Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relations between illness perceptions, coping, social support, and outcomes in the Chronic Fatigue Syndrome Sleigh, Kenna Marie 2004

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

Item Metadata


831-ubc_2004-931773.pdf [ 11.58MB ]
JSON: 831-1.0091905.json
JSON-LD: 831-1.0091905-ld.json
RDF/XML (Pretty): 831-1.0091905-rdf.xml
RDF/JSON: 831-1.0091905-rdf.json
Turtle: 831-1.0091905-turtle.txt
N-Triples: 831-1.0091905-rdf-ntriples.txt
Original Record: 831-1.0091905-source.json
Full Text

Full Text

Relations between Illness Perceptions, Coping, Social Support, and Outcomes in the Chronic Fatigue Syndrome by Kenna Marie Sleigh B S N , The University of British Columbia, 1990 M S N , The University of British Columbia, 1995 A THESIS SUBMITTED IN P A R T I A L F U L F I L M E N T OF THE REQUIREMENTS FOR T H E D E G R E E OF  DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES Experimental Medicine Program, Department of Medicine We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRTISH C O L U M B I A November 2004 © Kenna Marie Sleigh  11  Abstract Chronic fatigue syndrome (CFS) is a disorder of unknown etiology characterized by fatigue, flu-like symptoms, and neurocognitive complaints. Due to the severe impairment associated with the illness patients are often unable to work and must rely on disability income. Few patients recover spontaneously although a sizeable proportion has reported improvement. Since no definitive treatment is available for CFS, it is essential to identify effective ways for patients to adjust to the illness. The purpose of the study was to describe how patients adapt to CFS. A sample of 100 patients diagnosed medically with CFS and meeting the Centers for Disease Control and Prevention (CDC) case definition were recruited primarily from the practices of two hospitalbased internists. The respondents were screened for psychiatric conditions and completed a battery of questionnaires at four time points over a 21-month period. Guided by Leventhal's selfregulatory model, multiple regression analyses and A N O V A were utilized to assess the relations between respondents' illness representations, self-management strategies (coping and social support), perceived efficacy of self-management, and outcomes including functioning, psychological distress, and health care utilization. Analyses revealed pessimistic illness representations and coping characterized by catastrophizing, focusing on symptoms, avoidance, and venting predicted unfavorable outcomes. On the other hand, optimistic perceptions and coping by accommodating to the illness, approach strategies, and social support predicted favorable outcomes. Moreover, in addition to the direct associations between illness representations and outcomes, representations were also related to coping. Pessimistic representations predicted maladaptive coping (strategies that predicted unfavorable outcomes) and optimistic perceptions predicted adaptive coping (strategies that predicted favorable outcomes). Coping frequency was a better predictor of outcomes than perceptions of coping efficacy save for physical functioning.  Ill  To examine change, analyses of differences within and between groups were conducted. When the sample was stratified into short (ill four years) and long illness duration groups (ill four years or more) the long duration group reported fewer health care visits. As a whole the respondents experienced little change in outcomes over a 21-month period but stratification demonstrated that the short illness duration group evidenced a small but statistically significant improvement in physical functioning while the long illness duration group reported less vitality. Coping and social support efficacy decreased over time although for the most part the frequency of self-management behaviors was quite stable. The implications of the study results for the management of CFS and future research were discussed.  iv T A B L E OF CONTENTS Page ii iv vi viii ix  Abstract Table of Contents List of Tables List of Figures Acknowledgements C H A P T E R ONE: Introduction and Overview Introduction Background to the Study: Overview of Chronic Fatigue Syndrome Problem Statement Purpose Research Questions Theoretical Framework Chapter Summary  1 1 1 30 30 30 31 34  C H A P T E R TWO: Review of the Literature The Cognitive Behavioral Perspective of Illness Limitations of CBT Illness Perceptions Illness Representations in CFS Patients Relations between Illness Representations and Outcomes Coping Relations between Coping and Outcomes Changes in Coping over Time Relations between Illness Representations and Coping Social Support Relations between Social Support and Outcomes Relations between Illness Representations and Social Support Appraisals of the Efficacy of Coping and Social Support Chapter Summary  35 35 36 37 45 48 48 52 57 59 61 64 66 67 70  C H A P T E R THREE: Method Research Design Participants Sample Criteria Study Procedures Measurement Statistical Procedures Ethical Considerations Chapter Summary  :  71 71 71 72 72 73 82 83 83  V  C H A P T E R FOUR: Results Data Collection Descriptive Statistics Inferential Data Analyses Chapter Summary  ,  84 84 85 102 121  C H A P T E R FIVE: Discussion Characteristics of the Sample The Role of Cognitions and Coping in Adjusting to CFS Illness Representations Relations between Illness Representations and Outcomes Coping Relations between Coping and Outcomes Relations between Illness Representations and Coping Social Support Relations between Social Support and Outcomes Relations between Illness Representations and Social Support Relations between Social Support and Coping Efficacy of Coping and Social Support Determinants of Predictive Perceptions Adjustment to CFS over Time Significance of the Study Limitations of the Study Chapter Summary  123 123 130 130 134 138 139 146 149 151 155 156 159 162 166 173 175 176  C H A P T E R SIX: Summary, Conclusions, and Implications Summary and Conclusions Implications Implications for the Management of CFS Implications for Research Chapter Summary Concluding Remarks  177 177 179 179 184 186 186  •  REFERENCES  187  APPENDICES  213  vi List of Tables Table  Page  Table 1 - Comparison of NEO-PI-R scores for respondents and a normative sample....  86  Table 2 - Psychiatric diagnoses as determined by the CDIS-IV  88  Table 3 - Descriptive statistics for the Health Visits Questionnaire  90  Table 4 - Descriptive statistics for the SF-36  91  Table 5 - Descriptive statistics for the H A D S  91  Table 6 - Descriptive statistics for the IPQ-R  93  Table 7 - Descriptive statistics for the IMQ  95  Table 8 - Descriptive statistics for the emotion-focused subscales (adapted from the COPE)  96  Table 9 - Descriptive statistics for the Catastrophizing Scale  96  Table 10 - Descriptive statistics for the SSQ-N, SSQ-S, and ISSB  97  Table 11 - Coping and social support strategies  101  Table 12 - Efficacious coping and social support strategies  102  Table 13 - Comparison of frequency and efficacy of coping measures for predicting physical functioning Table 14 - Comparison of SF-36 raw scores for respondents with short (under four years) and long (four years and more) illness durations  103 104  Table 15 - Comparison of H A D S scores for respondents with short (under four years) and long (four years and more) illness durations  105  Table 16 - Comparison of health care visits (HCV) for respondents with short (under four years) and long (four years and more) illness durations  105  Table 17 - Efficacy of coping and social strategies for respondents with short (under four years) and long (four years and more) illness durations  107  Table 18 - Efficacy subscales that are statistically different: Short versus long illness duration groups  107  Table 19 - Variables that correlated significantly with outcomes and were used in sequential multiple regression  108  vii Table 20 - Determinants of predictive perceptions  116  Table 21 - Illness outcomes: Means and SDs over four time points  118  Table 22 - SF-36: Significant differences for the physical functioning and vitality subscales between- and within-groups  119  Table 23 - Coping and social support scores over four time points  120  Table 24 - Differences in matched pairs ratings of the efficacy of coping and social support  121  Vlll  List of Figures Figure Figure 1. Theoretical Framework  Page 34  IX  Acknowledgements To the people who participated in the study: Your hard work and persistence will not be forgotten. In addition, I owe a huge debt of gratitude to Drs. Grant Stiver and Tony Chow who gave me the opportunity and support to chase a dream.  1 C H A P T E R ONE Since the 1980s there has been a resurgence of interest in a medically-unexplained illness called the chronic fatigue syndrome (CFS). Also known as myalgic encephalomyelitis (ME), chronic fatigue immune dysfunction syndrome (CFIDS), post-viral fatigue syndrome, and epidemic neuromyasthenia, CFS is a disorder of unknown etiology characterized by fatigue, flulike symptoms, sleep disturbances, and neurocognitive complaints (Fukuda, Straus, Hickie, Sharpe, Dobbins, & Komaroff, 1994). In addition to causing significant morbidity, the economic ramifications of CFS are considerable (Lloyd, Hickie, Boughton, Spencer, & Wakefield, 1992). Adults are afflicted in their prime working years (Gunn, Connel, & Randall, 1993). However, only about 28% to 39% of patients maintain full-time employment during the illness (Bombardier & Buchwald, 1996; Buchwald, Pearlman, Umali, Schmaling, & Katon, 1996). Although the trajectory of CFS is not well understood, spontaneous recovery after 15 months is considered unlikely (van der Werf, de Vree, Alberts, van der Meer, & Bleijenberg, 2002). No cure or definitive treatment have been found for CFS (Afari & Buchwald, 2003) and identifying how patients can adapt to the condition has become a clinical and research imperative. Background to the Study: Overview of the Chronic Fatigue Syndrome Symptoms of CFS The hallmark of CFS is profound, persistent fatigue that is unrelieved by rest (Fukuda et al., 1994). About 70% - 97% of patients experience post-exertional malaise, a worsening of symptoms lasting 2 - 1 4 days, occurring 6-48 hours after mental or physical activity (De Becker, McGregor, & De Meirleir, 2001; Jain & DeLisa, 1998). Another feature estimated to afflict 74% - 95% of patients is perceived cognitive dysfunction (Komaroff & Buchwald, 1991; Vercoulen, Swanink, Ferris, Galama, van der Meer, & Bleijenberg, 1996). The manifestations of cognitive dysfunction are legion and may include cognitive "fog" or confusion, memory deficits, and difficulties concentrating (Carruthers et a l , 2003). Patients also complain of sleep  disturbances including hypersomnia, frequent wakening, and insomnia. Somatic symptoms are often experienced and may include sore throat, swollen or tender cervical lymph glands, headache, muscle pain, and arthralgia. Although not included in the prevailing CFS case definitions, patients have reported restless leg syndrome, periodic limb movement disorder, bowel complaints, and orthostatic intolerance with dizziness, weakness, and lightheadedness (Caruthers et al., 2003; Stiver & Vortel, 1991). Although CFS is extremely disabling (about 45% of patients are periodically bedridden) objective measurements of symptoms bear little relationship to patients' subjective experiences (Buchwald, Sullivan, & Komaroff, 1987; Moss-Morris, 1997). Fatigue, the principal complaint, is objectively measured by determining muscle contractility. Abnormalities are associated with primary muscle diseases and neuromuscular transmission disorders (Layzer, 1998). However, studies examining skeletal, muscle functioning using MRI, endurance testing, single fiber electromyography, and magnetic resonance spectroscopy have found no differences between CFS patients and controls in terms of the intracellular metabolism and functioning of muscle tissue (Lloyd, Phales, & Gandevia, 1988; Roberts & Byrne, 1994). Similarly, CFS patients' complaints of severe cognitive disturbances have not been well substantiated in objective testing. Slight impairments in reaction time, poorer performance on complex attentional and memory tasks, and slowness in the processing of new information have been reported but no clear evidence of intellectual decline or sensory or perceptual impairments have been demonstrated (Jain & DeLisa, 1998; Michiels & Cluydts, 2001). It has been hypothesized that patients' subjective experiences of cognitive difficulties may be related to affective disorders, fatigue, and/or covert encephalopathy (Michiels & Cluydts, 2001; MossMorris, 1997; Natelson & Lange, 2002). Wessely (2001a) suggested that the mismatch between patients' subjective experiences and researchers' objective measurements might be due to patients' misperception of symptoms.  3 At the heart of this lies patients' sense of effort. In patients with chronically fatiguing illnesses mental and physical tasks may seem to require an excessive degree of effort which results in the experience of exceptional muscle exertion and cognitive strain. The basis for this disorder of effort is unknown. Historical Background of CFS The core symptoms of CFS, fatigue and persistent exhaustion, are thought to date back as far as human records allow (Byrne, 2000). In 1750, Sir Richard Manningham published a report about 'febricula' (little fever), a condition characterized by numerous constitutional complaints that manifested few clinical signs (Demitrack, 1991). In the mid 19 century George Beard th  popularized the diagnosis of neurasthenia, a symptom complex resembling CFS and thought to be brought on by the frantic pace of industrialized society (Evengard, Schacterle, & Komaroff, 1999; Jones & Wessely, 1999). Neurasthenia was one of the most frequently diagnosed conditions until the increasing use of psychiatric diagnoses led to its decline in the early 20  th  century (Demitrack, 1991). Epidemic-like outbreaks of chronic fatigue were reported world wide from 1934 to 1958 (Henderson & Shelokov, 1959). The most famous outbreak occurred in 1950 at the Royal Free Hospital in London, England (Bell, 1992). At the time doctors thought the illness represented an unusual encephalitic illness and named it benign myalgic encephalomyelitis (Byrne, 2000). Later analyses of the clinical records led researchers to question the physiological basis for the disorder, prompting some to label the event as mass hysteria (Showalter, 1997). Scientific interest in CFS was sparked in North America (Aronowitz, 1997) in January, 1985 when the Annals of Internal Medicine published two studies describing a viral-like illness characterized by fatigue and other subjective symptoms associated with serological evidence of the Epstein-Barr virus (Jones et al., 1985; Straus et al., 1985). The same year the Centers for Disease Control (CDC) dispatched a team to investigate an outbreak of a viral-like, fatiguing  .  4  illness in Lake Tahoe. Many of the patients were found to have high antibody titers to various Epstein-Barr virus (EBV) antigens. Following the intensive study of 15 Tahoe patients the CDC expressed concerns regarding the reliability of E B V serological testing, the substantial overlap of case and control serological abnormalities, and the vagueness of reported symptoms (Aronowitz, 1997). Regardless, in April 1985 the diagnosis chronic E B V infection was launched in the lay and medical worlds following a consensus conference organized by the National Institute of Allergy and Infectious Diseases (Aronowitz, 1997). During the next three years chronic E B V infection attracted considerable public and medical attention. The diagnosis may have served a segment of the population experiencing unexplained fatigue coupled with somatic symptoms, which lacked the cachet of being a recognized disease. Skepticism mounted though, when researchers were unable to reproduce findings showing specific links between E B V and chronic fatigue (Buchwald, Pearlman, Kith, Katon, & Schmaling, 1987; Holmes, Kaplan, Stewart, Hunt, Pinsky, & Schonberger, 1987; Salit, 1985). Another conference was convened in 1988. The new consensus stressed the minimal diagnostic utility of the E B V serological tests and, since that etiological link had been severed, decided to rename chronic E B V infection 'chronic fatigue syndrome' (Holmes, Kaplan, Gantz, Komaroff, & Schonberger, 1988). Etiology of CFS  The cause of CFS is unknown which has led to a continuing debate: Is the illness physical or psychological? Studies in the 1990s cast doubts on the organic basis of CFS by demonstrating the high prevalence of psychiatric illness in sufferers. However recent work has suggested that the considerable symptom overlap between CFS and psychiatric disorders probably confounded investigators' efforts to isolate discrete psychiatric illnesses using standard research methods (Jason, Richman, Friedberg, Wagner, Taylor, & Jordan, 1997). Moreover, it has not been clearly established if psychiatric illnesses precede CFS, are a reaction to the disabling effects of CFS, or  5 are co-morbid with CFS and due to an unidentified, underlying mechanism (Ray, 1991). Further, 25% to 35%) of patients meeting the criteria for CFS have no psychiatric co-morbidity (Krupp, Mendelson, & Friedman, 1991). Support for an organic basis for CFS was garnered when researchers identified various D N A fragments which were thought to indicate microbial infection or reactivation (Demitrack, 1991). Complex aberrations in immunological responses, neurohormonal stress pathways (Parker, Wessely, & Cleare, 2001), and cardiac functioning have also been offered to explain the illness (Klonoff, 1992). Various abnormalities have been found in the central nervous system using magnetic resonance imaging (MRI), single photon emission computed tomography (SPECT), and positron emission tomography (PET) scans (Klonoff, 1992). Although intriguing none of the demonstrated abnormalities have been consistently replicated or have demonstrated clinical significance or are. considered sensitive and specific to CFS (Komaroff et al., 1996; Salit, 1996). In addition, the debilitating effects of CFS would seem to be out of proportion to the modest immunologic, endocrine, cardiac, and neurological abnormalities that have been documented (Abbey & Garfmkel, 1991). More importantly, various antiviral, immunoregulatory, and metabolic drug treatments targeting presumed organic dysfunctions have been subjected to randomized, controlled trials but have failed to demonstrate significant efficacy (Lloyd, Hickie, & Loblay, 2000). Lloyd (1998) criticized research efforts driven by a "biologically reductionistic approach to disease conceptualization" conforming to the philosophy of mind/body dualism (p. 130). Recent theories on the psychology of physical symptoms tend to refute the traditional biomedical notion of a direct and linear relationship between peripheral physiological and pathological changes that culminate in the experience of symptoms. Instead, the influence of attributional and attentional factors has been highlighted (Van Wijk & Kolk, 1997). Empirical findings from the  6 literature on adaptation to chronic illness have demonstrated that functional status is better predicted by perceptions than disease activity (Kaplan, 1987; McFarlane & Brooks, 1988). CFS may be etiologically heterogeneous, with several different pathogenic paths leading to the expression of symptoms (Natelson & Lange, 2002). Multifactorial models that describe the interaction of physical and psychological factors may have greater utility in describing the illness (Shafran, 1991). In accordance with this thinking the cognitive behavioral model predicts that a cycle of cognitive and behavioral responses mediates between an acute organic illness and the chronic syndrome (Wessely, Butler, Chalder, & David, 1991). That is, while organic factors may precipitate the illness, cognitive and behavioral factors perpetuate CFS symptoms. Therapy based on the cognitive-behavioral model has reduced disability and fatigue in some CFS patients (Whiting, Bagnall, Sowden, Cornell, Mulrow, & Pamirez, 2001). Case Definitions of CFS The lack of a specific biomarker to identify CFS has necessitated the development of a case definition to standardize research efforts and improve the comparability of findings (Salit, 1996). The lack of a "gold standard" against which the identification of cases can be validated plus the syndrome's complexity has led researchers to propose somewhat different criteria (Ray, Weir, Cullen, & Philips, 1992). Four definitions are extant worldwide and include: (1) America's C D C case definition (Fukuda et al., 1994), (2) the Oxford criteria developed in the United Kingdom (Sharpe et al., 1991), (3) the Australian definition (Lloyd et al., 1993), and most recently, (4) the Canadian definition (Carruthers et al., 2003). A l l case definitions require a principal complaint of fatigue of six months duration in addition to substantial curtailment in premorbid functioning not attributable to another illness. Differences in the definitions are manifested in the accompanying inclusion and exclusion criteria. Originally published in 1988 (Holmes et al.), the C D C definition has been revised twice • and made less restrictive by; (1) allowing for the inclusion of patients with affective and non-  7  psychotic disorders (Schluederberg et al., 1992) and (2) by decreasing the number of required symptoms from eight to four (Fukuda et al., 1994). The Oxford definition (Sharpe et al., 1992) is the least restrictive in that only fatigue that affects physical and mental functioning is required. For the Australian definition (Lloyd et a l , 1993) signs of immune dysfunction (abnormal cellmediated immunity and/or cutaneous anergy) must be present in addition to fatigue and cognitive impairment. The newly developed Canadian definition (Carruthers et al., 2003) has yet to be used in published reports and is the most restrictive. Post-exertional malaise, sleep dysfunction, and pain in addition to fatigue are the required principal complaints. Additional criteria include the presence of two signs or symptoms indicating neurological or cognitive dysfunction plus one sign or symptom from two of the following systems; autonomic, neuroendocrine, and immune. Epidemiology of CFS Prevalence CFS has been found in most industrialized countries (Blondel-Hill & Shafran, 1993) in patients ranging from adolescence to late middle age with the average age of onset reported as 30 - 36 years for subjects recruited from the community (Jason et al., 1999; Reyes et al., 1997). Because CFS is a syndromal diagnosis and based on subjective symptoms, pinpointing its prevalence has been difficult. Self-reported fatigue is distributed continuously in healthy community members and those accessing medical care. Out of necessity, epidemiologic studies have imposed arbitrary boundaries, creating potentially artificial illness categories. Prevalence rates are noteworthy for their broad range. In a randomly selected sample (N=3066) taken from a health maintenance organization in the state of Washington and diagnosed using the Holmes et al. (1988) criteria investigators estimated the prevalence of CFS to be 75 to 267 individuals per 100,000 (Buchwald, Umali, Umali, Kith, Pearlman, & Komaroff, 1995). On the other hand, a large (N=l 8,675) epidemiological survey using random selection in a multi-ethnic urban community calculated the prevalence rate at 420 cases per 100,000 using  8 the 1994 Fukuda criteria (Jason et al., 1999). Lastly, a group of British researchers (Lawrie, Manders, Geddes, & Pelosi, 1997) who applied the Oxford definition reported a prevalence of 740/100,000 in a random sample of adults (N=695) registered at a local Health Center (covering about 75% of the local population). Prevalence rates vary depending on the case definition, the setting from which the sample was obtained, and the diagnosing physician's specialty (Aaron & Buchwald, 2001). Higher prevalence rates are associated with less restrictive criteria. Rates are also higher in community samples than clinical samples due to the winnowing effects of help seeking. In addition, case finding and prevalence are confounded by the overlap of symptoms between CFS and fibromyalgia (FM; Buchwald & Garrity, 1994). If provided with the appropriate physical examination, it is thought that about 70% of CFS patients will meet the criteria for F M , a rheumatological illness marked by widespread musculoskeletal pain in the presence of soft tissue "tender points" (Goldenberg, Simms, Geiger, & Komaroff, 1990). Whether the patient is diagnosed with CFS or F M may have more to do with the specialty of the consulted physician (infectious diseases versus rheumatology) than with the illness being reported (Stewart, 1990). Age and gender. In the United States CFS is most prevalent in middle age with the highest rates occurring in the 40- to 49-year-old age range. With a prevalence of .52%, women are more likely to have CFS than men, who exhibit a prevalence of .29%. Jason et al. (1999) suggested that the predominance of CFS in women may reflect gender-related vulnerabilities including reproductive factors and stress-associated immune modulation. The female to male ratio rises when samples are drawn from clinical settings with about 70% - 89% of patients being women (Friedberg, Dechene, McKenzie, & Fontanetta, 2000; Torres-Harding, Jason, & Taylor, 2002). Illness behavior may account for this phenomenon. A higher proportion of women (60%) than men visit physicians for all reasons (Torres-Harding et al., 2002), perhaps due in part to gender differences in coping repertoires (Leventhal & Diefenbach, 1991) and/Or attention to  9 bodily cues (Van Wijk & Kolk, 1997). Moreover, women are more likely than men to report fatigue as a symptom (Steele et al., 1998). Finally, fatiguing illnesses of all kinds including arthritis, rheumatism, multiple sclerosis, and systemic lupus erythematosus are more predominant in women (Carruthers et al., 2003; Evengard et al., 1999). Ethnicity and socio-economic status. CFS samples obtained from the community and medical settings also demonstrate inconsistencies in terms of ethnicity and socioeconomic status. Jason et al. (1999) found that African-Americans and Latinos were more likely than Caucasians to meet the criteria for CFS. Although no differences were found in educational attainment between CFS patients and controls, skilled craftsmen and clerical workers demonstrated the highest prevalence rates. However, it has been reported that patients in clinical studies are more likely to be Caucasian, well educated, and of higher socio-economic status (Euba, Chalder, Deale, & Wessely, 1996; Shafran, 1991). Again, social factors and help-seeking behaviors may provide an explanation. Individuals of color are more likely to be poor, underserved by the medical system, and subject to health compromising stressors (Jason et al., 1999). On the other hand, Caucasians tend to be more affluent and better educated. They may have a greater awareness of CFS as a diagnostic entity and the resources to demand and pay for professional help (Shafran, 1991). Occupation. CFS may be more prevalent in health care workers. In a community-based sample, health care workers accounted for 15% of the CFS patients, a figure which was significantly higher than the composition of health care workers in the American population (Jason, Taylor, Kennedy, Song, Johnson, & Torres, 2000). In an earlier study, CFS prevalence rates in nurses were found to be 1088 out of 100,000 (Jason et al., 1993). Coulter (1988) reported that 40%o of the members of a CFS patients' organization were associated with health care professions. Occupational factors that may place health care workers at higher risk for CFS include excessive work loads and stressful shift work that disturbs circadian rhythms.  10 . Illness Course The illness course for CFS varies. For about 37% to 75% of patients, the illness begins suddenly while others experience an insidious onset with symptoms appearing over many months (Jason et al., 2000; Wilson et al., 2001; Salit, 1997). About 30% of sufferers state that the symptoms are continuous (Bell, 1991). However, most patients experience a waxing/waning pattern with the severity as well as the type of symptoms changing over time, sometimes within hours (Bell, 1991). Prognosis Although CFS is not fatal except for suicide, relatively few sufferers completely recover (Joyce, Hotopf, & Wessely, 1997). Early in the 1990s it was thought that the illness was self limiting but subsequent investigations have reported that less than 10% of patients regain premorbid functioning (Joyce et al., 1997; McCluskey, 1992; Milbon, 1991; Vercoulen et al., 1996). Factors associated with a poorer prognosis include a more chronic immobilizing illness (bed rest), longer duration of illness, co-morbid psychiatric disorder or fibromyalgia, older age, cognitive dysfunction, and the belief that the illness has a physical cause (Joyce et al., 1997; Ray, Jeffries, & Weir, 1997). About 17% to 60% of patients reportedly experience improvements in symptom severity and/or functioning over 1 to 48 months (Friedberg et al., 2000; Joyce et al., 1997; Saltzstein, Wyshak, Hubbuch, & Perry, 1998; Vercoulen et al., 1996). In a longitudinal study following patients for 2.5 years, investigators found that factors associated with better functioning and/or return to work included improvements in physical examination findings and psychological health (Russo, Katon, Clark, Kith, Sintay, & Buchwald, 1998). Clearly, studies report varying improvement rates. This is likely due to the heterogeneity of the CFS population and differences in sampling criteria. Less restrictive case definitions are related to better outcomes while restrictive criteria are thought to capture a more severe illness which may augur unfavorable outcomes.(Becker, McGregor, & De Meirleir, 2001). In addition  11 the duration of symptoms in patients varies between studies. Patients with shorter illness durations are more likely to improve over time. Finally, although rarely included in published reports, the presence of co-morbid illnesses is also related to a poorer prognosis. Subgroup analysis. In an attempt to improve prognostication, theorists have tried to identify homogeneous subsets of the CFS population based on illness presentation. The presence of post-exertional malaise has been found to be a prognostic indicator of poorer outcome (Taylor, Jason, & Curie, 2002). Levine (1997) surmised that CFS cases occurring as part of a cluster have a better prognosis as a group than sporadic cases. As well, patients with an acute onset may have a better prognosis than patients with a gradual onset. Wessely et al. (1991) suggested that a gradual onset might overlap with the psychiatric concept of somatization disorder while the acute onset may be associated with a more discrete etiologic trigger, like infection. Some support for Wessely's contention may be found in a study that compared patients with different onsets of CFS. Patients with a gradual onset were more likely to manifest a co-morbid psychiatric disorder (8 out of 11 subjects) than patients with a sudden onset (7 out of 25 subjects). In addition patients with sudden onset had greater memory impairment (DeLuca, Johnson, Ellis, & Natelson, 1997). Hypothesizing that patients with an infectious-type onset would experience better outcomes, a Japanese study (Masuda, Nakayama, Yamanaka, Koga, & Tei, 2002) compared patients with post-infectious (N=9) and non post-infectious CFS (N=9) all of whom met the C D C case definition (Fukuda et a l , 1994). Two years after multidisciplinary treatment (drug therapy, rehabilitation, counseling, CBT, family therapy, and exercise therapy) the two groups were compared. In the post-infectious group eight patients had returned to work, while in the non post-infectious group only three patients had returned to work. Although the findings are intriguing, methodological flaws limit the study's usefulness. The very small sample precludes generalizing the findings until replication is achieved in a larger  12 population. In addition there are marked differences between the two groups in illness duration. At study and treatment start, the post-infectious group had been ill for only about 8.2 months, while the non post-infectious group had illness durations of about 38.2 months. As discussed, shorter illness durations are associated with improvement. Further, treatment was initiated in the post-infectious group far earlier in the illness course. Early intervention may have kept postinfectious subjects from becoming deconditioned and/or developing entrenched, maladaptive cognitions and behaviors that have been associated with negative outcomes (Moss-Morris, 1997). At present, criteria-based approaches to the diagnosis of CFS do not select homogeneous groups. However, to date there is no consensus on how to best reduce the heterogeneity (Wilson et al., 2001). Theorists suggest researchers stratify subjects according to illness duration, comorbid fibromyalgia, psychiatric status, and rapidity of onset (Fukuda et al., 1994; Wessely, 2001a). Analyses of homogeneous subsets may lead to improved replication of findings and aid in the identification of potential causative agents (DeLuca et al., 1997). Diagnosis and Management of CFS The lack of a specific biomarker for CFS ensures diagnosis rests on the patient meeting one of the case definitions. However, the case definitions were primarily developed to guide research. Salit (1996) urged physicians to base clinical management on "an assessment of the patient and not solely on whether they meet published diagnostic criteria" of CFS (p. 540). The recommended medical assessment for diagnosing CFS includes a thorough medical history, physical examination, mental status examination, and laboratory tests to identify underlying illnesses that may exclude CFS or co-morbid conditions that may require treatment (Klonoff, 1992). Management is aimed at relieving symptoms, providing emotional support, and protecting the patient from harm (Klonoff, 1992; Masuda et al., 2002). Sleep disturbances, pain,  13 orthostatic intolerance, and cognitive dysfunction may be treated using pharmacological agents and cognitive behavioral therapy (Meisler, 2000). If concomitant depression is diagnosed, psychotherapy and antidepressants may be ordered. It is recommended that small doses of drugs be prescribed, as many patients are intolerant of medication. To improve physical functioning and reduce emotional distress, clinicians recommend light exercising and relaxation techniques like yoga or meditation (Stiver & Vortel, 1991). Patients are advised to pace activities (undertake 50% of what they think they can do) and avoid "boom or bust" activity patterns characterized by periods of prolonged rest interspersed with bursts of high activity that culminate in feelings of exhaustion and helplessness (Stiver & Vortel, 1991). Supportive counseling is needed to help patients adjust to the debilitating and chronic nature of the illness. Finally, education regarding CFS is essential in order to protect patients from exploitive and potentially dangerous therapies (Blondel-Hill & Shafran, 1993; Wessely, David, Butler, & Chalder, 1989). Challenges of CFS CFS has a highly negative effect on patients' lives. Suffering is compounded by the stigma attached to CFS, the impact of the illness on lifestyle, and difficulties obtaining supportive health care. The Stigma of Having a "Non-disease " The Oxford Dictionary (Allen, 1990) defines stigma as "a mark or sign of disgrace or discredit" (p. 1197). Stigma occurs when a discrepancy occurs between an expected, socially defined norm and an actual attribute (Goffman, 1963). Since disease deviates from the norm of good health, chronic illness is a potentially stigmatizing condition. However, for CFS patients, stigmatization occurs not because they have an illness, but because they have an illness that cannot be proven. The etiology of CFS is unknown and sufferers often appear healthy with no demonstrable physical abnormalities. In addition patients may be functional one day and sick in bed the next.  14 Such dramatic changes may be interpreted as meaning the symptoms are self-imposed rather than the result of unpredictable fluctuations of a poorly understood illness (Friedberg & Jason, 2001). Many in the medical and lay communities dispute the validity of CFS as a diagnostic entity, preferring to believe the condition is benign or that CFS is a fashionable label for people who are unwilling to deal with the stress of living (Clarke, 1999; Cooper, 1997). Indeed, in the 1980's the media labeled CFS "yuppie flu" and promoted the idea that the illness was a selfinflicted condition of the young, upwardly mobile (Ware, 1994). The veracity and accuracy of patients' symptom descriptions are often doubted and the symptoms patients find most disabling, like fatigue and cognitive dysfunction, are trivialized and dismissed (Asbring & Narvanen, 2002; Anderson & Ferrans, 1997). The de-legitimization of CFS may lead others to question the morality of CFS patients, especially when functional limitations result in work absences or the inability to fulfill family responsibilities. CFS patients find the psychologizing of their somatic symptoms especially stigmatizing (Asbring & Narvanen, 2002). The key to understanding this reaction lies in the mind-body dualism that informs Western medicine. Illnesses are considered either psychological or physical (Ware, 1994). With no signs of physical abnormality, the physician may diagnose patients' symptoms as a psychological illness. Dualism dictates that physical illnesses are real and not under the control of the sufferer while psychological illnesses are imaginary and personally controllable (Aronowitz, 1997). Since society prefers to confer the "sick role" on the deserving few with a clearly demonstrable problem for which they are blameless, patients with psychological illnesses are viewed with disdain, suspected of usurping the sick role because they cannot show tangible proof of disease and lack the will-power to control their pathological intrapsychic processes (Mechanic, 1995). Thus, the "demotion" of symptoms from a high status  15 physical disease to a low status psychological illness places sufferers at risk of incurring social contempt. Although patients are often relieved to obtain a medical diagnosis many are unhappy with the term CFS. Patients consider CFS to be a misleading appellation that does not properly signify the gravity of their condition, which may lead to misunderstanding, belittlement, and marginalization. According to interviewees in Asbring and Narvanen's (2002) study, CFS lacks legitimacy among caregivers. In the U . K . myalgic encephalomyelitis (ME) is the preferred term. In Canada efforts are underway to change the name from CFS to M E (Carruthers et al., 2003). For some patients, dealing with the stigma associated with CFS is as burdensome as dealing with the illness itself. Stigma is a source of suffering that may manifest as shame, humiliation, guilt, and the deterioration of self-esteem. CFS patients have developed strategies to avoid stigmatization like "passing" or trying to appear normal. Withdrawing from social situations, concealing the illness through impression management, and controlling information are other ways of avoiding stigmatization (Asbring & Narvanen, 2002; Ware, 1998). Regrettably such behaviors may lead to increased social isolation, alienation, and the loss of personal authenticity (Antoni et al., 1994). At the collective level numerous advocacy organizations have sprung up throughout the world to lobby for greater understanding and improved treatment of CFS patients (Couper, 2000). Lifestyle Changes One of the greatest challenges confronting CFS patients is changes in lifestyle caused by functional limitations. Quality of life instruments including the SF-36 Health Survey (SF-36; Ware & Sherbourne, 1992), the Sickness Impact Profile (SIP; Bergner, Bobbit, Carter, & Gilson, 1981), and the Quality of Life Index (Ferrans & Powers, 1985) have consistently demonstrated in Europe, Canada, and the United States that CFS patients experience severe impairments in physical, occupational, and social functioning (Antoni et al., 1994; Buchwald et al., 1996; Hardt,  16 Buchwald, Wilks, Sharpe, Nix, & Egle, 2001; Myers & Wilks, 1999). Since CFS primarily strikes relatively young, previously healthy and productive adults, lifestyle changes are drastic (Clark, 1999). Effects of CFS on employment. Due to the disabling nature of CFS patients have difficulty maintaining premorbid levels of employment. The majority of sufferers must quit or reduce the number of hours they work (Barrows, 1995; Packer, Foster, & Brouwer, 1997; Yeomans & Conway, 1991). In Australia, after the onset of CFS, participation in the work force for men dropped from 73% to 37%, women's participation dropped from 52% to 19% (Lloyd & Pender, 1992). Moreover, patients may remain out of the work force for long periods (Bombardier & Buchwald, 1996). In a longitudinal study (N=35) about 68% of subjects were unemployed at the start of the study, a figure that remained unchanged at follow-up 41 months later despite self-reported improvement (Tiersky et a l , 2001). Patients who try to maintain employment face numerous challenges. The unpredictable course of CFS makes it difficult for patients to conform to a work schedule (Anderson & Ferrans, 1997). Getting up in the morning and commuting to work represents a major undertaking for patients who experience sleep disturbances and disorientation severe enough to prohibit driving a car (Ware, 1994). Shift work is particularly problematic for patients because it disturbs sleep patterns and circadian rhythms which may lead to a worsening of symptoms (Jason & Wagner, 1998). Patients with orthostatic intolerance may find that standing or sitting for prolonged periods results in dizziness, palpitations, or fainting spells (Carruthers et al., 2003). Environmental factors in the workplace including sensory overload (light, sound, motion, confusion, and odors) as well as cold, pollutants and chemicals are poorly tolerated by patients and perceived by many to have a detrimental effect on their condition. Symptoms of CFS compromise patients' efficacy on the job. Perceived cognitive disturbances may impair critical communication skills like listening, speaking, and writing  17 (Barrows, 1995; Ware, 1993). Job-related learning may be inhibited by patients' problems processing and remembering new information (Barrows, 1995). Recent reports suggest CFS patients experience disruptive cognitive dysfunctions. Compared to healthy controls, CFS patients scored lower on measures of divided attention, defined as the reallocation of cognitive resources between visual and auditory response channels (Ross, Fantie, Straus, & Grafman, 2001). Thus, clerks, secretaries, and receptionists who multitask by answering phones and typing may be at high risk of performance deficits. Physical and mental fatigue may create difficulties for patients trying to meet deadlines, avoid mistakes, or for individuals in supervisory or decision-making roles (Schweitzer, Kelly, Foran, Terry, & Whiting, 1995). Sleep deprivation as well as profound fatigue worsened by exertion limit patients' stamina and their ability to perform physically demanding tasks (Fulcher & White, 1997): Even non-demanding tasks may pose difficulties. Activities characterized by repeated moving, lifting, bending, and unsupported extension of the arms seem to be troublesome and are considered work place "aggravators" (Carruthers et a l , 2003). CFS patients have reported using a variety of strategies to mitigate the effects of illnessrelated impairments on work performance. In a longitudinal study (N=66) Ware (1994) found that patients often prioritized work and eliminated many other activities to conserve energy. For some patients maintaining employment means spending each evening and weekend in bed (Asbring, 2000; Barrows, 1995). Finding "easier" (and probably lower paying) work may also be an option. To compensate for cognitive deficits, primarily memory problems, patients often write things down. "Passing" or cultivating the impression of health on the job, is another strategy for staying employed. Ware (1994) suggested that finding flexibility at work that allows patients to negotiate a variable schedule and incorporate "rest breaks" into the day is perhaps the most effective way for patients to maintain employment.  18 Effects of CFS on the family. Research describing the impact of CFS on the family is minimal. Like other chronic illnesses, though, CFS probably interferes with patients' ability to meet role demands (Barrows, 1995). Household tasks like cleaning, cooking, and shopping take longer to perform due to the need to pace activities (Schweitzer et al., 1995). Some tasks may have to be delegated to other family members. Unfortunately, families are not always accepting of the diagnosis and may resent providing support for an illness that isn't "real" (Schweitzer et al., 1995). Even when assistance is offered patients may be resistant. Accepting they are ill was considered an important adaptive task by CFS patients in a Dutch study (de Ridder, Schreurs, & Bensing, 1998). Reluctance to "give in to the illness" and accept its limitations may be dispositional and could stem from patients' pre-morbid personality which is thought to be characterized by "action-proneness" featuring an inability to set personal limits and a tendency towards perfectionism (White & Schweitzer, 2000). Saltzstein et al. (1998) found that in response to the stress of illness, the defense mechanism used by their CFS subjects 40% of the time was to "minimize their feelings and persevere with demanding tasks while ignoring affective signals of distress" (p. 313). Learning to "let go" and accept the help of others may represent a significant challenge to the CFS patient (Van Houdenhove, Neerinckx, Onghena, Lysens, & Vertommen, 2001; White & Schweitzer, 2000). Like other chronic illnesses, CFS has an impact on marital relationships due in part to sufferers' impairment and increased need for assistance. Although it has not been examined in the CFS population, investigators studied the adaptation of spouses of patients with fibromyalgia (FM). Husbands of F M patients (N=135) reported being "forced to make lifestyle changes, including taking on new roles and responsibilities and giving up activities" (Bigatti & Cronan, 2002). In addition, due to the disruption in their wives' professional careers one-third of husbands experienced financial strain. Moreover, because their wider social networks are  19 interrupted by the illness, CFS patients become more emotionally reliant on family members (Anderson & Ferrans, 1997). Finally, CFS patients' spouses often act as primary caregivers and, although studies are lacking in the CFS population, caregiver stress has been well documented in populations of the chronically ill (Schweitzer et al., 1995). Brigatte and Cronan (2002) found that husbands of F M patients scored higher on loneliness, depression, and subjective stress than a comparison group. Although marital strain often occurs in the aftermath of CFS, marital adjustment may not necessarily be jeopardized. In Goodwin's (1997) study, couples who completed a marital adjustment scale averaged about 110, well above the marital distress cutoff score of 100 determined in a normative sample. However, Goodwin (1997) surmised that fatigue had an aversive effect on couples' sexual relationship and shared activities which likely contributed to scores less than the normative average of 114. Studies have suggested that about 45% of patients experienced reduced sexual functioning or decreased libido (Friedberg et al., 2000; Cope, David, Pelosi, & Mann, 1994). Little has been documented about the effect of CFS on the parenting role. One-third of Schweitzer et al.'s (1995) interviewees (N=23) reported less energy to devote to child care activities stating: "I'm not myself, (I) don't want to play with them, (I) lose patience easily, and (I) don't want to help them with their homework" (p. 1370). However, other researchers have demonstrated that regardless of illness limitations, patients were content with the relationships they had with their offspring. Anderson and Ferrans (1997) found that 78% of the subjects who had children were highly satisfied with these relationships. Effects of CFS on social networks. The social networks of CFS patients are vulnerable to disruption. Anderson and Ferrans (1997) found that 32% of a sample subgroup (N=22) reported losing most, i f not all of their previous friendships. For 33% of the patients in Komaroff and Buchwald's (1991) sample, disclosing a diagnosis of CFS to friends and co-workers provoked  20 reactions of disbelief and non-acceptance. Fear of contagion is a justification that individuals may use to avoid an acquaintanceship with a CFS patient (Anderson & Ferrans, 1997). In response to such experiences, patients often withdraw socially or withhold illness-related information (Asbring & Narvanan, 2002). CFS exerts a detrimental effect on patients' ability to sustain and build social networks. Patients engage in fewer and less active leisure pursuits, which may mean forfeiting relationships based on strenuous activities like sports (Anderson & Ferrans, 1997; Schweitzer et al., 1995). Even sedentary pursuits like visiting with friends may prove taxing, with cognitive disturbances impairing patients' ability to follow conversations (Barrows, 1995). The scheduling of social events represents a genuine dilemma for many CFS patients because they can't predict whether or not they will feel well enough to attend. Due to lack of energy, sufferers may simply give up trying to seek out and establish new friendships (Asbring, 2000; Schweitzer et al., 1995). Ultimately, some CFS patients may become socially isolated. Although their social networks shrink, many CFS patients are able to maintain contact with a few close friends (Schweitzer et al., 1995) and some patients make new friends (Anderson & Ferrans, 1997). As well, the growing number of CFS self-help groups may provide a compensatory social venue. Fellow patients may be more sympathetically attuned and provide solidarity and acceptance in a social context that does not inhibit the sharing of information. However, patients' reactions to other CFS sufferers are not always positive. In their qualitative study Asbring and Narvanen (2001) reported that their sample (N=15) felt co-patients were "too focused on their illnesses and the problems associated with them" (p. 156). Furthermore, membership in a self-help group has been related to poorer outcomes for patients with fatiguing illnesses (Bentall, Powell, Nye, & Edwards, 2002; Sharpe, Hawton, Seagroatt, & Pasvol, 1992). Mechanic (1996) suggested that patient groups which endorse a particular theory of illness may reinforce unhelpful illness behaviors, such as the avoidance of activity.  21 Problems Negotiating the Medical System Although they experience profound functional impairment, when CFS patients seek medical attention they are often disappointed. Physicians are not always responsive to their needs. Deale and Wessely (2000) found that two-thirds-of patients referred to a specialist clinic were dissatisfied with the medical care they received. Patients' dissatisfaction centered on the confusing and often lengthy process of diagnosis, receiving an unacceptable psychiatric diagnosis, physicians' negative attitudes towards CFS, and physicians' lack of knowledge about the illness. Patients also complained about physicians who failed to provide emotional and informational support (Ax, Gregg, & Jones, 1997). Cooper's (1997) sample (N=10) recalled highly upsetting interactions with physicians who rejected or trivialized the symptoms. Such experiences motivated patients to "demonize" physicians or hold them in contempt, a reaction which was often generalized to all conventional medical services. Diagnostic issues. Obtaining a medical diagnosis of CFS can be a difficult process for many patients (Broom & Woodward, 1996). In a community-based U.S. epidemiological survey (N= 18,735) investigators reported that of the 32 individuals meeting the C D C criteria for CFS, 21 (65%) had consulted a physician but only three had been diagnosed with CFS (Jason et al., 2000) . Similar findings were reported in a CDC prevalence study. About 85% of CFS cases identified by the study investigators had not been previously medically diagnosed (Reeves, 2001) . Some physicians simply do not consider CFS a valid diagnosis and interpret symptoms as manifestations of a psychiatric illness, malingering, or factitious disorder (Davis, 1993). After all, it may be argued, no biomarker exists to diagnose the illness and there is considerable symptom overlap between CFS and depression, anxiety, and somatoform disorder. However, in addition to studies demonstrating important differences between CFS and psychiatric illnesses  22 (Lane, Manu, Dale, & Matthews, 1991; Moss-Morris & Petrie, 2001; Natelson & Lange, 2002; Komaroff et al., 1996) Sharpe (1994) questioned the usefulness of applying a psychiatric diagnosis to CFS. Not only do patients refuse to accept this diagnosis they respond to it with hostility which would severely limit any future therapeutic alliance between a patient and his/her physician. Physicians may fear that a diagnosis of CFS will act as a self-fulfilling prophecy and lead to chronic invalidism (Ffadler, 1997; Sharpe, 1998; Stewart, 1990; Woodward, Broom, & Legge, 1995). Although some evidence suggests that CFS patients who believe their condition is "purely" medical do not fare as well as patients who endorse a multifactorial model of etiology (Joyce et al., 1997; Sharpe et al., 1992), it may be the avoidance of activity rather than the belief in an organic cause that leads to unfavorable outcomes. Certainly at the time a physician is making the diagnosis it would be expected that he/she would explain the important role psychosocial factors play in the illness. Physicians are also reluctant to diagnose CFS because they fear they might overlook another more treatable disorder (Davis, 1993; Denz-Penhey & Murdoch, 1993). Fortunately, readily accessible guidelines from government agencies and medical journals enable a physician to conduct a thorough work-up and limit the possibility of misdiagnosis. In a review of 26 longitudinal studies Joyce et al. (1997) found it rare for new physical diagnoses to appear when subjects were properly investigated at the start of a study. However, physicians have been known to misdiagnose psychiatric disorders co-morbid with CFS. In a sample of 68 CFS patients in a hospital-based fatigue clinic, Deale and Wessely (2000) found that the referring physicians gave a psychiatric label to 31. Twenty-one (68%) of these individuals had been misdiagnosed with 17 individuals not meeting the criteria for any psychiatric illness and 4 meeting the criteria for a disorder other than that which was diagnosed  23 by the physician. Moreover, of the remaining 35 patients presumed to have no psychiatric illness 13 (35%) were found to have a treatable psychiatric condition. In contrast to Deale and Wessely's (2000) study, investigators found that few CFS patients recruited for a community-based study had been diagnosed with a psychiatric disorder when none existed (Torres-Harding, Jason, Cane, Carrico, & Taylor, 2002). Torres-Harding et al. (2002) concluded that tertiary-care patients may be more vulnerable to misdiagnosis due to the increased number of medical evaluations they have undergone and the misconception among some physicians that CFS is a primary psychiatric illness. Like the Deale and Wessely (2000) study, Torres-Harding et al. (2002) reported that referring physicians had missed about one-third of the lifetime psychiatric disorders later diagnosed in the CFS group. Although physicians may question the usefulness of a CFS diagnosis, patients do not (Denz-Penhey & Murdoch, 1993). Providing a diagnosis may be one of the most important services a physician can offer (Cooper, 1997; Deale & Wessely, 2000; Lehman, Lehman, Hemphill, Mandel, & Cooper, 2002). Diagnosis gives patients a structured and rational way to make sense of their suffering and is considered by some to be a key feature in the doctor-patient encounter (Wood, Bentall, Gopfert, & Edwards, 1991). Moreover, healing may be delayed until a diagnosis is rendered, because as Hadler (1997) pointed out: "If you have to prove you are ill, you can't get well" (p. 172). Diagnosis can have psychological, social, and financial benefits. In the early stages of the illness patients are often confronted with a frightening and confusing array of symptoms. A medical diagnosis provides an explanation for bodily changes which reduces a patient's anxiety over the possibility of even more frightening alternatives (Ax et al., 1997; Sharpe, 1994). A diagnosis also legitimates the individual's assumption of the sick role and reduces the stigma of having a psychiatric illness or being considered a malingerer (Sharpe, 1994). Through diagnosis "the physician formally acknowledges the reality of the patient's experiences and approaches the  24 illness as though it were any other process in need of medical attention" (Lehman et al., 2002, p. 462). Lower depression and anxiety scores have been recorded in patients who believed their physician legitimized their illness through diagnosis or belief (Lehman et al., 2002). Having a credible diagnosis also makes various entitlements possible, including disability pensions and legitimate reductions in work and family role demands. Finally, a prompt and accurate diagnosis reduces the risk of over-investigation and iatrogenesis which may occur i f CFS patients continue to seek out health care providers in order to obtain a satisfactory explanation for their symptoms. Issues in the patient-physician relationship. Once the diagnosis has been made, patients and physicians may have difficulty establishing a therapeutic relationship. A n area of potential conflict concerns the etiology and treatment of CFS. Due to the flu-like and somatic nature of the symptoms, most patients believe CFS has a physical, usually an infectious cause, beliefs which are reinforced by the media and support groups (MacLean & Wessely, 1994). However, physicians consult peer-reviewed medical journals and most research studies do not support an organic model of causation. Acting accordingly, physicians often state that they don't know the cause of CFS or espouse a multifactorial model that includes psychological factors (Byrne, 2000). Over time some patients acquire considerable information about CFS from the Internet, television, lay publications, and newspapers. Patients may attempt to update physicians on the latest findings regarding causal theories and new treatments (Cooper, 1997). Physicians may feel their professional expertise is being challenged and/or refute the validity of patients' information declaring that certain media sources are not credible and/or that patients are not academically prepared to properly critique research. Maclean and Wessely (1994) believe some of the blame for the polarized and contentious encounters between patients and physicians may be directed towards biased media coverage of CFS that amplifies and distorts divisions within the medical and research communities.  