| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Ladson Hinton, MD, Department of Psychiatry, School of Medicine, University of California, 2230 Stockton Boulevard, Sacramento, CA 95817. E-mail: ladson.hinton{at}ucdmc.ucdavis.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Neuropsychiatric symptoms • Caregiving • Hispanic • Dementia
Neuropsychiatric disturbances are common in dementia and often present unique challenges for caregivers (Cohen-Mansfield, Taylor, & Werner, 1998; Lyketsos & Steinberg, 2000; Mega, Cummings, Fiorello, & Gornbein, 1996; Ory, Hoffman, Yee, Tennstedt, & Schulz, 1999; Reisberg et al., 1987). Recent population-based studies have found that between 61% and 80% of elderly persons with dementia have one or more mental and neuropsychiatric disturbances (Lyketsos et al., 2000, 2002). A longitudinal study of 181 patients with Alzheimer's disease (AD) found that, at any given point in time, 23–29% had symptoms of depression, 28–32% had symptoms of aggression, and 12–25% had symptoms of psychosis (Levy et al., 1996). In another longitudinal study of community-dwelling patients with AD, neuropsychiatric symptoms, specifically aggression, were the most common and persistent; psychotic symptoms were intermediate; and depressive, vegetative symptoms the least likely to persist (Devanand et al., 1997). Longitudinal research has attempted to tease apart the interrelationships among different types of neuropsychiatric symptoms; for example, the presence of depression (Lyketsos et al., 2000) and the presence of delusions (Gilley, Wilson, Beckett, & Evans, 1997) have been found to be predictors of physical aggression among patients with AD.
There are conflicting data about whether the prevalence of dementia neuropsychiatric symptoms varies by ethnicity or race (Chen, Borson, & Scanlan, 2000; Cohen & Magai 1999; Hargrave, Stoeklin, Haan, & Reed, 2000; Harwood et al., 1998; Yeo & Gallagher-Thompson, 1996). One study of minority elderly persons evaluated at the California Alzheimer's Disease Center found the frequency of psychiatric symptoms to be similar in minority and nonminority elderly people (Yeo & Gallagher-Thompson, 1996). Another study also found quite similar rates of neuropsychiatric symptoms among patients (Hargrave et al., 2000). However, a third study found significant differences in the prevalence of neuropsychiatric symptoms across different ethnic minority groups (Chen et al., 2000). One of the limitations of existing studies is that they have focused on highly selected clinic-based populations or samples drawn from social service agencies, and they are unlikely to be representative of the general population of elderly individuals with dementia.
In nonminority populations, studies have consistently found that neuropsychiatric problems are associated with caregiver depression (Chappell & Penning, 1996; Clyburn, Stones, Hadjistavropoulos, & Tuokko, 2000; Donaldson et al., 1997; Schulz et al., 1995; Teri, 1997). A large community-based longitudinal study of dementia caregivers found support for a model in which neuropsychiatric disturbances in care recipients predicted increased caregiver burden, which in turn caused increased caregiver depression (Clyburn et al., 2000). What remains unexamined in the literature, however, is whether the impact of neuropsychiatric symptoms varies by caregiver characteristics, such as age, gender, or spousal status, or contextual factors, such as culture, social networks, and local community organizations. Such data could help identify high-risk subgroups for interventions and refine our understanding of the processes through which these disturbances lead to caregiver depression.
Mental and neuropsychiatric disturbances in care recipients may affect families of color differently than European American families. For example, Valle found that Mexican American caregivers were more "reactive" to neuropsychiatric symptoms compared with their European American counterparts (Valle, 1994). One study of Cuban American caregivers did not find a significant association between care recipient psychotic symptoms and caregiver depression (Harwood et al., 1998), leading the authors to conclude "patient psychopathology may play a decreased role in Hispanic caregiver burden" (p. 344). Although the studies by Harwood and Valle would seem to point in different directions, they were conducted with different Latino populations (Cuban Americans and Mexican Americans, respectively) and used very different measures to assess caregiver distress and care recipient neuropsychiatric symptoms. Clearly, more research is needed on the relationship of neuropsychiatric symptoms in demented patients to Latino caregiver mental health, especially in view of studies showing that Latino caregivers suffer higher rates of depression compared with European American caregivers (Cox & Monk, 1990; Harwood et al., 1998; Mintzer et al., 1992; Valle, 1994).
