| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Gwendolen T. Buhr, MD, Department of Medicine, Division of Geriatrics, Box 3003, Duke University Medical Center, Durham, NC 27710. E-mail: buhr0001{at}mc.duke.edu
 |
Abstract |
|---|
 TOP Abstract Design and MethodsResultsDiscussionReferences |
|---|
Key Words: Caregiver • Dementia • Task demand • Institutionalization • Nursing home placement
Three categories of patient and caregiver factors have been found to predict institutionalization in the context of dementia: patient demographics, caregiver characteristics, and patient health status. Regarding patient demographics, in several studies, married elders were less likely to enter a nursing home, suggesting the importance of spousal caregivers (Colerick & George, 1986; Hanley, Alecxih, Wiener, & Kennell, 1990; Pot, Deeg, & Knipscheer, 2001). However, in other studies, when the patient–caregiver relationship and household composition were included in multivariate models, only living alone related to institutionalization (Branch & Jette, 1982; Gaugler, Kane, Kane, Clay, & Newcomer, 2003; Greene & Ondrich, 1990; Wolinsky, Callahan, Fitzgerald, & Johnson, 1992; Yaffe et al., 2002). In yet another study, though living alone was included, lower rates of institutionalization were associated with the caregiver's being a spouse or child (Hebert, Dubois, Wolfson, Chambers, & Cohen, 2001). Thus, there is a complex relationship among living alone, the relationship of the caregiver to the patient, and institutionalization. This literature also shows that Whites compared with Hispanics and African Americans have had higher rates of institutionalization (Gaugler et al., 2003; Greene & Ondrich; Hanley et al.; McFall & Miller, 1992; Schulz et al., 2004; Wolinsky et al.; Yaffe et al.). Finally, older age has been a risk factor for permanent placement in some studies (Branch & Jette; Gaugler et al., 2000; Hanley et al.; McFall & Miller; Wolinsky et al.), but not in others (Colerick & George; Yaffe et al.).
Among caregivers, perceived stress or burden has been shown to be a key predictor of institutionalization (Pot et al., 2001; Gaugler et al., 2003; Hebert et al., 2001; McFall & Miller, 1992; Schulz et al., 2004; Yaffe et al., 2002), as have feeling trapped in the role (Gaugler et al., 2000), health problems (Hebert et al.), and poor self-rated health (Gaugler et al., 2000; Gaugler et al., 2003). Studies involving caregiver age have provided mixed results, with both younger (Colerick & George, 1986) and older caregivers (Gaugler et al., 2003; Hebert et al.; Yaffe et al.) choosing institutionalization.
Studies including dementia characteristics have revealed that higher levels of cognitive and functional impairment and the presence of difficult behaviors in care recipients predict institutionalization (Gaugler et al., 2000; Gaugler et al., 2003; Hebert et al., 2001; Pot et al., 2001; Yaffe et al., 2002). Greater risk also has been associated with declining functional status (Gaugler et al., 2003; Wolinsky, Callahan, Fitzgerald, & Johnson, 1993) and a prior nursing home stay (Hanley et al., 1990; Wolinsky et al., 1992).
In summary, previous research indicates that both patient and caregiver factors predict nursing home admission. Caregivers' stated reasons generally are not reported as part of this body of literature. The few studies that report caregivers' reasons are limited by small sample sizes and have used data gathered from a single interview, either retrospectively (Wingard, Williams-Jones, McPhillips, Kaplan, & Barrett-Connor, 1990) or around the time of admission (Arling & McAuley, 1983; Nolan & Dellasega, 2000). To better understand the complex decision-making process surrounding institutionalization, in this study we focused exclusively on caregivers' reasons for institutionalization and related these to caregiver and patient characteristics gathered in advance of the decision. The study population was a large nationally representative group of caregivers of elderly male veterans with dementia. Our study aims were as follows: (a) to identify the relative importance of reasons for institutionalization endorsed by caregivers of patients with dementia; (b) to examine the relationship between caregivers' reasons for institutionalization and indicators of caregiver and patient physical and emotional functioning measured in the prior year; and (c) to use these indicators to compare caregivers who institutionalized their care recipients with caregivers who did not. To our knowledge, this is the first study to examine how caregivers' stated reasons for institutionalization relate to more objective evidence collected prior to the placement transition.
|
Design and Methods |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
Caregivers were initially surveyed in 1998, and follow-ups were sent in 1999, 2000, and 2001. We excluded caregivers who were unreachable or no longer providing care by Year 2 from this analysis, leaving 2,211.