25 Lack of agreement about the cause of CFS can lead to conflict regarding its management. Patients are often desperate and may request treatments for organic dysfunctions, including immune system deficiencies, infections, or endocrine imbalances (Meisler, 2000). However, physicians are hesitant to prescribe drugs or surgery with potentially dangerous side-effects for conditions which are not objectively verifiable. On the other hand, due to the high rates of psychiatric co-morbidity associated with CFS, physicians may suggest psychotherapy (Natelson, 2002). Regrettably, patients often reject psychological interventions interpreting them to mean that the symptoms are "all in their head." Similarly, patients and physicians may dispute the appropriate role of exercise. Deconditioning caused by inactivity is thought to increase disability (McCully & Natelson, 1999). Moreover, clinical trials support the efficacy of graded exercise therapy for some patients (Fulcher & White, 1997). However, most patients are intolerant of exercise and experience a worsening of symptoms (Bell, 1992). For many sufferers rest or pacing are the preferred ways of dealing with CFS (Ax et al., 1997; Carruthers et al., 2003; Lehman et al., 2002; Sharpe, 2002). Whether due to patient/physician conflict or the high level of "need to know," CFS patients have been found to be frequent users of health care averaging about 22 visits a year (Bombardier & Buchwald, 1996; Buchwald & Garrity, 1994). Similar findings were published in an Australian study reporting that CFS patients visited general practitioners and specialists 18 times a year (Lloyd & Pender, 1992). It has been suggested that excessive health care utilization reflects patients' dissatisfaction with physicians who fail to provide adequate treatment. However, Deale and Wessely's study (2000) led the authors to conclude that dissatisfaction was more contingent on physicians' interpersonal and informational skills. The divergent illness schemas adopted by physicians and CFS patients may underlie nonsupportive relationships (Kleinman, Eisenberg, & Good, 1978). Physicians are trained in the biomedical model which holds that the "diagnostic" reality of a physical disease manifests as an  26 anatomical lesion or cellular dysfunction and that a one-to-one correlation exists between tissue destruction and impairment (Hadler, 1997). Because CFS rarely demonstrates any objectivelyverifiable abnormalities, physicians have difficulty taking the illness seriously (Hadler, 1997). Unfortunately, patients early in their illness career are ignorant of the special problems medically-unexplained illnesses pose to the diagnostician and assume that severe and disabling somatic symptoms warrant immediate and sustained medical attention (Byrne, 2000). Expecting a compassionate healer who can cure their illness, or at least a trusted authority figure who will legitimize and understand their suffering, patients' expectations are soundly dashed when the physician is unable to identify any pathology, minimizes their complaints, and/or questions the source of their disability (Cooper, 1997). At this point, the physician is seen to fail in his/her primary role and may be accused of violating a central tenet in the practice of medicine which is to provide help (Cooper, 1997). In addition to conventional therapies, CFS patients often utilize therapies derived from complementary and alternative medicine (CAM), defined as "self- or other-assisted health care strategies that are unconventional or historically have not been a component of western medical practice" (Afari et al., 2000, p. 98). A few therapies including massage, homeopathic remedies, and osteopathy have been subjected to randomized clinical trials (RCT) and demonstrated some benefits (Whiting et al., 2001). However, many C A M practices are not well-suited to the methodologies utilized in the RCT, which has probably led to their low regard among practitioners of evidence-based medicine (Thorne, Best, Balon, Keiner, & Rickhi, 2001). Nonetheless, the public has become increasingly interested in therapies outside the realm of conventional medicine. In a clinical sample of CFS patients (N=200) Lane et al. (1991) reported that one-third had utilized alternative care. In a smaller study (N=30) all patients had reported seeing a massage therapist or acupuncturist in the previous year (Buchwald & Garrity,  27 1994). In a study of CFS discordant twins (63 twin pairs) Afari et al. (2000) reported that 91% of twins with CFS and 71%> of non-CFS twins had used at least one alternative treatment. Although thought by some to represent an unrealistic search for a cure or a rejection of Western biomedicine, recent work suggests that the chronically ill believe C A M therapies may enhance conventional care strategies and provide a way for them to assume greater responsibility for their health (Thorne et al., 2002). CFS patients often rely on C A M treatments to reduce symptoms or prevent worsening of the illness (Afari et al., 2000; Bombardier & Buchwald, 1996). Contrary to the belief that the scientific naivete of the ill allows them to pursue unproven remedies (Beyerstein, 1997), it has been shown that patients engage in thoughtful and rigorous decision-making in order to assess and evaluate C A M interventions (Thorne et al., 2002). However, as opposed to physicians who rely on "gold-standard" evidence provided by RCTs, patients' decisions to adopt a remedy are generally based on anecdotal evidence and/or bodily cues (Thorne et al., 2001; Thorne et al., 2002). Further, instead of judging a treatment by how a group scores on standardized outcome measures, patients judge efficacy at the experiential level basing their evaluation of a treatment's usefulness on subjective (and difficult to measure) responses like mental outlook, healing potential, energy level, and feelings of balance (Thorne et a l , 2002). Patients often do not discuss their use of C A M therapies with conventional providers which may not be in the patient's best interests (Bombardier & Buchwald, 1996). Some C A M treatments are thought to have an adverse effect on health and/or may produce untoward effects when combined with conventional therapies. However, it is unclear how a conventional practitioner who is unfamiliar with C A M remedies would have sufficient knowledge to know when an interaction was possible (S. Thorne, personal communication, December 30, 2003). Participants in the Thorne et al. (2002) study reported a desire to disclose C A M treatments but were reluctant to do so if their provider held a negative attitude towards non-  28 conventional approaches. In a sample of family physicians (N=295) Bombardier and Buchwald (1996) found that about 70% to 90% believed counseling, behavioral medicine, and hypnotherapy to be to legitimate practices but did not hold positive views of Native American medicine, traditional Asian medicine, and/or homeopathy. The authors concluded that patients are generally not asked about their use of C A M therapies because physicians "lack the necessary interest or information to discuss this topic with their patients" (Bombardier & Buchwald, 1996, p. 104). Practitioners who are familiar with C A M therapies may enable patients to engage in a more critical analysis of information obtained from alternative sources (Thorne et al., 2002). However, additional information is needed about the types and perceived effectiveness of C A M practices utilized by CFS patients. The successful integration of C A M and conventional modalities in which credible information may be pursued and potential interactions between therapies may be detected is dependent upon the establishment of an open and respectful dialogue between the patient and a knowledgeable health care practitioner (Thorne et al., 2002). Financial Hardship Patients experience considerable financial hardship due to CFS (Anderson & Ferrans, 1997). In an Australian sample, Hickie and Pender (1992) calculated the cost per annum for each patient as $9436, with $1936 in direct costs and $7500 in lost wages. Expenses may also be incurred through the use of alternative treatments since these services are rarely covered by medical insurance (Afari et a l , 2000). In an American study, 12% of a CFS cohort reported spending more than $5000 on health care compared to 3% of the general medical care cohort (Twemlow, Bradshaw, Coyne, & Lerma, 1997). Over time CFS causes many patients to exhaust their savings, go into debt, relocate, and/or become financially dependent on friends or family. Often patients apply for disability pensions but find that the mental strain of the application process exceeds their limited energy  29 resources. Anderson and Ferrans (1997) reported that 18% of their sample experienced relapses as a result of trying to obtain disability benefits. In conclusion, CFS exerts a profoundly negative impact on numerous aspects of a patient's life and can lead to multiple losses including work, friendships, hobbies, and financial security. Indeed the intensity and accumulation of losses probably violates patients' sense of identity. Asbring's (2000) sample of CFS and fibromyalgia patients reported that the illness forced them to move from an identity of high activity to one of inactivity. Some patients have reported that "the losses and changes imposed by CFS were so complete that it often required a redefinition of oneself (Anderson & Ferrans, 1997, p. 364). Lives are described as coming to a complete standstill,-disrupted in all areas by an illness which is poorly understood and for which there is no cure. It is not surprising that patients have considerable difficulty accepting CFS and that the illness is associated with profound psychological distress (Bergmann & Stefansson, 1997). Anxiety and depression are the most common emotional disorders experienced by medical patients (Moorey et a l , 1991). A key factor in determining the emotional impact of chronic illness may be the nature of the symptoms (Ray et a l , 1992). Certain kinds of symptoms may be more related to affect than others. Weakness and fatigue have been found to be relatively highly correlated with negative affectivity, while symptoms such as musculoskeletal symptoms are not so clearly linked (Persson & Sjoberg, 1987; Vassend, 1989). Anxiety and depression have frequently been reported in CFS patients. About 50%> of CFS patients experience depression (David, Wessely, & Pelosi, 1991). Anxiety was found to be the predominant affective state among 55 patients who had experienced Akureyri disease (CFS) in 1948 (Bergmann & Stefansson, 1997). Ray (1991) suggested that rather than reflecting despondency or psychiatric disturbance, emotional sequelae like anxiety and depression that  30 accompany a debilitating illness like CFS may be a reaction to the illness, reflecting a state of demoralization. Problem Statement CFS is associated with multi-system complaints that are associated with considerable morbidity, dysfunction, and psychological distress. Patients' woes are further compounded by difficulties accessing supportive health care. As a result CFS patients may seek out a variety of conventional and alternative health care practitioners in order to obtain a diagnosis and treatment. Such behavior represents an economic burden to the system and the patient. Despite the daunting challenge of living with CFS, a sizeable percentage of patients improve. Although certain cognitions and coping efforts have, been associated with adaptation a thorough characterization of the salient constructs, their relations with each other, and outcomes has yet to be attempted. In addition, with few exceptions (Ray, Jeffries, & Weir, al., 1997) research on coping in the CFS population has been limited to a single wave of data collection which precludes describing the process of adjustment over time. Purpose The aim of the study is to describe in a population of CFS patients how cognitions and self-management behaviors are related contemporaneously and over time to adjustment as characterized by functioning, psychological distress, and health care utilization. Research Questions 1.  Which cognitions and coping behaviors relate to health care utilization, psychological distress, and functioning?  2.  How are cognitions related to coping?  3.  What personological factors are related to cognitions?  4.  What coping strategies do patients find efficacious?  5.  Does coping have to be efficacious in order to be effective?  31 6.  Are there differences in health care utilization, psychological distress, and functioning for patients with a short duration of illness compared to patients with a long duration of illness?  7.  Do patients with CFS improve over time?  8.  Do patients change their coping strategies over time?  9.  Do patients' appraisals of coping efficacy change over time? Theoretical Framework The self-regulatory model of illness representations (Leventhal, Nerenz, & Steele, 1984)  has been used by CFS researchers to guide empirical investigations of cognitions and coping (Moss-Morris, 1997). The model proposes that in response to health threats an individual develops parallel emotional and cognitive representations of illness. Cognitive representations address the objective features of the illness and are comprised of five components or perceptions; (a) illness identity (the label placed on the illness and the symptoms associated with it), (b) causes (ideas about how one gets the illness), (c) timeline (expectations about the duration of the illness), (d) consequences (the expected outcome and sequelae of the illness), and (e) controllability/cure (beliefs about the extent to which the illness is amenable to control or cure). Emotional representations reflect the individual's emotional reaction and serve to regulate the emotional responses prompted by the illness (Moss-Morris & Petrie, 2000). Cognitive and emotional representations of illness are based on patients' experiences and are thought to shape patients' self-management efforts. While each of the five cognitive dimensions has been shown to have implications for a wide range of illness-related behaviors including disability, psychological distress, and the seeking and evaluating of health care (Baumann, Cameron, Zimmerman, & Leventhal, 1989; Cooper, 1998; Heijmans & de Ridder, 1998a), the role emotional representations play in CFS has yet to be described. To operationalize self-management efforts and ensure consistency with earlier work it may be useful to borrow concepts from the cognitive model of coping which is considered  32 compatible with the self-regulatory model (Park, 1994). In the coping model the individual appraises; (1) the degree of threat implicit in the situation and (2) the personal resources available to manage the threat. Based on these appraisals coping may be elicited and social support mobilized. Coping functions to deal with situational demands deemed taxing or exceeding the individual's ability to adapt and may be categorized as problem-focused, which aims to change the stressor, or emotion-focused which is designed to control the emotional distress evoked by the stressor (Lazarus & Folkman, 1984). A coping strategy that has been found to be disadvantageous is catastrophizing (Diefenbach & Leventhal, 1996), which is characterized by negative self-statements and overly negative views about the future. Individuals who engage in catastrophic thinking unrealistically assume that the worst possible outcome will occur, given the current situation (Keefe, Brown, Wallston, & Caldwell, 1989). In addition to coping, individuals also draw on the resources of others by seeking social support, which refers to comfort, esteem, caring, or help received from others (Gentry & Kobasa, 1984). Social support may be differentiated as to type. Perceived support is defined as the cognitive appraisal of the availability of support while received support is the extent to which others enact behaviours intended to assist the individual (Finch, Barrera, Bryant, Pool, & SnowTurek, 1997). Although thought to be integral to the adjustment process, coping and social support have demonstrated relatively modest associations with outcomes. Theorists have suggested that to have an impact on outcomes psychosocial behaviours must be successful to some degree. Therefore in accordance with current thinking and congruent with the self-regulatory model, it is proposed that perceived efficacy mediates relations between coping, social support, and outcomes. A useful modification to the model consists of specifying outcomes of the self-regulatory process that reflect adjustment. In the coping literature adjustment generally refers to adaptive  33 mental functioning or psychological well being. However, theorists have expanded the concept to include dimensions related to the ability to carry out normal physical arid psychosocial activities (Jensen, Turner, Romano, & Karoly, 1991). Attention must also be paid to the social costs of CFS. The inexplicable nature of the illness has implications for the excessive use of medical services (Buchwald & Garrity, 1994). Since there is no effective treatment for CFS, patients probably do not benefit from frequent health care visits, which may burden the system unnecessarily. Thus, outcomes considered reflective of adjustment which are included in the framework are anxiety, depression, functioning, and the number of visits to health care providers (see Figure 1).  34 Figure 1 Theoretical Framework  I L L N E S S  COPING 1. Problem-focused 2. Emotion-focused  R E P R E S E N T A T I O N  E F F I C A C Y  FUNCTION  DEPRESSION  ANXIETY  H E A L T H CARE VISITS  SOCIAL SUPPORT 1. Perceived support 2. Received support  CATASTROPHIZE  Chapter Summary Chapter one described CFS and its effects on sufferers. Patients'difficulties obtaining supportive care from health care practitioners were detailed. In addition, the problem underlying the research, its purpose, and the questions being addressed were offered. The theoretical framework guiding the study was described. In the next chapter a review of the literature will be presented.  35 C H A P T E R TWO: Review of the Literature The Cognitive Behavioral Perspective of Illness The success of the medical model in advancing physicians' ability to treat disease has led to an emphasis on the distinction between disease and illness. Disease is defined as a physiological event characterized by pathology in structure and/or function of body organs. The manifestations of a disease may be objectively ascertained as physical signs by the senses of an observer (Aronowitz, 2001). Illness, on the other hand, reflects a sufferer's perceptions, thoughts, and reactions to disvalued changes in well-being. Such experiences manifest as the subjective perception of symptoms, which may occur without any physical signs of disease and which may not be observable to others (Schober & Lacroix, 1991). A major shortcoming of the medical model is its inability to deal with medicallyunexplained illnesses, symptom-based conditions that do not demonstrate objectively verifiable signs of disease. The medical model concentrates on biological processes and fails to consider patients' phenomenological experiences of illness and the impact of such experiences on outcomes (Diefenbach & Leventhal, 1996). Unlike the medical model, the cognitive behavioral model focuses on patients' non-physiological responses to illness, which may provide insight into factors that predispose and prolong a medically-unexplained condition like CFS. Wessely et al. (1991) proposed the formative CFS cognitive behavioral model hypothesizing that while organic factors like an infection or trauma may precipitate CFS, cognitive and behavioral factors prolong the condition. For instance, following an acute illness or stressful episode when individuals resume their normal activities symptoms caused by physical deconditioning may be experienced. If these symptoms are attributed to ongoing disease the individual may attempt to cure the symptoms by restricting activity. Future efforts to engage in activity will be met by more symptoms of deconditioning. A vicious cycle of avoidance,  36 deconditioning, loss of control, and emotional distress ensues which generates even more symptoms (Wessely et al., 1991). The cognitive behavioral model was expanded by researchers who described how certain individuals may be predisposed to CFS (Surawy, Hackmann, Hawton, & Sharpe, 1995). Confronted by an illness that limits their ability to perform, achievement-oriented individuals force themselves to press on with their usual activities which often results in exhaustion. In an effort to make sense of the situation and protect their self-esteem, individuals may attribute exhaustion to physical disease, which motivates them to rest. Eventually, the reduction in activity conflicts with the individual's need to achieve. The individual attempts to meet his/her performance expectations through extravagant bursts of activity, further exacerbating symptoms, and contributing to the idea that a serious illness is being experienced. Cognitive Behavioral Therapy (CBT), which is based on the cognitive behavioral perspective, has shown promise for improving functioning and fatigue in randomized, controlled trials of CFS patients (Raine, Haines, Sensky, Hutchings, Larkin, & Black, 2002; Whiting et al., 2001). CBT is a technique borrowed from the field of chronic pain in which the therapist and patient collaborate to identify and modify cognitions and behaviors counterproductive to the patient's recovery. Goals of therapy include; (1) altering beliefs that prolong disability, (2) introducing a predictable, consistent pattern of rest and activity, and (3) gradually increasing activity and decreasing rest (Sharpe, 1994). Limitations of CBT Critics have argued that the success of CBT in clinical trials may be somewhat over-rated because investigators have used case definitions that select subjects inherently more likely to improve. Critiquing a review of CFS treatments written by Whiting et al. (2001), Carruthers et al. (2003) pointed out that all the CBT studies "used the less restrictive Oxford criteria with the exception of one study (Prins et al., 2001) which used C D C criteria for prolonged fatigue but  37 eliminated other C D C criteria" (p. 48). As discussed in chapter one, the prognosis for improvement, regardless of treatment, is better when less restrictive criteria are used for patient selection. Moss-Morris (1997) opined that CBT "has largely been based on frameworks drawn from clinical observation and the underlying mechanisms involved have not been empirically confirmed" (p. 417). Moreover, as reflected in CBT's major aim which is to increase activity by encouraging patients to question their perception that symptoms are the result of a physical disease, the therapy has focused primarily on causal beliefs (Vercoulen et al., 1996). CBT's emphasis on improving physical functioning may have inadvertently diverted attention from identifying ways patients adapt emotionally to CFS (Ax, Gregg, & Jones, 2001). In order to "flesh out" and connect notions introduced by the cognitive behavioral model it may be helpful to examine a broader range of illness perceptions and determine their relations with important outcomes. Illness Perceptions It has long been suspected that an individual's beliefs can play a pivotal role in illness causation and its remedy. The most startling examples are cases of voodoo death and the placebo phenomenon (Hahn & Kleinman, 1984). Lay theories have also been associated with entry into the health care system, the evaluation of treatment, and the way individuals interpret and cope with illness-related events (Scharloo & Kaptein, 1997). Examination of illness models is a relatively recent development that may be considered part of a broader scientific inquiry into the role of cognition in illness behavior. Commonly researchers in the field conceptualize illness models as generic and organized cognitive representations of schemata that originate from prior experiences in the medical domain and guide information processing in a way that is consistent with prior knowledge (Schober & Lacroix, 1991). The following review of the literature will demonstrate that lay beliefs or illness  38 perceptions have been examined in the CFS population using a variety of theoretical frameworks and methodologies. Perceptions of Causal Attribution in CFS: A Double-edged Sword? A n early and consistent finding in the literature has been CFS patients' attribution of symptoms to physical factors, especially viruses (Clements, Sharpe, Simkin, Borrill, & Hawton, 1997; Wessely & Powell, 1989; Sharpe et al., 1992). However, this type of causal attribution has not been found to be a distinctive feature in samples drawn from community and primary care settings (Euba et al., 1996). Instead, patients from these settings espoused more interactive models of causation that included biological, psychological, and/or social factors (Chalder, Powell, & Wessely, 1996; Euba et al., 1996; Lawrie et al., 1997). Patients' causal attributions may differ as a function of help seeking. Compared to patients recruited from the community, tertiary-care patients reported greater symptom severity, higher socio-economic status, and were more likely to have self-diagnosed CFS (Torres-Harding et al., 2002). Such factors may increase the chances that a patient will present to a general practitioner and be referred onward to specialists, further reinforcing the idea of a physical cause (Lawrie et al., 1997; Torres-Harding et al., 2002). Additional factors that may contribute to patients' attributing symptoms to a physical illness include applying for disability (which entails specialist referrals and laboratory investigations), seeking information from the lay media, and exposure to self-help groups (Lawrie et al., 1997). Thus, it would appear that the tendency to attribute CFS to physical factors might be related to personal and social experiences, most notably ongoing contact with the health care system. Although it is not possible to judge the "correctness" of patients' causal attributions since the etiology of CFS is unknown "it is still valid to consider the differences in attributions and the consequences of such differences ..." (Powell, Dolan, & Wessely, 1990, p. 670). A risk factor for developing CFS after a viral infection may be patients' preference for attributing common  39 symptoms to physical factors (Cope, Mann, & Pelosi, 1996). Robbins and Kirmayer (1991) hypothesized that people attend to physical sensations differently and some are more inclined to somatize or attribute symptoms to a physical process. Such a process may allow an individual to use bodily complaints as a mask or substitute for psychological difficulties (Abbey & Garfinkel, 1991). Compared to fracture clinic patients CFS patients were more likely to make somatic, rather than normalizing attributions for symptoms (Butler, Chalder, & Wessely, 2001). Cross-sectional and prospective CFS studies have linked the tendency to make physical attributions to prolonged illness states characterized by increased symptom reporting (Cathebras, Jacquin, Le Gal, Fayol, Bourchou, & Rousset), impairment (Butler et al., 2001; Chalder et al., 1996; Sharpe et al., 1992), and lack of improvement (Vercoulen et al., 1996). Joyce et al. (1997) suggested that believing symptoms are caused by physical disease may be "predictive of avoiding exercise or reducing other activities which might prolong disability" (p. 232). Investigators reported that CFS patients' beliefs that exercise was unsafe or contraindicated by their condition explained 12% of the variance in exercise performance (Silver, Haeney, Vijayadurai, Wilks, Pattrick, & Main, 2002). In a study using structural equation modeling it was found for CFS patients that attributing symptoms to a somatic cause predicted physical activity level which in turn predicted fatigue (Vercoulen et al., 1998). Such a path could not be supported for patients with multiple sclerosis. Certain causal attributions may also be associated with compromised coping resources. Powell et al. (1990) surmised that patients who thought CFS was caused by an untreatable physical factor were vulnerable to feelings of helplessness, lack of selfefficacy, and diminished responsibility for their own health. Although physical attributions have been related to disability they may serve a protective function (Chalder et al., 1996; Wessely & Powell, 1989). In a study comparing hospital and primary care patients with CFS, 33% of the primary care cases but 56% of the hospital patients attributed their fatigue to physical causes. Hospital patients reported greater fatigue, poorer role  40 functioning, but less psychological distress than primary care patients. Certain advantages may accrue from attributing CFS to a physical cause. The explanation is simple, easy to understand and may protect self-esteem by absolving sufferers of guilt, blame, and stigma (Chalder et al., 1996; Powell et a l , 1990; Cope et al., 1994). In summary, it can be seen that due to the lack of knowledge regarding the etiology of CFS, patients are left to generate their own models of causation. These causal models may be complex and probably have differential relations with outcomes. Moreover, causal attributions likely evolve over time and are associated with personological factors and contact with the health care system. In light of the strong association between attributions and adjustment, additional work is needed to describe the determinants and mode of action of causal attributions. Perceptions of Illness in CFS: A Type of Abnormal Illness Behavior? Because CFS was associated with co-morbid psychiatric disorders and displayed no objective signs of organic disease, some researchers interpreted CFS as a form of abnormal illness behavior (Abbey & Garfinkel, 1991). Abnormal illness behavior (AIB) is defined as a maladaptive mode of responding to one's health based on beliefs that are inconsistent with the appraisal of the consulting physician (Pilowsky, 1969). Similar to hysteria or somatization, AIB is a broad concept that includes abnormal bodily perceptions, an inaccurate evaluation of health state, and inappropriate action in relation to these perceptions and evaluations. Studies of AIB Using a measure developed by Pilowsky and Spence (1975), the Illness Behavior Questionnaire (IBQ), Hickie et al. (1990) compared CFS patients (N=48) with patients with somatoform disorder (N=48) and found that CFS patients had high scores on the disease conviction subscale (affirmation that physical disease exists, preoccupation with symptoms, and rejection of a physician's reassurance) and low scores on the psychological versus somatic concern subscale (a bipolar factor referring to the attribution of illness to either psychological  41 {high scorejor physical causes {low score}). In a follow-up study of 139 subjects previously enrolled in two treatment trials Wilson et al. (1994) reported that disease conviction was more predictive of disability than immunological or demographic variables. Schweitzer et al. (1995) found that CFS patients (N=40) had high scores on a measure of general hypochondriasis (anxious concerns about the state of one's health). However, when CFS patients (N=98) were compared with patients with multiple sclerosis (MS; N=78), similar IBQ scores were found for both groups (Trigwell et al., 1995). The researchers concluded that scores on the IBQ could not be taken "as evidence that CFS is a variety of abnormal illness behavior because the same profile occurred in multiple sclerosis" (Trigwell et al., 1995, p. 15). Byrne (1975) questioned the validity of the IBQ suggesting the items reflected neuroticism, a trait known to be related to excessive medical complaints. In a small sample of coronary heart disease patients (N=15), Byrne (1975) correlated scores between a measure of neuroticism, Eysenck's Personality Inventory (EPI) and the IBQ. For four of the seven IBQ subscales, Pearson's r ranged from .46 to .52. In a sample of 1061 health-care and non healthcare seeking subjects, the convergent validity of IBQ factor scores using the EPI was found to be between .25 and .50 (Zonderman, Heft, & Costa, 1985). Zonderman et al. (1985) concluded that IBQ items were "saturated" with neuroticism. In addition to concerns about the validity of the IBQ critics decry the patriarchal nature of AIB which locates the "pathology" in the patient. Key assumptions of the model are that the physician is correct in his/her appraisal that the patient does not have a disease and that the patient errs by not accepting the physician's evaluation. However, the cause of CFS has not been established and researchers' inability to identify organic pathology in CFS does not mean the patient is disease free. The uncertainty regarding the etiology of CFS precludes characterizing patients' perceptions as abnormal illness behaviors (Zonderman et al., 1985).  42 If one disregards the assumption that CFS is an abnormal illness behavior, elevated IBQ scores on the disease conviction and general hypochondriasis subscales may be interpreted to mean that CFS patients' perceive their symptoms to be severe and the consequences of the illness distressing. Further, the findings suggest that the belief system of a CFS patient is strong since it is adhered to regardless of countervailing medical opinion. Due to the association between the IBQ subscales and disability, and the implications of patients' belief system for the patient-physician relationship, further inquiry into patients' illness perceptions is warranted. Perceptions of Illness in CFS: The Patient as a Lay Scientist? Although the preceding discussion affirms the predictive utility of patients' beliefs, criticism must be launched at the atheoretical or flawed models from which the findings were derived. A more acceptable framework for assessing illness perceptions may be found in Leventhal's self-regulatory model (Leventhal et al., 1984). The self-regulatory model makes no judgments about the medical accuracy of illness representations but conceptualizes them as attempts by the individual to give meaning to the experience of being ill. The model proposes that individuals respond to illness threats cognitively and emotionally which determine the type of coping strategy selected. Appraisals of the efficacy of coping strategies may be made and used to modify coping strategies and illness representations. Cognitive illness representations consist of five interconnected components or perceptions which address key questions individuals ask when faced with an illness; (1) what is it? (identity), (2) what caused it? (cause), (3) how will it affect me? (consequences), (4) how long will it last? (timeline), and (5) can it be controlled or cured? (controllability/curability). The consistency and validity of the five perceptions have been confirmed in a number of studies using a variety of clinical populations and different methodologies (Bishop, 1991; Cooper, 1998; Leventhal & Diefenbach, 1991). The unique etiologic, diagnostic, and management features of  43 CFS make the study of illness perceptions of relevance for theorists and therapists working within the context of the cognitive behavioral framework. 1. Identity Often, in the early stages, illnesses reveal themselves as symptoms. The individual is prompted to use prior experiences and knowledge to interpret and provide a label for the somatic change. The illness identity is the name an individual places on the somatic change and the symptoms associated with it. Establishment of an illness identity is the core around which the other perceptions develop. That is, one must first give the illness a name before one can consider the cause, consequences, timeline, and controllability of it. The difficulties associated with diagnosing CFS make "identity" an important perception to understand. The medically inexplicable nature of the illness and the lack of a biomarker ensure that diagnosis rests on the patient's self-report of subjective symptoms. Controversy continues regarding the symptoms that are specific to CFS. Diagnostic criteria and articles published in the lay media include numerous and diverse symptoms as part of the syndrome. Just as individuals will seek out a label when symptoms are experienced, individuals with an illness label will search for symptoms consistent with the label (Baumann et al., 1989). Thus, CFS patients may be highly vulnerable to interpreting all bodily sensations and neuropsychological and emotional responses as part of the illness, thereby increasing the symptom load that must be managed (Moss-Morris, 1997). A linear correlation has been found between the number of medicallyunexplained symptoms, functional disability, and psychological distress (Katon & Russo, 1992). 2. Cause The onset of an illness, particularly i f symptoms are severe and sudden, is a frightening experience. Individuals confronted with an unexpected negative event usually attempt to identify the cause, thereby making the experience less anxiety provoking. Perceptions of cause include an individual's ideas about how one gets the illness. Early research on causal attributions and illness  44 behavior suggested most CFS patients attributed the cause to a disease process (Clements et al., 1996; Hickie et al., 1990). However, it is now thought that patients' causal attributions are more complex. Although it is difficult to determine the veridicality of patients' causal beliefs, the relations between certain types of causal attributions and outcomes are a matter of record. 3. Consequences This dimension of the self-regulatory model reflects a patient's ideas about the short- and long-term effects of the illness. The severe, unpredictable symptoms associated with CFS likely exert a negative impact on all facets of patients' lives. As well, the stigma attached to the illness is considerable and constitutes an additional burden sufferers must bear. Studies using the IBQ indicate patients experience considerable distress as a result of CFS (Hickie et a l , 1990). 4. Controllability Individuals develop certain ideas about what they themselves or health care providers can do to manage the illness. The medically inexplicable nature of CFS and the unpredictability of the symptoms may lead patients to believe that little can be done to control the condition resulting in feelings of helplessness, psychological distress, and the use of ineffective coping behaviors (Chalder et al., 1996; Felton et al., 1984; Findley et al., 1998). The importance of attaining a sense of control was demonstrated by Vercoulen et al. (1996) who found control was a strong predictor of improvement in an 18-month longitudinal study (N=298). A clearer understanding of the determinants of control and its mode of action is needed. 5. Timeline Timeline concerns patients' expectations about the duration and expected course of the illness. Due to the lack of understanding about CFS, few physicians can tell a patient how long the illness will last, a factor that may cause uncertainty and distress. As well, the disconcerting, episodic "on-again/off-again" nature of CFS symptoms has implications for patients' appraisals of coping efficacy and outcomes.  45 Nerenz and Leventhal (1983) postulated that patients formed different models depending on perceptions of the illness' timeline including; (a) the acute episode model - short-term treatment followed by cure, (b) cyclic model - symptoms subside and then reappear, and (c) the chronic model - symptoms need continuing, long-term care. Patients are thought to shift from acute though cyclic to chronic representations of timeline based on the stage of illness. The perception that the illness is chronic has been associated with greater disability and depression (Moss-Morris, Petrie, & Weinman, 1996). Illness Representations in CFS Patients A major contribution to the study of illness representation was the development of the Illness Perception Questionnaire (IPQ; Weinman et al., 1996). Theoretically derived from the self-regulatory model and psychometrically sound, the IPQ quantifies the five illness perception components. Weinman et al. (1996) demonstrated that CFS patients hold highly negative perceptions but such negativity does not appear to be a consequence of the illness' chronicity. Compared to patients with diabetes, chronic back pain, and rheumatoid arthritis CFS patients had the highest identity scores (perceived a large number of symptoms) and consequences scores (the perception the illness has very serious consequences). CFS patients were also most likely to attribute the illness to a virus or pollution and least likely to attribute the illness to their own behavior (Weinman et al., 1996) Relations between Illness Representations and Outcomes in CFS Illness perceptions have been shown to explain considerable variance in outcomes. Using the IPQ two cross-sectional studies investigated the relations between perceptions, psychological distress, functioning, fatigue, and vitality (Edwards, Suresh, Lynch, Clarkson, & Stanley, 2001; Moss-Morris et a l , 1996). The studies differed in how CFS was defined (Oxford versus C D C criteria), the method of sampling (random selection versus convenience), and the instruments used to measure the dependent variables. However, multiple regression analyses generated  46 similar findings. The combined five perceptions were found to predict fatigue (Edwards et al., 2001; N= 126) and vitality (Moss-Morris et al., 1996; N=233) with adjusted R of .14 and .42, 2  respectively. Perceptions also explained a sizeable degree of variance in psychological adjustment and produced similar effect sizes (IV of about .30). Some perceptions were found to be better predictors than others. Identity was the most salient and shown to be a statistically significant predictor of adverse outcomes with standardized beta coefficients ranging from .19 to .50. Consequences produced greater effect sizes and was associated with more outcomes in the Edwards et al. (2001) study, which reported standardized betas of .30, .33, and .22 for fatigue, depression, and anxiety, respectively. On the other hand, Moss-Morris et al. (1996) found that consequences produced a significant beta of .18 for the dysfunction variable, only. This inconsistency may be due to sampling differences. Edwards et al. (2001) enrolled tertiary-care patients while Moss-Morris et al. (1996) recruited from the community. Perceptions of the consequences of CFS may be less dire, and generate weaker associations with outcomes in a community-based sample. Of interest was the researchers' inability to demonstrate that attributing the cause of CFS to physical factors was predictive of disability. Moss-Morris et al. (1996) suggested that identity and consequences may be such strong predictors that they limit the amount of variance available to other variables like attributions of a physical cause. However, consistent with earlier work, both studies linked attributions of an emotional or psychological nature to psychological distress. Although thought to be an important factor in chronic illness, control was not related to psychological adjustment in the Moss-Morris et al. (1996) study but found to be negatively related to psychological distress producing betas of -.20 for depression and -.17 for anxiety by Edwards et al. (2001). Again sampling differences may be responsible for the inconsistency. The perception of control may be more relevant and generate stronger associations with outcomes for  47 patients in tertiary care. Making an effort to obtain highly specialized care suggests that these patients have a greater need to obtain treatment that will control their symptoms. In general, the studies support the existence of an association between illness perceptions and outcomes. The findings also indicate that the perceptions of CFS patients are characterized by high negativity. It appears that negative perceptions are more predictive than positive perceptions, suggesting that pessimistic perceptions may be more strongly related to outcomes. Overall, illness representations provide a cogent way of categorizing and relating beliefs to important outcomes. Some may argue that perceptions, particularly identity, reflect the severity of the illness, and it is the severity of the illness rather than the individual's perception that predicts outcomes. However, this argument may be disputed for the following reasons. One, there is a considerable discrepancy between sufferers' subjective reports of illness severity and objective measures. Two, no known indicators of severity have been found for CFS. And three, evidence from other studies on chronic illness suggest that perceptions are stronger predictors of disability than objective measures of disease severity (Jensen et al., 1991; Moss-Morris et al., 1996). Unfortunately, methodical limitations diminish the utility of extant findings. Due to limitations associated with the study design, the findings cannot be used to conclude that perceptions result in psychological distress and fatigue. A plausible, alternative conclusion may be that psychological distress and fatigue produce pessimistic illness perceptions. Furthermore, the lack of association between physical attributions and outcomes requires greater explication, departing as it has from reports from other researchers. As well, it would be helpful to determine if causal attributions are related to health care utilization. Although it has yet to be studied in CFS patients, evidence gathered from a non-clinical population demonstrated that individuals who attributed their symptoms to psychogenic factors were less inclined to visit a physician than individuals who attributed their symptoms to organic factors (Bishop, 1987).  48 Hypotheses Testing the following hypotheses will provide information about the types of perceptions associated with outcomes; 1.  Pessimistic illness perceptions predict depression, anxiety, and decreased functioning.  2.  Perceived control predicts decreased levels of depression and anxiety.  3.  Attributing the cause of CFS to non-physiological factors predicts depression and anxiety.  4.  Attributing the cause of CFS to physical factors predicts greater health care utilization. Coping In addition to their direct association with outcomes, illness perceptions are thought to be  related to individuals' coping efforts. In the context of illness, Lipowski (1970) defined coping as "cognitive and motor activities which a sick person employs to preserve his bodily and psychic integrity, to recover reversibly impaired function and compensate to the limit for any irreversible impairment" (p. 93). Researchers have characterized coping in different ways. A trait approach has been adopted by some who defined coping as a dispositional characteristic consisting of habitual problem-solving thoughts and actions (Costa, Somerfield, & McCrae, 1996). However, most chronic illness researchers have utilized the general coping model which treats coping as a state. As such, coping may be defined as "constantly changing cognitive and behavioral efforts to manage the specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person" (Lazarus & Folkman, 1984, p. 114). Coping is thought to buffer or moderate, the negative effects of stress on an individual's psychological and physical health (de Ridder etal., 1998).  49 Types of Coping Since the potential number of coping strategies is infinite, an attempt has been made to parsimoniously categorize them. Folkman and Lazarus (1985) suggested two functions of coping. Problem-focused coping is aimed at changing the situation causing the stress while emotion-focused coping is designed to control the emotions caused by stress. However, the distinction between the two modes cannot be considered entirely clear-cut since successful resolution of the stress using a problem-focused strategy often reduces emotional distress (Katz, Ritvo, Irvine, & Jackson, 1996). Another important distinction is an individual's orientation and activity in response to a stressor. Approach coping describes an individual's active efforts to invest energy in addressing the problem. Avoidance is a passive response to stress in which an individual refrains from investing effort in trying to cope (Katz et al., 1996). Yet another distinction emphasizes the method of coping individuals use, that is whether the response is cognitive or behavioral (Holahan, Moos, & Schaefer, 1996). Cognitive coping includes efforts of an intra-psychic nature to deal with stress while behavioral coping emphasizes active responses to manage the situation. Although no consensus has been reached on which distinction is preferable, it has been suggested that combining the dimensions of problem-/ emotion-focused coping and avoidance/approach "is a good starting point for categorizing coping behavior" (de Ridder, 1997). The Adaptiveness of Coping Viewed from the perspective of the coping model adaptive is defined as the effectiveness of coping in improving adjustment (Folkman & Lazarus, 1994). Coping researchers generally expect problem-focused coping to be more beneficial to well-being than emotion-focused coping (Thoits, 1995). However, ascertaining whether a particular strategy is adaptive requires an examination of situational factors (Zeidner & Saklofske, 1996). In general, problem-focused  50 coping is related to fewer psychological symptoms in situations appraised as controllable. On the other hand, emotion-focused coping is associated with lower symptom levels in situations appraised as uncontrollable (Aldwin, 1994). Chronic illness may lead patients to believe they lack control, which suggests that emotion-focused strategies would be a more effective mode of coping. Indeed, many studies of the chronically ill have found that problem-focused strategies are not highly correlated with outcomes. The adaptiveness of a coping strategy may also vary depending on the acuity or brevity of a stressor. Avoidance strategies such as inattention or distraction may be adaptive when one is confronted with pain or uncomfortable medical procedures. In the long run, though, avoidance has been related to distress and the reduced adherence to medical requirements (Sullivan, Edgley, Mikail, Dehoux, & Fisher, 1992). Coping's adaptiveness also depends on whose perspective is being considered. Coping that is considered adaptive by the individual may be burdensome to society. The tendency to use active behavioral strategies that may include help seeking has been linked to increased health care utilization (Sullivan et al., 1992). Such a strategy would be considered adaptive in the acutely ill because it might lead to early diagnosis and treatment but may be considered maladaptive in the chronically ill for whom help seeking would not be expected to impact on illness severity or course (Sullivan et al., 1992). The timing of coping strategies may be related to adaptation. Since chronic illness makes different demands at different times, an individual's cognitions and behaviors may pass through qualitatively different phases during the illness experience (Cassileth et al., 1984; Weinstein, Rothman, & Sutton, 1998). It is likely that in order to be adaptive an individual's coping strategy must be congruent with the illness phase. Some studies have shown that chronically ill patients utilize multiple strategies and that it is the pattern of coping, rather than the action of individual strategiest that is more relevant to  51 adaptation. To be effective a coping strategy should be matched and capable of neutralizing a specific stressor. Individuals who are flexible in their choice of coping tend to adapt better than persons who have more restricted or rigid coping repertoires (Miller, 1990). Thus, a broad coping repertoire, consisting of both problem- and emotion-focused strategies invoked at the appropriate moment increases the likelihood that a response will be matched to a particular demand, thereby having an effect on resolving the stress (de Ridder & Schreurs, 2001). Coping Research in CFS CFS researchers have focused on three issues; (1) coping differences between CFS patients and other populations, (2) relations between coping and outcomes, and (3) changes in coping over time. Group Comparisons Although studies comparing CFS patients with other groups have used a variety of coping instruments some consistent findings have emerged. Differences in coping were found between CFS patients and patients with the medically-explained illness, Parkinson's disease (PD). For CFS patients the patterns of correlations between adaptive tasks, coping, and quality of life were much stronger. However, in the case of PD, adaptive tasks did not predict coping or quality of life. Disease characteristics like symptoms and duration were better predictors, possibly, because of the strong constraints PD imposed on coping and quality of life. The authors suggested that CFS patients may enjoy greater situational power which allows for more subjective evaluations. Thus, CFS patients' quality of life may be more strongly related to their cognitions of the disease and the actions associated with the cognitions (de Ridder et al., 1998). Few gender differences in coping have been found in CFS, suggesting that men and women experience similar difficulties (for example, disability, unemployment) and utilize comparable coping behaviors (Buchwald, Pearlman, Kith, & Schmaling, 1994). As well, researchers reported CFS patients with or without major depressive disorder utilized similar  52 coping strategies (Antoni et a l , 1994; Moss-Morris, 1997). Finally, no differences in emptionand problem-focused coping were found between patients with idiopathic chronic fatigue (N=33) and chronic fatigue syndrome (N=33). The limited number of subjects in each cell may have resulted in insufficient power to detect significant differences between the groups (Jason, Witter, & Torres-Harding, 2003). However, compared to controls and patients with chronic pain, it was found that CFS patients favored avoidant coping strategies, including behavioral disengagement, denial, and mental disengagement (Blakely, Howard, Sosich, Murdoch, Menkes, & Spears, 1991). Limiting coping was the strategy that differentiated CFS patients from patients with major depressive disorder (Moss-Morris, 1997). The use of avoidant coping has been a consistent finding in the CFS population and evidence suggests this type of coping is associated with fatiguing conditions (Afari et al., 2000; Sharpe et al., 1992). In a co-twin study that controlled for genetic and environmental effects, fatigue-discordant monozygotic twins were compared and found to cope very similarly with the exception of avoidance. Twins with chronic fatigue utilized more avoidance that their non-fatigued counterparts (Afari et a l , 2000). Findings from another study suggested avoidant strategies were utilized more after the onset of CFS. CFS patients reported using confrontational coping pre-illness but switched to more avoidant strategies after developing CFS (Lewis, Cooper, & Bennett, 1994). Such an alteration in coping patterns was not found for healthy controls or patients with irritable bowel syndrome (Lewis et al., 1994). Relations between Coping and Outcomes A cognitive strategy shown to be predictive of adverse outcomes is catastrophizing, a way of thinking characterized by "negative self-statements and overly negative thoughts and ideas about the future" (Keefe et a l , 1989, p. 51) that may include elements of rumination, magnification, and helplessness (Sullivan, Rodgers, Wilson, Bell, Murray, & Fraser, 2002). Researchers asked CFS subjects what consequences they would experience if they .pushed  53 themselves beyond their present physical state (Petrie, Moss-Morris, & Weinman, 1995). Subjects were coded as evidencing catastrophic thinking if their response was one of a highly negative and exaggerated nature. Analysis showed that catastrophizers experienced greater fatigue and dysfunction. The authors concluded that catastrophizers may perceive increasing activity as highly undesirable and were likely to avoid activity, a consequence that has been associated with deconditioning. In order to gauge the effects of problem- and emotion-focused coping most studies have used a cross-sectional design and sampled CFS patients with moderately long illness durations (6 - 1 0 years). Outcomes of interest have included psychological distress (anxiety and depression), functioning, fatigue, and vitality. For the most part, generic tools have been used like the Coping Orientation to Problems Experienced Scale (COPE; Carver, Scheier, & Weintraub, 1989), Ways of Coping Inventory (WOC; Folkman & Lazarus, 1988), and the Utrecht Coping Questionnaire (UCQ; Schreurs, van der Willige, Tellegren, & Brosschot, 1993). Emotion-focused coping was modestly predictive of outcomes while problem-focused coping appeared to be a poor predictor (Antoni et al., 1994; Heijmans, 1998; Moss-Morris et al., 1996). When emotion-focused coping was separated into approach and avoidance categories, avoidant strategies proved to be the better predictor. For American, British, and Dutch CFS patients tested with a variety of coping instruments disengagement and denial strategies were related to impairment, psychological distress, fatigue, and decreased functioning (Jason et al., 2003; Moss-Morris et al., 1996; Ray et a l , 1993; Ray et a l , 1995). Social functioning and psychological well-being demonstrated a consistently negative association with avoidance (Antoni et al., 1994; Heijmans, 1998a). To a lesser extent avoidance was negatively related to physical functioning (Antoni et al., 1994). One of the only emotion-focused strategies linked with favorable outcomes (greater psychological well-being) was positive reinterpretation, which  54 could be construed as an approach strategy since it describes efforts by the individual to view the problem at hand in a more positive manner (Moss-Morris et al., 1996). The weakness of prediction of problem-focused coping may be due to the reliance of researchers on generic instrumentation (Ray et al., 1993). Generic measures are probably satisfactory for measuring emotion-focused strategies. The emotional distress caused by an illness may share common features across different populations and may be managed using similar strategies (Costa, Somerfield, & McCrae, 1996). However, strategies intended to deal with the objective demands of an illness may be much more specific. Illness-specific measures may be needed in order to provide sufficient descriptive power to relate problem-focused coping to outcomes. To address this issue, Ray et al. (1993) developed a CFS-specific coping instrument. The Illness Management Questionnaire (IMQ) was designed to assess problem-focused coping and consists of four subscales; (1) maintaining activity (attempting to ignore symptoms and carrying on though unwell), (2) accommodating to the illness (organizing one's life to prevent stress and overexertion), (3) focusing on symptoms (preoccupation with symptoms and the perception of helplessness), and (4) seeking information (searching for information about CFS and the readiness to try remedies). In a CFS sample with an illness duration of 4 years, the IMQ predicted substantial variance in measures of impairment (26%), anxiety (27%), and depression (22%). The four subscales produced differential associations with outcomes. In addition, two of the subscales demonstrated that coping strategies might serve a "trade-off function. For example, maintaining activity predicted less impairment but greater anxiety. Conversely, accommodating to the illness was related to less psychological distress but resulted in more impairment. Focusing on the symptoms was uniformly linked to unfavorable outcomes including impairment, anxiety, and depression. Seeking information was positively correlated to  55 impairment. This pattern of findings was essentially replicated (N=147) in a cross-sectional analysis of data (Ray et al., 1995) from Ray et al.'s (1997) longitudinal study. Ray et al.'s (1993, 1995) work demonstrated the predictive utility of the IMQ but in order to model change over time the authors (Ray et al., 1997) conducted a longitudinal study in which 137 patients with illness durations of about 3.5 years completed the IMQ and were re-tested one year later. Controlling for prior status on outcomes by using residualized scores of fatigue and impairment the authors reported that seeking information predicted fatigue. However, maintaining activity and focusing on symptoms did not achieve statistical significance. Accommodating to the illness predicted impairment but only for subjects with a longer duration of illness. The authors interpreted this finding to mean that with increasing chronicity the individual tends to limit activity and is thus less able to function. Accommodating to the illness showed a positive association with fatigue and impairment but only for subjects with a low locus of control. Ray et al. (1997) surmised that under conditions of diminished control illness accommodation represented a strategy of disengagement that may be associated with unfavorable outcomes. However, under conditions of greater control accommodation may serve a goal-directed function. That is, the individual moderates activity with the expectation that recovery will occur. In addition, decreasing activity may temper symptoms thereby reducing the distress caused by the illness' unpredictability. Finally, illness accommodation may give patients a rationale for removing themselves from stressful situations (Moss-Morris et al., 1996). Clearly the challenge is to identify adaptive and maladaptive coping strategies in the CFS population. Achieving this goal may necessitate that investigators treat coping as a multidimensional construct consisting of problem- and emotion-focused components. Analyses may be further fine-tuned by including patients' perceptions of control and differentiating between avoidant and approach orientations in emotion-focused coping. Moreover, investigators would be  56 well advised to broaden the scope of coping behaviors to be studied. A noteworthy addition would be coping characterized by catastrophizing since this strategy is strongly linked to impaired functioning. Finally, information is needed about how coping relates to health care utilization. In a group of patients with multiple sclerosis and chronic pain, investigators observed that patients who actively coped by planning and initiating action toward problem resolution reported a higher frequency of visits to health care professionals (Sullivan et al., 1992). Ray et al. (1993) noted a high correlation between the active coping and planning subscales of the COPE and the seeking information subscale of the IMQ. It may be useful to determine i f seeking information is also related to increased health care visits. Hypotheses Based on the preceding discussion the following hypotheses have been formulated to identify coping strategies related to outcomes: 1.  