An Outcome in Need of Clarity: Building a Predictive Model of Subjective Quality of Life for Persons With Severe Mental Illness

Living in the Community

Peiying Sarah Chan, Terry Krupa, J. Stuart Lawson, Shirley Eastabrook
PURPOSE. The study purpose was to construct a predictive model of subjective quality of life for persons with severe mental illness living in the community with particular attention to participation in occupations. METHOD. Persons with severe mental illness (N = 154) rated their subjective quality of life. Several measures for each of the following categories of variables were completed: demographics, clinical, social participation, and self-measured well-being. Regression analysis was used to determine the signicant predictors for each category and then to build the predictive model from these signicant variables.

RESULTS. Symptom distress accounted for the most variance (33%) in subjective quality of life, followed
by psychological integration (3%) and physical integration (2%).

CONCLUSIONS. The study suggests that occupational therapists should attend to subjective experience of symptoms to inuence quality of life. Therapists are also in a good position to address their clients sense of belonging to their communities and to enable community participation.
Chan, P. S., Krupa, T., Lawson, J. S., & Eastabrook, S. (2005). An outcome in need of clarity: Building a predictive model of subjective quality of life for persons with severe mental illness living in the community. American Journal of Occupational Therapy, 59, 181190.

Peiying Sarah Chan, MSc, OT Reg (Ont), NRCS, is Doctoral Student, Department of Physiological Sciences, College of Veterinary Medicine, University of Florida, Gainesville, Florida. At the time of the study Ms. Chan was a graduate student in the School of Rehabilitation Therapy at Queens University, Kingston, Ontario, Canada. Terry Krupa, PhD, OT Reg(Ont), is Associate Professor and Chair, Occupational Therapy Program, School of Rehabilitation Therapy, Faculty of Health Sciences, Queens University, Kingston, Ontario, Canada KL7 3N6;.krupat@post.queensu.ca J. Stuart Lawson, PhD, is Professor, Department of Psychiatry, Queens University, Kingston, Ontario, Canada. Shirley Eastabrook, PhD, RN, is Assistant Professor, School of Nursing, Faculty of Health Sciences, Queens University, Kingston, Ontario, Canada.

uality of life has emerged as an important indicator of community adjustment for those with severe mental illness. A response to the notion that avoiding hospitalization is not, by itself, a meaningful measure of successful community adjustment, quality of life is viewed as a construct with the potential to capture important aspects of the community life experience. In the profession of occupational therapy the construct has gained prominence because the experience of participation in occupations is considered to be closely associated with quality of life (Laliberte-Rudman, Yu, Scott, & Pajouhandeh, 2000; Mayers, 2000; Velde, 2001). The quality of life construct has been developed as a complex interaction between multiple dimensions including the illness experience (Achat et al., 1998), an individuals state of physical, psychological, and social well-being (Ferrans & Powers, 1985; Lehman, 1983), satisfaction with life activities and relations and opportunities for autonomy (Keith & Schalock, 1994), and consistency between personal perceptions of the consistency between goals hopes and current situations (Hinds, 1990). Objective conditions of community life, such as those reected in societal standards for nances, housing, friendships, work, and safety are considered integral to quality of life, in that it is assumed that an individuals status in these areas will impact on the sense of well-being in the community (De Souza, 2000). Studies by occupational therapists support the view that both subjective and objective factors are part of the experience of quality of life (Laliberte-Rudman et al., 2000; Mayers, 2000). Although the relationship between objective and subjective aspects of quality of life seems intuitively logical, the nature of this relationship remains somewhat of a puzzle. For example, individuals who appear to be living in objectively poor
181

The American Journal of Occupational Therapy

conditions may still give high ratings to their experience of subjective quality of life. (Fakoury & Priebe, 2002; Ruggeri, Bisof, Fontecedro, & Warner, 2001). This suggests that there may be several mediating inuences on the subjective state of well-being as it is experienced by persons with severe mental illness in the community. It appears that the subjective experience and objective indicators of quality of life are distinct constructs (Ruggeri et al., 2001), both of which are required to provide a complete picture of the life circumstances of service recipients. Subjective quality of life, or an individuals own perception of well-being, happiness or satisfaction, is an important aspect of a client-centred service system. From an occupational therapy perspective, subjective quality of life, along with material life circumstance and activity and role performance, is one of the key dimensions of the quality of life construct. Subjective quality of life suggests that the evaluation of quality of life in occupational therapy should include objective observations of behaviours and life circumstances and self-report measures that attempt to capture the individuals experience of well-being. Previous research has indicated a lack of understanding of the factors that inuence subjective quality of life of individuals with serious mental illness living in the community (Ruggeri et al., 2001; Ruggeri, Gater, Bisof, Barbui, & Tansella, 2002). Prince and Prince (2001) stress that clarifying the construct is essential if evaluations of community services for persons with serious mental illness are to continue to use change in subjective quality of life as evidence of successful intervention. This is an important concern for occupational therapists who expect to see changes in the subjective experience of quality of life as a result of interventions that engage individuals in meaningful activities. Building and testing predictive models of subjective quality of life is one way to develop conceptual clarity. In this paper we present the results of a study in which we examined determinants of subjective sense of quality of life for individuals with severe mental illness who were receiving community-based mental health services. Consistent with the domain of occupational therapy, we were particularly interested in building a predictive model of subjective quality of life that included involvement in everyday occupations. The ndings of the study are discussed in relationship to the construct of subjective quality of life and the implications for occupational therapy practice.

