CAREGIVERS ’ SENSE OF COHERENCE: IMPLICATIONS ON DIRECT AND INDIRECT COSTS OF DEMENTIA CARE

BACKGROUND. Dementia care is associated with physical, emotional, and monetary impact on the informal carers providing unpaid care. Differences in the personal characteristics of caregivers may help explain the variations in the costs of dementia care. OBJECTIVE. The aim of this study was to analyze the effect of caregivers ’ sense of coherence (SOC) on direct and indirect costs in dementia care. METHODS. A cross-sectional study was conducted in community dwelling caregivers of patients with Alzheimer’s disease . Data of healthcare services were obtained from clinical registries, and information was collected from caregivers regarding their use of social care resources and time spent caregiving. The transformation of all costs into Euros was made assigning a fixed cost of 10.29 €/h and 16.24 €/h for assisting in instrumental and basic activities of daily living, respectively. Caregiver s’ SOC was assessed using the Orientation to Life Questionnaire (OLQ-13). Adjusted regression models were developed, with different types of costs as dependent variables. RESULTS. A sample of 147 caregivers was recruited. The mean OLQ-13 score was 73.3 points (SD = 11.6). The regression models showed a small association between caregivers ’ SOC and direct costs, mainly linked to the use of social care resources (r 2 = .429; β = -15.6 €/month) , and a greater association between SOC and indirect costs (r 2 = .562; β = -222.3 €/month). CONCLUSIONS. Increasing caregivers ’ SOC could reduce dementia care costs by decreasing the use of social care resources and caregiving time. Sociodemographic variables: A questionnaire was developed to collect information on both the patients and their caregivers. Age, education level, and the patient-caregiver relationship were recorded.


INTRODUCTION
with many studies translating this model into health action plans in oncology, physiotherapy, and dementia [13]. However, there are currently only a few of these studies in dementia, with the main focus being on caregivers rather than care recipients [14]. Factors such as anxiety and burden among caregivers have been analyzed and correlated with their SOC [15,16]. Although dementia can change the form of relations in a negative way, some caregivers express positive aspects associated with these changes [17]. These differences in perceived (expressed) burden are associated with the use of resources, including the early institutionalization of people with dementia.
Although there is no consensus on how costs should be calculated, especially unpaid costs and those associated with the emotional impact of dementia care, there is agreement on the need to shift policies from institutional settings to home-and community-based settings [18]. If a caregiver's ability to cope with the stressors associated with caring increases, this may additionally reduce the utilization of health care system resources, resulting in less caring time and/or lower levels of stress for carers. The aim of this study was to analyze the association between a caregiver's SOC and the costs associated with care, assessing the cost of using medical healthcare services and other resources (i.e., day care centers or domiciliary care services), as well as the cost of the time spent caregiving (informal cost/unpaid cost).

Design and population
The 'Sense of Coherence and Dementia Study' (SOC & DEM study) is a prospective observational study of family caregivers of people with Alzheimer's disease (AD) living in the community in Spain. The main purpose of the study is to assess the relevance of SOC in the personal and financial evolution of caregiving throughout 24 months. This present study was cross-sectional, using baseline data of people

Inclusion and exclusion criteria
Inclusion criteria: 1) Individuals with a primary caregiver defined as a family relative who attends to and supervises the patient in their daily activities for a minimum of 4 h/week.
2) Individuals with a clinical diagnosis of mild to moderate AD (4-6a according to the Global Deterioration Scale and the FAST scale) [19].
Exclusion criteria: 1) Individuals with expressive aphasia and/or with moderate-to-severe difficulties in comprehension that are related to AD.
2) Vision or hearing impairment, or an inability to read and write (illiteracy).
3) People looked after by different family caregivers in different homes. 4) Primary caregivers who provide care to more than one dependent.

Procedure
Specialists from the memory clinics informed patients and their caregivers about the study. Visits were scheduled by telephone, and an experienced psychologist conducted the interviews in the centers.