Procedures
The Year 2, 3, and 4 surveys began with questions concerning changes that may have occurred in the caregiving situation since the last survey. These included an alternate primary caregiver for the patient, a temporary stay in a care facility by the patient, a permanent move to a nursing home by the patient, or the death of the patient. When institutionalization occurred, caregivers were asked to provide the date of admission, name of the nursing home, distance from the caregivers' home, frequency of visitation, and whether the nursing home was private pay or a VA facility. The questionnaire then provided five possible reasons that the move occurred, and participants were asked to check all that applied: (a) "my health would not permit me to continue caregiving"; (b) "my loved one's behavior became too difficult to handle"; (c) "my loved one needed more advanced or skilled care than I could provide"; (d) "I needed more assistance from others and could not get it"; and (e) "family or friends thought it was the best thing to do." Caregivers could also provide a reason in their own words.
We took several measures to ensure that the move was permanent. First, there was a separate question regarding temporary admissions. Furthermore, if caregivers reported a permanent move, but on the following year's survey they were again caring for their loved one in the community, we did not code the patient as institutionalized. In addition, we cross-checked the institutionalization data with the VA Beneficiary Identification and Records Locator Subsystem death files and Extended Care databases. We excluded patients from this sample if they died within 30 days of entering a nursing home, or were found through the Extended Care database to have stayed fewer than 90 days.
Measures
We selected from the survey those indicators of caregiver and patient physical and emotional functioning that related to the reasons for institutionalization (Table 1). Indicator domains included (a) caregiver sociodemographics, (b) caregiver health, (c) patient behavior, (d) patient physical functioning, (e) need for skilled care, (f) caregiver social support, and (g) caregiver emotional health. Cronbach's alphas, calculated for the NLCS baseline data, are provided in parentheses for the scale measures.
|
We estimated caregiver physical health by using self-rated health on a 4-point scale from poor to excellent, and three measures covering the previous 6 months: (a) the number of days too sick to perform caregiving activities, (b) the number of doctors' visits, and (c) the number of days spent in the hospital. We also included the number of medications taken by the caregivers, the number of comorbid illnesses, and the extent to which these illnesses interfered with caregivers' usual activities (3-point scale ranging from 1 = not at all to 3 = a great deal).
We measured patient dementia-related problem behavior with the Behavior Rating Scale–Dementia (BRS-D), which assesses behavioral problems and psychiatric symptoms of individuals with no history of mental retardation but with some degree of cognitive impairment (
= 0.70). Six subscales measured irritability or aggression ( = 0.79), behavior dysregulation ( = 0.66), depressive symptoms ( = 0.74), inertia ( = 0.57), vegetative symptoms ( = 0.48), and psychotic symptoms ( = 0.86; see Mack, Patterson, & Tariot, 1999). For each item, caregivers reported the frequency of the problem behavior in the previous month (higher scores indicate greater frequencies).
We estimated patient physical functioning by using the seven-item instrumental activities of daily living (IADL) subscale ( = 0.89) and the seven-item physical activities of daily living (ADL) subscale ( = 0.93) of the OARS Multidimensional Functional Assessment questionnaire (Fillenbaum, 1988). We used three of six items originally identified by Rosow and Breslau (1966) and used by the Established Populations for Epidemiologic Studies of the Elderly (known as EPESE) as an estimate of advanced ADLs. These included performing heavy household work, walking up and down a flight of stairs, and walking half a mile ( = 0.75; see Rosow & Breslau; Smith et al., 1990). Caregivers were asked to rate each item (ADLs, IADLs, Rosow–Breslau) on a 3-point scale (1 = needs no help, 2 = needs some help, or 3 = unable to do).
Indicators that the patient needed skilled care were the number of illnesses present on a modified OARS comorbidity scale, whether the caregivers received help from home health care, the time spent supporting deficits in patient ADLs, and the number of tasks (0–22) caregivers performed for the patient (yes, no, or very rarely; Clipp & Moore, 1995). The modified OARS comorbidity scale was based on the original 26 medical conditions to which depression, need for tube feedings, incontinence of bowel or bladder, and leg amputations were added, and to which cerebral palsy, muscular dystrophy, and speech impediment were removed (Fillenbaum, 1988).
