Factors related to feelings of burden among caregivers looking after impaired elderly in Japan under the Long-Term Care insurance system
Dr Yumiko Arai, Research Unit for Nursing Caring Sciences and Psychology, National Institute for Longevity Sciences (NILS), 36-3 Gengo Morioka-cho, Obu-shi, Aichi 474-8522, Japan. Email: firstname.lastname@example.org
Abstract Since the 1970s, the burden of caregiving has been the subject of rather intense study, a trend that will continue with the rapid graying of populations worldwide. Since the Long-Term Care insurance system began in 2000, few cross-sectional studies have attempted to identify factors related to the feelings of burden among caregivers looking after the impaired elderly in Japan. In the present report, among 46 pairs of caregivers and impaired elderly, the elderly receiving regular nurses’ visits in Kyoto Prefecture, Japan were assessed for problems with activities of daily living, the severity of dementia, the presence of behavioral disturbance, and cognitive impairment. The caregivers were asked to complete questionnaires in relation to their feelings of burden and caregiving situation. The results indicated that caregivers of impaired elderly with behavioral disturbances were more likely to feel a ‘heavier burden.’ Those temporarily relieved of caregiving three or more hours a day were less likely to experience ‘heavier’ caregiver burden than those who were not. Moreover, caregivers who found it ‘inconvenient’ to use care services tended to be more likely to feel a ‘heavier’ caregiver burden than those who did not. Recourse to respite services, which are ideally positioned to help, proved inconvenient because of their advance reservation system. More ready access to respite services in emergencies could do much to reduce caregiver burden.
Informal care for the disabled elderly has proved to be a heavy burden for family caregivers in many countries, and numerous studies have been conducted in relation to the burden of caregiving since the 1970s.1–9 In all of these cross-sectional studies, behavioral disturbances of the impaired elderly were found to be factors related to caregiver burden. In contrast, factors such as activities of daily living (ADL) of the impaired elderly have not shown consistent relationships with caregivers’ burden.1,7,9 Longitudinal studies conducted in various countries have also yielded inconsistent results as to how caregivers’ burden changes over the course of caregiving.10–14
In April 2000, a new public Long-Term Care (LTC) insurance scheme (Kaigo Hoken) was launched in Japan, making it only the third country, after The Netherlands and Germany, to provide such insurance. For Japan, the new system departed from the Confucian-based tradition and policy according to which the family is considered the prime resource for the long-term care of its members.15 The new eligibility criteria do not take into account the extent of informal family care available to clients. As in Nordic countries, the ultimate responsibility lies with the State rather than with the family itself. In addition, the LTC insurance system has brought about a fundamental change in how care services are provided to the impaired elderly. Prior to the system, most services were made available by order of the municipal government. Hence, neither the impaired elderly nor caregivers were entitled to ask the municipal government what kind or the amount of services they wished to use. In contrast, after the system was in place, services have been allocated based on the Government Certified Index (GCDI)(Yokaigodo); the impaired elderly are entitled to decide both the kind and the amount of services they wish to use within the parameters which the GCDI allows.16 Under the LTC insurance system, the impaired elderly and/or caregivers are now in a position to express their views regarding the quality of services, for example, their convenience in using services.
There have been few cross-sectional studies regarding caregiver burden conducted after April 2000. Thus, the objectives of the present study were to identify factors related to the feelings of burden among a group of caregivers looking after the impaired elderly under the LTC insurance system. We were especially interested in exploring whether or not their feelings of burden were related to the inconvenience in using services experienced by caregivers looking after the impaired elderly under the new system.
As the first step, 53 impaired elderly were identified who had been registered with a practice nurse clinic attached to U general hospital in the vicinity of Kyoto City. In the second step, these 53 elderly and their family principal caregivers were contacted by letter to explain the objectives of the present study. Informed consent was obtained from all of these pairs in March 2001. This study was endorsed by the ethical committee of the National Institute of Longevity Sciences.