Ethical considerations
The SOC & DEM study protocol was approved by the hospital's clinical ethics committee (ref v2:18/12/147). All participants consented to participate in the study.

Variables and instruments
Sociodemographic variables: A questionnaire was developed to collect information on both the patients and their caregivers. Age, gender, education level, and the patient-caregiver relationship were recorded.

Variables associated with costs
Costs were divided into direct and indirect costs. Direct costs included those associated with the use of healthcare resources and services. Information on any medical visits in the previous 6 months before the start of the study (to primary care and specialized services such as neurology, geriatrics, and neuropsychology) and visits to an emergency department and/or the number of days of hospitalization were recorded. Direct costs also included those associated with the use of public and private social care resources (mainly domiciliary care, day care centers, and participation in cognitive stimulation treatment programs). Indirect costs included those attributed to specific costs for time spent by caregivers on assisting individuals with dementia in their instrumental and basic activities of daily living (ADL). The monetary cost of using healthcare resources and visits was extracted from Vela et al., who described it as "a fee or an indirect cost applied to all healthcare activities financed by Catalan Health care system (CatSalut) during 2014, computing the health expenditure made up by each person and adding it all up for the inhabitants of Catalonia" [20]. We recorded the number of each type of visit and its specific cost.
In the case of emergency visits and hospitalizations, we used only the mean cost. The cost of private services such as home caring and day care centers was also included. We used the standard costs established by CatSalut for domiciliary care (13.83 €/hour) and for part-time (370.39 €/month) or fullday (617.32 €/month) attendances at a day care center. Finally, the cost of services such as using a private physician or complementary private health insurance was included too.
As for indirect costs, we assigned a fixed monetary cost of 10.29 €/h for assisting with instrumental ADL (IADL) and 16.24 €/h for assisting with basic ADL (BADL) (www.idescat.cat). The caregiving time was recorded as the number of hours per week spent in the last two weeks providing assistance with IADL and BADL.

Demographic variables and SOC
Data on sociodemographic variables such as sex, age, education level, household income, civil status and kinship were recorded.
Sense of coherence (SOC): The Orientation to Life Questionnaire (OLQ-13) consists of 13 items, which include: "Have you ever been surprised in the past by the behavior of people who you thought you knew well?" or "Do you experience feelings that you would rather not have to endure?" [12]. The items are scored on a seven-point Likert scale (e.g., 1 = never, 7 = always). Negatively-worded items are reverse coded (items 1, 2, 3, 7, 10). Higher scores indicate higher levels of SOC. The scale has shown appropriate internal consistency and high construct and cross-cultural validity [21]. The OLQ-13 has been adapted to the Spanish older population [22,23].

Patient clinical variables
Information on the main sociodemographic variables such as sex, age, education level, and civil status was recorded.
Disability of the patient: This was evaluated with the Dependence Scale (DS), which has 13 items, 2 of which have 3 options for response and 11 have dichotomous answers to identify the degree of dependence from the minimum (range = 0-15) [24,25]. The DS scores are used to indicate the level of dependence ranging from 0 (no dependence) to 5 (complete dependence) [26].
Cognitive status: This was assessed with the Mini-Mental State Examination (MMSE) questionnaire, with scores ranging from 0 to 30 (the lower the score, the greater the cognitive impairment) [27].
Behavioral and psychological symptoms in dementia (BPSD): The presence of BPSD was evaluated with the Neuropsychiatric Inventory -Questionnaire (NPI-Q) [28]. The NPI-Q is a self-administered questionnaire completed by the caregiver. Each of the 12 NPI-Q domains contains a question that reflects the cardinal symptoms of that particular domain. The NPI-Q indicates symptom severity and distress levels for each symptom reported, with the total severity and distress scores reflecting the sum of the individual domain scores. A higher score means more symptoms and distress (range = 0-96).