eastern Ontario (Krupa, Eastabrook, & Gerber, 1998). Together these teams serve approximately 370 individuals in two small cities and their adjoining rural areas. ACT has received international attention as an effective community-based program for individuals with severe mental illness who have been heavy users of mental health services. Originally developed in the 1970s, the ACT model was designed as an alternative to hospital-based treatments, enhancing community adjustment by providing continuous and 24-hour, in vivo services for illness management, daily living, work and leisure activities, and crisis management (Stein & Test, 1980). The model has been widely researched and disseminated internationally. The ACT model has been subject to standardization in order to ensure consistency in its implementation. Two standardized delity measurements were used to ensure that services offered by the four teams were consistent with standard ACT practice guidelines: The Index of Fidelity of ACT (McGrew, Bond, Deitzen, & Salyers, 1994) and the Critical Components of ACT Interview (McGrew & Bond, 1995). The larger study received university research ethics approval and informed consent was obtained from all participants. The data of 154 randomly recruited clients from the four ACT teams were available for use in this study. The mean age of the study participants was 43 years (SD = 11) with a range from 17 to 68 years. Eighty-seven percent were presently single, and 78% were living in private houses or apartments in the community. Sixty-three percent had a secondary school education, whereas 28% had some post-secondary education. Schizophrenia (56%), and mood disorders (21%) were the most prevalent diagnoses. The mean age of the clients at rst hospitilization was 26 (SD = 9). The mean length of time that the clients participated in ACT programs was 38 months (SD = 33). This demographic prole is consistent with the characteristics of individuals with severe mental illness living in the community and served by ACT. Quality of Life Model We developed a predictive model to capture the multidimensional nature of the subjective quality of life construct. The model was constructed from a review of quality of life research studies in the area of psychiatry, ACT, community mental health, and psychosocial occupational therapy. Specically, we examined these studies to isolate variables that are believed to inuence subjective quality of life. This lead to the identication of 25 potential predictor variables that were than organized into categories to form a model. The model comprised four categories of potential predictor variables:
March/April 2005, Volume 59, Number 2

Method
This study was a secondary analysis of data from a 4-year study examining the processes and outcomes of four Assertive Community Treatment (ACT) teams in South182

(1) Demographics: This category included variables that are social descriptors of the participants and comprised age, gender, marital status, education level, residential status, length of stay in the program, and age at the rst psychiatric hospitalization. (2) Clinical variables: This category described the participants health situation. Clinical variables comprised diagnosis, clinical symptoms, days in hospital, physical health, alcohol and drug use, functional status, and medication compliance. (3) Social participation variables: This category comprised variables related to actual participation in the activities, events and locations in the community. Social participation variables included physical aspects and social aspects of community integration, community resource use, contact with family, contact with friends, and time spent in productive activities. (4) Self-measured variables of well-being: This category referred to the sense of well-being in a variety of life domains, and comprised the participants perception of the availability of social support, sense of personal empowerment, psychological aspects of community integration, and the personal experience of symptom distress. Data Collection Information was collected by trained research assistants using a standardized research protocol. The training included measures of interrater reliability for the instruments. Table 1 summarizes data collection instruments in relation to the proposed predictive model. The independent variable, subjective quality of life, was measured as the average of the two global well-being scores on the Quality of Life Interview (QOLI; Lehman, 1988). The QOLI is widely used in the mental health eld and has proven to have good psychometric properties (Nieuwenhuizen, Schene, Boevink, & Wolf, 1997). Global well-being is rated on a seven point scale ranging from 1 terrible to 7 delighted. Although the QOLI contains several subscales measuring subjective and objective quality of life in eight domains, we selected alternate measurements instruments that were consistent with our conceptual model and provided a more complete assessment of specic variables. For example, the QOLI health domain lacks specicity with respect to symptoms associated with mental illness, and the daily activities domain includes only a small number of activities as evidence of actual participation in the community. Demographics were collected from the ACT health records and interviews with the participants and case managers. Diagnoses were extracted from clinical records.
The American Journal of Occupational Therapy

A modied version of the Canadian Toolkit for Psychosocial Outcomes (Evaluation Centre @ HSRI, 1995; Ontario Federation of Community Mental Health and Addictions Programs, 2000) was used to collect the information about the following domains for the previous 9month period: nature of the type of living situation (private house or apartment, shelter, boarding home, foster home, rooming house, group home or co-op, retirement home, long-term-care facility, correctional facility, specialty hospital, psychiatric hospital, general hospital, chronic care hospital, or on the street); days spent in hospital; physical health problems (1 = no physical health problem, 5 = severe or complete incapacity due to physical health problem); medication compliance (1 = most of the time, 3 = less than half the time); and number of hours spent by participants in any paid or unpaid productive activities. A Community Services and Support Programs Log was developed to measure the number of hours in the previous month during which a participant used community services and support programs, beyond those provided by ACT. The log was designed for use by several models of community mental health service delivery and based on Beecham and Knapps (1992) Consumer Service Receipt Interview and a societal costing instrument developed by
Table 1. Summary of the Quality of Life Model and Measurements
Category Demographics Variables Age, gender, educational level, marital status, length of time in the program, age at rst psychiatric hospitalization, residential status Diagnosis, symptoms, days in hospital, alcohol use, drug use, functional status Measurements ACT records

