Self-regulation activities used by individuals with schizophrenia

The purpose of this study was to describe self-regulation processes individuals with schizophrenia used to understand their illness, manage symptoms, and maintain wellness. Theories of illness cognition and stress vulnerability served as guides for selection of study variables. A cross-sectional, descriptive design was used to determine subjects' beliefs about and experiences with their illness. Data were collected on measures of illness representation (Illness Perception Questionnaire [IPQ]), early warning symptoms (Early Symptom Scale [ESS]), daily hassles (Daily Hassles Scale-Revised [DHS-R]), mood (Profile of Mood States-Short Form-[POMS-SF]), current symptoms (Psychiatric Assessment Scale [PAS] and the Symptom Scale of the IPQ [IPQ-S]). Semistructured interviews provided information about illness history, knowledge of illness processes, and learned methods for illness management. Sixty-nine subjects were recruited from five settings. Their mean age was 44.3 years; 54% were female, and 86% were single, divorced, or widowed. Forty-nine percent were diagnosed with schizophrenia, 36% with schizo-affective disorders, and 14% with psychosis, NOS. Average years of treatment were 19.2 and mean number of hospitalizations were 6.14. The IPQ measured subjects' beliefs about cause, time line, consequences, and control/cure of their illness. Causes were attributed to stress, heredity, state of mind, and other people. Subjects noted consequences from their long-term illness but believed they could control aspects of their illness. Internal consistency reliability was noted for consequences and symptom subscales only. Subjects selected similar symptoms across mood states, early warning symptoms, and current symptoms. Common experiences were trouble concentrating, difficulty remembering things, depression, worry, fatigue, tension, and nervousness. Subjects reported daily hassles concerning weight; trouble relaxing; regrets over past decisions; troubling thoughts about the future; being lonely; trouble with spelling, reading, and arithmetic; and filling out forms. Study variables that correlated with IPQ consequence and symptoms subscales were DHS-R and three subscales of the POMS-SF (e.g., fatigue, tension, and confusion). Illness histories noted symptom awareness, consequences, and efforts to control symptoms. Self-regulation activities suggested subjects used early warning symptoms, stress reduction activities, and help seeking from providers or self-help groups. The cross-sectional design of the study did not allow for prospective confirmation of subjects' self-regulation activities during illness instability.

SELF-REGULATION ACTIVITIES USED BY INDIVIDUALS WITH SCHIZOPHRENIA by Sandra L. Talley A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Nursing The University of Utah December 1998 Copyright C Sandra L. Talley 1998 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITIEE APPROVAL of a dissertation submitted by Sandra L. Talley This dissertation has been read by each member of the following supervisory committee and by maj ority vote has been found to be satisfactory. ~ " Chair: B. Lee Walker Robert T. Croyle Maeona K. Kramer Paula M. Meek Lillian M. Nail THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the dissertation of Sandra L. Talley in its final form and have found that (1) its format, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. J/Hr. ::1., (7 7 >! Date B. Lee Walker Chair, Supervisory Committee Approved for the Major Department Approved for the Graduate Council ~~ <;;~ C-Qct-1'-,-- David S. Chapman , Dean of The Graduate School ABSTRACT The purpose of this study was to describe self-regulation processes individuals with schizophrenia used to understand their illness, manage symptoms, and maintain wellness. Theories of illness cognition and stress vulnerability served as guides for selection of study variables. A cross-sectional, descriptive design was used to determine subjects' beliefs about and experiences with their illness. Data were collected on measures of illness representation (Dlness Perception Questionnaire [IPQ)), early warning symptoms (Early Symptom Scale [ESS]), daily hassles (Daily Hassles Scale­Revised [DHS-R)), mood (profile of Mood States-Short Form-[pOMS-SF]), current symptoms (psychiatric Assessment Scale [p AS] and the Symptom Scale of the IPQ [IPQ-S)). Semistructured interviews provided information about illness history, knowledge of illness processes, and learned methods for illness management. Sixty-nine subjects were recruited from five settings. Their mean age was 44.3 years; 54% were female, and 86% were single, divorced, or widowed. Forty-nine percent were diagnosed with schizophrenia, 36% with schizo-affective disorders, and 14% with psychosis, NOS. Average years of treatment were 19.2 and mean number of hospitalizations were 6.14. The IPQ measured subjects' beliefs about cause, time line, consequences, and control/cure of their illness. Causes were attributed to stress, heredity, state of mind, and other people. Subjects noted consequences from their long-term illness but believed they could control aspects of their illness. Internal consistency reliability was noted for consequences and symptom subscales only. Subjects selected similar symptoms across mood states, early warning symptoms, and current symptoms. Common experiences were trouble concentrating, difficulty remembering things, depression, worry, fatigue, tension, and nervousness. Subjects reported daily hassles concerning weight; trouble relaxing; regrets over past decisions; troubling thoughts about the future; being lonely; trouble with spelling, reading, and arithmetic; and filling out forms. Study variables that correlated with IPQ consequence and symptoms subscales were DHS-R and three subscales of the POMS-SF (e.g., fatigue, tension, and confusion). Illness histories noted symptom awareness, consequences, and efforts to control symptoms. Self-regulation activities suggested subjects used early warning symptoms, stress reduction activities, and help seeking from providers or self-help groups. The cross-sectional design of the study did not allow for prospective confirmation of subjects' self-regulation activities during illness instability. v T ABLE OF CONTENTS ABSTRACT ....................................................... IV LIST OF TABLES ................................................. viii ACKNOWLEDGNlENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ix INTRODUCTION ................................................... 1 Chapter I. SCIllZOPIfRENIA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Nosology of Schizophrenia .................................. 5 Theoretical Models of Schizophrenia .......................... 9 Illness Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Symptom Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Coping Responses to Symptom Onset ........................ 22 Self-Regulation in Schizophrenia ............................ 25 Summary .............................................. 26 II. SELF-REGULATION .......................................... 29 Theoretical Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Self-Regulation Model .................................... 32 Illness Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Appraisal .............................................. 36 Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Studies of Self-Regulation/11lness Representation ................ 41 Summary .............................................. 45 III. METHODOLOGY ............................................ 49 Design ................................................ 49 Criteria for Subject Selection ............................... 51 Data Collection Procedures ................................ 51 Measurement Methods .................................... 55 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 IV. RESULTS ................................................... 69 Sample ................................................ 69 Characteristics of the Subjects .............................. 71 Illness Representation-Qualitative Results . . . . . . . . . . . . . . . . . . . . . . 94 Correlations Among Variables ............................. 106 Summary ............................................. 113 Self-Regulation Activities Based on Narrative Account ........... 114 V. DISCUSSION ............................................... 119 Characteristics of the Sample .............................. 119 Illness Representation Used by Individuals with Schizophrenia ................................... 121 Summary ............................................. 130 Limitations of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Implications for Further Study ............................. 137 Appendices A. CONSENT FORM ........................................... 139 B. DATA COLLECTION INSTRUMENTS .......................... 142 C. SE:MISTRUCTURED INTERVIEWS: EXAMPLES ................. 148 REFERENCES .................................................... 155 VIi LIST OF TABLES 1. Research Questions, Variables, Method of Measurement, and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 2. Selected Sample Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 3. Psychiatric Assessment Scale (PAS) Mean, SD, and % of Sample Per Selected Item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 4. Illness Perception Questionnaire-Symptom: Mean, SD, and % of Sample Per Selected Item ................................................ 77 5. Early Symptom Scale: Mean, SD, and Percentage of Items Selected ....... 79 6. Comparison of Selection Patterns for Similar Items on IPQ-S and ESS ..... 81 7. Profile of Mood States-Short Form: Internal Consistency Reliability, Mean, and SD for Total Scale and Subscales .............................. 83 8. Daily Hassles Scale -Revised: Mean, SD, and % of Sample Per Selected Item ........................................................ 85 9. Causes of Illness Scores (Mean and SD) ............................. 88 10. Illness Perception Questionnaire - Time Line, Consequences, and ControVCure Scores (Mean & SD) ................................ 89 11. Research Instruments Mean, SD, and Reliability Estimates Scores ......... 91 12. Correlations Among Variables of Self-Regulation . . . . . . . . . . . . . . . . . . ... 107 ACKNOWLEDGMENTS I would like to thank the following individuals who have been instrumental in helping this project to completion. Mae Kramer, for her willingness to chair my committee through the early stages of planning, developing, and launching of the project. B. Lee Walker, for chairing the committee through the final stages of organizing the defense and writing the final edition. Also, for her fierce editing and careful attention to detail. Paula Meek, for her advise and expert assistance with methodology, data analysis, and interpretation. Robert Croyle, for his consultation and expertise in the area of illness cognition and self-regulation of illness. Lillian Nail, for her oversight of the project in areas of writing, methodology, data analysis and interpretation, and presentation of results. To the staffat Community Nursing Service (Catherine Fasselin, Joyce Harris, and Marilyn Park), Valley Mental Health (Valley Plaza and Adult Day Treatment), Weber Human Services (Angela Keane and Laura Owen), and Four Comers Mental Health Center for allowing me access to clients in their respective systems of care. To my many colleagues and friends, both locally and nationally, who remained steadfast in the belief that I would complete this project. To my lifetime companion and friend, Nora Goicoechea, who had the patience and wisdom to know how and when to support me in this endeavor. Finally, to the subjects who shared very meaningful and private information about their experiences with a mental disorder. Without their courage and insight this project would not have been possible. INTRODUCTION Schizophrenia is a long-term mental disorder characterized by exacerbations, remissions, and continued symptomatology that result in disability and impaired ability to cope with the everyday demands of society (Bachrach, 1988; Breier, Schreiber, Dyer, & Pickar, 1991; Wiedl & Schottner, 1991). A stress-vulnerability theory of schizophrenia describes it as an enduring disorder with characteristic patterns of long-term illness in which individuals are sensitive to multiple sources of stress and require moderators to reduce symptom exacerbation and decompensation (O'Connor, 1994; Wiedl & Schottner, 1991; Zubin & Spring, 1977). Stressors come from biological alterations associated with the disorder and interactions in the psychological-biological and environmental-interpersonal domains. Moderators or interventions found helpful in achieving the best outcomes include psychoactive medications, social support, and self­regulation techniques (O'Connor, 1994). In order to manage a chronic illness such as schizophrenia, individuals require an understanding of their illness and symptom complexes in order to assess changing symptom patterns and to initiate appropriate behaviors when symptoms exacerbate. This is consistent with self-regulation theory, which aims to understand illness beliefs and resultant behaviors that affect outcomes of illness monitoring. Theoretical models of self-regulation (Carver & Scheier, 1982; Leventhal, Nerenz, & Steele, 1984) have delineated beliefs and processes used by individuals with physical illnesses to monitor and cope with their respective disorders. Structures of illness representation; cognitive, affective, and behavioral responses; and coping outcomes have accounted for differences in compliance, care seeking, and response to changing symptom patterns (Leventhal & Diefenbach, 1991). 2 The self-regulation model is comprised of individual components that form a decision model for dealing with changes in health. A target symptom or change in level of wellness initiates a process of cognitive and emotional problem solving aimed at understanding the significance of the symptom or change in health status (Cameron, Leventhal, & Leventhal, 1993; Carver & Scheier, 1982). Reference to an illness representation associated with the symptom provides a structure for evaluating the significance of the symptom. The illness representation is comprised of symptom, label, cause, timeline, consequences, and controVcure. Determination of the significance of the symptom then determines action necessary to remedy the problem. The outcome of this action is evaluated and modifies the illness representation and subsequent initiatives in similar illness episodes. The study of self-regulation activities used by individuals with schizophrenia has been limited. McCandless-Glimcher and colleagues (1986) and Hamera, Peterson, Handley, Plumlee, and Frank-Ragan (1991) tested Leventhal's (1984) self-regulation model with individuals diagnosed with schizophrenia. They identified symptoms of the disorder and coping mechanisms used, but were unable to correlate these findings with level of functioning. Research also has been conducted that described the individual components of self-regulation such as symptom recognition, prodromal symptoms, insight, and coping methods (Breier & Strauss, 1983; Herz & Lamberti, 1995; McEvoy et at, 1989). These efforts have been useful in delineating components of importance within the self-regulation model, but have not attempted to determine the presence of self-regulation activities used by this popUlation. 