Factors associated with caregiver burden in patients with Alzheimer's disease.

OBJECTIVE: Caregivers for patients with Alzheimer's disease (AD) suffer from psychological and financial burdens. However, the results of the relationship between burden and cognitive function, performance of activities of daily living, and depressive symptoms have remained inconsistent. Therefore, the aim of this study was to examine which factors are more significant predictors of heightened burden, cognitive impairment or functional decline, besides neuropsychiatric symptoms.
METHODS: A cross-sectional study was conducted in a sample comprised of 1,164 pairs of patients with AD and caregivers from the Clinical Research of Dementia of South Korea study cohorts. The cognitive function of each sub-domain, functional impairments, depressive symptoms, and caregiver burden were assessed using the dementia version of Seoul Neuropsychological Screening Battery (SNSB-D), Barthel Index for Daily Living Activities (ADL), Seoul-Instrumental Activities of Daily Living (S-IADL), the Clinical Dementia Rating Sum of Box (CDR-SB), the Global Deterioration Scale (GDS), the Korean version of the Neuropsychiatric Inventory (K-NPI), and the 15-item Geriatric Depression Scale.
RESULTS: We found that higher severity (higher CDR-SB and GDS scores) and more functional impairment (lower ADL and higher S-IADL scores) were significantly associated with higher caregiver burden. In addition, depressive symptoms of patients (higher Geriatric Depression Scale scores) were associated with higher caregiver burden.
CONCLUSION: Therefore, interventions to help maintain activities of daily living in patients with AD may alleviate caregiver burden and improve caregiver well-being.

INTRODUCTION

Alzheimer's disease (AD) is a progressive and degenerative disease of the brain that causes multiple cognitive areas to become seriously impaired and results in a decline in functional abilities and behavioral changes. The prevalence of AD among persons aged >65 years is 2.7-11.2%, and the annual costs for AD are $20 to $100 billion in the United States.1,2 As the prevalence and costs of AD grow rapidly, the burden to family, society, and the nation increases.

In particular, primary caregivers frequently suffer from enormous psychological and financial stress. This stress makes caregivers choose to institutionalize patients with AD earlier and, consequently, the burden on social and health services increases.3,4 Thus, it is important to focus on caregiver burden, because it increases social costs and the individual burden of dementia.

Several studies have been conducted with the aim of identifying factors related to caregiver burden. Mohamed et al.3 found that severity of psychiatric symptoms, behavioral disturbances, and patient quality of life have a stronger association with caregiver-experienced burden. Kim et al.5 reported that caregiver-related factors, such as gender, history of home care, and education level are correlated with caregiver burden. Additionally, patient-related factors such as performance of activities of daily living (ADLs) have been related to burden.

Although many studies have tried to show the factors associated with caregiver burden, there is a lack of comprehensive studies that have investigated which factor is the most significant predictor of caregiver burden. Many studies have focused on caregiver or patient-related sociodemographic characteristics,4,5,6 but few studies have conducted comprehensive neuropsychological testing and examined cognitive function subdomains to identify the association between patient cognitive function and caregiver burden.

We designed this study to explore specific factors that are associated with caregiver burden and to integrate current findings. In this study, we examined whether cognitive dysfunction or functional impairment is more related to caregiver burden. In addition, we examined the association among caregiver burden and demographic variables of patients and caregivers, patient depressive symptoms, insight, and underlying medical conditions using multivariate analysis in a large cohort of multicenter patients.

METHODS

Subjects and study design

The Clinical Research of Dementia of South Korea (CREDOS) study is a 9-year longitudinal, nationwide, hospital-based study that has been carried out in 56 hospitals. This study had a cross-sectional design. The sample comprised 1,164 patients diagnosed with incident AD and 1,164 caregivers. This study was based on baseline data derived from the CREDOS study from February 2005 to February 2011. All patients were diagnosed with AD or probable AD using the criteria of the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA).7

Patients were excluded if they had any neurodegenerative disease other than AD (dementia with Lewy bodies, Parkinson's disease, Huntington's disease, etc.). Patients with mild cognitive impairment were also excluded. Subjects with a severe medical condition (e.g., uncontrolled diabetes mellitus, severe liver failure, severe arrhythmia, and end-stage renal disease) were excluded due to safety issues. Additionally, we excluded subjects with significant hearing or visual impairments that rendered interviews difficult. Each caregiver met the following inclusion criteria: 1) was a relative of the patient; 2) was familiar with the patient's daily activities; and 3) agreed to be interviewed.

