Comparing Behavioral and Psychological Symptoms of Dementia and Caregiver Distress Caused Between Older Adults With Dementia Living in the Community and in Nursing Homes

Objectives To investigate differences in behavioral and psychological symptoms of dementia (BPSD) and caregiver distress caused between older adults with dementia living in the community and in nursing homes. Design A comparative cross-sectional study. Setting and Participants Participants were recruited from outpatient clinics of a tertiary psychiatric hospital and dementia units of a nursing home in Guangzhou, China. Methods Neuropsychiatric Inventory was used to assess symptoms and caregiver distress. Dementia severity was determined using the Clinical Dementia Rating. Results This study included 157 community and 112 nursing home residents with dementia. Clinically significant symptoms (item score ≥ 4) were found in 88.5% of the former and 75% of the latter. Caregivers of 79.6% of the former and 26.8% of the latter reported that at least one of these caused them moderate-to-severe distress (distress score ≥ 3). Among the community patients, anxiety was the most frequent “very severe” symptom, while sleep disorders and agitation caused the most frequent “very severe” caregiver distress. After controlling for dementia severity and medication use, family caregiving remained an independent risk predictor for clinically significant symptoms and moderate-to-severe caregiver distress. The prediction of caregiver distress based on symptom scores varied across caregiver types and individual symptoms (R2 0.36–0.82). Group differences in clinically significant symptoms and moderate-to-severe caregiver distress showed at the stage of moderate-to-severe dementia. Conclusions and Implications Tailored management strategies to relieve family caregivers’ BPSD-induced distress are needed, especially at the stage of moderate-to-severe dementia. An effective service system should be established for supporting family caregivers to cope with BPSD.


INTRODUCTION
Dementia affects approximately 55.2 million people and their families worldwide in 2019, and this is expected to rise to 139 million by 2050 (1). Dementia care may be a long journey characterized by significant neuropsychiatric symptoms, caregiver distress, and unmet needs (2). Behavioral and psychological symptoms of dementia (BPSD) are common noncognitive impairments occurring at all stages of dementia (3). Around 80%-90% of people with dementia have at least one of these symptoms (4,5). One of the greatest challenges in caring for people with dementia is managing these symptoms. Difficulty in doing so often leads to higher caregiver burden (6), impaired quality of life (7,8), increased healthcare utilization (9,10), and risk of institutionalization (11). The problem is likely to be even greater within the community (12).
Family caregivers are mainly responsible for managing BPSD of people with dementia living at home (13)(14)(15), but typically possess only limited knowledge about symptoms and how to handle them (16). The number of people with dementia in China is estimated to exceed 9 million and is expected to reach 40 million by 2050 (17). Around 90% of them live at home and are looked after by their families (18). Dementia care services and related support services are also still developing in China (2). The lack of dementia-related training and community-based support services designed for people with dementia and their families can lead to exhaustion among family caregivers (18). Family caregivers in low-and middle-income countries often experience a higher level of distress due to the lack of resources and support mechanisms in the public healthcare system to help them properly managing BPSD (15,19).
By comparison with their counterparts in long-term care facilities, we may get a better appreciation of clinically significant and moderate-to-severe distressing BPSD and their risk predictors in community-dwelling older adults with dementia cared by family caregivers. This is necessary to generate research evidence to inform the design of tailored management strategies and support family caregivers to enable people with dementia to stay at home longer, especially in lowand middle-income countries. Previous studies have investigated differences in caregiver distress caused by BPSD between community-dwelling older adults and residents of long-term care facilities (15,(20)(21)(22). Cheng and colleagues found that BPSD caused more distress to family caregivers than formal caregivers (20). On the contrary, Loi and colleagues suggest that formal caregivers may be exposed to more severe and frequent BPSD than informal caregivers (22). The discrepancies between the studies suggest that BPSD-induced distress may influence by a myriad of factors (such as country/culture, long-term care systems, outcome measures used, and sample characteristics). Especially, the use of samples with different severity of dementia, as well as a potential influence of medication used by participants, may lead to significant differences in the assessed severity of symptoms and the caregiver distress caused. We hypothesized that after controlling for dementia severity, as well as medication use, we might clarify the difference in BPSD and caregiver distress caused between community-dwelling older adults with dementia and nursing home residents with dementia.
This study therefore aimed to explore differences in BPSD and caregiver distress in responding to individual symptoms between community-dwelling older adults with dementia and their nursing home counterparts after controlling for dementia severity and medication use, and explore risk predictors of clinically significant symptoms and moderate-to-severe caregiver distress caused in the target population.