3 The study of illness representations and self-regulation in the severely mentally ill represents an attempt to understand issues of noncompliance, how symptoms of relapse are perceived, interaction of the individual with the illness in ongoing self-monitoring, and the use of lay theories and lay support networks to assist with symptom monitoring, coping, and self-regulation (Hamera et al., 1991; Leventhal & Cameron, 1986; McCandless-Glimcher et al., 1986; Strauss, 1989). An important function within psychiatric nursing's scope of practice is helping individuals diagnosed with schizophrenia (and other mental disorders) understand and cope with the nature of their long-term illness. The theoretical framework of self­regulation is useful in the development of interventions designed to help individuals improve coping efforts (Ward, 1993). Specifically, there is a need to understand how illness is conceptualized in order to promote self-regulation and to prevent relapse. This information can then be linked to daily stressors, early symptoms, affective states, and illness beliefs to determine relationships among components of self-regulation activities. The purpose of this study was to determine components of self-regulation used by individuals with schizophrenia. Specifically, the study focused on illness representations, illness symptoms (prodromal versus those currently active), affective states, stressors encountered in daily life, and illness beliefs about schizophrenia. A second purpose was to determine relationships among components of self-regulation. The research questions addressed by this study were the following: 1. Is there a schema or illness representation used by individuals with schizophrenia? 1 a. What is the relationship between symptoms reported as currently being experienced (Illness Perception Questionnaire-Symptoms) by individuals with schizophrenia and those selected as warning symptoms (Early Symptom Scale) of changes in their illness stability? 2. Are there correlations among daily stressors (Daily Hassles Scale), symptoms (Early Symptom Scale and Illness Perception Questionnaire-Symptoms), affective response (profile Of Mood States-Short Form), and illness beliefs (Illness Perception Questionnaire-Beliefs) consistent with self-regulation activities? 4 As background information for the study, literature reviews on schizophrenia and self-regulation follow. Since minimal literature exists on self-regulation activities of individuals with schizophrenia, each topic is covered in a separate chapter in order to acquaint the reader with the theoretical and research findings for the respective topics. CHAPTER I SCmzOPHRENIA This chapter provides an overview of schizophrenia. An historical perspective of schizophrenia describes how the diagnosis has become increasingly refined over the last 80 years. In spite of this increased clarity in symptoms of the disorder, the variability in clinical course and outcomes still relies on theoretical models for explanation. Contemporary theories of schizophrenia have integrated biopsychosocial variables known to influence the process and outcome of the disorder. Two complementary theories of schizophrenia are reviewed: (a) stress-vulnerability (O'Connor, 1994; Zubin & Spring, 1977) and (b) coping with schizophrenia (Wiedl & Schottner, 1991). Together, these theories present a conceptualization of the interplay of stressors, illness­associated dysfunction, and coping efforts in individuals with schizophrenia. Following the review of these theoretical models, specific research on illness course, symptom monitoring, coping responses, and self-regulation is presented. Nosology of Schizophrenia Kraepelin (1971) first described what is now called schizophrenia in 1919. His description of an early onset (praecox) disorder with general impairment in cognitive and behavioral functions (dementia) remains generally accurate (Andreasen & Carpenter, 1993). Bleuler (1950) continued the study of dementia praecox, labeling it 6 the "group of schizophrenias." He believed that the central disorder in schizophrenia was the loosening of associations. Bleuler categorized as the four "A's": (a) (loose) associations, (b) (lack of) affect, ( c) autism, and (d) ambivalence. He believed these to be the pathognomonic symptoms of schizophrenia (Andreasen & Flaum, 1991). Later he included two additional symptoms: (a) avolition and (b) attentional impairment (Andreasen & Carpenter, 1993). The most definitive list of symptoms used for diagnostic purposes were Schneider's (1959) "first rank symptoms," which appeared in the 1980 edition of the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1980). This list shifted diagnostic emphasis from Bleuler's (1950) four "A's" to specific symptoms of delusions and hallucinations. Symptoms of the disorder within this framework included thought withdrawal, thought insertion, thought broadcasting, voices conversing about the patient in the third person, and externally controlled thoughts and movements (American Psychiatric Association, 1980, 1987; Andreasen & Carpenter, 1993). The most recent format for the diagnosis of schizophrenia has categorized symptoms as "positive" and "negative." Only a few "first rank symptoms" remain in the 1994 edition (American Psychiatric Association, 1994) and are classified under positive symptoms. Positive symptoms are divided into two dimensions: (a) psychoticism and (b) disorganization. Symptoms within these dimensions include hallucinations, delusions, bizarre behavior, and formal thought disorder (Andreasen, 1984; Arndt, Andreasen, Flaum, Miller, & Nopoulos, 1995). 7 Andreasen's (1982, 1983) scale for negative symptoms (SANS) has delineated negative symptoms: (a) affective flattening, (b) alogia (poverty of thought and cognition), (c) avolition-apathy, (d) anhedonia-asociality (lack of pleasure-willingness to participate), and (e) inattentiveness. Negative symptoms have become one of the five characteristic symptom clusters (American Psychiatric Association, 1994) because of their impact on functioning (Bouricius, 1989; Eaton, Thara, Federman, Melton, & Liang, 1995). These diagnostic refinements have more narrowly defined and separated schizophrenia from other mental disorders that share similar symptoms during the course of illness such as mood disorders with psychotic features, schizo affective disorders, brief reactive psychosis, and organic psychosis. Other values of positive and negative symptom clusters are their potential for increasing specificity regarding pharmacologic treatment and possible sites of etiology. Positive symptoms respond more quickly to psychotropic agents, whereas negative symptoms take twice the amount of time to respond to medication, yet are the first to emerge when medication is discontinued or fails to control symptomatology (Gel enberg, 1991; Keith & Matthews, 1991). Since the absolute etiology remains to be determined, schizophrenia (or the schizophrenias) remains a syndrome with multiple heterogeneous subforms. With such variability in the course of the syndrome and symptom complexes, Andreasen and Carpenter (1993) proposed three conceptual models of schizophrenia. The first model is a single pathologic process that leads to a variety of manifestations. Their second model implies mUltiple diseases that result in a schizophrenia process via different pathways. 8 The third model suggests that specific symptoms reflect different disease processes that present in various ways in different individuals. These differing disorders may ultimately be distinguished by symptom complexes, illness course, disability, and response to treatment (Andreasen, Arndt, Alliger, Miller, & Flaum, 1995; Marengo, 1994). Although there is increased refinement in the diagnostic specificity for clinicians, these diagnostic symptoms of schizophrenia are not necessarily those noted by individuals as they monitor their sense of illness stability or well-being. Descriptive studies have identified a range of symptoms noted by individuals with schizophrenia during the prodromal phase of illness decompensation. McCandless-Glimcher et al. (1986) grouped symptoms described by individuals with schizophrenia into anxiety (41 %), depression (28%), or psychotic (31 %) categories. Others have noted symptoms of sleep disturbance, distrust, agitation, loss of appetite, restlessness, and hallucinations (Herz, 1985; Herz & Melville, 1980; Heinrichs & Carpenter, 1985; Subotnik & Nuechterlein, 1988). The Early Symptom Scale, developed by Herz and Melville (1980), includes 29 symptoms considered "prodromal" or likely to occur during the early phase of illness decompensation. Symptoms include somatic symptoms (changes in eating and sleeping, development of aches and pains); cognitive difficulties (trouble concentrating and difficulty making decisions); typical psychiatric symptoms (depression, hallucinations, and feeling controlled by others); and intrapersonal problems (being lonely and afraid). The percentage of individuals who respond to symptom onset and, subsequently, initiate coping responses may be fairly low. Herz and Melville (1980) noted that 70% of their subjects identified prodromal symptoms, but only 50% sought 9 mental health advice, and the other 50% did nothing. In addition, the ongoing deficits of schizophrenia such as cognitive, behavioral, and affective (which increase vulnerability) influence coping with the demands of everyday life. Malia (1992) noted these findings in his study of the impact of daily hassles versus life events on symptomatology. He found that daily hassles and perceived stress, but not life events, were correlated with symptoms (as measured by the positive and negative symptom scales). When controlling for symptoms that were active at the beginning of the study, correlations between daily hassles (Kanner, Coyne, Schaefer, & Lazarus, 1981) and eight symptoms were significant. These symptoms were grouped into the categories of somatic, anxiety, and depression. Theoretical Models of Schizophrenia Stress-vulnerability frameworks have been used to understand the nature of schizophrenia from several dimensions (O'Connor, 1994; Zubin & Spring, 1977). These dimensions include identification of problem areas, the need for moderators of stress, and the efficacy of outside interventions (O'Connor, 1994). Wiedl and Schottner's (1991) model focused on areas of stressors, coping efforts, efficiency of coping, and outcomes. Together these two models complement the study of schizophrenia. They also address the domains of illness, stressors, and coping etTorts. See Figure 1 for an integration of the variables from these two models. STRESSORS* OBJECTS OF COPING EFFORTS+ • Neurologic dysfunction* • Disease+ • Psychobiological Stressors* • Self-Personal situation+ • • *O'Connor (1994) EnvironmentallInterpersonal Stressors* Transactions+ +Wiedl and Schottner (1991) APPRAISAL OF STRAIN+ • Objects of coping • Control versus Changeability of straining items MODERATORS * APPRAISAL OF SATISFACTION WITH COPING+ • Stress & Symptom regulation skills • Perceived social support • Anti-psychotic medication Interventions* • Family CATEGORIES OF COPING+ • Peer • Problem vs • Clinician! consultation Non-problem centered • Behavioral, Cognitive, Emotional Figure 1. Integration of Variables from Stress-Vulnerability * and Coping With Schizophrenia+ Models. -.. o 11 Using terminology from O'Connor's (1994) stress-vulnerability model, stressors or vulnerabilities are related to (a) the biological or neurological areas of dysfunction, (b) intrapersonal conflicts, and ( c) environmental or interpersonal stressors. Wiedl and Schottner's (1991) terminology is similar, using terms such as (a) disease, (b) self- or personal conflicts, and (c) transactions with the environment or others. Many of the troublesome problems faced by individuals with schizophrenia are found in the psychological-biological and interpersonal-environmental domains. These problems arise from difficulties with disease management, managing intrapersonal responses to life situations, and conducting transactions within the environment. Whether these problems are primary or secondary to biological deficits remains unclear. Nonetheless, these difficulties pose challenges to problem solving, self-esteem, and competency in everyday life (Amador, Strauss, Yale, & Gorman, 1991; Geen, 1991). Deficits arising from these areas likely come from the long-term nature of schizophrenia that impacts social cognition and identity (Fabrega, 1989), the ability to perceive complex psychological messages (Brothers, 1989), belief structures, and the ability to form interpersonal relationships (Allen, 1990; Rawnsley, 1991). The sense of selfis often impaired by introjection of "schizophrenia" as an identity. Individuals come to see themselves as "schizophrenic" versus an individual with schizophrenia (Strauss, 1989; Strauss & Esteroff, 1989). Benjamin's (1989) findings are consistent with this change in self when symptoms such as hallucinations become integral and accepted by some individuals. For individuals who do not adopt a schizophrenic image or find hallucinations and delusions acceptable, additional explication is necessary to underscore 12 the burden of these symptoms and what may ameliorate their impact on changing states of well ness (Wiedl & Schottner, 1991). The ability to tolerate social stress, carry on social interactions, and process novel information :from the environment is severely compromised by these alterations in cognitive mechanisms (Bellack, Morrison, & Mueser, 1989). Without an adequately functioning information processing system, individuals have difficulty relating to the environment and utilizing feedback necessary for self-regulation (Croyle & Ditto, 1990). Stressors in the Psychological-Biological Domain Within the psychological-biological domain, cognition is the primary basis for evaluation of the perception of complex psychological information. Freedman (1974) described cognitive and perceptual problems experienced by individuals with schizophrenia. Perceptual problems included enhanced or muted sensory awareness; visual changes in clarity, depth perception, and illusions; and the sensation that voices sounded different. Cognitive experiences included increased or decreased mental efficiency; memory loss for word meanings and contexts in time; and loss of control in judgment, reasoning, associations between ideas, thought blocking, and disappearance of ideas. Other authors have identified problems secondary to schizophrenia such as the erosion of insight and distortion of reality (Amador et aI., 1993); deficiencies in the cognitive processing of memory, sustained attention, and executive functions (Eckman et aI., 1992); and increased sensitivity to cultural ~tigmatization of persons with mental illness (Lefley, 1990; RogIer & Cortes, 1993). 