| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Jung Hyun Kim, PhD, Department of Rehabilitation and Welfare for the Aged, Yongin University, 470 Samga-dong, Cheoin-gu, Yongin-si, Gyeonggi-do, South Korea 449-714. E-mail: jungh_kim{at}hotmail.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Blood pressure • Caregiving • Cortisol • Ethnicity • Self-reported general health
Most research on caregiver distress has looked at Whites and to a lesser extent at African Americans, but very little attention has been paid to other minority caregivers. Because of the Confucian tradition and a potential view of caring for older family members as a natural part of family life, the relationship between caregiving experiences and caregiving outcomes could be demonstrated in different ways among East Asian American caregivers. Nonetheless, only a few previous studies have explored caregiving experiences and their outcomes among Korean Americans (Chun, 2004; Chun, Knight, & Youn, 2007; Knight et al., 2002; Youn, Knight, Jeong, & Benton, 1999). None of these studies focused on physical health outcomes. Therefore, it is important to study the impact of caregiving experiences on the physical health of Korean American caregivers, as they constitute a rapidly growing population in the United States.
Although many studies have supported the adverse effects of caregiving on the physical health of caregivers, adverse effects have more frequently been found with self-reported general health than with such objective physical health markers as physical illness symptoms, health care utilization, and medication use (Pinquart & Sörensen, 2007; Schultz, Visintainer, & Williamson, 1990). The inconsistent effects of caregiving on the physical health of caregivers may also be explained by the complex interaction of variables within stress and coping models for caregiving. For example, the adverse effects of caregiving on physical health can be lessened by the cultural justification for caregiving (Dilworth-Anderson et al., 2004), the psychosocial resiliency of the caregiver (Dilworth-Anderson & Anderson, 1994), the caregiver's receipt of social support from others (Barusch & Spaid, 1989; Pinquart & Sörensen), and the coping style of the caregiver (Kim, Knight, & Longmire, in press).
The sociocultural stress and coping model for caregivers (Aranda & Knight, 1997) provides a framework for understanding the effects of caregiving stress on physical health outcomes as well as on mental health outcomes (Kim et al., in press). This model adapted the Lazarus and Folkman (1984) coping model for understanding ethnic differences in response to caregiving distress. Although the initial version placed considerable emphasis on the role of perceived burden in mediating cultural differences, more recent work has suggested that the influence of cultural values as an explanation of ethnic group differences is mediated more through coping styles and social support than through perceived burden (e.g., Kim et al.; Chun et al., 2007). That is, caregivers who are in similar caregiving situations but have different cultural backgrounds may choose dissimilar coping strategies that are congruent with their cultural values, receive different levels and types of social support based on different cultural norms, and thus demonstrate dissimilar outcomes of caregiving.
Researchers have found caregiver coping style to be an important intervening class of variables in the relationship between caregiving stressors and the physical health of caregivers, as ineffective coping styles lead to adverse caregiving outcomes (Knight, Silverstein, McCallum, & Fox, 2000). Lazarus and Folkman (1984) defined avoidant coping (formerly called emotion-focused coping) as thoughts and actions whose goal is to relieve the emotional impact of stress by distraction from the stressor. They argued that avoidant coping may adversely influence one's physical health, because avoidant coping only attempts to manage one's emotion and thus leaves stressful situations unresolved. In a study of White and African American caregivers, Kim and colleagues (in press) confirmed the intervening effects of avoidant coping on the relationship between caregiver burden and poor self-reported physical health outcomes.
Lazarus and Folkman (1984) also argued that avoidant coping is more likely to occur when people appraise the given situations as unavoidable, unchangeable, and threatening. Because caregiving for frail, older family members typically involves long-term caring for a person with a progressive and chronic disease, it is expected that, when we control for life stressors other than caregiving, caregivers are more likely than noncaregivers to use avoidant coping.
