121. Gielin, A. C , Fogarty, L., O'Campo, P., Anderson, J., Keller, J., & Faden, R. (2000). Women living with HIV: Disclosure, violence, and social support. Journal of Urban Health, 77, 480-491. Gielen, A. C, O'Campo, P., Faden, R. R., & Eke, A., (1997). Women's disclosure of HIV status: Experiences of mistreatment and violence in an urban setting. Women & Health, 25, 19-31. Gill, T. M., & Feinstein, A. R. (1994). A critical appraisal of the quality of quality-of-life measurements. Journal of the American Medical Association, 272, 619-626. Gioiella, M. E., Berkman, B., & Robinson, M. (1998). Spirituality and quality of life in gynecologic oncology patients. Cancer Practitioner, 6, 333-338. Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory: strategies of qualitative research. Chicago, IL: Aldine. Grabowski, H. G., & Hansen, R. W. (1996). Economic scales and tests. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2 ed.; pp. 79-106). nd  Philadelphia: Lippincott-Raven.  125  Guarnaccia, P. J. (1996). Anthropological perspectives: The importance of culture in the assessment of quality of life. In B. Spilker (Ed.), Quality of Life and Pharmacoeconomics in Clinical Trials (2 ed., pp. 523-528). Philadelphia: Lippincott-Raven. nd  Guyatt, G. H., Feeny, D. H., & Patrick, D. L. (1993). Measuring health-related quality of life. Annals of Internal Medicine. 118, 622-629. Guyatt, G. FL, Jaeschke, R., Feeny, D. FL, & Patrick, D. L. (1996). Measurements in clinical trials: Choosing the right approach. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical yrials (2 ed.; pp. 41-48). Philadelphia: Lippincott-Raven. nd  Haas, B. K. (1999a). A multidisciplinary concept analysis of quality of life. Western Journal of Nursing Research. 21. 728-742. Haas, B. K. (1999b). Clarification and integration of similar quality of life concepts. Image: Journal of Nursing Scholarship, 31, 215-220. Haberman, M.R., & Bush, N. (1998). Quality of life methodological and measurement issues. In C. R. King & P. S. Hinds (Eds.), Quality of life from nursing and patient perspectives - theory, research, practice (pp. 117-139). Toronto: Jones and Bartlett. Harrison, M. B., Juniper, E. F., & Mitchell-DiCenso, A. (1996). Quality of life as an outcome measure in nursing research. "May you have a long and healthy life." Canadian Journal of Nursing Research, 28, 49-68. Hilton, A., & Skrutkowski, M. (2002). Translating instruments into other languages: Development and testing processes. Cancer Nursing, 25, 1-7. Hjermstad, M. J., Fayers, P. M., Bjordal, K., & Kaasa, S. (1998). Using reference data on quality of life—the importance of adjusting for age and gender, exemplified by the EORTC QLQC30 (+3). European Journal of Cancer, 34, 1381-1389.  126  Holmes, W. C , & Shea, J. A. (1997). Performance of a new, HIV/AIDS-targeted quality of life (HAT-QoL) instrument in asymptomatic seropositive individuals. Quality of Life Research, 6, 561-571. Holmes, W. C , & Shea, J. A. (1998). A new HIV/AEDS-targeted quality of life (HATQoL) instrument: development, reliability, and validity. Medical Care, 36, 138-154. Holzemer, W. L., & Wilson, H. S. (1995). Quality of life and the spectrum of HIV infection. Annual Review of Nursing Research, 13, 3-29. Holzemer, W. L., Spicer, J., Skodol, W. H., Kemppainen, J. K., & Coleman, C. (1998). Validation of the Quality of Life Scale: Living with HIV. Journal of Advanced Nursing, 28, 622630. Hubley, A.M., & Zumbo, B.D. (1996). A dialectic on validity: Where we have been and where we are going. Journal of General Psychology, 123, 207-215. Hunt, S. M., McKenna, S. P., McEwen, J., Backett, E. M., Williams, J., & Papp, E. (1980). A quantitative approach to perceived health status: A validation study. Journal of Epidemiology and Community Health, 34, 281-286. Jadack, R. A., Hyde, J. S., & Keller, M. L. (1995). Gender and knowledge about HIV, risky sexual behavior, and safer sex practices. Research in Nursing and Health, 18, 313-324. Jenkinson, C , Coulter, A., & Wright, L. (1993). Short Form 36 (SF36) Health Survey Questionnaire: Normative data for adults of working age. British Medical Journal, 306, 14371440. Johnson, T. M. (1996). Cultural considerations. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2 ed.; pp. 511-516). Philadelphia: Lippincott-Raven. nd  Juniper, E.F., Guyatt, G.H., & Jaeschke, R. (1996). How to develop and validate a new health-related quality of life instrument. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2 ed.; pp. 41-48). Philadelphia, PA: Lippincott-Raven. nd  127  Kagawa-Singer, M. (1993). Redefining health: Living with cancer. Social Science & Medicine. 37, 295-304. King, C. R. (1998). Overview of quality of life and controversial issues. In C. R. King & P. S. Hinds (Eds.), Quality of life from nursing and patient perspectives - theory, research, practice (pp. 23-34). Toronto: Jones and Bartlett. Kinney, M. R., Burfitt, S. N., Stullenbarger, E., Rees, B., & DeBolt, M. R. (1996). Quality of life in cardiac patient research: A meta-analysis. Nursing Research, 45, 173-180. Kirkham C. M. & Lobb, D. J. (1998). The British Columbia Positive Women's Survey: A detailed profile of 110 HIV-infected women. Canadian Medical Association Journal, 158, 317-323. Kline, P. (1979). Psychometrics and psychology. London: Academic Press. Landgraf, J. M., Maunsell, E., Speechley, K. N., Bullinger, M., Campbell, S., Abetz, L., & Ware, J. E. (1998). Canadian-French, German and UK versions of the Child Health Questionnaire: Methodology and preliminary item scaling results. Quality of Life Research, 7, 433-445. Lenderking, W. R., Testa, M. A., Katzenstein, D., & Hammer, S. (1997). Measuring quality of life in early HIV disease: The modular approach. Quality of Life Research, 6, 515-530. Leplege, A. P., & Berzon, R. A. (1997). Introduction. Quality of Life Research, 6, 469470. Leplege, A., Rude, N., Ecosse, E., Ceinos, R., Dohin, E., & Pouchot, J. (1997). Measuring quality of life from the point of view of HIV-positive patients: The HTV-QL31. Quality of Life Research, 6, 585-594. Li, T. S., Tubiana, R., Katlama, C , Calvez, V., Ait, M. H., & Autran, B. (1998). Longlasting recovery in CD4 T-cell function and viral-load reduction after highly active antiretroviral therapy in advanced HIV-1 disease. Lancet, 351, 1682-1686. 128  Lichtenstein, B., Laska, M. K., & Clair, J. M. (2002). Chronic sorrow in the HIVpositive patient: Issues of race, gender, and social support. AIDS Patient Care, 16, 27-38. Linzer, M., Spitzer, R., Kroenke, K., Williams, J. B., Hahn, S., Brody, D., & deGruy, F. (1996). Gender, quality of life, and mental disorders in primary care: Results from the PRIMEMD 1000 study. American Journal of Medicine, 101, 526-533. Lubeck, D. P., & Fries, J. F. (1997). Assessment of quality of life in early stage HIVinfected persons: Data from the AIDS Time-Oriented Health Outcome Study (ATHOS). Quality of Life Research, 6, 494-506. Lynn, M. R. (1986). Determination and quantification of content validity. Nursing Research, 35, 382-385. MacKenzie, C. R., Charlson, M. E., DiGioia, D, & Kelley, K. (1986). Can the sickness impact profile measure change? An example of scale assessment. Journal of Chronic Diseases, 39,429-438. Marschner, I. C , Collier, A. C , Coombs, R. W., D'Aquila, R. T., DeGruttola, V., Fischl, M. A., Hammer, S. M., Hughes, M. D., Johnson, V. A., Katzenstein, D. A., Richman, D. D., Smeaton, L. M., Spector, S. A., & Saag, M. S. (1998). Use of changes in plasma levels of human immunodeficiency virus type 1 RNA to assess the clinical benefit of antiretroviral therapy. Journal of Infectious Diseases, 177, 40-47. Mathias, S. D., Fifer, S. K., & Patrick, D. L. (1994). Rapid translation of quality of life measures for international clinical trials: Avoiding errors in the minimalist approach. Quality of Life Research, 3, 403-412. McCall, S. (1975). Quality of life. Social Indicators Research, 2, 229-248. McClure, J. B., Catz, S. L, Prejean, J., Brantley, P. J., & Jones, G. N. (1996). Factors associated with depression in a heterogeneous HIV-infected sample. Journal of Psychosomatic Research, 40, 407-415. 