13 Green (1996) completed a review of studies on neurocognitive deficits in schizophrenia. His analysis aimed at determining the functional outcomes of neurocognitive deficiencies noted in schizophrenia. Two consistent findings were that secondary verbal memory and vigilance had the greatest impact on functional outcomes. Vigilance here means the ability to maintain an attentional focus on relevant versus irrelevant processes in learning. Secondary verbal memory (long-term memory for word lists or stories) was highly correlated with skill acquisition and community outcome. Vigilance was correlated with social skill acquisition and social problem solving. Owens and Johnstone (1980) noted deficits in the mental, cognitive, neurological, and behavioral capacities of 510 patients hospitalized for many years. These deficiencies likely arose from medication side effects, institutionalization, and the disorder itself Physical symptoms also are related to schizophrenia, and a majority of these result from the side effects of medication such as restlessness, sexual dysfunction, motor rigidity, and blunted thought processes. Other physical symptoms may arise from failure to maintain physical hygiene, monitor changes in physical status, and obtain healthcare in the face of physical symptoms (Breier et al., 1991). Stressors in the Environmental-Intetpersonal Domain In the environmental-intetpersonal domain, information-processing skills are also necessary to maneuver through the complex world of communication and problem solving. Processing environmental cues is dependent on past memory: (a) short term to review immediate circumstances of the situation and (b) long term to recall previous 14 problem-solving attempts and specific skills for the situation (Simon, 1979). Competent use of a response also depends on the belief that the selected intervention will be successful (self-efficacy). Learning, subsequently, is enhanced (or diminished) by the results obtained. Finally, if the individual is to select and then carry out selected steps, there must be a sustained attentional focus in order to proceed with the behavioral actions (Green, 1996). Deficits in the attainment or preservation of interpersonal relationships may • depend on the individual's ability to trust communications and to feel comfortable in interacting with another person. Carpenter and Strauss's (1991) study of 40 patients with schizophrenia noted correlations between employment, social contact, hospitalization, and symptom severity across their II-year follow-up. Anxiety at initial interview (year 1) was positively correlated with time employed. Individuals with stable heterosexual relationships and frequent social contacts had better outcomes in symptom severity, employment, and social contacts. A reanalysis of the findings of Brekke et al. (1993) in 193 individuals with schizophrenia was conducted to determine causal relationships among the variables of symptoms, intrapsychic deficit, friendships, satisfaction, and employment (Romney, 1995). The goal was to find a "goodness offit" model to predict quality of life. Using a LISREL model, data from several research instruments such as the BPRS, Index of Self- Esteem, Satisfaction with Life Scale, Brief Symptom Inventory, Community Adjustment Form, and Quality of Life Scale were evaluated to determine goodness offit for quality of life. Their findings indicated that the best goodness of fit model was similar to one for cardiac patients: adjusted goodness offit index = 0.975,11 = 0.50, df= 3, chi square = 2.38, and root-mean square residual = 0.029. Symptomatology and social functioning were strongly related, suggesting that symptoms of schizophrenia are a primary deficit and that psychosocial functioning is a secondary deficit. 15 Although there is considerable variation in the outcome of schizophrenia, the disorder has great impact on the cognitive, psychological, and interpersonal functioning of individuals. Since illness can reduce these capacities and increased stress can outstrip the abilities of the individual to cope, there is a need to monitor any illness exacerbation in response to stress or environmental demands. Understanding processes used by individuals to monitor symptom changes will provide direction for subsequent research investigation in how this information is related to various components of self-regulation. Illness Course The peak onset of schizophrenia occurs during late adolescence and into early adulthood, affecting approximately 1 % of the population (incidence) (American Psychiatric Association, 1994; Kessler et al., 1994). The natural course of this disorder may include many remissions and exacerbations of psychotic symptoms. One of the more enduring descriptions of the process of a psychotic decompensation in schizophrenia was provided by Docherty, Van Kammen, Siris, and Marder (1978). They reviewed a number of phenomenological studies describing decompensation and derived a five-stage model with two eventful circumstances (A and B) that lead t'.} decompensation: 16 A. An initiating condition such as stress or excess demand I. Feeling like one is not fitting in ll. Feeling of alienation B. A transition event such as an insoluble impasse ill. Feelings of panic IV. Reduced concentration, confusion, and psychotic disorganization V. Psychotic insight, delusions, and psychotic resolution. The overall estimation of the severity and persistence of schizophrenia has been based on the course of illness, return of functional ability after an illness episode, and response to treatment (Bachrach, 1988; Schinner, Rothbard, Kanter, & Jung, 1990). Three generally accepted criteria are used to determine severity: (a) diagnosis, (b) duration, and (c) disability (Bachrach, 1988). Individuals with severe and persistent schizophrenia are affected by ongoing impairment, limited return to a premorbid functional state, and only a partial or refractory response to treatment. These individuals go on to develop impairments in their ability to maintain interpersonal relationships, carry out activities of daily living, obtain or sustain employment, maintain nonwork activities, and control social behavior (Andreasen & Carpenter, 1993; Schinner et al., 1990). The percentage of individuals who have a downward course has been challenged, however, as longitudinal studies have shown disparate rates for the course of the disorder. Outcome studies of schizophrenia have estimated that 15-25% will be cured at some point and remain symptom free; 30-40% will require repeated hospitalizations, but have symptom-free periods in between with improved functional capacities; and 20% will stay chronically handicapped for the rest of their lives. (de Vries & Delespaul, 1989, p. 235) Longitudinal studies spanning 2 to 25 years have noted improvement in 20% to 50% of cohorts, with a flattening out of disability progression during the early years of 17 the disorder (Breier et al., 1991; Carpenter & Strauss, 1991; Harding, Zubin, & Strauss, 1987c). Breier et al. (1991), in their 12-year follow-up of 58 individuals with schizophrenia, found a 78% relapse rate; those with a concurrent affective disorder had a 24% relapse rate, and 38% attempted suicide. Harding, Brooks, Ashekaga, Strauss, and Breier's (1987a, 1987b) study of individuals released from the Vermont State Hospital noted that most of the stability achieved by these clients was secondary to intense outpatient and home care management. One half to two thirds showed improvement during their 32-year postdischarge period, with the greatest improvement noted at 20 to 25 years after initial diagnosis. As these and other outcome rates suggest, there may be different disorders within the "schizophrenia" complex that have varying trajectories and disabilities and that interventions have made significant differences in outcomes through intense therapy modalities. The failure of antipsychotic medication to suppress symptoms or noncompliance with medication (Weiden & Olfson, 1995) also have been implicated in affecting the course of illness in schizophrenia. In these instances, the protective effects of medication fail, and the individual becomes vulnerable to stressors occurring in the internal 18 (biological) or external (interactive) environments. Relapse rates have been estimated at 3.5% per month for individuals on neuroleptic medication and 11 % per month for those who discontinued medication (Weiden & Olfson, 1995). The cost of relapse and hospital readmission has been estimated at $2 billion during a 2-year, posthospital discharge follow-up study. Loss of neuroleptic effectiveness accounted for 60% of these costs, whereas noncompliance accounted for the remaining 400/0 (Weiden & Olfson, 1995). Over time, the individuals' level of disability may increase and multiple community resources may become necessary to support their tenure in the community (Bachrach, 1988). These community resources help individuals control symptoms, live with their disability, meet the obligations of everyday life, and adapt to social and psychological changes secondary to the illness (Corbin & Strauss, 1991). Many of the symptoms of schizophrenia may be present (in varying degrees) much of the time, including auditory and visual hallucinations, difficulty interacting with others, anergia, lack of motivation, anxiety, and depression. These symptoms can be categorized within the three wlnerabilities of the stress­wlnerability model (O'Connor, 1994; Wiedl & Schottner, 1991). Components of the model and associated symptoms are (a) biological (delusions, hallucinations, and anergia); (b) psychological-biological (altered sense of self and self-esteem, as well as a sense of worthlessness or grandiosity); and (c) interactional-environmental (difficulty interacting with others, stress over daily hassles, and poor coping responses). For some individuals, these symptoms are troublesome; whereas for others, they have become 19 integral parts of their experience and indistinguishable from nonillness states (Benjamin, 1989). Symptom Monitoring Symptom monitoring comprises an evaluation of changes in stable symptoms and an awareness of new symptoms that may require action on the part of individuals. Descriptive research studies on prodromal symptoms or "early warning symptoms" of persons with schizophrenia have been useful in describing the symptoms of concern in this disorder (Costello, 1992; Marder et al., 1991; McCandless-Glimcher et al., 1986). Prodromal signs and symptoms are considered the earliest changes in illness stability and may be predictive of an impending deterioration or relapse (Bustillo, Buchanan, & Carpenter, 1995; DeVries & Delespaul, 1989; Docherty, Van Kammen, Siris, & Marder, 1978; Heinrichs, Cohen, & Carpenter, 1985; Herz, 1985; Herz & Lamberti, 1995; Lieberman, 1995; Lufoff, Liberman, & Nuechterlein, 1986; Subotnik & Nuechterlein, 1988). Whether or not prodromal symptoms actually herald a psychotic relapse or merely represent changes in illness stability remains open to debate and further prospective study (Norman & MalIa, 1995). A prospective study of three treatment groups by Gaebel et al. (1993) investigated rates of sensitivity (true positives) and specificity (true negatives) for prodromal symptoms. Rates of specificity and sensitivity did not prove useful as indicators of impending relapse. When new symptoms are used as indicators of impending relapse, they avoid false negatives in predicting relapse; using changes in 20 existing symptoms has been of some value in avoiding false positive predictions of impending relapse. Changes in illness stability should be considered important indicators that illness stability has changed and that some type of intervention may be necessary to avoid further deterioration (Bustillo et al., 1995). Many of the symptoms noted are typical of the disorder itself such as "psychotic" symptoms of delusions and hallucinations that may be present, but controlled, during periods of illness stability. Other nonpsychotic symptoms such as subtle changes in relationships, thinking processes, ability to solve problems, and the "disorder of relating syndrome" may be the first changes noted that eventually may lead to exacerbation or worsening of delusions and hallucinations, as well as decreased reality testing (Cuesta & Peralta, 1995; Norman & MalIa, 1995). In other words, there may be a symptom change progression that begins with neurovegetative variations such as sleep disturbance, appetite changes, and dysphoria, eventually resulting in psychotic processes such as impaired reality testing, hallucinations, and delusions. O'Connor (1991) reviewed several studies of symptom monitoring during the prodromal phase and found that anxiety and depression were commonly noted by subjects as early prodromal symptoms. McCandless-Glimcher et al. (1986) conducted a study of symptoms noted by individuals. Using a self-regulation model (Leventhal, Nerenz, and Straus, 1982) and a symptom checklist, they found high use of symptom monitoring. The symptoms noted by their participants were nervousness, hypersomnia or insomnia, fatigue, and depression. Herz and Melville (1980) interviewed individuals who had relapsed, asking them 21 if they had noticed changes in cognition, affect, and behavior indicative of increasing illness. Seventy percent of the group noted symptoms, whereas 30% did not. Twenty­nine symptoms were reported. The seven most frequently cited symptoms were tense and nervous (80.4%), anorexia (71.7%), difficulty concentrating (69.6%), insomnia (67.4%), anhedonia (reduced pleasure) (65.2%), restlessness (63%), and decreased memory (63%). These same symptoms were again noted in a later study of 99 outpatients, 80 family members, and 46 inpatients (Herz, 1985). Rates and rankings were somewhat different for these clients. The seven most frequently cited rankings were tense and nervous (70.7%), depression (63.6%), insomnia (61.6%), restlessness (58.6%), trouble concentrating (56.6%), loss of interest (56.5%), and seeing friends less (54.5%). Examples of other categorical symptoms were affective changes, paranoia, anger/aggression, religiosity, and hallucinations. Currently this scale is being used with clients to monitor their well-being or changing symptom patterns (Herz, 1996). Heinrichs and Carpenter (1985) used staff observations of inpatients to conduct a prospective study of prodromal symptoms. Three of the 10 most commonly noted symptoms were related to dysphoria. Other symptoms were hallucinations, suspiciousness, changes in sleep, anxiety, and cognitive inefficiency. These studies indicate a capacity on the part of individuals with schizophrenia to monitor symptoms, label them, and potentially make efforts to seek interventions to cope with changing symptoms. It is less clear what some of the stress-initiated changes are that alter symptom stability and how these also must be included in the coping responses to symptom change and interventions aimed at reducing their longer lasting effects. 