Meanwhile, Lazarus and Folkman (1984) argued that active coping (formerly called problem-focused coping), which refers to efforts to alter the environment surrounding the stressors (e.g., seeking information about what to do), may favorably influence one's physical health, because active coping can eliminate the potential adverse effects of stressful situations. Goode, Haley, Roth, and Ford (1998) confirmed that the use of more active coping (called approach coping in their study) was significantly associated with the better physical health of caregivers who care for individuals with Alzheimer's disease. Kim and colleagues (in press) found a positive effect of active coping for African American caregivers with respect to blood pressure.
Previous findings have also confirmed the favorable influences of informal social support on the physical health of caregivers (Franks & Stephens, 1996; Goode et al., 1998; Monahan & Hooker, 1995; Pinquart & Sörensen, 2007). In a meta-analytic study on the physical health of informal caregivers, Pinquart and Sörensen argued that the worse physical health of caregivers was significantly associated with the receipt of less informal social support. Uchino, Kiecolt-Glaser, and Cacioppo (1992) found that caregivers with low levels of social support (considering both quantitative and qualitative aspects of social support) demonstrated significant increases in heart rate reactivity that were not explained by the normal aging process. Caregivers with high levels of social support, in contrast, showed only age-related increases in heart rate reactivity. Kiecolt-Glaser and associates (1991) reported that, among the spousal caregivers of dementia patients, those who reported lower levels of social support at the beginning of their longitudinal study were more likely than those who reported higher levels to show poorer immune functioning at the 1-year follow-up.
Although a number of studies have supported the notion that there are favorable effects of social support on the physical health of caregivers, previous research findings have found reduced involvement in social relationships after people become primary caregivers (Cantor, 1983; Moritz, Kasl, & Ostfeld, 1992). In a longitudinal study examining caregiving outcomes (Gold, Cohen, et al., 1995), the authors found smaller social networks among caregivers than among former caregivers whose care recipients had died.
In summary, there is theoretical and empirical evidence supporting the adverse relationship between caregiver status and physical health. However, the relationship has been inconsistent in previous research because there have been differences between subjective and objective measures of health outcomes and because of the influence of caregiver coping styles and social support. The previous studies were mainly based on White and African American caregivers, leaving the need for research on other minority caregivers. Therefore, our aim in this study was to investigate the direct and indirect effects of caregiver status on the physical health of Korean American caregivers. Based on the sociocultural stress and coping model as well as on empirical evidence, this study proposed a path model. The path model considered direct and indirect effects of caregiver status on physical health outcomes by means of caregiver coping styles and the quantity and the quality of informal social support (Figure 1<--CO?3-->).
|
On the basis of the sociocultural stress and coping model and the supporting literature, we proposed the following hypotheses. First, controlling for age, gender, marital status, education, life stressors other than caregiving, medication for high blood pressure (only for the levels of blood pressure), and sampling timing of cortisol (only for the levels of cortisol), we think that being a caregiver will be directly associated with poorer physical health outcomes. Second, being a caregiver will be indirectly associated with poorer physical health outcomes through a less frequent use of active coping styles. Third, being a caregiver will be indirectly associated with poorer physical health outcomes through having fewer numbers of informal social supporters and receiving a lower quality of informal social support.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
We recruited the caregiver sample in Los Angeles and Orange Counties, California, between May 2003 and August 2003. This sample of caregivers is the same as the Korean American caregiver sample used in a dissertation by Chun (2004). We recruited caregivers through various community agencies, including the Korean Health, Education, Information, and Research Center, churches, clinics, seniors' apartments, and Korean American newspaper advertisements.