Study Design
A comparative cross-sectional study.

Participants
Between March 2019 and September 2019, community-dwelling older adults with dementia and their primary family caregivers were recruited from the outpatient clinics of a 1920-bed tertiary psychiatric hospital in Guangzhou (the capital city of Guangdong province in China) using convenience sampling. The hospital provides neuropsychiatric services for 800 thousand outpatients per year. The inclusion criteria for community-dwelling older adults with dementia were (1) aged ≥ 60 years, lived at home and with support from family members; (2) had a physician diagnosis of dementia, following expert assessments. The exclusion criteria for older adults with dementia were (1) having past or present comorbidity of another major psychotic disorder, such as schizophrenia and bipolar disorders; (2) having a terminal illness (with life expectancy < 6 months); or (3) having significant variations in dosage of psychotropic medications during the past four weeks. The inclusion criteria for family caregivers were as follows: (1) the primary family caregiver of the patient; (2) had been in the role for at least 4 h per day during the past 3 months; (3) was able to read and speak Mandarin, and cognitively intact. The family caregiver who refused to participate was excluded. Nursing home residents with dementia were recruited from dementia units of nursing homes in Guangzhou, China. Eligibility for nursing home residents with dementia and their caregivers (the primary caregivers of the residents were formal caregivers who had the most day-to-day contact with the residents) has been described elsewhere (23).
Diagnosis of dementia was made using the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (24). The diagnosis of Alzheimer's disease (AD) was based on the NINCDS-ADRDA criteria (25), vascular dementia (VaD) was based on the NINDS-AIREN criteria (26), dementia with Lewy bodies (DLB) was based on DLB consensus criteria (27), and FTD was based on FTD consensus criteria (28).

Severity of Dementia
Severity of dementia was assessed using the Clinical Dementia Rating (CDR) (29). Clinical Dementia Rating is a global, informant-based, structured interview to determine the presence and severity of dementia. Information was collected from the persons with dementia and/or their informants about their performance in the following six domains: memory, orientation, judgment and problem-solving, community affairs, hobbies, and personal care. Each domain is rated on a 5-point scale (0 = none, 0.5 = questionable, 1 = mild, 2 = moderate, 3 = severe). A global rating of 1 (mild dementia), 2 (moderate dementia), and 3 (severe dementia) were used.

Behavioral and Psychological Symptoms of Dementia and Caregiver Distress Caused by Them
The Neuropsychiatric Inventory (NPI) was used to interview family caregivers. The NPI-Nursing Home Version (NPI-NH) was used to interview formal caregivers enrolled from nursing homes (23). Both NPI and NPI-NH examine 12 sub-domains of behavior functioning: delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/liability, aberrant motor behaviors, sleep, and appetite and eating disorders. The 12 items were registered as present or not present during the past 4 weeks. If present, they were scored by frequency (1-4), severity (1)(2)(3), and level of caregiver distress (0 = not distressing to 5 = extremely). A higher score indicated more severe symptoms and a higher level of caregiver distress. We used the established cut-off item score (frequency × severity) of ≥ 4 to indicate a clinically significant symptom (21,30,31).

Basic Activities of Daily Living
Basic activities of daily living (BADL) were measured using the Barthel index (BI). We used a BI score of 0-20 to indicate total dependency, 21-60 for severe dependence, 61-90 for moderate dependence, and 91-99 for slight dependence (32).

Data Collection
Clinical Dementia Rating (CDR) was rated by psychiatrists who worked in the tertiary psychiatric hospital and the nursing home from where the participants were enrolled. Primary caregivers were asked to complete the NPI-NH, and demographic information about care recipients was also collected through face-to-face interviews with the caregivers. The interviewers were three graduate students majoring in nursing, who received training in data collection, including all procedures and content of the assessments. Clinical data were obtained from medical records. Data about the actual use of medications were also collected based on the patients' medical records. Medication use was classified according to the Anatomical Therapeutic Chemical classification, including anti-dementia drugs, antipsychotics, anxiolytics, anticonvulsants, and antidepressants.