22 Coping Responses to Symptom Onset As noted above, descriptive research studies have established a number of prodromal symptoms based on the self-report of persons with schizophrenia (Costello, 1992; Marder et al., 1991; McCandless-Glimcher et al., 1986). Some studies also have described the initiation of coping activities by individuals able to recognize changes in a number of functions and to adopt self-help strategies to cope with these changes (Breier & Strauss, 1983; Heinrichs et al., 1985; Kumar, Thara, & Rajkumar, 1989; Subotnik & Nuechterlein, 1988). Coping mechanisms used by individuals with schizophrenia have included internal dialogue or talking to a close friend or relative (Kumar et al., 1989). McCandless­Glimcher et al. (1986) based their study on interviews with 61 subjects and categorized the coping efforts of 82% of the subjects who identified such efforts into self-treatment, seeking assistance from a mental health professional, or altering their behavior in some way. Other reports of coping mechanisms have come from Herz and Melville (1980). They noted that 70% of their clients identified prodromal symptoms, but only 50% sought mental health advice, and the other 50% did nothing. In their analysis of studies describing coping mechanisms, Hamera et al. (1994) noted coping efforts such as "doing nothing," "get busy," ''withdraw from activities," "seek out a friend," "increase medication," "self-talk," and "drug or alcohol." Personal coping styles used to deal with the early prodromal symptoms noted by Herz and Melville (1980) were studied by Cohen and Berk (1985). Responses included "fighting back" (25.4%), "isolation or time out" (24.4%), "prayer" (4.5%), "street drugs" (1.3%), "social diversion" (5.7%), 23 "medical" (7.8%), "doing nothing (helpless)" (13.1 %), or "doing nothing (acceptance)" (17.5%). Breier and Strauss (1983) studied the coping activities of individuals with psychotic disorders. They noted responses in three areas: (a) self-instruction, (b) reduced involvement in activity, or (c) increased involvement in activity. Wiedl and Schottner (1991) studied how individuals with schizophrenia coped with their symptoms. Coping was defined as "the person's cognitive and behavioral efforts to manage, reduce, minimize, or tolerate the internal and external demands of the person" (p. 526). Areas in which coping efforts were required included symptoms and impairments of the disorder, perceptions of the self and individual situations, and transactions with the environment. Individuals responded differently in how distressed they were by various problems. Some individuals were highly strained by the symptoms and impairments of the disorder, whereas others found perceptual dysfunctions to be more troubling. Regardless of the problem area, Wiedl and Schottner noted that more distressed individuals used fewer cognitive coping methods. The combination of illness and perceptual dysfunction may have interfered with cognitive efforts to plan for ameliorating interventions. Amador et al. (1991) suggested that deficiencies in insight may be secondary to disinhibition of cognitive biases rather than a deficiency of the disorder itself. In other words, cognitive biases are protective mechanisms preventing overwhelming access to the reality of the illness. These protective mechanisms are less effective in depressive disorders, allowing for a greater awareness of depressed mood. The effect of mood on self-appraisal may account for less effective responses to stress. Since affect mediates 24 cognition and behavior, excessive affect reduces the efficiency of self-appraisal (Croyle & Ditto, 1990). Hatfield (1989) analyzed numerous personal accounts of schizophrenia to delineate stress and coping methods. Four general categories of changes resulted from stress: (a) altered perception, (b) cognitive confusion, (c) attentional defects, and (d) impaired identity. Altered perceptions resulted in enhanced or blunted perceptions, distortions in reality, and hallucinations. Symptoms attributed to cognitive confusion included being disoriented; losing control over ideas; memory blanking; and a lag time between hearing a word, recalling its meaning, and coming up with an answer (Freedman, 1974). Attentional deficits and lacking volition over shifts in attention were the most debilitating problems in coping with schizophrenia. The identity of persons with schizophrenia was impaired due to overidentification with the illness and an inability to characterize what was wrong. TIlness attribution affects how individuals view their disorder. Sheehan and Kroll (1990) studied 52 psychiatric inpatients' religious beliefs and practices and the relationship to their illnesses. A majority attributed illness to "health factors" such as rest, diet, sleep, and exercise, whereas 23% believed that it was "sin related." Within the group believing that "sin" was a factor, 42% thought a moral life helped prevent illness, 23% thought illness was directly related to sinful acts, and 19010 believed that their illness was a punishment for sinful acts. Greenfeld, Strauss, Bowers, and Mandelkem (1989) interviewed 21 patients admitted to an inpatient setting for treatment. The focus of inquiry was a description of 25 insight from patients' perspectives. Five dimensions of insight were reported: "Views about symptoms; views about the existence of an illness; speculations about etiology; vulnerability to recurrence; and opinions about the value of treatment" (Greenfeld et al., 1989, p. 245). Patients who had reported psychotic symptoms were able to provide a causal hypothesis for being ill. Great variety was noted, ranging from physical illness, hereditary factors, and conflicts to moral or personal failures (Greenfeld et al., 1989). Illness attribution, symptom monitoring, and coping with changing symptom patterns are essential components in managing a long-term illness. Studies have described many symptoms of change and the responses to or attributions concerning these symptoms; however, there has not been a systematic evaluation of how these activities may exist within a structure of self-regulation. In the next section, application of self-regulation techniques to prodromal symptoms is reviewed. Self-Regulation in Schizophrenia The study of how seriously mentally ill individuals monitor symptoms and take action has used "illness cognition" theory (Leventhal & Diefenbach, 1991; Skelton & Croyle, 1991) only recently to better understand mental representations of illness and resulting self-regulatory behaviors (Hamera et al., 1991; McCandless-Glimcher et al., 1986). Individual components of illness representations have been the focus of several research efforts. Breier and Strauss (1983) studied self-control methods used by individuals with psychotic diSQrders. They noted three phases of self-regulation: (a) awareness of psychotic behavior from self-monitoring, (b) self-evaluation of behavior as 26 danger, and (c) mechanisms of control to reduce the health threat. Hamera et al. (1991) further elaborated on the study of symptoms using the "self­regulation" model developed by Leventhal, Nereoz, and Straus (1982). They asked clients a series of questions that encompassed knowledge of illness recurrence such as how they knew what was happening and what they did in response to the symptoms. A comparison of the general level of functioning (GAS scores) with the type of symptom identified by the subject (categorized according to psychotic, anxiety, or depression) noted that higher functioning individuals had fewer psychotic symptoms, and were more anxious or depressed. Cognitive inefficiency, dysphoria, anorexia, impaired sociability, agitation, sleep disturbances, and anxiety also have been noted as early or prodromal symptoms (Hamera et al., 1994). Coping responses were coded as (a) medical such as taking medication and seeking a therapist or (b) nonmedical such as resting and drug or alcohol use. Individuals used 3.1 coping responses, including "getting busy," "concentrating on activities," "resting/withdrawing," or "drug and alcohol." The investigators also found that types of coping actions used were unrelated to level of functioning. Summary This section reviewed schizophrenia from a number of perspectives. The criteria for diagnosis of the disorder have undergone numerous revisions in symptom complexes aimed at describing a psychotic disorder mutually exclusive from other psychotic syndromes. To the extent this goal is achieved, there may be increased diagnostic sensitivity to schizophrenia, its hallmark symptoms, the predictability of the illness trajectory, best interventions, and outcomes. 27 Nonetheless, any long-term disorder requires adaptation on the part of individuals experiencing the illness for them to achieve control over symptoms and stressors that may cause exacerbation and initiate interventions that reduce further decline. As studies reviewed here have indicated, individuals are capable of symptom identification and monitoring and of making differential efforts to cope with these changing symptom complexes. Clearly, symptom monitoring does not uniformly result in coping efforts that lead to better outcomes because many "do nothing" or take drugs and alcohol in response to symptom escalation or new symptom appearance. A range of important symptoms occur during early changes in illness states. Many studies have categorized symptoms into psychotic, anxious, or affective groups, whereas others have noted constitutional symptoms such as insomnia, anorexia, and fatigue. Some symptoms might onset earlier than others, some may have special meaning to individuals in relation to their illness state, or both might be necessary in order for individuals to evaluate changes in illness stability. Determining if there is a conceptual framework or illness representation within which individuals evaluate their illness has yet to be the focus of research efforts. Symptom identification in the absence of an illness representation framework does not inform individuals or their care providers what meaning symptoms may have for engaging in self-regulation activities (Karoly, 1993). Understanding the "architecture" of relapse, symptom changes, and coping efforts within a self-regulation system provides a 28 framework for applying these principles to a disorder such as schizophrenia (Bustillo et al., 1995). Self-regulatory systems must possess many dimensions such as awareness of signs and symptoms, consequences of illness, attributions about the illness and its signs and symptoms, self-concept formation, and psychological defensiveness (Amador et al., 1991). The ability of individuals to differentiate illness stability from changes indicative of worsening mental health is a fundamental step in self-regulation. In addition to awareness of changing symptom complexes, familiarity with the impact of stressors and affective states also provides information from other domains known to affect the course of schizophrenia. CHAPTER II SELF-REGULATION This chapter reviews self-regulation from the standpoint of theory development, components of the self-regulation model, and research findings. Development of theory in this area has been largely informed by the study of individuals' responses to physical illness. Retrospective and prospective analyses of how physical illness is categorized through cognitive and emotional responses have served as the database for evolving self­regulation theory. A review of theoretical frameworks that have contributed to self­regulation is followed by a discussion of selected components of the model and research in this area. Theoretical Review Components that have become part of self-regulation frameworks evolved from a series of early research endeavors (Leventhal & Nerenz, 1985). Interviews, rating scales, and checklists were used to gather information from patients (subjects). The authors relied on the following basic assumptions about human behavior: "People are active problem-solvers, cognitive processing activities are not directly observable, behavior is episodic in organization, and situational and individual factors influence experience and behavior" (p. 519). 30 Furthennore, Leventhal and Nerenz (1985) determined that structural models depicting problem solving or cognitive processing must contain content relevant to the task. This provided the foundation for the stages of processing: (a) problem representation, (b) action plim, and ( c) appraisal. According to Leventhal and Nerenz, the problem representation was necessary to define the parameters of concern and expectations for resolution. The action plan represented the activities of attaining resolution. Appraisal of the situation occurred at the point of problem identification (need for action) and postcoping efforts to determine their efficacy. Appraisal was also an integral part of Carver and Scheier's (1982) theory of self-regulation. In their model, regular appraisal occurred between the individual's current state and a prototype "wellness" state. Disparity between the two resulted in a plan for achieving correction of health changes. Data collection activities were shaped by specification of the illness under study and methods that would garner the most relevant (and perhaps unexpected) descriptions by subjects (Leventhal & Nerenz, 1985). For each illness, specific symptoms, illness processes, and interventions were the necessary parts of identifying illness representation and self-regulation activities, which became the basis for content to be ascertained from subjects about their illness. In order to prevent bias or search too narrowly for explanations, the process of structuring inquiry and questioning of subjects had to be relatively free of expected outcomes. Based on these research strategies, researchers found that an individual's mental representation for illness( es) was derived from cognitive and affective learning during 31 illness episodes. The perception of symptoms or changes in well ness were incorporated into a "cognitively oriented psychological approach" for considering the cause, effects, and need for action related to symptoms or illness experiences (Skelton & Pennebaker, 1982). The resulting model for self-regulation included perceptual, emotional, cognitive, and behavioral responses to illness (Leventhal, Meyer, & Nerenz, 1980). Building upon these early efforts at defining self-regulation and symptom monitoring frameworks, illness cognition theory refined the activities and research directions for this body of knowledge. The goals of illness cognition research, according to Skelton and Croyle (1991) are to define the structure of lay illness representations; deployment of structures in everyday life; the relationship of representational structures and processes in the domain of illness to other cognitive structures; and how to modify the undesirable effects of inadequate mental representation structures. (p. 2) Relevant concepts from illness cognition include subjectively perceived symptoms and attributions regarding the cause of symptoms such as information in the social environment and past learning experiences (Skelton & Croyle, 1991). The value of a cognitive approach to illness representation is that it is "an iterative process by which the person integrates both internal and external stimulus information with existing cognitive structures to give meaning to the person's experience" (Bishop, 1991, p. 33). Lacroix (1991) differentiated illness from disease for the purposes of clarifying respective components of illness schemata. Disease is narrowly defined as the signs or symptoms of the disorder, largely serving a medical function. lllness is a more broad concept inclusive of the individual's experience with the disease. The individual's 32 experience of the illness episode includes symptom recognition and the impact of the illness. Impact of the illness includes behavioral consequences such as compliance, inability to perform role functions, and treatment options. These components become the content of the schema or illness representation. Individuals' schema becomes the "link between disease and illness" for purposes of monitoring symptom onset or changes (Lacroix, 1991, p. 198). These concepts and their interactions have been represented in a "self-regulating processing system" designed to understand how various processes affect a person's health care choices (Leventhal, Nereoz, & Steele, 1984, p. 221). Self-Regulation Model The process of self-regulation (as applied to care seeking) is initiated by the appraisal of a target symptom that leads to operations aimed at constructing or retrieving an appropriate illness representation (Cameron et al., 1993). At the next stage, coping procedures begin within cognitive and emotional structures of the individual in which symptom meaning is evaluated. Emotional states are further modified by life stressors that have a direct effect on coping procedures. Outcome appraisals (determined by how well self-regulation has proceeded) feedback onto coping procedures, the illness representation, emotional state, and life stressors in order to modify one or more aspects of the model. If a determination is made that care-seeking should be undertaken, then the model proceeds in a linear direction. The model by Leventhal et al. (1984) included a process of evaluation and feedback from coping procedures to care-seeking. These feedback mechanisms are 33 necessary to inform the illness representation and self-regulatory processes (Cameron et al., 1993). Life stressors, in the physical illness self-regulatory model, may contribute to an evaluation of the seriousness of the symptoms or need for immediate care. This factor is not representative of stress related to health threats or symptom distress (see Figure 2). As noted, the Leventhal et al. (1984) model of self-regulating processes is comprised of several operational components based on underlying assumptions of how the processing of information occurs. Assumptions of the model include an active processing that takes in new experiences and understands them based on previous learning experiences of different natures and a parallel processing that occurs in the cognitive and emotional domains of coping responses. Each path (cognitive and emotional) is engaged in coping with a response (objective and rational or emotional and preconscious). Stages of processing involve "representation of the problem and the emotion accompanying it [and] then development and execution of response plans for coping with the problem and emotion" (Leventhal et al., 1984 p. 220). A third stage is one of evaluation or appraisal to determine how the regulatory process is functioning. Should earlier stages require revision of plans or coping responses the system allows for this. Each successful adaptation to a health threat, coping with it, and having a successful outcome modifies the illness representation for the health threat. Last, hierarchical processing represents the abstract, concrete, and conceptual operations involved in self-regulation. At the most basic level, stimuli are perceived concretely. Abstraction of the stimulus involves intermediate or perceptual memory TARGET SYMPTOM ILLNESS REPRESENTATION • Symptoms & label • • Cause Time line • Consequences • ControVCure COPING PROCEDURES • Illness related • Affect EMOTIONAL STATES • Negative mood • Ruminations • Depression I LIFE S~SSORS I I SEEK I CARE OUTCOME APPRAISALS • Well • Stressed • Sick Note. "Symptom Representations and Affect as Determinants of Care Seeking in a Community-Dwelling, Adult Sample Population" by L. Cameron, E. A. Leventhal, and H. Leventhal, 1993, Health Psychology, 12, p. 172. Copyright 1993 by Health Psychology. Reprinted by permission. Figure 2. Variables of Self-Regulation Model. I"H ~ processes. The conceptual level requires the use of language and represents the most flexible, dynamic processes in which information can be rearranged and environments appraised in order to anticipate future outcomes. Representation of the illness, coping responses, and mechanisms for the evaluation of coping outcomes comprises the negative feedback/self-regulatory system for the control of danger (health threats). Illness Representation An essential feature of the self-regulation model is the illness representation. 