PSR toolkitResidential log ACT records BPRS-E Hospitalization Log Alcohol Use Scale Drug Use Scale Multinomah Community Ability ScaleAdjustment to Living PSR ToolkitHealth, Education, and Legal Domains PSR ToolkitEmployment Domain Community Resource Use Scale Physical Integration Scale Social Integration Scale

Clinical

Physical health, medication compliance

Social participation

Time spent on productive activities, Community resource use, Physical integration, Social integration, Contact with family, Contact with friends Social support, Empowerment, Symptom distress, Psychological integration

Self-measured well-being

Social Support Scale Empowerment Scale Symptom Distress Scale Psychological Integration Scale

183

Wolff, Helminiak, & Diamond (1995; see Dewa, Horgan, Russell, & Keates, 2001). The Expanded Brief Psychiatric Rating Scale (BPRS-E; Lukoff, Nuechterlein, & Ventura, 1986) was used to rate 24 psychiatric symptoms (1 = not present, 7 = extremely severe). The BPRS is widely used in psychiatric research and has been shown to have good psychometric properties. The Alcohol Use and Drug Use Scales (Drake & Wallach, 1989), were used to measure clinicians ratings of participants substance use in the previous 6 months (1 = abstinent, 5 = dependence with institutionalization). Both the alcohol and drug scale are reported to have good sensitivity and specicity when used by case managers following clients with mental illness in the community (Drake et al., 1990; Drake & Wallach, 1989). The adjustment to living subscale of the Multnomah Community Ability Scale (MCAS; Barker, Barron, McFarland, Bigelow, & Carnahan, 1994) was used to measure clinicians ratings of the functional level of participants, specically their ability to money management, acceptance of illness and independence in daily living. Ratings on the MCAS range from 1 (almost never) to 5 (almost always). The MCAS has good psychometric properties (Barker et al., 1994). Three distinct scales were used to measure different aspects of self-report of community integration as proposed by Aubry and Myner (1996) The Physical Integration Scale (Segal & Aviram, 1978) a 12-item scale was used to measure the frequency of activities (0 = never, 4 = very often) outside the household over the previous month across a variety of settings and locations (e.g., eating in a restaurant, visiting a library, walking in the park). The Social Integration Scale (Aubry, Tefft, & Currie, 1995) was used to measure the frequency and nature of social contacts with neighbors. The scale has 13 items ranging from 1 = never to 5 = frequently. The Psychological Integration Scale (Perkins, Florin, Rich, Wandersman, & Chais, 1990) was used to measure participants beliefs and feelings about their neighborhood and neighbors, including their sense of emotional investment, feelings of inuence, and sense of belonging. The 12 items are rated as yes or no. Cronbachs alpha for internal consistency for the three scales, respectively, are reported to be .73, .87, and .71 (Aubry & Myner, 1996). The total score of the Social Provision Scale (Cutrona & Russell, 1987) was used to measure the participants perceptions of availability of, and level of trust and comfort with, his or her social network. Two individual questions in the Friendship Network Scale (Humphreys & Noke, 1997) were used to measure the participants frequency of contact with their family and friends over the past month.
184

The Symptom Distress Scale was used to gather information regarding the participants experience of being bothered by psychiatric symptoms during the previous 7 days (0 = not at all, 5 = extremely). The scale is a combination of 15 items from the Symptom Checklist (Nguyen, Attkisson, & Stegner, 1983) and 5 items from the anxiety dimension of the SCL-90 (Derogatis & Cleary, 1977). Although psychometric information is not yet available, the SDS is currently in use as a consumer-oriented outcome measure in a number of states (Teague, Ganju, Hornik, Johnaon, & McKinney, 1997; Ohio Department of Mental Health, 1998). Finally, empowerment was measured by a scale that combined Rosenbergs (1965) well-known 10-item SelfEsteem Scale and Rogers, Chamberlin, Ellison, & Creans (1997) power/powerlessness scale, using a 5-point measurement (strongly agree to strongly disagree). The reliability, construct validity, and responsiveness to mental health populations of the Rosenberg scale has been supported empirically (Arns & Linney, 1993; Morse, Calsyn, Allen, Tempelhoff, & Smith, 1992; Van Dongen, 1996). Initial psychometric study of the empowerment scale demonstrated adequate internal consistency and evidence for validity (Rogers et al., 1997). Data Analysis The data were initially analyzed to ensure that the underlying assumptions of regression analysis were met. Frequency distributions for the scores of the dependent variable demonstrated a range of responses and a normal distribution of quality-of-life residuals. Variable ination values were checked and the correlation coefcients among predictor variables were found not to be so high as to suggest collinearity. Finally, the data were checked for non-zero variance in each independent variable and a linear relationship of the predictor variables and the outcome variable was retained. We selected the forced-entry, stepwise regression model. Although we selected variables based on a theoretical framework, the analysis was still exploratory and a theoretically based hierarchical ordering of predictor variables was not possible. The analysis was executed in two phases. In the rst phase, a forced entry regression analysis was conducted for the demographic variables. Demographic variables were forced in because of their potential as modifying factors. Clinical, social participation, and self-measured well-being variables were entered as three separated blocks. The variables within each block were analyzed using the stepwise method in order to select the predictors from each block that signicantly contributed to the regression equation. When entering the variables as a block using stepwise
March/April 2005, Volume 59, Number 2