All of the primary caregivers were co-residing with the patients at the time. Among these 53 pairs, 51 pairs participated in the present study. As the third step, the medical records of these 51 impaired elderly were reviewed by the first author. Of these 51 elderly, 46 were found to suffer from dementia. These 46 pairs were analyzed in the present study.
In the present investigation, the impaired elderly were assessed for problems in carrying out ADL, the severity of their dementia, the presence of behavioral disturbances, and cognitive impairment. The nurses assessed the patients’ physical disability using the Barthel Index (BI), a widely used 10-item ADL scale.17 A cut-off point of nine on the BI represented the threshold between moderate and severe dependence.17 Severity of dementia was assessed by nurses using criteria developed by The Ministry of Health, Labor and Welfare; severity of dementia was rated from I (very mild) to IV (very severe), and M was defined as ‘very severe with extremely disruptive behavior.’18 The same Ministry suggested that those who were rated I or II were relatively easy to care for at home albeit having dementia, and hence we used this classification for the statistical analyses in the present study.
The frequency of behavioral disturbances observed by primary caregivers was assessed using the Troublesome Behavior Scale (TBS), which was designed to quantify the specific observable behavior usually associated with dementia.19,20 Cognitive impairment of the elderly was assessed with the Japanese version of the Short Memory Questionnaire (SMQ),21 which was developed as an objective tool for the assessment of memory difficulties found in dementia22 and has proved useful for screening of dementia in a Japanese population.23 The SMQ consists of 14 items concerning everyday memory problems; the scoring range is 4–46 points. A score of less than 40 is suggestive of dementia.
These caregivers were asked to complete the following questionnaires in relation to their feelings of burden, depression and caregiving situation. The questionnaires included: the Japanese version of the Zarit Burden Interview (J-ZBI); questions regarding demographic variables of the caregivers and patients; and questions regarding the hours spent in caregiving and the duration of caregiving.
Caregivers’ burden was assessed by the J-ZBI, a 22-item self-report inventory that examines the burden associated with functional behavioral impairments in the home care setting.24 A short version of the J-ZBI (J-ZBI_8) has recently been released.25–27 The original version of the ZBI is one of the most common scales used in North America and European countries for assessing the burden of caregiving.28 Caregivers were asked to indicate how many hours per day they provided care for the patient as well as how many months’ duration they had cared for him or her. Caregivers were also asked to estimate the number of hours per day they were able to be temporarily relieved of their duties or to leave the side of the patient to go out.
The caregivers in the present study were divided into two groups: those who felt a relatively heavier burden; burden score of 40 or more (i.e. group 1), and those who felt less burden; burden score under 40 (i.e. group 2). The score was put at 40, since the cut-off point as the median ZBI score in the present study was 40.
First, the χ2 test was conducted to determine differences between the more heavily and less burdened groups (i.e. associations between ZBI and the variables concerned). Second, Spearman's rank correlation tests were conducted on the variables found to be statistically significant on the above univariate analysis (P < 0.05). Then, the variables found to be highly correlated with one another were excluded from further analyses in order to avoid multicollinearity. Third, a multiple logistic regression analysis was employed to determine which of the remaining variables was significantly related to the caregivers experiencing heavier feelings of burden. Then the Statistical Package for Social Science (SPSS, version 11.5) was used for the above statistical analyses.
Characteristics of subjects
Table 1 shows the characteristics of the subjects in our study. The average age of the caregivers was 60.5 years old, and three-quarters were female. The mean age of the elderly (patients) was 79.8 years old, and two-thirds were female. The average hours of caregiving were 5 h/day, and the average duration of caregiving (months) was 55 months (i.e. approximately 5 years). All of these 46 impaired elderly scored less than 40 in the Short Memory Questionnaire (SMQ), indicating that all of them were cognitively impaired.