35 This term comprises six factors that are used to evaluate the target symptom: (a) concrete symptom experience, (b) application of a label to facilitate identification and beliefs, (c) determination of immediate and long-term consequences of the problem, (d) setting a temporal course and attributions concerning the symptom( s), ( e) cause of the problem, and (f) means of effecting a cure (Lau & Hartman, 1980; Leventhal et aI., 1980, 1984; Skelton & Croyle, 1991). Leventhal et aI. (1984) described major sources of information used in illness representations. General information is available in the culture that provides linguistic meaning to symptoms and illness. Language creates novel expressions of illness that may not correspond to medical facts, but have meaning for the individual. Contributing factors that may inform the illness representation include appraisal of symptoms (health threat), interaction with stored memories of previous illness experiences, emotional reactions, and social factors. These factors were derived from "lay illness theories" that describe the stages individuals experience during an illness experience. The illness representation was comprised of these factors and served as a guide to coping with illness, treatment, and life problems (Leventhal et al., 1980). 36 The symmetrical relationship of body sensations with disease labels is at the core of these processes (Leventhal et al., 1980, 1984). People seek labels to define their body status. In the presence of symptoms, a label or diagnosis is applied (Baumann, Cameron, Zimmerman, & Leventhal, 1989; Croyle, 1992). These labels may originate from previous experiences such as a cold or the flu or from social knowledge acquired in any number of experiences such as the news media, illness of friend of relative, or an illness prototype (Bishop, 1991). In the event of "silent" diseases such as hypertension, people define symptoms that they believe are associated with the label; that is, they reverse the process (Croyle & Barger, 1992; Leventhal et al., 1980, 1984). TIlness representation is coded in three mechanisms: (a) sensory motor, (b) perceptual memory, and (c) semantic and conceptual (Croyle & Barger, 1992). The perceptions, structure, and meaning of the representation are based on the specific symptoms, the emotional reaction, external events at the time, and the overall illness episode. Semantic representation occurs immediately by organizing the experience and any commonality according to previous experiences. Abstract meaning is developed and then preserved as the totality of the illness experience is concluded. Appraisal In the appraisal process, responses occur "along two parallel and semiindependent pathways" (Croyle, 1992, p. 176). An emotional response to the health 37 threat and a cognitive response to the details of the threat are present. In the self­regulation model proposed by Leventhal et al. (1984), feedback loops can be found from the cognitive and emotional "tracts" in which coping responses from the respective systems can interact. Individuals use appraisal in recognizing symptoms or health threats (Croyle, 1992). Second, these appraisals mediate the effects of the threat on emotion, compliance, and service utilization. With awareness of symptoms, a course of action is commenced. The perceived seriousness or unfamiliarity with the health threat predicts the need for consultation with family or lay support members for verification and validation of the individual's assessment of the symptom(s). Delay in seeking help has been found in each of the stages of the model: (a) appraisal, (b) care-seeking, and (c) utilization (Safer, Tharps, Jackson, & Leventhal, 1979). Appraisal delay comprises the time from symptom notice to decision of illness. If there is a weak or inadequate formulation of the health threat or some other reason the individual chooses not to seek help, this will delay the appraisal phase. The time from illness identification until the decision to seek professional help is labeled care-seeking delay. Negative emotions about the illness and treatment could delay illness-related help seeking. This may be related to minimization of the disorder (a form of denial in the context of positive illness confirmation without knowledge of treatment) (Croyle, 1992), stigma of the illness (association with negative attributes for the person) (Croyle & Jemmott, 1991), downward comparison with others (have known others with more serious symptoms) (Croyle, 1992), or imagery of unpleasant or serious interventions 38 (Safer et al., 1979). Utilization delay begins with the decision to seek professional help until the intake interview in a health setting. This stage also is influenced by attribution of the problem, its seriousness, curability, and the best place to receive care (Horwitz, 1987; Safer et al., 1979). When seeking help is considered necessary, members of the lay network often suggest the best intervention or place to receive care. According to Horwitz (1987), help-seeking processes with mental health problems may be influenced by the individual's beliefs and interpretation of symptoms, the presence of somatic symptoms, cultural beliefs about mental illness and care, stigmatization, and symptoms of the illness that make accurate interpretations of "the problem" difficult. From a cultural perspective, common sources of influence include beliefs about mental illness, primary and secondary family group influence, cognitions and meanings about the symptoms from the individual's perspective, and determination about the source of greatest assistance in resolving the issue (Rogier & Cortes, 1993). Selection of specific agencies may be based on a preference for anonymity with the agency such as an emergency room or an existing close relationship with care providers in a familiar setting such as mental health center or primary care practice (Regier, Goldberg, & Taube, 1978). Research Cognitive operations have been delineated through various research efforts and theoretical speculation on the use of prototypes and exemplars in mental representation activity leading to self-regulation (Bishop, 1991; Croyle, 1992; Croyle & Ditto, 1990; 39 Leventhal & Diefenbach, 1991). Bishop (1991) suggested that an illness prototype is useful as a referent for symptoms. This "standard" is used for comparison purposes with health threats. Social comparison provides individuals an opportunity to compare themselves with others who are ill or who seek advice from members of their social and familial networks (Croyle, 1992; Croyle & Barger, 1992; Leventhal et al., 1984). This data search is useful in defining the factors of causation, consequences, time line, interventions, and cure (Leventhal et al., 1984). In addition to the influence of social comparison on health threats, social cognition provides a structural format for processing symptoms. Social cognition is based on attributions, mental simulations, memory, self-concept, and mental control (Schneider, 1991). Attributions have been explored in the areas of explaining relationships, accounting for depression, and determining responsibility. Heuristics, social judgment, and reasoning are the components of mental simulation. Maintenance of the self-concept is accomplished through selective interaction and processing of motivational and cognitive factors, which is a particularly important skill in processing many of the components of self-regulation. Without an ability to utilize the mechanisms of social cognition, reduced resources are found for the development and retrieval of illness representations and learning from past errors and successes. According to social cognition theory, knowledge structures such as schemata, scripts, frames, and world knowledge are used in information processing; whereas mental constructs are used to summarize past experiences and to guide subsequent behaviors (Schneider, 1991). Two examples of categorizations from past memory are prototypes 40 and exemplars. The purpose of prototypes is for the storage of abstract representations, with attributes of a particular thing or person. The utility of prototypes is that it enhances the ability to type or classify new objects in the environment. Exemplars differ; that is, they represent stored experiences or instances of activities and are used for comparative purposes to guide behavior in similar or near similar situations. Personal illness experience is considered to be the most powerful attribution because there is an identical, previous illness experience from which to evaluate the current health threat (Leventhal & Diefenbach, 1991). Past illness experiences provide the most reliable prototypes, whereas "new" symptoms may require an iterative process of data search and hypothesis testing. In the development of a new illness representation, feedback mechanisms are utilized to reformulate appraisal, apply new coping mechanisms, and reduce emotional distress (Cameron, Leventhal, & Leventhal, 1993). These cognitive processes are influenced by past personal learning, social information from others, and an acceptable hypothesis that fits the data in awareness. With the emphasis in hea1thcare on the medical language of diagnosis and cure, less attention is paid to distinguishing the language of individuals' long-term illness experiences, and less sensitivity is given to needs for changing treatment regimens. Thus, individuals have ample room to formulate idiosyncratic explanations of illness experiences that exist over time, but these rarely inform the medical perspective of care providers. Leventhal and Diefenbach (1991) noted additional influences that contribute to the cognitive processing of target symptom appraisal and attribution. Many of these 41 influences are related to the "body schema" or physical self and the degree of "stress" in one's current life at the time of symptom appearance. The following "rules" have been noted as influential in the identification and seriousness of symptoms: (a) The symptoms of stress often get incorporated into the assessment of the seriousness of the target symptom or the "stress-illness rule"; and (b) as individuals age, attributions may be related to normal aging changes or normal sexual changes or the "age-illness rule" (Leventhal & Diefenbach, 1991). Studies of Self-RegulationlIllness Representation In an effort to determine if a "generic illness representation" could be discerned that would cross populations and illnesses, Turk, Rudy, and Salovey (1986) developed a 38-item questionnaire attempting to "operationalize constructs of the illness representation" such as symptoms, label, time line, consequences, cause, and cure. A purposeful sample of 165 subjects (diabetes educators, diabetic patients, and college students) completed the questionnaire. For some, the specific diseases were personally relevant; whereas for others, knowledge, but not personal experience, provided the information. Third, all were asked to relate to a disease such as cancer that was known only through social knowledge. Random samples of35 subjects from each of the groups also were asked to rate flu, diabetes, or cancer. This request increased the sample to 210 cases. Using exploratory and then confirmator:: factor analyses, four factors were identified as "general" dimensions of illness representation. Seriousness referred to the 42 individual's knowledge about the degree of contagion, incurability, chronicity, and requirements for medical care. Personal responsibility measured the amount of oneself involved in the onset of illness and cure. Controllability of the disease was dependent on the ill person and the treatment system. Last, changeability referred to the course of illness and how it might progress in terms of predictability. The final questionnaire included 24 of the original 38 items. These dimensions were considered "self.. involving" and necessary components of illness schemata. Weinman et al. (1996) noted they were criticized for their data analysis procedures and the use of others (students and diabetic educators-nurses) to describe illness representations versus those of patients and their own disease. Nonetheless, the outcomes were seen as consistent with illness representation theory. An example of mental representation processes from a study with specific illness cohorts came from Meyer, Leventhal, and Gutmann (1985). In their study of230 patients with hypertension (and 50 nonhypertensive control subjects), they noted that illness representation occurred at concrete ( symptom specific) and abstract (disease label) levels of thinking. For clients who presented with symptoms and were subsequently diagnosed hypertensive, the belief that symptoms were perceptible remained operative. For another group of clients with hypertension, dropout from treatment was based on an "acute illness model" of hypertension. These individuals believed that they required treatment only when symptomatic and that treatment was intended for a short course. Other clients were classified as "intermittent" users of health care and treatment. 