Statistical analysis
A descriptive analysis was performed for the clinical and sociodemographic characteristics. Absolute and relative frequencies were used for the qualitative variables, while central tendency and dispersion measures were used for the quantitative variables. Parametric and non-parametric tests were used for quantitative variables, while chi-squared test was used for categorical variables. Bivariate analyses were conducted to compare the measures of cost by categorical variables. Correlation analyses were performed for the OLQ-13 score and cost variables, taking into account the minimum sample size for a moderate effect size of 0.3 (89 people). The classification of the correlation effect was according to Cohen's criteria (0.1, low; 0.3, moderate; and ≥ 0.5, high).
The direct cost was divided into healthcare costs (the cost of using all services in the previous 6 months) and the costs of using social care resources (during the month of the baseline visit). A linear model was used, with each cost as the dependent variable and the caregiver variables as the independent variables.
The analysis of tolerance and collinearity was used to adjust the model. Variables such as the DS and NPI-Q scores were also included. The indirect cost (caregiving time) was analyzed by comparing users of resources and non-users.
All statistical comparisons were bilateral, and confidence intervals were calculated using a 95% reliability level. Data processing and analysis were performed using IBM SPSS v. 21 for Windows.

Description of the sample
The study sample consisted of 294 participants, of whom 147 were caregivers and 147 were care recipients. The percentage of women was higher in both groups. Spouses represented 55.1% of the caregivers, while the remaining caregivers were sons/daughters (except in eight cases, where the caregiver was a different type of family member). Three-quarters of the caregivers cohabited with the care recipient, while only 18% of the non-partner caregivers cohabited with the care recipient. Among our study participants, 91.2% of the individuals with dementia continued to live at their own home, of whom 76.6% were classified as having a level 1 (36.3%) or level 2 (40.3%) of dependence according to the DS score. A full description of the characteristics of the sample is shown in Table 1.
The mean total OLQ-13 score was 73.3 points (SD = 11.6). There were no differences in SOC according to the caregivers' gender, education level, kinship or cohabitation with the person with dementia, nor between the users and non-users of social care resources (73.7 vs. 72.9; Mann-Whitney U = 2,502.0; p = .754). However, there was a significant difference in the relationship between SOC and the use of resources by gender, with SOC being higher in male and lower in female users (76.9 vs. 71.2; Mann-Whitney U = 260.0; p = .016).
Primary care was the most used healthcare service. Over 70.1% of the sample had visited their primary care center in the previous 6 months on an average of 1.8 occasions (SD = 1.9; median = 1). Of note, all the patients had attended memory clinics in that period, with specialist appointments generally comprising a visit to a neurologist (91.8% of the participants) that involved an additional neuropsychological assessment in 63.3% of the appointments. Almost all the diagnoses had been made with neuroimaging, mainly computed tomography (62.6%) and MRI (36.1%), while 58.5% had laboratory tests conducted such as the analyses of complete serum biochemistry and vitamin B12 levels.
Only 10 individuals with dementia had been an inpatient in the previous 6 months, while 27 (18.5%) had attended the emergency department once (or more often). Domiciliary care and day care centers were used by 22 (15%) and 21 (17 part-time and 4 daily) participants, respectively. Finally, 27 individuals with dementia used other types of healthcare resources such as cognitive stimulation therapy and assistive technology (e.g., telecare alarm).
Caregivers reported providing less than half an hour a day (mean = .4; SD = .7) of support for BADL and more than 2 hours a day (mean = 2.1; SD = 1.7) of support for IADL. Table 2 shows the costs associated with these different services and assistance, while Figure 1 shows the distribution of these costs.
Differences in direct and indirect costs according to the characteristics of the caregiver When analyzing the three types of costs, namely those associated with healthcare services, social care resources and caregiving time, we observed differences that correlated with the characteristics of the care recipients and caregivers. Only adult child cohabitating caregivers showed a small statistically significant relationship (€108.1 vs. €80.2; Mann-Whitney U = 757.5; p = .016). The effect of SOC on costs SOC showed a small correlation with healthcare service costs (rho = -.195; p = .022), but a larger correlation with the costs of using social care resources (rho = -.226; p = .007). In terms of direct costs, spouses showed a significant correlation (rho = -.295; p = .007), while adult children did not, and only cohabitating caregivers showed a small correlation (rho = -.211; p = .036). Thus, a higher SOC of the spouses correlated with lower medical costs (rho = -.321; p = .004), but this was not observed for adult children. Likewise, a higher SOC in cohabiting caregivers correlated with lower costs of using social care resources (rho = -.261; p = .009) and healthcare services (rho = -.250; p = .013). Sole caregivers showed a moderate correlation between SOC and the cost of using social care resources (rho = -.308; p = .023). For indirect costs, sole caregivers also showed a correlation between SOC and costs associated with caregiving time (rho = -.208; p = .045). A stronger correlation between SOC and costs were observed in caregivers who used social care resources (rho = -.336; p = .010) than in those who did not use them. Table 3 shows the results of the different linear regression models used that were adjusted for SOC, DS scores, NPI-Q scores, household income, sole caregivers, and the caregiver's education level.