This study's specific aims were to (a) evaluate neuropsychiatric symptom frequency and intensity in Latino elderly people who were demented and in Latino elderly people who had mild cognitive impairment but did not meet criteria for dementia (CIND), (b) evaluate whether neuropsychiatric symptom intensity would be associated with higher levels of caregiver depression, and (c) identify factors that modify the relationship between neuropsychiatric symptoms and caregiver depression. To our knowledge, this is the first population-based study to examine the prevalence and caregiver impact of neuropsychiatric symptoms in Latino elderly people with dementia and those who had CIND.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
In the SALSA study, cognitive impairment at baseline was assessed in five steps. First, all SALSA participants were screened by use of the Modified Mini-Mental State Exam, or 3MSE (Teng & Chui, 1987), and a word-list learning test that is part of a battery of psychometrically matched English- and Spanish-language neuropsychological assessment instruments (Mungas, Reed, Marshall, & Gonzalez, 2000). Second, participants who scored below the 20th percentile (education, age, gender, and language adjusted) on either screening test and a random subsample of those scoring above the thresholds were referred for testing with five other neuropsychological tests (Mungas et al., 2000). Third, participants scoring below the 10th percentile on one or more neuropsychological tests and showing evidence of functional decline on a standardized instrument, the Informant Questionnaire on Cognitive Decline in the Elderly (Jorm, 1994), were referred for neuroclinical examination. Fourth, all participants undergoing a neuroclinical examination were case adjudicated by a team of neurologists and a neuropsychologist to determine the presence or absence of dementia according to Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R; American Psychiatric Association, 1987) criteria. Fifth, participants meeting criteria for dementia were referred for magnetic resonance imaging and then case adjudicated again to determine their dementia subtype.
The present study used the SALSA evaluation of cognitive impairment and dementia to identify a subsample of SALSA elderly people with dementia and a group who had CIND. We drew a sample of all SALSA elderly people with dementia, all SALSA participants who were referred for neuroclinical exam but were found not to be demented, and a random subsample of those who scored below the 20th percentile but were not referred for neuroclinical examination. For the purposes of this study, the CIND group is defined as the latter two subsamples (i.e., participants who scored below the 20th percentile on either of cognitive screening tests but were not referred for further examination as well as all participants who were referred for neuroclinical examination but were determined not to be demented). To be eligible for the present study, elderly persons also needed to be noninstitutionalized (i.e., not in a nursing home or long-term care facility) and to be currently living in one of the target counties.
The present study consisted of two interviews: (a) an initial in-home interview with the cognitively impaired elderly person (or a proxy) that included a standardized set of questions to identify the primary, informal caregiver, and (b) a semi-structured interview with the primary informal caregiver. All interviews were conducted in either Spanish or English by a bilingual, bicultural research assistant who was trained by the principal investigator (L. Hinton). The study interviews were conducted from November of 1999 through August of 2000. This corresponded to the time when SALSA participants were undergoing their first annual follow-up visits.