129  McDonnell, K. A., Gielen, A. C , Wu, A. W., O'Campo, P., & Faden, R. (2000). Measuring health related quality of life among women living with HIV. Quality of Life Research. 9, 931-940. Meeberg, G. A. (1993). Quality of life: A concept analysis. Journal of Advanced Nursing, 18,32-38. Melnick, S. L., Wertheimer, W. J., & Pinn, V. W. (1995). Women and HIV/AIDS research in the United States. Journal of the American Medical Women's Association, 50, 137138. Metcalfe K. A., Langstaff, J. E., Evans, S. J., Paterson, H. M., & Reid, J. L. (1998). Meeting the needs of women living with HIV. Public Health Nursing, 15, 30-34. Moneyham, L., Seals, B., Demi, A., Sowell, R., Cohen, L., & Guillroy, J. (1996). Experiences of disclosure in women infected with HIV. Health Care for Women International, i L 209-221. Multidimensional Quality of Life Questionnaire for Persons with HIV/AJDS (MQOLHIV) administration and scoring manual. Watertown, MA: New England Research Institute (NERI), 1997. Naughton, J., Shumaker, S. A., Anderson, R. T., & Czajkowski, S. M. (1996). Psychological aspects of HRQOL measurement: Tests and scales. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2nd ed.; pp. 117-132). Philadelphia: LippincottRaven. Nieuwkerk, P. T., Gisolf, E. H., Colebunders, R., Wu, A. W., Danner, S. A., & Sprangers, M. A. (2000). Quality of life in asymptomatic and symptomatic HIV infected patients in a trial ofritonavir/saquinavirtherapy. The Prometheus Study Group. AIDS, 14, 181-187. Norusis, M. J. (1988). The SPSS guide to data analysis for SPSS/PC+. Chicago, IL: SPSS. 130  Nussbaum, M., & Sen, A. (1993). The qualify of life. Oxford: Clarendon. O'Brien, M. E., & Pheifer, W. G. (1993). Physical and psychosocial nursing care for patients with HIV infection. Nursing Clinics of North America, 28, 303-316. Oleson, M. (1990). Subjectively perceived quality of life. Image: Journal of Nursing Scholarship, 22, 187-190. Osoba, D. (1994). Lessons learned from measuring health-related quality of life in oncology. Journal of Clinical Oncology, 12, 608-616. Osowiecki, D. M., Cohen, R. A., Morrow, K. M., Paul, R. H., Carpenter, C. C , Flanigan, T., & Boland, R. J. (2000). Neurocognitive and psychological contributions to quality of life in HJV-1-infected women. AIDS, 14, 1327-1332. Palella, F. J., Jr., Delaney, K. M., Moorman, A. C , Loveless, M. O, Fuhrer, J., Satten, G. A., Aschman, D. J., & Holmberg, S. D. (1998). Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HJV Outpatient Study Investigators. New England Journal of Medicine, 338, 853-860. Parse, R. R. (1994). Quality of life: Sciencing and living the art of human becoming. Nursing Science Quarterly, 7, 16-21. Piette, J., Wachtel, T. J., Mor, V., & Mayer, K. (1995). The impact of age on the quality of life in persons with HIV infection. Journal of Aging and Health, 7, 163-178. Ragsdale, D., & Morrow, J. R. (1990). Quality of life as a function of HIV classification. Nursing Research, 39, 355-359. Revicki, D. A., Wu, A., & Brown, R. (1995). Change in clinical status, health status and health utility outcomes in HIV-infected patients. Medical Care, 33, AS173-AS182. Riedinger, M. S., Dracup, K. A., Brecht, M. L., Padilla, G., Sarna, L., & Ganz, P. A. (2001). Quality of life in patients with heart failure: Do gender differences exist? Heart and Lung. 30, 105-116. 