DISCUSSION
This study's main objective was to assess the effect of the caregiver's SOC on the direct and indirect costs of taking care of a relative with dementia. We observed that the caregiver's SOC may reduce the costs of care. Direct and indirect costs were also affected by the dependence of the individual with AD, the type of relationship (spouse or adult child), cohabitation with the caregiver, and being a sole caregiver. The main differences in the costs were in those associated with the use of social care resources like domiciliary care or day care centers.
Dementia care is associated with increasing medical and social care costs. It also has a physical, emotional, and monetary impact on the relatives providing unpaid care. Dependence and BPSD are the main variables associated with increasing costs in AD care in terms of both the costs of using healthcare services and the costs associated with caregiving time [4,29,30]. European multinational studies have reported an average cost ranging from €14,500 to €17,000 per year in dementia patients at home with some autonomy in ADL [31,32]. This is more than double the figures obtained in another study in which the participants had mainly mild Alzheimer's disease, like in our study, and baseline costs were €535 per month [33].
Caregiver's personal characteristics associated with dementia care costs SOC had a significant influence on the use of social care resources, with a reduction of €15 per month in direct costs per point of increased SOC and a reduction of €222.3 in indirect costs. These results are in accordance with those of previous studies showing greater increases in costs associated with caregiving time during the course of the disease [34]. Although one SOC point may not represent a large change, it could increase in significance during the whole process of dementia. Moreover, a 3point increase is a relevant change in SOC (10% improvement) that represents a greater reduction in the costs associated with using social care resources. In fact, the use of social care resources and paid carers for people with dementia is growing in Western societies. This has been linked to changes in the characteristics of caregivers, such as having a paid job, not cohabiting with the care recipient [6,35], and, more specifically, having a high education level and good employment [36]. The use of social care resources (e.g., day care centers and paid carers) involves an increase in care costs that in general are not included in public insurances (or only partially). Our study confirmed these observations, with greater use of resources linked to non-cohabiting adult child caregivers, in contrast to spouses who used fewer healthcare and social care resources. Non-cohabitation, in addition to increasing monetary care costs, are also known to cause a greater feeling of guilt (emotional cost) in caregivers [37,38].

Implications of sense of coherence in dementia care costs
In our study, the use of social care resources was not generalized, and, as usual, the largest part of the care costs was unpaid [39,40]. Therefore, if two-thirds of the global cost is not reimbursed money One interpretation could involve caregiver's burden. Previous research has linked burden to a lower SOC in caregivers [41,42]. Our results indicated that an improvement of one point in the caregiver's SOC reduced care costs up to €2,600 a year (extrapolation of monthly use of resources and the amount of caregiving time per week). Furthermore, caregivers' burden is a risk factor for greater and earlier use of social care resources and the institutionalization of the person with dementia [43,44].
In line with Antonovsky's theory, wellbeing is associated with generalized resistance resources (GRR) and a sense of coherence [45]. This health model encourages people to reflect on stressful situations in order to make them able to understand them, think about how to face them, and identify the best way to think about how dealing with a stressor can be meaningful [46,47]. Thus, one plausible explanation is that a greater sense of coherence promotes better use of social care resources. The reductions in costs are not due to caregivers with a better SOC not using social care resources, but are due to the SOC in caregivers allowing them to use the resource better. At the same time, handling the situation more efficiently reinforces self-esteem and self-efficacy perceptions that increase carers' sense of coherence [48].