Measures
Care recipient characteristics that were not measured by SALSA (i.e., neuropsychiatric symptoms) or that might change significantly over time (i.e., functional impairment) were assessed through a direct interview with the caregiver. Data on other care recipient characteristics were drawn from measures administered as part of the SALSA study. Care recipient mental and neuropsychiatric disturbances were measured by the Neuropsychiatric Inventory, or NPI(Cummings et al., 1994). The instrument encompasses neuropsychiatric disturbances that are common in neuropsychiatric disorders, such as dementia. The instrument was administered to caregivers, who were asked to rate the presence, frequency, and severity of 10 symptom domains during the previous month: depression, anxiety, apathy, inappropriate elation, aggression/agitation, irritability, disinhibition, hallucinations, delusions, and aberrant motor activity. In previous studies the NPI has been shown to be a valid and reliable instrument (Cummings et al., 1994; Lyketsos & Steinberg, 2000). In the present study, an existing Spanish version developed and validated in Spain (Vilalta-Franch et al., 1999) was further refined through a translation and back-translation process to ensure that it was appropriate for use in a predominantly Mexican American population. Questions that asked the caregiver about the care recipient's ability to perform 13 common activities of daily living (ADLs) were used to assess functional limitations. For each item, the caregiver was asked whether the care recipient didn't need any help (0), needed some help (1), or needed help all the time (2). For the purposes of analysis, ADLs were scored 0 if the care recipient didn't need any help and 1 if at least some help was needed. Scale items were summed to yield a scale with a range of 0–13. In order to minimize participant burden, other care recipient characteristics were derived from data collected by SALSA, including basic sociodemographics and two measures of cognitive functioning, that is, the 3MSE (range 0–100) and the presence or absence of dementia based on DSM-III-R criteria (as described earlier). For regression analyses, age was centered by subtracting 75 (sample median age) from care recipient age (i.e., age – 75). Because data collection occurred in the year following the SALSA baseline assessment, a decision was made to include in our demented group those elderly persons who met criteria at baseline as well the incident cases at the first annual follow-up visit (n = 6).
Caregiver characteristics were measured through the caregiver survey and, for those caregivers who were also SALSA participants, by drawing on SALSA data. The Center for Epidemiological Studies Depression scale (CES-D; range 0–60), a depression measure that has been extensively used in Latino populations, and questions about the characteristics of the caregiving situation were administered to all caregivers (Radloff, 1977). For a few who were missing the CES-D as part of the substudy, we used their SALSA CES-D scores from SALSA's first annual follow-up visit that was being conducted in parallel with this study. The caregiver's relationship to the care recipient was assessed and dichotomized as spouse versus nonspouse (two caregivers who identified themselves as unmarried partners of the care recipients were included in the spousal group). Sociodemographic characteristics and health status of caregivers were assessed in person or, for those 36 caregivers who were SALSA participants, by drawing on their SALSA data. Caregivers were asked to rate their health status on a scale of 1 = excellent to 5 = poor, and this variable was later collapsed to yield three levels: 0 = excellent or good, 1 = fair, and 2 = not good or poor.
Statistical Approach
To address the first study aim, the prevalence and intensity of individual neuropsychiatric symptoms were compared across the demented and CIND groups. A symptom "intensity" score was calculated for each neuropsychiatric symptom by multiplying its frequency (range 0–4) and severity (range 0–3). A total neuropsychiatric symptom score (NPI total score) was calculated by summing the individual symptom intensity scores (range 0–120). A chi-square test or Fisher exact test was used to compare the prevalence of individual symptoms in the two groups. The 95% confidence intervals for individual symptoms were also calculated. Because the distribution of individual symptom intensity scores and NPI total score was skewed, nonparametrics (i.e., Wilcoxon rank sum test) were used to compare intensity scores for the demented and CIND groups.
To evaluate whether the association of care recipient's NPI total score and caregiver depression was robust, we examined the relationship between neuropsychiatric symptom score and caregiver depression before (Model 1) and after (Model 2) we adjusted for potential confounders. Because regression diagnostics (i.e., a plot of the residuals of depression and neuropsychiatric symptoms) revealed a significant problem with heteroscedascity, a generalized linear model using a log link and Poisson distribution (also known as Poisson regression) was used (Liang & Zeger, 1986). Preliminary analyses conducted prior to the regression (i.e., examining univariate frequency distributions of the NPI total score and of the key confounders, followed by crude associations, cross-tabulation, and graphs) detected potential problems with collinearity among independent variables that might influence interpretation of the results. Highly associated pairs of variables were as follows: caregiver age and caregiver spousal status; number of ADL or instrumental ADL impairments with care recipient 3MSE score, the NPI total score, dementia status, and the care recipient's age; care recipient's gender with both caregiver spousal status and the caregiver's gender; and caregiver spousal status and caregiver health status. A decision was made to exclude caregiver age, care recipient ADL impairments, and care recipient gender from subsequent analyses. Caregiver spousal status and health status were retained in the modeling because both variables have been shown in previous studies to be strongly correlated with caregiver depression. Although retaining the latter two variables in the model limited our ability to assess their separate effects on caregiver depression, it permitted us to evaluate the main goal of this part of the analysis: to determine the strength of the relationship between neuropsychiatric symptoms and caregiver depression.
