| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
Swinburne University of Technology, Melbourne, Australia.
Address correspondence to: Simon Knowles, Swinburne University of Technology, John St, Hawthorn, Melbourne, Victoria 3122, Australia. E-mail: sknowles{at}swin.edu.au
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
2 = 67.87, df = 55, p >.05; 2/df = 1.23, Comparative Fit Index = 0.98, Root Mean Square Error of Approximation = 0.04). This study provides an important contribution to a growing field of research that applies theoretical models to investigate the stressors, appraisals, and protective factors that impact caregiver well-being.
Key Words: Parkinson's disease • Caregiver • Burden • Quality of life
Parkinson's disease can be physically and mentally disabling, not only for the patient but also for the caregiver (Aarsland, Larsen, Karlsen, Lim, & Tandberg, 1999
; Caap-Ahlgren, Lena, & Ove, 2002). Furthermore, caregivers absorb a great deal of the indirect costs (e.g., lost wages, in-home support) of this disease (Vitaliano, & Katon, 2006). Consequently, several researchers (i.e., Chappell & Reid, 2002 and Yates, Tennstedt, & Chang, 1999) have attempted to develop and evaluate theoretically based models to explain the relationships between caregiver stressors, protective factors, and caregiver outcomes.
Although many of these models have been validated and applied to a diverse range of caregiving populations (e.g., persons with dementia, elderly persons), very few studies have addressed the relationship between burden and well-being specifically in Parkinson's disease caregivers. Nonetheless, it is known that different caregiving factors are related to unique caregiver groups (Hooker, Manoogian-O'Dell, Monahan, Frazier, & Shifren, 2000; Hooker, Monahan, Bowman, Frazier, & Shifren, 1998; Miller et al., 2001; Thommessen et al., 2002). For example, the cognitive function of individuals with Parkinson's disease or who have suffered from a stroke has been found to adversely impact the burden of the caregivers who care for them. However, only a weak association has been identified in caregivers of individuals with dementia (Thommessen et al., 2002).
One reason for the inconsistent results may be that Parkinson's disease is a progressive neurological disorder with a range of symptoms that can constantly fluctuate on a daily or even hourly basis within the individual (Gunal, Nurichalich, Tuncer, Bekiroglu, & Aktan, 2002). Furthermore, for many people the medication for Parkinson's disease is only effective for approximately 10 years. Following this period of time, the therapeutic benefits do not outweigh the side effects and the medication is discontinued (Marjama-Lyons & William, 2001). Consequently, sufferers live with this disease for the rest of their lives and, as the disease advances, patients become more and more disabled, until eventually very little movement is possible (Brod, Mendelsohn, & Roberts, 1998).
Persons with Parkinson's disease may even be confined to a wheelchair or bed and may unable to speak or feed themselves (Marjama-Lyons & William, 2001). These situations place a high demand on the caregivers of people with Parkinson's disease. Consequently, Parkinson's disease caregivers have been found to have increased depression scores, poor social lives, and generally low quality-of-life scores (Martinez-Martin et al., 2005; Schrag, Hovris, Morley, Quinn, & Jahanshahi, 2006).
One of the limitations of the caregiving models discussed in previous literature (Pearlin, Mullan, Semple, & Skaff, 1990; Lawton, Moss, Kleban, Glicksman, & Rovine, 1991) is that they underemphasize the complexity of relationships between stressors, protective factors, and appraisal variables (Miller et al., 2001; Spruytte, Audenhove, Lammertyn, & Storms, 2002). To progress beyond this limitation, Yates and colleagues (1999) developed a conceptual model based on the stress model by Pearlin, Mullan, Semple, and Skaff and an appraisal model by Lawton, Kleban, Moss, Rovine, and Glicksman (1989). However, the research by Yates' group was limited to only one aspect of caregiver well-being (depression).