This study protocol was approved by the Institutional Review Boards of all participating hospitals (#2005-02-008). Signed informed consent was obtained from both caregivers and patients.

Dementia assessment

The clinical workup included complete medical histories, physical and neurological evaluations, routine laboratory tests, including complete blood counts, blood chemistry profiles, vitamin B12/folate levels, syphilis serology, thyroid function tests, and a brain magnetic resonance imaging (MRI) evaluation. Genotyping for the APOE polymorphism was undertaken in 733 subjects. Brain MRI scans included transaxial T2, T1-weighted scans, and fluid-attenuated inversion recovery slices. CREDOS rating scales were employed to assess white matter changes.8 Patients were classified into three groups (mild, moderate, and severe) based on their white matter changes around the lateral ventricles or in deep white matter.

All patients underwent a standardized neuropsychological battery (the dementia version of Seoul Neuropsychological Screening Battery, SNSB-D) (Table 1).9 The SNSB-D consists of sub-domains that assess attention (score of 0-17), language and related function (score of 0-27), visuospatial function (score of 0-36), memory (score of 0-150) and frontal/executive function (score of 0-70). Higher scores indicate intact function in all sub-domains. The Barthel Index for Daily Living Activities (rating 0-20, higher scores indicate more independence) was used to evaluate basic ADLs.10 We used the Seoul-Instrumental Activities of Daily Living (S-IADL) for the instrumental ADL tasks.11,12 The S-IADL was developed to assess patient ability to perform social and instrumental ADLs. All items were scored on a 4-point scale ranging from 0 to 3, and the possible score ranged from 0 to 24. A lower score indicates better functioning. The Korean version of the Neuropsychiatric Inventory (K-NPI) was administrated to the caregivers of the subjects.13 The K-NPI evaluates 12 domains of common behavioral and psychological symptoms, including delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, euphoria/elation, apathy/indifference, disinhibition, irritability/lability, aberrant motor behavior, sleep/nighttime disturbances, and appetite/eating abnormalities. The caregivers were asked about symptoms that had occurred within the preceding 4 weeks. After the questions were answered, the caregivers were asked about symptom frequency (scale of 1 to 4, higher scores indicate more frequent) as well as severity (scale of 1 to 3, higher scores indicate more severe). The composite score (rating, 0-12) for each domain was calculated by multiplying frequency by severity. A total score (range, 0-144) was calculated by adding the 12 composite scores. Depressed mood was assessed using the 15-item Geriatric Depression Scale,14 and the Korean version of the Mini-Mental State Examination (K-MMSE) was used to assess cognitive impairment.15,16 Global severity of disease was assessed using the Clinical Dementia Rating Sum of Box (CDR-SB)17 and the Global Deterioration Scale (GDS).18

Table 1

Construction of the SNSB-D

SNSB-D: Seoul Neuropsychological Screening Battery-Dementia Version, K-BNT: Korean-Boston Naming Test, RCFT: Rey-Complex Figure Test, SVLT: Seoul Verbal Learning Test, GCF: global cognitive function

Caregiver assessment

Caregiver burden was evaluated using the Neuropsychiatric Inventory Caregiver Distress Scale (NPI-D).19 The NPI-D was developed to provide a quantitative measure of distress experienced by caregivers in relation to the individual symptom domains assessed by the K-NPI. After rating each symptom item of the K-NPI, caregivers were asked to rate their distress (range, 0-5, higher score indicates more distress). The NPID has a moderate correlation (r=0.51, p<0.01) with the Zarit Caregiver Burden Inventory (ZBI), which is a widely used assessment scale for measuring caregiver perceived burden as reported by caregivers.20,21

Data analysis

Continuous variables that were not normally distributed are presented as medians and interquartile ranges. Non-parametric methods were applied to analyze these variables. We employed the Wilcoxon rank-sums test or the Kruskal-Wallis test to compare groups of continuous variables. Spearman's rank correlation test was used to test correlations between continuous variables. Median regression was employed to test the multivariate association between NPI-D and the dependent variables. We examined potential collinearity by using Variance inflation factor (VIF) scores, and we did not detect any variables with collinearity (VIF>5). Results were considered significant at p<0.05. p-values were corrected using Bonferroni's correction if necessary.