The gender role
In the bivariate analysis, we observed differences between men and women in the relationship between using resources and a sense of coherence. It is known that the role of caregiving has an important gender component. Thus, we observed that women with paid jobs used fewer resources than men (25/75 vs. 40/60). Previous analysis of the profiles of caregivers in a similar context showed similar results. In both cases, when males were the partners of care recipients, they used more external services than female partners. In addition, maintaining a paid-job outside the home increased the use of external support services and, consequently, the cost of care. Again, women used these support services less often [49]. In fact, the role of women seems perpetuated by social pressure, the job market, and social culture [38,[50][51][52].
Gender did not appear as a significant factor in the adjusted regression models. However, our results did not allow us to conclude any effect of gender on dementia care costs. Previous studies have not found differences by gender, but by household or social class [53]. Maybe, we should have reported characteristics of the non-participants, such as the number of women or low-income families. Moreover, cultural differences should have been taken into account when assessing, for example, time spent caring for an individual. Some people might consider caregiving a normal part of life and a duty to their family.
Perceptions like these might be seen more often in spouses than in sons/daughters and more often in women than in men [54].
We found a significant relationship between a caregiver's sense of coherence and monetary care costs, but there were some limitations of the study that should be taken into account. The first limitation was the real clinical impact of the sense of coherence on care costs. The monetary costs assigned to the services could vary even though our results were similar to those of previously published studies from other countries [30]. A limitation of our study was that the use of social resources was not generalized by country or by healthcare system, with specific monetary costs being higher in other European places [55].
Second, this was a cross-sectional study that did not allow the causal effects of SOC or its influence on costs to be determined. For that reason, to clarify our results, we calculated the differences in costs for 3 SOC points. For a middle-income family in Catalonia (€13,338/year), a monthly reduction of more than €200 would represent a significant change in their household costs [56]. However, this study was undertaken in a specific Mediterranean region and, as mentioned previously, there were some cultural differences that could limit the generalization of our results. Comparative studies have shown differences between different countries and regions of the EU in terms of the values associated with caregiving [54].

Future lines of research
This paper is part of the SOC & DEM study. The aim of this study is to assess the effect of the personal characteristics of caregivers on dementia care costs, and the main hypothesis is sense of coherence has an important influence on these costs (monetary and personal costs) and may reduce them. To our knowledge, this is the first study that has translated OLQ-13 scores into monetary figures associated with the care of dependent people with AD in the community. Although there are some limitations to generalize these results, this work could complement previous studies on the cost of healthcare in dementia, focusing on the need to improve the sense of coherence in caregivers to reduce indirect costs. This was a cross-sectional study and it will be interesting to observe these relationships and their strengths longitudinally. Social support, self-esteem, and self-efficacy are direct factors associated with this improvement [58][59][60][61]. Cultural values of caregiving or clustering caregivers by their sense of coherence could be used to enhance the effects of SOC on care costs.

Conclusions
Based on our results, we conclude that investing in caregivers could reduce care costs in dementia, especially by working on the differences between genders in caregiving. Although dementia is commonly described in a negative way (dependence, disability, and loss), some caregivers and professionals are changing this perception [62]. A greater sense of coherence could promote this change. Although it will not significantly reduce direct and indirect monetary costs, a greater sense of coherence may reduce the emotional and social costs associated with caring for people living with dementia.

DISCLOSURE STATEMENT
None of the authors have reported a potential conflict of interest.