43 These persons continued in treatment, but took medication intermittently, believing the disorder had a cyclic course. Last, others reentered treatment, believing that they had a chronic episodic need for treatment of their illness. The best predictor for compliance with treatment was the belief that medication (or any treatment intervention) had a beneficial effect on the illness or that new symptoms in the presence of high blood pressure caused reevaluation of the ability to tell when blood pressure was elevated. Otherwise individuals applied their own representation for compliance based on a time factor: (a) acute, (b) cyclic, or (c) chronic. The acute model was most often applied by new patients entering treatment who believed that a short course of medication would "cure" the problem. For individuals who have received treatment for cancer, the label and symptomatology of the illness are important determinants of follow-through and emotional response. When obvious symptoms/signs exist or there is knowledge of the potential for metastatic processes to return, emotional reactions to chemotherapy are lower. This is in contrast to persons receiving chemotherapy after tumor evidence has disappeared and the need for continued, unpleasant treatment seems unjustified (Easterling & Leventhal, 1989; Leventhal et al., 1984). Cameron and associates (1993) matched 366 participants in a medical clinic and then compared them on several variables on a day of care-seeking by one of the pair. For the member care-seeking, there was a greater, more highly formulated illness representation (label), the presence of pain, and an evolving notion of greater seriousness in the symptom over time. The increase in daily symptom reports of both members 44 versus the day of care-seeking was only 1.5 symptoms. They concluded that symptom increase or presentation of a new symptom is a necessary, but not sufficient, motivator to seek care. Interference with social factors also proved important in care-seeking, presumably due to the seriousness factor of the symptom(s). In their final analysis of factors (using logistic regression), the following items were significantly related to care­seeking: (a) stress load of predicted care-seeking after controlling for illness representation and coping factors, (b) seriousness of the symptom, (c) symptom duration, (d) social communication from others to seek care, and ( e) appraisal of a coping failure (Cameron et al., 1993). Some of the variance noted in outcome studies of illness representation may be related to self-defense motives that reduce threats through a process termed "minimization" (Croyle, 1992; Croyle & Barger, 1992). These findings were based on experiments with subjects who were given fabricated test results of a nonexistent illness, along with objective components of the health threat. The researchers found that subjects who were given positive results of the test (had the disease) lessened the seriousness of the illness or the reliability of the test. Another mechanism of denial was termed "false consensus effect" in which positive results led subjects to "overestimate the commonality of the illness or its prevalence" (Croyle & Barger, 1992, p. 11). Denial of illness seriousness also was lessened by information on treatment for the health threat. Treatment options were incorporated into the illness representation, which brought additional information to bear on the seriousness factor. 45 Summmy The illness representation component of self-regulation provides individuals with a referent for symptom onset and determination of various elements of the illness episode, and it determines the potential for symptoms to be considered representative of an ongoing chronic illness. Matching symptoms or early warning signs to the mental representation is critical in mobilizing interventions associated with treatment of the disorder in question such as chronic schizophrenia. Investigation of the individual's understanding and coping responses for the specific symptom or illness in question identified the following patterns and trends: 1. The time line for illness treatment such as acute, cyclic, or chronic. 2. The symmetry rule such as asymptomatic illnesses will cause a search for symptoms and a health threat will create a search for a label/diagnosis. 3. Symmetry between beliefs about the illness and expected symptoms. 4. Selection of symptom of most concern. 5. Age-illness rule such as persons 65+ years old attributed gradual onset conditions to aging. 6. Stress-illness rule such as tendency to have more attributions to stress in the context of a stressful situation and vague, undefined symptoms. 7. Use of lay members for infonnation and validation of health threat. 8. The prevalence factor such as more common illnesses are judged less serious and, once diagnosed, the fear factor lessens or reliability of the test diminishes. 46 9. Social comparison factor such as downward comparison with those more ill versus an upward comparison if the only one diagnosed with rare disorder (Baumann et aI., 1993; Croyle, 1992; Croyle & Barger, 1992; Croyle & Jemmott, 1991; Keller, Leventhal, Prohaska, & Leventhal, 1989; Leventhal & Diefenbach, 1991). The study of illness representation and self-regulation in the severely mentally ill is an attempt to understand how these processes and structures may be similar to, or different from, those utilized by persons with symptoms of physical illness. The uniqueness of schizophrenia is that it is a long-term disorder in which the individual is sensitive to multiple sources of stress and requires moderators to reduce symptom exacerbation and decompensation (O'Connor, 1994; Wiedl & Schottner, 1991; Zubin & Spring, 1977). Stressors come from biological alterations associated with the disorder ( symptoms) and interactions in the psychological-biological and environmental­interpersonal domains. Important symptoms of the illness may occur in the processing of emotion and cognition and be worsened by stressors from the psychological and environmental domains, which may subsequently affect the activities of adaptation. Specific issues of concern in this population are noncompliance; how symptoms of relapse are perceived; interaction of the individual with the illness in ongoing self­monitoring; and the use of lay theoties and lay support networks to assist with symptom monitoring, coping, and self-regulation (Hamera et al., 1991; McCandless-Glimcher et al., 1986; Leventhal & Cameron, 1986; Strauss, 1989). According to O'Connor (1994), moderators or interventions found helpful in achieving the best outcomes with these concerns include psychoactive medications, social support, and self-regulation techniques. 47 As the reviews on schizophrenia and illness representation have noted, little systematic study has been conducted to describe self-regulation activities (from a theoretical perspective) in this population. Leventhal and Nerenz (1985) suggested that variables for study based on theoretical frameworks provide a structure for investigating the phenomenon of interest. The stress-vulnerability theory of schizophrenia has delineated factors representative of the disorder, secondary consequences associated with illness, and coping efforts necessary to alter the illness course (0 ' Connor, 1994; Wiedl & Schottner, 1991). Self-regulation theory has provided a decision model of how individuals process changing states of health (Leventhal & Nerenz, 1985). The content of self-regulation frameworks generally includes target symptoms, illness representation (cognitive evaluation), life stressors, affective response, and outcomes of coping. The focus of this study was to determine how individuals with schizophrenia (and related long-term mental disorders) have come to know their illness and (in this process) learned how and what to do when changes occurred in symptoms, affect, behaviors, and thinking. Based on stress-vulnerability theory, attention was focused on symptoms (current and early warning); stressors (intrapersonal, interpersonal, and environmental); and coping efforts (medications, social support, and self-regulation). Self-regulation activities were drawn from the conceptual model of Cameron, Leventhal, and Leventhal (1993) and evaluated for illness schema, life stressors, coping efforts, and evaluation of coping outcomes. Three research questions were used to guide the study: 1. Is there a schema or illness representation used by individuals with schizophrenia? 1 a. What is the relationship between symptoms reported as currently being experienced (Illness Perception Questionnaire-Symptoms) by individuals with schizophrenia and those selected as warning symptoms (Early Symptom Scale) of changes in their illness stability? 2. Are there correlations among daily stressors (Daily Hassles Scale), symptoms (Early Symptom Scale and Illness Perception Questionnaire-Symptoms), affective response (profile Of Mood States-Short Form), and illness beliefs (Illness Perception Questionnaire-Beliefs) consistent with self-regulation activities? 48 CHAPTERll METHODOLOGY This chapter reviews the methodology used to determine the schema or illness representation used by individuals with schizophrenia and the relationships with stressors, symptoms, affect, and level of pathology. Variables from self-regulation and stress-vulnerability frameworks were selected in an attempt to determine if a schema is distinguishable in individuals with schizophrenia. The study design, criteria for subject selection, data collection procedures, instruments for measurement and semistructured interviews, and procedures for data analysis are addressed. Design A cross-sectional, descriptive design was used to answer research questions concerning an illness schema in individuals with schizophrenia. Key study variables included stressors, mood states, illness beliefs, and symptoms of schizophrenia (see Table 1), along with the instruments used for measurement and methods of evaluating data results. Reliability (internal consistency) of the instruments used in this investigation was conducted to determine their trustworthiness with this sample. 'Interviews were used to collect qualitative data in order to obtain multiple measures of variables utilized and to generate additional factors of illness monitoring used by this population. RESEARCH QUESTION Question I: Is there a schema or illness representation used by individuals with schizophrenia? Question IA: What is the relationship between symptoms reported as currently being experienced (Illness Perception Questionnaire .. Symptoms) and those selected as warning symptoms (Early Symptom Scale) of changes in illness stability? Question 2: Are there correlations among: stressors (Daily Hassles Scale), symptoms (Early Symptom Scale; Illness Perception Questionnaire- Symptoms), affective response (Profile of Mood States-Short Form), and illness beliefs Questoinnaire- Beliefs) that are consistent with the ~Self-Regu1ation Model? Table 1 Research Questions, Variables, Method of Measurement, and Analysis VARIABLES METHOD OF MEASUREMENT Illness Representation Illness Perception Symptoms Questionnaire-Beliefs Cause Time Line Consequences Control/Cure Current Symptoms Illness Perception Questionnaire-Symptoms Prodromal Symptoms Early Symptom Scale Interviews Stressors Daily Hassles Scale Symptoms-Current IPQ-S & Prodromal ESS Affective State POMS-SF Illness Beliefs IPQ-B Interviews ANALYSIS Subscale Analysis of IPQ Reliability of IPQ Scores-Items Description of IPQ Subscales Content Analysis Scale Total Correlations Item Comparisons Content Analysis Correltions: IPQ-Reliable Subscales with Sample Variables I i Ul o 51 Criteria for Subject Selection Eligibility criteria for participation in the study included demographic and clinical factors. Males and females, between the ages of 18 years old and 75 years old, were eligible to participate in the study. This age span of subjects was selected to increase sample size and to include subjects in varying stages of their illness. Subjects had to be in active psychiatric treatment and diagnosed with schizophrenia or related disorders such as schizo affective disorder or psychosis not otherwise specified. The length of time for the symptoms of schizophrenia had to be at least 2 years (consistent with DSM­IV criteria). Type of diagnosis was used to obtain a sample of individuals who had experience with an illness (severe and chronic). All subjects had to be able to speak, understand, read, and write English at the equivalent of a sixth-grade level. The ability to speak and understand English was required for the semistructured interviews, and reading English was necessary to complete self-administered questionnaires. Data Collection Procedures Data collection sites included a nonprofit mental health center in rural Utah, two urban mental health centers, and a psychiatric home care agency. This nUluber of sites was needed to recruit an adequate sample size. Institutional Review Board approval was obtained through the University of Utah. Each data collection site completed its own research review, and the State Health Division required prior approval of the research design before respective mental health centers could approve it. 52 Staffwere informed of the criteria for subject eligibility. They subsequently approached clients in person or during meetings, informed them of the study, and invited them to participate. Those who agreed were taken to an office or group room in which an explanation of the study requirements and purpose was given verbally by the investigator. Each subject received a verbal explanation of the study (including an estimate of the amount of time required to complete the respective data collection activities), was told the interview would be tape recorded, informed of financial remuneration in the amount of $5 upon completion of data collection, and given an explanation of the written informed consent form. Permission to take breaks such as for a cigarette, use of the bathroom, or a drink of water was given. Permission not to answer questions was affirmed. Written consent was obtained, and subjects were given a copy of the consent form. At the conclusion of data collection, subjects were paid the agreed-upon fee, and they initialed the researcher's copy of the consent form indicating payment (see Appendix A). In order to decrease the influence of data collection activities on the results, random assignment of subjects to the order of data collection activities was conducted. Subjects began with the interview or the questionnaires after demographic information was collected. After completion of the first 30 interviews, this random assignment was dropped, and focused, rather than semistructured interviews, were conducted with the remaining subjects prior to completing the questionnaires. The decision to drop random assignment of data collection order was based on the following experiences with the first 30 subjects. Subjects often continued talking 53 about their illness experiences once demographic information was concluded, regardless of the interview order. Others wanted the interviewer to know about their illness history before proceeding to answer the questionnaires. From the interviewer's perspective, it was easier for subjects to complete the questionnaires once they had discussed their illness history, since this allowed for prompts back to specific language of the subject if there was confusion about items or time frames on the questionnaires. Weinman et al. (1996) noted similar expedences in their interviews with 52 diabetic subjects. Subjects found it easier to complete the !PQ subscales following an interview about their illness experiences. The authors speculated that this may have been due to a "priming effect" which elicited information about the illness schemata, making it easier to focus on beliefs during the questionnaire part of data collection. Differences between the focused and semistructured interviews were largely in the amount of further inquiry performed within the interview. The initial semistructured interviews used more open-ended prompts to allow for further information that the subject might present. After 30 interviews, the redundancy of information allowed development of a more focused interview