We estimated instrumental social support by using a multi-item scale from the Duke Social Support Index ( = 0.87). Caregivers were asked to report the frequency with which family and friends provided 13 specific types of assistance (4-point scale from 1 = never to 4 = regularly). Services included helping when sick; helping with shopping or running errands; helping with money, meals, household repairs, and housework; giving financial or personal advice; providing transportation; giving gifts; listening to problems; and providing companionship for the patient (Landerman, George, Campbell, & Blazer, 1989). A measure of tasks and time provided by others was a multi-item scale of how often secondary helpers assisted with 18 tasks (4-point scale from 1 = never to 4 = regularly). The specific tasks included bathing, eating, dressing, toileting, grooming, transferring from one spot to another, walking inside and outside, cooking, doing housework, doing laundry, shopping, driving, taking medication, providing companionship, making repairs, doing outside work, and dealing with social agencies.
We measured subjective social support by using a 10-item scale from the Duke Social Support Index ( = 0.64) that requested information on caregivers' satisfaction with their relationships; whether caregivers felt lonely, understood, useful, listened to, included in the lives of family and friends, and as if they had a definite role in the family and among friends; and whether the caregivers had someone they could count on in times of trouble, and someone with whom they could talk about their deepest problems. The final item in the subjective-support scale was a single item concerning caregivers' desires for more help from family and friends (3-point scale from 1 = very much to 3 = not at all; George & Gwyther, 1986). We summed responses to yield a total score, with higher scores representing greater subjective support (Landerman et al., 1989). Also included was a single item measuring whether there were other people who helped care for the patient.
We measured caregiver emotional health by using the Center for Epidemiologic Studies–Depression scale short form, in which the caregivers rate the presence or absence of 20 feelings in the previous week ( = 0.72), a single-item life-satisfaction measure (3-point scale from not satisfying to very satisfying; Bradburn, 1969), and a 22-item emotional symptom scale ( = 0.78). We derived the 22-item emotional symptom scale, which summed the presence or absence of stress symptoms, from the Short Psychiatric Evaluation Schedule (Pfeiffer, 1979) and the Twenty-Two Item Screening Scale (Langer, 1962).
Statistical Analyses
Initially, we examined bivariate relationships between caregivers' reasons for institutionalization and indicators of caregiver and patient physical and emotional functioning from the survey completed in the year prior to placement. Using t tests for continuous variables and chi-square for categorical variables, we compared caregivers who chose a reason with those who did not choose that reason across the corresponding indicators. We analyzed caregiver sociodemographic and emotional health covariates with each reason, because the caregivers' emotional well-being may have influenced the selection; in addition, spousal caregivers, older caregivers, and caregivers of different racial backgrounds may have chosen different reasons. The fifth reason, "family or friends thought it was the best thing to do," did not correspond closely with any survey indicators and thus was reported only as the number of caregivers who endorsed this reason.
To determine the independent associations between caregiver and patient characteristics and institutionalization, we performed a multivariate analysis using a proportional hazards model (Cox, 1972). This powerful method incorporates time to the event (e.g., nursing home placement) rather than simply whether or not the event occurs. The method allows researchers to look at the relationship between the time to the event and other variables associated with it. We incorporated the variables with statistically significant bivariate association with institutionalization into a stepwise proportional hazards model, with institutionalization as the dependent outcome. The model included variables derived from each of the seven domains (see Table 1). The multivariate analysis included the fixed covariates of race and the caregiver–patient relationship (spouse or other). All other covariates were time varying. The variable for the number of minutes per day that caregivers spent supporting deficits in patients' ADLs was not normally distributed; therefore, we dichotomized it at the median (23.0) and entered it into the model as a categorical variable. We conducted all analyses in SAS 9.1 (SAS Institute, 2004).
|
Results |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
|
We performed a bivariate analysis to determine caregiver and patient indicators that were associated with the reasons given for institutionalization. For every reason, there was at least one indicator that differentiated caregivers who endorsed the reason from those who did not. Thus, indicators that supported the ultimate decision were present in the year before the decision was reached (Table 3). For example, caregivers who indicated that their health was the impetus for institutionalization were more likely to have lower self-rated health, more visits to the doctor, more sick days, more medications, and more comorbid illnesses. In addition, caregivers who cited their health as a reason were more likely to be a spouse (96.4%; 81.1%; p <.0001), older (71.1 ± 7.4; 68.0 ± 10.9; p =.0001), have a lower income (23,419 ± 14,027; 27,290 ± 16,831; p =.004), lower life satisfaction, and higher stress symptoms. When caregivers cited patient behavior as the reason, their patients were more likely to be their spouses (93.8%; 84.0%; p =.0003) and to have received higher scores on four of the BRS-D subscales: behavior dysregulation, irritability or aggression symptoms, inertia, and psychotic symptoms. Caregivers reporting the need for more assistance from others as their reason were younger than those who did not choose that reason (68.0 ± 9.7; 70.0 ± 9.4; p =.036), and they also reported lower perceptions of social support and desire for more help from family or friends in the year prior to placement. Caregivers reporting the need for skilled care were caring for patients with higher scores on the Rosow–Breslau scale, suggesting deficits in lower body strength.