Problem-focused coping strategies exhibit differential associations with outcomes as follows; a.  Accommodating to the illness predicts lower levels of depression and anxiety.  b.  Maintaining activity predicts higher levels of depression, anxiety, and functioning.  2.  The problem-focused strategy of focusing on symptoms predicts increased depression, anxiety, and poorer functioning.  3.  The problem-focused strategy of seeking information predicts increased health care utilization.  4.  Emotion-focused coping demonstrates differential associations with outcomes depending on whether the orientation of the strategy is one of avoidance or approach. a.  Avoidance predicts higher levels of anxiety, depression, and poorer functioning.  57 b. 5.  Approach predicts lower levels of anxiety, depression, and improved functioning.  Coping by catastrophizing predicts higher levels of anxiety, depression, and poorer functioning. Changes in Coping over Time Although the IMQ demonstrated associations between problem-focused coping and  outcomes previously unrevealed in studies using generic instruments, the IMQ's findings may have been related to subjects' relatively short illness duration, which averaged about 40 months. The "newness" of illness demands may have been associated with certain types of coping in these individuals, perhaps dictated by prior experiences or dispositional characteristics. In response to continuing ill-health, patients may modify their coping efforts, eliminating strategies that are ineffective and/or implementing new ones (Lewis, 1996). Evidence suggests that patients' coping behaviors and illness outcomes differ depending on the duration of the illness. In a community sample it was reported that compared to individuals with a longer duration of illness, individuals with shorter illness durations reported better health on a measure of functioning and were more optimistic about recovery even though both groups experienced similar levels of symptom severity (Jason et al., 2000). The authors surmised that individuals with shorter illness durations may cope with their new and debilitating illness by denying illness chronicity and expecting a cure. Friedberg et al.'s (2000) findings lend support to this hypothesis. Compared to patients ill for three to seven years, CFS patients who were ill for less than three years favored the strategy of denying the illness by maintaining activity. Friedberg et al. (2000) opined that an "initial illness coping strategy of denial of illness may be replaced over time by illness accommodation, as illness limitations are accepted" (p. 65). Similar to Ray et al.'s (1993) sample, patients in Friedberg et al.'s (2000) study who coped by maintaining activity reported higher levels of fatigue, depression, and stress.  58 Sullivan et al. (1992) found that for patients with chronic illness, the duration of illness was related to coping and health care utilization. Patients with a shorter duration of illness utilized more health care resources, possibly due to medical referrals and attempts to obtain diagnostic clarification. It was concluded that as illness duration lengthened individuals began to acknowledge the chronicity of their condition and reduced their health-seeking behavior. To further investigate the longitudinal processes involved in coping with the demands of CFS, Fennell (1995) developed a model of CFS that included four phases; (1) crisis, when the individual experiences the traumatic aspects of a new illness, (2) chaos, a time of intense upheaval and dissembling that concludes with the stabilization of symptoms, (3) resolution and acceptance of chronic illness when the individual is able to create meaning out of the illness experience, and (4) integration, the final phase when the individual is able to integrate pre-and post-illness self-concepts and respond to the illness in a planful way. Analysis of Fennell's phase model (N=65) was performed using cross-sectional comparisons of CFS subjects with varying illness durations (Jason et al., 2000). Three clusters were identified and labeled as a crisis group, a stabilization group, and a resolution group. Significant differences were found between the groups with the crisis group experiencing the greatest illness severity, psychological distress, and impairment. The stabilization group reported the least symptoms, dysfunction, and distress (thereby demonstrating a plateau in the CFS experience). Intermediate between the crisis and stabilization groups was the resolution group who reported a resurgence of CFS symptoms. However, patients appeared to be better able to manage the illness as reflected in psychological distress scores that were lower than patients in the crisis group. Coping also varied in the three groups with accommodating to the illness and seeking information predominant in the stabilization and resolution groups. In conclusion, it would seem that the demands of CFS and individuals' coping strategies change over time. However, the episodic nature of CFS symptoms may confound researchers'  59 attempts to clearly identify different phases and the coping behaviors associated with each phase. Greater clarity may be gained by using a longitudinal design that follows patients' coping efforts over time. Although the literature consistently demonstrates that patients with longer illness durations cope by accommodating to the illness and experience less distress, no study has examined how emotion-focused coping functions over time in the CFS population. In addition, the implications of each illness phase for the utilization of health care have not been studied. Hypotheses  The following hypotheses may shed light on how coping changes over time and the association of these changes to outcomes; 1.  Compared to CFS patients ill for four years or more (long duration), patients ill for less than four years (short duration) cope by maintaining activity. (Because the analyses were conducted in order to examine a progressive-stage model of coping that does not specify time intervals for each stage, a four year demarcation was arbitrarily chosen to stratify the sample into two groups).  2.  Compared to short-duration illness patients, long-duration illness patients cope by accommodating to the illness and by utilizing emotion-focused coping characterized by approach strategies.  3.  Compared to long-duration illness patients, short-duration illness patients report greater anxiety, depression, functioning, and health care utilization. Relations between Illness Representations and Coping Many theorists have stressed the importance of understanding how the appraisal of  a stressor affects an individual's choice of coping strategy. With the advent of the self-regulatory model and the development of the IPQ a few CFS studies have examined this process in a systematic way. Appraisals of illness threats, specifically illness perceptions, were shown to be logically related to coping; specifically pessimistic perceptions were related to coping associated  60 with unfavorable outcomes (maladaptive strategies), while more positive perceptions were linked to coping associated with favorable outcomes (adaptive strategies). Using the IPQ to quantify illness perceptions and the COPE, a generic measure of problem- and emotion-focused coping, Moss-Morris et al. (1996) reported that patients who believed they exercised some control over CFS utilized more approach strategies including active coping, positive reinterpretation, and less avoidance coping, like behavioral disengagement. Pessimistic perceptions like a strong illness identity, belief the illness will last a long time, and the perception that CFS had serious consequences were correlated with negative emotion-focused coping including venting and avoidant behaviors like disengagement. In a Dutch sample (N=98) Heijmans (1998) confirmed the findings of Moss-Morris et al. (1996) using the IPQ and the Utrecht Coping Questionnaire. More pessimistic beliefs including a strong illness identity, longer timeline, and limited beliefs in the possibility of cure predicted the use of avoidant coping and were negatively related to problem-solving coping. Overall, the findings support the validity of the self-regulatory model and the IPQ. However, the cross-sectional nature of the study designs precludes determining change over time. In addition, due to the heterogeneity of the CFS population, additional replication will be necessary before the findings may be introduced into clinical practice. Finally, the complete absence of any information on the determinants of catastrophizing in the CFS population is a significant oversight that must be addressed. Hypotheses Testing the following hypotheses may help characterize the relations between illness perceptions and coping; 1.  Pessimistic perceptions predict maladaptive coping strategies including catastrophizing, focusing on symptoms, and avoidance.  61 2.  Perceptions of control predict adaptive coping strategies including accommodating to the illness and approach strategies. Social Support In addition to coping, an important way of managing the stress of illness is by utilizing  social support. Research into social support was sparked by findings from epidemiological studies published in the mid-1970s demonstrating that social factors were associated with the duration and course of many disorders (Caplan, 1974; Cassel, 1976; Cobb, 1976). Social support is commonly defined as functions provided to a distressed individual by family, friends, coworkers, and/or neighbours. Several classification schemes have been developed for categorizing the provisions of social support. At a basic level Pattison (1977) described support as being either affective or instrumental. Affective support includes emotional support, recognition, social reinforcement, and esteem-building. On the other hand, instrumental support describes more tangible forms of assistance, for instance material aid and/or financial help. Types of Social Support In general, social support is viewed as a generic construct that includes the structural characteristics of social networks (the existence and quantity of social relationships) and the functional aspects of social interactions (Schreurs & de Ridder, 1997). Most theoretical formulations further differentiate functional aspects by characterizing support as either perceived or received. Perceived support is an individual's cognitive appraisal of his/her linkages to others. Measures of perceived support quantify the individual's confidence that when support is needed it will be available, sufficient to meet the need, and offered in a beneficial way. Received support refers to specific actions or behaviours actually performed by others as they exhibit expressions of support and assistance, for example; listening, expressing concerns, helping with a task, giving advice, and showing affection (Streeter & Franklin, 1992).  62 O'Reilly (1988) argued the case for measuring both received and perceived social support. Such a strategy would avoid the problem of relying solely on subjective perceptions of support which might be influenced by intrapsychic characteristics, like personality or mood and would provide insight into the specific interactions resulting in the perception of support. In 1  addition, limited evidence provided by cardio-vascular studies suggests the utility of including an assessment of the structural elements of social support, like the number of contacts. Some theorists have recommended that researchers measure all three social support, approaches whenever possible (Penninx, van Tilburg, Boeke, Deeg, Kriegsman, & van Eijk, 1998). Mode of Action Like coping, social support is thought to buffer the effects of stress, although how this is accomplished remains a matter of debate (Pierce, Sarason, & Sarason, 1996). Support may alter an individual's appraisal of stressors or contribute to feelings of personal mastery and selfesteem (Vaux, 1988). Thoits (1986) argued that social support represents a form of coping assistance pointing out that coping and social support share a number of functions. For instance, problem-focused coping and instrumental support are both aimed at changing the stressful situation. Emotion-focused coping and affective support serve to ameliorate the negative feeling that typically accompanies stress exposure. Because social support seems to parallel coping strategies, Thoits (1986) hypothesized that the same coping methods employed by individuals in response to stress may be utilized by others as assistance. For example, an individual's choice to use the emotion-focused coping strategy of distraction to reduce distress may have his/her efforts augmented by a friend who provides entertaining company. Or, a problem-focused strategy of conserving energy may be operationalized by obtaining transportation to work from a friend. Thus, an individual's efforts at stress management may be supplemented by the assistance of others.  63 Empirical support exists for Thoits' (1986) conceptualization of social support as coping assistance. In a variety of correlational studies investigating the psychosocial behaviours of the chronically ill it has been demonstrated that social support is generally associated with adaptive strategies, such as problem-focused coping and that diminished social support is associated with less adaptive coping strategies (Friedman et al., 1992; Kvam & Lyons, 1991). However, some studies have produced contradictory results. Among cancer patients the perceived availability of, and satisfaction with, social support were not related to coping strategies such as avoidance (Ell, Mantell, Hamovich, & Nishimoto, 1989; Behan & Rodrigue, 1994). Moreover, in an M S population the perception of received support was not correlated with problem-focused coping (O'Brien, 1993). Such inconsistencies may be explained by the fact that social support may only be helpful when it is in accordance with the preferred coping strategy (Schreurs & de Ridder, 1997). To be effective social support should also be matched to the stressor. For individuals experiencing stressors involving resource deficits (e.g., financial hardship, illness) the most salient effects would be expected from instrumental assistance (money and/or help with household chores). Stressors that lead to internal attributions of failure or inadequacy may be buffered more effectively by affective support. Similarly, different types of support may assume greater or lesser significance in chronic illness. Researchers have reported that the perception of support is more important than received support. Possibly, this occurs because perceived support exerts a generalized, non-specific, favourable influence on the course of illness. Timing of Social Support Cutrona (1990) emphasized the importance of utilizing longitudinal designs in the study of social support. Stressors change over time making the timing of social support of some import. Specific types of support may be more or less adequate at different points of time. Since  64 illnesses tend to go through different phases, a longitudinal examination of social support is needed in order to better understand the relations between support and outcomes. Relations between Social Support and Outcomes Social support research in the CFS population is somewhat flawed methodologically. Instead of utilizing a separate tool, many investigators have measured support using coping instruments like the COPE, the Utrecht Coping Questionnaire (UCQ), or the Ways of Coping Checklist that operationalize social support as a coping strategy (Afari et al, 2000; Antoni et al., 1994; Heijmans, 1998; Moss-Morris et al., 1996). Such coping measures treat support as a unidimensional construct that captures patients' intention to seek social support but fail to indicate if support was deemed adequate. Other studies have relied on brief, investigatordeveloped support measures of unproven validity (Lewis et al., 1994). In a few studies social support was unrelated to outcomes (Afari et al, 2000; Antoni et al., 1994). For the studies that produced statistically significant results it was unclear how to interpret the findings in the context of social support theory (Heijmans, 1998; Moss-Morris et al., 1996). Only the most basic information could be deduced, that some sort of social support was linked to outcomes in some way. Information of this nature cannot be used to develop meaningful interventions. There is reason to believe that social support deserves greater attention. CFS patients consider support necessary for recovery (Saltzstein et al., 1998). Due to the controversial nature of the illness patients often require the esteem building and validation that may only be forthcoming from supportive family and friends (Cordingley, Wearden, Appleby, & Fisher, 2001). In patients with a short duration of CFS, support dissatisfaction has been related to worsening or no change in symptoms (van der Werf et al., 2002). Finally, social support has been linked to decreased levels of anxiety (Moss-Morris et al., 1996; Ray, 1992; Saltzstein et al., 1998).  65 Unfortunately, as discussed in chapter one, patients may find it difficult to obtain support. Compared to healthy controls patients with CFS have reported less satisfaction with their support systems (Servaes, Verhagen, & Bleijenberg, 2001). Friends and family may not be accepting of the diagnosis (Anderson & Ferrans, 1997). Further, CFS has a detrimental effect on relationships and reduces patients' opportunities to interact socially. A clearer understanding of social support is needed because evidence suggests that certain types of social support may be related to unfavorable outcomes (Jang, Haley, Mortimer, & Small, 2003). Chronically ill patients receiving instrumental support reported more depressive symptoms, perhaps due to feelings of helplessness or dependency (Penninx et al., 1998). Studies from the chronic pain literature have shown that assistance from family members can reinforce illness behaviors (Gil, Abrams, Philips, & Williams, 1992; Lousberg, Schmidt, & Groenman, 1992; Paulsen & Altmaier, 1995). In a case study of an adolescent with CFS Rikard-Bell and Waters (1992) opined that an overprotective family can increase dependency and perpetuate disability. Two cross-sectional studies found that the supportive behaviors of significant others predicted more severe symptoms and disability in patients with CFS (Schmaling, Smith, & Buchwald, 2000; Goodwin, 1997). Solicitous partners who assist patients with tasks of daily living may inadvertently reduce patients' activity levels and increase deconditioning. However, the nature of the study designs did not allow the investigators to conclude that high support resulted in increased symptoms. The findings could be interpreted in an opposite direction, that is, sufferers with severe symptoms may offer more opportunities for concerned responses. Hypotheses In order to clarify support's relations with outcomes, indices used to measure the construct must respect its multidimensionality, specify the perceived adequacy of support, and itemize the supportive behaviours. In addition, it may be wise to use longitudinal data to model  66 relations between variables since the perceived quality of support may be related to concurrent levels of psychological adjustment (Sandler & Barrera, 1984). The following hypotheses are offered to clarify the relations between social support and outcomes; 1.  Perceived social support predicts lower depression and anxiety scores.  2.  Received social support predicts poorer functioning and lower depression and anxiety scores. It is also important to examine support at various points in the illness experience.  Different amounts of support may be utilized at different times. Since the literature suggests patients may experience some difficulties initially accepting the illness and mobilizing social support, the following hypothesis will be tested: 3.  Compared to patients ill for four years or more, patients ill for fewer than four years report lower levels of perceived and received support. Relations between Illness Representations and Social Support In general, little research has been done to identify the factors that motivate an individual  to mobilize support (Hobfoll & Freedy, 1990). CFS studies that examined the determinants of support have done so in the form of illness perceptions. Two studies identified statistically significant relations between perceptions and support (Heijmans, 1998; Moss-Morris et al., 1996). Although both used samples recruited from patient support groups, the authors produced contradictory results, perhaps due to differences in how social support was measured. In a sample of 98 CFS patients tested with the U C Q , Dutch investigators reported a negative association between seeking social support and the pessimistic perceptions that CFS was serious, had a strong identity, and was likely to last a long time. However, in a New Zealand sample (N=233) Moss-Morris et al. (1996) demonstrated positive correlations between seeking social support (as measured by the COPE) and pessimistic perceptions including the belief that CFS would last a long time and has serious consequences. To their credit, Moss-  67 Morris et al.'s (1996) results are more consistent with the notion that social support is mobilized when stress is experienced. In summary, additional research using longitudinal data is needed to clarify how perceptions relate to social support and to permit an examination of this process over time. Based on the preceding discussion it is hypothesized that highly pessimistic perceptions predict perceived and received social support. Appraisals of the Efficacy of Coping and Social Support Overall, researchers' have found that coping and social support are only modestly related to outcomes which may be due to shortcomings in measurement (Heijmans, 1998; Moss-Morris, 1997, Scharloo et al., 1998). Extant instruments generally only measure the frequency of behaviors. Respondents are presented with a list of items and asked to indicate how often each tactic is used. Correlating the frequency of use of a particular strategy with an outcome does not take into account how successful the effort was at solving the problem or modifying distress (Aldwin & Revenson, 1987). Coping and social support may need to be successful in order to demonstrate a significant association with outcomes. Aldwin and Revenson (1987) differentiated between coping effectiveness and efficacy. Effectiveness was defined as "the relation between coping efforts and some outcome measure while efficacy referred to the perception that the effort was successful in achieving the individual's goals in a particular situation" (p. 339). In a longitudinal study using a community sample of adults (N=291), the authors demonstrated that specific problem-focused coping strategies were associated with fewer symptoms but only when deemed efficacious. Researchers reported in a sample of M S patients (N=56) that psychological distress was positively correlated with emotion-focused coping and lower levels of perceived efficacy of coping (Jean, Paul, & Beatty, 1999). For CFS patients (1NN68) higher self-efficacy for symptom management was predictive of lower disability and distress (Findley et al., 1998).  68 Similarly, social support may not correlate with outcomes because it lacks efficacy. Regardless of a provider's best intentions, supportive behaviors are not always perceived as helpful by the recipient. Ideally support should be congruent with the requirements of the situation and the needs of the recipient. If differences exist between the kind of support offered and what the recipient needs, or if support is offered when it is not expected, support may not be considered efficacious (Coyne & DeLongis, 1996). In addition, support may come with "strings attached," or be provided by others who are annoying, critical, or over-involved. Due to the constraints of illness, recipients of support may not be able to reciprocate assistance and experience guilt. Ray (1992) found that non-supportive behaviors reflecting expressed irritation, frustration, misunderstanding, and pressure were positively related to anxiety and depression. Although a better understanding of the efficacy of patients' coping and social support efforts is important for clarifying the relations between psychosocial behaviors and outcomes, there are additional reasons for studying efficacy in the CFS population. Due to the lack of treatments for CFS, patients are often advised to go home and cope with their illness. However, cognitive deficits, both perceived and actual, may limit a patient's ability to effectively utilize certain strategies, for instance the cognitive strategy of distraction (reading a book or watching television). Cognitive strategies are included in most coping instruments but it is unknown to what extent CFS patients find such strategies useful. As well, CFS patients are faced with a long-term illness that induces multiple stressors. Evidence suggests coping may be less efficacious among people exposed to chronic as opposed to acute stressors. The power of coping to facilitate adjustment may become less pronounced as stress continues. For instance, certain problem-focused strategies, like seeking information and maintaining activity may become less useful as patients adapt to their condition. Over time the efficacy of social support may also wane (Stewart et al., 1997). Supportive others may become "burned out" and provide less helpful assistance or patients' inability to reciprocate may  69 discourage them from requesting help. It would be instructive to determine if patients' coping and social support efforts were perceived to be uniformly efficacious or i f certain strategies lose potency over time. Investigation into CFS patients' perceptions of the efficacy of coping and social support is needed to guide clinical practice. It has been suggested that patients consider the avoidance of activity useful. However, other coping efforts may also serve important purposes but may have failed to attract attention because they haven't produced statistically significant results (Ax et al., 2001). Further, next to nothing is known about what supportive behaviors are considered helpful by patients. Identifying coping and social support strategies patients themselves find efficacious may aid practitioners who wish to develop interventions. Patients may be more likely to adopt strategies perceived to be efficacious by other sufferers (Ax et al., 2000; Thoits, 1986). Hypotheses Clarification of the association between outcomes and the perceived efficacy of social support and coping may be gained by testing the following hypothesis: 1.  Variables measuring the perceived efficacy of coping and social support strategies are better predictors of outcomes than variables measuring the frequency that coping and social support strategies are used. In addition, a better understanding of how patients' appraisals of efficacy change may be  evaluated by testing the following hypothesis: 2.  Over time patients' perception of the efficacy of the following strategies diminish; (a) seeking information, (b) maintaining activity, and (c) received support. Finally, it will be possible to identify the coping and social support strategies perceived  by patients to be efficacious by testing the following hypothesis; 3.  Patients attribute greater efficacy to; (a) accommodating to the illness, (b) perceived support, and (c) emotion-focused strategies characterized by approach.  70 Chapter Summary Chapter two presented a critical review of the literature in order to support the study problem and methodology. The critique outlined the theoretical and empirical links between the constructs of interest. The findings and shortcomings of extant studies were used to generate hypotheses designed to address the research questions and guide statistical analyses. In the next chapter a comprehensive description of the study methods is offered.  71 C H A P T E R THREE: Method In this chapter, the methods used to answer the research questions are presented. The chapter describes the study design, sample selection, recruitment process, procedures, instrumentation, and the statistical treatment of the data. In addition, the ethical considerations of the study are discussed. Research Design The study is informed by a survey design with four data collection points, each scheduled six months apart. Participants Recruitment Patients from British Columbia, Canada were eligible for inclusion as long as the patient could visit the investigator's office at the Vancouver General Hospital for screening. Patients were recruited from the community and physicians' practices. To recruit from the community, advertisements were placed in local newspapers and professional journals. In addition, appeals were directed to members of a CFS patient support organization, Myalgic Encephalomyelitis of British Columbia (MEBC), through the group's newsletter. As well, family practice physicians in Vancouver who treated CFS patients were contacted by mail and asked to inform patients about the study. Patients were also recruited from the practices of two Vancouver, British Columbia internists, one specializing in Infectious Diseases and the other specializing in Rheumatology. To recruit patients from the specialty practices, letters from the physicians describing the study were mailed to patients who met the CDC criteria for CFS (Fukuda et a l , 1994). Interested individuals were invited to contact the investigator.  72 Sample Criteria Prospective participants who had not been seen by either of the participating specialists were requested to provide written evidence from their attending physician to verify the diagnosis. For all potential respondents, an interview was conducted to screen for the following inclusion and exclusion criteria: Inclusion Criteria 1.  Fulfill the CDC criteria of CFS (Fukuda et al., 1994; see Appendix 1)  2.  English language fluency  3.  18 years or older  Exclusion Criteria 1. Ongoing cognitive behavioural therapy Study Procedures Eligible patients signed an informed consent form and completed the computerized version of the Diagnostic Interview Schedule IV (CDIS-IV; Robins, Cottier, Bucholz, Compton, North & Rourke, 2000) which was used to determine patients' psychiatric status. Following successful screening participants received and briefly reviewed a set of 11 questionnaires measuring independent, control, and dependent variables. Participants were instructed to complete the materials at home and return them by mail within one month. They were invited to contact the investigator at any time to resolve queries. Upon receipt, the questionnaire packs were inspected by the investigator for completeness and the participant contacted immediately i f any forms were not received. Six months after receiving the completed questionnaires, the investigator contacted the respondent by telephone to verify the correct address and the next questionnaire set was mailed out. In total, respondents were expected to complete four sets of questionnaires over an 18-month period. Follow-up phone calls were made to respondents who failed to return completed materials within one month.  73 Measurement The following measures were used to operationalize the study constructs. For measures not in the public domain, written permission was obtained from the authors. Independent Variables Demographics At time one, respondents recorded the following information: age, gender, ethnicity, marital status, number of children living in the home, number of hours each week of work and/or school, income pre- and post-CFS onset (five category scale), and highest school grade achieved. Psychiatric Status The CDIS-IV (Robins et al., 2000) was used to screen patients for lifetime and current episodes of major depressive disorder (MDD), phobia, panic disorder, generalized anxiety, mania (used in conjunction with M D D to identify bipolar disorder), hypomania, pain, dysthymia, and somatization disorder. The CDIS-IV is a computer program based on the Diagnostic Interview Schedule (DIS), a highly structured interview originally developed to make diagnoses using the following systems; (1) the Diagnostic and Statistical Manual of Mental Disorders (DSM-III; American Psychiatric Association, 1980), (2) the Feighner criteria (Feighner et al., 1972), and (3) the Research Diagnostic Criteria (RDC; Spitzer, Endicott, & Robins, 1978). The DIS was intended for use by lay interviewers to estimate the prevalence and incidence of specific psychiatric disorders in the general population (Robins, Helzer, Ratcliff, & Seyfried, 1982). Psychometric testing by the authors demonstrated that the overall agreement between DIS results and a psychiatrist's impression exceeded 90% (Robins, Helzer, Croughan, & Ratcliff, 1981). The sensitivity of the DIS has been calculated at between 93% and 95% (Helzer et al., 1985). The CDIS-IV was necessitated by the publication of the D S M - I V (American Psychiatric Association, 1994). The program covers the criteria for each diagnostic category in the form of one or more pre-coded, close-ended questions that the respondent reads on the computer monitor  74 and keys in responses. Each positive symptom invokes a probe that determines if a symptom should be counted towards diagnosis or is caused by physical illness or substance use. (To control for measurement bias in the somatization disorder scale, respondents were instructed to attribute medically-unexplained, CFS-related symptoms that occurred within six months of fatigue to physical disease.) Scoring is performed by the computer according to the explicit rules in the diagnostic systems served. Diagnostic categories are not mutually exclusive and respondents may meet criteria for more than one diagnosis. Personality Diefenbach and Leventhal (1996) have suggested that personality may be an important determinant of patients' illness perceptions. Thus, the N E O Personality Inventory (NEC* PI-R; Costa & McCrae, 1992) was used to measure normal personality traits. The instrument was selfadministered at time 1 and included 181 statements that were answered on a five-point scale from "strongly disagree" to "strongly agree." The NEO PI-R measures the "big five" dimensions of personality including extroversion (E), a disposition to being sociable, upbeat, and energetic; neuroticism (N), emotional vulnerability and the tendency to experience negative affects; openness to experience (O), curiosity about inner and outer worlds; conscientiousness (C), the tendency to be purposeful,, strong-willed, and determined; and agreeableness (A), the interpersonal tendency to be altruistic. To score the NEO PI-R, the authors advised checking for missing data, acquiescence, nay-saying, and random responding. Item scores may then be combined and summed to represent the final tally for each domain. Normative scores are available for the N E O PI-R which were based on a sample of 2273 American adults who completed the questionnaire. Five hundred men and 500-women were selected, screened for validity and random responding then matched to 1995 U.S. Census projections for the distribution of age and race (Costa & McCrae, 1992).  75 The advantage of using the NEO PI-R is that the instrument contains no items reflective of health thereby avoiding the possibility of content contamination. Illness Representations Derived from the self-regulatory model (Leventhal et al., 1984), the Revised Illness Perception Questionnaire (IPQ-R; Moss-Morris, Weinman, Petrie, Home, Cameron, & Buick, 2002) is an 82-item scale used to measure respondents' emotional and cognitive representations of illness. Although based on the five cognitive components; identity, consequences, control, timeline and causes, psychometric analyses of the original IPQ suggested that the subscales control and timeline required further differentiation. Another departure from the IPQ involved adding a scale to measure emotional representations. Finally, the authors included a scale to measure the extent to which an individual's representations provided a coherent understanding of the illness. The first part of the questionnaire assessed illness identity, which is operationally defined as the number of symptoms endorsed by respondents as part of the illness. Fourteen core symptoms were included and participants responded by indicating whether or not they experienced each symptom. The number of symptoms endorsed by the respondent was summed to produce an identity score. The second part of the questionnaire consisted of items designed to measure the following components; (a) emotional representations, higher scores indicate greater distress; (b) timeline (acute/chronic), higher scores suggest greater chronicity, (c) timeline (cyclical), higher scores indicate the symptoms are episodic, (d) consequences, higher scores reflect the belief the illness is serious; (e) personal control, higher scores are indicative of greater control; (f) treatment control, higher scores indicate the respondent believes the illness may be controlled by health care interventions; and (g) illness coherence, higher scores reflect the individual's belief that the illness is puzzling. A l l items were presented in a mixed order and rated on a five-  76 point scale ranging from strongly disagrees to strongly agrees. After reverse scoring was performed on the appropriate items the mean for each subscale was calculated. The remaining subscale, causal attributions consisted of 18 statements listing possible causes of the illness to which the respondent expressed agreement on a 5-point likert scale. The responses were summed and categorized according to factor analyses performed by the IPQ-R originators (Moss-Morris et al., 2002) on data obtained from chronically ill patients (N = 711). Using principal components analyses with a Varimax rotation the authors extracted four factors which accounted for 57% of the total variance. The factors were interpreted as; (1) psychological attributions, (2) risk factor attributions, (3) immune attributions, and (4) chance attributions (the illness was caused by an accident or bad luck). A total of 12 variables were rendered by the IPQ-R which was administered at each of the four time points. Coping 1. Problem-focused coping. The 55-item Illness Management Questionnaire (IMQ; Ray et al., 1993) quantifies four ways patients manage CFS; (a) maintaining activity, defined as "attempting to ignore symptoms (and) carrying on even though unwell" (17 items), (b) accommodating to the illness, "organizing and planning one's life in order to avoid overexertion and control stress" (13 items), (c) focusing on symptoms, "a preoccupation with symptoms, linked with an appraisal of helplessness and of one's life as dominated by the illness" (9 items), and (d) seeking information, searching "for relevant information and a readiness to try remedies" (6 items; p. 387). Respondents were requested to rate how frequently each strategy was utilized on a six point scale (1 = never to 6 = always). Final scores reflect the mean of each subscale. The IMQ was administered at each data point. 2. Emotion-focused coping. The COPE is a 60-item instrument that includes problemand emotion-focused coping which are incorporated into 14 conceptually distinct scales with four items each (Carver et al., 1989). Each item is a statement describing a strategy. The  77' frequency that the strategy is used is rated by the respondent on a four-point likert scale ranging from 1 = I usually don't do this at all, to 4 = I usually do this a lot. The subscales include, active coping, planning, suppression of competing activities, seeking support, religion, positive reinterpretation and growth, restraint coping, acceptance, focus on and venting of emotions, denial, mental disengagement, behavioural disengagement, and alcohol use. Since the original publication, a additional scale, humour, was added. To avoid duplication with items from the IMQ, ISSB, and SSQ, the following COPE problem-focused and social support subscales were deleted; active coping, planning, suppression of competing activities, restraint coping, and seeking social support. In addition, one item was deleted from the mental disengagement subscale, "I sleep more than usual," since this is a symptom often associated with CFS. Further, in order to simplify analyses and be consistent with the work of other CFS researchers using the COPE, subscales were combined into two emotion-focused meta-strategies, approach and avoidance (Antoni et al., 1994). Approach included the acceptance and positive reinterpretation subscales while avoidance consisted of the behavioural disengagement, denial, and mental disengagement subscales. The COPE was administered at each data point. 3. Catastrophizing. The six-item catastrophizing subscale of the Coping Strategies Questionnaire (CSQ: Keefe et al., 1989) was adapted to assess catastrophizing. Instead of assessing beliefs prompted by pain as formulated in the original CSQ, respondents were asked to indicate on a seven-point scale (0 = never to 6 = always) how often they had certain thoughts when symptoms worsened in response to over-exertion. Specific items included the following beliefs; (a) it's not going to get better, (b) it overwhelms me, (c) life isn't worth living, (d) I worry that it will never end, (e) I can't stand it anymore, and (f) I feel I can't go on. Although critics were concerned that catastrophizing overlapped with depression, studies in chronic pain suggest catastrophizing is a conceptually distinct entity. Path analysis demonstrated that catastrophizing mediated the relationship between depression and outcomes  78 (Geisser, Robinson, Keefe, & Weiner, 1994). In addition, a longitudinal study found that initial catastrophizing scores were able to predict long-term depression even when initial depression scores were controlled (Keefe et al., 1990). Catastrophizing is a unidimensional concept so items were summed to form a final score with higher scores indicating a greater tendency to catastrophize (Keefe et al., 1989). The CSQ was administered at each data point. Social Support 1. Received social support. The Inventory of Socially Supportive Behaviours (ISSB; Barrera & Sandler, 1981) is a 40-item scale that may be used as a unidimensional instrument (Streeter & Franklin, 1988). The items include instrumental and affective components of social support (Barrera & Sandler, 1981). Each item measures the number of times respondents were the recipients of a supportive action during the previous four-week period. Respondents were asked to rate the frequency of action on a five-point likert scale ranging from (1 = not at all to 5 = about every day). The scale was scored by obtaining the average frequency of the actions. The ISSB was administered at each data point. 2. Perceived social support. The Social Support Questionnaire (SSQ; Sarason, Levine, Basham, & Sarason, 1983) is a 27-item scale that investigates two aspects of social support, (a) the number of supportive others (SSQ-N) and (b) the degree to which support is satisfying (SSQS). Each item asks a question to which a two-part answer is requested. For example, respondents are asked, "who can you really count on to listen to you? Respondents are then prompted to; (a) list the initials of all the people who can be relied upon to provide the support and, (b) on a sixpoint likert scale indicate how satisfied they are with the support (1 = very dissatisfied to 6 = very satisfied). N and S scores are obtained by dividing the sum of N or S scores for all items by 27 (Sarason et al., 1983). The SSQ was administered at each data point.  79 Appraisal of the Efficacy of Self-management Strategies The coping and social support scales were adapted to allow respondents' to rate the efficacy of their efforts. The operationalization of efficacy requires situation- or disease-specific contexts (Bandura, 1977). Ray et al. (1993) suggested that CFS patients differentiated between a strategy's capacity to control symptoms and its capacity to aid in recovering from the illness. Thus, each strategy described in the coping and social support questionnaires (SSQ, ISSB, IMQ, and COPE) was paired with two questions asking respondents to rate on a six-point scale (1 = not applicable to 6 = extremely effective) the strategy's efficacy for; (a) relieving symptoms and (b) helping the patient recover from CFS. For example, the IMQ inquires about ways of dealing with CFS and asks respondents to say what is true for them. One of the items states, "I tell myself I don't feel so bad." Following this statement respondents are asked to rate the efficacy of such a strategy for; (1) helping control CFS symptoms and (2) helping to recover from CFS. The efficacy scales were administered at each data point. Outcome Variables Functioning The SF-36 (Ware & Sherbourne, 1992) was used to measure physical and mental functioning at each of the four data points. Although in the past CFS researchers have used the Sickness Impact Profile (SIP; Bergner et al., 1976), the SF-36 may be the preferred instrument. In a clinical population, compared to the SIP, the SF-36 was found to be more acceptable to respondents (due to its brevity) and more responsive to change (Stucki, Liang, Phillips, & Katz, 1995). The SF-36 (Ware & Sherbourne, 1992) is a quality of life instrument composed of 36 statements that are answered on a two to six response scale indicating the degree of impairment from a patient's ideal state of health. The responses are aggregated into eight subscales, and when transformed range from 0 to 100 with 0 representing the poorest health. The subscales  80 include; physical functioning (10 items), role-physical (four items), bodily pain (two items), general health perceptions (five items), vitality (four items), social functioning (two items), roleemotional (three items) and general mental health (five items). Algorithms for a total score have not been published but factor analysis has identified two main factors, physical and mental health (Ware, Snow, & Kosinski, 2000). For the physical factor the following subscales were found to be strongly associated (r > .7); physical functioning, role-physical, and bodily pain. The subscales with strong associations with mental health (r > .7) included mental health, roleemotional, and social functioning. Vitality and general health perceptions loaded on both mental and physical factors. For the purposes of the study, the subscales were aggregated to form two superordinate scales; (1) total physical functioning and (2) total mental health. Normative scores have been provided for the SF-36 (Ware et al., 2000) which are based on responses from the National Survey of Functional Status (NSFHS), a 1990 American survey (N = 2909) using probability sampling methods. In addition, SF-36 scores have been published for chronically ill patients who participated in the Medical Outcomes Study (N=l8,762; Ware et a l , 2000). Health Care Visits At each data point participants were asked to report all health care visits made over the preceding six months. The date, type of health care service, and reason for each visit was recorded. Three variables were created; (1) the total number of health care visits, (2) the number of visits to conventional care providers, and (3) the number of visits to alternative care providers. Psychological Distress The Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983) is a 14item scale consisting of two subscales measuring depression (7 items) and anxiety (7 items). Respondents were asked on a four-point scale how often or how much they experienced an aspect of depression or anxiety. Items in each subscale were summed and the higher the score the  81 greater the anxiety or depression. In the original validation study the authors suggested that for either subscale non-cases scored 7 or less, doubtful cases 8-10, and definite cases 11 or more (Zigmond & Snaith, 1983). The H A D S was specifically devised for use with physically ill patients and the depression scale was constructed to exclude somatic items. Thus, the scale measures anhedonia and should not lead to the overinflation of depression scores caused when physical symptoms are included as subscale items. The HADS was administered at each data point. Covariates The following variables have been linked with outcomes in CFS and must be controlled for in statistical analyses; Illness-related Variables At time point one data on the following variables were collected; illness duration (time of onset and time to diagnosis), the type of onset (sudden or insidious), non-psychiatric comorbid illnesses, and presence or absence of post-exertional malaise. Control Variables Disability pension status. At each time point, respondents were asked to describe their disability pension status including whether or not; (1) they were in the process of applying for disability, (2) a disability pension had been granted, (3) a disability pension had been denied, or (4) a disability pension decision was being appealed. Concomitant medications. Respondents recorded their use of prescribed, over-the-counter, and herbal medications at time point one. Medications were grouped according to indication and included the following categories; pain, indigestion, antidepressants, sleep, blood pressure, hormones, fatty acids, vitamins, cholesterol-lowering agents, herbal preparations, tranquilizers, allergy medications, bowel medications, antimicrobials, osteoporosis, thyroid, and stimulants. Summary scores of medication use included the total number of medications used, number of  82 herbal remedies, antidepressant use, and vitamin use. Statistical Procedures SPSS (version 8) was used to analyse the data. The following section describes the statistical procedures performed, the data used, and the purpose of the analysis. Descriptive and Inferential Statistics 1. Univariate statistics: To characterize the sample. 2. Analysis of Variance (ANOVA): To compare the sample characteristics of the short and long illness duration groups and to assess the impact of attrition. 3. Cronbach's alpha: To assess reliability as reflected by internal consistency. 4. Factor analysis: To assess the validity of the efficacy scales. 