format. This process allowed for more efficient data collection, while using the essence of interview questions that seemed most productive in eliciting subjects' descriptions of their illness experiences and self­regulation activities. Demographic information and illness history were collected on each subject (see Appendix B). The Demographic and Dlness History Form included age, race, marital status, religion, living arrangement, psychiatric diagnosis, length of treatment history, 54 and number of hospitalizations. Current treatment interventions, including pharmacotherapy and psychotherapy, were recorded. Information was self-reported by subjects; medical records were not examined. Records were not examined because compilation of historical data is rarely reliable, especially for individuals with many years of treatment or those treated in other locales. The interview focused on each subject's illness history, portions of which included hospitalizations, medication, and age at first episode. The semistructured interview served as a guide for obtaining additional information related to illness history, symptom monitoring, and illness representation (see Appendix C). Information was elicited from a chronological perspective about each subject's mental health problems. In this way, an attempt was made to understand how individuals had come to learn about their illness, specific symptoms, and the type of treatment received over the years. A second interviewer completed data collection on 10 subjects who had focused interviews prior to completing the questionnaires. This individual had 20 years of advanced psychiatric nursing experience, was versed in the study focus, and had listened to completed interviews. Two complete data collection sessions were attended with the researcher prior to commencing data collection alone. For the questionnaire part of the study, subjects were given response cards. For each questionnaire, a card with directions for answering the respective questionnaire and the numeric/written choices they could make to the items was provided. If there was any doubt about subjects' ability to read, they were asked to read aloud the content of the 55 card. Explanations about the focus of the questionnaire, the time frame covered, and the range of scores were discussed until the subject and researcher had come to a mutual understanding. Each item or belief statement was read aloud, with subjects selecting a score. Words that were not understood by the subject were briefly explained. The researcher/interviewer recorded subject responses on data sheets. The following order was used in describing the study to subjects and in collecting data: (a) introduction of study, (b) use of tape recording, (c) need for breaks, (d) financial remuneration, (e) sign informed consent, (f) order of activities, (g) demographic information, (h) random assignment of subjects to begin with questionnaires or in-depth interview [30 subjects], and (n payment of subject. Measurement Methods Questionnaires Psychiatric Assessment Scale. In order to establish a rating of illness severity at the time of the study, subjects were assessed using the Psychiatric Assessment Scale (Krawiecka, Goldberg, & Vaughn, 1977). This scale is brief, yet gives a simple, clinical rating of chronically mentally ill individuals across four psychiatric symptoms, four behavioral obseIVations, and five medication side effects. Criteria for rating the severity of the respective items are provided (Krawiecka et al., 1977). The scale was selected for its clinical value in determining illness severity at the time of data collection. It was the only single reviewer, easily administered scale, that measured symptoms in a time frame similar to the Illness Perception Questionnaire and 56 the Profile of Mood States. The simplicity of the scale was a major advantage, given its' ease of administration, limited time commitment, and focus on chronic psychotic symptoms and observer ratings. Since the focus of the study was not on "accuracy" of diagnosis (individuals had known illnesses), it was chosen as a reasonable scale to use. Another advantage was that the PAS evaluates subjects across three different dimensions: (a) major symptoms, (b) behavioral observations, and (c) medication side effects. The total score for the scale is dependent on illness severity across these different dimensions. The scale has 13 items and covers "the past week" as a time period for determining a score for each item. Four items consist of psychiatric symptoms: (a) depression, (b) anxiety, (c) delusions, and (d) hallucinations. Since subjects were not previously known to the interviewer, these symptoms were rated by subjects on a 0- to 4-point scale. In discrepancies between the researcher and subject, the researcher prevailed. For example, when a subject had talked about delusional content, but selected o for that symptom, a rating was given by the researcher. Possible scores for this section of the questionnaire were from 0 to 16. Another four items were "observed" and rated by the researcher. These items covered (a) speech, (b) affect, (c) speaking, and (d) behavior. Scoring was on a 0- to 4- point scale. This section had possible scores from 0 to 16. Five items were medication related and were scored on a 0- to 2-point scale, with possible scores of 0 to 10. Symptoms were indicative of short-term symptoms of medication side effects such as dystonia and long-tenn effects such as akathisia or tremor. Questioning, as well as observations, of subjects resulted in scores assigned by the researcher. 57 High scores on the scale were indicative of severely ill individuals with psychotic symptoms such as delusions and hallucinations who were unable to communicate and had several types of medication side effects. Severe depressive symptoms also accounted for high scores on depression and perhaps delusional categories, with withdrawn behavior and limited medication side effects. Lower scores represented chronic symptoms of psychosis or affective disorders, with an ability to communicate and only minimal medication side effects. Early Symptoms Scale. This scale was used to measure symptom onset or prodromal symptoms. Herz and Melville (1980) and Herz and Lamberti (1995) developed the Early Symptom Scale through interviews with 145 individuals with schizophrenia and 80 family members. Spearman rank-order correlations between two different patient groups (e.g., Buffalo, NY and Atlanta, GA) were r = 0.85 (n = < 0.001) and between family members and patients were r = 0.78 (12 = < 0.001). The 29 symptoms that comprise the Early Symptoms Scale cover cognitive factors such as trouble concentrating and remembering things; psychotic symptoms such as more religious thinking, hearing voices, being laughed at by others, and thoughts of killing self and others; physical changes of insomnia and anorexia; and affective symptoms such as anxiety, feeling tense and nervous, and depression. The Early Symptoms Scale was developed for the context "symptoms experienced when you begin to have problems." Subjects selected the items on the questionnaire if they occurred "when you begin to have problems" and then rated them according to a 4-point severity scale. Scores could range from 0 to 116. 58 Illness Perception Questionnaire. The Illness Perception Questionnaire (Weinman et al., 1996) was designed with five subscales to evaluate the concepts inherent in Leventhal's illness representation model: (a) symptoms, (b) cause, (c) time line, (d) consequences, and (e) control/cure. Belief statements are given for each of the subscales, except symptoms, which are listed separately. The belief statements are rated on a 5-point scale from 1 = strongly disagree to 5 = strongly agree. Cause has 10 belief statements regarding various attributions of illness. Although the cause subscale was rated according to the 5-point rating scale, the items are not assumed to relate together. Therefore internal consistency reliability was not appropriate for this subscale. The remaining subscales contain 3 items for time line, 7 for consequences, and 6 for control/cure. Fourteen symptoms are used in the symptom sub scale (illness Perception Questionnaire), and the items are rated on a 4-point scale from never to all the time. Mean score for the symptom sub scale is calculated according to the number of positive responses greater than zero (Weinman et al., 1996). Minimum and maximum scoring using this method would be 0 to 14. This scoring method was not used in this study since elimination of actual scores (e.g., 1, 2, & 3) would not allow for comparisons of subjects' scores on the Illness Perception Questionnaire-Symptom subscale with other instruments measuring similar dimensions. Therefore, for comparison purposes with other instruments used in the study, internal consistency reliability, mean, and standard 59 deviation were based on all scores (0 to 4). Comparisons made between individual items from scales used item means and standard deviations based on actual scores greater than zero (e.g., a severity score). Reliability has been evaluated by internal consistency (Cronbach's alpha) and test-retest procedures for each subscale, except cause (Weinman et al., 1996). Internal consistency for each subscale was symptoms. 82, time line. 73, consequences .82, and controVcure .73. Test-retest reliability was obtained over 1-, 3-, and 6-month time periods and was highest among the factors of controV cure (.68, .54, and .46) and consequences (.68, .55, and .55). Decreasing values are consistent with a variable that is changing over time. Concurrent validity of the scale was determined using a group of myocardial infarction subjects' responses to the Sickness Impact Profile, recent doctor visits, recovery self-efficacy, self-rated health, likelihood offuture myocardial infarction, and control over heart problems (Weinman et al., 1996). Correlations with the subscales of the Illness Perception Questionnaire noted logical relationships between and among the respective scales. Symptom identity was positively related to the Sickness Impact Profile and recent doctor visits. ControVcure was positively related to the recovery self-efficacy scale. Time line was negatively correlated with control over heart disease and self-rated health, indicating that time would be longer rather than shorter in bringing about positive changes. The consequences scale was positively related to health distress, disability, doctor visits, and the likelihood of a future heart attack. For subjects who continued to feel ill, the sub scales of symptoms and consequences correlated with other scales measuring similar dimensions such as health distress, doctor visits, Sickness Impact Profile, disability, and the likelihood of a heart attack. ControVcure and recovery self­efficacy related to the amount of impact subjects believed they could have over their illness. 60 In testing the questionnaire, Weinman et al. (1996) used it with different medical popUlations such as rheumatoid arthritis, renal problems, asthma, myocardial infarction, chronic fatigue syndrome, chronic pain syndrome, and diabetes to determine discriminant validity of the instrument with different illness groups. Their findings indicated that various illness groups were distinguished by variations in sub scale scores consistent with the character of the respective illness. Chronic fatigue patients had higher illness identity, seriousness consequences, and time line than pain patients. Chronic fatigue patients also had a stronger illness identity and seriousness ratings than diabetic patients. Rheumatoid arthritis patients had greater illness identity than diabetic or pain patients and a greater time line than pain patients. Chronic fatigue patients were more likely to attribute illness to virus or pollution, whereas pain patients blamed themselves for their illness. Predictive validity was evaluated with the myocardial infarction patients between hospitalization and 3- and 6-month time intervals. Time line was positively related to beliefs about future myocardial infarction at 3 and 6 months. ControVcure scores were related to 3- and 6-month ratings of control over heart problems but were negatively related to the likelihood of a subsequent myocardial infarction. Identity was negatively related to self-rated health at 3 months but not at 6 months. Consequences were negatively related to self-rated health at 3 months but were positively related to the perceived likelihood of another myocardial infarction at 6 months. 61 The instrument has not been tested on chronically mentally ill patients. In order to alter the questionnaire for this population, the symptom scale had to be constructed of symptoms known to exist in chronic mental disorders. Fourteen items were selected for the symptom sub scale by the researcher. Seven items were taken :from the top-rated symptoms on the Early Symptoms Scale. Four somatic items were included: (a) headaches, (b) upset stomach, ( c) fatigue, and (d) stiff joints. Three items represented internal emotional states: (a) worry, (b) being afraid, and (c) loneliness. The only adjustment made to the Dlness Perception Questionnaire beliefs was substituting "sin" as a cause in place of "pollution. " Internal reliability evaluation was carried out on the individual subscales, except cause, using Cronbach' s alpha procedures. Profile Of Mood States-Short Form. This scale was developed to measure "transient, distinct mood states" (McNair, Lorr, & Droppleman, 1992). The original Profile Of Mood States had 65 items and was rated on a 5-point scale :from not at all to extremely. Six factors emerged: (a) tension-anxiety, (b) depression-dejection, (c) anger­hostility, (d) fatigue-inertia, (e) vigor-activity, and (f) confusion-bewilderment. A Total Mood Disturbance score is calculated :from the scores for each of the six subscales (McNair et al., 1992). The score is derived :from the addition of all scale means, except vigor, which is subtracted :from the total. Reliability of the Profile of Mood States was established through test-retest and internal consistency methods (McNair et al., 1992). Test-retest reliability was measured 62 in subjects awaiting psychotherapy, with time between repeat tests averaging 20 days. Reliabilities ranged between r=. 65 for vigor to r=. 