131  Rosenberg, E. S., Billingsley, J. M., Caliendo, A. M., Boswell, S. L., Sax, P. E., Kalams, S. A., & Walker, B. D. (1997). Vigorous FflV-1-specific CD4+ T cell responses associated with control of viremia. Science, 278, 1447-1450. Sackett, D. L. (1979). Bias in analytic research. Journal of Chronic Diseases, 32, 51-63. Sarna, L., van Servellen, G., Padilla, G., & Brecht, M. L. (1999). Quality of life in women with symptomatic HIV/AEDS. Journal of Advanced Nursing, 30, 597-605. Sartorius, N. (1987). Cross-cultural comparisons of data about quality of life: A sample of issues. In N. K. Aaronson & J. Beckmann (Eds.), The quality of life of cancer patients (pp. 19-24). New York: Raven. Schag, C. A., Ganz, P. A., & Heinrich, R. L. (1991). Cancer Rehabilitation Evaluation System-Short Form (CARES-SF). A cancer specific rehabilitation and quality of life instrument. Cancer, 68, 1406-1413. Schag, C. A., Ganz, P. A., Kahn, B., & Petersen, L. (1992). Assessing the needs and quality of life of patients with HIV infection: Development of the HIV Overview of ProblemsEvaluation System (HOPES). Quality of Life Research, 1, 397-413. Schipper, H , Clinch, J. J., & Olweny, C. L. M. (1996). Quality of life studies: Definitions and conceptual issues. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2nd ed.; pp. 11-24). Philadelphia: Lippincott-Raven. Schipper, H., Clinch, J., & Powell, V. (1990). Definitions and conceptual issues. In B. Spilker (Ed.), Quality of life assessments in clinical trials (pp. 11-24). New York: Raven Press. Seals, B. (1996). The overlapping epidemics of violence and HIV. The Journal of the Association of Nurses in AIDS Care, 7, 91-93. Semple, S. J., Patterson, T. L., Straits-Troster, K., Atkinson, J. H., McCutchan, J. A., & Grant, I. (1996). Social and psychological characteristics of HIV-infected women and gay men. HIV Neurobehavioral Research Center (HNRC) Group. Women & Health. 24, 17-41. 132  Shumaker, S. A., & Berzon, R. (Eds) (1995). The International Assessment of HealthRelated Quality of Life: Theory. Translation, Measurement & Analysis. Oxford, NY: Rapid Communications. Shumaker, S. A., Ellis, S., & Naughton, M. (1997). Assessing health-related quality of life in HIV disease: Key measurement issues. Quality of Life Research, 6, 475-480. Smith, K. W., Avis, N. E., Mayer, K. H., & Swislow, L. (1997). Use of the MQoL-HlV with asymptomatic HIV-positive patients. Quality of Life Research, 6, 555-560. Songwathana, P. (2001). Women and ADDS caregiving: Women's work? Health Care for Women International. 22. 263-279. Sowell, R. L., Lowenstein, A., Moneyham, L., Demi, A., Mizino, Y., & Seals, B. (1991). Resources, stigma, and patterns of disclosure in rural women with HIV infections. Public Health Nursing, 14. 302-312. Sowell, R. L., Seals, B. F., Moneyham, L., Demi, A., Cohen, L., & Brake, S. (1997). Quality of life in HIV-infected women in the south-eastern United States. AIDS Care, 9, 501512. Spiegelhalter, D. J., Gore, S. M., Fitzpatrick, R., Fletcher, A. E., Jones, D. R., & Cox, D. R. (1992). Quality of life measures in health care. HI: Resource allocation. British Medical Journal. 305. 1205-1209. Spilker, B (Ed). (1996). Quality of life and pharacoeconomics in clinical trials (2nd ed.). NY: Lippincott-Raven. Stewart, A. L. (1992). Conceptual and methodologic issues in defining quality of life: State of the art. Progress in Cardiovascular Nursing, 7, 3-11. Stewart, A. L., Hays, R. D., & Ware, J. E., Jr. (1988). The MOS Short-Form General Health Survey. Reliability and validity in a patient population. Medical Care, 26, 724-735.  133  Streiner, D.L., & Norman, G.R. (1995). Health measurement scales: A practical guide to their development and use (2 ed.). New York: Oxford University Press. nd  Tate, H., & George, R. (2001). The effect of weight loss on body image in HIV-positive gay men. AIDS Care, 13, 163-169. Thorne, S. E. & Jillings, C. R. (1996). Quality of life science: Implications for cancer nursing. Canadian Oncology Nursing Journal, 6, 72-74. Tsasis, P. (2000). Health-related quality-of-life measurements in HIV/AIDS care. AIDS Patient Care and STDs. 14, 427-438. Vallerand, A.H., Breckenridge, CM., & Hodgson, N.A. (1998). Theories and conceptual models to guide quality of life related research. In C. R. King & P. S. Hinds (Eds.), Quality of life from nursing and patient perspectives - theory, research, practice (pp. 37-53). Toronto: Jones and Bartlett. van Servellen, G., Sarna, L., & Jablonski, K. J. (1998). Women with HIV: Living with symptoms. Western Journal of Nursing Research, 20, 448-464. Wachtel, T., Piette, J., Mor, V., Stein, M., Fleishman, J., & Carpenter, C. (1992). Quality of life in persons with human immunodeficiency virus infection: Measurement by the Medical Outcomes Study instrument. Annals of Internal Medicine, 116, 129-137. Wagner, G. J., & Rabkin, J. G. (1999). Development of the Impact of Weight Loss Scale (IWLS): A psychometric study in a sample of men with HIV/AIDS. AIDS Care, 11. 