Sample Size Calculation
The sample size was calculated using a formula for calculating sample size in a prevalence study (33). The allowable error was assumed to be 10%. The level of confidence was set at 95%. Previous studies report the prevalence of BPSD ranged from 50.1 to 99% in Chinese community-dwelling persons with dementia (15,21), and from 63.6 to 92.9% in Chinese nursing home residents with dementia (4,23). According to the lowest prevalence rate of BPSD in previous studies, the sample size of this study should be at least 96 cases at each group. Considering 15% invalid questionnaires, 110 samples were required for this study, and 157 samples were finally collected from a community and 112 from a nursing home.

Statistical Analysis
Statistical analysis used SPSS 25.0 (IBM SPSS Statistics, IBM Corp., Armonk, NY). Continuous variables were described as mean and standard deviation (SD), and categorical variables as number (n) and percentage (%). Non-parametric analyses were used when the data showed a non-normal distribution. Betweengroup comparisons used the Chi-square test (with Fisher's exact test for multiple analyses) for categorical variables and Mann-Whitney U tests for continuous variables. Linear regression was conducted with continuous caregiver distress scores as a dependent variable and NPI item scores as independent variables. The NPI item score was dichotomized into ≥ 4 and < 4 to identify clinically significant symptoms, and caregiver distress score was dichotomized into ≥ 3 and < 3 for moderateto-severe caregiver distress. Multivariate logistic regression (forward: conditionally) was used to test whether the odds of a clinically significant symptom and moderate-to-severe caregiver distress differed by CDR, caregiver type (family versus formal), medication use (anxiolytics, anticonvulsants, anti-depressants, anti-dementia, and antipsychotics), physical dependence, age, sex, and education. An additional independent variable for the latter was clinical significance of the symptoms (item score ≥ 4). As for controlling for the medication use, first, we compared medication use between the community-dwelling older adults with dementia and nursing home residents with dementia. And then, the multivariate logistic regressions above-mentioned were used in the full sample. A P value of less than 0.05 was considered statistically significant.

Ethical Considerations
The Ethics Committee of N Hospital (No. NFEC-201511-K2) approved this study. Participants or their legal representatives were given a written and verbal explanation about the study, the benefits, and any risks. Written informed consent was obtained from all participants or their legal representatives. All personal information was kept confidential and anonymous as required.

RESULTS
The study included 157 community-dwelling older adults with dementia [age: 76. 8   They were also more likely to use anxiolytics, anti-depressants, and anticonvulsants ( Table 1).

Predictors for Caregiver Distress Caused by Behavioral and Psychological Symptoms of Dementia
After controlling for dementia severity, medication use, and symptom severity, family caregiving remained an independent risk predictor for moderate-to-severe caregiver distress caused by BPSD (OR 2.11∼18.16), especially eating disorders and anxiety ( Table 2).
Linear regression analysis shows that caregiver distress was highly predicted by NPI item scores (R 2 0.63-0.82), except for apathy, euphoria, disinhibition, hallucination, and eating disorders (Figure 2). The apathy score did not highly predict caregiver distress for both the community (R 2 = 0.37) and nursing home patients (R 2 = 0.46). Item scores for euphoria (R 2 = 0.36), disinhibition (R 2 = 0.48), and hallucination (R 2 = 0.59) did not highly predict caregiver distress in the community patients. Eating disorders scores did not highly predict caregiver distress in nursing home patients (R 2 = 0.55).

Comparisons Across Severity of Dementia
Among participants with mild dementia (CDR = 1), there were no group differences in prevalence, clinical significance, and moderate-to-severe caregiver distress caused by the symptoms (P > 0.05).