74 for depression. Variability in scores from time one to time two was considered normal for mood states which are expected to vary over time and circumstance. Internal consistency reliability was studied on two groups of psychiatric outpatient subjects. Scores from the study of male outpatients reported reliabilities (K-R20) of near .90 or above for all six factors. Similar results were obtained for a sample of female outpatients (McNair et aI., 1992). Data from the studies reported above, and other large studies using the long form of the Profile of Mood States were used to develop a 30-item, shortened version of the scale (McNair et al., 1992). The same six subscales are measured by the 30-item scale with internal consistency reliabilities ranging between r=.67 (confusion in a sample of college students) to r=.95 (fatigue in female outpatients). Validity of the Profile Of Mood States has been established through "six factor analytic replications in the development of the Profile Of Mood States ... as evidence of the factorial validity of the six mood factors" (McNair et al., 1992, p. 8). Validity was further tested in clinical drug trials with outpatients using placebo versus an anxiolytic for anxiety reduction and monitoring drug withdrawal in addiction. Changes in mood were evident in drug-treated groups versus those on placebo. Concurrent validity was determined by comparison of the original 65-item form with similar instruments such as the Hopkins Symptom Distress Scale, Manifest Anxiety Scale, and Beck Depression Scale (McNair et al., 1992). Subscales of the Profile Of Mood States were compared with the respective scale for comparison. Correlations 63 ranged from .43 to .86 on the Hopkins Symptom Distress Scale for tension-anxiety and depression-dejection with the Profile Of Mood States. The Beck Depression Scale correlated .61 with the depression subscale of the Profile Of Mood States. The Manifest Anxiety Scale's highest correlation (.80) with the tension subscale was with a sample of outpatients. Daily Hassles Scale-Revised. Stressors found in the environment, the subjects' internal state, health status, and interpersonal relationships were measured by the Daily Hassles Scale (Kanner et al., 1981). The Daily Hassles Scale was developed on the assumption that ( a) hassles are part of the individual's life, (b) specific hassles will have distinct levels of impact depending on how taxing the item is, (c) the experience of multiple simultaneous hassles will increase the stressor experience, and (d) daily stressors may have greater impact on psychological health than life events. Since variation in attention to hassles may change under differing circumstances, the overall perception of hassles and their severity may be more meaningful than specific hassles and their ratings. In contrast to life events, hassles may serve as "event mediators" between life change and health outcomes (Kanner et al., 1981). The original scale was composed of 117 items. These items were selected by the researchers, their staff, and family members (Kanner et al., 1981). This list was mailed to 100 randomly selected participants (normal) monthly. These subjects were selected from a larger sample of individuals who participated in a health research project that had been conducted in the area several years earlier. Subjects were asked to select items they had experienced during the past month and then rate them on a 3-point severity scale (1 = 64 somewhat severe, 2 = moderately severe, and 3 = extremely severe). Scores were generated in (a) a frequency score noting the number of items selected (0 to 117), a cumulated severity score determined by multiplying the number of items by their 3-point severity rating (0 to 351), and an intensity score based on dividing the cumulated severity score by the frequency (0 to 117). Since correlations between cumulated severity scores and intensity scores were high (r = 0.95), only frequency and intensity scores were further calculated. Kanner et al. (1981) conducted monthly correlations on test-retest measures over the 9-month data collection period. They noted differences between frequency and intensity scores. Over the course of the 9 months, the frequency scores remained fairly stable (r = .79), whereas intensity scores decreased significantly (r = 0.48). The authors concluded that the same number (not necessarily the same type) of hassles was consistent, whereas the intensity seemed to drop off. Test-retest correlations (according to gender) found differences between men and women (r = 0.76 for women and r = 0.66 for men). Comparison of the Daily Hassles Scale findings and psychological symptoms reported by the Hopkins Symptom Checklist noted correlations ofr = 0.60 for the total sample, r = 0.55 for men, and r = 0.66 for women. These correlations were considered evidence of construct validity for the scale and its relationship to adaptational outcomes and psychological symptoms. A stepwise regression analysis of the impact of hassles on psychological symptoms noted that daily hassles accounted for a greater amount of variance than life events (Kanner et al., 1981). 65 The Daily Hassles Scale was modified for use in a study with persons with severe mental illness such as schizophrenia (Segal & VanderVoort, 1993a). The authors reduced the number of items from the original 117 to 46. Items eliminated were considered inappropriate for the level of functioning of the population such as work, too many meetings, yard work, and home maintenance. Internal consistency (Cronbach's alpha) of the modified scale was 0.91 (Segal & VanderVoort, 1993b). Use of the modified Daily Hassles Scale in a follow-up study of their original sample noted significant interaction between daily hassles and psychopathology, as well as daily hassles and physical health outcomes. Using regression analysis with psychopathology as the dependent variable, the authors noted a significant (although not strong) relationship with daily hassles (8 = .19, R = .05) and emotional support (8 = -.26, oR < .01) with an adjusted R2 = .07, f = 3.0. For purposes of the present study, an additional seven items were added from the original Daily Hassles Scale to the Segal and VanderVoort (1993b) modification. These items were selected because of their known association with the course of illness and treatment in schizophrenia. The goal was to select specific items for the scale (not selected by Segal and VanderVoort) that might be considered hassles for this population. In the cognitive domain, one added item was "trouble making decisions." The use of nonprescribed substances in the control of symptoms or worsening of symptoms was reflected in the use of alcohol and personal use of drugs. Somatic hassles may arise from medication effects and were reflected in the items "trouble relaxing,'" "concerns about weight," and "not enough personal energy." Last, interpersonal hassles were indicated 66 by "problems with your lover/friend/spouse." These items increased the total to 53. Internal consistency (Cronbach's alpha), frequencies, and means were calculated on this revised scale (Daily Hassles Scale - Revised) and are reported in the results section. Interview A semistructured interview was used to elicit subjects' descriptions of their illness and how they had come to know and experience it over the length of time since they were first diagnosed and first knew they had problems. Based on the theoretical frameworks of stress-wlnerability and self-regulation, questions were constructed to obtain information in these areas. A common focus was how subjects had learned about their illness and what measures they took to maintain wellness or to seek help when their illness/symptoms worsened. The questions designed to elicit the patterns of illness were derived from the domains of the stress-wlnerability model (O'Connor, 1994) and from factors identified in illness representation (Leventhal & Diefenbach, 1991). What subjects did to cope with illness or stressors was based on self-regulation and coping with mental illness models (Cameron, Leventhal, & Leventhal, 1993; Wiedl & Schottner, 1991). These descriptors were used to validate illness representation terms that exist in the literature for physical illness or provide new dimensions of terminology unique to schizophrenia. 67 Data Analysis A sample size of 70 subjects was determined necessary for two purposes. First, the use of 14 new items such as symptoms of schizophrenia for the TIlness Perception Questionnaire required 5 subjects/items for reliability analysis. Second, using a power analysis, a sample size of 70 was needed to achieve a power of .80 with a significance level of. 05 using Cohen's formula for Pearson correlation at r=. 50. Although the Illness Perception Questionnaire was new to this population, data were found to suggest that correlations in the .50 range were possible between scores on the other measurement instruments such as Early Symptoms Scale and Daily Hassles Scale-Revised. Procedures for the data analysis are listed in Table 1. (See page 50.) Reliability for the respective instruments was based on internal consistency (alpha coefficients). Correlations between instruments and the respective subscales of the Illness Perception Questionnaire were computed to examine relationships among the variables of self­regulation. Concurrent validity of symptom scales was examined by computing correlations between the Early Symptoms Scale and the symptom scale of the Illness Perception Questionnaire (although the time frame for measuring symptoms of these two scales differed). Qualitative data were obtained from interviews with subjects. An analysis of the completed interviews was conducted to describe the language, terminology, and processes used by subjects in describing their experiences with mental illness. Specific attention was paid to the variables of interest from the stress-wlnerability (O'Connor, 1994), coping with schizophrenia (Wiedl & Schottner, 1991), and self-regulation frameworks (Cameron et al., 1993) (see Figures 1 and 2). 68 Content was categorized according to illness representation, currently active versus warning symptoms, stressors, moods, coping efforts, and methods of learning how to understand and, subsequently, manage their mental illness used by individuals with schizophrenia. These data were used to triangulate the quantitative findings and to generate differing perspectives associated with self-regulation used by individuals with schizophrenia. Data were coded from questionnaires and entered into a Statistical Package for the Social Sciences, V6.1 (June 24, 1994) program for evaluation. Frequency distributions and mean scores were examined to evaluate the data for errors and adequate variation of each item. All items for the respective scales showed adequate variation had been attained in the data analysis. A significance level of .05 was used for all calculations. Internal consistency reliability was obtained for all questionnaires and subscales, with the exception of the Psychiatric Assessment Scale (which was used for clinical purposes only) and the cause subscale of the IPQ (which contained nominal variables not assumed to related together). CHAPTER IV RESULTS The results of the data analysis are presented in this chapter. All instruments were evaluated in two ways. Internal consistency reliability (Cronbach's alpha), total scale mean, and standard deviation were calculated using all scores, including O. The means and standard deviations for individual items in each scale were computed only on subjects scoring greater than 0 (e.g., having the item in question) so the mean is based on scores greater than 0 (e.g, 1-2, 3, or 4, respectively). The percentage of subjects categorized as greater than 0 also are reported for individual items. Missing data were minimized by data collection procedures. All subjects' responses to items were recorded by the interviewer during data collection. Only one client refused to answer two items on one questionnaire, and these two items were scored as '0' during data entry since there was not confirmation the item was experienced. Demographic and clinical characteristics of the sample are presented first, followed by scale evaluations. Each research question is addressed with results presented from the quantitative and qualitative data analyses. Sample Sixty-nine (69) subjects were recruited from five data collection sites. All but five subjects were receiving treatment in the public sector mental health system. Four 70 subjects received care in their homes and one saw a private psychiatrist for treatment. Potential subjects were identified by staff who had been informed of the eligibility criteria for the study by the researcher. In the presence of the researcher, 15 clients who were approached by staff declined to be in the study. Reasons for declining were not pursued in order to avoid the appearance of pressuring individuals to be subjects or demanding a reason for not participating. The interactions that were observed between staff and clients who declined suggested that reasons might be related to exercising autonomy (e.g., being able to say "No" to someone without recourse), shyness, or unwillingness to talk to a stranger about their illness. There may have been other potential subjects who were approached outside the presence of the researcher, but declined to be in the study. Based on known numbers of potential subjects declining to enter the study, the agreement rate was approximately 82%. All subjects who agreed to be in the study were able to complete the data collection processes. Subjects receiving home care required two appointments to complete their interviews, and one appointment to do the questionnaires. The additional time for home care subjects was due to the longevity of their illness history, difficulty keeping subjects focused, and their appreciation of having an interested person listen to their life history. Total time for interview and questionnaires was approximately 2 hours for the first 30 subjects, 3 hours for home care subjects, and 60- 75 minutes for remaining subjects with more focused interviews. Characteristics of the Subiects Demographic Characteristics 71 Selected characteristics of the sample are listed in Table 2. Gender, race, and marital status are reported along with age, number of hospitalizations and years of treatment. Over half the subjects were female (54%), a majority were Caucasian (87%), and most (85%) were unmarried (e.g., single, divorced, or widowed). The mean age of the subjects was 44.3 years ~=12.94; minimum 18, maximum 74). A majority of subjects listed a religious affiliation (n=57) while the remainder indicated they were not practicing or had no religion. Most subjects were in living arrangements under the auspices of the mental health system. Residential housing accounted for 10%, board and care 42%, while 28% lived in arrangements such as apartments subsidized by the housing authority. Those married or those living with family members (e.g., parents, siblings, or boyfriend/girlfriend) accounted for 9% and 10%, respectively. One subject was living in a shelter at the time of the study. Clinical Characteristics Clinical characteristics were derived from subject report and responses to the Psychiatric Assessment Scale (PAS), the Illness Perception Questionnaire-Symptoms (IPQ-S), Early Symptom Scale (ESS), Profile of Mood States-Short Form (pOMS­SF), and the Daily Hassle Scale-Revised (DHS-R). Beliefs about illness representation were derived from the illness Perception Questionnaire (IPQ}subscales of cause, time line, consequences, and controVcure. 72 Table 2 Selected Sample Characteristics Subject Characteristics Total Sample N=69 %N Gender Female 54% (37) Male 46% (32) Race Caucasian 87% (60) Other 13% (9) Marital Status Single 52% (36) Divorced 30% (21) Widowed 3% (2) Married 15% (10) Mean SD Minimum- Maximum AGE 44.300 12.940 18-74 Hospitalizations 6.140 7.411 0-50 Years of 19.217 10.867 1.0-50 Treatment 73 Eighty-six percent of the subjects reported their diagnosis as schizophrenia or schizo affective disorder. The remaining subjects (14%) were categorized as psychosis-not otherwise specified (NOS). Sixty-eight of the subjects were being treated with psychotropic medication, and all but one were ill enough to require treatment in the rehabilitation portion of the mental health system or home care (psychiatric). The mean number of years of treatment was x= 19.2 <m 10.88, minimum 1, maximum 50), and the mean number of hospitalizations for mental health care was x= 6.15 <m= 7.41, minimum 0, maximum 50). The number of hospitalizations was simply a frequency count, with no determination of the length of time for each hospitalization. There was wide variability in the number of hospitalizations and their length. For example, one elderly subject had 3 hospitalizations that accounted for 35 years of her life, a young man had several hospitalizations during a 6-year period, and another was hospitalized once for 5 years when he was a teenager, but had received outpatient care for the last 30 years. Symptoms The Illness Perception Questionnaire - Symptom sub scale (IPQ-S) and the Psychiatric Assessment Scale (PAS) were used to determine the types and severity of symptoms reported as occurring "within the past week" by subjects in the sample. The Early S