453-457. Ware, J. E., Jr., Brook, R. H., Davies, A. R., & Lohr, K. N. (1981). Choosing measures of health status for individuals in general populations. American Journal of Public Health, 71, 620625. Weinberger, M., Oddone, E. Z., Samsa, G. P., & Landsman, P. B. (1996). Are healthrelated quality-of-life measures affected by the mode of administration? Journal of Clinical Epidemiology. 49. 135-140. 134  Wiggins, J. D. (1979). Childbearing. Physiology, experiences, needs. Toronto, ON: The C. V. Mosby Company. Wiklund, I., Dimenas, E., & Wahl, M. (1990). Factors of importance when evaluating quality of life in clinical trials. Controlled Clinical Trials, 11, 169-179. Williams, A. B. (1997). New horizons: Antiretroviral therapy in 1997. The Journal of the Association of Nurses in AIDS Care. 8, 26-38. Wilson, I. B. & Cleary, P. D. (1995). Linking clinical variables with health-related quality of life. A conceptual model of patient outcomes. Journal of the American Medical Association, 273, 59-65. Wood-Dauphinee, S. (1999). Assessing quality of life in clinical research: From where have we come and where are we going? Journal of Clinical Epidemiology, 52, 355-363. Wu, A. W. (1996). MOS-HIV Health Survey: Scoring guidelines. Boston, MA: Medical Outcomes Trust. Wu, A. W., & Cagney, K. A. (1996). The role of quality of life assessments in medical practice. In B. Spilker (Ed.), Quality of life and pharmacoeconomics in clinical trials (2nd ed.; pp. 517-522). Philadelphia: Lippincott-Raven. Wu, A. W., Revicki, D. A., Jacobson, D., & Malitz, F. E. (1997). Evidence for reliability, validity and usefulness of the Medical Outcomes Study HIV Health Survey (MOS-HIV). Quality of Life Research, 6. 481-493. Wu, A. W., Rubin, H. R., Mathews, W. C , Ware, J. E., Jr., Brysk, L. T., Hardy, W. D., Bozzette, S. A., Spector, S. A., & Richman, D. D. (1991). A health status questionnaire using 30 items from the Medical Outcomes Study. Preliminary validation in persons with early HJV infection. Medical Care, 29, 786-798.  135  00  o  c  c  CD  e  <  < C <U CH CH  Apalone & Mosconi (1998) Berzon, Leplege, Lohr, Lenderking, & Wu(1997) Fava (1990) Ferrans (1996) Haas (1999b) Harrison et al. (1996) Molin & Arrigo (1995) Smith et al. (1997) Spilker (1996) Wiklund et al. (1990) Wu et al. (1991) Osoba (1994)  CH  Mental/ Cognitive  o  Psychological  <D  Energy/ Vitality  •S 13 c _o *+-»  Physical function  Social function  <D  X  Role function  "S >  X  Physical or health  Economic  3 O  HIVspecific  <  Author  Family  X  Spiritual  o .£ Symptoms/ Side effects  Sexual function  X  X  X  X X  X  X  X'  X  X X  X X  X  X  X X  X  X  X  X X  X X  X  X  X  X  X X  X X  03  VH  X  X  X  X  X X  X X  X X  X X  X  X  X  X  X  X  X  X  X  X  X  cn  Appendix B: Socio-Demographic Questionnaire Participant # This questionnaire contains some questions about your background, education, job, family, lifestyle, and health. Some of these questions are very personal and may make you uncomfortable. We want to emphasize that all of the information you provide on this questionnaire is completely confidential, and will not be linked with any information that identifies you or with your medical record. Please feel free to ask the investigator if there are any questions you don't understand.  Today's date: How old are you?  (years)  Where were you born (check one)? Canada USA Caribbean Central/South America  Africa Asia Australia/Pacific Islands other - (specify)  Western Europe Eastern Europe If you were born outside of Canada, what year did you arrive in Canada? How would you describe your background (check one)? Caucasian Black South Asian Asian Hispanic First Nations/aboriginal Other - (specify) How would you describe your religious preference (check one)? Christian - Catholic Buddhist Christian - Protestant Mormon Muslim Jehovah's Witness Hindu Agnostic  Atheist Jewish Sikh Other (describe)  137  What was the highest schooling you COMPLETED (check one)? elementary school college/university: years completed high school: last grade completed What is the combined, total yearly income (before taxes) of everyone in your household one)? less than $10,000 $30,000 - 50,000 $10,000-29,999 more than $50,000 What is your current job status (check all that apply)? Working fulltime (occupation) Working part-time (occupation) Unemployed Disability insurance Welfare recipient Other: How would you describe your current marital status (check all that apply)? Single Separated Other (explain) Married Divorced  Widowed Living with partner > 1 year  If you have a spouse/partner, is he or she HIV infected? Yes No Don't know Do you have children? Yes - how many? No (skip next two questions) If you have children, are any of them HIV infected? Yes No Don't know If you have children, where do they live (check all that apply)? With me With their father (not with me) With other family members Foster care  Adopted Other (describe) W h e n d i d y o u r e c e i v e y o u r f i r s t p o s i t i v e H I V test ( y e a r ) ? W h e n do y o u think y o u b e c a m e H I V infected (year)?  H o w d i d you become infected with H I V (check one)? Sexual contact Sharing dirty needles w h e n using drugs Blood/blood product transfusion Other (explain) I don't know In the last 6 m o n t h s , h a v e y o u u s e d a n y street d r u g s ? No Yes  -*• i f " y e s , " h o w d i d y o u t a k e t h e d r u g s ( c h e c k a l l t h a t a p p l y ) ? Injected Snorted Swallowed Smoked  In the last 6 m o n t h s , h a v e y o u been i n a d e t o x i f i c a t i o n / r e h a b i l i t a t i o n p r o g r a m m e f o r d r u g o r alcohol addiction? Yes No I n the past 2 w e e k s , h a v e y o u e x p e r i e n c e d a n y o f the f o l l o w i n g s y m p t o m s ( c h e c k a l l that a p p l y ) ? Nausea/vomiting  P o o r appetite  Diarrhea  Abdominal pain  Fatigue  Headaches  Fever  General body pain  Shaking chills  N u m b n e s s o r t i n g l i n g i n f i n g e r s o r toes  N i g h t sweats  Depression  S k i n rashes/itchiness  Problems seeing  M o u t h sores  Problems breathing  Dry cough  Thank you for taking the time to complete Your participation  this  in this study is greatly  questionnaire! appreciated 139  Appendix C: Rating Form for Health-Care Professional Members PART I: Item validity Please rate each of the 40 questions of the MQOL-HIV as to its relevance to the quality of life of HIV infected women using the following ordinal scale: 1 2 Not relevant  3 4 Unable to assess Relevant but needs minor Very relevant and relevance without item revision succinct revision - OR - item is in need of such revision that it would no longer be relevant  Please indicate your numerical response corresponding with the MQOL-HTV item numbers below: Item number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  Rating (1-4)  Item number 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40  Rating (1-4)  For each item you rated as '2' or '3,' please indicate in the space provided below and on the following page your recommended revisions:  140  (item revisions con't)  141  Part II: Dimension Validity The MQOL-HIV purports to assess 10 dimensions of quality of life, each dimension being represented by four items in the questionnaire as shown below: DIMENSION Mental health  Physical health  Physical functioning  Social functioning  Social support  