To address the final aim, a series of analyses were conducted to identify interactions between neuropsychiatric symptoms and other independent variables. Interaction terms significantly (<.10) associated with caregiver depression were included in the subsequent models. Both forward and backward stepwise regressions were conducted, and only those main effects and interaction terms that were significantly (i.e., <.05) associated with caregiver depression were retained. The stability of the final model was confirmed by repeating the analysis after several outliers were removed; this analysis yielded consistent results.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
|
We examined the interaction of the NPI total score with other predictors of caregiver depression (see Table 4). The final model includes three significant two-way interaction effects. The impact of two of these interactions effects (whether the caregiver is a spouse and care recipient dementia status) is illustrated in Figure 2. This figure illustrates the association between caregiver depression and care recipient NPI total score, using values for a woman caregiver in good health who is the caregiver of a 75-year-old care recipient. For this combination of values, the steeper the curve, the greater the impact of neuropsychiatric symptoms on caregiver CES-D score. The greatest impact of care recipient NPI on depression is for nonspousal caregivers of elderly persons who have CIND. Spousal caregivers of CIND elderly persons and nonspousal caregivers of demented elderly persons experience similar effects on depression from increasing care recipient neuropsychiatric symptoms. However, spousal caregivers of elderly people with dementia actually show a modest decrease in CES-D depression score with increasing neuropsychiatric symptoms. It is important to note that being a spousal caregiver conferred a higher risk for depression at lower or absent levels of neuropsychiatric symptoms. Finally, greater care recipient age appears to have a buffering effect on caregiver depression.
|
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
To our knowledge, this study is the first (in Latinos or non-Latinos) to report spousal status as a mediator of the relationship between neuropsychiatric symptoms and caregiver depression. The relationship between the caregiver and care recipient (i.e., whether the caregiver was the care recipient's husband or wife) influenced both the strength and direction of the association between NPI score and caregiver depression. For example, in the subgroup of spousal caregivers of elderly people with dementia, there was an inverse relationship between NPI total score and caregiver depression. There are several possible explanations for these findings. It is plausible that spouses have been caregivers for a longer period of time compared with nonspouses, resulting in a greater degree of "adaptation" to mental and neuropsychiatric disturbances. Combined with a selection effect, in which caregivers who were not able to adapt are more likely to relinquish their role to other family members or utilize long-term care, this would result in a more resilient group of spousal caregivers. Second, "protective" factors, such as a larger social support network, religiosity, or a more traditional Mexican cultural orientation, may be more common in spouses than in nonspouses. The possible influence of acculturation is intriguing because it is associated with age (i.e., older adults are more likely to be more traditional in their cultural orientation), and a recent large community study of Mexican Americans has found that immigrants, who are more likely to hold a traditional cultural orientation, have lower rates of a number of psychiatric disorders, including depression (Vega et al., 1998). In-depth, qualitative studies of caregiving among Latino spousal caregivers of elderly individuals with dementia might also shed light on social and psychological processes that underlie this intriguing association among Latino spousal caregivers. It will also be important to see if spousal status is an important mediator of caregiver depression and burden in other racial and ethnic groups.