In a more recent study, Chappell and Reid (2002) extended the conceptual model of Yates and colleagues (1999) in which caregiver well-being (life satisfaction) is influenced by primary stressors (e.g., functional dependency of the care recipient) and their appraisal (e.g., hours of caregiving), mediators (e.g., perceived social support, self-esteem), and finally a secondary appraisal process (perceived level of burden). Despite the support for Chappell and Reid's model, it should also be noted that there are several limitations.
First, their measure of caregiver well-being was limited to one domain of well-being (life satisfaction). Furthermore, it was a nonspecific measure rather than a measure specifically for caregivers. Another limitation of their research was that it focused on a population of caregivers who care for those with a variety of chronic illnesses, consequently failing to address the specific stressors or factors unique to a particular chronic illness. Finally, their study failed to incorporate the quality of the caregiver–care recipient relationship as a possible protective factor. Several studies have provided evidence for the importance of the caregiver relationship as a protective factor (Beeson, Horton-Deutsch, Farran, & Neundorfer, 2000; Chee & Mancini, 2002; Fried, Bradley, O'Leary, & Byers, 2005).
Our focus in this study was to address the outlined limitations and extend Chappell and Reid's model by investigating the relationships between caregiver stressors (cognitive impairment, behavioral problems, and functional dependency of the care recipient, as well as the caregiver's informal hours of caregiving) and their impact on the burden of Parkinson's disease caregivers. In addition, this study explored the protective role of perceived social support, frequency of breaks, formal service hours, self-esteem, and quality of the caregiver–care recipient relationship on caregiver burden. We attempted to address the limitations of past research by using more specific caregiver measures with strong psychometric properties and investigating the additional protective factor of quality of the caregiver–care recipient relationship on caregiver burden and quality of life. We added this variable on the basis of previous research (Chee & Mancini, 2002; Yates et al., 1999), which found that the quality of the caregiver–care recipient relationship was an important variable one that had yet to be investigated by use of a stress-appraisal model.
Our study unfolded in the following way. First, on the basis of previous research (Aarsland et al., 1999; Chappell & Reid, 2002; Pinquart & Sorensen, 2003b; Thommessen et al., 2002; Yates et al., 1999), we hypothesized that increased caregiver stressors and increased informal hours of caregiving would be adversely related to caregiver burden. In extension to Chappell and Reid's model, we also hypothesized that increases in these variables would be adversely related to caregiver quality of life.
Second, we hypothesized that perceived social support, the frequency of breaks, formal service hours, and self-esteem would act as protective factors ameliorating the adverse effects of caring on caregiver burden (Chappell & Reid, 2002; Chee & Mancini, 2002; Fried et al., 2005; Heru, Ryan, & Iqbal, 2004; Yates et al., 1999). In a further extension to Chappell and Reid's model, we also hypothesized that these protective factors would ameliorate the adverse effects of caring on caregiver quality of life. A further extension to their model was the hypothesis that the quality of the caregiver–care recipient relationship would act as a protective factor ameliorating the adverse effects of caring on caregiver burden and quality of life (Beeson et al., 2000; Chee & Mancini; Fried et al.; Pinquart & Sorensen, 2003b; Yates et al.).
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
The majority of caregivers were married to their care recipient (81.6%), and 88.2% lived in the same house as their care recipient. On average, the caregivers had known their care recipient for 44.23 years and had been caring for them for 8.29 years. Of the caregivers, 64% were not working. However, 22.1% were paid for caregiving. Considering these demographic differences, we included the following caregiver variables in our analyses: gender, age, how far away the care recipient lived, and whether the caregiver was paid.
Measures
Demographic questions were related to the caregiver's age, gender, marital status, length of relationship with the care recipient, work status, and medical conditions. The caregivers were also asked about the length of time they had been caregiving, how long they had known the care recipients, and how close they lived to their care recipients. We provide a brief summary of the scales used in this study in the following paragraphs; however, further information relating to scale to internal reliability, scoring, and interpretation are provided in Table 1.
|
), which we modified to include only the thinking section (14 items). We made further adjustments to make each item more understandable. For example, "Disorientation on place" was accompanied by "Fails to remember places and directions."