Cognitive functioning  Financial status  Partner intimacy  Sexual functioning  Medical care  ITEMS ("How much of the time in the past 2 weeks ...") You felt depressed You felt anxious You felt needed You worried about things You had health problems You were too tired to do things you wanted to do You were bothered by nausea You were bothered by tingling sensations You were able to do things around the house You were able to get from place to place, such as shopping and to the doctor You were able to climb several flights of stairs without help You could perform strenuous sports like running or weight lifting You felt isolated You have withdrawn from socializing You lacked energy to socialize with friends You spent quality time with friends You have received enough emotional support from people close to you You had someone who could help you in an emergency You had someone you could talk to about problems You had someone who could give you useful advise You were bothered by trouble remembering things You were bothered by forgetting what you started to do You have found that your mind wanders more than usual You were bothered by a short attention span You had to put off paying regular living expenses You have not had enough money to pay for medication You have not had enough money to pay for recreational activities You were concerned about your financial future You felt satisfied with the activities you do with your spouse or partner You felt satisfied with the amount of affection your spouse of partner expressed towards you You were able to confide in your spouse or partner You felt mistrust for your spouse or partner You felt satisfied with your sexual activity You have wanted sex more often than you have had sex You felt afraid of infecting someone through sexual contact You felt afraid of sexual rejection you wanted more information from your health care providers you wanted more support from your doctors and other health providers you doubted that your health providers are doing all they can for your health you were able to get the medical care you needed  142  Please rate each of the 10 dimensions of the MQOL-FflV as to its relevance to the quality of life of HIV infected women using the following ordinal scale: 1 Not relevant  Unable to assess relevance without revision of items OR - items within dimension are in need of such revision that it would no longer be relevant  Relevant but needs minor Very relevant and succinct revision of some of the items  Please indicate your numerical response corresponding with the MQOL-HIV dimensions: Dimension  Rating (1-4)  Mental health Physical health Physical functioning Social functioning Social support Cognitive functioning Financial status Partner intimacy Sexual functioning Medical care  143  Part III: Identification of Omissions Please indicate in the space provided below any omissions in the MQOL-HIV instrument, either with regards to dimensions of quality of life or the items selected to tap these dimensions. For example, is there an important dimension of QOL that you feel is relevant to HIV-infected women that is not reflected in this instrument? If so, how would you label this dimension and what types of items would you suggest that would effectively tap it? If you feel the current 10 dimensions capture the essence of quality of life in HIV infected women, do you believe that additional items would strengthen the validity of these dimensions?  Thank you for your time and effort!  