DISCUSSION
In this study, community-dwelling older adults with dementia showed more frequent clinically significant BPSD than nursing home patients, and these caused more severe caregiver distress. This was especially true for those with moderate-to-severe dementia. After controlling for dementia severity and medication use, family caregiving remained an independent risk predictor for presenting clinically significant symptoms and moderateto-severe caregiver distress. The prediction of caregiver distress based on symptom severity varied across caregiver types. The strength of this study is that our findings were controlled for dementia severity and medication used by participants. Our findings add to the growing understanding of BPSD and caregiver distress they cause in low-and middle-income countries.
Clinically significant symptoms were found in 88.5% of community-dwelling older adults with dementia and 75% of nursing home patients in this study. After controlling for dementia severity and medication use, we found that caregiver type remained an independent risk predictor for presenting clinically significant BPSD. One potential explanation is that BPSD rely on caregiver proxies to report, and family caregivers may have more opportunities to observe these symptoms. Moreover, family caregivers might be more involved in affective disturbances of care recipients, and the internalizing symptoms of dementia, such as anxiety and depression, might be overreported by family caregivers because of their increased risks for stress, depression, and burden. Another explanation is that formal caregivers are more likely to have had some training in managing BPSD (34). Training has been found to have an impact on the way staff behave toward residents with dementia (35). Family caregivers' improper management of behavior problems might result in aggravation of the symptoms (36). Delivery appropriate training to caregivers has been proposed as a key component of good dementia care (34).
Impressively, up to 70% of the community-dwelling older adults with dementia had at least one symptom causing severe caregiver distress in this study. However, only 9.8% of the nursing home residents with dementia had at least one symptom causing severe caregiver distress. Family caregivers caring for community-dwelling older adults with dementia experienced more severe distress caused by symptoms than formal caregivers caring for nursing home residents with dementia. After controlling for symptom severity, family caregiving remaining an independent risk predictor for moderate-to-severe caregiver distress in this study, especially eating disorders and anxiety. One explanation is that family caregivers provide longer company for care recipients and undertake additional caregiving tasks than formal caregivers, while formal caregivers are more likely to work cooperatively to manage patients in their workplace than family caregivers. Another explanation is that there is currently insufficient quality and quantity of family based care services in China to meet the needs of community-dwelling older adults with dementia (2). Family caregivers may therefore experience a higher level of distress due to the lack of resources and support in the public health care system to manage and treat dementia symptoms (18).
Previous studies suggest that information on caregiver distress caused by individual symptoms is crucial to inform the design of tailored management strategies for supporting family caregivers (6,15,37). In this study, agitation and sleep disorders were the fifth and sixth frequent of "very severe" symptom but caused the most frequent "very severe" distress to family caregivers caring for community-dwelling older adults with dementia. However, the community-dwelling patients had a similar prevalence of agitation during the moderate stage of dementia, compared with their nursing home counterparts. Agitation is one of the most common symptoms across studies (6,7,37,38). In the context of biopsychosocial approaches (36,39,40), agitation often occurs during personal care, and may be closely linked to expressions of unmet need (e.g., can be aggressive if attention is not given) and inadequate interactions between caregivers and care recipients (41). Sleep disorders in patients with dementia include fragmented sleep at night, daytime sleepiness, inversion of the sleep-wake cycle, and sleep behavior disorders (42). Various factors, such as the need to urinate during the night, brain stimulants (coffee and bronchodilators), pain, and medications (diuretics), may contribute to sleep disorders (43). Family caregivers have more exposure to care recipients at night (42,44), and sleep disorders can therefore cause or exacerbate sleep deprivation and exhaustion in family caregivers. We also found that euphoria was more prevalent and clinically significant, but had no difference in the distress for caregivers of community patients with severe dementia, compared with the nursing home patients. Our findings should encourage healthcare providers to focus on clinically significant symptoms that produce severe distress in family caregivers, rather than those that are more prevalent. Similarly, a recent study suggests that clinicians should distinguish relatively untroubling symptoms from more exhausting symptoms (37).
In this study, community-dwelling older adults with dementia had more than ten times the risk of presenting clinically significant anxiety and depression, which causing five to ten times the risk of moderate-to-severe distress to family caregivers. Affective disturbances should be taken into account in alleviating distress for family caregivers. We also found that the severity of euphoria and disinhibition did not strongly predict caregiver distress caused in community-dwelling older adults, while the severity of eating disorders did not strongly predict caregiver distress caused in nursing home residents with dementia. Euphoria and disinhibition, which have been classified as "Frontal" endophenotype (45), were the rarest symptoms in both this study and previous studies (37,45,46). A recent study suggests that sexual disinhibition appears to be a particularly difficult symptom for the family caregiver (47). However, there were 125 (11.8%) FTD patients in that study, but only 2 (1.3%) FTD patients in this study. The low prevalence of euphoria and disinhibition might partially contribute to the low prediction of caregiver distress caused in this study. Formal caregivers may be more likely to cooperate with colleagues, and be less involved in providing daily care (such as support with eating). Therefore, eating disorders may not place such a physical and psychological burden on formal caregivers. Our findings suggest that intervention strategies for reducing caregiver distress caused by BPSD should consider caregiver types and individual symptoms.  ≥ 4), and moderate-to-severe caregiver distress (distress score ≥ 3) caused by these symptoms, between community-dwelling older adults with dementia and nursing home residents with dementia, by severity of dementia. BPSD, behavioral and psychological symptoms of dementia; CDR, Clinical Dementia Rating. * P < 0.05, * * P < 0.01, * * * P < 0.001.
From a public health perspective, optimizing dementia care in the home may provide the maximum population-level benefit (such as decreasing the huge economic costs associated with transitions to other care settings) and is desirable given the preference of older people to remain in their familiar communities for as long as possible (48). The majority of older people with dementia live at home with unpaid family caregivers, and these family caregivers are at risk for poor physical, mental, emotional, and socioeconomic outcomes (49). The present study found that family caregivers caring for communitydwelling older people with dementia were exposed to more frequent clinically significant and moderate-severe distressing BPSD than formal caregivers (nursing home staff). BPSD are the most treatable aspects of dementia and might be alleviated or eliminated through comprehensive interventions to target contributing factors or triggers (34). There are complex overlaps or interplay of reactive, responsive, and/or organic causes of BPSD (50). Caregiver factors are some of the causes and triggers of BPSD (36), for example, inadequate interactions between caregivers and care recipients may contribute to the occurrence or exacerbation of BPSDs. Reciprocally, exacerbation of BPSD may aggravate caregiver distress (15). Training and supporting family caregivers, especially for proper management of BPSD may help to break this vicious circle. Our findings suggest that an effective support service system should be established to give family caregivers access to resources, information, and knowledge to improve their ability to cope with BPSD. The support service system, such as Dementia Support programs proposed in Australia (51), might enable people with dementia to stay at home longer.
This study had several limitations. First, the cross-sectional design of this study does not allow for a test of causal relationships. Second, modest sample size and insufficient information on dementia subtypes limited further analysis. Third, our data were collected from family caregivers who voluntarily visited the outpatient clinic of a tertiary psychiatric hospital, which may have introduced referral bias. Fourth, the study did not investigate factors related to caregiver burden or mental health status (such as depression and anxiety). The above-mentioned factors might limit the generalizability of our findings. Future studies with a larger sample size should consider these factors.