As a comparison group, we recruited 87 noncaregivers who had not provided care during the previous 24 months or more. Because Gold, Feldman-Reis, and colleagues (1995) reported long-term (more than 1 year) effects of caregiving stress on the physical and mental health of caregivers, we set the period of 24 months for determining the eligibility of noncaregivers. We used three matching criteria for noncaregivers: age, gender, and timing of cortisol samples. Matched noncaregivers and caregivers were of the same gender and were within 5 years (±5 years) of the age of the given caregiver. We collected saliva samples of the matched noncaregiver within half an hour (plus or minus 30 minutes) of the sampling timing of the given caregiver. Sampling timing and gender have been reported to be the most influential factors affecting levels of cortisol (Lovallo & Thomas, 2000). Recruitment procedures were similar for both caregiver and noncaregiver samples, although we recruited the noncaregivers 1 year later than the caregivers. There was no reason to expect that the 1-year difference between the groups created any differences in stress levels.
All caregivers were interviewed in Korean for about 60 to 90 minutes with a structured questionnaire, and all of them were given $20 as compensation. Over 90% of the caregivers were interviewed in their homes for their convenience. All noncaregivers were interviewed for about 30 to 50 minutes with a structured questionnaire, and all of them were given a set of soaps as compensation. Although there were differences in the nature of compensation between caregivers and noncaregivers, this did not result in differences in responses or attitudes between the two groups of subjects. As with the caregivers, the majority of noncaregivers (over 80%) were interviewed at their homes for their convenience.
Measures
Self-Reported General Health
We operationalized the self-reported general health of the caregivers by asking how an individual perceives his or her current physical health. We assessed the caregiver's self-rated general health by using one standard, single-item scale (1 = excellent, 2 = good, 3 = fair, 4 = poor, 5 = very poor), with a higher score being indicative of worse self-reported general health. We used the same measure for noncaregivers.
Blood Pressure Levels
We used self-reported levels of systolic and diastolic blood pressure as a second physical health indicator. Cardiovascular indicators are psychophysiological variables that may serve as intervening variables between objective stress and its appraisal and health outcomes. According to theories of allostatic load, a complex set of hormonal and biological responses including cardiovascular responses are essential for adaptation, maintenance of homeostasis, and survival (allostasis; see McEwen & Seeman, 1999). However, repeated cycles of allostasis as well as the inefficient turning on or shutting off the body's responses increase vulnerability to diseases. Previous studies have reported that cardiovascular indicators demonstrate one's risks for stress-related cardiovascular illnesses such as hypertension (Fredrickson & Matthews, 1990) and coronary heart disease (Blascovich & Katkin, 1993). In a longitudinal study examining work stress and risk of cardiovascular mortality, Kivimäki and colleagues (2002) reported that, after adjustment for age, gender, occupational group, and biological and behavioral risks at baseline, participants with high job strain (a combination of high demands at work and low job control) had double the cardiovascular mortality risk of their colleagues who had low job strain.
In our study, we asked each caregiver what the values of his or her blood pressure were at the most recent examination. The responses were given as continuous blood pressure values. If respondents did not know what the exact values were but knew only that the values were normal, then we entered systolic and diastolic blood pressure as 120 and 80, respectively. We did this for approximately 15% of the samples for systolic blood pressure and 20% for diastolic blood pressure. If a caregiver did not know his or her levels of blood pressure, then we treated the answer as missing. We applied the same measure for noncaregivers. Among both caregivers and noncaregivers, fewer than 7% of the blood pressure variables were missing (5.2% for systolic blood pressure and 6.3% for diastolic blood pressure). The values of missing cases were automatically calculated by the AMOS program.
Although the low percentage of missing data was surprising, it may be related to the easy access to services for blood pressure measurement in Koreatown, Los Angeles. Most drug stores, acupuncture clinics, physical therapy clinics, and alternative food and herb stores in Koreatown provide free measurement services of blood pressure. As a result, older Korean American individuals, who are mostly beneficiaries of Medicaid and Medicare and who regularly visit these places, tend to be aware of their blood pressure levels. The self-reported levels of systolic and diastolic blood pressure were significantly correlated with the use of medication for high blood pressure (r =.469, p <.001 for systolic blood pressure; r =.303, p <.001 for diastolic blood pressure), thus supporting the validity of the self-reported measure.