We measured care recipient behavioral problems by using the Behavioral Problems Scale (Chappell & Reid, 2002). We measured care recipient functional dependency by using a combination of the scales for activities of daily living (Katz, Ford, Moskowitz, Jackson, & Jaffe, 1963) and instrumental activities of daily living (Lawton & Brody, 1969). Participants assessed whether their care recipient was independent for each of 15 activities.
We also asked caregivers how many hours of formal services they received per week, and how many hours they spent providing informal care per week. For each item, we took the log of the score. We measured social support by using the Perceived Social Support Scale (Blumenthal et al., 1987).
We measured the frequency of caregiver breaks by using the Frequency of Breaks from Caregiving Scale (Chappell & Reid, 2002). Consistent with Chappell and Reid's method, we conducted a factor analysis using a varimax rotation on this scale. For the present study, we produced one main factor (eigenvalue = 5.50) including all items, except for Item 37 ("Going to work or school or volunteering"). A reliability analysis revealed that Item 37 was the least reliable item, and therefore we removed it from the scale. We obtained an overall frequency of breaks score by summing the second component of each item, excluding Item 37 and then dividing the sum by the number of nonzero responses.
We measured caregiver self-esteem by using the 10-item Self-Esteem Scale (Rosenberg, 1965). We measured the quality of the relationship between the caregiver and the care recipient by using the 7-item Relationship Assessment Scale (Hendricks, 1988). We measured caregiver burden by using the 24-item Caregiver Burden Inventory (Novak & Guest, 1989). We had participants assess the statements according to a 5-point Likert scale, from 0 (strongly disagree) to 4 (strongly agree).
We measured caregiver quality of life by using the Scale of Quality of Life of Caregivers (Glozman, Bicheva, & Fedorova, 1998). This scale measures three aspects of the quality of life of Parkinson's disease caregivers: professional activity, social and leisure activities, and responsibilities to help the patient in everyday living.
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
model by using maximum likelihood criterion, and by incorporating the Bollen–Stine bootstrap and standardized estimates (Tabachnick & Fidell, 2001).
Overall, we found the model shown in Figure 1 to be a good fit of the data: 2 = 67.87, df = 55, p >.05; 2/df = 1.23, Comparative Fit Index = 0.98, Incremental Fit Index = 0.98, Tucker–Lewis Index = 0.96, Parsimonious Goodness-of-Fit Index = 0.50, and Root Mean Square Error of Approximation = 0.04. (Consistent with Chappell and Reid's model, we controlled for several caregiver and care receiver characteristics, e.g., age, gender, and how far away the care recipient lives, in the model. These control variables were not our focus, though, so they are not shown in Figure 1. One can obtain a full model including these control variables by contacting S. Knowles). Testing the null hypothesis that the model is correct, we found that the Bollen–Stine bootstrap was not significant (p =.36), which demonstrated that the model is appropriate. The square multiple correlation indicated that 64.2% and 68.9% of the variance in quality of life and burden, respectively, were explained by this model.
|
A number of protective factors had significant direct effects on caregiver quality of life and burden. The quality of the caregiver–care recipient relationship, perceived social support, and the frequency of breaks were negatively associated with caregiver burden. A higher frequency of breaks was also directly associated with better caregiver quality of life. Perceived social support and formal service hours were directly influenced by the cognitive impairment of the care recipient. As this cognitive impairment increased, formal service hours increased and perceived social support decreased. However, perceived social support and formal service hours did not act as mediators between care recipient cognitive impairment and caregiver quality of life. Other protective factors such as the quality of the caregiver–care recipient relationship and the frequency of breaks also affected the relationships between caregiver stressors and caregiver well-being.