5. Cox regression: To identify i f specific variables are predictive of respondents' withdrawal from the study. Cross-sectional Data Analyses 1. Correlations: Demographic, illness-related, and control variables were examined. Variables with significant correlations with outcome variables were used as predictors in sequential regression equations. 2.1. Sequential multiple regression: For each of the outcome variables sequential regression models were run which included the following predictors; emotional representations, cognitive representations, coping (problem- and emotion-focused), catastrophizing, and social support (perceived and received). Selected demographic, illness-related, and control variables were entered as covariates. 2.2. Sequential multiple regression: Each illness representation component found to be a statistically significant predictor was regressed on selected personality traits and psychiatric diagnoses.  83 3. A N O V A for normally distributed data and Kruskal-Wallis for non-normal data: To determine if differences in outcomes and selected predictors exist between the short and long duration illness groups. Longitudinal Data Analyses 1. Sequential multiple regression: For the residualized scores of the outcome variables a series of sequential regression analyses were run which included demographic, illness-related, and control variables as covariates. 2. Test-retest reliability: To assess the stability of each instrument used. 3. Repeated measures analysis for four data points to determine i f outcomes, illness representations, coping, social support, and efficacy change over time. 4. Repeat measures analysis using time 2 and 3 data to determine the stability of results obtained from the analyses of time 1 and 2 data. Ethical Considerations The study used a variety of strategies to protect the rights of the participants. First, a proposal of the study was submitted and approved by the University of British Columbia Behavioural Research Ethics Board. Second, prospective participants reviewed and signed an informed consent before study procedures were initiated. Third, to protect respondents' confidentiality, questionnaires were numbered. Only the investigator had access to respondents' names. Chapter Summary In chapter three, the methods used to address the research questions and test the hypotheses were presented. A description of the sampling criteria and the strategies used to recruit participants were offered. The procedures used to collect the data were described and the plan for statistical analysis outlined. Ethical considerations of the study were furnished. In the next chapter, the results of the statistical analyses are discussed.  84 C H A P T E R FOUR: Results In chapter four, the results from the statistical analyses are presented. In the first section information about the data collection process is provided which is followed by a profile of the sample. Descriptive statistics are offered to characterize the study variables and furnish results from psychometric assessments of the instruments. The final section includes findings from inferential statistics used to answer the research questions and test the study hypotheses. P-values are reported as p < .05, p < .01, p < .005, or p <.0001. The author acknowledges that many authorities recommend reporting the actual p-value calculated instead. Data Collection Between February 2, 2000 and November 21, 2003, 108 patients who had been medically diagnosed with CFS were screened and enrolled into the study. Two individuals were immediately excluded because they met the diagnostic criteria for bipolar disorder as rendered by the CDIS-IV. Six subjects agreed to participate but failed to return any data. One hundred respondents (92%) completed the first set of questionnaires, 92 individuals (85%>) completed two questionnaire sets, 85 (78%) completed three questionnaire sets, and 61 participants (56%) returned four sets of questionnaires. On average, respondents returned the first questionnaire set 36 days after screening (range 3 - 102 days). Each following questionnaire set was received an average of 7.3 months after its predecessor was received (range 6.1 - 11.9 months). After the first set was received by the investigator a mean of 21 months (range 17.1 - 30.4) elapsed before the second, third, and fourth sets were completed and returned. Respondents found the questionnaire sets burdensome and time consuming. Fourteen individuals (14%») cited fatigue as prompting their withdrawal from the study. In one case, an individual's fourth questionnaire set was destroyed by fire but the respondent was judged too distressed to complete a duplicate. Finally, due to the need to collect data in a timely manner, about 24 respondents (24%>) were not given sufficient time to finish questionnaire set four.  85 However, compared to respondents who completed four sets of questionnaires (N=61), respondents who completed one to three sets of questionnaires (N=39) were statistically comparable in terms of the demographic, illness-related, control, and dependent variables measured at time point one. Cox regression analysis was also used to identify statistically significant predictors of withdrawal. None were noted. Descriptive Statistics Analyses were conducted to provide summary statistics about the sample, allow a preliminary assessment of the distribution of the data, and to evaluate the reliability of the instruments. Two types of reliability were assessed; (1) internal consistency, using Cronbach's alpha and (2) stability, based on test-retest correlation with retest scores obtained from time two (about seven months after time one). Although test-retest reliability is best suited to constructs that don't change much over time, like personality and attitudes (Schwartz & Daltroy, 1999), the technique is considered appropriate when comparing individuals coping with the same kind of stress at two distinct points in time (Schwarzer & Schwarzer, 1996). In addition, bivariate correlations between the demographic, illness-related, control, and outcome variables were examined. Variables that correlated significantly with the outcome variables were included as covariates in multiple regression models and entered sequentially as blocked sets. Characteristics of the Sample Demographics Ninety-seven percent of the respondents who completed one or more questionnaire sets (N=100) were Caucasian. Eighty-five percent were female and the average age of the respondents was 49 years (SD=9). On average the respondents had completed 15 years (SD=2.8) of schooling (about third-year university). Fifty percent were married or in a common-law relationship and 22% had one child or more living in the home. About 76% of the respondents  86 were not engaged in paid labor and 45% earned less than $20,000 per year. (Only 12% had earned less than $20,000 in the year preceding the onset of CFS.) At time one 58% of the sample reported receiving a disability pension while 11%> were appealing an unfavorable disability pension decision. Personality Profile A l l respondents (N=100) completed the personality inventory, N E O PI-R. (Costa & McCrae, 1992). Mean scores demonstrated that the sample was within the normal range for each of the personality domains (see Table 1). Table 1 Comparison of NEO PI-R scores for respondents and a normative sample  Personality domain  Sample (N=100)  Normal range for adults (N=500)*  Mean (SD)  Range of Means  Agreeableness  134(15)  119- 141  Conscientiousness  124 (17)  124 - 144  Extroversion  107(18)  103 - 123  Openness  119(15)  97-116  80 (22)  59- 81  Neuroticism  * Costa and McCrae, 1992 N E O PI-R scores displayed minimal skewness (.11 - .30) and alpha coefficients were high ranging from .85 to .93. The neuroticism subscale was negatively related to the other subscales including extroversion (r = -.31, p < .05, two-tailed), openness (r = - .25, p < .05, twotailed), agreeableness (r = -.23, p < .05, two-tailed), and conscientiousness (r = -.34, p < .01, onetailed). Openness was positively related to agreeableness and extroversion, (r = .33, p < .01 and r  87 = .21, p < .05, two-tailed, respectively). Conscientiousness was positively related to extroversion (r = .3 l , p < . 0 1 , two-tailed). Clinical Characteristics On average respondents were 42 years old (SD=9) when CFS symptoms first presented and 49% reported a sudden onset. The mean time from symptom onset to medical diagnosis was two years (SD=2.8). The average duration of illness was 7.9 years (SD=5.3) and most respondents (94%) experienced post-exertional malaise. Co-morbidities. Of the 99 respondents who completed the CDIS-IV, about 70% (69/99) met the diagnostic criteria for the lifetime occurrence of one or more psychiatric disorders (Table 2) with 40% having two or more diagnoses. Forty-five percent had experienced an episode of psychiatric illness within the year prior to study enrollment. Pain disorder predominated with a frequency rate of 48.5%. The anatomical regions or types of pain cited included; stomach (26), headache (24), joints (21), back (17), menstrual (16), chest (14), arms and legs (10), pain on intercourse (11), and pain on urination (1). The criteria for Major Depressive Disorder (MDD) were met by 37% (37/99) with 54% (20/37) of respondents experiencing the first episode after the onset of CFS. For 37% (14/37) M D D was concurrent and an episode had occurred within one year of study screening. For the concurrent depressives 86% (10/14) were on antidepressants and 35% (5/14) were visiting either a psychologist or a psychiatrist. However, 28% (4/14) may have had an undetected depression. They reported no visits to psychotherapists, were not on antidepressant therapy, and did not include depression as a co-morbid illness. As evidenced by A N O V A no differences were found between the type of CFS onset (sudden or insidious) and lifetime or current psychiatric morbidity. In addition no differences were found in outcomes for respondents with or without a psychiatric diagnosis save for anxiety.  88 Respondents with no psychiatric diagnosis demonstrated greater anxiety (mean = 7,6, SD = 4) than respondents with one or more psychiatric diagnoses (mean - 6, SD = 3.7; F = 5, p = .02). Table 2 Psychiatric diagnoses as determined by the CDIS-IV (N = 99). Categories are not mutually exclusive and respondents can have more than one diagnosis  Diagnosis  Lifetime  a  Current  Diagnosis  Lifetime (%)  a  Current  (%)  (%)  (%)  Pain disorder  48 (48.4)  32 (32.3)  Phobia  9(9)  5(5)  Major Depressive Disorder  37(37.3)  14(14.1)  Dysthymia  2 (2.2)  1(1)  Panic  33 (33.3)  14(14.1)  Somatization Disorder  2 (2.2)  2 (2.2)  Anxiety  13 (13.1)  10(10.1)  Hypomania  1(1)  1(1)  Any diagnosis  69 (69.7)  45 (45.4)  2 or more diagnoses  40 (40.4)  23 (23.2)  'Respondents who experienced an episode within the year prior to screening Fifty-two percent (52/100) of respondents had at least one non-psychiatric co-morbidity with the average number of co-morbidities being two (N=52; SD=1.2). The most frequently reported illnesses in order included, digestive disorders (12/52), allergies (12/52), hypertension (10/52), arthritis (9/52), osteoporosis (4/52), hypothyroidism (4/52), and diabetes (3/52). (All non-psychiatric disorders were stable under treatment and not considered the cause of patients' fatigue). Twenty-two respondents reported having other medically-unexplained syndromes including fibromyalgia (19/22), irritable bowel syndrome (2/22), and multiple chemical sensitivity (1/22). Overall, 58% (58/100) of subjects self-reported one or more psychiatric or non-psychiatric illnesses with the average for the entire group (N=100) being 1.1 (SD = 1.36).  89 Medication Use Respondents listed all the medications and herbal preparations they were using at screening and 92% reported taking at least one medication (mean = 4.73; SD = 3.4). Forty-two percent of respondents used five or more medications. The types of medications used covered a broad range (see Appendix 2) but for the most part respondents relied.on pain medications (55%), anti-depressants (45%), and sleep aids (41%). Distribution of scores for vitamin and herbal preparation use were positively skewed so a log ten transformation was applied (with a constant of one added to each individual score to counteract the presence of zeros). Outcome Variables Health Care Visits In a six month period, respondents (N=100) averaged about 15 visits to various conventional and alternative care providers. Ninety-six percent of the sample (N=97) made at least one visit to a conventional care practitioner with respondents who used conventional care averaging 13 visits. Forty-seven percent of conventional care visits were to general practitioners. Fifty-eight respondents (58%) made at least one visit to an alternative care provider with eight visits being the average. Body work was most often sought. Massage (N=29; 41%) and chiropractic treatments (N=24; 24%) were the preferred treatments. Descriptive statistics for the Health Care Visits Questionnaire are listed in Table 3.  90 Table 3 Descriptive statistics for the Health Care Visits Questionnaire and correlations between scales  Mean(SD) (N=100)  Reliability Cronbach'sa (N=100)  1. Total health care visits 15.4(15.8)  .41  2. Conventional care 3. Alternative care  10.2 (9.1) 5.1 (11.8)  •  '  Test-retest  Intercorrelations Skew  1  2  (N=92)  .58**  2.8  .35  .46**  1.4  .67**  .41  .53**  5.6  .82**.12 **  ** p < .01 (two-tailed) Alpha coefficients were low for the three subscales. Nunnally (1978) suggested that instruments achieve a minimum internal consistency of .70 to be considered acceptable for group-level comparisons. Stability was moderate. Since the variables were positively skewed, a log ten transformation was applied. A small correlation was found between the conventional and alternative care visit subscales while strong correlations were shown between total number and alternative or conventional visits. Functioning Transformed scores for the SF-36 subscales (see Appendix 3) were very low for physical functioning (mean = 41.4; SD = 21), role-physical (mean = 4.8; SD = .14), social functioning (mean = 41; SD = 24), vitality (mean = 26; SD = 14.7), and bodily pain (mean = 44; SD = 21). As well as being far below the norm for their age group, respondents' scores were lower than that reported for patients with congestive heart failure (CHF). However, the respondents achieved higher scores on role-emotional than either the healthy females or the CHF group. Descriptive statistics for the SF-36 super-ordinate subscales used in the regression analyses are listed in Table 4.  91 Table 4 Descriptive statistics for the SF-36  Reliability a  Mean (SD)  Cronbach's a  Test-retest  Skewness  (N=100)  (N=100)  (N=92)  1. Total physical functioning  42.3 (8.1)  .86  .73**  .32  2. Total mental health  45.2 (8.1)  .74  .58**  1.0  a  raw scores  ** p < .01 (two-tailed)  Reliability coefficients were acceptable and the total mental health subscale found to be slightly skewed. The two subscales were moderately intercorrelated (r = .47, p < .01, two-tailed). Psychological distress Means for the subscales were below eight indicating that the sample, as a group, did not meet the criterion for "caseness" for anxiety or depression (see Table 5). For the depression subscale, 16% of respondents scored over 11 and could be considered depressed while 17% scored over 11 meeting the criteria for anxiety. Table 5 Descriptive statistics for the HADS  Mean (SD) (N=100)  Reliability Cronbach's a Test-retest (N=92) (N=100)  Anxiety  6.6 (3.9)  .83  74**  .53  Depression  7.2 (3.2)  .76  .78**  .27  **p<.01 (two-tailed)  Skewness  92 Alphas and test-retest coefficients were acceptable and scores were not skewed. The subscales were moderately correlated at .45 (p < .01, two-tailed). Correlations between Outcomes A n examination of the correlation matrix of outcome variables suggested some redundancy between functioning and psychological distress, particularly for mental health which was negatively associated with depression (r = -.55, p < .01, two-tailed) and anxiety (-.56, p < .01, two-tailed). Health care visits were not significantly correlated with either functioning or psychological distress. Predictor Variables Illness Representations Scores for cognitive and emotional representations as reflected by the IPQ-R ranged between 3.1 and 4.1 (see Table 6). Respondents endorsed an average of 10 symptoms. Of the four attributions describing the cause of CFS immune attributions were most favored. Many of the IPQ-R subscales were reliable and scores were not skewed. However, for the timeline (cyclical) and identity subscales, alpha coefficients were in the mid-fifties and must be considered suboptimal. Alphas were extremely low for the causal attribution subscales, immune and chance, probably reflecting an insufficient number of items included in each subscale. The high stability found for several of the subscales (particularly consequences and coherence), demonstrated that respondents' perceptions were not subject to much revision over a sevenmonth period. However, test-retest correlations of less than .45, suggested that attributing cause to psycholocial or risk factors changed to a certain extent over time.  93 Table 6 Descriptive statistics for the IPQ-R  Reliability Mean (SD) IPQ-R subscales  (N=100)  1 "Consequences  4.1 (.55)  2. Timeline (chronic/acute) 3. Personal control  Cronbach's a  (N=92)  Skewness  .84  .75**  -.84  3.6 (.76)  .83  .65**  -.41  3.5 (.57)  .79  73**  -.30  4. "Emotional representation 3.2 (.70)  .85  .69**  -.06  5. "Coherence  3.2 (.90)  .88  .75**  -.11  6. Treatment control  3.1 (.63)  .80  .65**  -.69  7. Timeline (cyclical)  3.1 (:54)  .50  .59**  -.95  8. "Identity  10.2 (2.1)  .56  70**  -.34  9. Immune attributions  3.7 (.61)  .20  .50**  .50  10. Chance attributions  2.1 (.89)  .01  .50**  1.2  11. Psychological attrib.  2.4 (.85)  .84  43**  .26  12. Risk factor attrib.  1.9 (.63)  .70  .24**  1.1  a  b  b  (N=100)  Test-retest  "higher scores suggest a more pessimistic perception higher scores suggest a more optimistic perception  ** p < .01 (two-tailed)  The IPQ-R subscales were intercorrelated in a logical manner. Pessimistic perceptions like timeline, consequences, coherence, and emotional representations correlated positively with each other but were negatively correlated with optimistic perceptions like personal and treatment control (see Appendix 4). In terms of perceptions of cause, psychological and risk factor attributions were positively and moderately correlated (r = .50, p < .01, two-tailed).  94 Coping Descriptive statistics for the problem- and emotion-focused measures are presented in tables 7 and 8. To allow a comparison between problem- and emotion-focused coping in terms of the frequency that each type of coping was performed (the measures use different likert scales) the percentage of the respondents who used the behavior at least "sometimes" (IMQ) and "a little bit" (COPE) were calculated. Problem-focused coping. Statistics for the problem-focused measure, the IMQ, and the correlations between its subscales are presented in Table 7. The mean scores for the subscales suggested that all problem-focused strategies were used at least "sometimes." Respondents favored accommodating to the illness, reporting they used this strategy "quite often." A l l the subscales demonstrated considerable internal consistency with alphas that ranged from .83 to .91. Test-retest correlations ranging from .64 (for accommodate) to .83 (maintain activity) indicated considerable stability in the IMQ almost matching the test-retest correlations reported for personality traits (Costa & McCrae, 1992). The scores were normally distributed and the highest value for skewness was -.72. Moderate intercorrelations between subscales were detected. Maintaining activity was negatively related to accommodating to the illness and positively related to seeking information. Accommodating to the illness, seeking information, and focusing on symptoms were positively correlated with each other.  95 Table 7 Descriptive statistics for the IMQ  Mean (SD) (N=100)  % of sample using strategy sometimes" (or more)  Intercorrelations 1 2 3  1. Accommodate  4.26 (.68)  92%  -  2. Seek information  3.89 (.91)  80%  .09  3. Focus on symptoms 3.43 (.86)  63%  2y**  31**  4. Maintain activity  61%  -.29**  .20*  3.22 (.76)  -.13  *p<.05 (two-tailed) **p<.01 (two-tailed) Emotion-focused coping. Table 8 describes the descriptive statistics for emotion-focused coping as measured by the COPE and aggregated into three subscales for the purposes of the study. With a mean of 3.04 (indicating respondents used the strategy a "medium amount"), approach was preferred by the sample. Avoidance was least favored (mean = 1.83) and venting was used a "little bit." As demonstrated by A N O V A (F = 21.2, p < .0001) men (mean = 2.5, SD = .54) were less likely to use approach than women (mean = 3.1, SD = .51). Although avoidance produced a suboptimal alpha of .65, venting and approach were internally consistent with coefficients of .82 and .83, respectively. Test-retest correlations from .54 to .62 suggested the scales had moderate stability but were more likely to change over time than problem-focused coping strategies. Scores were normally distributed with minimal skewness. The three subscales were positively intercorrelated with the largest correlation occurring between the approach and avoidance subscales (r = .28, p < .01, two-tailed).  96 Table 8 Descriptive statistics for the emotion-focused subscales (adaptedfrom the COPE)  Subscales  Mean (SD) (N=100)  % of sample using strategy "a little bit" (or more)  Intercorrelations 1 2  1. Approach  3.04 (.60)  96%  2. Venting  2.22 (.91)  72%  .13*  3. Avoidance  1.83 (.38)  34%  .28**  .25*  * p < .05 (two-tailed) ** p < .01 (two-tailed ) Catastrophizing. The mean catastrophizing score was. 1.8, which corresponded to rating a catastrophic thought as "rarely" occurring when symptoms were exacerbated by overexertion. The scale produced an alpha coefficient of .86 and a correlation coefficient of .73 in test-retest assessment. No skewness was noted (see Table 9). Table 9 Descriptive statistics for the Catastrophizing scale  Mean (SD) (N=94)  Catastrophizing scale  1.8(1)  Reliability Cronbach's a Test-retest (N=94) (N=85)  .86  .73**  Skewness  .24  ** p = .01 (two-tailed) Social support. Descriptive statistics for the perceived and received social support measures are summarized in Table 10. On average, the respondents identified 3.9 (SD = 1.9) supportive others in their lives and were fairly satisfied with the support provided (SSQ-S mean  97 = 5.0; SD = .87). For received support the mean of 2.03 (SD = .91) suggested respondents obtained instrumental assistance rather infrequently (once or twice in four weeks). However, the following behaviors were received at least once a week; (1) expressions of interest in the respondent's well-being (mean = 3.5, SD =1.1), (2) discussing respondent's interests (mean = 3.2, SD = 1.2), and (3) comforting the respondent with physical affection (mean = 3, SD = 1.4). Internal consistency coefficients for the SSQ and ISSB were excellent and test-retest coefficients ranged from .66 to .72 suggesting the scales were stable. Support satisfaction scores were slightly negatively skewed (-1.2) but no skewness was found for the SSQ-N or the ISSB. The two SSQ scales, number and satisfaction were moderately correlated at .48 (p < .01, twotailed). Table 10 Descriptive statistics for the SSQ-N, SSQ-S, and ISSB  Social Support  Mean (SD)  Cronbach's a  1. Number of supportive people  3.9(1.99)  .82**  2. Satisfaction with support  5.01 (.87)  .96**  2.03 (.48)  .91**  Perceived support  Received support Frequency of supportive behaviors  ** p = .01 level (two-tailed) Correlations between Coping and Social Support In order to explore relations between coping and social support a correlation matrix of the variables was generated (see Appendix 5). Accommodating to the illness was positively related to approach (r = .34, p < .01, two-tailed), number of supportive others (r = .22, p < .05, twotailed), and received support (r = .28, p < .01, two-tailed). The perceived (satisfaction) and  98 received support scales were also positively related (r = .36, p < .01, two-tailed). Venting was negatively related to the perceived support scores with r = -.22 (p < .05, two-tailed) for the number of supportive others and r = -.28 (p < .01, two-tailed) for support satisfaction. Catastrophizing was positively related to focusing on symptoms (r =.45, p < .01, two-tailed), avoidance (r = .30, p < .01, two tailed), and venting (r = .39, p < .01, two-tailed). The Efficacy of Coping and Social Support In addition to describing their coping and social support efforts, respondents were asked to evaluate the degree to which each coping and support behavior helped them to: (1) recover from CFS and (2) control the symptoms. Thus, there are eight efficacy scales, two for each of the coping (IMQ, COPE) and social support (ISSB, SSQ-S) instruments. Efficacy scales and subscales were designed to parallel the "parent" measures and were aggregated into approach, avoidance, and venting subscales from the COPE and the four subscales found in the IMQ. For the ISSB and SSQ the efficacy of received support and support satisfaction are measured. In order to determine the usefulness of the efficacy measures psychometric testing was performed. Validity of Efficacy Measures Validity refers to the degree to which a scale measures what it is intended to measure. Of considerable interest is a scale's construct validity that answers the question: Does the scale adequately measure the concept under investigation? Construct validity may be assessed by investigating the measure's factorial structure and by examining the convergent validity of the instrument. Factor analysis (FA). The purpose of F A is to "summarize patterns of correlations among observed variables to reduce a large number of observed variables to a smaller number of factors..."(Tabachnick & Fidell, 2001, p. 582). Preliminary examination revealed the efficacy scores were normally distributed with no outliers of importance. A n exploratory principal components F A with Varimax rotation and Kaiser normalization was performed to assess the  99 underlying factor structures. The analyses demonstrated that each of the efficacy subscales and scales essentially reflected one factor (Appendix 6 & 7). Although analyses for some scales revealed more than one factor, typically the first factor produced a high eigenvalue and subsequent factors achieved eigenvalues of two or less. Furthermore, the scree plots demonstrated sharp breaks in trajectory between the first and subsequent factors which provided additional support for a one-factor solution (Tabachnick & Fidell, 2001). Assuming a one-factor solution, the variance explained by the coping and support efficacy scales ranged from 32% to 73%. The highest and lowest total variances were produced respectively by the problem-focused coping efficacy subscale, accommodate to illness - recovery (73%) and the emotion-focused coping efficacy scale, avoidance - symptom control (32%). In general scales measuring efficacy for recovery accounted for greater variance than the scales measuring efficacy for symptom control. Convergent validity. A scale's convergent validity rests on its ability to relate to other, . similar constructs. Personal control and treatment control are two variables included in the study which are likely to be similar to the concept of efficacy. As measured by Pearson's productmoment, correlations between perceived control and the efficacy measures were all significant ranging from .19. to .56. This pattern was not found for the frequency measures, which for the most part, were uncorrected with control. Reliability of Efficacy Scales In general, the alpha coefficients for the efficacy subscales were high. The lowest alpha was .75, which was produced by the emotion-focused coping subscale venting - efficacy for controlling symptoms. Test-retest correlations were obtained which ranged from .54 (venting efficacy for controlling symptoms) to .70 (maintaining activity - efficacy for recovery) suggesting the stability of efficacy evaluations.  100 In summary, although the efficacy measures were created in a rather unconventional manner, it appears that the scales are reliable and demonstrate some construct validity. The internal consistency of many of the scales exceeded .90, a figure that suggested considerable redundancy among items (Streiner & Norman, 1995). Test-retest correlations hovering around .50 showed that the scales possessed a degree of stability even after seven months. Such moderate correlations tend to support the validity of the scales because the episodic nature of CFS would likely result in changes in efficacy perceptions over time. Evidence for construct validity was found through significant correlations between the efficacy measures and measures of perceived control (personal and treatment). The correlations are high enough to indicate efficacy is related to control but not so high as to suggest the two instruments are measuring the same construct. Moreover, although the efficacy measures parallel the coping and social support items, the efficacy and frequency instruments appear to measure different attributes since the frequency measures did not correlate with perceptions of control. Although the small sample size renders factor analytic techniques tentative, a one-factor solution confirms the hypothesis that the scales are unidimensional. In addition, the scales accounted for a fair amount of variance in the scores. Efficacy of coping and social support: Descriptive statistics Comparison of the means for the 18 efficacy scales and subscales showed that the highest rating assigned to the coping and social support strategies was about three which corresponded to "somewhat effective" on the likert scale (see Table 11). The following strategies achieved a score of three or higher and were considered efficacious; (a) accommodating to the illness symptom control, (b) approach - symptom control, (c) perceived social support - symptom control, (d) accommodating to the illness - recovery, and (e) seeking information - symptom control. Because approach is composed of two subscales, both were assessed separately. Positive  • 101 reinterpretation - symptom "control was considered more efficacious than acceptance - symptom control. Specific behaviors considered efficacious are presented in Table 12. Table 11 Coping and social support strategies: In descending order of efficacy for recovering from CFS and controlling symptoms (N=100). Bolded text denotes strategies considered efficacious.  Efficacy scale or subscale  a  Mean(SD)  Cronbach's a  Accommodate to illness - symptom control  3.7 (.91)  .98  Approach - symptom control  3.2 (.99)  .90  Perceived social support - symptom control  3.1 (1.1)  .98  Accommodate to illness - recovery  3.1 (1.2)  .93  Seeking information - symptom control  3.0 (.80)  .81  Approach - recovery  2.9(1.2)  .95  Focus on symptoms - symptom control  2.7 (.80)  .84  Seeking information - recovery  2.7 (.98)  .89  Perceived social support - recovery  2.6(1.07)  .98  Maintaining activity - symptom control  2.4 (.66)  .91  Focus on symptoms - recovery  2.4 (.90)  .90  Venting - symptom control  2.3 (.73)  .75  Maintaining activity - recovery  2.3 (.85)  .96  Venting - recovery  2.2 (.91)  .87  Received social support - symptom control  2.1 (.73)  .94  Avoidance - symptom control  2.1 (.60)  .77  Avoidance - recovery  1.9 (.64)  .81  Received social support - recovery  1.8 (.65)  .95  "Respondents asked to rate the strategy for controlling symptoms or in recovering from CFS as; 1 = not applicable, 2 = not effective at all, 3 = sometimes effective, 4 = quite effective, 5 = very effective, 6 - extremely effective  102 Table 12. Efficacious coping and social support strategies: Two specific behaviors perceived most efficacious in each category  1. Accommodate (symptom control)  Mean (SD)  2. Approach (symptom control)  Mean (SD)  I manage my time so I don't overdo  3.9(1.2)  I learn to live with it  3.6(1.1)  I balance activity and rest  3.9(1.1)  I learn from the experience  3.3 (1.2)  3. Support satisfaction (symptom control)  Mean (SD)  4. Accommodate (recovery) Mean (SD)  M y innermost feelings are listened to uncritically  3.3 (1.4)  I know what's best for me  I'm appreciated as a person  3.3 (1.4)  I manage my time so I don't 3.24(1.4) overdo  5. Seeking information (symptom control)  Mean (SD)  I find out as much as I can about CFS  3.5(1.1)  I look for new information  3.2(1.1)  3.2(1.4)  Inferential Data Analyses Comparison of Frequency and Efficacy Measures for Predicting Outcomes To determine whether it is the efficacy, rather than the frequency of coping and social support efforts that is more predictive of outcomes multiple regression models were run and the adjusted R of the two modalities compared. For the most part efficacy did not outperform 2  frequency. One exception was physical functioning. Compared to the frequency of emotion- and problem-focused coping the efficacy scales predicted more variance in the physical functioning scores (see Table 13).  103 Table 13 Comparison offrequency and efficacy coping measures for predicting physical functioning F  Measure  Adjusted R  Frequency of emotion-focused coping  0.12  5.5**  Efficacy of emotion-focused coping (recovery)  0.17  6.5***  Efficacy of emotion-focused coping (symptoms)  0.13  5.9**  Frequency of problem-focused coping  0.03  1.9  Efficacy of problem-focused coping (recovery)  0.16  5 7***  Efficacy of problem-focused coping (symptoms)  0.18  6.5***  *p<.05  2  **p<.005 ***p<.0001 Differences in Illness Duration Groups  Data were further analyzed to determine if differences existed in outcomes, coping, social support and/or efficacy for subjects with short (mean = 2 years, SD =1) and long illness durations (mean =10.1 years, SD = 4.6). The two groups were found to be statistically equivalent for disability status and illness-related variables. However, with an average age of 51 years (SD = 8.9) the long duration group was older (F = 8.1, p =.005) than the short duration group (mean = 45 years, SD = 9.6) as calculated by A N O V A (F = 8.9, p = .005). In addition A N O V A demonstrated (F = 4.4, p = .04) that the short duration group used more antidepressants (mean = .77, SD = .89) than the long duration group (mean = .43, SD = .84). Outcomes Functioning. Using A N O V A a comparison of the two illness groups showed no differences in any of the SF-36 subscales or the super-ordinate scales, total mental health and total physical functioning (Table 14).  104 Table 14 Comparison of SF-36 raw scores for respondents with short (under four years) and long (four years and more) illness durations Short illness duration (N=31)  Long illness duration (N=69)  Mean (SD)  Mean (SD)  Physical functioning  18.5 (3.9)  18.1 (4.5)  Role-physical  4.1 (.63)  4.2 (.58)  Bodily pain  6.8 (2.2)  6.2 (2)  General health  12.1 (3.5)  12.2 (3.9)  Vitality  8.7 (2.6)  9.4 (3)  Social functioning  5(2)  5.4(1.8)  Role-emotional  5(1.2)  5.7(3.1)  Mental health  21.9(4.1)  22.4 (4.3)  Total mental health  44.2 (6.5)  45.8 (8.7)  Total physical functioning  42.3 (7.8)  42.2 (8.3)  Psychological distress. As demonstrated by A N O V A a comparison of the two illness groups showed no differences in psychological distress (see Table 15).  105 Table 15 Comparison of HADS scores for respondents with short (under four years) and long (four years and more) illness durations  Short duration group (N=31)  Long duration group (N=69)  Mean (SD)  Mean (SD)  Anxiety  6.7 (3.5)  6.6(4.1)  Depression  7.5 (3.1)  7.1 (3.2)  Health care visits. Because of the difficulty in interpreting transformed scores, the untransformed means of the health care visit scores were used in the analysis. Since the data were not normally distributed a non-parametric technique was used to test for statistically significant differences between the groups. Analyses showed that respondents with shorter illness durations made more visits to all practitioners, primarily conventional care, than respondents with longer illness durations (Table 16). Table 16 Comparison of health care visits (HCV) for respondents with short (under four years) and long (four years and more) illness durations  Short illness duration group (N=31)  Long illness duration group (N=69)  Mean (SD)  Mean rank  Mean (SD)  Mean rank  H C V Total  23.3 (23.3)  61.82  11.9(9.5)  45.4*  H C V Conventional  16.7(14.7)  61.71  8.6(7.5)  45.5*  6.7(11.7)  54.85  3.2(5.1)  48.5  H C V Alternative  *p = .009 (Kruskal-Wallis test)  106 Illness Representations As determined by A N O V A , no differences in illness representations were found between the short and long illness duration groups (see Appendix 8). Coping and Social Support As determined by A N O V A , no differences were found in the coping or social support variables between the short and long illness duration groups (see Appendix 9 &10). Perceptions of Efficacy Comparison of the two illness groups (see Table 17) demonstrated that respondents with a short illness duration rated more coping and social support strategies as "somewhat effective" than the long duration illness group. However, as determined by A N O V A (table 18) statistically significant differences in efficacy ratings were found for only four scales; (1) venting - recovery, (2) maintaining activity - recovery, (3) seeking information - recovery and (4) received social support - recovery.  107 Table 17 Efficacy of coping and social strategies for respondents with short (under four years) and long (fouryears and more) illness durations: Strategies considered" somewhat effective" in descending order of efficacy Long illness duration (N=69)  Short illness duration (N=31)  Mean (SD)  Efficacy scale  Mean (SD)  Efficacy scale  Accommodate - symptoms  3.7 (.99)  Accommodate - symptoms  3.6 (.88)  Accommodate - recovery  3.3 (1.1)*  Approach - symptoms  3.1 (1.06)  Approach - symptoms  .3.3 (.78)  Seek information - symptoms  3.2 (.74)  Approach - recovery  3.1 (1.15)*  Support satisfaction - symptoms  3.1 (.80)  Seek information - recovery  3.0 (.94)*  Support satisfaction - symptoms 3.07(1.25)  * differences between short and long duration group significant at p < .10 Table 18 Efficacy subscales that are statistically different: Short versus long illness duration groups Illness group  Efficacy scale VentingRecovery  a b  Maintain activity Recovery  a  b  Seeking information Recovery Received support Recovery  l  N = 44  b  N=17  a  b  a b  Mean (SD)  Significance  >4year <4year  2.0 (.87) ,2.5(2.5)  F = 5.1*  >4year < 4 year  2.2 (.84) 2.5 (.82)  F = 4*  >4year < 4 year  2.5 (.97) 3.0 (.94)  F = 4.2*  >4year <4year  1.7 (.58) 2.1 (.74)  F = 8.6**  *p<.05  **p<.005  108 Relations between Cognitions, Coping, Social Support, and Outcomes Multiple regression using cross-sectional data collected at time one (N=100) were used to determine the extent to which cognitions, coping, and support predicted functioning, psychological distress, and health care visits. Separate sequential regression models were run for each outcome. Demographic, illness-related, and control variables found to correlate significantly with outcomes (see Table 19) were forced into each model in successive steps. In the final step each of the outcome variables was entered (see Appendices 11 - 24). Table 19 Variables that correlated significantly with outcomes and were used in sequential multiple regression models Variable Demographic -  Mental Physical Health Function  M a l eg e n d e r #w o r kh o u r s .47 #s c h o o lh o u r s H i g h e s t g r a d e. 2 1 N o tw o r k i n g -.24  .36  Depression Anxiety Total Conventional H C V H C V  - # comorbidities.  Control  -.20 -.26 -A p p e a l disability - #medications - S l e e p i n gp i l l s - A n t i d e p r e s s - a. 2 n2 t s - H e r b a lp r e p s . -V i t a m i n s  -.21  -.22 -.24 -.24 .21  .31 -.38  Illnessrelated  Alternative H C V  .19  .39  .20 .21  . .  2  2  7  2 .40 .26  In the next analysis, coping and support variables found to be statistically significant predictors of outcomes were regressed on cognitive and emotional representations to identify significant determinants (see Appendix 25 & 26). All models were then re-run using longitudinal data (N=92) to control for baseline status of outcomes, and to identify variables important for adjustment over time (Appendices 27 - 40). In each model the score of the dependent variable at  109 time one was partialled out of the time two score and the residual regressed on the same predictors used in the cross-sectional analyses. For models that contributed statistically significant variances, the first-order correlations (r), standardized beta coefficients (b), and t values (t) of significant predictors are presented. Finally, to evaluate the stability of the results, cross-sectional and longitudinal analyses were repeated using time two (N=92) and time three data (N=85). Noteworthy departures from the results obtained from time one data were found only for the cross-sectional predictions of mental health, physical functioning, depression, and anxiety (see Appendix 41). Mental Health Cross-sectional analyses (time one data). Since no illness-related variable correlated with mental health, none was entered into the models. The combined variance in mental health provided by demographic and control variables was 15%. When entered in step three, cognitive representations added another 24%) (p < .005). A number of cognitive perceptions achieved substantial betas but only consequences was statistically significant with a standardized beta of -.39 (p < .0001). Emotional representations also contributed to the prediction of mental health contributing a variance of 12% (p < .001) revealing a beta of-.38 (p < .0001). The causal attribution subscales were unrelated to mental health. Emotion-focused coping contributed about 16% (p < .005) to the variance in mental health. Avoidance and venting were negatively related to mental health with betas of-.24 (p < .05) and -.20 (p < .05), respectively. Catastrophizing added 14% to the variance (p < .0001) and was negatively related to mental health (b = -.40, p < .0001). Problem-focused coping explained 13%) of the variance (p < .005) with a beta of -.29 (p < .05) for focusing on symptoms. The only coping subscale positively related to mental health was accommodating to the illness (b = .27, p < .05).  110 Perceived and received support were positively related to mental health and contributed to its variance. The greatest variance (11%, p < .005) was provided by perceived support with the satisfaction subscale achieving a beta of .39 (p < .005). Received support explained 5% of the variance (p < .05) in mental health (b = .24, p < .05). Repeat cross-sectional analyses (comparison of time one, two, and three data). Notable fluctuations in the adjusted variances and standardized betas of some of the variables were found over the three time points. The variance of cognitive representations was greatest at time two and there was a substantial change in the beta of personal control which rose from .02 at time one to .35 (p <.05) at time two. At time three the beta for personal control was found to be .14. The coping variables also fluctuated with the beta for approach rising from .16 at time one to .29 (p <.005) at time two. However, it dropped back to .19 at time three. Although the problem-focused strategy of maintaining activity produced a beta of only .07 at time one the beta rose to -.21 (p < .05) at time two, then dropped to -.007 at time three. Catastrophizing fluctuated in explained variance but maintained betas ranging from -.34 (p <.005) to -.67 (p < .0001) through the three time points. The satisfaction subscale of perceived social support demonstrated some instability ranging from .39 (p <.0001) at time one to .03 at time, three. Longitudinal analyses. After controlling for pre-scores, demographics, and control variables, emotional representations remained a significant predictor of mental health explaining 3% of the variance (p < .05) in the residualized scores (b = -.22, p < .05). In addition, emotionfocused strategies predicted change over time with a variance of .07 (p < .05). Approach and avoidance were differentially related to mental health with betas of .24 (p < .05) and -.27 (p < .005), respectively. As well, catastrophizing explained 3% of the variance (p < .05) and was negatively related to mental health (b = -.23, p < .05). Neither perceived nor received social support was predictive of residualized scores.  Ill Physical functioning Cross-sectional analyses (time one data). Demographic variables including highest grade achieved and number of paid work hours were positively related to physical functioning and contributed 22% (p < .0001) to its variance. A n additional 19% (p < .005) was provided by cognitive representations. Statistically significant betas were achieved by consequences (b = -.22, p < .05), identity (b = -.19, p < .05), and personal control (b = .26, p < .05). Emotional representations, coping, and perceived social support did not contribute any statistically significant variance to physical functioning. However, received support was positively correlated with physical functioning (b = .22, p < .05) and added 4% (p < .05) to its variance. Repeat cross-sectional analyses (comparison of time one, two, and three data). As with mental health the adjusted R of cognitive representations varied somewhat over the three time 2  points. In addition, variations were noted in the individual subscales including a decrease in the beta of consequences from -.22 (p <.05) at time one to .01 at time two. The beta of consequences at time three was -.16. Personal control dropped from .26 (p < .05) to -.03 and remained near zero at time three. On the other hand, the beta of cyclic timeline rose from -.001 at time one to .25 (p <.005) at time two but dropped to .03 at time three. For the coping variable of catastrophizing betas ranged from -.14 at time one, to -.27 (p < .005) at time two to -.01 at time three. Longitudinal analyses. No variable was able to predict residualized scores in physical functioning. Depression Cross-sectional analyses (time one data). Demographic, illness-related, and control variables explained about 19% of the variance in depression scores. However, cognitive representation added another 36% (p <.0001). Consequences was the most potent predictor of depression with a beta of .49 (p <.0001). Depression was negatively related to personal control (b  112 — -.21, p < .05). Similarly, emotional representations added 16% to the total variance in depression producing a beta of .42 (p < .0001). Eight percent of variance (p < .05) in depression was explained by emotion-focused coping. Approach was negatively related (b = -.22, p < .05) and avoidance positively related to depression (b = .24, p < .05). Catastrophizing contributed 22% (p < .0001) of the variance with a beta of .48 (p < .0001). Problem-focused coping was responsible for 17% (p < .0001) of the variance in depression scores with differential associations demonstrated by accommodating to the illness (b = -.21, p < p .05) and focusing on symptoms (b = .45, p < .0001). Perceived (b = -.27, p < .005) and received support (b = -.29, p < .005) were negatively related to depression contributing 10% (p < .005) and 8% (p <.005) of the variance, respectively. Repeat cross-sectional analyses (time one, two, and three data). Slight fluctuations in the variances and betas of catastrophizing were noted between time two and time three. Longitudinal analyses. No variable was able to predict residualized scores in depression. Anxiety Cross-sectional analyses (time one data). About 4% (p < .05) of the variance in anxiety was contributed by demographic variables with an additional 3% produced by antidepressant use. Twenty percent (p < .05) of the variance was explained by cognitive representations, although no individual perception demonstrated a statistically significant beta. Emotional representations added 23% (p < .0001) to the variance and demonstrated a positive relationship with anxiety (b = .48, p<.0001). Emotion-focused coping explained 22% (p < .0001) of the variance in anxiety with the following subscales producing significant betas; approach (b = -.37, p < .0001), avoidance (b = .23, p < .05), and venting (b = .30, p < .005). Catastrophizing was also a strong predictor contributing 13% (p < .0001) of the variance in anxiety with a beta of .38 (p < .0001). Nineteen percent (p <.0001) of the variance was explained by problem-focused coping with the following  113 subscales producing differential associations: (1) maintaining activity was positively related to anxiety (b = .22, p <.05) and accommodating to the illness was negatively related (b = -.33, p < .005). Of the social support variables, only perceived social support contributed any statistically significant variance (9%, p < .005) with betas negatively related to anxiety (b = -.25, p < .05). Repeat cross-sectional analyses (time one, two, and three data). Although the variance explained by cognitive representations was similar between the three time points, the beta of personal control rose from -.09 at time one to -.32 (p < .05) at time two. It dropped back to -.09 at time three. The emotion-focused coping strategy of venting also varied with betas of .30 (p < .005) at time one dropping to .18 at time two. Longitudinal analyses. No variable was able to predict residualized scores in anxiety. Health Care Visits Cross-sectional analyses (time one data). Neither the cognitive nor the coping variables explained any variance in the total number of visits, or the number of visits to conventional or alternative health care providers. The only predictor of total health care visits was the total number of medications used which accounted for about 15% (p < .0001) of the variance. Antidepressant use was associated with conventional care visits (b =.22, p < .05) while the use of herbal preparations and vitamins explained 13%> (p < .005) of the variance in alternative care visits. However, only the use of herbal preparations achieved a significant beta (.26, p < .05). Perceived support (number of supportive others), was positively related to conventional care visits accounting for 7% (p < .05) of the variance in scores. A l l analyses were re-run with the sample divided into short and long illness duration groups but no significant predictors (other than medication use and support) emerged. Longitudinal analyses. Medication use continued to predict the total number of health care visits over time (R = .04, b = .24, p < .05). As in the cross-sectional analysis the use of vitamins 2  114 and herbal preparations explained some variance in alternative care visits (R = .09, p < .005) with a significant beta produced by herbal preparations (.28, p < .005). However, cognitive representations also achieved significance as a predictor (R = . 12) with causal attributions of immune dysfunction producing a beta of .21 (p < .05). Furthermore, emotion-focused coping contributed significantly to the variance in total health care visits (R = .05, p < .05) and visits to conventional providers (R = .07, p < .05). Approach was negatively linked to both variables producing a beta of -.21 (p < .05) for total visits and a beta of -.24 (p < .05) for conventional care visits. Although perceived support scales again explained some variance in health care visits, the type of care and direction of the correlation differed from the cross-sectional analysis. Perceived support produced differential associations with the number of supportive others negatively related (b= -.20, p < .05) and satisfaction with support positively related (.20, p < .05) to the number of visits made to alternative care providers. Relations between Cognitions, Coping, and Social Support Although it has been shown that cognitions, as reflected by representations of illness, were directly related to outcomes, analyses were conducted to determine if representations were also related to coping and social support. Thus, cognitive and emotional representations were regressed on the coping and social support variables found to be statistically significant predictors of outcomes. Cross-sectional analyses (time one data). Cognitive representations explained no statistically significant variance in emotion-focused coping. However, emotional representations contributed 7% (p < .005), 9% (p < .005), and 10% (p < .005) respectively to the variance of the three emotion-focused subscales approach, avoidance, and venting. Betas demonstrated the emotional representations subscale was negatively related to approach (b = -.26, p < .005) and positively related to avoidance (b = .30, p < .005) and venting (b = .32, p < .05). Cognitive and emotional representations respectively explained 24% (p < .05) and 30% (p < .0001) of the  115 variance in catastrophizing. Two perceptions positively related to catastrophizing included consequences (b = .35, p < .005) and emotional representations (b = .55, p < .0001). For problem-focused coping, illness representations explained no variance in maintaining activity or accommodating to the illness. However, both cognitive and emotional representations were predictive of focusing on the symptoms, respectively explaining 24% (p < .005) and 13% (p < .0001) of the variance in the scores. Consequence was the most salient cognitive perception and was positively related to focusing with a beta of .29 (p < .005). Emotional representations was also positively related to focusing (b = .36, p < .001). Variances in the perceived and received social support scores were explained by cognitive perceptions (R = .25, p < .005) and emotional representations (R = .19, p < .05). Attribution to 2  2  risk factors was negatively related to perceived support satisfaction (b = -.31, p < .005) while attribution to immune dysfunction was positively related to receiving support (b =.25, p <.05). Emotional representation was negatively related to perceived support satisfaction (b = -.28, p < .005) but unrelated to received support. Longitudinal analysis. Only one statistically significant model emerged from the longitudinal analyses. Emotional representation was positively related to venting (b = .43, p < .005) and explained 6%> (p < .005) of its variance. Determinants of Predictive Perceptions The preceding analyses identified the following perceptions as directly or indirectly related to illness outcomes; consequences, identity, emotional representations, personal control, immunity attributions, and risk factor attributions. To permit greater contextualization, an attempt was made to isolate personological variables that were associated with predictive perceptions. A correlation matrix was prepared that included the perceptions, and demographic, illness-related, control, personality, and psychiatric status variables. The results demonstrated that the perceptions consequences, personal control, and emotional representations almost  116 exclusively correlated with personality and psychiatric status. No significant correlations were found between the personological variables and identity, immune attributions, and/or risk factor attributions. Sequential regression models were then run using the statistically significant variables as predictors of the perceptions. As can be seen from Table 20 neuroticism explained 6% (p < .05) of the variance in consequences and produced a beta of. 17 (p < .05). Neuroticism was also linked to emotional representations with an R o f .23 (p < .0001) and a beta of .45 (p < .0001). To 2  a lesser extent the trait of agreeableness explained some variance in personal control. Table 20 Determinants of predictive perceptions •——  Dependent variable  Step/Predictor  Consequences  1 Personality traits Neuroticism  Adj. R  0.06*  *  r  b  t  0.24  0.17*  1.6  .28  .23*  2.2  2 Psychiatric diagnosis Phobia Personal control  1 Personality traits  0.02 .14*  Agreeableness Emotional representations 1  Demographic Highest grade  .04*  ns  2 Personality traits Neuroticism 3 Psychiatric diagnosis *p<.05  **p<.005  ***p<.0001  .23*** 0.03  0.51  .45*** ns  4.9  117 Changes in Outcomes over Time In order to identify the extent to which respondents' functioning, psychological distress, and health care visits changed over time, scores from the four time points were subject to repeated measures analysis. The sample as a whole (see Table 21) did not demonstrate any statistically significant changes in functioning as demonstrated by the SF-36 or any changes in anxiety and depression over the study period (approximately 21 months). However, a decline was shown over time for total number of health care visits and visits to conventional providers, most notably between time 1 and time 2. When the respondents were stratified as to illness duration some differences emerged. For the SF-36 subscales the short illness duration group experienced a statistically significant improvement in physical functioning that was not evidenced by the long illness duration group. In addition, the long duration group reported a decline in vitality while the short duration group showed a non-significant trend towards greater vitality (Table 22). The analyses of health care visits demonstrated that both illness groups reported fewer health care visits as time progressed. Between-groups comparisons suggested that compared to the short duration group the long duration group (N=44) made fewer visits to alternative care practitioners.  