inclusion, the statistical program develops the most parsimonious model, beginning with the variable with the highest correlations, subsequently entering variables with the highest partial correlations and eliminating those variables that no longer contribute signicantly to the prediction. In the second phase of the analysis the demographic variables were again analyzed as a block using the forced entry method, while the predictors from phase 1 were analyzed as a block using the stepwise method. Those predictors that signicantly contributed to the regression equation were selected as the best predictors of the ACT clients quality of life. This procedure allowed us to examine quality of life from the perspective of our proposed conceptual model while maintaining a reasonable subject-variable ratio throughout the analysis.

Results
Scores on the independent variable showed a wide range of responses to ratings of subjective global well-being. Fortythree percent of the respondents rated they were satised (or more then satised), 28% had mixed feelings, and 29% indicated that they were dissatised (or more) with their lives. Table 2 presents a summary of the results of the phase 1 analysis. None of the demographic variables made a signicant contribution to the regression equation, F (7, 122) = 1.008, ns). The clinical symptom variable as rated by the score on the BPRS-E was the only predictor from the clinical category which signicantly contributed to the regression equation, F (1, 121) = 24.931, p < .05. The incremental variance accounted for by the BPRS score was 16.1%. Physical integration was the only variable from the social

participation category that signicantly contributed to the regression equation, F (1, 120) = 5.157, p < .05. The incremental variance that was accounted for by the physical integration variable was 3.2%. Two variables from the subjective measures of well-being category made a signicant contribution to the regression equation. Symptom distress had a signicant incremental contribution, F (1, 119) = 30.839, p < .05. and accounted for an incremental predictive variance of 15.5%. Psychological integration had a signicant incremental contribution to the equation, F (1, 118) = 4.928, p < .05. The incremental predictive variance of the psychological integration variable was 2.4%. Table 3 presents a summary of the phase 2 analysis. Demographics still did not signicantly contribute to the variance of the outcome variable. Symptom distress was the predictor that contributed most to the nal equation. The correlation coefcient between this model and the outcome variable was .571. In this regression model, the symptom distress variable accounted for 32.6% of the variance of the outcome variable, F (1, 134) = 64.729, p < .05. The psychological integration variable was next strongest signicant predictor. It accounted for an incremental 3.3% of the variance in the outcome variable F (1, 133) = 6.866. Physical integration was the third strongest predictor, F (1, 132) = 5.195 accounting for 2.4% of the variance. In summary, three variables were included in the nal regression equation as the most predictive variables (symptom distress, psychological integration, and physical integration) accounting for, respectively, 32.6%, 3.3%, and 2.4% of the variance. Clinical symptoms that signicantly contributed to the equation in the phase 1 analysis no longer made a signicant contribution to the nal equation. It appears that the predictive capacity of the clinical symptom variable was subsumed under other signicant predictors in this second phase of analysis.

Table 2. Summary of the Results of Phase 1 Analysis (N = 130)

Model 1 2 3 4 5 R2 .055 .216 .248 .403 .427 R 2 change .055 .161 .032 .155 .024 B value .0156 .04415 .0435 .08933 df (7, 122) (1, 121) (1, 120) (1, 119) (1, 118) F change 1.008 24.931 5.157 30.879 4.928 Sig.* .429 (ns) .000* .025* .000* .028*

Table 3. Summary of the Results of Phase 2 Analysis (N = 136)

Final Model Model 1 2 3 R2 .321 .349 .369 R 2 change .326 .033 .024 B value .0495 .09677 .04721 df (1, 134) (1, 133) (1, 132) F change 64.729 6.866 5.195 Sig.* .000* .010* .024*

*Signicant level: p .05 Model 1. Predictors: (constant), age, sex, marital status, educational level, age at the rst psychiatric hospitalization, length of stay, residential status Model 2. Predictors: (constant) age, sex, marital status, educational level, age at the rst psychiatric hospitalization, length of stay, residential status, BPRS Model 3. Predictors: (constant), age, sex, marital status, educational level, age at the rst psychiatric hospitalization, length of stay, residential status, BPRS, physical integration Model 4. Predictors: (constant), age, sex, marital status, educational level, age at the rst psychiatric hospitalization, length of stay, residential status, BPRS, physical integration, symptom distress Model 5. Predictors: (constant), age, sex, marital status, educational level, age at the rst psychiatric hospitalization, length of stay, residential status, BPRS, physical integration, symptom distress, psychological integration

Excluded Variable BPRS T to enter 1.551 df (1,131)

*Signicant level: p .05 Model 1. Predictors: (constant), symptom distress Model 2. Predictors: (constant), symptom distress, psychological integration Model 3. Predictors: (constant), symptom distress, psychological integration, physical integration