|
|
|
Discussion |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
The most commonly cited reasons for institutionalization included caregivers' perceptions that skilled care was needed, that their own health would not allow them to continue, and that patient dementia-related problem behaviors were too difficult to handle. Each of these reasons related significantly to indicators of caregiver and patient characteristics from the prior year's survey, suggesting that caregivers' reasons for institutionalizing have valid underpinnings. This effort to substantiate caregivers' reasons for placement is important, because it informs clinicians that a majority of caregivers can assess their situation objectively. In addition, more clearly understanding the needs and perceptions of caregivers should enable clinicians to intervene and meet these needs, perhaps delaying institutionalization.
The most common reason chosen, "my loved one needed more advanced or skilled care than I could provide," related significantly to the care recipients' scores on the Rosow–Breslau scale. This measure is similar to the scale used by Wolinsky and colleagues (1992, 1993) to measure lower body limitations, which significantly predicted institutionalization. This result is logical, as the inability of caregivers to lift nonambulatory patients may be more difficult to offset than other functional limitations.
Previous studies that examined reasons for nursing home admission collected data from a single interview and described more general information than that revealed in the present study (Arling & McAuley, 1983; Nolan & Dellasega, 2000; Smallegan, 1985; Wingard et al., 1990). Although these studies included patients with various diagnoses, the results reflected ours in that the most common reasons for placement involved the care recipient's declining health and function, which became too much for the caregiver to handle. Compared with these studies, the current study offers several advantages: (a) the large national NLCS sample; (b) the longitudinal nature of the NLCS, which permitted analysis of patient and caregiver characteristics prior to institutionalization; and (c) the restriction of the sample to older individuals with clinically diagnosed dementia who predominately were cared for by spouses. Spousal caregivers of persons with dementia may experience more trauma and guilt associated with institutionalization than those placing a partner because of a physical disability (Tilse, 1998).
The initial analysis of the caregivers' reasons was limited to caregivers who chose institutionalization. Using a proportional hazards model, we compared caregivers who institutionalized with those who continued caregiving in the community as a way to examine predictors of institutionalization. As in previous research, the results showed that race (Gaugler et al., 2003; Yaffe et al., 2002), patient behavioral problems (Gaugler et al., 2000, 2003; Yaffe et al.), and fewer patient comorbidities (Eaker, Vierkant, & Mickel, 2002) predicted institutionalization. The fact that increasing comorbidity was not a risk factor suggests that, rather than number of medical conditions, the timing of placement is determined by dementia severity and concomitant behavioral and dependency problems. Relatedly, those patients with a greater burden of comorbid illness may succumb to one of those illnesses before the dementia progresses to the point of requiring formal care.
The study also contributes to the caregiving literature by indicating more precisely which problem behaviors tend to lead to institutionalization. By entering the BRS-D subscales into the multivariate model instead of the total score, we demonstrated that only the behavior dysregulation and psychotic symptoms subscales reached significance. This suggests that compared to wandering or psychotic behaviors in their care recipients, caregivers are more able to manage depressed, agitated, or apathetic behaviors or alterations in the sleep or appetite patterns. Greater task burden and lower life satisfaction in the caregivers also were found to be associated with a greater risk of institutionalization. Few studies have included a measure of caregiving hours or tasks. At least one study (Gaugler, Kane, Kane, Clay, & Newcomer, 2005) assessed the number of caregiving hours, but it found no relation to institutionalization. In contrast, this study comprehensively assessed caregivers' task demands and found them to be a significant predictor. To our knowledge, the life-satisfaction variable used in our study has not been used in other studies of nursing home placement. However, caregiver life satisfaction likely relates to caregiver burden (Gaugler et al., 2003; Yaffe et al., 2002), self-perceived pressure (Pot et al., 2001), or feeling trapped in their caregiving role (Gaugler et al., 2000), all of which have been associated with increased risk of institutionalization. Clinically, it would be very useful to have a single question: "How would you say that you find life these days?" Results of this study suggest that a caregiver report of low life satisfaction could signal institutionalization within the next year.