118 Table 21 Illness outcomes: Means and SD over four time points and statistical significance of the differences between and within groups Time 2 Mean (SD)  Time 1 Outcome Variable Mean (SD)  Time 3 Mean (SD)  Time 4 Mean (SD)  Within Within Subjects Subjects Difference Contrasts  Between Subjects Difference  Total mental health  47.9 (6.6) 45.5 (8.0) 45.5 (7.4) 45.8(9.1) ns  na  na  Total physical function  42.7 (7.3) 41.9 (7.6) 42.0 (7.4) 42.9 (9.9) ns  na  na  Depression  7.3 (3.1)  8.1(4.4)  7.8 (4.0)  7.7 (4.3)  ns  na  na  Anxiety  6.1 (3.8)  6.4 (3.9)  6.1 (3.4)  6.4 (4.2)  ns  na  na  Total number HCV (raw score)  14.9(11)  15 (20.8)  10.4 (8.8)  10(8.3)  F = 5.1* (lowerbound p)  d  F= 16.2  na  i l l > 4 years  1.0 (.28)  .96 (.39)  .92 (.31)  .87 (.31)  F = 3.6*  p = 9 9**  ns  i l l < 4 years  1.2 (.27)  1.1 (.29)  .98 (.42)  1.0 (.45)  F = 4.3** F = 5.0*  11.3 (10.2) 8.9 (7.3)  7.6(6.5)  F = 5.6* (lower bound p)  d  ab  ac  Conventional HCV 10.9 (7.9) (raw score)  d  d  H  F = 6.7*  ab  i l l > 4 years  .93 (.27)  .92 (.37)  .88 (.31)  .79 (.29)  F = 2.9*  d  ac  i l l < 4 years  1.1 (.32)  .98 (.44)  .88 (.41)  .90 (.41)  ns  ns  4(5.7)  3.76 (4.9)  1.4 (3.1)  2.4(3.1)  ns  na  Alternative HCV (raw score) i l l > 4 years  .38 (.43)  .13 (.33)  .12 (.28)  .23 (.37)  F = 6.7*  e  ac  ill<4years  .57 (.55)  .45 (.60)  .38 (.44)  .41 (.50)  ns  ns  *p<.05  **p<.005  b  N= 44 :  C  N=17  ***p<.0001  d  linear  Quadratic  ns  na  F = 4.1*** F = 6.7*  ab  "Transformed scores  4**  119 Table 22 SF-36: Significant differences for the physical functioning and vitality subscales between and within groups SF-36 Subscale  Illness group  Time 1 Mean (SD)  Time 2 Mean (SD)  Time 3 Mean (SD)  Time 4 Mean (SD)  Physical >4 Functionin year g <4 year  18.5(4)  18.7(4)  18.7(4.2)  18.9(4.7) ns  na  18.5(3.9)  19.1(4.4)  19.7(4.7)  20.8(5.2)  F=4*  a  Vitality  9.5(2.9)  7.5(2.8)  7.7(2.6)  7.5(3.1)  F=9.3***  a  b  >4 year <4 year  a  b  a  N = 44  b  Within Within Subjects Subjects Difference Contrast  Between Subjects Diff. ns  F = 6.4  F=16.6*** ns F= 8.9** na  C  d  .8.9(2.5)  N = 1 7 'linear  9.2(3.3)  d  10.1(3.3)  10.7(3.9)  ns  quadratic * p < .05 ** p < .005 ***p<.0001  Changes in Coping and Social Support over Time Repeated measures analysis was used to determine if respondents altered their coping and social support strategies over time (see Table 23). During the course of the study most scores decreased slightly and statistically significant declines were observed for the emotion-focused strategies; approach and venting, and the problem-focused strategies; accommodate to the illness, focus, and seeking information. With the exception of accommodate to the illness, significant changes transpired between time one and time two. For accommodate to the illness significant changes occurred between time two and time three.  120 Table 23 Coping and social support scores over four time points (approximately seven months between time points) and statistical significance of the differences Time 1  Time 2  Time 3  Time 4  Within Subjects Difference  Within Subjects Contrasts  Coping/Support Subscale  Mean (SD)  Mean (SD)  Mean (SD)  Mean (SD)  Significance  Significance  Approach  3.0 (.62)  2.9 (.54)  2.8 (.60)  2.8 (.65)  F = 4 9**  a  Avoidance  1.8 (.37)  1.7 (.42)  1.7 (.35)  1.7 (.41)  ns  na  Venting  2.2 (.59)  2.1 (.59)  2.0 (.56)  2.0 (.69)  F = 3.0*  a  Catastrophizing  11.5 (6.6)  10.7 (6.5)  10.3 (6.2)  11.6 (7.8)  ns  na  Maintain Activity  3.0 (.70)  2.9 (.67)  2.9 (.71)  2.9 (.68)  ns  na  Accommodate  4.3 (.61)  4.2 (.63)  4.1 (.62)  4.2 (.69)  F = 3.1*  b  F = 4.8*  Seek information  3.7 (.87)  3.5 (.93)  3.3 (.98)  3.2(1.0)  F = 22 \***  a  F= 27.4  Focus on symptoms  3.5 (.86)  3.3 (.83)  3.2 (.88)  3.1 (.95)  F = 7 4**  a  F = 13 g***  # of supportive other  4.0 (2.1)  3.8(2.1)  3.8 (2.2)  3.9 (2.1)  ns  na  Support Satisfaction  4.9 (.87)  5.0 (.75)  4.8(1.1)  4.9(1.0)  ns  na  Received support  2.0 (.49)  2.0 (.47)  1.9 (.51)  1.9 (.51)  ns  na  linear  b  quadratic  * p < .05  **p<.005  F = 11.7**  F = 6.1*  ***p<.0001  Changes in the Perceived Efficacy of Coping and Social Support over Time Comparison of respondents (N=61) at time 1 and time 4 revealed that 77% (14/18) of the efficacy measures demonstrated a drop in scores and 22%> (4/18) remained essentially unchanged. Using matched pairs t-tests, statistically significant declines (Table 24) were shown for the following subscales: (1) approach - symptoms, (2) seek information - symptoms,  121 (3) approach - recovery, (4) focus on symptoms - symptoms, and (5) seek information recovery. At time point four only the following strategies retained a score of three or higher; (1) accommodate to the illness - symptoms and (2) satisfaction with support - symptoms. Table 24 Differences in matched pairs (N=61) ratings of the efficacy of coping and social support: Statistically significant declines from time 1(T1) to time 4 (T4)  Efficacy scale  Tl:Mean(SD)  T4: Mean (SD) t  Approach - recovery  2.8(1.1)  2.5 (1)  2.1*  Approach - symptoms  3.1 (.99)  2.9 (.96)  2.7**  Seeking information- symptoms  3.0 (.71)  2.6 (.85)  4 2***  Seeking information- recovery  2.7 (.98)  2.3 (.97)  Focus on symptoms - symptoms  2.8 (.76)  2.6 (.85)  2.4*  *p<.05 **p<.001 ***p<.0001 Chapter Summary This chapter outlined the results of the statistical analyses. Although only 61 of 100 respondents completed all four data sets no significant differences were found between these respondents and respondents who completed one to three sets of data (N=39). Characteristics of the respondents were described and the psychometric properties of the instruments presented. Preparatory to inferential analyses, the central tendencies, dispersion, and distribution of the data were calculated and the results detailed. Results from hypothesis-testing revealed that respondents with short illness durations visited health care practitioners more often than respondents with long illness durations. The short duration group also perceived more coping and social strategies to be efficacious.  122 Illness representations, coping, and social support were shown to predict functioning and psychological distress in the cross-sectional analyses. The effect sizes of certain cognitions, especially perceptions of personal control, fluctuated at various time points. Although fewer statistically significant results emerged when longitudinal data were used, an important finding was that cognitions and coping predicted health care visits. Findings demonstrated that in addition to being directly related to outcomes, the association between negative perceptions and outcomes may be mediated by coping and social support. Pessimistic perceptions predicted the use of strategies associated with unfavorable outcomes or were negatively related to advantageous strategies. Trait neuroticism was shown to be a statistically significant predictor of the negative perceptions consequences and emotional representations of distress. Examination of outcomes over time revealed that, for the most part, respondents' level of psychological distress remained unchanged. However, respondents in the short illness duration group experienced improvement in physical functioning, while respondents with a long illness duration reported a decline in vitality. Both illness groups showed a tendency to visit health care practitioners less often as time progressed. In addition, although coping and social support behaviors were quite stable, their efficacy (particularly seeking information) seemed to dwindle over time. In the next chapter, the results are discussed as they pertain to the literature.  123 C H A P T E R FIVE: Discussion In this chapter a discussion of the study findings is presented. The characteristics of the respondents are detailed and compared to samples from other studies to determine the generalizability of the results. As guided by the theoretical framework, the illness representations and self-management behaviors of the respondents are detailed and the interplay between the constructs explored. Respondents' perceptions of the efficacy of self-management are described as are changes in the adaptive process over time and at different points in the illness trajectory. Characteristics of the Sample Demographics The sample (N=100) included a high percentage of women, Caucasians, and the welleducated. Although differing from profiles of CFS patients sampled from the community, the characteristics were consistent with data gathered from study subjects in secondary- and tertiarycare settings, probably reflecting the effects of referral filtering (Friedberg et al., 2000; Komaroff et a l , 1996; Manu, Lane, & Matthews, 1992; Wilson et al., 2001). Almost 60% of the study respondents were married or in a common-law relationship which is similar to the 40% to 60% rate found in both community and clinical studies (Jason et al., 1999; Buchwald & Garrity, 1994). The study respondents were at the upper limit of the age range (49) associated with CFS and, at 42 years, had been afflicted later in life than has been reported in epidemiological surveys (Jason et al., 1999; Reyes et al., 1997). However, the sample's age and onset were comparable to subjects acquired by university-based, referral centers in Boston and Miami (Wilson et a l , 2001). The study sample's income (45% earned less than $20,000 a year) was considerably lower than the income reported by Reyes et al. (1997) in their surveillance of four American cities. In the Reyes et al. (1997) study, only 15% of CFS patients earned less than $20,000 a year. However, the inconsistency in income may be related to differences in employment. Although difficult to  124 verify because Reyes et al. (1997) did not provide information on employment status, the 76% unemployment rate in the present study was fairly high, in excess of the 37% - 60% figure reported by other referral centers (Buchwald et al., 1997; Lloyd, Hickie, Boughton, Spencer, & Wakefield, 1990; Lane et al., 1991; Wilson et al., 2001). In addition, only 58% of the respondents received disability pensions. The respondents' high unemployment begs the question: "Why are so many respondents not working?" Using logistic regression, with outcome, demographic, illness-related, and control variables as predictors, it was shown that 'not working' was predicted by impaired physical functioning as reflected by the SF-36 (Wald 6.8, R = -.21, p =.008) and older age (Wald 5.9, R = .19, p = .01). Tiersky et al. (2001) also found older age was associated with unemployment. The respondents' physical functioning score (mean = 41) corresponded to a physician-assessed rating of "unable to work but self-caring" reported by Myers and Wilk (1999). Clinical Profile About 95% of respondents experienced post-exertional malaise, a common symptom of CFS. In addition, 49% of the sample reported a sudden illness onset, a figure somewhat higher than the 37.5% rate observed by Jason et al. (2000) in a community sample. However, contrary to Jason et al.'s (2000) findings, respondents in the present sample with an insidious onset were no more likely to report psychiatric illness than those with a sudden onset. With a mean illness duration of 7.9 years the sample was chronically ill. Fifty-two percent reported a non-psychiatric co-morbidity with the average being two conditions per person. Including all self-reported co-morbidities each person averaged 1.1 (SD=1.36) conditions. This was less than the average number of co-morbidities (2.4 - 3.4) found in two community-based studies (Heijmans & de Ridder, 1998a; Jason et al., 2000).  125 Illnesses reported in the present sample were of a chronic nature but, with one exception, were not life-threatening. Twenty-two percent (22/100) of respondents suffered other medicallyunexplained syndromes including irritable bowel syndrome (2/100), multiple chemical sensitivities (1/100), and fibromyalgia (FM). F M was pre-eminent with a rate of 19%, well within the 12%> to 24% range reported for CFS patients sampled from community and tertiary care settings (Bombardier & Buchwald, 1996; Buchwald et al., 1996; Jason et al., 2000). Unlike Natelson and Lange's sample (2002), CFS respondents in the present study who reported another medically-unexplained syndrome were no more likely to have a psychiatric diagnosis than respondents with CFS alone. Psychiatric Co-morbidities Seventy percent (70/99) of the respondents met the criteria for the lifetime occurrence of at least one psychiatric disorder. While this is greater than a rate of 21 % found for healthy controls administered psychiatric interviews in CFS studies (Torres-Harding et al., 2002; Lawrie, et al., 1997), it was not an unusual finding in the CFS population. In two clinical studies and a community survey, using structured Diagnostic Interview Schedules-III-A and IV, 80% to 82% of chronic fatigue and CFS subjects were diagnosed as having had a psychiatric disorder in their lifetime (Buchwald et al., 1997; Jason et al., 1999; Lane et al., 1991). Similarly, the rate of current psychiatric disorder (45%) found in the present study was consistent with rates of up to 55%) reported by other researchers using psychiatric interview techniques that included computerized modalities (Buchwald et al., 1997; DeLuca et al., 1997). Moreover the 23% (23/99) rate of having two or more active psychiatric diagnoses was similar to the rate of 18%> found by Lane et al. (1991). Unlike other investigations in which Major Depressive Disorder (MDD) was the most prevalent diagnosis, pain disorder (with a lifetime prevalence of 45%) was diagnosed most often in the study respondents. However, this diagnostic category is highly inclusive requiring that the  126 following five criteria be met; (1) pain in one or more anatomical sites, (2) pain causes clinically significant distress or impairment in functioning, (3) psychological factors are judged to play an important role in pain onset, severity, and/or maintenance, (4) pain is not feigned, and (5) pain is not better accounted for by affective disorders, psychosis, and/or dyspareunia (American Psychiatric Association, 1994). In the present study the usefulness of pain disorder as a diagnostic category must be considered limited since it overlaps with CFS symptoms and symptoms of co-morbid illnesses. If pain disorder is excluded only 45% (45/99) of the respondents meet the criteria for lifetime psychiatric disorder with 33% (33/99) having experienced an episode in the year prior to screening. About 37% (37/99) of respondents had at least one episode of M D D in their lifetime, which was higher than the 3.7% to 6.7% rate reported by Robins et al. (1984) in three American cities evaluated by lay interviewers using the DIS. However the respondents' rate was similar to that reported by other CFS referral centers (Lane et al., 1991). For 54% (20/37) of those with M D D the onset occurred after developing CFS, suggesting a reactive disorder, perhaps brought on by resultant disability, as hypothesized by Ray (1991). In 37% (14/37) of the M D D group the disorder persisted and was present within 12 months prior to screening. Somewhat concerning was the 28% (4/14) of respondents who met the criteria for M D D but who did not self-report the diagnosis and were not receiving therapy. This group may have an undetected depression, a proportion reminiscent of the 35% rate of undiagnosed psychiatric disorders found by Deale and Wessely (2000). The following anxiety disorders were diagnosed in the respondents; panic disorder (PD, lifetime rate = 33%; current rate = 14%), phobia (lifetime rate = 9%; current rate = 5%), and generalized anxiety disorder (GAD; lifetime rate = 13%; current rate = 10%). Respondents' PD and G A D rates were higher than those reported by Robins et al. (1984) for the general population (PD: lifetime rate = 1.5%; G A D : lifetime rate = 4%). However similar rates for PD and G A D  127 have been reported in hospital outpatient samples of CFS patients (Johnson, DeLuca, and Natelson, 1996c). Other disorders evidenced by the respondents that did not differ from rates found in the general population included dysthymia (2%), somatization disorder (2%), and hypomania (1%). The sample's low SD rate was probably due to the fact that patients were instructed to code symptoms of CFS in the CDIS-IV as caused by a physical illness. Psychiatric co-morbidity in CFS and other fatiguing illnesses. Although the respondents' rates of psychiatric morbidity were similar to rates reported by other CFS researchers, the rates for a number of diagnoses are higher than has been reported in the general population (Kruesi et al., 1989; Robins et al., 1984). Why CFS patients have elevated rates of psychiatric morbidity has been an ongoing question. However, this phenomenon has been reported in patients with other fatigue-related conditions. In a population of English patients registered at various health centers (N=2798) Pawlikowska et al. (1994) noted an association between the number and severity of fatigue symptoms and psychological morbidity. The authors concluded that fatigue was not caused by psychological morbidity, but that the two conditions overlapped and the "concordance between fatigue and psychological morbidity, and between the chronic fatigue syndrome and psychiatric disorder, is inevitable given the similarities of the criteria and measures used to define them" (Pawlikowska et al., 1994, p. 765). In addition to fatigue, concentration difficulties, sleep disruption, mood disturbances, and memory impairment are also experienced by CFS patients which overlap with criteria used to diagnose psychiatric disorders (Lloyd, 1998). Indeed, another fatiguing illness, multiple sclerosis (MS), exhibits an affinity with psychiatric disturbance. Lifetime prevalence estimates for M D D in a population of M S patients ranged from 47% to 54% (Sadovik et al., 1996; Joffe et al., 1987; Minden et al., 1987). Some researchers have attributed depression to the psychological stress associated with impairment and  128 lack of a cure (Aikens et al., 1997). Others have suggested that depression is due to illnessrelated brain changes (Whitlock & Siskind, 1980). In a study comparing CFS and M S patients (N=65), CFS patients (N = 69) were not found to differ on Axis I or Axis II psychiatric disorders as measured by the Millon Clinical Multiaxial Inventory - Revised (Pepper, Krupp, Friedberg, Doscher, & Coyle, 1993). However in structured psychiatric interviews CFS patients revealed a trend toward higher rates of lifetime prevalence of psychiatric disorder, particularly with respect to the rate of a current episode of MDD. When addressing psychiatric rates in CFS, a few commentators have pointed to the heterogeneity of the CFS population. The lack of a definitive biomarker for diagnostic testing and the syndromal nature of the case definitions have created a heterogeneous population that could include patients with atypical depression and/or somatoform disorder (Lane et al., 1991; Wilson et al., 2001). Several studies have identified differences between patients with CFS and M D D that may enable clinicians to more precisely screen patients. The depressed are more likely to feel guilt, express suicidal ideation, experience low self-esteem, make internal attributions for negative interpersonal experiences, have fewer somatic complaints, and produce higher scores on the physical and role limitations subscales of the SF-36 (Evengard et al., 1999; Moss-Morris & Petrie, 2001; Powell et al., 1990). Differentiating CFS and somatization disorder (SD) may be more difficult (Natelson & Lange, 2002). A diagnosis of SD in CFS depends on whether the evaluator believes CFS symptoms are of a psychiatric nature. If so, most CFS patients meet the criteria for SD. When C D C symptoms of CFS were considered physical, investigators who limited their population to CFS patients with no history of psychiatric illness reported the SD rate dropped from 55% to 12% (Johnson et al., 1996a). In another group of patients whose scores were re-analyzed by Johnson et al. (1996a), SD declined from 10% to 3%. Despite the decrease in SD when CFS  129 symptoms were coded as physical, patients were still more likely to meet the criteria for SD than healthy controls. However, Johnson et al. (1996a) did not find a difference between CFS and M S patients for full syndromal SD, suggesting the disorder may be related to the experience of chronic illness. Although Johnson et al. (1996a) questioned the utility of SD in a controversial illness like CFS, Natelson (2002) suggested that SD may be diagnosed in a CFS patient with lifelong multisystem complaints, but not in a previously asymptomatic patient with symptoms following an acute influenza-like illness. On the other hand, Demitrack (1991) opined that the overlap between psychiatric disorders and CFS may reflect the existence of a shared, final, biological pathway, the hypothalamic-pituitary-adrenal axis (HPA). Once compromised by infectious or non-infectious agents the H P A axis may become incapable of mounting an adequate defense to stressors. Consequently, a relative or absolute reduction of adrenal glucocorticoids ensues, impeding the counter-regulatory effects of the H P A axis resulting in a secondary biological response of sleep dysregulation and immune activation. At present explanations for CFS patients' psychiatric morbidity remain somewhat speculative. Ensuring that the instruments used to diagnose psychiatric disorders do not include items that overlap with CFS symptoms may allow investigators to achieve greater precision when measuring psychopathology. Development of an objective diagnostic test or further refinement of the case definition may provide greater homogeneity in samples while continuing work on the neuroendocrinology of CFS may lead to a better understanding of the etiology of the syndrome. In the meantime, clinicians might adopt a pragmatic approach, as suggested by Salit (1996). The presence of a co-morbid psychiatric disorder may be viewed as an exigency that provides more management options than a diagnosis of CFS alone (Salit, 1996).  130 The Role of Cognitions and Coping in Adjusting to CFS Having characterized the study sample and demonstrated that it is representative of CFS patients from tertiary care settings, it is time to turn the reader's attention to findings bearing on the purpose of the study which is to describe patients' adjustment to CFS. Guiding the investigation is the self-regulatory model developed by Leventhal et al. (1984) which proposed that illness representations affect outcomes directly and indirectly through coping. Presumably, differences in representations and coping can explain why some individuals are more disabled than others. Based on this perspective it was hypothesized that highly negative illness representations would be related to unfavorable outcomes and predict the use of maladaptive coping strategies, while positive perceptions would be related to favorable outcomes and predict adaptive coping. Illness Representations In the following section the respondents' illness representations are described and compared to those of other illness groups. As detailed in the self-regulatory model, individuals develop parallel cognitive and emotional representations when confronted by health threats. Cognitive representations consisting of identity, timeline, consequences, control/cure, and causal attributions are meant to deal with the objective features of the illness. On the other hand, emotional representations reflect an individual's emotional reactions to the illness. Instrumentation The IPQ-R was used in the study to measure respondents' cognitive and emotional representations. It is a revision of the IPQ, an instrument that has been used in a number of studies. Unfortunately, due to its recent development only one study using the IPQ-R has been published (Moss-Morris et al., 2002). In order to allow a comparison of the present sample's illness representations with those of other illness groups, the IPQ-R was re-scored so that it replicated the IPQ. As stated by the authors, items from the original IPQ scale were included  131 unchanged in the IPQ-R (Moss-Morris et al., 2002). Direct comparison of the two questionnaires verified this claim. Therefore, transforming the IPQ-R into the IPQ was accomplished by excluding the newly added items. The four causal attribution subscales from the IPQ-R were "deconstructed" and each belief treated as a separate scale (except for "other people's behavior" which wasn't included in the IPQ-R). The following subscales were re-scored and are equivalent to those found in the IPQ; identity, consequences, timeline (acute/chronic), control, and causal attributions (7 subscales). Data analyses revealed that correlations between the IPQ and IPQ-R subscales for the. identity, consequences, and timeline (acute/chronic) subscales exceeded .95 (p < .01, two-sided). Cure/controllability produced a correlation of .80. Correlations between the causal attribution subscales were not performed due to substantial differences in item aggregation. Unfortunately, re-scoring necessitated the exclusion of coherence, emotional representations, and timeline (cyclic) and did not allow differentiation between treatment and personal control. Thus, IPQ-R scores are also included in the discussion since the measure represents a fuller explication of the illness representation construct. Perceptions of Illness: CFS versus Other Chronic Illnesses Respondents' IPQ scores for identity, timeline (acute/chronic), consequences, and control closely approximated those generated by CFS patients tested by Weinman et al. (1996). Compared to cognitive representation scores obtained from patients with diabetes (N=88), rheumatoid arthritis (RA; N=22), and chronic pain (N=60), the respondents in the present study had the strongest illness identity, considered the consequences most serious (men more than women {F = 4.4, p < .05}), and perceived the most chronic timeline. Marginally lower scores for control/cure were found for the CFS respondents compared to the other illness groups (Weinman etal., 1996).  132 These findings were partially confirmed when the IPQ-R scores of the present sample were compared to those obtained from chronic and acute pain patients (Moss-Morris et al., 2002). The CFS group demonstrated a considerably stronger illness identity, the perception the illness was cyclic, and the belief the consequences were more serious. In addition, the emotional representations of the CFS sample reflected the most distress and respondents experienced the least illness coherence. For timeline (acute/chronic) and treatment control the CFS respondents' scores were intermediate between the two pain samples. They perceived greater chronicity and less treatment control than the acute pain patients, but less chronicity and more treatment control than patients with chronic pain. The CFS respondents' perception of personal control was considerably greater than that demonstrated by either of the pain groups. In terms of causal attributions as reflected by the IPQ, both the present sample and Weinman et al.'s (1996) CFS group (N=l 15) rated a germ or virus as the most likely cause of the illness. This stands in contrast to the beliefs of patients with diabetes, R A , or chronic pain who were more likely to attribute their illness to chance or genetics. Moreover, this finding was confirmed by the IPQ-R data. Compared to chronic and acute pain patients, the CFS respondents were far more likely to endorse immune attributions (a subscale composed of the belief the illness was caused by a germ or virus, pollution, and/or altered immunity). However, unlike the CFS participants in the Weinman et al. (1996) study, the present sample rated stress as the second most likely cause of CFS and respondents were more likely than either of the pain groups to cite psychological attributions as etiological factors. As reflected by test-retest correlations the present sample's immune attributions were relatively fixed demonstrating little change after seven months while attributions of risk and psychological factors showed some variation. Respondents probably created complex, possibly interactional models of their illness, as hypothesized by Ray et al. (1992).  133 The tendency of CFS patients to make external attributions has been a consistent finding in the literature (Butler et al., 2001; Clements et al., 1997; Moss-Morris & Petrie, 2000). Although patients appreciate the relevance of stress and overwork, physiological factors seem to play a dominant role in their causal models. Typically, patients believe CFS is a manifestation of immune dysfunction caused by infection, stress, and/or pollution (Clements et al., 1997; MossMorris & Petrie, 2000). While this explanation is not well-supported in the scientific community, it probably fits patients' experience of the illness which is often characterized by a sudden onset with flu-like symptoms, fatigue, and new allergies (Surawy et al., 1995). It is also an explanation that is strongly endorsed by CFS support organizations and promulgated in the media (Clements . etal., 1997). Thoughtful commentators worry about the types of illness perceptions that may be communicated to CFS patients (Moss-Morris & Petrie, 2000). After analyzing the linguistic patterns of different illness groups on the Internet, Davison and Pennebaker (1997) remarked that "CFS sufferers have the most rigidly defined boundaries about illness prototypes.. .suspected psychological factors or psychosocial treatment strategies are viewed as anathema ... Such  x  intolerance reduces the information available to members, distilling and distorting, in turn, the collective illness schema" (p. 479). Given that an organic etiology cannot be dismissed, one must still be concerned that rigidly held causal attributions may prompt some CFS patients to refuse and/or respond poorly to treatment. In a study of fatigued patients (N=160) from a primary care setting who were randomized to either cognitive behavioral therapy (CBT) or counseling, follow-up assessments at three and six months revealed that patients who attributed fatigue to a physical cause and who had never consulted their general practitioner for an emotional reason were less likely to improve with either psychotherapeutic modality (Chalder, Godfrey, Ridsdale, King, & Wessely, 2003).  134 Summary In conclusion it can be seen that the respondents' illness perceptions were negative, exceeding the pessimism displayed by patients with other chronic conditions. Moreover, respondents' perceptions were very similar to those obtained from a 1996 New Zealand sample Weinman et al., 1996), suggesting a certain homogeneity in the perception of CFS symptoms across time and nationalities. In the present sample, as measured by the IPQ, illness identity and consequences were the two cognitive perceptions that inspired the greatest negativity. Respondents' emotional representations were also found to be highly pessimistic reflecting even greater distress than that experienced by pain patients. In addition, the study sample was most likely to attribute causation to immune dysfunction, a pattern which was distinctly different from the attributions of patients with other chronic illnesses. While a description of the respondents' illness representations has implications for CBT, it is the associations between representations, outcomes, and coping that are key to understanding the adjustment process. In the next section, the topic of illness representations is more fully explored with a detailed discussion of the relations between perceptions, functioning, psychological distress, and health care visits. Relations between Illness Representations and Outcomes Overall, bivariate and regression analyses demonstrated that illness representations are highly stable over time and predict outcomes even after controlling for demographics, illnessrelated variables, medication use, and disability status. Cognitive representations demonstrated a strong association with mental health, physical functioning, anxiety, and depression on the crosssectional analyses and predicted total health care visits in the longitudinal analysis. On the other hand, while being related to mental health and psychological distress, emotional representations did not demonstrate a statistically significant relationship with physical functioning or health care visits.  135 Certain perceptions were found to be more salient than others. A s hypothesized, pessimistic perceptions including the belief the illness; (1) has serious consequences, (2) prompts emotional representations of a distressing nature, and/or (3) has a strong identity (includes a large number of symptoms) predicted decreased functioning and psychological distress. Conversely, the optimistic perception of personal control was negatively related to depression and positively related to physical functioning. Curiously, analyses of time 2 and 3 data revealed that the effect sizes of personal control differed from those calculated at time 1 and 2. Since personal control and physical functioning were highly stable, the variations suggest mediation by another variable. Finally, in the longitudinal analysis an unexpected finding was that causal attributions of immune dysfunction predicted increased visits to health care providers. The results highlighted the important role cognitive representations play in psychological adjustment and functioning. In addition the findings suggested that perceptions which predict outcomes may vary depending on whether the patient is from a community- or tertiary-care setting. For example, the belief the illness has serious consequences was strongly predictive of depression in the present study and in another tertiary-care sample (Edwards etal., 2001). However, in Moss-Morris et al.'s (1996) community sample consequences were not related to psychological well-being. Further, the authors found no association between control and outcomes, which contrasted to findings from the tertiary-care sample. One explanation for the differences may be that individuals who believe the consequences of CFS are serious may be more likely to pursue medical care and gravitate to specialists. In addition, patients at the tertiary-care level may require a greater sense of control in order to maintain psychological well-being. It has been shown that cognitive behavioral therapy is most effective in secondary or tertiary care and that one of its most important benefits is increasing patients' feelings of control (Raine et al., 2002).  136 Only one study hypothesis was not confirmed, that psychological attributions were related to psychological distress, as reported by both Edwards et al. (2001) and Moss-Morris et al. (1996). However, this inconsistency may be due to differences in measurement. The preceding authors used the IPQ while the present study used the IPQ-R. When the IPQ-R was rescored to replicate the IPQ and the psychological attribution subscale reconfigured to meet the specifications of the emotional cause subscale, in a series of regression analyses emotional causes showed a significant association with anxiety (b = .22; p = .02) and a near-significant association with depression (b = .16; p = .057). Similarly, an attempt was made to replicate the analysis performed by Moss-Morris et al. (1996) by regressing the IPQ on the mental health subscale of the SF-36. (For psychological well-being Moss-Morris et al.{1996}used the Five Item Mental Health Scale derived from the SF-36 mental health subscale). Re-analysis confirmed Moss-Morris et al.'s (1996) findings demonstrating that attribution of illness to emotional causes was negatively and significantly related to psychological well-being (b = -.21, p =.03). It must be noted that the effect sizes for emotional causes were lower in the present sample than those reported by Edwards et al. (2001). However, the Edwards sample experienced greater psychological morbidity with 29% of their participants meeting the criteria for depression and 32% the criteria for anxiety (as opposed to about 16.5% of respondents in the present study who were anxiety or depression "cases"). Thus, psychological attributions may have a greater impact on psychological well-being in a distressed population. Since the result is based on crosssectional analysis only, the causal direction cannot be established. It may be that emotional distress increased the likelihood that the illness will be attributed to psychological factors, or it may be that attributing the cause of the illness to psychological factors increased emotional distress. Although attributing causes to physical factors is thought to be psychologically protective, longitudinal analysis in the present study demonstrated that attributions of immune  137 dysfunction were related to increased health care utilization. Despite reports that CFS patients are heavy consumers of health care, averaging between 21 and 23 visits per year (Bombardier & Buchwald, 1996; Buchwald & Garrity, 1994) and that patients often attribute the illness to a virus or immune dysfunction (Powell et al., 1990), only one other study has demonstrated a link between causal attributions and health care utilization. However, in their structural equation model of chronically fatigued individuals sampled from the community (N=780), Torres-Harding et al. (2002) found that participants with mixed or more psychological attributions were more likely to seek medical care than those attributing cause to physical factors, a finding inconsistent with Bishop's (1987) observation that subjects were less inclined to visit a doctor if they attributed symptoms to a psychological cause. In their discussion, Torres-Harding et al. (2002) opined that fatigued individuals who perceived their condition had both psychological and physical underpinnings may have been more distressed which prompted them to seek medical care. Based on the results from Torres-Harding et al. (2000) and the present investigation it is proposed that the type of causal attribution endorsed by an individual may not be as critical to health care utilization as previously thought. Rather, the strength of the causal attribution may be more relevant and function as a marker for disease conviction, as suggested by Lawrie et al. (1997). In both the Torres-Harding et al. (2000) and the present study causal attributions were assessed with a 5-point likert scale which may have inadvertently measured the strength of an attribution in addition to its type (physical or psychological). Perhaps strongly felt causal attributions mirror the individual's belief that something is "wrong," be it at a psychological or physical level, and that consultation with a health care practitioner is warranted. On the other hand, weakly endorsed causal attributions may reflect the individual's lack of certainty about the illness' causation and/or the tendency to "normalize" (as opposed to somatize or psychologize) symptoms which may curtail help seeking. Support for this contention may be found in a study  138 that found infrequent attenders of a general medical practice were more likely to make normalizing explanations for symptoms than frequent attenders (Sensky, MacLeod, and Rigby, 1996). Although not indicative of causality, longitudinal analyses suggested that emotional representations of distress are associated with mental health over time. In addition, causal attributions of immune dysfunction were found to predict total health care visits rather than vice versa. Thus, partial support has been provided for the self-regulatory model. Coping In addition to illness representations, coping has been related to adjustment. Three types of coping were evaluated; (1) problem-focused coping (accommodating to the illness, focusing on the symptoms, maintaining activity, and seeking information), (2) emotion-focused coping (approach, avoidance, and venting), and (3) catastrophizing. In the following section the respondents' coping profile is presented. The relative adaptiveness and maladaptiveness of coping strategies are discussed after having been ascertained by determining the association between coping strategies and outcomes. Coping Profile To the degree that the IMQ can be considered a measure of problem-focused coping (the focusing on the symptoms subscale does contain items that may relate to emotions generated by CFS), the sample appeared to use problem-focused strategies quite often. The disease-specific IMQ may have afforded greater predictive power in the present study. Unlike most other studies of CFS patients, gender differences were found with men more likely to focus on symptoms (F = 4.2, p < .05) and less likely to use approach strategies (F = 21, p < .0001). Problem-focused coping scores were similar to those obtained from another tertiary-care sample with an illness duration of 2.5 years (Ray et al., 1993). However the scores reported by Ray et al. (1993) and the present sample differed from those generated by patients recruited from  139 the community who had an illness duration of 2.5 years. The community sample (N=32) was more likely than the present sample to maintain activity (mean = 69 versus mean = 54) and less likely to accommodate to the illness (mean = 46 versus mean = 55). Moreover, the community sample tended to engage in less focusing and information seeking (Jason et al., 2000). The findings suggested that samples from community and clinical settings may differ with respect to coping. Perhaps variations in the ethnic and/or gender composition of the samples or differential exposure to the health care system accounted for the differences in coping profiles. Of the three emotion-focused coping strategies, the respondents favored approach with 96% of the sample utilizing this method at least sometimes. Avoidance was not a popular choice and two-thirds of respondents rarely used it. Intermediate between approach and avoidance was venting, a strategy used a "medium" amount by about 12% of the sample. Comparison of the COPE subscales in the present study with those reported in samples taken from tertiary care and the community revealed no substantial differences (Jason et al., 2000; Ray et al., 1983). About 22% (21/93) of respondents engaged in catastrophic thinking at least "sometimes." Using a qualitative method, Petrie et al. (1995) calculated the proportion of catastrophizers in their CFS sample to be 31%. Although data on catastrophizing in normal populations is not available, the mean score of 12.8 produced by the present sample was about mid-point compared to scores obtained from five samples of patients with chronic pain (range 8.4 to 15.6; Lawson, Reesor, Keefe, & Turner, 1990). Relations between Coping and Outcomes Coping is an individual difference that has been associated with adaptation. Rather than judging the adaptiveness of a coping strategy a priori Pearlin and Schooler (1978) suggested investigators evaluate the adaptiveness of a strategy by examining its relationship to an outcome.  140 Maladaptive Coping As hypothesized, in addition to catastrophizing, problem- and emotion-focused strategies including maintaining activity, focusing on the symptoms, avoidance, and venting were related to impaired mental health and/or psychological distress. Maintaining Activity Although at first glance maintaining activity would seem to be a laudatory approach, coupled as it has been with reduced impairment (Jason et al., 2003; Ray et al., 1993; Ray et al., 1995), it predicted anxiety in the present sample. Ray et al. (1993) concluded that maintaining activity is actually a form of denial. Rather than dealing with CFS, individuals apply their energy to minimizing the illness' impact. While this may protect everyday functioning it often leads to a destructive pattern in which the individual exhausts him/herself trying to maintain pre-morbid activity levels (Suraway et al., 1995). As a consequence anxiety rises and a sense of frustration and hopelessness ensues (Suraway et a l , 1995; Ray et al., 1993). The failure of maintaining activity to predict improved functioning in the present sample may be due to the respondents' long duration of illness and a preference for using the strategy of accommodating to the illness. Focusing on Symptoms Consistent with the work of other CFS researchers (Ray et al., 1995; Vercoulen et al., 1998) focusing on symptoms was found to be maladaptive, negatively related to mental health functioning and positively related to depression. Preoccupation with symptoms has been associated with greater morbidity in other clinical populations. Perhaps symptom preoccupation is similar to vigilance which has been related to a heightened experience of pain (Goubert, Crombez, & Van Damme, 2004). However, simply being aware of bodily sensations may not be as critical as the meaning an individual assigns to the sensation. Individuals who perceive bodily sensations as dangerous may get into the habit of selecting threatening somatic experiences over other competing demands thereby increasing the overall burden of illness (Goubert et al., 2004).  141 Avoidance Another consistent finding in the literature is that chronically ill patients who rely on avoidant strategies produce lower scores on measures of adjustment (Felton et al., 1984; Kemp, Morley, & Anderson, 1999; Maes et al., 1996; Sullivan et al., 1992). CFS patients are no exception and sample respondents who used avoidance (which included denial and mental and behavioral disengagement) experienced greater anxiety, depression, and lower levels of mental health. Some evidence has suggested that in the face of an unchangeable stressor or when confronted by acute stressors avoidance is advantageous, helping the individual maintain emotional balance and sparing him/her from futile attempts to alter an intractable situation (Maes et al., 1996). However, when adjustment is beyond the initial crisis period, the situation is remediable, and avoidance entails resignation and withdrawal the individual may experience greater distress and the amplification of future problems (Holahan et al., 1996). Avoidance is common in fatiguing illnesses. M S patients experiencing an unpredictable illness course utilized avoidant strategies to control the anxiety accompanying the destructive insecurity regarding future events (Warren, Warren, & Cockerill, 1991). However, escapeavoidance predicted future distress in another sample of M S patients (Aikens et al., 1997). Similarly, avoidance is an oft-reported strategy in CFS patients which has been associated with the disruption of social relationships, functional impairment, and persistent disability (Antoni et al., 1994; Heijmans, 1998; Ray et al., 1993; Surawy et al., 1995; Sharpe et al., 1992). It must be noted that avoidance is almost the antithesis of the premorbid personality profile of many CFS patients who have been described as hard-driving, perfectionistic, and action-prone (Lewis, 1996; Van Houdenhove et al., 2001). Thus, although it has not been established empirically, for CFS patients avoidance may be a particularly repugnant and demoralizing way of coping.  142 Venting Emotions Sample respondents who vented emotional distress also experienced greater anxiety and decreased mental health. In a healthy population venting has been associated with physical functioning (Jason et al., 2003), suggesting emotional expression may be benefical. Carver et al. (1989) found venting to be functional in some situations, for example, providing catharsis following an acute loss. However, in the chronically ill venting may not be adaptive. Jason et al. (2003) reported a positive correlation between venting and reduced functioning in a group of individuals with idiopathic chronic fatigue while Moss-Morris et al. (1996) demonstrated a negative relationship between venting emotions and psychological adjustment in CFS patients. . Carver et al. (1989) surmised that over time focusing on upsetting emotions is deleterious because it discourages an individual from using more constructive coping strategies and impedes movement away from distress. Catastrophizing Consistent with the study hypothesis, catastrophizing predicted impaired mental health, anxiety, and depression, paralleling findings from other studies (Keefe et al., 1989; Petrie et al., 1995; Rosenstiel & Keefe, 1983). In the chronic pain population catastrophizing has been identified as one of the most important factors in maladaptive coping, explaining more variance in disability than disease variables (Turk & Rudy, 1992). In CFS catastrophizing may motivate patients to place limits on functioning. Surawy et al. (1995) noted a tendency for patients to believe that aggravation of symptoms was harmful and increased the chances of a relapse. Thus, the tendency to catastrophize in response to symptom exacerbation discourages patients from increasing their tolerance of exertion. Coping Unrelated to Outcomes Seeking information. The only coping strategy not associated with outcomes was seeking information, which was hypothesized to predict health care utilization. CFS patients have shown  143 a marked interest in the etiology and management of the illness. In the present sample seeking information correlated only with the personality traits of conscientiousness (r = .31, p < .05, twotailed) and openness (r = .20, p < .05, two-tailed). Perhaps the respondents sought informal rather than formal sources of information. Clements et al. (1997) noted a predilection in their sample for information from journalists, patient assistance organization, and self-help books. The Internet has also become a popular venue. In general, few health care providers have an in-depth knowledge of CFS and therefore may not have been considered a rich source of information (Evengard & Klimas, 2002). As well, sample respondents' "need to know" may have become tempered by a long illness duration. Information seeking is likely more prevalent in the early stages of the illness. Adaptive Coping Supporting the study hypotheses and qualifying as adaptive strategies are accommodating to the illness and approach that predicted favorable outcomes. Accommodating to the Illness As a problem-focused strategy accommodating to the illness describes a process whereby patients organize their life to avoid stress and over-exertion. Sample respondents who used the strategy experienced greater mental health functioning and less depression. O f concern in the literature have been the limits on activity imposed by accommodating. Cross-sectional and longitudinal studies have demonstrated that "limiting" strategies correlated with greater dysfunction, perhaps confirming the old adage that "the less one does, the less one can do" (Ray et al., 1995; Ray et al., 1997; Sharpe et al., 1992). However, accommodating has also been consistently related to increased psychological well-being (Ray et al., 1993; Ray et al., 1997). Explaining this paradox, are theorists who suggested that accommodating may lessen symptom exacerbations and decrease the stress due to the illness' unpredictability (Moss-Morris & Petrie,  144 2000; Ray et al., 1997). Thus, accommodating may give patients a sense that they can control the illness (Ray etal., 1997). Approach The emotion-focused strategy of approach (consisting of acceptance and positive reappraisal) was negatively associated with depression and health care utilization. Carver et al. (1989) suggested that acceptance motivates an individual to actively deal with the situation and may be particularly important in circumstances when the stressor must be accommodated, as opposed to situations in which the stressor can be eliminated. Accepting the illness was the task deemed most important to accomplish by a group of Dutch CFS patients (de Ridder et al., 1998). Patients may find it very difficult to accept CFS since it entails giving up hope of regaining their pre-morbid status. However, acceptance may enable patients to extricate themselves from the cycle of boom or bust activity that has been found to be highly debilitating in newly diagnosed CFS patients (Surawy et al., 1995). Finally, patients who accept the illness may relinquish ineffective help-seeking behaviors (de Ridder et al., 1998) like "doctor-shopping" which probably reflects ongoing attempts to obtain an acceptable diagnosis. Positive reappraisal is a strategy in which the individual tries to view the stressful situation in positive terms. Carver et al. (1989) noted that the utility of positive reappraisal lies in its capacity to motivate an individual to persist in his/her coping efforts. Although not unique to CFS anecdotal reports and findings from qualitative studies have shown that patients can reappraise their situation and find something positive (Affleck & Tennen, 1996). For some sufferers, CFS provided an opportunity to "take charge - to slow down a breakneck pace, transform over-commitment into manageable levels of activity, reject externally imposed standards of performance, and finally, focus on themselves" (Ware & Kleinman, 1992, p. 11). Asbring's (2000) CFS participants stated they gained new insights including a deeper  145 understanding of life and themselves. Positive reinterpretation was related to psychological adjustment in Moss-Morris et al.'s (1996) sample. The Scope and Impact of Coping Longitudinal analyses that controlled for baseline outcome status as well as illness-related and demographic variables confirmed cross-sectional findings which demonstrated that catastrophizing and avoidance strategies predicted decreased mental health. Further, longitudinal data showed that approach coping predicted mental health functioning and was related to fewer total and conventional health care visits. However, the limitations of an observational design and correlational analysis preclude interpreting the data as proof of a causal association between the variables. In addition, the association between outcomes, maintaining activity, and venting appeared to vary at different time points. This suggests that the maladaptiveness of these strategies may be somewhat unstable, perhaps due to mediation by an unknown variable. Like the findings reported by Heijmans (1998) and Jason et al. (2003), in the present study coping was unrelated to physical functioning. Coping's lack of correlation with physical outcomes has been noted in a recent meta-analysis (Penley, Tomaka, & Wiebe, 2002). Some have argued that coping may be more highly correlated with proximal, rather than distal outcomes. On the other hand, it may be that there is little the chronically ill can do in terms of spontaneous efforts to mitigate the impact of the illness on physical functioning (Ray et al., 1993; de Ridder & Schreurs, 1996). A n interesting finding was that although the adaptive strategies were related to either improved mental health or decreased psychological distress, maladaptive strategies demonstrated greater effect sizes and were related to more domains. The observation that maladaptive coping is more predictive of outcomes than adaptive coping has been consistently reported in the literature and may be artefactual, reflecting the use of measures heavily laden with dysphoric content and a dearth of positively-valenced items (Watson & Hubbard, 1996). However, the  146 finding may be valid, demonstrating that the value of coping may not lie in its capacity to increase well-being but in its ability to prevent a worse situation (de Ridder & Schreurs, 1996). When designing coping interventions for CFS patients it may be more essential to curtail the use of maladaptive coping than foster the use of adaptive strategies (Petrie et al., 1995). Finally, compared to women, male respondents were more likely to utilize problemfocused strategies like focusing on symptoms and less likely to engage in emotion-focused coping like approach. This probably reflects men's tendency to use problem-focused coping and women's inclination to engage in emotion-focused coping (Hobfoll, Dunahoo, Ben-Porath, & Monnifer, 1994). Few investigators have found coping differences between genders in CFS. Blakely et al. (1991) reported differences in the types of emotion-focused strategies used with women preferring escape/avoidance and men relying on distancing strategies. Although problemfocused coping is generally considered superior to emotion-focused strategies, in the context of CFS this may not be the case. In the present study focusing on symptoms was related to unfavorable outcomes while approach strategies were related to favorable outcomes. Relations between Illness Representations and Coping Clearly coping must be considered a factor in how individuals adjust to CFS. However, coping behaviors do not develop in a vacuum. Coping is elicited by the appraisal of threat that is conceptualized in the self-regulatory model as a collection of perceptions representing an individual's representation of illness. As such, illness representations may, in part, determine the type of coping adopted by an individual. Illness representations may crystallize the degree of threat implicit in a situation. Perceived high threat in concert with feelings of high response- and self-efficacy are thought to lead to the adoption of adaptive strategies (Rippletoe & Rogers, 1987). On the other hand, a pessimistic appraisal of the situation coupled with an individual's belief that he/she is helpless to cope may trigger maladaptive strategies (de Ridder & Schreurs, 1996; Rippletoe & Rogers, 1987).  