Table 1. . Characteristics of subjects (n = 46)
| Age|| ||79.8 (9.8)|
| Male||16|| |
| Female||30|| |
| Barthel Index|
| Less than 9||30|| 7.5 (6.1)|
| Severity of dementia|
| I|| 5|| |
| II||14|| |
| III|| 6|| |
| IV||19|| |
| M|| 2|| |
| Troublesome Behavior Scale|| || 6.0 (8.1)|
| Short-Memory Questionnaire|| || 9.0 (8.1)|
| Age|| ||60.5 (13.7)|
| Male||11|| |
| Female||35|| |
| Wife||13|| |
| Husband|| 4|| |
| Daughter||10|| |
| Son|| 5|| |
| Daughter-in-law||11|| |
| Other|| 3|| |
| Japanese version of the Zarit Burden Interview|| ||40.1 (18.3)|
| Hours of caregiving/day|| || 5.0 ( 4.1)|
| Duration of caregiving/months|| ||55.0 (52.7)|
| Hours caregivers can be relieved|| || 2.9 ( 2.9)|
| No. of services used|| || 2.4 ( 1.1)|
Table 2 compares group 1 (less burdened caregivers) and group 2 (more burdened caregivers) regarding the variables concerned. The following four variables were significantly different between the two groups: the proportion of disabled elderly who had behavioral disturbances; the proportion of disabled elderly with III, IV or M severity of dementia; the proportion of caregivers who found it convenient to use services; and the proportion of caregivers who were able to go out (leave the patients) more than 3 h/day.
Table 2. . Comparisons between ‘less burdened’ and ‘more burdened’ caregivers
| Sex (male vs. female)|| 9 vs. 14||n = 23|| 7 vs. 16||n = 23||0.38||0.54|
| Age (years) (−79 vs. 80+ )|| 8 vs. 15||n = 23||13 vs. 10||n = 23||2.19||0.14|
| Barthel Index (0-9 vs. 10+)||13 vs. 10||n = 23||17 vs. 6||n = 23||1.53||0.22|
| Behavioral disturbances (0 vs. 1+)||15 vs. 8||n = 23|| 6 vs. 17||n = 23||7.10||0.01***|
| Severity of dementia (I, II vs. III, IV, M)||13 vs. 9||n = 22|| 6 vs. 16||n = 22||4.54||0.03**|
| Sex (male vs. female)|| 5 vs. 18||n = 23|| 6 vs. 17||n = 23||0.12||0.73|
| Age (years) (−59 vs. 60+)||12 vs. 11||n = 23||13 vs. 10||n = 23||0.09||0.77|
| Hours of caregiving/day (−4 vs. 5+)||15 vs. 7||n = 22|| 9 vs. 14||n = 23||3.81||0.05*|
| Duration of caregiving/ months (−54 vs. 55+)||17 vs. 6||n = 23||13 vs. 10||n = 23||1.53||0.22|
| Find it convenient to use services (+vs.–)||13 vs. 8||n = 21|| 6 vs. 16||n = 22||5.23||0.02**|
| Support from others (+vs.–)||12 vs. 11||n = 23||13 vs. 9||n = 22||0.22||0.64|
| Occupation (+vs.–)||10 vs. 13||n = 23|| 4 vs. 18||n = 22||2.28||0.13|
| Caregiving experiences (+vs.–)|| 9 vs. 13||n = 22||12 vs. 11||n = 23||0.57||0.45|
| Belong to support group (+vs.–)|| 2 vs. 21||n = 23|| 6 vs. 17||n = 23||2.42||0.12|
| Consulted physicians within last 1 month (yes vs. no)||17 vs. 6||n = 23||17 vs. 6||n = 23||0.00||1.00|
| Spouse as caregiver (yes vs. no)||10 vs. 11||n = 21|| 7 vs. 16||n = 23||1.37||0.24|
| Child as caregiver (yes vs. no)|| 7 vs. 14||n = 21|| 8 vs. 15||n = 23||0.01||0.92|
| Hours caregivers can be relieved (−2 vs. 3+)|| 9 vs. 13||n = 22||17 vs. 6||n = 23||5.02||0.03**|
Table 3 presents the Spearman's rank correlation between the above four variables. First, caregivers’ burden was correlated with the patients having behavioral disturbances (ρ = 0.39, P = 0.01), severity of dementia (ρ = 0.32, P = 0.03), level of convenience the caregivers experienced in using services (ρ = 0.35, P = 0.02), and hours caregivers were temporarily relieved of their duties (ρ = − 0.33, P = 0.03). Second, the severity of the patients’ dementia was correlated with the level of convenience the caregivers experienced in using services (ρ = 0.31, P = 0.05). In summary, there was a close relationship between the severity of dementia and the level of convenience caregivers encountered in using services. As we were interested in investigating whether or not their feelings of burden were related to the convenience expressed by caregivers looking after the impaired elderly under the LTC insurance system, we excluded ‘the severity of patients’ dementia’ variable from the following multivariate analysis.