The American Journal of Occupational Therapy

185

Discussion
This is an exploratory study of a predictive model of subjective quality of life among persons with severe mental illness living in the community. The study participants were all served by ACT Teams. ACT is designed to meet the needs of individuals who have been high users of intensive health services and caution should be exercised in generalizing these ndings across community mental health. The study has several limitations. As we were unable to access data related to the characteristics of the entire ACT population served by the four teams we cannot assume that the sample is representative. Previous studies completed on this specic sample of ACT clients suggested that their rates of substance use are lower then expected (Eastabrook et al., 2003; Gerber, Krupa, Eastabrook, Gargaro, 2003), and this may account for the poor predictive capacity of this variable. Finally, the study uses a snapshot of subjective quality of life at one point in time and therefore may be biased by factors such as mood, social expectations, and recent events that have been known to inuence subjective quality of life. Future studies that use longitudinal analysis or time sampling strategies of subjective quality of life would serve to reduce this bias and examine the stability of our predictive model. The most powerful predictor of subjective quality of life was symptom distress that accounted for a large portion (33%) of the variance. Studies that examine clinical symptoms in relation to quality of life typically use objective ratings by trained researchers. In our model, the predictive capacity of clinical symptoms as rated by an outside observer was appropriated by the predictive capacity of self-reported symptom distress, suggesting that there is something to be gained by including the personal experience of symptoms. This distinction between objective clinical measures and subjective responses to pathology of disorder is consistent with trends in quality-of-life research related to persistent physical disease (Miller, 2000). Psychiatric symptoms have been found to have a relationship to subjective quality of life for those with serious mental illness, both in studies using quantitative research designs (Evenson & Vieweg, 1998) and in qualitative studies examining the perspectives of recipients of community mental health services (DeSouza, 2000; Laliberte-Rudman et al., 2000). Although this study did not examine the particular symptom or symptom distress patterns associated with quality of life, other studies have suggested that mood symptoms, in particular depression and anxiety, are associated with lower subjective quality-of-life scores (Ruggeri, Gater, Bisof, Barbui, & Tansell, 2002). An occupational therapy study by Goldberg, Brintnell, & Goldberg (2002),
186

found that depression accounted for more of the variance of quality of life then engagement in meaningful activities. In the current study psychological integration emerged as a predictor of subjective quality of life. Although there were no studies found that directly examined the relationship between psychological integration and subjective quality of life, a common perspective in studies of community life with serious mental illness is the extent to which individuals perceive rejection, and are marginalized in the community (Davidson, & Staynor, 1997; Prince & Prince, 2002). The variable psychological integration appears to be similar to the theme of connecting and belonging that emerged in a qualitative study by occupational therapists exploring the perspectives of individuals with schizophrenia regarding factors important to quality of life (LaliberteRudman et al., 2000). Physical integration was the only variable that emerged as a signicant predictor in the social participation category of the quality of life model. Previous occupational therapy research has suggested that participation in activities is related to quality of life (Laliberte-Rudman et al., 2000). These ndings are, however, confounded by different interpretations of meaningful activity, varying from participation in a range of selected activities (Goldberg, Brintell, & Godlberg, 2002), specic activities such as horticulture (Perrins-Margalis, Ruglectic, Schepis, Stepanski, & Walsh, 2000), to participation in valued roles (Eklund, 2001). To date there has been little attention paid to the extent to which actual participation in everyday community settings affects subjective quality of life. This is a remarkable omission in view of published research suggesting that the community lives of individuals with serious mental illness do not demonstrate activity patterns that are consistent with satisfaction and well-being (Krupa, McLean, Eastabrook, Bonham, & Baksh, 2003). Employment and social contacts did not emerge as predictive variables in the social participation category, although they have been identied as important factors by persons with mental illness (Laliberte-Rudman et al., 2000; Mayers, 2000). The physical integration measure used in this study asked participants about their participation in a variety of community settings, including work. This broad conceptualization of community participation may have displaced the predictive power of employment and social contact by including an array of contacts and activities that more comprehensively represents the community lives of persons with serious mental illness. Two of three nal predictors (i.e., symptom distress and psychological integration) were variables of self-measured well-being in the predictive model. This is consistent with previous ndings examining quality of life in severe mental
March/April 2005, Volume 59, Number 2