Cortisol Levels
We assessed the level of cortisol by analyzing saliva samples, as they are a convenient and unobtrusive specimen source for cortisol measurement (Kirschbaum & Hellhammer, 1989). The increased use of saliva sampling over the past 10 years suggests its widespread acceptance, simplicity, and safety (Lovallo & Thomas, 2000). Because cortisol is essential for all the forms of physiological regulation, cortisol levels have been considered to be important indicators of one's physical health. Cortisol can integrate widespread physiological functions by reaching all tissues via the systematic circulation, thereby significantly altering the background environment in which those tissues operate. In addition, cortisol modulates and coordinates activity and responses across many tissues (Lovallo, 2005). Vitaliano and associates (2003) argued that cortisol levels function as a proxy in the linkage between caregiver stress and disease outcomes. In a study examining the associations between biomarkers of stressful experience and profiles of physical and mental functioning based on a national sample of middle-aged and elderly Taiwanese, Seplaki, Goldman, Weinstein, and Lin (2004) reported that those individuals with substantial impairment in physical and mental functioning were more likely to have extreme levels of cortisol than were their counterparts without such impairments.
We collected saliva samples in our study twice—before and after the participant interview. However, because of reliability issues, we used only the second sample. Because the participants had never experienced the saliva-sampling process, the sampling process itself could be perceived as stressful and could thus influence the participants' levels of cortisol during the first measure. Because the participants felt more comfortable with the second experience of the saliva-sampling process than the first one, we thought that the level of cortisol measured after the interview was more reliable than the level measured before it.
We stored the collected saliva samples in the Andrus Gerontology Center lab freezer, which maintained a stable temperature before the samples were shipped to a designated lab in Germany (Dr. Clemens Kirschbaum's lab, Institute of Physiological Psychology II, University of Dusseldorf, Universitaetsstrasse 1, D-40225 Duesseldorf, Germany). Because cortisol is an extremely stable compound, samples can remain at room temperature for several hours or can remain frozen until processing (Lovallo & Thomas, 2000). We shipped the frozen samples to the lab with wet ice by way of a 2-night delivery. The conditions of the samples were reported as satisfactory when they arrived at the lab. We measured and analyzed the cortisol levels as values. Missing values of cortisol, which, as we stated before, was measured after the interview (5.0% among both groups), were automatically calculated by the AMOS program.
Caregiver Coping Styles
We measured coping styles, defined as ways to deal with stressful situations, by using the 28 items from the Brief COPE scale (Carver, 1997). Participants were asked to answer questions about ways they were dealing with stressful problems. For example, the items included "I've been concentrating my efforts on doing something about the situation I'm in," and "I've been getting emotional support from others." The response choices scored from 0 to 3, respectively. Fewer than 3% of missing cases in both groups of subjects were imputed with mean values at an item level. In previous studies that have used the COPE scale, the subscales of Active, Avoidant, and Faith-Oriented Coping were identified through factor analysis by McCallum (2002) and modified by Kim and colleagues (in press).
In a confirmatory factor analysis including all the subscales of McCallum (2002), that is, Active Coping, Avoidant Coping, and Faith-Oriented Coping subscales, indicators of goodness of fit of the measurement model were not satisfactory (ratio of chi square to degrees of freedom or 
2/df = 3.130; Comparative Fit Index or CFI = 0.789; root mean square error of approximation or RMSEA = 0.111). After exploring the alternative two- and three-factor structures and deleting subscales with low loadings on the factors, Chun (2004) confirmed a two-factor model for Korean American caregivers with Active-Cognitive Coping and Social Support Coping factors (Figure 2). Chun's two-coping-factor model fit the Korean American data of this study well (2/df = 0.859; CFI = 1.00; RMSEA = 0.00). These results did not change when we separately conducted confirmatory factor analyses for noncaregivers (2/df = 0.745; CFI = 1.00; RMSEA = 0.00) and for caregivers (2/df = 0.617; CFI = 1.00; RMSEA = 0.00). For the Active-Cognitive Coping factor, Cronbach's alpha was 
= 0.76 for noncaregivers and = 0.81 for caregivers. For the Social Support Coping factor, Cronbach's alpha was = 0.64 for noncaregivers and = 0.69 for caregivers. Thus, our work shows cultural noninvariance for coping factors, with Korean and Korean American caregivers showing a distinct set of coping styles as compared with African American and White caregivers. This difference in factor structures is in itself an important influence in that cultural values can shape coping styles and so lead to ethnic group differences in the stress and coping process.