147 To a limited extent, this pattern may be seen in a study of coping using Lipowski's (1970) framework. Based on his clinical experience, Lipowski (1970) hypothesized that patients generated personal meanings of illness, which included perceiving illness as; a challenge, an enemy, a punishment, a weakness, a relief, a strategy, irreparable loss or damage, and as value. In a sample of the chronically ill (N=205) Schussler (1992) reported that patients who considered illness as a challenge or a value used more adaptive strategies like cognitive reconstructing and problem analysis. On the other hand, when illness was construed as an enemy or punishment patients adopted maladaptive coping strategies like wishful thinking. As hypothesized, for study respondents who endorsed pessimistic cognitive representations, particularly the belief that the consequences of illness were serious, there was a tendency to utilize maladaptive coping strategies including catastrophizing and focusing on the symptoms. Further, distressing emotional representations predicted the maladaptive strategies of catastrophizing, avoidance, and venting. While no.other study has linked catastrophizing with illness representations, investigators have reported that pessimistic representations includinga chronic timeline, low perceived control, more numerous symptoms (identity), and serious consequences were related to avoidance and disengagement (Moss-Morris et al., 1996; Heijmans, 1998). Findings in the literature were mixed in regard to venting with some investigators linking pessimistic representations to venting (Moss-Morris et a l , 1996) and others reporting a negative association between pessimistic beliefs and venting (Heijmans, 1998). Since Moss-Morris et al. (1996) sampled New Zealanders and Heijmans (1998) recruited Dutch patients it may be that inconsistencies reflect cultural differences in the expression of emotions. Not confirmed by the data was the hypothesis that personal control predicts adaptive coping. However, the perception the illness is less distressing predicted the adaptive.strategy of approach. This finding was consonant with studies demonstrating that optimistic illness  148 representations like perceived control correlated positively with reappraisal and negatively with disengagement (Moss-Morris et al., 1996). Presumably, less pessimistic illness representations enable an individual to deal with the situation in a constructive manner. Although the degree of pessimism associated with an illness representation has been consistently linked with whether an individual adopts an adaptive or maladaptive coping strategy, there is considerable variation among studies as to which perception is predictive. Likely this is due to differences in how illness representations and coping are measured. In addition, the effects of illness representations on coping may vary depending on the setting in which the study is conducted. The role of single perceptions in predicting coping may be less important than the overall degree of pessimism attached to an individual's illness representation. In the present sample pessimistic perceptions of chronic timeline, serious consequences, and emotional representations of distress were positively correlated with each other and negatively correlated with perceptions of control. This pattern of correlations is congruent with findings from other studies that have examined the structure of illness representations in the chronically ill. Heijmans and de Ridder (1998b) opined that strong intercorrelations among the various illness representation components indicate that the construct may be better understood as a group of beliefs or schemata (Heijmans & d e Ridder, 1998). However, based on Heijman's (1998) analytic model, it may be necessary to treat causal attributions as a separate schemata. Causal attributions did not correlate well with either the pessimistic or optimistic perceptions (except for attribution of cause to psychological factors, which was weakly related to personal and treatment control). In addition, while pessimistic perceptions were negatively related to favorable outcomes and optimistic perceptions positively related to favorable outcomes, the same causal attribution exhibited differential associations, positively and negatively correlated with favorable outcomes.  149 An unexpected finding was that cognitive representations predicted only problem-focused coping while emotional representations primarily predicted emotion-focused coping. This is of interest from a theoretical perspective because it lends support to the self-regulatory model that posits parallel processing by each representation in response to threat (Leventhal et al., 1997). Cognitive representations deal with the objective aspects of illness and emotional representations deal with emotional reactions engendered by illness. Thus, it is fitting that each representation should relate to a corresponding focus of coping. The finding also highlights the importance of addressing the objective and emotional demands of the illness experience when describing adjustment. Social Support As well as eliciting coping, it is thought that stress prompts individuals to marshal social support. Three types of support were examined in the study; (1) structure (SSQ-N), the number of supportive others in the network; (2) perceived support (SSQ-S), respondents' satisfaction with support; and (3) received support (ISSB), the frequency with which supportive behaviors were enacted. Profile of Social Support Perceived Support The respondents' perceptions of support closely approximated those reported by Jason et al. (2003) in a study of CFS patients recruited from the community. Moreover, compared to other groups tested with the SSQ by Jason et al. (2003), the number of supportive others reported by respondents in the present study (mean = 3.9) was higher than that of healthy controls (mean = 3.05) and for patients with medically-explained illnesses (mean = 2). In addition, with a mean of 5.01 the respondents experienced greater support satisfaction than patients with medicallyexplained illnesses (mean = 4.6) and slightly less satisfaction than healthy controls (mean = 5.5). Further the respondents' perceived support rivaled that of university undergraduates who  150 identified 4.2 supportive others and produced a mean satisfaction score of 5.3 (Sarason et al., 1983). Unlike Lewis et al.'s (1994) findings which showed that CFS patients perceived less support than a comparison group of IBS sufferers, respondents in the present sample enjoyed a fairly substantial degree of support satisfaction, a circumstance which has been observed in other chronically ill populations (de Ridder & Schreurs, 1996). Received Support The respondents' mean of 2.03 on the ISSB indicated that supportive behaviors were enacted rather infrequently. Although similar to the results obtained by investigators who enrolled CFS sufferers from a patient assistance organization (Kelly, Soderlund, Albert, & McGarrahan, 1999), the respondents' received support (mean = 83.4) was lower than that experienced by college undergraduates (mean = 100.8) and pregnant teenagers (mean = 103.8 {Barrera, 1981; Stokes & Wilson, 1984}). However when juxtaposed against the relatively high support satisfaction scores, low levels of received support need not be alarming. Despite the limited frequency of enacted behaviors, respondents were relatively satisfied with their support. Further, high instrumental support levels are not always associated with favorable outcomes. Some chronically ill patients do not want a great deal of help. Rheumatoid arthritis patients who received more help than they desired experienced greater depression (Manne & Zautra, 1989). Differences in received support levels may be due in part to the larger social networks possessed by adolescents and young adults. Received support may be dependent on the quality and frequency of social interactions (de Ridder & Schreurs, 1996). CFS often disrupts patients' social networks. While the number of supportive others would suggest the respondents were able to maintain a viable support system, the constraints imposed by the illness on patients' lifestyles may have impeded their ability to interact socially (Antoni et al., 1994). Furthermore, some patients may not care to solicit social support. Receiving support from others requires that the individual recognize his/her status as an impaired person. CFS patients  151 may not accept their condition and/or attempt to "pass" in order to conceal their stigmatized identity from others. Some may exercise a degree of social detachment perceiving the costs of self-disclosure to be greater than the benefits that can be derived from increased assistance. As well, patients may be reluctant to accept help due to a perceived inability to reciprocate the aid received (Affleck et al., 1988). Chronic illness is especially likely to engender a sense of indebtedness. Unfortunately, if the recipient elects to resolve this dilemma by rejecting assistance relationships may be jeopardized. No gender differences were found in SSQ scores or in the composite ISSB scores but when the ISSB was split into different dimensions, male respondents were shown to receive emotional support on a more frequent basis than women. This differed from a finding reported in a study of university undergraduates in which women received more emotionally supportive behaviors than men (Forbes & Roger, 1999). The chronically ill have a strong need for emotional support that may out-weigh the gender differences shown in non-clinical samples. In a study of M S patients, men perceived higher levels of affect, affirmation, and aid support than women (Gulick, 1994). Relations between Social Support and Outcomes Perceived Support, Functioning, and Psychological Distress Both received and perceived support predicted mental health functioning and decreased depression. The positive effect of perceived support on respondents' mental state confirmed the study hypothesis and resonated with other studies of the chronically ill (Brummett, Barefoot, Siegler, & Steffens, 2000; Brenner, Melamed, & Panush, 1994; Kerns & Turk, 1984; Scharloo et al., 1998). The perception that support is available enables individuals to confront stressors more effectively since it is believed that others may help if they become overwhelmed (Pierce et al., 1996). In addition, supportive relationships are thought to buffer the effects of illness-related stress (Pierce et al, 1996) which may be particularly important for CFS patients due to the  152 stigmatization, demoralization, and disability associated with the illness. Studies in CFS patients have linked seeking support with psychological adjustment (Heijmans, 1998; Moss-Morris et a l , 1996) The data revealed that perceived and received support make different, potentially complementary contributions to functioning and well-being. Perceived support was more strongly related to mental health than received support, a finding commonly reported in the literature (Affleck, Pfeiffer, Tennen, & Fifield, 1988). This may reflect method variance due to conceptual overlap between measures of perceived support and psychological adjustment (Felton & Shinn, 1992). Received Support and Physical Functioning Only received support was related to physical functioning. Further data analysis revealed that the component primarily influenced by received support was physical role functioning which measured health-related limitations in the kind or amount of work performed. Instrumental assistance may allow patients to conserve energy allowing them to focus on work-related projects. In addition, support may function to increase an individual's feelings of control and confidence (Schreurs & de Ridder, 1997). Perhaps greater confidence leads to the perception that CFS is less of a barrier to activity. Expectancy theory predicts that an individual is more likely to attempt a task he/she believes can be accomplished (Bandura, 1977). Partial support for this interpretation may be gleaned from other studies and post-hoc data analysis. In a study of patients with rheumatoid arthritis, the unavailability of instrumental support was associated with a decreased perception of competence in patients (Smith, Dobbins, & Wallston, 1991). Moreover, higher self-efficacy for the management of CFS symptoms was found to be predictive of improved work and social functioning (Findley et al., 1998). MossMorris et al. (1996) reported that perceptions of control predicted greater vitality in CFS patients.  153 Finally, in the present study, received support was significantly correlated with personal control (r = .23, p < .05, two-sided) and personal control was strongly predictive of physical functioning. Although it was heartening to identify a self-management behavior that was positively related to physical functioning the study hypothesis predicted that received support would be associated with decreased functioning. It was thought that instrumental support might decrease patients' activity levels. The study's failure to demonstrate this phenomenon may be due to an erroneous characterization of the process through which support wields its deleterious effects. Studies that have linked support and unfavorable outcomes focused on how spouses' solicitous responses reinforced illness behaviors which predicted subsequent disability (Paulsen & Altmaier, 1995; Goodwin, 197; Schmaling et al., 2000). Operant conditioning has been proffered as the explanatory model positing that specific illness behaviors act as operants in the environment increasing the frequency which an individual obtains a desired contingency or is able to avoid undesirable consequences. Supportive others may inadvertently reinforce unhelpful illness behaviors by responding solicitously to patients' bids for increased attention. Perhaps, received support is associated with outcomes but only in the context of close relationships as a function of solicitous expressions. (In partial support of this hypothesis was the finding that the tangible assistance dimension of the ISSB did not predict physical functioning.) The study may have overestimated the adaptiveness of received support. The ISSB only utilized items describing positive behaviors. Practical help, while useful, may be given in a begrudging way or in a manner that demeans the recipient. Advice that is unwanted or rejected may create conflict and additional stress. Moreover, the ISSB is a generic measure that does not include items specific to the experience of CFS, which may have caused respondents' to endorse fewer items. In order to truly assess the association between received support and outcomes, a more finely-nuanced instrument may be needed that contains negative as well as positive support  154 behaviors and which includes the recipient's appraisal of the behavior (Heller, Swindle, & Dusenbury, 1986; Ray, 1992). Perceived Support and Health Care Visits A somewhat complex pattern of correlations emerged for perceived support. The data showed that the number of supportive others was positively related to conventional health care visits cross-sectionally and negatively associated with alternative care visits over time. Support satisfaction was also found to predict increased use of alternative care visits over time. While at first glance it may appear that perceived support produced contradictory effects, the results may be reconciled if perceived support is treated as a multidimensional construct. Leventhal et al. (1996) addressed the impact of social contexts on health care utilization. The authors claimed the "use of professional services for diagnosis and treatment is often a direct product of social influence; people ask and/or are told by family members and friends to "go to the doctor and find out what (is wrong)..."(p. 39). Illness representations held by an individual's cohort can generate social pressures, supporting the adoption of specific procedures for the management of a health problem (Leventhal et al., 1997). Such a process may have been demonstrated in the present study. The data may reveal that respondents' supportive others exerted a conservative influence on help seeking, favoring conventional over alternative strategies. The relationship between support satisfaction and alternative care may reflect a different process. Patients who experience greater support satisfaction may feel emboldened and more open to (and less suspicious of) unconventional treatment modalities. Further, qualitative work has suggested that in addition to providing treatment for CFS, alternative care providers serve as a source of support. Its practitioners are considered better listeners and more responsive to patients' concerns (Cooper, 1997). Thus, patients' health care utilization may be related to;  155 (1) external factors including supportive others who suggest patients visit a particular type of provider and (2) internal factors like the patient's direct and positive experiences with a provider that motivates him/her to maintain contact. However, it must be kept in mind that limitations imposed by the study design preclude an interpretation of causality and require that the preceding explanations be subject to further investigation. Relations between Illness Representations and Social Support Social support is thought to be elicited by the presence of stress, which may be conceptualized in the self-regulatory model as cognitive and emotional representations of illness. As hypthesized, representations predicted support, albeit in a differential manner. Causal attributions of risk predicted less support satisfaction, while attributions of immune dysfunction predicted more received support. Although not observed by Kelly et al. (1999), causal attributions of a biological nature predicted support seeking in Heijmans' (1998) sample. CFS patients with physical attributions may report greater support for two reasons. One, network members may be more willing to provide support for individuals suffering from an organic disease (Lewis, 1996). Two, individuals may be more likely to seek social support if they attribute their need to a physical infirmity believing such an exigency makes appeals more valid (Pearlin, 1985). Both reasons reflect the wide-spread belief that physical illnesses are real and deserving of attention, while illnesses without a physiological basis are self-inflicted and should be ignored (Sharpe, 1993). The positive relationship between physical attributions and increased support also suggests a mechanism through which physical attributions enhance psychological well-being. Although pessimistic representations were hypothesized to predict support, the results revealed that emotional representations of distress predicted less support satisfaction. This finding is congruent with Heijmans (1998) who reported that pessimistic perceptions of identity and chronic timeline predicted less support seeking but at odds with the work of Moss-Morris et  156 al. (1996) who found that pessimistic perceptions like serious consequences and a chronic timeline predicted support seeking. However, neither Moss-Morris et al. (1996) nor Heijmans (1998) had the advantage of using the IPQ-R, which, by including a scale measuring emotional representations, permitted a more fine-tuned analysis. In addition neither group differentiated between the types of social support actually utilized by the sample. Likely the association between representations and support is multi-faceted, functioning at an intra-psychic and interpersonal level. At the intra-psychic level pessimistic representations may reflect the appraisal of greater threat that may be associated with the mobilization of support. On the other hand, at the interpersonal level, pessimistic perceptions as characterized by emotional distress may be related to a perceived reduction in satisfaction with the support being provided by others. Highly disturbed emotional reactions may drive supportive others away and/or create greater expectations of support in the individual which lead him/her to be less satisfied with the support provided (Affleck et al., 1998). Relations between Social Support and Coping To further understand how patients manage CFS, it may be instructive to describe the interplay among social support and coping strategies. To accomplish this goal the correlations between the two self-management modes (Appendix 5) are discussed. Moderate and positive correlations were found between catastrophizing with other maladaptive strategies including focusing, avoidance, and venting. Catastrophizing is likely a particularly toxic way of coping demonstrating strong and negative relations with favorable outcomes, in addition to possibly facilitating and/or sharing qualities associated with other maladaptive strategies. The ruminative nature of catastrophizing in addition to its association with feelings of helplessness probably overlaps substantially with the concept of focusing (as conceptualized by Ray et al., 1993). In chronic pain patients catastrophizing has also been associated with focusing on symptoms (Geisser et a l , 1994). Catastrophizing may reflect the individual's inclination to give up rather  157 than persist in activities that require exertion (Petrie et al., 1995). Like avoidance, this may reflect a general and maladaptive orientation to stress, which falls under the rubric of disengagement related to feelings of low self-efficacy. The relationship between catastrophizing and venting probably involves an interpersonal process. In a recently developed coping model theorists proposed that catastrophizing is characteristic of individuals with a communal and expressive orientation towards managing stress (Thorn, Ward, Sullivan, & Boothby, 2003). Through venting, the catastrophizer may display distress and communicate an inability to deal with the situation thereby maximizing the probability that companions or potential caregivers will maintain proximity and provide assistance. Unfortunately, such a strategy may be counter-productive and drive supportive others away. Both catastrophizing and venting were negatively related to social support in the present study. Some of the maladaptive strategies were positively correlated with the adaptive strategies. The association between the problem-focused strategies, focusing and accommodating, may have occurred because each subscale reflects an awareness of internal states. However, a sense of helplessness pervades the focusing subscale while the accommodating subscale captures the respondent's proactive intention to limit stress and exertion. Compared to accommodating, focusing may represent a form of disengagement, borne of a sense of futility or helplessness (Moss-Morris et a l , 1996). Also of interest was the positive correlation between the two emotion-focused coping scales, approach and avoidance. Although one might expect the scales to be polar opposites, the positive correlation may reflect respondents' uncertainty about how to manage CFS and their tendency to vacillate between opposing tendencies (Lewis, 1996). Social support demonstrated a positive relationship with adaptive coping as shown by the correlations between accommodating to illness, number of supportive others, and received support. While the correlations were modest, they provided some support for the notion that  158 support functions by facilitating coping. Instrumental assistance may bolster patients' attempts to modify their life in order to limit stress and overexertion. Patients probably need help to accomplish routine tasks so energy may be expended on other activities. Moreover, a larger circle of supportive others may be advantageous since it spreads the burden and decreases the risk of burnout. In general it appears that disengagement related to a sense of helplessness is an overarching theme that may lie at the heart of an individual's unsuccessful attempts to adjust. Giving up and withdrawing from the situation may occur when an individual has strongly negative expectations of coping outcomes (Heijmans, 1998). Disengagement may actually represent the decision to refrain from coping. Summary Although establishing causal relations is beyond the scope of the present work it may be instructive to provide the reader with a concise description of the major findings with the understanding that additional investigation is necessary to validate the conclusions drawn. With this caveat, it may be seen that CFS patients adopt one of two different pathways to manage the illness (Moss-Morris & Petrie, 2000). Patients who appraise the situation as being severe and who tend to catastrophize are likely to deal with the objective demands of CFS by focusing on and becoming preoccupied with the illness. If the illness causes excessive distress and triggers the response of catastrophizing, patients may become overwhelmed and deal with the emotional repercussions by venting, withdrawing from meaningful activities, and alienating supportive others. Patients following this pathway will likely suffer extensive losses and experience psychological distress. On the other hand, although just as disabled, patients with less pessimistic beliefs who are not inclined to catastrophize will deal with the objective demands of CFS with the help of supportive others, by balancing rest and activity, and by avoiding stress. In the face of minimal  159 emotional distress and by not catastrophizing they will cope emotionally by accepting limitations and finding something positive in the situation. Eventually they may be able to weave the illness into the pattern of their life. For these patients greater psychological well-being may be expected. Efficacy of Coping and Social Support As has been found by other investigators, coping and social support produced modest correlations with outcomes, considerably weaker than the correlations produced by illness representations. Aldwin and Revenson (1987) opined that the efficacy of self-management efforts must be determined in order to appraise their association with outcomes. Thus, to better understand the role of self-management in the adjustment process respondents were requested to rate the efficacy of problem- and emotion-focused coping in addition to received and perceived support. As hypothesized accommodating to the illness, perceived support, and approach strategies were considered efficacious by the respondents. However, analyses revealed that these strategies were only of marginal benefit in controlling the symptoms or recovering from CFS. Moreover, only 5 of 18 self-management efforts were rated as "somewhat effective." This may be related to the narrow focus of the efficacy scales, which only captured the efficacy of selfmanagement behaviors for dealing with CFS. It must be assumed that coping and social support serve other non-illness related functions (Findley et a l , 1998). In addition the efficacy scales only measured respondents' expectations of the outcome of self-management efforts. No attempt was made to capture self-efficacy, an individual's perception of whether or not he/she was capable of performing a behavior. This may have introduced measurement error. The low efficacy ratings of some strategies may have been due to the inclusion of "not applicable" as a response that was likely used by respondents to code strategies they were unable to perform. Further, self-efficacy may be a better predictor of adjustment than outcome expectations (Hampson, 1996).  160 For the most part respondents considered strategies primarily efficacious for controlling symptoms. Clements et al. (1997) reported a similar finding with participants in the study stating they "could control their symptoms to some degree, but could not alter the course of the underlying disease process" (Clements et al., 1997, p. 622). In the present study, only accommodating to the illness was perceived to assist in recovery, specifically the behavior; "knowing what's best for me." Such a sentiment reflects a high degree of autonomy in the respondents which may develop when patients realize there is little medicalscience can offer them. With three out of five strategies included in the list referring to problem-focused efforts, this coping mode was considered most efficacious by respondents. Although consistent with a large body of literature, demonstrating the utility of problem-focused approaches in controllable situations, it has been thought that the chronically ill experienced a limited degree of control which forced them to rely on more emotion-focused strategies. Perhaps CFS patients enjoy greater perceived control than has been previously recognized. The problem-focused strategy of accommodating to the illness, specifically balancing rest and activity, was considered the most efficacious strategy and is likely a defining feature of CFS (Moss-Morris & Petrie, 2000). Moreover, astute clinicians usually advise CFS patients to adopt a "disciplined program of rest and carefully paced activity" (Stiver and Vortel, 1991, p. 549). More recently Jason (1999) incorporated Envelope Theory into cognitive behavioral therapy. Envelope Theory posits that CFS patients who "overexert themselves and push the limits of their energy resources might need to reduce their activity levels; whereas those who are not active enough might need to increase their activity level..." (p. 433). Jason (1999) stressed the need to customize the treatment plan to fit the individual with the goal of helping "patients maintain their energy expenditure within the range of their daily available or perceived energy" (p. 433).  161 Seeking information was also considered somewhat efficacious although it did not predict outcomes in this study. Ray et al. (1993) found seeking information predicted greater fatigue, opining that it might give rise to an exaggerated concern with symptoms. However, respondents . in the present sample differentiated between types of information seeking rating a general search for knowledge and new information as efficacious. Perhaps this reflected their need to confirm the validity of the illness (Moss-Morris & Petrie, 2000). On the other hand the respondents rated "trying anything I hear of to get better" or "following the advice of others" as non-efficacious. This suggested that a cautious and reasoned approach to information gathering may be more helpful than engaging in a desperate attempt to find a cure or a tendency towards gullibility which could make an individual vulnerable to quacks or charlatans. Emotion-focused coping and perceived support were also considered efficacious. Accepting the illness, by "learning to live with it" and engaging in positive reappraisal by "learning from the experience" were cited as specific efficacious behaviors. In terms of perceived support the specific behaviors most valued by the respondents were being listened to and accepted unconditionally. The study hypothesis proposed that coping and support efforts considered efficacious would be more predictive of outcomes than efforts not considered efficacious. The hypothesis was supported only for the association between problem- and emotion-focused coping and physical functioning. A n interesting explanation for this finding may be found in the work of Bandura, Reese and Adams (1982). Using an experimental design the investigators observed that phobic subjects high in self-efficacy experienced no changes in blood pressure or heart rate when asked to perform a task that exposed them to the focus of their phobic fears (Bandura, Reese, & Adams, 1982). However, subjects who were moderately insecure about their coping efficacy experienced elevated heart rates and blood pressure readings.  162 In CFS patients, coping efficacy may moderate physiological responses to stress. CFS patients often fear activity will increase their symptoms (Silver et al., 2002) and over time sufferers may become more sensitive to ever-lower-intensity symptom flare-ups (Surawy et al., 1995). It is possible that low perceived efficacy might be associated with autonomic arousal that might be interpreted as an exacerbation of symptoms leading the individual to avoid physical activity. Patients with greater expectations of efficacy may not experience autonomic arousal and thus be more inclined to engage in physical activities. In conclusion it must be noted that the effectiveness of self-management efforts (their association with positive outcomes) was congruent with respondents' perceptions of the strategies' efficacy. Accommodating to the illness, approach, and perceived support all predicted mental health and these same strategies were considered efficacious by the respondents. Similarly focusing on symptoms and avoidance were related to psychological distress and impaired mental health. Moreover, both coping strategies received very low efficacy ratings. Efficacious and effective strategies, like accommodating and approach, were performed with far greater frequency than ineffective and non-efficacious strategies like catastrophizing, focusing, and avoidance. Such a pattern of findings suggests that the illness behaviors of the respondents, as reflected in self-management efforts, were constructive and geared towards maximizing adaptation. Determinants of Predictive Perceptions To better contextualize the findings, an attempt was made to identify the determinants of illness perceptions that predicted outcomes and coping. After controlling for demographic, illness-related, and control variables predictive perceptions were regressed on personality traits and psychiatric status. Two perceptions were found to be related to the personological variables. However, in sequential regression models only personality traits were statistically significant  163 predictors. Neuroticism predicted serious consequences and emotional representations of distress while agreeableness predicted personal control. Few investigators have assessed the determinants of illness perceptions. Heijmans and de Ridder (1998) compared CFS and Addison's Disease (AD) patients. In the CFS group a strong illness identity (more symptoms) was predicted by pessimism, a younger age, and less education. Chronic timeline was also predicted by pessimism while control/cure was related to optimism. The investigators were unable to isolate a variable that predicted consequences. For A D patients less optimism and more pessimism were associated with a stronger illness identity. Greater pessimism implied a chronic timeline and less optimism predicted more serious consequences. Neuroticism Congruent with the findings of Heijmans and de Ridder (1998) the present study found an association between the negative personality trait neuroticism and the pessimistic illness perceptions, emotional representations of distress and serious consequences. Likely this reflects the interaction between disposition and the situation in the perception of stress (Costa et al., 1996). The degree to which individuals perceive events as stressful has been related to neuroticism (Costa et al., 1996). Compared to university students with low neuroticism, students with high neuroticism reported more daily conflicts and a tendency to react to them with depression and anger (Bolger & Zuckerman, 1995). Costa et al. (1996) defined neuroticism as "the propensity to experience negative affects" (p. 51) and a hallmark of neuroticism is the tendency of those with the trait to report emotional distress. In a non-clinical sample individuals high in trait neuroticism exhibited greater distress when confronted by home or work overload, or i f faced with interpersonal stress than individuals with low trait neuroticism (O'Brien & DeLongis, 1996). Individuals high in neuroticism may also be prone to negativistic appraisals of the environment. "They tend to interpret ambiguous stimuli in a negative or threatening manner and  164 therefore, are likely to see threats, problems, and crises where others do not." (Watson & Hubbard, 1996, p. 748). Thus, neuroticism may act as a filter distorting an individual's perception of the illness experience leading him/her to perceive more serious consequences. Researchers studying a construct similar to neuroticism, negative affectivity (NA), suggested that individuals high in N A may have an over-active Behavioral Inhibition System (BIS) that causes them to classify all incoming stimuli as important and requiring careful inspection. Combined with the tendency to be negativistic, hypervigilence in high trait N A individuals may be associated with the interpretation of health threats in two ways; (1) individuals may be more likely to notice untoward experiences and (2) because their scanning is fraught with anxiety and uncertainty the individual may be more likely to interpret such experiences as harmful and distressing (Watson & Pennebaker, 1991). Studies vary regarding the degree of neuroticism present in the CFS population. The present study found respondents scored within the normal range but CFS patients have been found to score higher than healthy controls on the "neurotic triad" of the Minnesota Multiphasic Personality Inventory (MMPI; Blakely et al., 1991; Millon et al., 1989). In recent years, the M M P I has been somewhat discredited due to its tendency to overestimate neuroticism in the chronically ill (Johnson et al., 1996c). However, in studies using psychometrically validated personality measures CFS patients were still found to exhibit higher levels of trait neuroticism than healthy normals but lower levels than those experienced by patients with major depressive disorder (Johnson et al., 1996c; Buckley, MacHale, Cavanagh, Sharpe, Deary, & Lawrie, 1998). Investigators have found that CFS patients were equivalent to M S and R A patients in terms of trait neuroticism (Johnson et a l , 1996c; Wood & Wessely, 1999) suggesting that neuroticism may be related to the experience of chronic illness. Such a proposition was supported by Buckley et al. (1999) who reported that a comparison of pre- and post-illness neuroticism scores in CFS patients indicated that neuroticism increased following illness onset.  165 Some caveats were attached to this conclusion since retrospective reports of pre-illness neuroticism may be susceptible to subjects' tendency to emphasize their premorbid psychological health. Johnson et al. (1996c) opined that the high degree of stability associated with trait neuroticism probably makes changes due to life stresses unlikely. Agreeableness In the present study, the predictive illness perception of personal control was related to the personality trait, agreeableness. Costa et al. (1996) characterized agreeableness as "a stoic and compliant attitude" (p. 53). Individuals high in agreeableness are altruistic, sympathetic to others, and eager to help. Low agreeableness is tantamount to antagonism and describes an individual who is cynical, hostile, and intolerant of others (Costa & McCrae, 1991). Agreeableness may be predictive of personal control because the quality of effortful control is likely the developmental substrate of agreeableness (Jensen-Campbell, Rosselli, Workman, Santisi, Rios, & Bojan, 2002). The work of a group of Finnish investigators strongly suggested that agreeableness arises from an individual's ability to practice self-control in childhood (Laursen, Pulkkinen, & Adams, 2002). In a 25-year prospective study (N=194) that included teacher, peer, and self-reports of adjustment and personality attributes, Laursen et al. (2002) reported that "childhood compliance, aggression, and self-control discriminated high-agreeable from low-agreeable adults... (p. 596). Low-agreeableness may be related to an inability to regulate behaviors and emotions. Conversely, "compliance and restraint may foster positive social interactions, which promote an accommodating interpersonal style and a benevolent disposition" (p. 599). High agreeableness may anchor a resilient personality type whose cooperative and pleasant nature facilitates and enhances feelings of personal control.  166 Adjustment to CFS over Time Stages of CFS: Differences between Illness Duration Groups Now that a picture has emerged about how respondents' illness representations, coping, social support, and perceptions of efficacy relate to the adjustment process, the reader's attention is turned to another important issue. Do patients at different points on the illness trajectory adjust in the same way? To address this question respondents with short (ill less than four years) and long illness durations (ill four years or more) were compared to see i f any self-management differences existed and, if so, did the differences relate to functioning, psychological distress, and health care utilization. Differences in Coping and Social Support The hypothesis that respondents with a short duration of illness coped by maintaining activity while respondents with a long duration of illness coped by using accommodating and approach strategies was not confirmed. Moreover, contrary to the study hypothesis that respondents with a short duration of illness would experience less support, respondents in short (mean = 2.6 years; SD = 1) and long duration (mean = 10.1 years; SD = 4.6) groups were comparable in regards to coping and social support. Similar non-significant findings were reported by Friedberg et al. (2000) in a study comparing CFS patients (N=258) stratified as to illness duration (ill under seven years versus ill for over ten years). However, based on a median split of the short duration group (ill less than three years versus ill three or more years) Friedberg et al. (2000) reported a trend towards higher scores for maintaining illness in the under-threeyear duration group (p < .10). The authors concluded that the lack of coping differences in patients ill over three years reflected the propensity for coping to stabilize within 12-18 months of diagnosis as has been reported for heart disease and cancer.  167 Differences in Perceived Efficacy of Coping and Social Support When coping was dichotomized as efficacious or non-efficacious the data suggested that respondents with shorter illness durations found more coping and support strategies efficacious than respondents with a long illness duration. In addition, although both groups reported that accommodating to the illness, approach, and support satisfaction were helpful for controlling symptoms, only the short duration group considered the accommodation and approach strategies useful for recovering from CFS. Moreover the subscales that were statistically higher in the short duration group including venting, maintaining activity, seeking information, and received social support all referred to the strategies' efficacy in recovering from CFS. In the long illness duration coping strategies were considered efficacious only for symptom control. The data suggested that the short duration group is somewhat more optimistic regarding their chances of recovering from CFS. Perhaps, patients initially deny the chronicity of the illness and try out multiple strategies hoping to "beat" CFS. However, patients with a longer duration may recognize the intractability of the illness process. They have learned that selfmanagement is primarily of use for controlling symptoms, which may assist in rendering the illness more predictable. Differences in Outcomes Differential outcomes between the two illness duration groups were hypothesized to occur as a function of coping that varied depending on the respondent's illness phase. As previously discussed, similar coping strategies were used by the two illness groups and, as might be expected no differences were observed between the illness groups for functioning and psychological distress. Friedberg et al. (2000) also reported no differences in fatigue, functioning, or depression between short and long illness duration CFS patients. In the present study health care utilization did vary according to illness group, with more numerous visits reported by the short duration group who averaged 23 visits to health care  168 practitioners in six months, a figure considerably higher than 22 visits/year reported by Buchwald and Garrity (1994) for a sample of tertiary-care CFS patients (N=30) who had been ill for about four years. However, the authors' figures are consistent with the present study's long duration illness group which averaged about 12 visits in six months. Such differences in health care utilization probably reflect changes in help-seeking behavior that may have run the gamut from a dedicated search for diagnosis and treatment requiring numerous visits to GPs and specialists to a sense of resignation that health care providers can do little to change the course of CFS and the decision to confine visits to select practitioners. In conclusion, the findings provided little support for Fennell's (1995) four-phase model of CFS. Only perceptions of efficacy and health care visits differentiated short- and long-duration illness groups, differences which are relatively non-specific and likely integral to the experience of other chronically ill patients. In a sample of M S and chronic pain patients Sullivan et al. (1992) reported that individuals with a short illness duration reported a higher frequency of visits to health care professionals. Changes over Time: Within-Individual Differences In order to properly characterize the adjustment process, an attempt was made to measure changes in individuals over time. Coping and Social Support Respondents' (N=61) self-management scores at four points in time (a duration of 21 months) were subject to repeated measures analyses. As hypothesized seeking information decreased significantly. Also consistent with the study hypotheses were findings that maintaining activity and received support were less frequently utilized (although the decline in use did not reach statistical significance which may have been due to type II error caused by a small sample size).  169 In general the trend was for all self-management efforts to decrease slightly with significant differences observed for problem- and emotion-focused strategies. However, the decline was probably clinically insignificant representing a drop of less than one-half of a standard deviation (SD). As such, coping and social support demonstrated remarkable stability, a finding consistent with the work of researchers in the fields of CFS and chronic pain (Keefe et al., 1989; Ray etal., 1993). Folkman and Lazarus (1985) proposed that coping varied depending on the stressor. The relative lack of change in respondents' coping suggested that the demands of CFS may stabilize requiring the individual to make fewer alterations. Coping stability may be desirable when it is confined to adaptive efforts. However, the intransigent use of maladaptive strategies like catastrophizing, avoidance, and focusing on symptoms may be dysfunctional and reflect coping inflexibility. It has been argued that flexible "copers" adjust more successfully to illness than inflexible "copers" (Maes et al., 1996). Catastrophizing's stability may be related to another factor. The construct probably reflects personality characteristics (Diefenbach & Leventhal, 1996) thereby making its use more dependent on an individual's dispositional qualities than the demands of the situation. Efficacy of Coping and Social Support About 77% of the efficacy ratings of coping and social support dropped from time one to time four with five strategies achieving a statistically significant decline; the problem-focused strategies of seeking information (recovery and symptoms) and focusing on symptoms (symptoms). In addition ratings of the emotion-focused strategy of approach (recovery and symptoms) decreased. The overall drop in efficacy ratings suggested a subtle erosion of respondents' sense of outcome efficacy which is of concern. To the extent that coping is perceived as ineffective subjective stress may rise and the motivation to actively respond to a  170 stressor may diminish, perhaps leading to a sense of helplessness. As has been discussed, helplessness may lead patients to disengage which has been shown to be maladaptive. Decline in the efficacy of seeking information is not surprising. Once patients' queries about management and prognosis are adequately addressed information seeking may come to be seen as less of a boon. Indeed, new information may be somewhat disturbing. Through the 1990s media interest in CFS was fairly high. New treatments and diagnostic tests were regularly showcased on the television and the Internet. Some patients felt driven to keep up with all the socalled medical breakthroughs in the hopes of an eventual cure. In so doing they ran the risk of boarding a roller-coaster in which high expectations were quickly followed by bitter disappointment. CFS "old-timers" have learned that in last ten years little of practical significance has occurred to change their situation and that engaging in activities other than information gathering makes better use of their limited energy resources. Respondents' perception that focusing on symptoms is not efficacious is an adaptive change. As discussed previously, focusing has been associated with unfavorable outcomes in a number of different illness populations. However, the drop in efficacy for approach strategies is not so adaptive. Changes in approach's efficacy may be expected given the loss and sacrifice patients with CFS must tolerate. Very few patients are consistently able to accept their condition and find something positive in the experience. Instead, patients probably vacillate to a certain extent and fail to fully integrate CFS into their life. Outcomes Psychological distress. Comparison of respondents' depression and anxiety scores at the four time points revealed no differences. Respondents' low baseline scores on the psychological distress measure H A D S suggested the sample did not experience an excessive degree of anxiety or depression. Thus, the H A D S may have not have been sensitive enough to register changes in emotional well-being.  171 Functioning. Although the sample as a whole did not experience differences in functioning during the course of the study, when respondents were stratified the data revealed that respondents ill for less than four years (average duration 2.5 years) experienced a slight improvement in physical functioning. Moreover respondents ill for four years or more (average duration ten years) reported a decline in vitality. Such findings are consistent with substantial body of work associating improvement with shorter illness durations (Joyce et al., 1997; Vercoulen et al, 1996). In a Dutch study van der Werf et al. (2002) compared CFS patients ill for less than two years (N=78) with a comparison group (N=199). About 46% of patients with a short duration of illness (mean 1.4 years) self reported either full recovery (N=6) or fewer complaints (N=30) at the one-year follow up. However, 19% of the comparison group (mean duration of illness 8.8 years) self reported improvement. Post hoc analysis within the comparison group confirmed the authors' findings with 36% of patients with a short duration of illness (N=51) self reporting improvement compared to a 12% improvement rate in the long duration group. The authors found that greater fatigue, concentration problems, weaker psychosocial causal attributions, and support dissatisfaction predicted non-improvement in the short duration group (N=78). Russo et al. (1998) completed a longitudinal study of tertiary care patients (N=78) with chronic fatigue (average illness duration = 5.5 years). As reflected by the SF-36 significant differences between baseline and a 2.5 year follow-up were demonstrated for health perceptions and the mental health scales only. No changes were observed in the physical or role-functioning scales. The authors associated self-reported recovery with younger age, a decrease in the number of physical signs, lower scores on the SF-36 and the General Health Questionnaire (a measure of somatic and psychological distress). Return to work was predicted by changes in the number of physical signs and changes in psychiatric status (only 4.3%) of patient with a current psychiatric diagnosis at time 2, or time 1 and time 2, resumed work).  172 Although symptom severity was not assessed and physical examinations were not performed in the present study no differences in social support or causal attributions were found between the illness duration groups at time one. Statistically significant differences were found for; (1) age (with the mean age of the short and long duration groups, 45 and 51 years, respectively) and (2) antidepressant use (with the mean antidepressant use of the short and long duration groups, .77 and .43, respectively). Thus, younger age and pharmacological treatment for depression may be related to improvement in the short duration group. Certainly age has been identified as a prognostic indicator by Joyce et al. (1997) but the relationship between antidepressant use and improvement is less straightforward. Randomized trials have failed to demonstrate the efficacy of antidepressants in non-depressed CFS patients (Vercoulen et al., 1996; Wearden et al., 1998; Hickie, Wilson, Wright, Bennett, Wakefield, & Lloyd, 2000). Moreover, in the present study neither lifetime nor a current episode of M D D differentiated the short- and long illness duration groups. However, based on Russo et al.'s (1996) work and that of others (Sharpe et al., 1992; Wilson et al., 1994) which revealed that the presence of a psychiatric disorder at follow-up predicted poor outcomes, it is possible that antidepressant therapy may have had a salubrious effect on respondents. It is also possible that the short and long duration groups differed in illness severity and that this factor was responsible for differences in improvement between the two groups. A n oversight in the present study may have been in not capturing the respondents' subjective evaluations of change. Self-reports have been successfully utilized by a number of investigators (Sharpe et al., 1992; Friedberg et al., 2002; van der Werf, 2002) and may be more sensitive than extant objective measures. Komaroff et al. (1996) administered the SF-36 on three occasions to tertiary care CFS patients (N=233) and found no significant change. Critics have argued that due to the restriction of scores and a floor-effect in some subscales (especially role  173 limitations), the SF-36 does not allow a fine-tuned evaluation of change over time in disabled populations (Myers & Wilks, 1999). Measurement is one of the greatest challenges in CFS research. Determining the severity of CFS and therefore, changes in the condition over time has been a problem for investigators. Of note is the difficulty quantifying psychological factors and illness outcomes using extant instruments. Until a suitable measurement strategy can be developed, patients' self-reports may provide the most sensitive way of assessing the illness experience. Health care visits. Over the 21 month follow-up period respondents reduced their health care visits. This change was particularly marked when the sample was stratified into short and long duration groups. Conventional care visits dropped, especially in respondents who had CFS for over four years. The findings echoed the between-groups results suggesting that patients may become somewhat disenchanted with health care services over time. In addition, physicians probably make fewer referrals as diagnostic clarification is attained. Decreased utilization of alternative care, particularly in the long duration group, could be construed as a reflection of the limited efficacy of C A M practices. However, the decrease is more likely due to recent government cut-backs. Many CFS patients are no longer eligible to receive alternative care treatments under the Medical Services Plan of B C and few respondents are in a financial position to pay privately for health care. Significance of the Study The study described links between illness representations, self-management, and outcomes. A clearer understanding of this process may offer points of leverage that can be exploited for the development of interventions that may be implemented with greater precision and predictability. For instance, coping demonstrated negative and positive associations with outcomes. A better understanding of the complex relations between psychosocial behaviors and  174 outcomes may lessen the risk of introducing interventions that increase functional independence but inadvertently lead to psychological distress. In addition, the study's longitudinal design permitted an examination of representations and coping as they related to patients' adjustment at different points in the illness trajectory. Although few differences were apparent the findings provided information that may narrow the window when changes are likely to occur. The longitudinal design also allowed the analyst to "partial out" baseline outcome status to provide a more rigorous test of relations over time (Ray etal., 1997). For the first time, the efficacy of patients' coping and social support efforts were quantified and compared to the frequency with which these self-management behaviours were performed thus allowing a comparative evaluation of the predictive utility of the two modalities. Such information may be of interest to coping theorists who have been under some pressure to explain why coping frequency often exhibits only a modest association with outcomes. Moreover, the study identified coping and social strategies that patients, themselves, found efficacious providing practitioners with information that may assist in the development and testing of interventions that may be more acceptable because they have been endorsed by other sufferers. The study generated new information by describing the dispositional determinants of perceptions. Although not highly amenable to change, knowledge of personological factors can help clinicians (and the client) better understand the context in which cognitions are generated. Finally the study broke new ground by examining the impact of perceptions and selfmanagement on health care utilization. Conventional and alternative treatments for CFS have shown little promise in effecting recovery. Yet, patients are heavy users of health care. Identifying the cognitions and behaviors associated with patients' help seeking may lead to the development of more cost-effective ways of meeting patients' needs. As well, a better  175 understanding of the types of alternative care utilized by CFS patients may help investigators prioritize future research regarding the safety and efficacy of non-conventional modalities for the management of CFS. Limitations of the Study The study had limitations that posed threats to internal and external validity. Since the subjects were not randomly selected, sampling biases may have been introduced. Volunteer bias due to self-selection is a paramount issue. The self-selected may experience outcomes that differ from non-participants. Attrition was a problem due to the burdensome nature of the questionnaire sets and the frequency of measurement. Since it was not possible to control for attrition, demographic data and respondents' reasons for withdrawal were presented so the reader may judge if bias was introduced. The study may also have been subject to attentional biases. Knowing they were part of a study may have led respondents to alter their cognitions and behaviors. In addition, participation in the study may have affected outcomes through non-specific treatment factors like validation of the illness and supportive contact with the investigator. Data collection was confined to self-report measures that may be vulnerable to measurement bias and/or error; Retrospective assessment of health care visits may have been adversely affected by faulty recollection and/or fatigue. No attempt was made to verify comorbid illnesses by examining medical records. Functional status was not measured using objective instruments and respondents may have over-reported the degree of impairment they experienced. As well, self-report may have allowed individuals to under-report the use of coping strategies thought to be socially undesirable and/or over-report strategies deemed desirable. Data were obtained from a mail survey and there was no assurance that respondents completed the questionnaires unaided. Further, patients who experienced severe cognitive dysfunction could not complete the questionnaires and were unable to participate which limited the generalizability  176 of the findings to patients functioning at moderate or high levels. Patients with severe symptoms may have different representations of the illness and/or use other coping strategies. Finally, the study had statistical limitations. Numerous statistical analyses were performed although the possibility of experiment-wise type I error was somewhat reduced by using empirical criteria to select regression variables. Due to the study's use of subjective measures method variance may have been a problem. Significant correlations between variables may have occurred because the instruments shared similar items and assessments were conducted at the same point in time. In addition, the reliabilities of some of the instruments were suboptimal which may have attenuated the correlations. Although it is possible to correct for attenuation this procedure was not applied to the data. Chapter Summary In the preceding chapter the demographic, psychiatric, personality, and illness-related characteristics of the respondents were presented. Compared to other CFS samples described in the literature, the respondents' profile was typical of patients recruited from tertiary-care settings. As modeled by the self-regulatory framework, relations between the respondents' cognitions, coping, social support, perceptions of efficacy, and outcomes were discussed as they pertained to the literature. The personological determinants of predictive illness perceptions were discussed. A n attempt was made to characterize changes in the adjustment process over time by determining if CFS entails different stages and whether respondents' self-management and outcomes varied over a 21-month period. Little support for a specific stage theory of CFS was found and only respondents with relatively short illness durations were shown to improve. Lastly, the significance and limitations of the study were detailed. In the next and final chapter a brief summary of the study is presented along with the implications of the findings for clinical management and future research.  