illness (Roeder-Wanner, Oliver, & Priebe, 1997; Trauer, Duckmanton, & Chiu, 1998). This relationship between subjective quality of life and self-rated well-being may be inuenced by methodology, because relationships might be expected to be higher when common methods (i.e., selfappraisal) are used to measure variables (Nunnally & Bernstein, 1994). Fakhoury & Priebe (2002) demonstrated that different subjective evaluation criteria may not be fully distinct and that they may reect the presence of a single, underlying self-appraisal factor that requires further conceptual development and research. Our predictive model accounted for approximately 38% of the variation in subjective global quality of life. It is likely that important variables were excluded from our quality-of-life model. For example, we did not include any data related to nances. The study sample primarily received xed amounts of disability benets with little variation in total individual income. Studies have shown a startling lack of relationship between economic status, self-appraisals of nancial well-being in the community, and overall subjective ratings of quality of life (Trauer et al., 1998), even though individuals with mental illness have identied these as important factors (Laliberte-Rudman et al., 2000, Mayers, 2000). Conceptualizing economic well-being in a model of subjective quality of life is an important challenge for researchers and service providers. Its absence undermines our ability to recognize that the community lives of individuals with mental illness as characterized by poverty and systemic nancial constraints that are probably deeply embedded in the experience of subjective quality of life (Krupa, Lagarde, & Carmichael, 2003). Satisfaction with services was also not included, although this may influence subjective well-being. Fakhoury & Priebe (2002) argue that any model of subjective quality of life should include neuropsychological evaluation of executive functioning. They speculate that such psychobiological factors may predispose individuals to inated positive measures of subjective experience in response to only small improvements in their life circumstances. Similarly, coping and adaptation have been associated with the experience of subjective quality of life in health-related conditions, perhaps by affecting health outcomes, reframing the health experience, and readjusting expectations (OConnor, 1993). It may be that subjective sense of quality of life is inuenced by context. For example, it has been suggested that individuals with severe mental illness will judge their quality of life relative to changes in life circumstance (Prince & Prince, 2001). If this is the case, then conceptualizing a subjective quality of life in a manner that is useful to community mental health services may involve the development of
The American Journal of Occupational Therapy

multiple predictive models that consider changing life contexts. The participants in this study were largely individuals who had been receiving ACT services in the community for some time and this may help to explain why community integration emerged as positive predictors. It may be, for example, that factors such as residential status will be relatively more important in the context of pending or recent discharge from hospital. Consistent with previous ndings in the occupational therapy and community mental health literature, this study suggests the importance of this multidimensional perspective on subjective quality of life in research and service delivery. The importance of symptom distress as a predictor of subjective quality of life lends support to clinical practices that are both geared to reducing symptoms of mental illness and working directly with the individuals appraisal and experience of symptoms. For example, approaches focused on inuencing personal appraisal of symptoms such as developing acceptable meanings for symptoms, anxiety and stress management techniques and developing behavioural controls to reduce symptoms all capitalize on the clients ability to examine and change their appraisals of experience. The study also suggests the importance of service recipients sense of belonging and integration within their communities. Occupational therapists on community mental health teams such as ACT work in natural community environments and are in a good position to understand and inuence the processes of marginalization that leave a service recipient feeling disconnected. The study suggests that occupational therapists examine to the extent to which individuals experience their activities and efforts inuential within the context of their local communities. For example, therapists may direct interventions toward the persons experience of discrimination or disadvantage, the internalized sense of stigma, or towards the development of occupational participation that is meaningful in the community context. Occupational therapists are also in a good position to attend to the extent to which service recipients are actually engaging in the activities, places, and situations that constitute community life, a factor that appears to be related to subjective quality of life. Although the specic nature of occupational engagement is highly individual, occupational therapy research has highlighted the importance of occupational participation that is characterized by personal meaning, purpose, providing connections with others, and facilitating a sense of personal control as particularly salient in promoting subjective quality of life in the context of severe mental illness (Laliberte-Rudman et al., 2000). Therapists are encouraged to use caution in their application of the model. Although this predictive model contributes to our understanding of subjective quality of life in
187

severe mental illness, a great deal of the variance of the construct was not accounted for by the model. Perhaps one of the most important benets of such predictive models is the promotion of well-developed theoretical frameworks to explain potential relationships between therapeutic interventions and approaches and complex outcomes. It is hoped that the model will serve this function and facilitate theory-based practice and research related to subjective quality of life in occupational therapy.

Conclusions
In this study the predictors that were associated with subjective quality of life of persons with severe mental illness living in the community were symptom distress, psychological integration, and physical integration. The three predictor variables together explained 38.3% of the variance of subjective quality of life. The study contributes to our understanding of construct of subjective quality of life by developing a model that captured the complexity of the construct, including its relationship to aspects of daily occupation. The study suggests that occupational therapists should attend to the clients subjective experience of symptoms, sense of community belonging and community participation to inuence quality of life.L

Acknowledgments
This project was funded by the Ontario Ministry of Health and Long-Term Care in collaboration with the Ontario Mental Health Foundation, the Centre for Addictions and Mental Health, and the Canadian Mental Health Association-Ontario. The authors thank the two anonymous reviewers for their feedback and suggestions.

References
Achat, H., Kawachi, I., Levine, S., Berkey, C., Coakley, E., & Colditz, G. (1998). Social networks, stress and health-related quality of life. Quality of Life Research, 7, 735750. Arns, P., & Linney, J. A. (1993). Work, self and life satisfaction for persons with severe and persistent mental disorders. Psychiatric Rehabilitation Journal, 17, 6379. Aubry, T., & Myner, J. (1996). Community integration and quality of life: A comparison of persons with psychiatric disabilities in housing programs and community residents who are neighbours. Canadian Journal of Community Mental Health, 15(1), 520. Aubry, T., Tefft, B., & Currie, R. (1995). Public attitudes and intentions regarding tenants of community mental health residences who are neighbours. Community Mental Health Journal, 31, 3952.
188