Table 3. . Correlations between ADL, behavioral disturbances, no. of family members and caregivers’ experience of having consulted physicians
|J-ZBI score (−39 vs. 40+)||1.00||0.39||0.32||0.35||−0.33|
| ||P = 0.01||P = 0.03||P = 0.02||P = 0.03|
|Behavioral disturbances (0 vs. 1+)|| ||1.00||0.13||0.11||0.04|
| || ||P = 0.42||P = 0.49||P = 0.79|
|Severity of dementia (1–2 vs. 3-4)|| || ||1.00||0.31||−0.29|
| || || ||P = 0.05||P = 0.06|
|Find it convenient to use services (+vs.–)|| || || ||1.00||−0.21|
| || || || ||P = 0.18|
|Hours caregivers can be relieved (−2 vs. 3+)|| || || || ||1.00|
A logistic regression analysis was employed to determine which of the following three remaining variables were significantly related to the caregivers’ heavy feelings of burden: behavioral disturbances (0 vs.. 1); hours during which caregivers were temporarily relieved (< 3 vs. ≥3 h/day); and caregivers’ convenience in obtaining services (Yes vs.. No).
As shown in Table 4, two of these three variables proved to be significant factors related to the caregivers’ heavier burden: behavioral disturbances and hours during which caregivers can be relieved. Those who looked after the disabled elderly with behavioral disturbances were more likely to experience ‘heavier’ caregiver burden (odds ratio = 7.16; CI = 1.48–34.70, P = 0.01). In contrast, those able to leave their patients for 3 h or longer were less likely to experience ‘heavier’ caregiver burden (odds ratio = 0.20, CI = 0.47–0.97, P = 0.04). Caregivers who found it convenient to use services tended to belong to the ‘less burdened’ group, although the difference between groups was not statistically significant (odds ratio =3.62; CI = 0.81–16.22, P = 0.09).
Table 4. . Factors related to caregivers’‘heavier’ burden
|Behavioral disturbances (0 vs. 1+)||7.16||1.48–34.70||0.01|
|Hours caregivers can be relieved (−2 vs. 3+)||0.20||0.04–0.97||0.04|
|Find it convenient to use services (+vs.–)||3.62||0.81–16.22||0.09|
The present study confirmed the well-documented relationship between caregiver burden and behavioral disturbances.7–9,12,29 Some studies have even reported that behavioral disturbances were the strongest correlate of caregiver burden.7,8 On this point, our study is thus consistent with the above studies conducted in the West as well as with our previous study conducted in Japan prior to the implementation of the LTC insurance system. This result can be interpreted to mean that behavioral disturbances have remained a strong correlate of caregiver burden even after the LTC insurance system was in place. One of the goals of this system is to reduce caregiver burden.30 Most of the care services currently available under the LTC insurance scheme are intended to assist the impaired elderly with ADL. The services are not intended to provide medical/pharmacological treatments. Thus, all interventions possible under the present system are non-pharmacological. The problem, of course, is that there is little unequivocal evidence that such interventions are effective.