|
We measured the quantity of informal social support by using the two items from the Support Questionnaire, which made inquiries regarding the number of instrumental helpers and the number of emotional helpers. For example, the items included "How many people help you with tasks of caregiving (i.e., keeping an eye on the person, helping with bathing, feeding, etc.) at any time?" and "How many people can you talk to about the things related to caregiving that are difficult for you?"
We measured the quality of informal social support by using the 18 items from Main Helper Questionnaires 1 and 2; of these items, 9 measured positive and negative support from Main Helper 1 (instrumental support provider) and 9 measured this from Main Helper 2 (emotional support provider). The participants were requested to answer questions related to how they feel about their main instrumental social supporter and their main emotional social supporter. If a participant answered the quantity of social support as "0" and the participant was not able to answer the questions regarding the quality of social support, we then coded the quality of social support of the participant as zero, indicating the lowest quality of social support. Some of the scales represented positive support (e.g. "How much does your main helper make you feel loved and cared for?") and negative support (e.g., "How much does your main helper argue with you?"). We scored the response choices from 0 to 4 (0 = not at all, 1 = quite a bit, 2 = some, 3 = a little, and 4 = a great deal); the total scores for instrumental and emotional social support ranged from 0 to 36 for each type. We recoded six items representing negative social support, so that higher scores indicated a higher quality of instrumental and emotional social support. For the quality of instrumental social support, Cronbach's alpha for this study was = 0.81 for each group of noncaregivers and caregivers. For the quality of emotional social support, Cronbach's alphas for this study were = 0.79 for noncaregivers and = 0.76 for caregivers.
We analyzed the two dimensions of social support (quantity and quality of social support) in separate models; we included the two types of social support (instrumental and emotional social support) in each of them. Thus, this study had four variables of informal social support: the quantity of instrumental social support, the quantity of emotional social support, the quality of instrumental social support, and the quality of emotional social support.
Control Variables
We included age, gender (0 = male, 1 = female), education, and marital status (0 = married, 1 = unmarried) as background variables because they are known to be strong correlates of physical health. We excluded income, which has been reported as an important factor predicting physical health, in these path models because it was significantly correlated to education (r =.31, p <.001). We did not consider coresidence and the relationship of caregivers to their care recipient (spouse, daughter, or daughter-in-law) as background variables in this study because of multicollinearity issues. Most physical health indicators except self-reported general health were not significantly correlated to the variables (Table 1)<--CO?4--><--CO?5-->. Furthermore, coresidence and the relationship variables were significantly correlated to age (r = –.444 for coresidence; r =.772 for spouse caregiver; r = –.516 for daughter caregiver; r = –.427 for daughter-in-law caregiver), which was already included as a control variable.
|
Analysis
In the path model, we set some control variables (age, gender, education and marital status, caregiver status, and daily life stress other than caregiving) to predict coping styles, informal social support, and physical health indicators. However, we did not set the control variables of use of medication for high blood pressure and cortisol-sampling timing to predict the intervening variables because the paths were not theoretically reasonable. We set medication for high blood pressure to predict only the levels of blood pressure levels. We set cortisol-sampling timing to predict only cortisol levels.
