Factors influencing caregiver burden by dementia severity based on an online database from Seoul dementia management project in Korea

Seoul Dementia Management Project (SDMP) provides comprehensive healthcare programs under the supervision of Seoul Metropolitan Center for Dementia (SMCD) [17]. The SDMP emphasizes community-based integrative management for dementia, which encompasses education, preventive programs, early detection, therapeutic interventions, and proper care services that are closely linked to online case-registration & management systems. The online database from SDMP provides socio-demographic characteristics of patients, results of screening examination, and history of care programs such as detailed information of services to prevent worsening dementia symptoms, improve patients’ cognitive function, and provide social support for patients’ families.

The Clinical Dementia Rating (CDR) scale, a generally used dementia staging instrument, was used by a panel of neurologists and psychiatrists from 25 districts of Dementia Safe Centers. CDR score was assessed using a collected clinical instrument during phase 1 and phase 2. This study followed global CDR criteria, a 5-point ordinal scale. CDR scores of 0.5 ~ 1, 2, and 3 refer to mild, moderate, and severe dementia, respectively, while CDR 0 indicates no dementia. In phase 1, a trained nurse performed a mental status examination to identify cognitive impairment. The cognitive impairment group had completed either Consortium to Establish a Registry for Alzheimer’s Disease or Seoul Neuropsychological Screening Battery [18]. In phase 2, neurologists and psychiatrists diagnosed patients’ symptoms under the basis of clinical assessment or Diagnostic and Statistical Manual of Mental Disorders-IV. Therefore, a total of 12,292 patients were analyzed after excluding those who had no information about their major caregiver (38,055), no Clinical Dementia Rating (CDR) information (1527), and patients with no dementia (CDR = 0) [19] from 51,908 patients from 2010 to 2016 in an online database (Fig. 1).

The severity of dementia was assessed with the Seoul Dementia Assessment Packet (SDAP), a brief screening instrument that was developed to screen patients’ symptoms and caregiver burden in multidimensional aspects. SDAP was used to evaluate patients’ symptoms based on the summary score of cognitive impairment, Behavioral And Psychological Symptoms In Dementia (BPSD), Activities of Daily Living (ADL), and Instrumental Activities of Daily Living (IADL) [20].

Tools used for assessing the functional ability of patients with dementia are as follows. Cognitive impairment was measured by four items: memory, orientation, problem-solving, and communication skill. The severity of cognitive impairment was categorized into three levels based on total scores from the questionnaire: mild (0–4), moderate [5‐8], and severe [9‐12]. The severity of BPSD included six domains: delusion, hallucination, agitation, apathy, irritability, aberrant motor behavior, and sleep disturbance. The severity of BPSD was classified into three categories after summing scores from each questionnaire: mild (score of 0 ~ 6), moderate (score of 7 ~ 12), and severe (score of 13 ~ 18). ADL consisted of nine items: bathing, dressing, grooming, mouth care, walking, climbing stairs, eating, transferring bed or chair, and getting toilet hygiene. The severity of ADL was determined as follows. Patients who gained a total score of 0 to 9 were considered to have mild symptoms. Those with a total score of 10 to 18 and 19 to 27 were defined as having moderate and severe symptoms, respectively. IADL included seven areas: ability to use a telephone, laundry, shopping, food preparation, housekeeping, taking medication, and handling finances. The severity of IADL was classified into three categories: mild (score of 0–7), moderate (score of 8–14), and severe (score of 15–21).

Korea version of Zarit Burden Interview (ZBH-K) was used to measure caregiver burden [21]. Trained nurses were educated about burden interviews in advance. They recorded responses of caregivers during a 1:1 interview. Assessment tool was comprised of 22 domains, ranging from a score of 0 (not at all) to a score of 4 (have always been).