144  Appendix D: Patient Panel Scoring Sheet  PART I: Item validity Please rate each of the 40 questions of the MQOL-HIV as to its relevance to the quality of life of HIV-infected women using the following rating system: 1 2 Item is not relevant  3 4 Item is not relevant unless Item is relevant but needs Item is very it receives major revision minor revision relevant and to the point  Please place the number of your score in the blank to therightof the corresponding MQOL-HIV item numbers below: Item number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  Rating (1-4)  Item number 21 22 23 .24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40  Rating (1-4)  145  Part II: Dimension validity The MQOL-HIV claims to assess 10 areas (called "dimensions") of quality of life, each dimension being represented by four items in the questionnaire as shown below: DIMENSION Mental health  Physical health  Physical functioning  Social functioning  Social support  Cognitive functioning  Financial status  Partner intimacy  Sexual functioning  Medical care  ITEMS ("How much of the time in the past 2 weeks ...") You felt depressed You felt anxious You felt needed You worried about things You had health problems You were too tired to do things you wanted to do You were bothered by nausea You were bothered by tingling sensations You were able to do tings around the house You were able to get from place to place, such as shopping and to the doctor You were able to climb several flights of stairs without help You could perform strenuous sports like running or weight lifting You felt isolated You have withdrawn from socializing You lacked energy to socialize with friends You spent quality time with friends You have received enough emotional support from people close to you You had someone who could help you in an emergency You had someone you could talk to about problems You had someone who could give you useful advise You were bothered by trouble remembering things You were bothered by forgetting what you started to do You have found that your mind wanders more than usual You were bothered by a short attention span You had to put off paying regular living expenses You have not had enough money to pay for medication You have not had enough money to pay for recreational activities You were concerned about your financial future You felt satisfied with the activities you do with your spouse or partner You felt satisfied with the amount of affection your spouse of partner expressed towards you You were able to confide in your spouse or partner You felt mistrust for your spouse or partner You felt satisfied with your sexual activity You have wanted sex more often than you have had sex You felt afraid of infecting someone through sexual contact You felt afraid of sexual rejection you wanted more information from your health care providers you wanted more support from your doctors and other health providers you doubted that your health providers are doing all they can for your health you were able to get the medical care you needed  146  Please rate each of the 10 dimensions of the MQOL-FflV as to its relevance to the quality of life of HIV infected women using the following rating system: 1 2 Dimension is not relevant  3 4 Dimension is not relevant Dimension is relevant but Dimension is very without major revision of needs minor revision of relevant and to the some of the items some of the items point  Please place the number of your score in the blank to the right of the corresponding MQOL-HTV dimension: Dimension Mental health Physical health Physical functioning Social functioning Social support Cognitive functioning Financial status Partner intimacy Sexual functioning Medical care  Rating (1-4)  Thank you for your time and effortl  147  Appendix E: Chart Review Data Collection Form Subject # Questionnaire(s) date: FJJ.V infection stage (as per CDC classification): Current viral load: Current CD4: CD4 nader:  (copies/ml) Date of test: (absolute) Date of test:  Current antiretroviral regimen: Duration of current regimen: Duration of TOTAL ART exposure: Other medications:  Chronic, comorbid medical conditions:  Current opportunistic infections:  Appendix F: Informed Consent Form - Patient Version  Appendix G: Informed Consent Form - Patient Work Group Version  Appendix H: Informed Consent Form - Health-Care Professional Version  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090275/manifest

Comment

Related Items