Barker, S., Barron, N., McFarland, B. H., Bigelow, D. A., & Carnahan, T. (1994). A community ability scale for chronically mentally ill consumers: Part II. Applications. Community Mental Health Journal, 30(5), 459472. Beecham, J., & Knapp, M. (1992). Costing psychiatric interventions. In G. Thornicroft, C. R. Brewin, & J. Wing (Eds.), Measuring mental health needs. (pp. 163183). London: Gaskell/The Royal College of Psychiatrists. Cutrona, C. E., & Russell, D. (1987). The provisions of social relationships and adaptation to stress. In W. H. Jones & D. Perlman (Eds.), Advances in personal relationships (Vol. 1, pp. 3768). Greenwich, CT: JAI Press. Davidson, L., & Staynor, D. (1997). Loss, loneliness and the desire for love: Perspectives on the social lives of people with schizophrenia, Psychiatric Rehabilitation Journal, 20(3), 312. Derogatis, L. R., & Cleary, P. A. (1977). Conrmation of the dimensional structure of the SCL-90: A study in construct validation. Journal of Clinical Psychology, 33(4), 981989. DeSouza, H. R. (2000). Use of the Delphi technique and qualitative method to compare the way in which mental health consumers and providers understand the meaning of quality of life. Unpublished masters thesis. Queens University, Kingston, Ontario, Canada. Dewa, C., Horgan, S., Russell, M., & Keates, J. (2001). What? Another form? The Process of measuring and comparing service utilization in a community mental health program model. Evaluation and Program Planning, 24(3), 239247. Drake, R. E., Osher, F. C., Noordsy, D. L., Hurlbut, S. C., Teague, G.. B., & Beaudett, M. S. (1990). Diagnosis of alcohol use disorders in schizophrenia. Schizophrenia Bulletin, 16(1), 5767. Drake, R. E., & Wallach, M. A. (1989). Substance abuse among the chronic mentally ill. Hospital and Community Psychiatry, 40, 10411046. Eastabrook, S., Krupa, T., Horgan, S., Gerber, G., Grant, R., Leder, M., et al. (2003). Prevalence and nature of substance use among clients of four ACT teams in southeastern Ontario. Canadian Journal of Nursing Research, 35(1), 2443. Eklund, M. (2001). Psychiatric patients occupational roles: Changes over time and associations with self-rated quality of life. Swedish Journal of Occupational Therapy, 8(3), 125130. Evaluation Centre @ H.S.R.I. (1995). Toolkit for measuring psychosocial outcomes. Columbia, MD: Human Services Research Institute. Evenson, R. C., & Vieweg, G. W. (1998). Using a quality of life mesure to investigate outcome in outpatient treatment of severely impaired psychiatric clients. Comprehensive Psychiatry, 39, 5762. Fakhoury, W. K. H., & Priebe, S. (2002). Subjective quality of life: Its association with other constructs. International Review of Psychiatry, 14, 219224. Ferrans, C., & Powers, M. (1985). Quality of Life Index: Development and psychometric properties. Advances in Nursing Science, 8, 1524. Gerber, G. J., Krupa, T., Eastabrook, S., & Gargaro, J. (2003). Substance use among persons with serious mental illness in
March/April 2005, Volume 59, Number 2

eastern Ontario. Canadian Journal of Community Mental Health, 22(1), 113128. Goldberg, B., Brintnell, E. S., & Goldberg, J. (2002). The relationship between engagement in meaningful activities and quality of life in persons disabled by mental illness. Occupational Therapy in Mental Health, 18(2), 1744. Hinds, P. S. (1990). Quality of life in children and adolescents with cancer. Seminars in Oncology Nursing, 6, 285291. Humphreys, K., & Noke, J. M. (1997). The inuence of posttreatment mutual help group participation on the friendship networks of substance abuse patients. American Journal of Community Psychology, 25, 116. Keith, K., & Schalock, R. (1994). The measurement of quality of life in adolescence: The quality of student life questionnaire. American Journal of Family Therapy, 22(1), 8387. Krupa, T., Eastabrook, S., & Gerber, G. (1998). Variations in Assertive Community Treatment: A study of approaches and client outcomes of four teams in South Eastern Ontario [Funding proposal]. Queens Universtiy, Kingston, Ontario, Canada. Krupa, T., Lagarde, M., & Carmichael, K. (2003). Transforming sheltered workshops into afrmative businesses: An evaluation of outcomes. Psychiatric Rehabilitation Journal, 26, 359367. Krupa, T., McLean, H., Eastabrook, S., Bonham, A., & Baksh, L. (2003). Daily time use as a measure of community adjustment for persons served by Assertive Community Treatment Teams. American Journal of Occupational Therapy, 57, 558565. Laliberte-Rudman, D., Yu, B., Scott, E., & Pajouhandeh, P. (2000). Exploration of the perspectives of persons with schizophrenia regarding quality of life. American Journal of Occupational Therapy, 54, 137147. Lehman, A. F. (1983). The well-being of chronic mental patients. Archives of General Psychiatry, 40, 369373. Lehman, A. F. (1988). A quality of life interview for the chronically mentally ill. Evaluation and Program Planning, 11, 5162. Lukoff, D., Nuechterlein, K. H., & Ventura, J. (1986). Manual for the Expanded Brief Psychiatric Rating Scale. Schizophrenia Bulletin, 1(2), 594602. Mayers, C. A. (2000). Quality of life: Priorities for people with enduring mental health problems. British Journal of Occupational Therapy, 63(12), 591597. McGrew, J. H., & Bond, G. R. (1995). Critical ingredients of assertive community treatment: Judgments of the experts. Journal of Mental Health Administration, 22, 113125. McGrew, J. H., Bond, G. R., Deitzen, L., & Salyers, M. (1994). Measuring the delity of implementation of a mental health program model. Journal of Consulting and Clinical Psychology, 62(4), 670678. Miller, J. F. (2000). Coping with chronic illness: Overcoming powerlessness (3rd ed.). Philadelphia: F. A. Davis. Morse, G. A., Calsyn, R. J., Allen, G., Tempelhoff, B., & Smith, R. (1992). Experimental comparison of the effects of three treatment programs for homeless mentally ill people. Hospital and Community Psychiatry, 43(10), 10051010. Nguyen, T. D., Attkisson, C. C., & Stegner, B. L. (1983).
The American Journal of Occupational Therapy