RESULTS

Subject characteristics

The clinical and demographic characteristics of the patients are shown in Table 2. The median age of the patients was 74 years (interquartile range, 68-79 years). We recruited 445 (38.1%) male patients and 719 (62.8%) female patients (male: female, 1:1.62). Almost all of the patients were educated more than 6 years (median, 7 years; interquartile range, 6-12 years). The median K-NPI and K-MMSE scores were 7 (interquartile range, 2-18) and 20 (interquartile range, 17-23), respectively. The median GDS score was 4 (interquartile range, 4-5), comparable with moderate cognitive decline. The median age of the caregivers was 51 years (interquartile range, 42-66 years), and they consisted of 402 males (34.5%) and 762 females (65.5%). The median NPI-D score was 4 (interquartile range, 1-12). In total, 33% of patient caregivers were spouses, 20.6% were sons, 28.5% were daughters, and 17.5% were others.

Table 2

Clinical and demographic characteristics of the subjects

Continuous variables that were not normally distributed are presented as medians and interquartile ranges. ADL: activities of daily living, S-IADL: Seoul-Instrumental Activities of Daily Living, K-NPI: Korean Neuropsychiatric Inventory, K-MMSE: Korean Mini Mental State Examination, CDR-SB: Clinical Dementia Rating Sum of Boxes, GDS: Global Deterioration Scale, NPI-D: Neuropsychiatric Inventory-caregiver Distress Scale

Univariate analysis for the factors associated with caregiver burden

No notable association with caregiver burden was found for patient age, patient gender, patient education level, or current alcohol drinking/smoking in patients (Table 3).

Table 3

Correlations among clinical and demographic characteristics of patients, caregivers, and caregiver burden (NPI-D)

Continuous variables that were not normally distributed are presented as medians and interquartile ranges. *Spearman's rank correlation test, †Wilcoxon rank-sum test, ‡Kruskal-Wallis test. NPI-D: Neuropsychiatric Inventory-caregiver Distress Scale, CDR-SB: Clinical Dementia Rating Sum of Boxes, ADL: activities of daily living, S-IADL: Seoul-Instrumental Activities of Daily Living, K-MMSE: Korean Mini Mental State Examination, GDS: Global Deterioration Scale, APOE genotype: apolipoprotein E genotype, SNSB-D: Seoul Neuropsychological Screening Battery-Dementia version

The patient's ability to perform ADLs was highly associated with caregiver burden. The caregivers of more dependent patients had higher NPI-D scores. The scales for dementia severity (CDR-SB and GDS) were correlated with higher caregiver burden. The Geriatric Depression Scale showed a positive correlation with NPI-D. In addition, patient insight for their disease was associated with caregiver burden. The impaired insight of patients was associated with higher caregiver burden (impaired insight group vs. partial insight group: corrected p=1.45×10-4; impaired insight group vs. intact insight group: corrected p=0.02). No significant difference was observed for the NPI-D score between the partial insight and intact insight groups (corrected p=0.11). Patient cognitive function was associated with caregiver burden. In the SNSB-D test, memory function and frontal and executive function were more significantly associated with caregiver burden than the other sub-domain scores. Among the underlying medical histories of patients, heart disease, renal disease, lung disease, cancer, and epilepsy were associated with caregiver burden (Table 3).

Among the caregiver-related variables, age, gender and the relationship with the patient were significantly associated with NPI-D score. However, we did not find a significant association between living with the caregiver and NPI-D score in this analysis.

Neither apolipoprotein E genotype nor brain ischemic changes were significantly associated with NPI-D score. Additionally, we did not find an association between neurological deficit (motor weakness, sensory loss, or parkinsonian features) and caregiver burden (Table 3).

Median regression analysis for factors associated with caregiver burden

We tested a multivariable model that included cognitive variables (SNSB-D sub-domains) and functional variables (ADL and S-IADL) as representative variables of cognitive parts and functional parts of dementia severity, respectively. In this test we adjusted for demographic characteristics, the Geriatric Depression Scale score, insight of patients, and underlying medical history, which were significant in the univariate analysis.

The results of the multivariate analysis are shown in Table 4. The more dependent patients (patients with lower ADL scores or higher S-IADL scores) had higher caregiver burden. In contrast with the result of the functional part of dementia severity, cognitive impairment in patients was not significantly associated with caregiver burden. None of the SNSB-D subdomains were significantly associated with caregiver burden.

Table 4

Median regression analysis for the factors associated with caregiver burden (NPI-D)

NPI-D: Neuropsychiatric Inventory-caregiver Distress Scale, ADL: activities of daily living, S-IADL: Seoul-Instrumental activities of daily living

A higher Geriatric Depression Scale score was significantly associated with higher caregiver burden in both models. Among underlying medical conditions, renal disease was significantly associated with a higher NPI-D score. Caregiver age was significantly associated with caregiver burden. Patient insight was not significantly associated with caregiver burden.