Another unanticipated and noteworthy finding is the significant impact of neuropsychiatric symptoms on caregivers of CIND elderly persons. Little is known about either the prevalence of noncognitive symptoms in CIND elderly persons or their impact on family caregivers. This group of elderly people is a "high-risk" population from the standpoint of developing AD or a related dementia, and one that is becoming increasingly important as a target for prevention studies (Petersen et al., 2001). Why should neuropsychiatric symptoms have more of an impact on caregivers of CIND elderly people? One possible explanation for our findings is that prior to diagnosis, neuropsychiatric symptoms are more stressful to caregivers because their etiology is unclear, leading to higher caregiver depression. After diagnosis, neuropsychiatric symptoms are likely to be attributed to a disease process such as AD or a related dementia. As the focus of AD research moves in the direction of comprehensive early intervention and prevention models, it will be important to learn more about sources of strain for caregivers of elderly persons who are CIND.
Our results support the notion that neuropsychiatric pathology is more common and intense in Latino elderly persons with dementia compared with CIND Latino elderly, a finding that is consistent with studies of White non-Hispanic elderly persons with and without dementia (Lyketsos et al., 2000). More surprising are the relatively high levels of neuropsychiatric symptoms in both the dementia and CIND groups, at least in comparison with European American elderly people. The Cache County Study, which used methods very similar to those of the present study, including their method of assessing both dementia and neuropsychiatric symptoms, found much lower levels of almost all neuropsychiatric symptoms for their frequency and intensity scores (Lyketsos et al., 2000). One recent study found that Latino elderly persons manifested higher levels of neuropsychiatric symptoms compared with African American, Asian American, and Native American elderly persons (Chen et al., 2000). Clearly, systematic cross-ethnic comparisons of neuropsychiatric symptoms in older minority and nonminority elderly people drawn from population-based samples are needed in order to determine if neuropsychiatric symptom burden is higher in particular ethnic groups.
It is important to note strengths and limitations of this study. Unlike many previous studies of caregivers, this sample was drawn from a relatively large population-based sample, making it more representative of the general population of Latino elderly people with dementia and their caregivers who are living in the target areas. Another strength is the uniform and comprehensive ascertainment of dementia status and cognitive impairment. A potential limitation is that this study relied on caregiver reports of care recipient neuropsychiatric symptoms rather than a more "objective" clinical assessment. The NPI has been validated in nonminority populations through comparison with other established instruments (Cummings et al., 1994). The Spanish-language version used in this study has been shown to have reliability and face validity (Vilalta-Franch et al., 1999). Additional validation studies of instruments measuring behavioral disturbances in minority populations are needed. Another limitation of this study is that the sample size is relatively small, at least by epidemiological standards, resulting in wide confidence intervals for estimates of neuropsychiatric symptom prevalence.
The results of this study may have important clinical and public health implications. The high prevalence of behavioral symptoms found in these Latino elderly people suggests a substantial need for formal services to assist families. Based on previous community studies of Latinos examining access to mental health services (Vega, Kolody, Aguilar-Gaxiola, & Catalano, 1999), it is likely that many of the families participating in this study are receiving little if any formal treatment for these neuropsychiatric symptoms. Clinicians, particularly those in primary care settings where most care for dementia occurs, should also be aware of the potential adverse mental health impact of behavioral symptoms on Latino caregivers. These data underscore the need to develop and disseminate interventions to help Latino caregivers manage these dementia-related problems more effectively. Data presented here may also help identify high-risk subgroups of caregivers who may benefit from early interventions. Finally, it is equally important to deepen our understanding of those who appear to be adapting well in the face of considerable stress. Additional research is needed that is longitudinal in nature and that is broader in its view of outcomes, both negative and positive, that result from the caregiving process.
|
Footnotes |
|---|
1 Department of Psychiatry, University of California, Davis.
2 Department of Epidemiology, University of Michigan, Ann Arbor.
3 Department of Mathematics, Texas A&M University, College Station.
4 Department of Neurology, University of California, Davis.
Decision Editor: Laurence G. Branch, PhD
Received for publication July 31, 2002. Accepted for publication October 23, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. Ortiz, L. J. Fitten, J. L. Cummings, S. Hwang, and M. Fonseca Neuropsychiatric and behavioral symptoms in a community sample of Hispanics with Alzheimer's disease. American Journal of Alzheimer's Disease and Other Dementias, August 1, 2006; 21(4): 263 - 273. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