Because the sample size was small, we calculated separate path models for the classes of potential mediating variables (coping styles, the quantity of social support, and the quality of social support) and for the different classes of health outcomes in order to keep the number of parameters in each model to a reasonable number. For coping styles, we used active-cognitive coping and social support coping. For informal social support, we used two types of informal social support in separate models: the quantity and the quality of social support. For physical health outcomes, we used self-reported general health, systolic and diastolic blood pressure levels, and cortisol levels separately; in other words, we tested the physical health indicators separately. Because the intervening variables (coping styles, the quantity of informal social support, and the quality of informal social support) and the three physical health indicators were employed separately, there was a total of nine analyses (3 intervening variables x 3 health indicators). We tested the proposed path models and study hypotheses by using the AMOS statistical program (Arbuckle, 1997). We assessed the path models by examining the statistical significance of estimated path coefficients and several statistical indices indicating goodness of fit for the model as a whole: CFI, 2/df, and the RMSEA. In general, better fitting models have higher CFI and lower 2/df and RMSEA values. Although the rules for determining the goodness of fit of models vary, we considered values of CFI > 0.90 to be representative of a well-fitting model (Bentler, 1990). We considered values of 2/df < 2.0 to represent a good fit. Values for RMSEA up to 0.08 represent reasonable errors of approximation in the population (Browne & Cudeck, 1993).
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
24% of care recipients had dementia). Caregivers reported on whether the care recipients had any of nine types of physical ailments (respiratory, circulatory, digestive, nervous, endocrine, rheumatoid or arthritic, tumor, chronic illness, and other disorders). The most frequently mentioned physical illness was circulatory system disorders (40.2%), followed by rheumatism or arthritis (37.9%) and gastritis problems (37.9%). In terms of functional abilities, most of the care recipients received at least a little help for the four IADLs (shopping, using transportation, managing medication, and doing housework). More than half of the care recipients received help for at least three out of the five ADLs (bathing, dressing, and getting up from a chair). Approximately one third of care recipients received help for toileting and eating. Table 2 contains additional information on demographic and descriptive characteristics of care recipients and caregiving-related experiences.
|
|
2/df = 1.627, CFI = 0.973, and RMSEA = 0.056 for the self-reported general health model; 2/df = 1.475, CFI = 0.973, and RMSEA = 0.052 for the blood pressure model; 2/df CMIN/DF = 1.592, CFI = 0.957, and RMSEA = 0.059 for the cortisol model). When we included the quantity of informal social support as intervening variables, the proposed models also fit the Korean American sample of this study (2/df = 1.621, CFI = 0.945, and RMSEA = 0.060 for the self-reported general health model; 2/df = 1.499, CFI = 0.973, and RMSEA = 0.054 for the blood pressure model; 2/df = 1.469, CFI = 0.930, and RMSEA = 0.052 for the cortisol model). When we included the intervening variable of the quality of social support, the proposed models were well fitted to the Korean American sample of this study (2/df = 1.633, CFI = 0.950, and RMSEA = 0.060 for the self-reported general health model; 2/df = 1.677, CFI = 0.967, and RMSEA = 0.063 for the blood pressure model; 2/df = 1.612, CFI = 0.925, and RMSEA = 0.059 for the cortisol model).
In the interest of parsimony, we estimated final models after we deleted nonsignificant paths from the originally proposed models. Because the coping variables and the quantity of social support were not found to mediate caregiver status effects on health outcomes, we proposed three out of the nine final path models including the quality of social support as their intervening variables with standardized path coefficients for self-reported general health, the levels of systolic and diastolic blood pressure, and the levels of cortisol (Figures 3–5
). Because direct effects of caregiver status on physical health indicators were shown in similar ways across the path models, we only report the path models supporting indirect effects of caregiving status on physical health. The unstandardized coefficients and their significance levels for the path models, which are not shown in this article, are available from J-H. Kim on request.