We found the quality of the caregiver–care recipient relationship to mediate the relationship between caregiver burden and care recipient behavioral problems. As the number of these behavioral problems decreased, the quality of the caregiver–care recipient relationship increased, which led to a decrease in caregiver burden. The frequency of breaks also acted as a mediator. A decrease in care recipient functional dependency was associated with an increase in the frequency of breaks and led to a decrease in caregiver burden and an increase in caregiver quality of life.
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
model of caregiver well-being was demonstrated. We did find support for our hypotheses. First, we found that caregiver stressors (e.g., behavioral problems of the care recipient) were adversely related to caregiver burden and in turn care recipient quality of life. Second, we found the frequency of breaks and perceived social support to ameliorate the effects of caring on burden and in turn quality of life. Finally, in support of an extension of Chappell and Reid's (2002) model, we found the caregiver–care recipient relationship to ameliorate the effects of caring on burden and quality of life.
In contrast to Chappell and Reid's model, in the current study we found that an increase in care recipient functional dependency adversely affected caregiver burden and quality of life. One possible explanation for this discrepancy is that in the current study we used a sample of Parkinson's disease caregivers, rather than a nonspecific population of caregivers. However, our findings are consistent with other studies exploring functional impairment and caregiver well-being using a sample of Parkinson's disease caregivers (e.g., Martinez-Martin et al., 2005; Thommessen et al., 2002). Therefore, the results may indicate that the negative relationship between care recipient functional dependency and caregiver well-being may be specific to caring for Parkinson's disease patients or those with similar health problems.
Regarding extended components evaluated in this study, the quality of the caregiver–care recipient relationship was directly related to caregiver burden and an increase in care recipient behavioral problems was associated with a lower caregiver–care recipient relationship quality. These results suggest that specific interventions such as support and social groups that involve caregivers and their care recipients could focus on improving the quality of the caregiver–care recipient relationship. Alternatively, programs that provide both the care recipient and the caregiver with advice about how to communicate more effectively and how to better understand the problems associated with the disease may be beneficial.
We also found that, as care recipient functional dependency increased, the frequency of breaks decreased and this adversely impacted caregiver quality of life and burden. Therefore, caregiver breaks may be especially important when the caregiver is caring for those with higher levels of functional dependency, as among Parkinson's disease populations. To further increase caregiver well-being, interventions could focus on increasing the frequency that caregivers receive breaks.
Limitations of this study included the finding that the majority of care recipients (57.4%) had serious medical conditions other than Parkinson's disease. It is likely that at times it was hard to distinguish between the cognitive, behavioral, and functional symptoms that were due to Parkinson's disease and those that resulted from another medical condition. Cultural factors, the economic situation, and other sources of caregiver stress were also uncontrolled in this study. In addition, we acknowledge that there was a relatively small sample size for structural equation modeling. This makes it difficult to generalize the results. However, the model was theoretically based and the results were consistent with previous research based on the stress-appraisal model. Lastly, not all caregivers lived with their care recipients, and therefore the interpretation of some variables used in the model (e.g., frequency of caregiver breaks) raises difficulties. For example, frequency of breaks has a different meaning for spouses or live-in carers than for non-live-in carers, although breaks are likely to be important for all carers. Nevertheless, future studies should explore the efficacy of this model with a larger sample size; they should either use a more homogenous group of caregivers or compare caregivers of different categories, such as spousal and nonspousal caregivers.
In spite of the study limitations, this study forms the basis for future comparative, longitudinal, and qualitative investigations into a model that represents the influence of stressors, appraisals, and protective factors on the well-being of Parkinson's disease caregivers and caregivers of those with other similar illnesses. Consistent with previous research, it was found that increases in care recipient behavioral problems and functional dependency are adversely related to caregiver well-being. Furthermore, protective factors such as the quality of the caregiver–recipient relationship, the frequency of caregiver breaks, and perceived social support can help to ameliorate the adverse effects of caregiving, as demonstrated by their mediating effect in the model.
If extended prospectively, these results have implications for improving caregiver well-being and therefore the quality of care that Parkinson's disease sufferers receive.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication July 9, 2007. Accepted for publication May 23, 2008.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