Several limitations of this study should be considered in future research. First, self-reported data are considered less reliable than structured interviews and clinical examinations. In addition, mailed surveys are associated with relatively low response rates, potential for poor-quality data if the questions are easily misunderstood or open ended, and inability to confirm that the intended respondent filled out the survey. Furthermore, caregiver assessments of patient functional and behavioral status may be biased. This limitation may be less important because caregivers are the key decision makers regarding institutionalization and their perceptions form the basis of reality in this process. In addition, the study may not be generalizable to other caregiving scenarios because it focused on a male veteran sample predominantly cared for by female spouses (e.g., dementia-related problems behaviors may not predict institutionalization among male caregivers who are physically more able to handle disruptive behaviors). Finally, although not as homogeneous in race as some caregiving studies, the majority of NLCS caregivers were White, and extending conclusions to caregivers with diverse ethnic backgrounds should be done with caution.
Strengths of the study include the numerous valid and reliable measures available to relate to the major "caregiver reason" domains. Moreover, the NLCS is a large national level sample, followed longitudinally. To our knowledge, this is the first study to use such data to examine informal caregivers' reasons for institutionalizing their family members with dementia, and to compare these reported reasons with indicators of caregiver and patient physical and emotional functioning prior to institutionalization. Findings suggest that clinicians should be able to identify caregivers who are more likely than others to institutionalize their care recipients within a year. Examples of probes and direct questions to facilitate this process and engage these caregivers include the following:
If a caregiver is considering nursing home placement for one of the issues raised in these questions, results from this study suggest that the caregiver's reasons are reality based, and the caregiver should be encouraged to examine options for formal long-term care. The American Medical Association (AMA; n.d.) has developed a caregiver self-assessment questionnaire as part of its project to help primary care physicians identify and help patient caregivers (American Medical Association Caregiver Self-Assessment Questionnaire). This study supports the psychological emphasis of the AMA's questionnaire but also expands the questionnaire to include some patient characteristics. With the knowledge gained from this study, clinicians can offer support to caregivers to help diffuse some of the negative feelings of guilt, thus facilitating outcomes that are more positive for both caregiver and patient. Future studies should investigate the clinical impact of prompting such discussions with caregivers at appropriate times in the dementia disease course, with an overall goal of easing the burden associated with transitioning from home to formal care.
|
Footnotes |
|---|
1 Department of Medicine, Duke University Medical Center, Durham, NC.
2 Center for the Study of Aging and Human Development, Duke University Medical Center, Durham, NC.
3 Department of Biostatistics and Bioinformatics, Duke University Medical Center, Durham, NC.
4 School of Nursing, Duke University Medical Center, Durham, NC.
5 Geriatric Research Education and Clinical Center, Durham Veterans Affairs Medical Center, Durham, NC.
Decision Editor: Linda S. Noelker, PhD
Received for publication March 31, 2005. Accepted for publication August 22, 2005.
|
References |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. Iecovich Caregiving Burden, Community Services, and Quality of Life of Primary Caregivers of Frail Elderly Persons Journal of Applied Gerontology, June 1, 2008; 27(3): 309 - 330. [Abstract] [PDF]
|
|
|
|
 |
|
 C. K. Chen, M. Sabir, S. Zimmerman, J. Suitor, and K. Pillemer The Importance of Family Relationships With Nursing Facility Staff for Family Caregiver Burden and Depression J. Gerontol. B. Psychol. Sci. Soc. Sci., September 1, 2007; 62(5): P253 - P260. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D.-C. Chan, J. D. Kasper, B. S. Black, and P. V. Rabins Clinical Diagnosis of Dementia, Not Behavioral and Psychologic Symptoms, Is Associated With Psychotropic Drug Use in Community-Dwelling Elders Classified as Having Dementia J Geriatr Psychiatry Neurol, June 1, 2007; 20(2): 100 - 106. [Abstract] [PDF]
|
|
|
|
|
|
D.-C. Chan, J. D. Kasper, B. S. Black, and P. V. Rabins Clinical Diagnosis of Dementia, Not Presence of Behavioral and Psychological Symptoms, Is Associated With Psychotropic Use in Community-Dwelling Elders Classified as Having Dementia J Geriatr Psychiatry Neurol, March 1, 2007; 20(1): 50 - 57. [Abstract] [PDF]
|
|
|
|
 |
|
 R. L. Kane Ptolemaic bathing. J. Gerontol. A Biol. Sci. Med. Sci., August 1, 2006; 61(8): 819 - 820. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