177 C H A P T E R SIX: Summary, Conclusions, and Implications Summary and Conclusions The purpose of the study was to describe how patients adjust to CFS. The impetus for the study arose from the literature that suggested while few CFS patients completely recover a considerable proportion experience improvement over time. Although cross-sectional studies have related psychological factors to outcomes, a systematic examination of the adjustment process using psychometrically sound instruments and longitudinal data had yet to be attempted. The study adopted a cognitive behavioral perspective due to limitations associated with the medical model. Leventhal's self-regulatory framework was selected as the most appropriate way of conceptualizing how patients' illness beliefs and coping determine adjustment which was operationalized as functioning, psychological distress, and health care utilization. Data were obtained at four time points from respondents who were medically diagnosed with CFS and whose symptoms were consistent with the C D C case definition (Fukuda et a l , 1994). A n examination of the demographic and clinical characteristics of the sample revealed that the respondents were representative of patients in tertiary care. Multivariate regression, correlational analyses, and A N O V A were used to test hypotheses generated from a review of the literature. Interpretation of the findings resulted in the following conclusions; 1.  The self-regulatory model has considerable utility for describing how patients adjust to CFS.  2.  Representations of illness, coping, and social support are better predictors of outcomes than illness-related variables, demographics, or medication use.  3.  Illness representations appeared to be more strongly related to adjustment than either coping or social support.  4.  Negative perceptions are stronger predictors of outcomes than positive perceptions.  5.  Causal attributions demonstrate differential associations. Physical attributions  178 predict greater support while emotional attributions predict psychological distress. 6.  Problem- and emotion-focused coping demonstrate differential associations with outcomes. The problem-focused strategy of accommodating to the illness related to favorable outcomes and may be considered adaptive while focusing on symptoms related to unfavorable outcomes and may be maladaptive. Similarly, emotion-focused coping characterized by an approach orientation predicts favorable outcomes and may be considered adaptive while avoidance and venting predict favorable outcomes and may be maladaptive.  7.  As evidenced by their relations with outcomes, catastrophizing is maladaptive while perceived support is adaptive.  8.  Maladaptive coping is a better predictor of outcomes than adaptive coping.  9.  Gender differences in coping exist. Men utilize more problem-focused coping while women prefer emotion-focused coping.  10.  Age is unrelated to illness representations or self-management.  11.  Illness representations and self-management are primarily related to psychological health.  12.  Relations between illness representations, self-management, and outcomes are largely contemporaneous.  13.  Personality is a better predictor of illness perceptions than psychiatric, demographic, or illness-related variables. Neuroticism and agreeableness are the most salient dispositional qualities.  14.  CFS respondents with a short illness duration experience slight improvements in physical functioning while respondents with a long duration of illness experience reductions in vitality.  15.  Health care utilization declines as illness duration lengthens.  179 16.  CFS respondents with a short illness duration perceive greater coping efficacy and are more likely to believe their efforts aid in recovery from CFS as well as help control symptoms. Implications Implications for the Management of CFS The reader is reminded that the findings were obtained from a correlational study  intended to identify the salient constructs and processes associated with patients' efforts to adjust to CFS. As such, the findings serve a more hypothesis-generating than prescriptive function. However, bearing in mind the challenges patients confront coping with the illness, the following suggestions are offered. 1. Diagnosis Although more applicable to patients in secondary or tertiary care, the findings have some relevance for primary-care physicians. The significance patients attach to the label applied to their illness underscores the sensitivity required by practitioners when providing a diagnosis. A n awareness of the stigma, life-style changes, and occupational difficulties confronting patients must inform practitioners when a diagnosis of CFS is entertained. Given the high degree of impairment and chronicity demonstrated by respondents in the study it must be seen that for some patients CFS is a disabling, life-altering condition. Practitioners are advised to refrain from making light of the illness or considering the diagnosis of CFS to be more favorable to the patient than a diagnosis that reflects a medically-explained disease. In addition, study findings associating acceptance with positive outcomes suggest that patients may benefit from a timely diagnosis. Patients with a medical diagnosis may be better equipped to deal with the financial and life-style challenges of CFS. Furthermore, patients who do not receive a diagnosis may persist in their search for an explanation of their ill-health leading to greater health care consumption.  180 2. Psychiatric Assessment of CFS Patients At present few physicians perform a comprehensive psychiatric assessment of CFS patients (G. Stiver, personal communication, May 26, 2004). Instead, a cursory mental status exam may be utilized for the purposes of determining i f the patient suffers from a psychiatric illness that excludes a diagnosis of CFS. However, as demonstrated by the study, psychiatric comorbidity (particularly M D D ) is high in CFS and a small percentage of disorders may have escaped detection. Psychiatric illnesses contribute to the overall burden of illness and have the potential for effective treatment. Thus, it is imperative that practitioners include a psychiatric assessment of all CFS patients. Some patients may be resistant, interpreting psychiatric examination to mean the symptoms are all in "their head." However, i f the patient's symptoms are treated as genuine and it is explained that psychiatric disorders may be the natural consequence of a life-altering illness, patients may become less intractable. 3. Patient Education . One of the most common strategies used by health care practitioners to facilitate adaptation to illness has been to provide patients with information. Although a cause and effect relationship has not been demonstrated, the study findings have shown that patients' illness models are related to adjustment. Practitioners may consider inquiring about patients' illness representations and challenging highly negativistic perceptions. Astute clinicians and theorists have suggested benefits may accrue i f the diagnostician offers a multifactorial illness model that includes non-physical determinants of CFS that may be controlled by the patient. Practitioners must be mindful that challenging patients' causal attributions is potentially alienating. In addition, practitioners should be aware of differences between their perceptions of the illness and the beliefs held by the patients. Discrepancies can lead to a patients' withdrawal from care and a continuing search for a supportive health care provider.  181 The study indicates that a sizeable percentage of the respondents utilized alternative care services. Practitioners may be in a position to assist patients' in evaluating C A M practices and help them avoid remedies that are harmful, useless, and/or expensive. Since patients' strongest link with the health care system is through the family physician, it would be important for him/her to have a working knowledge of C A M practices to facilitate discussion. 4. Clinical Management CFS presents a challenge to most physicians whose models of disease and healing may not be well-suited to a medically-unexplained illness. Rather than cure, it may be more helpful for physicians to focus their efforts on assisting the patient to adjust to a chronic illness. A multidisciplinary team is likely the most effective way of addressing patients' varied and complex needs but such a model of service delivery has only been established in a few large urban settings. In most areas practitioners working "solo" are expected to provide the bulk of care. Physicians may support the adjustment process by decreasing patients' overall illness burden. Providing effective, accurate, and timely documentation on disability claim forms may be an important way to help patients avoid financial impoverishment. In addition, practitioners can decrease the symptom load by treating non-CFS disorders and reducing symptoms associated with CFS. 5. Counseling Although further studies are needed to establish a causal link, evidence from the present study provides a rationale for an examination and possible modification of patients' health behaviors. Since this is a field generally outside a physician's area of expertise referral for psychological counseling may be beneficial. As part of a counseling program patients may be encouraged to use positive reinterpretation and acceptance as alternatives to avoidance and catastrophizing strategies. One issue that must be addressed is the degree to which one can  182. modify and reduce patients' tendency to catastrophize. The study results suggest catastrophizing is very stable over time and linked to neuroticism. However, C B T in chronic pain has demonstrated that catastrophizing can be reduced with consequent improvements in function and decrease in pain (Keefe et al., 1989). To refrain from catastrophizing patients can be taught to control their "internal dialogue." Analyses showed that the relations between coping and outcomes may be primarily of a contemporaneous nature. Although further testing is needed, this finding suggests that it may be important to teach patients specific strategies that deal with an immediate stressor rather than teaching strategies aimed at long-term illness consequences. The positive association between psychological well-being and social support provides some justification for encouraging patients to maintain adequate support by avoiding behaviors that may alienate others. 6. CFS Workshops Specific techniques for enhancing the lifestyle of CFS patients are beyond the scope of the present study. However, practitioners may find that the presentation and modeling of cognitive and coping behaviors found adaptive at the experiential level by other sufferers may be provided by a CFS workshop. Although they have not been subjected to rigorous examination of efficacy, such venues may provide patients with a supportive framework for dealing with CFS. Instruction often includes educational information, nutritional advice, pain management techniques, and relaxation exercises. However, before recommending a group, practitioners should "vet" CFS programs carefully. Some self-help groups can be problematical with members using the opportunity to share negative experiences and disseminate destructive beliefs. 7. Practitioner-Patient Relationships Practitioners are urged to build a supportive relationship with their patients. Based on data suggesting that patients develop a degree of autonomy over time, health care practitioners are advised to respect the skills and knowledge a CFS patient may bring to the health care  183 interaction. Thorne (1990) advised practitioners to re-examine their domain of expertise and priorities for health care relationships. The "ideal of a health care 'partnership' may be exceedingly arrogant. At best health care professionals should probably be aiming for credibility in the role of consultant to chronically ill clients, the experts" (p. 68). 8. Cognitive Behavioral Therapy Although correlational, the results from the study suggest that the self-regulatory model may serve as a guide for helping a practitioner select patients who may benefit from CBT. Contrary to the opinion of some who have assumed that CBT succeeds in reducing disability because CFS is caused by "wrong thinking," CBT is not a cure (Wessely, 2001b). Rather, CBT functions to modify psychological factors that may perpetuate certain aspects of the illness. The study findings suggest that it may be time to move beyond a C B T model that equates decreased activity with dysfunction. A distinction may be drawn between the individual who chooses to moderate stress and activity in a planful manner versus the individual whose avoidance of activity reflects disengagement brought on by a sense of helplessness in the face of seemingly overwhelming stress. In the former case, such behavior may indicate accommodating to the illness, a response found to be related to favorable outcomes. Certainly, benefits might accrue by attempting to gradually increase patients' tolerance of activity but Friedberg and Krupp (1994) warned that increasing activity without considering the patient's available energy resources may be detrimental. The individual may already be exerting as much energy as he/she can. Rather than arbitrarily increasing patients' activity the authors suggested that it may be more beneficial for patients to prioritize, focusing energy on activities that are personally meaningful and restorative. Patients whose avoidance of activity is related to disengagement may require stress reduction as opposed to exercise. Problem-solving and hypothesis-testing may assist them to identify pessimistic perceptions associated with fear and anxiety. Interventions to enhance the  184 patient's interpersonal relationships may also be helpful. However, increased activity may be counterproductive in the disengaged. Such patients may be deconditioned and slight increments in activity will likely exacerbate symptoms and prompt them to withdraw from therapy. Failure to accomplish tasks prescribed by their cognitive behavioral therapist may decrease a patient's feelings of self-efficacy and increase his/her tendency to disengage. CBT's emphasis on altering the beliefs of patients who endorse a physical attribution of cause also requires re-examination. Findings from the present study (and others conducted in the past seven years) have shown that the link between physical attributions and outcomes is inconstant and not well understood (Edwards et al., 2002; Moss-Morris, 1997). Conversely, attributing cause to psychological factors has been consistently related to emotional distress (Moss-Morris et al., 1996; Weinman et al., 1996). Rather than assuming that a patient's belief in the physiological etiology of CFS leads inexorably to an unfavorable outcome, it would behoove the therapist to identify the consequences of the patient's belief system. If the patient thinks CFS is caused by physical factors and increased symptoms reflect the exacerbation of a disease state brought on by over-exertion, then the therapist may wish to assist the patient develop a more constructive model. If, on the other hand, the patient endorses a causal model that does not inspire excessive distress or significantly proscribe activity then that belief system may be considered benign and not require intervention. Respecting patients' illness appraisals may allow therapists to tailor CBT programs accordingly and increase the technique's acceptability. Implications for Research The study results demonstrated that cognitions and coping are related to adjustment in CFS. The challenge is to further clarify the processes involved. Specifically, a better understanding is needed about the way illness representations and coping develops over time. Prospective studies that begin research at the earliest possible stage may be useful to depict the evolution of illness representations and coping responses. It must be recognized that since CFS is  185 not a life-threatening illness, funding for such an investigation might be difficult to attract. However, a possible solution would be collaboration with researchers conducting large-scale studies of patients with viral illnesses. Perhaps in that population the relevant variables could be measured and the later incidence of CFS captured. Additional research is needed to examine the personal and social context of adjustment. Although personality was found to predict perceptions in the present study, undoubtedly dispositional qualities are related to other factors associated with adjustment. In addition, the social environment in which patients cope requires closer examination. Studies have concentrated on patients' coping but their coping is likely shaped by health care professionals and family members. Measures that assess how health care providers and families cope may be used as determinants or outcomes of patients' coping. Furthermore, the role financial stressors and social policy play in the illness process deserves greater attention. Such structural elements are largely beyond the control of patients but need to be understood and likely changed if patients are to be provided with a context that facilitates adaptive and health-enhancing behaviors. The study also demonstrated a need for valid instruments in CFS research. Instrumentation is lacking that can; (1) measure outcomes relevant to the CFS experience (that is; emotional reactivity, self-ratings of mood, feelings of malaise, perceived energy levels, ability to communicate, social interaction, functioning within the family), and (2) quantify CFS-specific emotion-focused coping and support strategies. Preparatory to the development of new instruments would be qualitative studies that can identify salient domains and/or behaviors. Methodologies like grounded theory that make no a priori assumptions about behaviors would be appropriate. Using such a method would allow CFS patients to describe important aspects of adjustment from their own experience. Finally, it is imperative that researchers identify a more precise way of diagnosing CFS. Due to the subjective nature of the symptoms, and similarities between CFS and other clinical  186 entities like depression, the illness poses a considerable diagnostic challenge. Patients are frequently shuffled from physician to physician, subjected to prolonged and expensive medical evaluations, and may be undiagnosed for months or years. Skepticism about the validity of CFS abounds in the lay and professional communities, further compounding sufferers' distress. Moreover, with no clear-cut way of establishing a diagnosis of CFS, insurance companies frequently reject or delay applications for disability pensions. Furthermore, since the diagnosis of CFS depends on the qualitative examination of non-specific symptoms, researchers enlist heterogeneous samples despite rigorous screening. Consequently, the interpretation and replication of findings is problematical. Chapter Summary This chapter presented a brief summary of the study and itemized conclusions drawn from the data analysis and discussion of findings. Implications of the findings for the management of CFS and future research were offered. Concluding Remarks New disciplines like psychoneuroimmunology are demonstrating the complex manner in which somatic and psychological factors interact to produce disabling conditions like CFS. Identification of a method to objectively measure the salient processes and describe the biological substrate of CFS may eventually render moot the Cartesian-fuelled controversies surrounding the illness' etiology. The study findings support the contention that certain cognitions and coping behaviors are associated with less favorable illness outcomes. At present, these factors warrant further investigation in order to develop interventions designed to improve patients' quality of life.  187 REFERENCES Aaron, L., & Buchwald, D. (2001). A review of the evidence for overlap among unexplained clinical conditions. Annals of Internal Medicine, 134, 868-881. Abbey, S., & Garfinkel, P. (1991). Neurasthenia and chronic fatigue syndrome: The role of culture in the making of a diagnosis. American Journal of Psychiatry, 148(12), 16381645. Afari, N . , & Buchwald, D. (2003). Chronic fatigue syndrome: A review. American Journal of Psychiatry, 160(2), 221-236. Afari, N . , Eisenberg, D., Herrell, R., Goldberg, J., Kleyman, E., Ashton, S., et al. (2000). Use of alternative treatments by chronic fatigue syndrome discordant twins. Integrative Medicine, 2(2-3), 97-103. Affeck, G., Pfeiffer, C , Tennen, H., & Fifield, J. (1988). Social support and psychosocial adjustment to rheumatoid arthritis. Arthritis Care and Research, 1(2), 11-11. Affleck, G., & Tennen, H. (1996). Construing benefits from adversity: Adaptational significance and dispositional underpinnings. Journal of Personality, 64(A), 899-922. Aikens, J., Fischer, J., Namey, M . , & Rudick, R. (1997). A replicated prospective investigation of life stress, coping, and depressive symptoms in multiple sclerosis. Journal of Behavioral Medicine, 20(5), 433- 446. Aldwin, C. (1994). Stress, coping, and development. New York: Guilford. Aldwin, C , & Revenson, T. (1987). Does coping help? A re-examination of the relation between coping and mental health. Journal of Personality and Social Psychology, 53(2), 337-348. Allen, R. (Ed.). (1990). The concise Oxford dictionary. (8 ed.). Oxford: Clarendon Press. th  American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders (3 ed.). Washington, DC: Author. rd  American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4 ed.). Washington, DC: Author. th  Anderson, J., & Ferrans, C. (1997). The quality of life of persons with chronic fatigue syndrome. The Journal of Nervous and Mental Disease, 185(6), 359-367. Antoni, M . , Brickman, A., Lutgendorf, S., Klimas, N . , Imia-Fins, A., Ironson, G., et al. (1994). Psychosocial correlates of illness burden in chronic fatigue syndrome. Clinical Infectious Diseases, 75(supp 1), S73-8. Aronowitz, R. (1997). Making sense of illness: Science, society, and disease. Cambridge: Cambridge University Press.  188 Aronowitz, R. (2001). When do symptoms become a disease? Annals of Internal Medicine, 134(9: Part 2), 803-808. Asbring, P. (2001).. Chronic illness - a disruption in life: Identify-transformation among women with chronic fatigue syndrome and fibromyalgia. Journal of Advanced Nursing, 34(3), 312-319. Asbring, P., & Narvanen, A . (2002). Women's experiences of stigma in relation to chronic fatigue syndrome and fibromyalgia. Qualitative Health Research, 12(2), 148-160. A x , S., Gregg, V., & Jones, D. (1997). Chronic fatigue syndrome: Sufferers' evaluation of medical support. Journal of the Royal Society of Medicine, 90, 250-254. Ax, S., Gregg, V . , & Jones, D. (2001). Coping and illness cognitions: Chronic fatigue syndrome. Clinical Psychology Review, 21(2), 161-182. Bandura, A . (1977). Self-efficacy: Toward a unifying theory of behaviour change. Psychological Reports, 84, 191-215. Bandura, A . , Reese, L., & Adams, N . (1982). Microanalysis of action and fear arousal as a function of differential levels of perceived self-efficacy. Journal of Personality and Social Psychology, 43(\), 5-21. Barker, J., & Peterson, P. (1996). A case-control study to assess possible triggers and cofactors in chronic fatigue syndrome. The American Journal of Medicine, 100, 548-554. Barrera, M . (1981). Social support in the adjustment of pregnant adolescents: Assessment issues. In B. H . Gottlieb (Ed.), Social networks and social support (pp. 69-96). Beverly Hills: Sage. Barrera, M . , & Sandler, I. (1981). Preliminary development of a scale of social support: Studies on college students. American Journal of Community Psychology, 9, 435-447. Barrows, D. (1995). Functional capacity evaluations of persons with chronic fatigue immune dysfunction syndrome. The American Journal of Occupational Therapy, 49(4), 327-337. Baumann, L., Cameron, L., Zimmerman, R., & Leventhal, H . (1989). Illness representations and matching labels with symptoms. Health Psychology, 8(4), 449-469. Becker, P., McGregor, N . , & De Meirleir, K. (2001). A definition-based analysis of symptoms in a large cohort of patients with chronic fatigue syndrome. Journal of Internal Medicine, 250, 234-240. Behen, J., & Rodrigue, J. (1994). Predictors of coping strategies among adults with cancer. Psychological Reports, 74, 43-48. Bell, D. (1991). CFIDS: The disease of a thousand names. New York: Pollard Publications.  189 Bell, D. (1992). Chronic fatigue syndrome: Recent advances in diagnosis and treatment. Postgraduate Medicine, 91(6), 245-250. Bentall, R., Powell, P., Nye, F., & Edwards, R. (2002). Predictors of response to treatment for chronic fatigue syndrome. British Journal of Psychiatry, 181, 248-252. Bergmann, L., & Stefansson, T. (1997). Anxiety disorders: A result of long-term fatigue - the psychiatric characteristics of the sufferers of Iceland disease. Acta Neurologica Scandinavia, 96, 158-162. Bergner, M . , Bobbit, R., Carter, W., & Gilson, B. (1981). The Sickness Impact Profile: Development and final revision of a health status measure. Medical Care, 19, 787-805. Beyerstein, B. (1997). Alternative medicine: Where's the evidence? Canadian Journal of Public Health, 88 (3), 149-150. Bigatti, S., & Cronan, T. (2002). A n examination of the physical health, health care use, and psychological well-being of people with fibromyalgia syndrome. Health Psychology, 21(2), 157-166. Bishop, G. (1987). Lay conceptions of physical symptoms. Journal of Applied Social Psychology, 17, 127-146. Bishop, G. (1991). Understanding the understanding of illness: Lay disease representations. In J.A. Skelton & R.T. Coyle (Eds.), Mental representation in health and illness (pp. 3260). New York: Springer-Verlag. Blakely, A., Howard, R., Sosich, R., Murdoch, J., Menkes, D., & Spears, G. (1991). Psychiatric symptoms, personality and ways of coping in chronic fatigue syndrome. Psychological Medicine, 21, 347-362. Blondel-Hill, E., & Shafran, S. (1993). Treatment of the chronic fatigue syndrome: A review and practical guide. Drugs, 46(4), 639-651. Bolger, N . , & Zuckerman, A . (1995). A framework for studying personality in the stress process. Journal of Personality and Social Psychology, 69(5), 890-902. Bombardier, C , & Buchwald, D. (1996). Chronic fatigue, chronic fatigue syndrome, and fibromyalgia: Disability and health-care use. Medical Care, 34(9), 924-930. Brenner, G., Melamed, B., & Panush, R. (1994). Optimism and coping as determinants of psychosocial adjustment to rheumatoid arthritis. Journal of Clinical Psychology in Medical Settings, 1(2), 115-134. Broom, D., & Woodward, R. (1996). Medicalisation reconsidered: A collaborative approach to care. Sociology of Health and Illness, 18, 357-378. Brown, M . (1986). Social support, stress and health: A comparison of expectant mothers and fathers. Nursing Research, 35(2), 72- 76.  190 Bmmmett, B., Barefoot, J., Siegler, I., & Steffens, D. (2000). Relation of subjective and received social support to clinical self-report assessments of depressive symptoms in an elderly population. Journal of Affective Disorders, 61, 41-50. Buchwald, D., & Garrity, D. (1994). Comparison of patients with chronic fatigue syndrome, fibromyalgia and multiple chemical sensitivities. Archives of Internal Medicine, 154(26), 2049-2053. Buchwald, D., Pearlman, T., Kith, P., Katon, W., & Schmaling, K . (1997). Screening for psychiatric disorders in chronic fatigue and chronic fatigue syndrome. Journal of Psychosomatic Research, 42(1), 87'-94. . Buchwald, D., Pearlman, T., Kith, P., & Schmaling, K. (1994). Gender differences in patients with chronic fatigue syndrome. Journal of General Internal Medicine, 9, 397-401. Buchwald, D., Pearlman, T., Umali, J., Schmaling, K., & Katon, W. (1996). Functional status in patients with chronic fatigue syndrome, other fatiguing illnesses, and healthy individuals. The American Journal of Medicine, 101, 364-3 70. Buchwald, D., Sullivan, J., & Komaroff, A . (1987). Frequency of "chronic active epstein-barr virus infection" in a general medical practice. Journal of the American Medical Association, 257, 2297-2302. Buchwald, D., Umali, P., Umali, J., Kith, P., Pearlman, T., & Komaroff, A . (1995). Chronic fatigue and the chronic fatigue syndrome: Prevalence in a Pacific Northwest health care system. Annals of Internal Medicine, 123(2), 81-88. Buckley, L., MacHale, S., Cavanagh, J., Sharpe, M . , Deary, I., & Lawrie, S. (1998). Personality dimensions in chronic fatigue syndrome and depression. Journal of Psychosomatic Research, 46(4), 395-400. Butler, J., Chalder, T., & Wessely, S. (2001). Causal attributions for somatic sensations in patients with chronic fatigue syndrome and their partners. Psychological Medicine, 31, 97-105. Byrne, D. (1975). Neuroticism, illness behaviour and coronary heart disease. Psychotherapy and Psychosomatics, 26,317-321. Byrne, E. (2000). Aetiological considerations on some conditions in the borderlands of neurology: Chronic fatigue syndrome, pan allergy syndrome, and repetitive strain injury, a personal view. Journal of Clinical Neuroscience, 7(1), 9-12. Caplan, G. (1974). Support systems and community mental health. New York: Behavioral Publications. Carruthers, B., Jain, A., De Meirleir, K., Peterson, D., Klimas, N . , Lerner, A., et al. (2003). Myalgic encephalomyelitis/chronic fatigue syndrome: Clinical working case definition, diagnostic and treatment protocols. Journal of Chronic Fatigue Syndrome, 77(1), 7-115.  191 Cassel. J. (1974). Psychosocial processes and "stress": Theoretical formulations. International Journal of Health Services, 4, 471 -482. Cassileth, B., Lusk, E., Strouse, T., Miller, T., Brown, D., Cross, L., et al. (1984). Psychological status in chronic illness. The New England Journal of Medicine, 311, 506-511. Cathebras, P., Jacquin, L., Le Gal, M . , Fayol, C , Bourchou, K . , & Rousset. H . (1995). Correlates of somatic causal attributions in primary care patients with fatigue. Psychotherapy and Psychosomatics, 63, 174-180. Carver, C , Scheier, M . , & Weintraub, J. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality & Social Psychology, 56, 267-287. Chalder, T., Godfrey, E., Ridsdale, L., King, M . , & Wessely, S. (2003). Predictors of outcome in a fatigued population in primary care following a randomized controlled trial. Psychological Medicine, 33, 283-287. Chalder, T., Power, M . , & Wessely, S. (1996). Chronic fatigue in the community: A question of attribution. Psychological Medicine, 26, 791-800. Clarke, J. (1999). Chronic fatigue syndrome: Gender differences in the search for legitimacy. Australian and New Zealand Journal of Mental Health Nursing, 8, 123-133. Clements, A . , Sharpe, M . , Simkin, S., Borrill, J., & Hawton, K . (1997). Chronic fatigue syndrome: A qualitative investigation of patients' beliefs about the illness. Journal of Psychosomatic Research, 42(6), 615-624. Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 38, 300314. Cooper, A . (1998). Whose illness is it anyway? Why patient perceptions matter. International Journal of Community Practice, 52(8), 551-556. Cooper, L . (1997). Myalgic encephalomyelitis and the medical encounter. Sociology of Health & Illness, 19(2), 186-207. Cope, L., David, A., Pelosi, A., & Mann, A . (1994). Predictors of chronic post viral fatigue. Lancet, 344, 864-868. Cope, L., Mann A., & Pelosi, A . (1996). Psychosocial risk factors for chronic fatigue and chronic fatigue syndrome following presumed viral illness: A case-control study. Psychological Medicine, 26, 1197-1209. Cordingley, L., Wearden, A . , Appleby, L., & Fisher, L . (2001). The Family Response Questionnaire: A new scale to assess the responses of family members to people with chronic fatigue syndrome. Journal of Psychosomatic Research, 51,417-424. Costa, P., & McCrae, R. (1992). NEO PI-R: Professional Manual. Odessa, F L : Psychological Assessment Resources.  192 Costa, P., Somerfield, M . , & McCrae, R. (1996). Personality and coping: A reconceptualization. In M . Zeidner and N . Endler (Eds.),' Handbook of coping: Theory, research, applications (pp. 44-61). New York: Wiley. Coulter, P. (1988). Chronic fatigue syndrome: A n old virus with a new diagnosis. Journal of Community Health Nursing, 5, 87'-95. Couper, J. (2000). Chronic fatigue syndrome and Australian psychiatry: Lessons from the U . K . experience. Australian and New Zealand Journal of Psychiatry, 34(5), 762-773. Coyne, J., & DeLongis, A . (1986). Going beyond support: The role of social relationships in adaptation. Journal of Consulting and Clinical Psychology, 54, 454-460. Cutrona, C. (1990). Stress and social support - in search of optimal matching. Journal of Social and Clinical Psychology, 9(1), 3-14. David, A., Wessely, S., & Pelosi, A . (1988). Postviral fatigue syndrome: Time for a new approach. British Medical Bulletin, 296, 696-699. Davis, H. (1993). Chronic fatigue syndrome: Fact or fiction? BC Medical Journal, 35(8), 582584. Davison, K., & Pennebaker, J. (1997). Virtual narratives: Illness representations in online support groups. In K . Petrie & J. Weinman (Eds.), Perceptions of health and illness (pp. 463-486). Toronto, Canada: Harwood Academic. . • Deale, A., & Wessely, S. (2000). Diagnosis of psychiatric disorder in clinical evaluation of chronic fatigue syndrome. Journal of the Royal Society of Medicine, 93(6), 310-312. De Becker, P., McGregor, N . , & De Meirleir, K. (2001). A definition-based analysis of symptoms in a large cohort of patients with chronic fatigue syndrome. Journal ofInternal Medicine, 250, 234-240. DeLuca, J., Johnson, S., Ellis, S., & Natelson, B. (1997). Sudden vs gradual onset of chronic fatigue syndrome differentiates individuals on cognitive and psychiatric measures. Journal of Psychiatric Research, 31(\), 83-90. Demitrack, M . (1991). Chronic fatigue syndrome: The need for an integrative approach. Biological Psychiatry, 30,747-752. Denz-Penhey, PL, & Murdoch, J. (1993). General practitioners acceptance of the validity of chronic fatigue syndrome as a diagnosis. New Zealand Medical Journal, April, 122-124. de Ridder, D. (1997). What is wrong with coping assessment? A review of conceptual and methodological issues. Psychology and Health, 12, 417-431. de Ridder, D., & Schreurs, K. (1996). Coping, social support, and chronic disease: A research agenda. Psychology, Health, & Medicine, 1(1), 71-82.  193 de Ridder, D., & Schreurs, K . (2001). Developing interventions for chronically ill patients: Is coping a helpful concept? Clinical Psychology Review, 21(2), 205-240. de Ridder, D., Schreurs, K., & Bensing, J. (1998). Adaptive tasks, coping, and quality of life of chronically ill patients. Journal of Health Psychology, 3(1), 87-101. Diefenbach, M . , & Leventhal, H . (1996). The common-sense model of illness representation: Theoretical and practical considerations. Journal of Social Distress and the Homeless, 5(1), 11-37. Edwards, R., Suresh, R., Lynch, S., Clarkson, P., & Stanley, P. (2001). Illness perceptions and mood in chronic fatigue syndrome. Journal of Psychosomatic Research, 50, 65-68. Ell, K., Mantell, J., Hamovich, M . , & Nishimoto, R. (1989). Social support, sense of control, and coping among patients with breast, lung, or colorectal cancer. Journal of Psychosocial Oncology, 7, 63-89. Euba, R., Chalder, T., Deale, A., & Wessely, S. (1996). A comparison of the characteristics of chronic fatigue syndrome in primary and tertiary care. British Journal of Psychiatry, 168, 121-126. Evengard, B., & Klimas, N . (2002). Chronic fatigue syndrome: Probable pathogenesis and possible treatments. Drugs, 62(17), 2433-2446. Evengard, B., Schacterle, R., & Komaroff, A . (1999). Chronic fatigue syndrome: New insights and old ignorance. Journal of Internal Medicine, 246, 455-469. Feighner, J., Robins, E., Guze, S., et al. (1972). Diagnostic criteria for use in psychiatric research. Archives of General Psychiatry, 26, 57-63. Felton, B., Revenson, T., & Hinrichsen, G. (1984). Stress and coping in the explanation of psychological adjustment among chronically ill adults. Sociology in Science & Medicine, 75(10), 889-898. Felton, B., & Shinn, M . (1992). Social integration and social support: Moving "social support" beyond the individual level. Journal of Community Psychology, 20, 103-115. Fennell, P. (1995). A systemic, four-stage progressive model for mapping the CFIDJ experience. CFIDJ Chronicle, Summer, 40-46. Ferrans, C , & Powers, M . (1985). Quality of Life Index: Development and psychometric properties. Advances in Nursing Science, 8, 15-24. Findley, J., Kerns, R., Weinberg, L., & Rosenberg, R. (1998). Self-Efficacy as a psychological moderator of chronic fatigue syndrome. Journal of Behavioural Medicine, 21(4), 351362.  194 Finch, J., Barrera, M . , Bryant, W., Pool, G., & Snow-Turek, A . (1997). The factor structure of received social support: Dimensionality and the prediction of depression and life satisfaction. Journal of Social and Clinical Psychology, 16(3), 323-342. Folkman, S., & Lazarus, R. (1985). If it changes it must be a process: A study of emotion and coping during three stages of a college examination. Journal of Personality and Social Psychology, 48, 150-170. Folkman, S., & Lazarus, R. (1988). Manual for the Ways of Coping Questionnaire. Palo Alto, C A : Consulting Psychologists Press. Forbes, A., & Roger, D. (1999). Stress, social support and fear of disclosure. British Journal of Health Psychology, -/(Part 2), 165-179. Friedberg, F., Dechene, L., McKenzie, M . , & Fontanetta, R. (2000). Symptom patterns in longduration chronic fatigue syndrome. Journal of Psychosomatic Research, 48, 59-68. Friedberg, F., & Jason, L. (2002). Chronic fatigue syndrome and fibromyalgia: Clinical assessment and treatment. Journal of Clinical Psychology, 57(4), 433-455. Friedberg, F., and Krupp, L. (1994). A comparison of cognitive behavioral treatment for chronic fatigue syndrome and primary depression. Clinical Infectious Diseases, 75(Suppl. 1), S105-S110. Friedman, L., Nelson, D., Baer, P., Lane, M . , et al. (1992). The relationship of dispositional optimism, daily life stress, and domestic environment to coping methods used by cancer patients. Journal of Behavioral Medicine, 15, 127-141. Fukuda, K., Straus, S., Hickie, I., Sharpe, M . , Dobbins, J., & Komaroff, A . (1994). The chronic fatigue syndrome: A comprehensive approach to its definition and study. Annals of Internal Medicine, 121, 953-959. Fulcher, K., & White, P. (1997). Randomised controlled trial of graded exercise in patients with the chronic fatigue syndrome. British Medical Journal, 314, 1647-1652. Geisser, M . , Robinson, M . , Keefe, F., & Weiner, M . (1994). Catastrophizing, depression and the sensory, affective and evaluative aspects of chronic pain. Pain, 59, 79-83. Gentry, W., & Kobasa, S. (1984). Social and psychological resources mediating stress-illness relationships in humans. In W.D. Gentry (Ed.), Handbook of Behavioural Medicine (pp. 231-245). New York: Guilford. Gil, K., Abrams, M . , Philips, G. & Williams, D. (1992). Predicting health care use and activity level at 9-month follow-up. Journal of Consulting Clinical Psychology, 60, 267-273. Goffman, E. (1963). Stigma: Notes on management of soiled identity. Englewood Cliffs, NJ: Prentice-Hall.  195 Goldenberg, D., Simms, R., Geiger, A., & Komaroff, A . (1990). High frequency of fibromyalgia in patients with chronic fatigue seen in a primary care practice. Arthritis and Rheumatism, 33,381-386. Goodwin, S. (1997). The marital relationship and health in women with chronic fatigue and immune dysfunction syndrome: Views of wives and husbands. Nursing Research, 46(3), 138-146. Goubert, L., Crombez, G., & Van Damme, S. (2004). The role of neuroticism, pain catastrophizing and pain-related fear in vigilance to pain: A structural equations approach. Pain, 107, 234-241. Gulick, E. (1994). Social support among persons with multiple sclerosis. Research in Nursing & Health, 17, 195-206. Gunn, W., Connel, D., & Randall, B . (1993). Epidemiology of chronic fatigue syndrome: The Centers for Disease Control study. Cuba Foundation Symposium, 173, 83-92. Hadler, N . (1997). Fibromyalgia, chronic fatigue, and other iatrogenic diagnostic algorithms. Postgraduate Medicine, 102(2), 161-177. Hahn, R., & Kleinman, A . (1984). Belief as pathogen, belief as medicine: "Voodoo Death" and the placebo phenomenon in anthropological perspective. Medical Anthropology Quarterly, 3, 16-19. Hampson, S. (1997). Illness representations and the self-management of diabetes. In K. Petrie & J. Weinman (Eds.), Perceptions of health and illness (pp. 323-347). Toronto, Canada: Harwood Academic. Hardt, J., Buchwald, D., Wilks, D., Sharpe, M . , Nix, W., & Egle, U . (2001). Health-related quality of life in patients with chronic fatigue syndrome: A n international study. Journal of Psychosomatic Research, 51, 431-434. Heijmans, M . (1998). Coping and adaptive outcome in chronic fatigue syndrome: Importance of illness cognitions. Journal of Psychosomatic Research, 45(1), 39-51. Heijmans, M . , & de Ridder, D. (1998a). Assessing illness representations of chronic illness: Explorations of their disease-specific nature. Journal of Behavioural Medicine, 21(5), 485-503. Heijmans, M . , & de Ridder, D. (1998b). Structure and determinants of illness representations in chronic disease. Journal of Health Psychology, 3(4), 523-537. Heller, K., Swindle, R. & Dusenbury, L. (1986). Component social support processes: Comments and integration. Journal of Consulting and Clinical Psychology, 4, 446-470.  196 Helzer, J., Robins, L., McEvoy, L., Spitznagel, E., Stoltzman, R., Farmer, A., et al. (1985). A comparison of clinical and diagnostic interview diagnoses. Physician re-examination of lay-interviewed cases in the general population. Archives of General Psychiatry, 42, 657666. Henderson, D., & Shelokov, A. (1959). Epidemic neuromyasthenia—clinical syndrome? New England Journal of Medicine, 260, 757-764. Hickie, I., Lloyd, A., Wakefield, D., & Parker, G. (1990). The psychiatric status of patients with the chronic fatigue syndrome. British Journal of Psychiatry, 156, 534-540. Hickie, I., Wilson, A., Wright, J., Bennett, B., Wakefield, D., & Lloyd, A . (1998). A randomized double-blind, placebo-controlled trial of moclobemide in patients with chronic fatigue syndrome. Journal of Clinical Psychiatry, 61, 643-648. Hobfoll, S., Dunahoo, C., Ben-Porath, Y . , & Monnier, J. (1994). Gender and coping: The dualaxis model of coping. American Journal of Community Psychology, 22(\), 49-82. Hobfoll, S., & Freedy, J. (1990). The availability and effective use of social support. Journal of Social and Clinical Psychology, 9(1), 91-103. Holahan, C., Moos, R., & Schaefer, J. (1996). Coping, stress resistance, and growth: Conceptualizing adaptive functioning. In M . Zeidner and N . Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 24-43). New York: Wiley. Holmes, G., Kaplan, J., Gantz, N . , Komaroff, A., & Schonberger, L. (1988). Chronic fatigue syndrome: A working case definition. Annals of Internal Medicine, 108, 387-389. Holmes, G., Kaplan, J., Stewart, J., Hunt, B., Pinsky, P., & Schonberger, L . (1987). A cluster of patients with a chronic mononucleosis-like syndrome: Is Epstein-Barr virus the cause? Journal of the American Medical Association, 257, 2297-2302. Jain, S., & DeLisa, J. (1998). Chronic fatigue syndrome. A literature review from a physiatric perspective. American Journal of Physical Medicine & Rehabilitation, 77(2), 160-167. Jang, Y . , Haley, W., Mortimer, J., & Small, B. (2003). Moderating effects of psychosocial attributes on the association between risk factors and disability in later life. Aging & Mental Health, 7(3), 163-170. Jason, L. (1999). Integrating cognitive behavioural therapy and envelope theory in the treatment of a person with chronic fatigue syndrome. Cognitive and Behavioral Practice, 6, 429436. Jason, L., Fricano, G., Taylor, R., Halpert, J., Fennell, P., Klein, S., et al. (2000). Chronic fatigue syndrome: A n examination of the phases. Journal of Clinical Psychology, 56(12), 14971508. Jason, L., Richman, J., Friedberg, F., Wagner, L., Taylor, R., & Jordan, K . (1997). Politics, science, and the emergence of new disease. American Psychologist, 52(9), 973-983.  197 Jason, L., Richman, J., Rademaker, A., Jordan, A., Plioplys, A., Taylor, R., et al. (1999). A community-based study of chronic fatigue syndrome. Archives of Internal Medicine, 159, 2129-2137. Jason, L., Taylor, S., Johnson, S., Goldston, S., Salina, D., Bishop, P., et al. (1993). Prevalence of chronic fatigue syndrome-related symptoms among nurses. Evaluation & the Health Professions, 16(4), 385-399. Jason, L., Taylor, S., Kennedy, C , Song, S., Johnson, D., & Torres, S. (2000). Chronic fatigue syndrome: Occupation, medical utilization, and subtypes in a community-based sample. The Journal of Nervous and Mental Disease, 188(9), 568-576. Jason, L., & Wagner, L. (1998). Chronic fatigue syndrome among nurses. American Journal of Nursing, 98(5), 16B-16H. Jason, L., Witter, E., & Torres-Harding. (2003). Chronic fatigue syndrome, coping, optimism and social support. Journal of Mental Health, 12(2), 109-118. Jean, V., Paul, R., & Beatty, W. (1999). Psychological and neuropsychological predictors of coping patterns by patients with multiple sclerosis. Journal of Clinical Psychology, 55(1), 21-26. Jensen, M . , Turner, J., Romano, J., & Karoly, P. (1991). Coping with chronic pain: A critical review of the literature. Pain, 47, 249-283. Jensen-Campbell, L., Rosselli, M . , Workman, K., Santisi, M . , Rios, J., & Bojan, D. (2002). Agreeableness, conscientiousness, and effortful control processes. Journal of Research in Personality, 36(5), 476-489. Joffe, R., Lippert, G., Gray, T., Sawa, G., & Horvath, Z. (1987). Mood disorder and multiple sclerosis. Archives of Neurology, 44, 376-378. Johnson, S., DeLuca, J., & Natelson, B. (1996a). Assessing somatisation disorder in the chronic fatigue syndrome. Psychosomatic Medicine, 58, 50-57. Johnson, S., DeLuca, J., & Natelson, B. (1996b). Depression in fatiguing illness: Comparing patients with chronic fatigue syndrome, multiple sclerosis, and depression. Journal of Affective Disorders, 39, 21-30. Johnson, S., DeLuca, J., & Natelson, B. (1996c). Personality dimensions in the chronic fatigue syndrome: A comparison with multiple sclerosis and depression. Journal of Psychiatric Research, 30(1), 9-20. Jones, J., Ray, C , Minnich, L., et al. (1985). Evidence for active Epstein-Barr virus infection in patients with persistent, unexplained illnesses: Elevated anti-early antigen antibodies. Annals of Internal Medicine, 102, 1-7. Jones, E., & Wessely, S. (1999). Case of chronic fatigue syndrome after Crimean and Indian mutiny. British Medical Journal, 319, 18-25.  198 Joyce, J., Hotopf, M . , & Wessely, S. (1997). The prognosis of chronic fatigue and chronic fatigue syndrome: A systematic review. Quarterly Journal of Medicine, 90, 223-233. Kaplan, S. (1987). Patient reports of health status as predictors of physiologic health measures in chronic disease. Journal of Chronic Disease, 40 (Supp.), 27s-35s. Katon, W., & Russo, J. (1992). Chronic fatigue syndrome critieria. A critique of the requirement for multiple physical complaints. Archives of Internal Medicine, 152, 1604-1609. Katz, J., Ritvo, P., Irvine, M . , & Jackson, M . (1996). Coping with chronic pain. In M . Zeidner and N . Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 252-278). New York: Wiley. Keefe, F., Brown, G., Wallston, K., & Caldwell, D. (1989). Coping with rheumatoid arthritis pain: Catastrophizing as a maladaptive strategy. Pain, 37, 51-56. Kelly, K . , Soderlund, K., Albert, C , & McGarrahan, A . (1999). Social support and the chronic fatigue syndrome. Health Communication, 11(1), 21, 35. Kemp, S., Morley, S., & Anderson, E. (1999). Coping with epilepsy: Do illness representations play a role? British Journal of Clinical Psychiatry, 38, 43-58. Kerns, R., & Turk, D. (1984). Depression and chronic pain: The mediating role of the spouse. Journal of Marriage and the Family, 46, 845-852. Kleinman, A., Eisenberg, L., & Good, B. (1978). Culture, illness, and care: Clinical lessons from anthropological and cross-cultural research. Annals of Internal Medicine, 88, 251-258. Klonoff, D. (1992). Chronic fatigue syndrome. Clinical Infectious Diseases, 15, 812-823. Komaroff, A., & Buchwald, D. (1991). Symptoms and signs of chronic fatigue syndrome. Review of Infectious Diseases, 1 J(Supp.), S8-S11. Komaroff, A . , Fagioli, L., Doolittle, T., Gandek, B., Gleit, M . , Gueriero, A . , et al. (1996). Health status in patients with chronic fatigue syndrome and in general population and disease comparison groups. The American Journal of Medicine, 101, 281-290. Komaroff, A . , Fagioli, L., Geiger, A., Doolittle, T., Lee, J., Kornish, J., et al. (1996). A n examination of the working case definition of chronic fatigue syndrome. The American Journal of Medicine, 100, 56-64. Kruesi, M . , Dale, J., & Straus, S. (1989). Psychiatric diagnoses in patients who have chronic fatigue syndrome. Journal of Clinical Psychiatry, 50(2), 53-56. Krupp, L., Mendelson, W., & Friedman, R. (1991). A n overview of chronic fatigue syndrome. Journal of Clinical Psychiatry, 52(10), 403-410. Kvum, S., & Lyons, J. (1991). Assessment of coping strategies, social support, and general health status in individuals with diabetes mellitus. Psychological Reports, 68, 623-632.  199 Lane, T., Manu, P., Dale, A., & Matthews, D. (1991). Depression and somatization in the chronic fatigue syndrome. The American Journal of Medicine, 91, 335-344. Larsen, R. (1992). Neuroticism and selective encoding and recall of symptoms: Evidence from a combined concurrent-retrospective study. Journal of Personality and Social Psychology, 62, 480-488. Laursen, B., Pulkkinen, L., & Adams, R. (2002). The antecedents and correlates of agreeableness in adulthood. Developmental Psychology, 38(4), 591-603. Lawrie, S., Manders, D., Geddes, J., & Pelosi, A . (1997). A population-based incidence study of chronic fatigue. Psychological Medicine, 27, 343-353. Lawson, K., Reesor, K., Keefe, F., & Turner, J. (1990). Dimensions of pain-related cognitive coping: Cross-validation of the factor structure of the Coping Strategy Questionnaire. Pain, 43, 195-204. Layzer, R. (1998). Asthenia and chronic fatigue syndrome. Muscle and Nerve, 21, 1609-1611. Lazarus, R., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Lehman, A., Lehman, D., Hemphill, K., Mandel, D., & Cooper, L . (2002). Illness experience, depression, and anxiety in chronic fatigue syndrome. Journal of Psychosomatic Research, 52, 461-465. Leventhal, H., Benyamini, Y . , Brownlee, S., Diefenbach, M . , Leventhal, E., Patrick-Miller, L., et al. (1997). Illness representations: Theoretical foundations. In K . Petrie & J. Weinman (Eds.), Perceptions of health and illness (pp. 19-45). Toronto, Canada: Harwood Academic. Leventhal, H., & Diefenbach, M . (1991). The active side of illness cognition. In J. Skelton and R. Croyle (Eds.), Mental representation in health and illness, (pp. 247-272). New York: Springer-Verlag. Leventhal, H., Nerenz, D., & Steele, D. (1984). Illness representations and coping with health threats. In A. Baum and J. Singer, (Eds.), A handbook of psychology and health Vol.4, Hillsdale, NJ: Erlbaum. Levine, P. (1997). Epidemiologic advances in chronic fatigue syndrome. Journal of Psychiatric Research, 31, 7-18. Lewis, S. (1996). Personality, stress, and chronic fatigue syndrome. In C. Cooper (Ed.), Handbook of stress, medicine, and health, (pp. 233-249). New York: CRC. Lewis, S., Cooper, C , & Bennett, D. (1994). Psychosocial factors and chronic fatigue syndrome. Psychological Medicine, 24, 661-671. Lipowski, Z. (1970). Physical illness, the individual, and the coping process. Psychiatric Medicine, 1, 91-102.  200 Lloyd, A . (1998). Chronic fatigue and chronic fatigue syndrome: Shifting boundaries and attributions. The American Journal of Medicine, 105(3A), 7S-10S. Lloyd, A., Hickie, I., Boughton, C , Spencer, O., & Wakefield, D. (1990). Prevalence of chronic fatigue syndrome in an Australian population. Medical Journal of Austalia, 153, 522528. Lloyd, A., Hickie, C , & Loblay, S. (2000). Illness or disease? The case of chronic fatigue syndrome. Medical Journal of Australia, 172, 471 -472. Lloyd, A., Hickie, C , & Wakefield, D. (1993). Immunologic and psychologic therapy for patients with chronic fatigue syndrome: A double-blind, placebo-controlled trial. The American Journal of Medicine, 94, 197-203. Lloyd, A., & Pender, H . (1992). The economic impact of chronic fatigue syndrome. The Medical Journal of Australia, 157, 599-601. Lloyd, A., Phales, J., & Gandevia, S. (1988). Muscle strength, endurance and recovery in the post-infection fatigue syndrome. Journal of Neurology, Neurosurgery, and Psychiatry, 51, 1316-1322. Lousberg, R., Schmidt, A., & Groenman, N . (1992). The relationship between spouse solicitousness and pain behaviour. Pain, 51, 75-79. MacLean, G., & Wessely, S. (1994). Professional and popular views of chronic fatigue syndrome. The British Medical Journal, 308, 776-777. Maes, S., Leventhal, H . , & de Ridder, D. (1996). Coping with chronic diseases. In M . Zeidner and N . Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 221-251). New York: Wiley. Manne, S. & Zautra, A . (1989). Spouse criticism and support: Their association with coping and psychological adjustment among women with rheumatoid arthritis. Journal of Personality and Social Psychology, 4, 608-617. Manu, P., Lane, T., & Matthews, D. (1992). The pathophysiology of chronic fatigue syndrome: Confirmations, contradictions, and conjectures. International Journal of Psychiatry in Medicine, 22(4), 397-408. Masuda, A . , Nakayama, T., Yamanaka, T., Koga, Y . , & Tei, C. (2002). The prognosis after multidisciplinary treatment for patients with postinfectious chronic fatigue syndrome and noninfectious chronic fatigue syndrome. Journal of Behavioral Medicine, 25(5), 487497. McCluskey, R. (1992). Chronic fatigue syndrome. Comprehensive Therapy, 18, 13-16. McCully, K., & Natelson, B. (1999). Impaired oxygen delivery to muscle in chronic fatigue syndrome. Clinical Science, 97, 603-608.  201 McFarlane, A., & Brooks, P. (1988). Determinants of disability in rheumatoid arthritis. British Journal of Rheumatology, 27(7), 7-14. Mechanic, D. (1995). Sociological dimensions of illness behaviour. Social Science & Medicine, 41(9), 1207-1216. Mechanic, D. (1996). The concept of illness behaviour: Culture, situation, and personal predisposition. Psychological Medicine, 16, 1-7. Meisler, J. (2000). Toward optimal health: The experts discuss chronic fatigue syndrome. Journal of Women's Health & Gender-Based Medicine, 9(5), 477-482. Michiels, V., & Cluydts, R. (2001). Neuropsychological functioning in chronic fatigue syndrome: A review. Acta Psychiatrica Scandinavica, 103, 84-93. Milbon, C. (1991). Clinical observations of chronic fatigue and chronic fatigue syndrome. Canadian Diseases Weekly Report 17(Supp. 1), 27-31. Miller, S. (1990). To see or not to see: Cognitive informational styles in the coping process. In M . Rosenbaum (Ed.), Learned resourcefulness: On coping skills, self-control and adaptive behaviour (pp. 123-145). New York: Spring Press. Millon, C , Salvato, F., Blaney, N . , Morgan, R., Mantero-Atienza, E., Klimas, N . , et al. (1989). A psychological assessment of chronic fatigue syndrome/chronic Epstein-Barr virus patients. Psychology and Health, 3, 131-141. Minden, S., Orav, J., & Reich, P. (1987). Depression in multiple sclerosis. General Hospital Psychiatry, 9, 426-434. Moorey, S., Greer, S., Watson, M . , Gorman, C , Rowden, L., Tunmore, R., et al. (1991). The factor structure and factor stability of the Hospital Anxiety and Depression Scale in patients with cancer. British Journal of Psychiatry, 158, 255-259. Moss-Morris, R. (1997). The role of illness cognitions and coping in the etiology and maintenance of the chronic fatigue syndrome (CFS). In K . Petrie & J. Weinman (Eds.), Perceptions of health and illness (pp. 19-45). Toronto, Canada: Harwood Academic. Moss-Morris, R. & Petrie, K . (2000). Chronic fatigue syndrome. New York: Routledge. Moss-Morris, R. & Petrie, K. (2001). Discriminating between chronic fatigue syndrome and depression: A cognitive analysis. Psychological Medicine, 31, 469-479. Moss-Morris, R., Petrie, K., & Weinman, J. (1996). Functioning in chronic fatigue syndrome: Do illness perceptions play a regulatory role? British Journal of Health Psychology, 1, 15-25. Moss-Morris, R., Weinman, J., Petrie, K., Home, R., Cameron, L., & Buick, D. (2002). The Revised Illness Perception Questionnaire (IPQ-R). Psychology and Health, 17, 1-16.  202 Myers, C , & Wilks, D. (1999). Comparison of Euroquol EQ-5D and SF-36 in patients with chronic fatigue syndrome. Quality of Life Research, 8, 9-16. Natelson, B . (2002). Linking evidence and experience. Journal of the American Medical Association, 285(20), 1-13. Natelson, B . , & Lange, G. (2002). A status report on chronic fatigue syndrome. Environmental Health Perspectives, 770(Suppl.4), 673-677. Nerenz, D., & Leventhal, H. (1983). Self-regulation theory in chronic illness. In T. Burish (Ed.), Coping with chronic disease (pp. 13-37). New York: Academic Press. Nunnally, J. (1979). Psychometric theory. New York: McGraw-Hill. O'Brian, T., & DeLongis, A. (1996). The interactional context of problem-, emotion-, and relationship-focused coping: The role of the big five personality factors. Journal of Personality, 64(6), 775-809. O'Brien, M . (1993). Multiple Sclerosis: The relationship among self-esteem, social support, and coping behaviour. Applied Nursing Research, 6, 54-63. O'Reilly, P. (1988). Methodological issues in social support and social network research. Social Science & Medicine, 8, 863-873. Packer, T., Foster, D., & Brouwer, B. (1997). Fatigue and activity patterns of people with chronic fatigue syndrome. The Occupational Therapy Journal of Research, 17(3), 186199. Park, D. (1994). Self-regulation and control of rheumatic disorders. In S. Maes, H . Leventhal, M . Leventhal & M . Johnston, (Eds.), International Review of Health Psychology (Vol. 3). New York: Wiley. Parker, A., Wessely, S., & Cleare, A. (2001). The neuroendocrinology of chronic fatigue syndrome and fibromyalgia. Psychological Medicine, 31, 1331-1345. Pattison, E. (1977). A theoretical-empirical base for social system therapy. In E. Foulks, R. Wintrob, J. Westermeyer, & A . Favazza (Eds.), Current perspectives in cultural psychiatry (pp. 217-253). New York: Spectrum. Paulsen, J., & Altmaier, E. (1995). The effects of perceived versus received social support on the discriminative cue function of spouses for pain behaviours. Pain, 60, 103-110. Pawlikowska, T., Chalder, T., Hirsch, S., Wallace, P., Wright, D., & Wessely, S. (1994). Population based study of fatigue and psychological distress. British Medical Journal, 308, 763-766. Pearlin, L. (1985). Social structure and processes of social support. In S. Cohen & S. Syme (Eds.), Social support and health (pp. 43-60). Toronto, Canada: Academic Press.  203 Pearlin, L., & Schooler, C. (1978). The structure of coping. Journal of Health and Social Behavior, 19 (March), 2-21. Penley, J., Tomaka, J., & Wiebe, J. (2002). The association of coping to physical and psychological health outcomes: A meta-analytic review. Journal of Behavioral Medicine, 25(6), 551-603. Penninx, B., van Tilburg, T., Boeke, A., Deeg, D., Kriegsman, D., & van Eijk, J. (1998). Effects of social support and personal coping resources on depressive symptoms: Different for various chronic diseases? Health Psychology, 17(6), 551-558. Pepper, C , Krupp, L., Friedberg, F., Doscher, C , & Coyle, P. (1993). A comparison of neuropsychiatric characteristics in chronic fatigue syndrome, multiple sclerosis, and major depression. The Journal of Neuropsychiatry and Clinical Neurosciences, 5, 200205. Persson, L. & Sjoberg, L. (1987). Mood and somatic symptoms. Journal of Psychosomatic Research, 37,499-511. Petrie, K., Moss-Morris, R., & Weinman, J. (1995). The impact of catastrophic beliefs on functioning in chronic fatigue syndrome. Journal of Psychosomatic Research, 39(1), 3137. Pierce, G., Sarason, I., & Sarason, B. (1996). Coping and social support. In M . Zeidner & N . Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 434- 451). New York: Wiley. Pilowsky, I. (1969). Abnormal illness behavior. British Journal of Medical Psychiatry, 42, 347351. Pilowsky, I., & Spence, N . (1975). Patterns of illness behaviour in patients with intractable pain. Journal of Psychosomatic Research, 19, 279-287. Powell, R., Dolan, R., & Wessely, S. (1990). Attributions and self-esteem in depression and chronic fatigue syndrome. Journal of Psychosomatic Research, 34(6), 665-673. Prins, J.,.Bleijenberg, G., Bazelmans, E., Elving, L., de Boo, T., et al. (2001). Cognitive behavior therapy for chronic fatigue syndrome; a multicenter randomised controlled trial. Lancet, 357, 841-847. Raine, R., Haines, A., Sensky, T., Hutchings, A., Larkin, K., & Black, N . (2002). Systematic review of mental health interventions for patients with common somatic symptoms: Can research evidence from secondary care be extrapolated to primary care? British Medical Journal, 9(325), 1082-1099. Ray, C. (1991). Interpreting the role of depression in chronic fatigue syndrome. In R. Jenkins & J. Mowbray (Eds.), Post-viral fatigue syndrome (pp. 93-113). New York: Wiley.  204 Ray, C. (1992). Positive and negative social support in a chronic illness. Psychological Reports, 77,977-978. Ray, C , Jeffries, S., & Weir, W. (1995). Coping with chronic fatigue syndrome: Illness responses and their relationship with fatigue, functional impairment, and emotional status. Psychological Medicine, 25, 937-945. Ray, C , Jefferies, S., & Weir, W. (1997). Coping and other predictors of outcome in chronic fatigue syndrome: A 1-Year follow-up. Journal of Psychosomatic Research, 43(4), 405415. Ray, C , Weir, W., Cullen, S., & Philips, S. (1992). Illness perception and symptom components in chronic fatigue syndrome. Journal of Psychosomatic Research, 36(3), 243-256. Ray, C , Weir, W., Stewart, D., Miller, P. & Hyde, G. (1993). Ways of coping with chronic fatigue syndrome: Development of an illness management questionnaire. Social Science & Medicine, 37(3), 385-391. Reeves, B . (2001). CDC and prevention planning session for optimizing the CFS case definition: Practical applications, methodology. Paper presented at A C F S meeting, Seattle, W A . Reyes, M . , Dobbins, J., Steele, L., Holmes, G., Mawle, A., Stewart, J., et al. (1997). Surveillance for chronic fatigue syndrome - Four U.S. cities, September 1989 through August 1993. Morbidity and Mortality Weekly Report, 46, (No. SS-2), 1-13. Rikard-Bell, C , & Waters, B. (1992). Psychosocial management of chronic fatigue syndrome in adolescence. Australian and New Zealand Journal of Psychiatry, 26, 64-72. Rippetoe, P., & Rogers, R. (1987). Effects of components of protection-motivation theory on adaptive and maladaptive coping with a health threat. Journal of Personality and Social Psychology, 52(3), 596-604. Robbins, J., & Kirmayer, L. (1991). Attributions of common somatic symptoms. Psychological Medicine, 21, 1029-1045. Robins, L., Cottier, L., Bucholz, K., Compton, W., North, C , & Rourke, K . (2000). Computerized Diagnostic Interview Schedule IV [Computer software]. Department of Psychiatry, Washington University of St. Louis: St. Louis, M O . Robins, L., Helzer, J., Croughan, J., & Ratcliff, K. (1981). National Institute of Mental Health Diagnostic Interview Schedule. Archives of General Psychiatry, 38, 381-389. Robins, L., Helzer, J., Ratcliff, & Seyfried, W. (1982). Validity of the Diagnostic Interview Schedule, Version II: DSM-III diagnoses. Psychological Medicine, 12, 855-870. Robins, L., Helzer, J., Weissman, M . , Orvaschel, H., Greunberg, E., Burke, J., et al. (1984). Lifetime prevalence of specific psychiatric disorders in three sites. Archives of General Psychiatry, 41, 949-958.  205 Roberts, L., & Byrne, E. (1994). Single fiber E M G studies in chronic fatigue syndrome. Journal of Neurology, Neurosurgery, and Psychiatry, 57, 375-376. Rosenstiel, A., & Keefe, F. (1983). The use of coping strategies in chronic low back pain patients: Relationship to patient characteristics and current adjustment. Pain, 17, 33-44. Ross, S., Fantie, B., Straus, S., & Grafman, J. (2001). Divided attention deficits in patients with chronic fatigue syndrome. Applied Neuropsychology, 8(1), 4-11. Russo, J., Katon, W., Clark, M . , Kith, P., Sintay, M . , & Buchwald, D. (1998). Longitudinal changes associated with improvement in chronic fatigue patients. Journal of Psychosomatic Research, 45(\), 67-76. Sadovnick, A., Remick, R., Allen, J., Swartz, E., Yee, I., Eisen, K . , et al. (1996). Depression and multiple sclerosis. Neurology, 46, 628-632. Salit, I. (1985). Sporadic postinfectious neuromyasthenia. Canadian Medical Association Journal, 133, 659-663. Salit, I. (1996). The Chronic Fatigue Syndrome: A position paper. The Journal of Rheumatology, 23(3), 540-543. Salit, I. (1997). Precipitating factors for the chronic fatigue syndrome. Journal of Psychiatric Research, 31, 59-65. Saltzstein, B., Wyshak, G., Hubbuch, J., & Perry, J. (1998). A naturalistic study of the chronic fatigue syndrome among women in primary care. General Hospital Psychiatry, 20, 307316. Sarason, I., Levine, PL, Basham, R., & Sarason, B. (1983). Assessing social support: The Social Support Questionnaire. Journal of Personality and Social Psychology, 44(1), 127-139. Scharloo, M . , & Kaptein, A . (1997). Measurement of illness perceptions in patients with chronic somatic illness: A review. In K . Petrie & J. Weinman (Eds.), Perceptions of health and illness (pp. 103-153). Toronto, Ontario, Canada: Harwood. Scharloo, M . , Kaptein, A., Weinman, K., Hazes, L., Willems, L., Bergman, W., et al. (1998). Illness perceptions, coping, and functioning in patients with rheumatoid arthritis, chronic obstructive pulmonary disease, and psoriasis. Journal of Psychosomatic Research, 44(5), 573-585. Schluederberg, A., Straus, S., Peterson, P., Blumenthal, S., Komaroff, A . , Spring, S., et al. (1992). Chronic fatigue syndrome research: Definition and medical outcome assessment. Annals of Internal Medicine, 117(A), 325-331. Schmaling, K., Smith, W., & Buchwald, D. (2000). Significant other responses are associated with fatigue and functional status among patients with chronic fatigue syndrome. Psychosomatic Medicine, 62, 444-450.  206 Schober, R., & Lacroix, M . (1991). Lay illness models in the enlightenment and the 20 century: Some historical lessons. In J. Skelton and R. Croyle (Eds.), Mental representation in health and illness (pp. 10-31). New York: Springer-Verlag. th  Schreurs, K., & de Ridder, D. (1997). Integration of coping and social support perspectives: Implications for the study of adaptation to chronic diseases (1997). Clinical Psychology Review, 17(A), 89-112. Schreurs, K., van der Willige, G., Tellegren, B., & Brosschot, J. (1993). The Utrecht Coping List - Manual. Lisse, Netherlands: Swets & Zeitlinger. Schussler, G. (1992). Coping strategies and individual meanings of illness. Social Science & Medicine, 34, All All. Schwartz, C , & Daltroy, L. (1999). Learning from unreliability: The importance of inconsistency in coping dynamics. Social Science & Medicine, 48, 619-631. Schweitzer, R., Kelly, B., Foran, A., Terry, D., & Whiting, J. (1995). Quality of life in chronic fatigue syndrome. Social Science & Medicine, 41, 1367-1372. Schweitzer, R., Robertson, D., Kelly, B., & Whiting, J. (1994). Illness behaviour of patients with chronic fatigue syndrome. Journal of Psychosomatic Research, 38, 41-49. Sensky, T., MacLeod, A., & Rigby, M . (1996). Causal attributions about common somatic sensations among frequent general practice attenders. Psychological Medicine, 26, 641646. Servaes, P. Verhagen, C , & Bleijenberg, G. (2001). Determinants of chronic fatigue in diseasefree breast cancer patients. A cross-sectional study. Annals of Oncology (September), 1185-1193. Shafran, S. (1991). The chronic fatigue syndrome. The American Journal of Medicine, 90, 730739. Sharpe, M . (1993). Non-pharmacological approaches to treatment. In Ciba Foundation Symposium, 173: Chronic fatigue syndrome (pp. 298-317). Chichester, England: Wiley. Sharpe, M . (1994). Neuropsychiatric practice and opinion. Journal of Neuropsychiatry, 6(1), 189-194. Sharpe, M . (1996). Chronic fatigue syndrome. Psychiatric Clinics of North America, 19, 549573. Sharpe, M . (2002). The English chief medical officer's working parties' report on the management of CFS/ME: Significant breakthrough or unsatisfactory compromise. • Journal of Psychosomatic Research, 52, 437-438.  207 Sharpe, M . , Archard, L., Banatvala, J., Borysiewicz, L., Clare, A., David, A., et al., (1991). A report-chronic fatigue syndrome: guidelines for research. Journal of the Royal Society of Medicine, 84, 118-121. Sharpe, M . , Hawton, K., Seagroatt, V . & Pasvol, G. (1992). Follow up of patients presenting with fatigue to an infectious diseases clinic. British Medical Journal, 305, 147-152. Showalter, E. (1997). Chronic fatigue syndrome. In E. Showalter, (Ed.), Hystories: Hysterical epidemics and modern culture (pp. 115-132). London: Picador. Shumaker, S., & Hill, D. (1991). Gender differences in social support and physical health. Health Psychology, 10(2), 102-111. Silver, A., Haeney, P., Vijayadurai, P., Wilks, D., Pattrick, M . , & Main, C. (2002). The role of fear of physical movement and activity in chronic fatigue syndrome. Journal of Psychosomatic Research, 52, 485-493. Smith, G., Dobbins, C , & Wallston, K. (1991). The mediational role of perceived competence in psychological adjustment to rheumatoid arthritis. Journal of Applied Social Psychology, 21, 1218-1247. Spitzer, R., Endicott, J., & Robins, E. (1978). Research Diagnostic Criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773-782. Stewart, D. (1990). Emotional disorders misdiagnosed as physical illness: Environmental hypersensitivity, candidiasis hypersensitivity, and chronic fatigue syndrome. International Journal of Mental Health, 19(3), 56-68. Stewart, M . , Ellerton, M . , Hart, G., Hirth, A., Mann, S., Meagher-Stewart, D., et al. (1997). Insights from a nursing research program on social support. Canadian Journal of Nursing Research, 29, 93-110. Stiver, G., & Vortel, J. (1991). Chronic fatigue syndrome. BC Medical Journal, 33(10), 547-550. Stokes, J., & Wilson, D. (1984). The Inventory of Socially Supportive Behaviours: Dimensionality, prediction, and gender differences. American Journal of Community Psychology, 12(1), 53-69. Straus, S., Tosato, G., Armstrong, G., et al. (1985). Persisting illness and fatigue in adults with evidence of Epstein-Barr virus infection. Annals of Internal Medicine, 102, 7-16. Streeter, C , & Franklin, C. (1992). Defining and measuring social support: Guidelines for social work practitioners. Research in Social Work Practice, 2, 81-98. Streiner, D., & Norman, G. (1995). Health measurement scales. (2 ed.). New York: Oxford University Press. nd  208 Stokes, J., & McKirnan, D. (1989). Affect and the social environment: The role of social support in depression and anxiety. In P. Kendall & D. Watson (Eds.), Anxiety and depression (pp. 253-283). San Diego, C A : Academic Press. Stucki, G., Liang, M . , Phillips, C , & Katz, J. (1995). The Short Form-36 is preferable to SIP as a generic health status measure in patients undergoing elective total hip arthroplasty. Arthritis Care and Research, 5(3), 174-181. Sullivan, M . , Edgley, K., Mikail, S., Dehoux, E., & Fisher, R. (1992). Psychological correlates of health care utilization in chronic illness. Canadian Journal of Rehabilitation, 6(1), 1321. Sullivan, M . , Rodgers, W., Wilson, P., Bell, G., Murray, T., & Fraser, S. (2002). A n experimental investigation of the relations between catastrophizing and activity intolerance. Pain, 100, 47-53. Surawy, C , Hackmann, A., Hawton, K., & Sharpe, M . (1995). Chronic fatigue syndrome: A cognitive approach. Behavioral Research Therapy, 33(5), 535-544. Tabachnick, B., & Fidell, L . (2001). Using multivariate statistics. (4 ed.). London: Allyn and Bacon. th  Taylor, R., Jason, L., & Curie, C. (2002). Prognosis of chronic fatigue in a community-based sample. Psychosomatic Medicine, 64(2), 319-327. Thoits, P. (1986). Social support as coping assistance. Journal of Consulting and Clinical Psychology, 54, 416-423. Thoits, P. (1995). Stress, coping and social support processes: Where are we? What next? Journal of Health and Social Behavior, (Extra Issue), 53-79. Thorn, B., Ward, L., Sullivan, M . , & Boothby, J. (2003). Communal coping model of catastrophizing: Conceptual model building. Pain, 106 (1-2), 1-2. Thorne, S. (1990). Constructive noncompliance in chronic illness. Holistic Nurse Practitioner, 5(1), 62-69. Thorne, S., Best, A., Balon, J., Keiner, M . , & Rickhi, B . (2001). Ethical dimensions in the borderland between conventional and complementary/alternative medicine. Journal of Alternative & Complementary Medicine, 5(6), 907-915. Thorne, S., Paterson, B., Russell, C , & Schultz, A . (2002). Complementary/alternative medicine in chronic illness as informed self-care decision making. International Journal of Nursing Studies, 39(1), 671-683. Tiersky, L., DeLuca, J., Hill, N . , Dhar, S., Johnson, S., Lange, G., et al. (2001). Longitudinal assessment of neuropsychological functioning, psychiatric status, functional disability and employment in chronic fatigue syndrome. Applied Neuropsychology, 5(1), 41-50.  209 Torres-Harding, S., Jason, L., Cane, V., Carrico, A., & Taylor, R. (2002). Physicians' diagnoses of psychiatric disorders for people with chronic fatigue syndrome. International Journal of Psychiatry in Medicine, 32(2), 109-124. Torres-Harding, S., Jason, L., & Taylor, R. (2002). Fatigue severity, attributions, medical utilization, and symptoms in persons with chronic fatigue. Journal of Behavioral Medicine, 25(2), 99-113. Trigwell, P., Hatcher, S., Johnson, M . , Stanley, P., & House, A . (1995). "Abnormal" illness behaviour in chronic fatigue syndrome and multiple sclerosis. British Medical Journal, 311, 15-18. Turk, D., & Rudy, T. (1992). Cognitive factors and persisting pain: A glimpse into Pandora's Box. Cognitive Therapy and Research, 16, 99-122. Twemlow, S., Bradshaw, S., Coyne, L., & Lerma, B. (1997). Patterns of utilization of medical care and perceptions of the relationship between doctor and patient with chronic illness including chronic fatigue syndrome. Psychological Reports, 80, 643-658. van der Werf, S., de Vree, B., Alberts, M . , van der Meer, J., & Bleijenberg, G . (2002). Natural course and predicting self-reported improvement in patients with chronic fatigue syndrome with a relatively short illness duration. Journal of Psychosomatic Research, 53, 749-753. Van Houdenhove, B., Neerinckx, E., Onghena, P., Lysens, R., & Vertommen, B. (2001). Premorbid "overactive" lifestyle in chronic fatigue syndrome and fibromyalgia. A n etiological factor or proof of good citizenship? Journal of Psychosomatic Research, 51, 571-576. Van Wijk, C , & Kolk, A. (1997). Sex differences in physical symptoms: The contribution of symptom perception theory. Social Science & Medicine, 45(2), 231-246. Vassend, O. (1989). Dimensions of negative affectivity, self-reported somatic symptoms and health-related behaviours. Social Science & Medicine, 28, 29-36. Vaux, A. (1988). Social support: Theory, research, and intervention. New York: Praeger. Verbrugge, L., & Ascione, F. (1987). Exploring the iceberg: Common symptoms and how people care for them. Medical Care, 25, 539-569. Vercoulen, J., Swanink, C , Fennis, J., Galama, J., van der Meer, J., & Bleijenberg, G. (1996). Prognosis in chronic fatigue syndrome: A prospective study on the natural course. Journal of Neurology, Neurosurgery, and Psychiatry, 60, 489-494. Vercoulen, J., Swaniuk, C , Galama, J., Fennis, J., Jongen, P., Hommes, S., et al. (1998). The persistence of fatigue in chronic fatigue syndrome and multiple sclerosis: Development of a model. Journal of Psychosomatic Research, 45(6), 507-517.  210 Vercoulen, J., Swaniuk, C , Zitman, F., Vreden, S., Hoofs, M . , Fennis, J, et al. (1996). Randomised, double-blind, placebo-controlled study of fluoxetine in chronic fatigue syndrome, Lancet, 30(347), 858-861. Ware, J., & Sherbourne, C. (1992). The MOS 36-item Short-Form Health Survey (SF-36). Conceptual framework and item selection. Medical Care, 30(6), 473-483. Ware, J., Snow, K., & Kosinski, M . (2000). SF-36® Health Survey: Manual and interpretation guide. Lincoln, RI: QualityMetric. Ware, N . (1998). Sociomatics and illness course in chronic fatigue syndrome. Psychosomatic Medicine, (50,394-401. Ware, N . (1994). A n anthropological approach to understanding chronic fatigue syndrome. In S. Straus (Ed.), Chronic fatigue syndrome (pp. 62-82). Chichester, England: Wiley. Ware, N . & Kleinman, A . (1992). Culture and somatic experience: The social course of illness in neurasthenia and chronic fatigue syndrome. Psychosomatic Medicine, 54, 546-560. Warren, S., Warren, K., & Cockerill, R. (1991). Emotional stress and coping in multiple sclerosis (MS) exacerbations. Journal of Psychosomatic Research, 35, 31-41. Watson, D., & Hubbard, B . (1996). Adaptational style and dispositional structure: Coping in the context of the five-factor model. Journal of Personality, 64(4), 131-114. Watson, D., & Pennebaker, J. (1991). Situational, dispositional, and genetic bases of symptom reporting. In J. Skelton and R. Croyle (Eds.), Mental representation in health and illness (pp. 60-84). New York: Springer-Verlag. Wearden, A., Morriss, R., Mullis. R., Strickland, D., Pearson, D., Appleby, L., et al., (1998). Randomised, double-blind, placebo-controlled treatment trial of fluoxetine and graded exercise for chronic fatigue syndrome. British Journal of Psychiatry, 172, 485-490. Weinman, J., Petrie, K., & Moss-Morris, R., & Home, R. (1996). The Illness Perception Questionnaire: A new method for assessing the cognitive representation of illness. Psychology and Health, 11, 431 -445. Weinstein, N . , Rothman, A., & Sutton, S. (1998). Stage theories of health behaviour: Conceptual and methodological issues. Health Psychology, 17(3), 290-299. Wessely, S. (2001a). Chronic fatigue: Symptom and syndrome. Annals of Internal Medicine, 134(9), 838-843. Wessely, S. (2001b). Chronic fatigue syndrome - Trials and tribulations. Journal of the American Medical Association, 256(11), 1378- 1379. Wessely, S., Butler, S., Chalder, T., & David, A . (1991). The cognitive behavioural management of the post-viral fatigue syndrome. In R. Jenkins & J. Mowbrey (Eds.), Post-viral fatigue, syndrome (pp. 305-334). Chichester, England: Wiley.  211 Wessely, S., David, A., Butler, S., & Chalder, T. (1989). Management of chronic (post-viral) fatigue syndrome. Journal of the Royal College of General Practitioners, January, 2729. Wessely, S., & Powell, R. (1989). Fatigue syndromes: a comparison of chronic 'post-viral' fatigue with neuromuscular and affective disorders. Journal of Neurology, Neurosurgery and Psychiatry, 52, 940-948. White, C , & Schweitzer, R. (2000). The role of personality in the development and perpetuation of chronic fatigue syndrome. Journal of Pychosomatic Research, 48, 515-524. Whiting, P., Bagnall, A., Sowden, A., Cornell, J., Mulrow, C , & Pamirez, G. (2001). Interventions for the treatment and management of chronic fatigue syndrome. A systematic review. Journal of the American Medical Association, 286(11), 1360-1368. Whitlock, F., & Siskind, M . (1980). Depression as a major symptom of multiple sclerosis. Journal of Neurology, Neurosurgery, and Psychiatry, 43, 861-865. Wilson, A., Hickie, I., Lloyd, A., Hadzi-Pavlovic, D., Boughton, C , Dwyer, J., et al. (1994). Longitudinal study of outcome of chronic fatigue syndrome. British Medical Journal, 308, 756-759. Wilson, A., Hickie, I., Hadzi-Pavlovic, D., Wakefield, D., Parker, G., Straus, S., et al. (2001). What is chronic fatigue syndrome? Heterogeneity within an international multicentre study. Australian and New Zealand Journal of Psychiatry, 35, 520-527. Wood, G., Bentall, R., Gopfert, M . , & Edwards, T. (1991). A comparative psychiatric assessment of patients with chronic fatigue syndrome and muscle disease. Psychological Medicine, 21, 619-628. Wood, B., & Wessely, S. (1999). Personality and social attitudes in chronic fatigue syndrome. Journal of Psychosomatic Research, 47(4), 385-397. Woodward, R., Broom, D., & Legge, D. (1995). Diagnosis in chronic illness: Disabling or enabling - the case for chronic fatigue syndrome. Journal of the Royal Society of Medicine, 88, 325-329. Yeomans, J., & Conway, S. (1991). Biopsychosocial aspects of chronic fatigue syndrome (myalgic encephalomyelitis). Journal of Infection, 23, 263-269. Zeidner, M . , & Saklofske, D. (1996). Adaptive and maladaptive coping. In M . Zeidner and N . Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 505-531). New York: Wiley. Zigmond, A., & Snaith, R. (1983). The Hospital Anxiety and Depression Scale. Acta Psychiatrica Scandinavia, 67, 361-370.  212 Zonderman, A., Heft, M . , & Costa, P. (1985). Does the Illness Behavior Questionnaire measure abnormal illness behaviour? Health Psychology, 4(5), 425-436.  APPENDIX 1 Case Definition of CFS (Fukuda et al., 1994) Clinically evaluated, persistent, relapsing, or unexplained fatigue which is of a new or definite onset. Further, fatigue is not substantially alleviated by rest, is not the result of ongoing exertion and results in a reduction in previous levels of educational, occupational, personal or social activities. Concurrent with fatigue, the individual must have four or more of the following symptoms. The symptoms must not have predated the fatigue and must have persisted or recurred during six or more consecutive months of illness. a. impairment in concentration or short-term memory b. sore throat c. tender lymph nodes d. myalgia e.  multijoint pain without swelling or redness  f.  headaches of a new type, pattern, or severity  g. unrefreshing sleep h. post exertional malaise lasting more than 24 hours Conditions which exclude CFS include any medical condition that may explain the presence of the fatigue, for example, sleep apnea, narcolepsy, untreated hypothyroidism, and iatrogenic conditions such as the side effects of medications. Past or current diagnosis of a major depressive disorder with melancholic or psychotic features; schizophrenia, bipolar disorders; delusional disorders; dementia; anorexia nervosa or bulimia nervosa. Alcohol or other substance abuse occurring within two years of fatigue onset. Obesity defined as a body mass index equal to or greater than 45.  APPENDIX 2 Respondents' use of medications  Medication  Percentage of sample  Medication  Percentage of sample  Pain pills  55  Indigestion  10  Antidepressants  45  *Other  10  Sleeping pills  41  Blood pressure pills  8  Hormones  23  Fatty acids  6  Vitamins  22  Anti-cholesterol pills  6  Herbal preparations  14  Tranquilizers  5  Allergy medications  13  Bowel medications  4  Antimicrobials  11  Osteoporosis  4  Thyroid  11  Stimulants  2  * includes guaifenesin, sinemet, diabetes preparations, immune modulators, diclofenac, melatonin, propecia, cardiac anti-arrhythmics  214  APPENDIX 3 Transformed scores for the SF-36 subscales compared to a normative sample of females and patients with congestive heart failure (CHF) Sample  Healthy females  CHF Patients  (N=100)  (N=1412)  (N=216)  SF-3 6 subscales  Mean(SD)  Mean(SD)  Mean (SD)  Physical functioning  41.4 (21.4)  81.4 (24.6)  47.5 (31)  4.8 (.14)  77.7 (36.2)  34.4 (39.7)  Bodily pain  44.1 (21.1)  73.6 (24.3)  62.8(31)  General health  36.1 (18.8)  70.6 (21.5)  47 (24.1)  Vitality  26.2 (14.7)  58.4 (21.4)  44.3 (24.4)  Social functioning  41.4 (24)  81.5 (23.7)  71.3 (33)  Role-emotional  82.3 (90)  79.5 (34.4)  63.7 (43)  Mental health  69.1 (16)  73.3 (18.6)  74.7(21.2)  a  Role-physical  "Norms for total physical functioning and total mental health subscales not available  215  APPENDIX 4 Correlation matrix of IPQ-R subscales  Perceptions .35**  1.  Timeline  2.  Consequences  3.  Personal control  4.  Treatment control  5.  Coherence  6.  Cyclical  7.  Emotional representation  8.  Identity  9.  Psychological attributions  10  Risk factor attributions  11  Immune attributions  12  Chance attributions  p < .05 (2-tailed)  10  11  12  -.25* -5.1*  -.04  .02  .13  .09  .16  .01  .03  .03  -.36"  .25"  .02  .51**  .07  ,04  .09  .01  .21'  .26**  -.23"  .04  .19"  .01  .04  -.16  .13  .11  .02  .24"  .07  .11  -.04  ,06  .37**  -.13  -.10  .02  -.002  .09  -.06  .02  -.03  -.07  .16  .04  ,02  .14  .06  -.19  .24  .03  .08  .14  .08  -.16  .02  ,29"  .65  -.28 -.15  .50"  .12  .26" -.05  ** p <.01 (2-tailed)  216  APPENDIX 5 Correlation matrix of coping and social support strategies Strategies  1  2  3  1. Maintain  -  -.29**  .20*  -  .09  2. Accommodate 3. Seek information 4. Focus 5. Approach 6. Avoidance 7. Venting 8. Number of supportive others 9. Support satisfaction 10. Received support  -  4  5  6  7  8  9  10  -.13*  .07  .18  05  -.01  .10  -.05  .12  .27**  .34**  .08  -.13  .22*  .09  .28**  .02  .31**  .02  .15  .004  .20*  .02  -.03  .17  -.17  .12  .23*  .05  -.08  .11  .45**  .28**  .13  .10  .004  .16  -.08  .25*  .001  -.08  .14  .30**  -.22*  -.28**  .07  .39**  .48**  .41**  -.07  .36**  -.21*  -  -  -  -  -  -  -  TT  0.12  11. Catastrophizing  217  APPENDIX 6 Factor analyses of efficacy measures. Principal components extraction with varimax rotation Efficacy type  Components  Eigenvalues/  Total  Extracted  (% of Variance)  Variance  Interpretation  Problem-focused coping Maintain activity  Recovery  2  10.8(64); 1.2 (7)  71%  Break in scree plot suggests 1 factor  Symptoms  4  7.6(45); 1.6 (9)  67%  Break in scree plot suggests 1 factor  1.2 (7); 1(6) Accommodate  Seek information  Focus on symptoms  Recovery  1  9.5  73%  Subscale unidimensional  Symptoms  2  7.4 (57); 1.0 (7)  64%  Break in scree plot suggests 1 factor  Recovery  2  3.2 (54); 1.0(17)  71%  Break in scree plot suggests 1 factor  Symptoms  2  7.4 (57); 1.0 (7)  64%  Break in scree plot suggests 1 factor  Recovery  1  5.2 (58)  58%  Symptoms  2  4.1 (46); 1.05(12)  58%  Break in scree plot suggests 1 factor  218  APPENDIX 7 Factor analyses of efficacy measures. Principal components extraction with varimax rotation  Efficacy focus  Components Extracted  Eigenvalues/ (% of Variance)  Total Variance  Interpretation  Recovery  1  5.9  75%  Subscale unidimensional  Symptoms  1  4.9  61%  Subscale unidimensional  Recovery  3  3.9(36); 1.6(15); 1.2(11)  62%  Break in scree plot suggest 1 factor  Symptoms  3  3.5(32); 1.6(15); 1.3 (22)  69%  Break in scree plot suggests 1 factor  Recovery  1  2.9  73%  Subscale unidimensional  Symptoms  1  2.3  58%  Subscale unidimensional  Recovery  3  18.7 (69); 1.3 (5); 1.1 (4)  78%  Break in scree plot suggests 1 factor  Symptoms  3  18.2(67); 1.2(4)  75%  Break in scree plot suggests 1 factor  71%  Break in scree plot suggests 1 factor  71%  Break in scree plot suggests 1 factor  Emotion-focused coping Approach  Avoidance  Venting  Social Support Satisfaction  1.1 (4) Received  Recovery  10  13.8(35); 1.0-2.4 (36)  Symptoms  15 (37)  219  APPENDIX 8 Representations of illness: Comparison of short and long duration illness groups. No statistical differences between groups.  Sample  Short duration  Long duration  (N=100)  (N=31)  (N=69)  Mean (SD)  Mean (SD)  Mean (SD)  Identity  10.2 (2.2)  10.4(1.8)  10.2 (2.3)  Psych. Attribution  2.5 (.85)  2.6 (.94)  2.5 (.81)  Risk attribution  1.9 (.63)  1.8 (.58)  2.0 (.64)  Immune attribution  3.7 (.61)  3.6 (.59)  3.7 (.63)  Chance attribution  2.2 (.89)  2.1 (.75)  2.1 (.95)  Timeline (chr./acute)  3.7 (.76)  3.5 (.77)  3.7 (.75)  Consequences  4.2 (.56).  4.1 (.50)  4.2 (.58)  Personal control  3.5 (.58)  3.6 (.58)  3.52 (.58)  Treatment control  3.1 (.63)  3.3 (.53)  3.0 (.66)  Coherence  3.2 (.91)  3.1 (.81)  3.2 (.95)  Timeline (cyclical)  3.1 (.55)  3.0 (.62)  3.2 (.51)  Emotional represent.  3.2 (.70)  3.3 (.70)  3.1 (.70)  APPENDIX 9 Catastrophizing: Comparison of short and long duration illness groups. No statistical differences between groups.  CSQ score  Sample  Short duration illness group  Long duration illness group  *(N=94)  *(N=28)  *(N=66)  Mean (SD)  Mean (SD)  Mean (SD)  12.8(7)  14.4(6.6)  12.1 (.9)  *Six respondents did not experience exacerbation of symptoms in response to activity and were not eligible to complete the catastrophizing questionnaire  APPENDIX 10 Coping: Comparison of short and long duration illness groups. No statistical differences between groups.  Sample  Short duration illness  Long duration illness  Mean (SD)  Mean (SD)  Mean (SD)  Maintain activity  2.2 (.76)  3.3 (.83)  3.2 (.74)  Accommodate to illness  4.3 (.69)  4.2 (.68)  4.3 (.69)  Seek information  3.9 (.92)  4.0 (.85)  3.8 (.95)  Focus on symptoms  3.4 (.87)  3.7 (.91)  3.3 (.84)  Avoidance  1.8 (.38)  1.7 (.37)  1.8 (.38)  Approach  3.0 (.60)  3.0 (.60)  3.0 (.60)  Venting  2.3 (.63)  2.4 (.55)  2.2 (.66)  2.0 (.48)  2.0 (.58)  2.0 (.44)  Number  3.9 (2.0)  3.8(1.7)  3.9 (2.1)  Satisfaction  5.0 (.88)  5.1 (.49)  4.9(1.0)  Problem-solving coping  Emotion-focused coping  Social support Received support Perceived support  APPENDIX 11 Summary of sequential multiple regression equations of illness representations on functioning controlling for selected confounding variables  1 Demographic  Total physical functioning  Total mental health  Step/Predictor A d j . ' R 2 AR2  df  .06  2/97  .09**  r  b  t  Adj.R2 AR2  df  .21  2/97  .22***  b  t  "highest grade achieved  NS  .31  24**  2.7  "number of paid work hours/week  NS  .36  .20*  2.3 .  -.41  -.22*  -2.4  Illness-related  Not entered  Not entered  2 Control  .12  .06%  2/95  .23  .02  1/96  3 Cognitive Representations  .29  .24**  11/84  .35  .19**  11/84  - 49  "consequences  - 39***  -3.9  "identity  NS  -.26  -.19*  -2.2  "personal control  NS  .43  .26*  2.2  3 Emotional representations  a  r  as calculated in full model  .24 '  *p<.05  .12***  1/94  **p<.005  -.46  -.38***  -4.0  .24  .02  1/95  NS  ***p <.0001  223  APPENDIX 12 Summary of sequential multiple regression equations of illness representations on psychological distress controlling for selected confounding variables Step/Predictor  Anxiety  Depression Adj.R2  AR2  df  .13  .15**  3/96  1 Demographic  r  b  t  male gender  .22  .20*  2.2  number of school hours/week  -.24  -.19  .2.1  a  a  Adj.R2  AR2  df  .03  .04*  1/98  2 Illness-related  .15  .03*  1/95  Not entered  3 Control  .16  .01  1/94  .06  .03  1/97  4 Cognitive Representations  .49  .36***  11/83  .18  .20*  11/86  b  consequences  .60  49***  5.7  NS  "personal control  -.37  -2.1*  -2.0  NS  .47  42***  4.9  a  4 Emotional representations a  r  as calculated in full model  .33  1 g*** *p<.05  1/93  **p<.005  .26  23***  1/97  .50  t  48**  p<.0001  224  APPENDIX 13 Summary of sequential multiple regression equations of illness representations and coping on health care visits controlling for confounders  Step/Predictor  Total number of health care visits Adj.R2  AR2  df  r  b  Number of visits to conventional provider t  Adj.R2  AR2  Demographic  Not entered  Not entered  Illness-related  Not entered 25*** .13  Not entered  1 Control  .39  "total number of medications Antidepressants  .13  1/98 .32**  3.2  .14  df  t  Not entered  Not entered .13  .08  11/87  .04  .11  11/87  3 Emotional representations  .13  .008  1/97  .03  .001  1/97  4 Problem-focused coping  .13  .02  4/94  .03  .03  4/94  5 Emotion-focused coping  .12  .10  3/95  .02  .01  3/95  as calculated in full model  b  1/98  2 Cognitive representations  a  r  *p<.05  **p<.005  ***p<.0001  225  APPENDIX 14 Summary of sequential multiple regression equations of illness representations and coping on health care visits controlling for confounders  Step/Predictor  Number of visits to alternative care provider Adj.R2 .03  1 Demographic  df  .04*  1/98  1.3**  2/96  Not entered  Illness-related  .15  2 Control a  AR2  .36  herbal preparations  3 Cognitive representations  .18  .12  11/85  3 Emotional representations  .14  0  1/95  3 Emotion-focused coping  .15  .02  3/93  3 Problem-focused coping  .13  .01  4/92  *p<.05  **p<.005  ***p<.0001  r  b  t  APPENDIX 15 Summary of sequential multiple regression equations of coping on functioning controlling for selected confounding variables  1 Demographic  Total Physical Functioning  Total Mental Health  Step/Predictor Adj.R2  AR2  df  .06  .09*  2/97  ;  t  Adj.R2 .21  AR2 .22"  df 2/97  "highest grade achieved  NS  .31  .24**  2.7  "number of paid work hours/week  NS  .36  .20*  2.3  Illness-related  Not entered  Not entered  2 Control  .12  .06*  2/95  .23  .02  1/96  3 Emotion-focused coping  .16  .16**  3/94  .23  .02  3/93  .237  .03  4/92  Approach "Avoidance  -.30  -.24*  -2.3  "Venting  -.27  -.20*  -2.1  3 Problem-focused coping  .22  .13**  4/91  "Accommodating  .22  "Seeking information  .27**  .30  *p<.05  **p<.005  -.29*  NS  2.8 NS  NS  "Focusing on the illness  'as calculated in full model  NS  NS  "Maintaining activity  -2.7  NS  ***p<.0001  227  APPENDIX 16 Summary of sequential multiple regression equations of coping on psychological distress controlling for selected confounding variables Depression  Step/Predictor  1 Demographic  Adj.R2  AR2  df  .13  .15**  3/96  Anxiety  r  b  t  "Male gender  .22  .20*  2.2  "Number os school hours/week  -.24  -.19  -2.1  Adj.R2  AR2  df  .03  .04*  1/98  b  t  Not entered  2 Illness-related  .15  .03*  1/95  3 Control  .16  .01  1/94  .06  .03  1/97  4 Emotion-focused coping  .23  .08*  3/91  .27  22***  3/94  Approach  -.25  -.22*  -2.2  -.29  ..37***  -4.0  "Avoidance  -.30  -.24*  2.6  .21  .23*  2.5  .33  .30**  3.34  .23  .22*  2.2  -.33**  -3.4  NS  "Venting 4 Problem-focused coping  .32  17***  .23  4/90 NS  "Maintaining activity -.13  "Accommodating  .44  "Focusing on symptoms as calculated in full model  -.21  -2.2  *p<.05  ** p<.005  45***  jo,***  4/93 * -.35  NS  NS  "Seeking information  a  r  4.5  NS  *** p<.0001  228  A P P E N D I X 17 Summary of sequential multiple regression equations of catastrophizing on functioning controlling for selected confounding variables Step/Predictor  Total Physical Functioning  Total Mental Health Adj.R2  AR2  df  .07  .09"  2/91  1 Demographic  t  Adj.R2  AR2  df  .22  .23***  2/91  "Highest grade  .29  "Number of paid work hours  .40  Illness-related  .12  .07*  2/89  3 Castastrophizing  .27  .14***  1/88  *p<.05  2.4 3.7  Not entered  Not entered  2 Control  'as calculated in full model  .22*  -.50  -.40***  4.2  .24  .03  1/90  .25  .02  1/89  * * *p<.0001  229  APPENDIX 18 Summary of sequential multiple regression equations of catastrophizing on psychological distress controlling for selected confounding variables Step/Predictor  1  Demographic  Depression Adj.R2  A R 2  df  .13  .16**  3/90  3 4  a  Illness-related Control Castastrophizing  as calculated in full model  r  b  .22  "male gender 2  Anxiety  .17  .05*  1/89  .16  .003  1/88  .39  .22***  1/87  *p<.05  **p<.005  .21  t  Adj.R2  A R 2  df  .04  .05*  1/92  r  b  t  .42  .38***  3.9  2.5 Not entered .06  .56  5.8  .19  .02  1/91 1/90  ***p<.0001  230  APPENDIX 19 Summary of sequential multiple regression equations of catastrophizing on health care visits controlling for selected confounders  Adj.R2  AR2  t  df  Adj.R2  AR2  Demographic  Not included  Not included  Illness-related  Not included  Not included  1 Control  .14  .14***  1/92  "Number of medications 2 Catastrophizing  as calculated in full model  l  Number of visits to conventional provider  Total Number of Health Care Visits  Step/Predictor  .39 .13  *p<.05  1/91  .38 ***  df  .03  .04*  1/92  .02  .003  1/91  4.0  ***p<.0001  231  A P P E N D I X 20 Summary of sequential multiple regression equations of catastrophizing on health care visits controlling for selected confounders Step/Predictor  Number of visits to alternative care provider Adj.R2  1 Demographic a  2 Control  as calculated in full model  r  b  t  -.21  -.21*  -2.0  1/92  Not entered .17  .14**  2/90  herbal preparations  3 Catastrophizing a  .05*  df  male gender  Illness-related  a  .04  AR2  .38 .16 *p<.05  .003  1/89  **p<.005  .29**  2.9  APPENDIX 21 Summary of sequential multiple regression equations of social support on functioning controlling for selected confounders Step/Predictor Adj.R2 1 Demographic  Total Physical Functioning  Total Mental Health  .06  AR2 .09  t  df 2/97  Adj.R2 .21  AR2 .22"  df 2/97  NS  "highest grade achieved "number of paid work hours/week Illness-related  Not entered .12  .06*  2/95  .23  .02  1/96  3 Perceived social support  .22  .11**  2/93  .25  .03  2/94  "Support satisfaction  as calculated in full model  2.7  .36  .20*  2.3  NS  NS  "Support number  a  .24**  Not entered  2 Control  3 Received social support  .31  .16  *p<.05  .05"  1/94  .35  .39**  3.8  .19  .24*  2.5  **p<.005  NS .27  .04"  1/95  .24  .22*  ***p<.0001  233  APPENDIX 22 Summary of sequential multiple regression equations of social support on psychological distress controlling for selected confounders Depression  Step/Predictor  1 Demographic  Adj.R2  AR2  df  .13  .15**  3/96  r  Anxiety . b  t  "Male gender  .22  .20*  2.2  "Number os school hours/week  -.24  -.19  -2.1  Adj.R2  AR2  df  .03  .04*  1/98  .15  .03*  1/95  3 Control  .16  1/94  .06  4 Perceived social support  .26  .01 19**  2/92 '  .14 NS  "Support number 4 Received social support  a  as calculated in full model  .24  .08**  *p<.05  1/93  b  t  -.32  -.25*  -2.3  Not entered  2 Illness-related  "Support satisfaction  r  -.40  -.27**  -2.6  -.31  -.29**  -3.2  .03 09**  1/97 2/95  NS .09  .03  1/96  ** p<.005  234  A P P E N D I X 23 Summary of sequential multiple regression equations of social support on health care visits controlling for selected confounders Step/Predictor  Total Number of Health Care Visits Adj.R2  AR2  df  r  b  Number of visits to conventional provider t  Adj.R2  AR2  Demographic  Not entered  Not entered  Illness-related  Not entered  Not entered  1 Control  .13  .15***  .39  "total number of meds. Antidepressants 2 Perceived social support'  .13  1/98 .32**  3.2  .14  df  .02  .07*  as calculated in full model  .22*  2.3  .28  :25*  2.3*  NS  "Support satisfaction  l  .22 2/96  "Support number 2 Received social support  t  Not included .09  2/96  b  1/98  Not included .14  r  .13  *p<.05  .001  .03  1/97  **p<.005  .007  1/97  ***p<.0001  235  APPENDIX 24 Summary of sequential multiple regression equations of social support on health care visits controlling for selected confounders Step/Predictor  Number of visits to alternative care provider Adj.R2  1 Demographic Illness-related 2 Control a  .03  AR2  df  .04*  1/98  1.3**  2/96  Not entered .15  .36  herbal preparations  3 Perceived social support  .14  .006  2/94  3 Received social support  .15  .002  1/95  a  as calculated in full model  r  *p<.05  **p<.005  .26*  2.5  APPENDIX 25 Summary of sequential multiple regression equations of illness representations on predictive coping strategies Approach  Step/Predictor  Avoidance t  Adj.R2  AR2  df  1 Cognitive representations  .06  .16  11/88  1 Emotional representations  .06  .07**  1/98  -.26  -.26**  -2.7  Adj.R2  AR2  df  .02  .13  11/88  .08  .09**  1/98  .01  .12  .14  11/88  .24*  *p<.05  ** p<.005  .09  .10**  1/98  .32  3.1  .38  .35**  3.1  .55  .55***  6.4  11/82  Consequences 1 Emotional representations  .30**  Catastrophizing  Venting 1 Cognitive representations  .30  .32**  3.4  .30  .30***  1/92  ***p<.0001  237  A P P E N D I X 26 Summary of sequential multiple regression equations of illness representations on predictive coping and social support strategies Focus on symptoms  Accommodating to the illness  Step/Predictor Adj.R2  A R 2  df  .06  .17  11/88  1 Cognitive representations  t  Adj.R2  AR2  df  .14  .24**  11/88  Consequences 1 Emotional representations  .02  .16  .25**  .09 -.29  Causal attributions: Immune  NS  p<.05  .  **p<.005  .13***  .07  .08**  1/98  -.28  -.31** -.28**  1/98  .34  .29**  2.7  .36  .36***  3.8  Received social support  11/88  Causal attributions: Risk 1 Emotional representations  |:  .12  1/98  Perceived social support (Satisfaction)  Step/Predictor 1 Cognitive representations  .03  t  .19*  11/88  -2.6 -2.9**  .006  .01  -.11  -.25*  -2.0  .16  .25*  2.4  1/98  *** p<.0001  238  A P P E N D I X 27 Summary of sequential multiple regression equations of illness representations on residualized functioning scores controlling for confounders Step/Predictor  1 Dependent 2  variable  Total Mental Health (T2)  (Tl)  Demographic  b  t  Control  Adj.R2  Adj.R2  A R 2  df  .32  33***  .1/90  .53  .31  .009  2/88  .53  A R 2  df 1.90  .01  2/88  .32  .02  2/86  .54  .01  1/87  .56  .07  11/76  .54  .005  1/86  4 Cognitive  representations  .37  .12  11/75  4 Emotional  representations  .35  .03*  1/85  -.43  -.22*  -2.2  4 Perceived  social  support  .38  .02  2/84  .53  .001  2/85  4 Received  social  support  .32  .008  1/85  .54  .002  1/86  *p<.05  (T2)  Not entered  Not entered  Illness-related 3  r  Total Physical Functioning  ***p<.0001  239  APPENDIX 28 Summary of sequential multiple regression equations of illness representations and support on residualized psychological distress scores for confounders Step/Predictor  Depression (T2) Adj.R2  AR2  Anxiety (T2) df  Adj.R2  AR2  df  1 Dependent variable (Tl)  .59  .60***  1/90  .54  .55***  1/90  2 Demographic  .58  .005  3/87  .54  .002  1/89  3 Illness-related  .58  0  1/86  4 Control  .58  .009  1/85  .53  0  1/88  5 Cognitive Representations  .61  .06  10/75  .51  .03  11/77  5 Emotional Representations  .59  .01  1/84  .53  .005  1.87  5 Perceived social support  .58  .009  2/83  .52  0  2/86  5 Received social support  .58  0  1/84  .53  .001  1/87  ***p<.0001  Not entered  APPENDIX 29 Summary of sequential multiple regression equations of illness representations and support on residualized health care visit scores controlling for confounders Step/Predictor  1 Dependent variable (Tl)  Total Number of Health Care Visits (T2) Adj.R2  AR2  df  .36  37***  1/90  r  b  Number of visits to conventional provider (T2) t  Adj.R2  AR2  df  .24  24***  1.90  Demographic  Not entered  Not entered  Illness-related  Not entered  Not entered  2 Control  .40  .04*  1/89 .40  "Number of medications  .24*  .24  .01  1/89  2.9  3 Cognitive representations  .34  .02  11/78  .22  .08  11/78  3 Emotional representations  .40  .006  1/88  .23  .001  1/88  3 Perceived social support  .39  .004  2/87  .23  .01  2/87  3 Received social support  .39  .005  1/88  .23  .003  1/88  a  as calculated in full model  *p<.05  ***p<.0001  241  APPENDIX 30 Summary of sequential multiple regression equations of illness representations and support on residualized health care visit scores controlling for confounders Step/Predictor  Number of visits to alternative care provider (T2) Adj.R2  A R2  df  1 Dependent variable (Tl)  .37  .37***  1/90  2 Demographic  .36  .002  1/89  Illness-related 3 Medication use  Not entered .44  .09 **  2/87 .43  .25 **  3.0  .30  .21 =  2.6  Number of supportive others  .03  -.20*  -2.1  "Support satisfaction  .12  .20*  2.2  Vitamin use 4 Cognitive representations a  .51  .12"  11/76"  Immune attributions  4 Emotional representations  .44  .004  1/86  4 Perceived social support  .47  .03*  2/85*  a  4 Received social support  as calculated in full model  l  .44  1/86  *p<.05  **p<.005  ***p<.0001  APPENDIX 31 Summary of sequential multiple regression equations of coping on residualized functioning scores controlling for confounders Step/Predictor  Total Mental Health (T2) Adj.R2  AR2  df  1/90  .53  53***  1.90  2/88  .53  .01  2/88  Adj.R2  AR2  df  1 Dependent variable (Tl)  .32  33***  2 Demographic  .31  .009  Illness-related  Total Physical Functioning (T2)  r  b  t  Not entered  Not entered  3 Control  .32  .02  2/86  .54  .01  1/87  4 Emotion-focused coping  .38  .07*  3/83  .58  .01  3/84  .52  .002  4/83  "Approach  .17  .24*  2.5  "Avoidance  -.35  -.27**  -2.8  4 Problem-focused coping  a  as calculated in full model  .33  *p<.05  .04  4/82  **p<.005  ***p<.0001  243  A P P E N D I X 32 Summary of sequential multiple regression equations of coping on residualized psychological distress scores controlling for confounders Step/Predictor  Depression (T2) Adj.R2  AR2  Anxiety (T2) df  Adj.R2  AR2  df  1 Dependent variable (Tl)  .59  .60***  1/90  .54  .55***  1/90  2 Demographic  .58  .005  3/87  .54  .002  1/89  3 Illness-related  .58  000  1/86  4 Control  .58  .009  1/85  .53  0  1/88  5 Emotion-focused coping  .58  .01  3/82  .52  .005  3/85  5 Problem-focused coping  .59  .01  4/81  .56  .04  4/84  ***p <.0001  Not entered  APPENDIX 33 Summary of sequential multiple regression equations of coping on residualized health care visit scores controlling for confoundersi  1 Dependent variable (Tl)  Number of visits to conventional provider (T2)  Total Number of Health Care Visits (T2)  Step/Predictor Adj.R2  AR2  df  .36  .37***  1/90  t  Adj.R2  AR2  df  .23  .24***  1/90  Demographic  Not entered  Not entered  Illness-related  Not entered  Not entered  2 Control  .40  . .04*  1/89  "Number of medications 3 Emotion focused coping  .40 .43  .05*  'as calculated in full model  -.16  .40  *p<.05  .03  •.21"  1/89  .29  .07*  3/86 -.19  -2.4 .21  4/85  **p<.005  .01  2.9  3/86  "Approach 3 Problem-focused coping  .24  .24  .005  -.24*  -2.4  4/85  ***p<.0001  245  A P P E N D I X 34 Summary of sequential multiple regression equations of coping on residualized health care visit scores controlling for confounders Step/Predictor  Number of visits to alternative care provider (T2) Adj.R2  A R2  df  1 Dependent variable (Tl)  .37  .37***  1/90  2 Demographic  .36  .002  1/89  Illness-related 3 Control  Not entered .44  .09**  2/87 .43  Vitamin use 4 Emotion-focused coping  .43  .01  3/84  4 Problem-focused coping  .46  .04  4/83  a  as calculated in full model  *p<.05  **p<.005  .25 **  ***p<.0001  3.0  A P P E N D I X 35 Summary of sequential multiple regression equations of catastrophizing on residualized functioning scores controlling for confounders  A d j . R 2 1 Dependent 2  variable  (Tl)  Demographic Control  4  Catastrophizing  *p<.05  A R 2  t  df  **p<.005  Adj.R2  AR2  d  (T2)  f  .31  .32***  1/84  .56  .56***  1/84  .31  .01  2/82  .56  .008  2/82  Not entered  Not entered  Illness-related 3  Total Physical Functioning  Total Mental Health (T2)  Step/Predictor  .31  .02  2/80  .35  .03*  1/79  -.44  -.23*  -2.2  .57  .02*  1/81  .57  .008  1/80  ***p<.0001  247  A P P E N D I X 36 Summary of sequential multiple regression equations of catastrophizing on residualized psychological distress scores controlling for confounders Step/Predictor  Depression (T2) Adj.R2  AR2  Anxiety (T2) df  Adj.R2  AR2  df  1 Dependent variable (Tl)  .59  .60***  1/84  .53  54***  1/84  2 Demographic  .58  .006  3/81  .53  .002  1/83  3 Illness-related  .58  000  1/80  4 Control  .58  .006  1/79  .53  .001  1/82  5 Catastrophizing  .59  .01  1/78  .53  .01  1/81  ***p<.0001  Not entered  APPENDIX 37 Summary of sequential multiple regression equations of catastrophizing on residualized health care visit scores controlling for confounders  1 Dependent variable (Tl)  Number of visits to conventional provider (T2)  Total Number of Health Care Visits (T2)  Step/Predictor Adj.R2  AR2  df  .36  .36***  1/84  t  Adj.R2  AR2  df  .23  .24***  1/84  Demographic  Not entered  Not entered  Illness-related  Not entered  Not entered  .39  2 Control  .04  1/83 .39  Number of medications 3 Catastrophizing  *p<.05  **p<.005  .39  .004  1/82  .21  .24  .01  1/83  .23  .002  1/82  2.3  ***p<.0001  249  APPENDIX 38 Summary of sequential multiple regression equations of catastrophizing on residualized health care visit scores controlling for confounders Step/Predictor  Number of visits to alternative care provider (T2) Adj.R2  A R2  df  1 Dependent variable (Tl)  .40  .40***  1/84  2 Demographic  .39  .001  1/84  Illness-related  Not entered  3 Control  .46  .07**  2/81  Vitamin use  .43  4 Catastrophizing  *p<.05  **p<.005  .46  ***p<.0001  .006  1/80  .25 **  3.0  A P P E N D I X 39 Summary of sequential multiple regression equations of illness representations on residualized scores of predictive coping variables Step/Predictor  Approach (T2) Adj.R2  AR2  df  1 Dependent variable (Tl)  .28  29***  2 Cognitive representation  .35  2 Emotional representation  .30  Adj.R2  AR2  df  1/90  .38  3g***  1/90  .14  11/79  .32  .02  11/79  .02  1/89  .38  .01  1/89  Step/Predictor  r  Avoidance (T2) b  t  Venting (T2)  Catastrophizing (T2)  1 Dependent variable (Tl)  .24  25***  1/90  .44  2 Cognitive representation  .26  .10  11/79  .43  .07  11/78  2 Emotional representation  .30  .06**  1/89  .46  .02  1/78  *p<.05  **p<.005  ***p<.0001  .38  r  43**  2.8  1/79  b  t  APPENDIX 40 Summary of sequential multiple regression equations of illness representations on residualized scores of predictive coping and social support variables  Step/Predictor  Accommodating to the Illness (T2) Adj.R2  AR2  Focus on the Symptoms(T2)  df  Adj.R2  AR2  df  1 Dependent variable (Tl)  .40  41***  1/90  .64  .65***  1/90  2 Cognitive representations  .44  .10  11/79  .63  .02  11/79  2 Emotional representations  .41  .01  1/89  .64  0  1/89  Perceived Social Support  Received Social Support  (Satisfaction) T2  (T2)  Adj.R2  AR2  df  Adj.R2  AR2  df  1 Dependent variable (Tl)  .43  43***  1/90  .51  .52**  1/90  2 Cognitive representations  .47  .10  11/79  .53  .07  11/79  2 Emotional representations  .42  .001  1/89  .51  .002  1/89  **p<.005  ***p<.0001 252  APPENDIX 41 Comparison of variance and standardized betas for regression models composed of time 1, 2, and 3 data Mental Health Predictor Cognitive Represent.  EmotionFocused Coping Catastroph. Problemfocused Coping  Subscale  a  2  Consequence Identity Personal control Coherence Cyclic  2  Approach Avoidance Venting  beta beta beta R A Catastrophize beta 2  T1 - Time 1 (N= 100)  b  .39*** -.37 -.30** .35*  C  T3  .30** -.22 -.29 0.14  .16**  24***  .16**  .16  .29**  0.19  .14*** .38*** -.40*** -.67***  T1  a  b  T2  C  T3  Depression a  T1  *T2  Anxiety  T3  T1  C  a  .38*** 0.01 '-.25* -.03  0.23 .36*** .32*** .29*** .22* -0.16 0.06 .06 -.09  -.001  .25**  0.03  .06** 0 -.27** -.01  22*** .30*** .05* 48*** .58*** .24*  .08*  .17*  .31***  .14**  beta  0.07  -.21*  -0.007  R A  .11**  .11*  .06  beta beta  -.13 .39***  .26* .12  .24* .03  0.03  .01  .28*  .16*  beta beta beta  T2 - Time 2 (N=92) T3 - Time 3 (N=85) C  b  .19** -.22* -.19 0.26*  .10** .02 -.34** -.14  .22**  2  Number Satisfaction  T2  R A 2  Maintain activity Accommodate Seeking info Focus on symptoms  .24** -.39** -.16 .02  b  beta beta R A  Perceived support  a  R A beta beta beta  T1  Physical Function  *p<.05  **p<.005  ***p<.0001  T2  C  T3  .20*  0.18  -.32*  -.09  .22*** .12  0.06  .30** .18 .13*** 22*** .38*** 43***  .17 .12** .35**  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items