Assessment of patient satisfaction: Development and renement of a service evaluation questionnaire. Evaluation and Program Planning, 6, 299313. Nieuwenhuizen, C. V., Schene, A. H., Boevink, W. A., & Wolf, J. R. (1997). Measuring the quality of life of clients with severe mental illness. Psychiatric Rehabilitation Journal, 20(4), 3341. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill. OConnor, R. (1993). Issues in the measurement of health-related quality of life. NHMRC National Centre for Health Program Evaluation, Melbourne, Australia, Retrieved March 24, 2004, from www.rodoconnorassoc.com/issues Ohio Department of Mental Health. (1998). Vital signs: A statewide approach to measuring consumer outcomes in Ohios publicly supported community mental health system. Columbus, OH: Ohio Mental Health Outcomes Task Force. Ontario Federation of Community Mental Health and Addiction Programs. (2000). Toolkit for Measuring Psychosocial Rehabilitation Outcomes. Users Manual. The Canadian Version. Toronto, Ontario, Canada: Ontario Federation of Community Mental Health and Addiction Programs. Perkins, D. D., Florin, P., Rich, R. C., Wandersman, A., & Chavis, D. M. (1990). Participation in the social and physical environment of residential blocks: Crime and community context. American Journal of Community Psychology, 18, 83115. Perrins-Margalis, N., Rugletic, J., Schepis, N., Stepanski, H., & Walsh, M. (2000). The immediate effects of a group-based horticulture experience on the quality of life of persons with chronic mental illness. Occupational Therapy in Mental Health, 16(1), 1532. Prince, P. N., & Prince, C. R. (2001). Subjective quality of life in evaluation of programs for people with serious and persistent mental illness. Clinical Psychological Review, 21(7), 10051036. Prince, P., & Prince, C. (2002). Perceived stigma and community integration among clients of assertive community treatment. Psychiatric Rehabilitation Journal, 25, 323331. Roeder-Waner, U. U., Oliver, J. P., & Priebe, S. (1997). Does quality of life differ in schizophrenic women and men? An empirical study. International Journal of Social Psychiatry, 43(2), 129143. Rogers, E., Chamberlin, J., Ellison, M., & Crean, T. (1997). A client constructed scale to measure empowerment among users of mental health services. Psychiatric Services. 48(8), 10421047. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University. Ruggeri, M., Bisof, G., Fontecedro, L., & Warner, R. (2001). Subjective and objective dimensions of quality of life in psychiatric patients: a factor analytical approach. British Journal of Psychiatry, 178, 268275. Ruggeri, M., Gater, R., Bisof, G., Barbui, C., & Tansella, M. (2002). Determinants of subjective quality of life in patients attending community-based mental health services. The South-Verona Outcome Project 5. Acta Psychiatrica Scandinavia, 105, 131140.
189

Segal, S., & Aviram, V. (1978). The mentally ill in communitybased sheltered care: A study of community care and social integration. New York: Wiley. Stein, L. I., & Test, M. A. (1980). Alternatives to mental hospital treatment, I: Conceptual model, treatment program, and clinical evaluation. Archives of General Psychiatry, 37, 392397. Teague, G. B, Ganju, V., Hornik, J. A., Johnson, J. R., & McKinney, J. (1997). The MHSIP Mental Health Report Card. A consumer-oriented approach to monitoring the quality of mental health plans. Evaluation Review, 21(3), 330341.

Trauer, T., Duckmanton, R. A., & Chiu, E. (1998). A study of the quality of life of the severely mentally ill. International Journal of Social Psychiatry, 44(2), 7991. Van Dongen, C. J. (1996). Quality of life and self-esteem in working and nonworking persons with mental illness. Community Mental Health Journal, 32, 535548. Velde, B. P. (2001). Quality of life issues in community occupational therapy practice. Occupational Therapy in Health Care, 13(3/4), 149155. Wolff, N., Helminiak, T. W., & Diamond, R. J. (1995). Estimated societal costs of assertive community mental health care. Psychiatric Services, 46(9), 898906.

Coming in May/June 2005

L L L L

The Occupation of Caring for Children and Adults Environmental Adaptations and Occupation Professional Practice and Education Strategies Habits and Stroke

Turn to AJOT for the latest information on occupational therapy

Update

treatment modalities, aids and equipment, legal and social issues, education, and research.

190

March/April 2005, Volume 59, Number 2