Some evidence suggests that caregiver intervention can possibly reduce the severity of patients’ behavioral disturbances31–33 or at least improve the caregivers’ ability to tolerate particular symptoms.34 Thus, it would be desirable if such caregiver intervention would be possible under the LTC insurance system.
Second, those who were completely relieved of their caregiving duties for 3 h/day (or longer) were less likely to feel caregiver burden than those who did not, even after statistically controlling for confounding factors. Indeed, the break from caregiving duties in order to spend time away from the patient and engage in activities other than caregiving is certainly a key reason why caregivers were less likely to feel a ‘heavier’ burden in the present study.35 Therefore, steps must be taken to temporarily relieve caregivers, allowing them more time away from their patients or duties. A first step would be for caregivers to have someone they could ask to take over for them when necessary, for example, a family member (informal help). Indeed, informal instrumental support has been demonstrated to be effective in alleviating caregiver burden.36 When asking for professional help, of course, the use of more home help services would be desirable. A second step would be for caregivers to rely on respite services. However, virtually no studies to date have shown that respite care directly alleviates caregiver burden itself, although its use is reportedly related to caregivers’ satisfaction37 and to delay institutionalization.38,39 Thus, care services which could completely release caregivers from home-caregiving temporarily would serve to alleviate their burden.
Third, although the difference was not significant, those who felt it inconvenient to use care services tended to be more likely to feel a ‘heavier’ caregiver burden than those who did not. In order to clarify what makes currently available services ‘inconvenient’, the first author interviewed four ‘heavily burdened’ caregivers who had expressed inconvenience in using services in the present study. When this subsample of caregivers was asked just what they meant by ‘inconvenient’ services, they defined them as ‘services which are not available when you want them.’ For example, in order to use respite services, caregivers many times had to apply several months in advance to avail themselves of such services; and even when these caregivers themselves suddenly became ill and became unable to look after the elderly in their care, neither respite nor home help services were an option because the requisite ‘reservation’ had not been made. Also, caregivers who suddenly had to attend funerals found such services unavailable. In particular, respite services were very often beyond reach because of the need for long-standing reservations. In the present LTC insurance system, services must be requested well in advance. Unexpected conditions are not foreseen. This appears to be why caregivers found them ‘inconvenient’, especially the respite services.
The issue of the convenience of access to services deserves further discussion.40 As mentioned at the outset, prior to the LTC insurance, neither caregivers nor the elderly were in any position to express their views regarding the quality of services, for example, their inaccessibility or so-called ‘inconvenience.’ Apparently, the ‘caregiver inconvenience’ first emerged as a factor after the LTC insurance system was in place. This welcome 180-degree change in the history of social policy in Japan may have made more caregivers aware of not only the availability of services but also their quality, especially in terms of accessibility. Thus, measures must be taken to allow caregivers immediate access to services when the unexpected happens. For example, a fixed number of respite beds must be set aside by each municipality to accommodate the emergency needs of the impaired elderly on a reservation-free basis.
In conclusion, behavioral disturbances and the length of time caregivers were totally relieved of caregiving were identified as factors related to caregiver burden even after the LTC insurance was implemented. Also, the ‘caregiver inconvenience’ in using such care services first emerged as a factor after the LTC insurance system was in place. Thus, services which can totally release caregivers from caregiving and also allow them emergency access as needed may well serve to alleviate their feelings of burden.
Our special thanks to the participants in the present study, which was in part supported by research grants provided by the Ministry of Health, Labor and Welfare (Grant No. H15-25), the Ministry of Education, Culture, Sports, Science and Technology, Japan (Grant no. 14570375) and the Uehara Memorial Foundation.