|
|
|
2/df, CFI, and RMSEA (Table 4). The chi-square differences between the first models and the reduced models were not significant, with 
2(10) = 12.57 for the self-reported health model, 2(13) = 11.61 for the blood pressure model, and 2(11) = 15.89 for the cortisol model. Other goodness-of fit indicators confirmed that the reduced models are acceptable.
|
The proposed indirect effects of caregiver status were supported only for cortisol levels through the quality of informal social support (Figure 5). Being a caregiver was associated with higher levels of cortisol in terms of receiving lower levels of quality of instrumental social support. The hypothesized indirect effects of caregiver status on physical health outcomes in terms of coping styles or the quantity of informal social support were not supported.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
The fact that the levels of blood pressure were higher among Korean American caregivers than among those of their matched noncaregivers was consistent with the previous findings of higher systolic blood pressure among caregivers (Schulz et al., 1997; Vitaliano et al., 2002). Higher levels of cortisol among Korean Americans also replicated findings reported among White caregivers in the United States (Vedhara et al., 1999).
However, the adverse effects of caregiver status on the physical health of caregivers were not found with the measure of self-reported general health. This finding was consistent with the finding of Haley and colleagues (1995) regarding the worse self-reported physical health among African Americans than among Whites, which were not explained by caregiver status. These findings, taken together, support an understanding of the effects of caregiving on health in which consistent effects on biomarkers of stress reaction serve as evidence of the effects of caregiving on the physical health of caregivers, even when these effects have not entered the realm of the conscious appraisal of health by the caregivers. Knight, Longmire, Dave, Kim, and David (in press) found a similar result for the specific effects of caregiving on African Americans' blood pressure, although they did not note a specific effect on subjective health. Vitaliano and associates (2003) placed these types of measures early in the causal model leading to poor health outcomes, and so it is also conceptually consistent that these measures would change before health would be perceived as having changed.
Only one finding of this study supported the hypothesized mediators of the effects of caregiver status on physical health outcomes. Caregivers who reported a lower quality of instrumental social support had higher levels of cortisol, indicating greater physiological stress. The lower quality of instrumental social support among Korean American caregivers than among noncaregivers was consistent with findings from studies of other ethnic groups. Haley and colleagues (1995) noted lower levels of satisfaction with informal social support among both White and Black caregivers compared with their noncaregiver counterparts, despite a higher number of visits from relatives. This may be due to the subjective nature of the quality of instrumental social support. Caring for older, frail family members or relatives may result in a situation in which caregivers expect a large quantity of instrumental social support. The imbalance between the supply of instrumental social support and the perceived need for it could result in a lowered perception of the quality of instrumental social support, leading in turn to increases of stress hormones.
A few caveats have to be considered with the present study. First, the small sample sizes result in low statistical power. If the magnitude of population effect sizes is low to medium, then these smaller effects may not be detected in the present study. Second, the cross-sectional design is not optimal for detecting the physical health effects of caregiving, which may take time to develop (Vitaliano et al., 2002). Third, one could overestimate the adverse effects of caregiver status on objective physical health by not considering other influential factors on one's physical health (e.g., body mass index, health behavior, and other health conditions). Fourth, self-reported blood pressure levels are also a limitation, even though the self-reported levels used in this study were significantly correlated with the use of medication for high blood pressure.
Nonetheless, to our knowledge, the present study is the first to examine the effects of caregiver status, social support, and coping styles on the physical health of Korean American caregivers. In addition, the present study is one of the few studies of any ethnic group having a matched reference group of noncaregivers and employing both subjective and objective physical health indicators. The demonstration of the physical health effects of caregiving in one of the nation's fastest growing ethnic groups, and the finding that these physiological effects occur without self-reported poor health, call attention to a potentially serious health problem in an understudied group providing family care to frail older family members.
|
Footnotes |
|---|
2 Department of Psychology, University of Southern California, Los Angeles.
Decision Editor: William J. McAuley, PhD
Received for publication March 19, 2007. Accepted for publication July 24, 2007.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