DISCUSSION

This study examined which factors were more significant predictors of heightened burden, cognitive impairment, and functional decline. The results indicated that caregiver burden had a stronger association with functional impairment than that of cognitive functioning. In the correlation analysis, both ADLs and cognitive function were associated with caregiver burden. However, only functional ability was related to caregiver burden after controlling for other variables. In previous studies, the correlation between caregiver burden and patient ADLs and cognitive function has been controversial. For example, Mohamed et al.3 found only limited relationships among cognitive function, ADLs, and caregiver burden. Additionally, Coen et al.22 reported that neither cognitive function nor functional impairment in patients with AD predicts caregiver burden, whereas Huang et al.23 found that poorer cognitive function and instrumental ADLs and ADL performance are associated with higher caregiver burden. Therefore, we controlled for other variables that could significantly impact caregiver burden to identify the association between functional ability and cognitive function. Our findings support that functional impairment of patients with AD was more related to caregiver burden.

Christofoletti et al.24 also suggested that caregivers of physically active patients with AD exhibit reduced burden. They noted that patients with dementia require increasing supervision and personal care depending on their functional ability, and that this results in negative effects on caregiver health. Therefore, interventions to help maintain ADLs in patients with AD can decrease caregiver burden compared to those that focus on cognitive decline. For example, Rolland et al.25 suggested that a collective exercise program leads to improved ability to perform ADLs, and that this could reduce caregiver burden.

In the present study, depressive symptoms of patients were highly associated with increased caregiver burden. This finding is similar to those of previous studies.3 Many studies have suggested that depressive symptoms are common in patients with dementia.26,27 Karttuten et al.28 evaluated 240 patients with AD and showed that 32.1% of patients with very mild and 39.6% of patients with mild AD have depressive symptoms. Enache et al.29 reported that 20-30% of patients with AD have depression.

Diagnosing depression in patients with dementia is important for adequate patient management, and depression of patients is significantly related to caregiver burden. However, diagnosing depression correctly may be complicated due to overlap of symptoms between dementia and depression.29 Therefore, clinicians should take more notice of depressive symptoms in patients with dementia.

The results of the present study indicate that dementia patient and caregiver demographic factors did not significantly impact caregiver burden except the caregiver age was related to caregiver burden, indicating that the younger the caregiver, the greater the burden. This result is inconsistent with findings from some studies. Lim et al.30 reported that older caregivers have much more caregiver burden, because older caregivers have more stress from multiple roles than that of young caregivers. In contrast, the results of Hope et al.31 were consistent with ours. They found that younger caregivers are less committed to care because they may have jobs or greater demands from other family members. Additionally, Fitting et al.32 reported that younger caregivers are lonelier and more resentful of their role. In particular, Koreans have more sense of responsibility to care for elderly family members due to the influence of filial piety. Thus, our findings suggest that less emotional and social support in younger caregivers leads to a greater burden.

We tested the association between caregiver burden and patient underlying medical history. The results showed that renal disease was associated with high caregiver burden in patients with AD. Renal disease is a representative chronic physical illness and caregivers of patients with chronic mental or physical illnesses experience significant distress and psychological burden.33 Our results indicate that the co-morbidity of medical illness in patients with AD could place additional burden on caregivers. The non-significant results of other medical history may be due to the limitation of our data, because we were unable to consider the severity, specific diagnosis, or clinical stage of underlying medical conditions.

Several limitations should be noted in this study. It was a cross-sectional study, not a longitudinal study; thus, we could not draw conclusions regarding causality. A longitudinal study may have been more appropriate for an accurate assessment of predictors related to caregiver burden. Another limitation is that the majority of the patients were early stage AD. Therefore, the findings may not be generalizable to more severely demented patients. In addition, we used the NPI-D to evaluate caregiver burden. This is an adjunct scale to the NPI for assessing the impact of neuropsychiatric symptoms on caregiver burden. However, this scale has a moderate correlation with the ZBI, which is the most widely used instrument for assessing caregiver distress.20 Despite these limitations, we think our findings are quite reliable, because we examined patients with AD in a large multicenter cohort, which was assessed comprehensively including neuropsychological testing, brain MRI, and diverse underlying medical history.

In conclusion, caregiver burden had a stronger association with functional impairment than that of cognitive functioning. Therefore, interventions to help maintain ADLs in patients with AD may alleviate caregiver burden and improve caregiver well-being.