Categorical data are presented as numbers and percentages. Continuous variables are presented as means and standard deviations. A chi-squared test was used to test the relationship between categorical variables. Multiple linear regression was used to analyze the relationship between CDR and characteristics of patients and caregivers. Patients’ CDR was used as a dependent variable which was stratified by severity level of dementia symptoms. All variables included in the bivariate analysis were entered in the multiple linear analysis except for patients’ and caregivers’ marital status and caregivers’ educational level. The marital status of the patients had strong collinearity with residence type. Therefore, we replaced marital status to residence type as one of the entry variables to understand the relationship between caregiver and patient. The reason for not including caregivers’ educational level is because more than 2000 caregivers did not answer for their education experience. Additionally, when we added the caregivers’ educational level to the regression model, no level of education was statistically significant, and the regression model was not improved. The number of missing in both patients’ and caregivers’ marital status were not large as that of caregivers’ educational level. However, it was not included in the model because of the same reason with the caregivers’ educational level.

Table 1 shows general characteristics of patients with dementia. The average age of patients was 80.5 years. The majority (68.4%) of them were females, which was twice as many as males (31.6%). The highest percentage was found for those with at least elementary education (36.9%) regarding education level, those who were bereaved (53.8%) regarding marriage status, those who were co-residing with other family members (36.9%) regarding co-resident type, and those who had no religion (36.5%) regarding religion status. Among patients classified by the severity of dementia, most patients had at least CDR 1 level (70.6%), followed by those with CDR 2 (18.5%) and CDR 3 (10.9%).

Results of SDAP are shown in Table 2. Most patients showed a moderate level of cognitive impairment (49.4%), behavior problems (82.6%), and ADL dependency (73.6%). However, a severe level of IADL dependency was the most frequent case (40.5%) among patients.

As shown in Table 3, most (68.5%) caregivers were females with an average age in the early 40s to late 50s (49.9%). They had 10 years to 12 years of upper secondary education (27.9%), co-residing with patients (63.8%), non-religion (45.8%), married (80.5%), unemployed (56.3%), and a moderate level of self-rated health status (56.0%). Mostly daughters were taking care of patients (27.8%) with no secondary caregiver (57.5%).

General characteristics of patients with dementia and caregivers were used as independent variables in multiple regression analysis to identify significant factors affecting caregiver burden (Table 4). Before stratifying the severity of dementia, significant determinants of patients and caregivers were as follows. Male patients (β = − 1.396), and living with spouse and other family members (β = − 1.487) or others such as friends or volunteer workers (β = − 6.889) showed negative relationships with caregiver burden. Caregivers who were not co-residing with patients (β = − 3.769), who had Christian religion (β = − 1.071), who had no family relationship with patients (β = − 1.795), who answered that their recent health status was more than moderate (moderate: β = − 7.019, good: β = − 10.299), and who shared their works with secondary caregivers (β = − 2.399) showed similar results. However, patients with cognitive impairment (β = 1.312), limited ADL (β = 0.072), limited IADL (β = 0.698), and severe BPSD (β = 1.545) had positive relationships with caregiver burden. In the case of caregiver characteristics, those who were females (β = 3.386) or having parent and child relationships (daughter: β = 1.838, son: β = 2.835) had more caregiver burden.

After stratifying the severity of dementia, results of CDR 1 were similar to overall patients’ caregiver burden. However, in case of CDR 2, patients who were living with others (β = − 6.451), caregivers who were not cohabitating with patients (β = − 4.706), who had moderate (β = − 7.452) or good (β = − 10.380) health status, or those who were working with secondary caregivers (β = − 3.460) showed less caregiver burden. On the other hand, female caregivers (β = 1.131) or patients who had limited IADL (β = 0.267) and severe BPSD (β = 1.368) had positive relationships with caregiver burden. Results of CDR 3 showed that patients who were co-residing with others (β = − 5.720) and caregivers who answered their health status as moderate (β = − 9.470) or good (β = − 9.924), and who were living independently with patients (β = − 5.726) showed less caregiver burden. In contrast with this, patients who were experiencing cognitive impairment (β = 0.806) or severe BPSD (β = 1.206) and female caregivers (β = 4.172) showed more burden.