CONCLUSION AND IMPLICATIONS
This study found that family caregivers experienced more severe distress caused by behavioral and psychological symptoms of dementia than formal caregivers; family caregiving was an independent risk predictor for clinically significant and moderate-to-severe distressing symptoms; the prediction of caregiver distress based on symptom scores varied by individual symptoms and caregiver types. Our findings highlight that tailored management strategies are needed to relieve family caregivers' distress caused by BPSD, and that an effective support service system should be established to give family caregivers access to resources, information, and knowledge to improve their ability to cope with BPSD.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The Ethics Committee of Nanfang Hospital (No. NFEC-201511-K2) approved this study. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
XJL, JW, and RY gathered the data, completed the data analysis and interpretation, and took part in drafting of the manuscript.
DR helped with acquisition of data and clinical diagnosis. XYL advised on the design of the study, secured funding, supervised the research, took responsibility for the accuracy of the data analysis and interpretation, and provided critical and substantive commentary to both the process and final product. All authors granted approval for submission and publication, contributed to revising the manuscript critically for important intellectual content, and were in agreement with the content of the manuscript.

FUNDING
This work was supported by the National Natural Science Foundation of China [grant number 82171202]. The funding agents had no role in the study design; collection, analysis, and interpretation of the data; writing of the manuscript; or decision to submit the manuscript for publication.