Caring for a Family Member with Dementia: Evidence from a Cross Sectional Comparative Study on Caregiver Burden and Psychological Well-being


Background and aims: Previous research is not entirely consistent with regard to whether dementia caregivers are more adversely affected by their role than non dementia caregivers.
The present study further explored this hypothesis by comparing 84 dementia dyads and 53 non dementia dyads in terms of caregiver burden and psychological well-being. Methods: Dementia and non dementia dyads were initially described in detail to evaluate the comparability of the two sub-samples. Five multiple regression analyses were run to see whether patient dementia status significantly predicted the criterion variables (caregiver objective and subjective burden, stress, anxiety and depression), once the effect of potentially confounding factors has been taken into account. Results: Patient dementia status emerged as a significant predictor of caregiver stress and depression, as dementia caregivers reported higher levels of stress and a more depressed mood than non dementia caregivers. Moreover, it was also found that spouses of care recipients, rather than children or other relatives, reported the most negative consequences to themselves. Conclusions: These findings support the view that caring for a dementia patient may be a particularly troublesome experience and that dementia caregivers — especially if they are spouses — are in need of more support from formal care services.
Keywords: caregiving spouses, dementia, burden, stress, anxiety, depression

1. Introduction

Providing care to a dementia patient is a highly demanding task. This activity, which in the great part of cases is carried out by informal caregivers (Kneebone and Martin, 2003; Stone et al., 1987; Wiles, 2003), often causes them high levels of burden and distress, which in turn may lead to appreciable physical as well as psychological consequences (Barinaga, 1998; Di Gregorio et al., 2002; Macdonald and Dening, 2002; Marshall, 2001; Pinquart and Sorensen, 2003; Schulz, et al., 1995), especially when an unfavourable physical and psychological status characterises the caregivers (Onishi, et al., 2005). In this respect, spouses of care recipients appear as those reporting the most considerable burden, given that they were repeatedly found to be more distressed and depressed than other relatives (Pinquart and Sorensen, 2003; Schulz, et al., 1998, 2004). Furthermore, in the case of spouses the emotional strain resulting from caregiving has been found to be an independent risk factor for mortality (Christakis and Allison, 2006; Schulz and Beach, 1999).
A question that is receiving increasing attention in this research area is whether caring for a dementia patient is comparatively worse than caring for a patient with other debilitating and chronic conditions, but a sufficient level of mental lucidity. It has been suggested (Light, et al., 1994) that there may be something unique about caring for a demented person, independently of caregiver characteristics and level of caregiving involvement. This may be related to the fact that dementia, contrary to other chronic conditions, progresses by erasing judgment, memory and the sense of self, thus finally compromising the bases for any form of interaction between the caregiver and the care recipient (MacRae, 2002), so that the severity of cognitive impairment has been clearly linked with depression in carers (Ballard, et al., 1995). Moreover, the behavioural changes associated with the onset of the disease (Hooker, et al., 2002) often render the care recipient much more difficult to look after. As a consequence, the caregiver may no longer manage to find benefits in his/her role.
However, while it appears understandable that the (cognitive, behavioural and affective) losses associated with the progression of the disease make dementia caregiving comparatively more stressful and burdensome, empirical evidence remains inconclusive on this. Clear-cut differences between dementia and non dementia caregivers emerged in some studies (Clipp and George, 1993; Holmen, et al., 2000; Ory, et al. 1999), with dementia caregivers reporting worse consequences in terms of self-rated health, emotional health, social life and financial status. Furthermore, dementia caregivers were also found to make use of more psychotropic agents than carers of cancer affected relatives (Clipp and George, 1993), suggesting that coping with the cognitive and behavioural decline of a person with dementia may be a particularly difficult task for his/her carer(s).
Such striking differences between the two groups of caregivers, however, were not replicated in other studies. For example, in a comparison between Alzheimer disease caregivers and Parkinson disease caregivers without coexisting dementia (Hooker, et al., 1998), it was found that while the former reported worse mental health consequences (anxiety, depression and a combination of positive and negative affect) than the latter, the pattern of results was inverted with respect to physical health consequences (perceived health and number of reported chronic conditions). In another study, which compared mild dementia caregivers with stroke and Parkinson disease caregivers (Tommessen, et al., 2002), no differences were found across the three diagnostic groups with respect to the psychosocial burden (perceived stress) experienced by the caregiver. Few, if any differences between caregiving groups emerged in a series of other studies (Canadian Study of Health and Ageing Working Group, 2002; Cattanach and Tebes, 1991; Draper, et al., 1992; Rabins, et al., 1990), and evidence has been also provided of an advantage of caregivers of cognitively impaired patients compared with caregivers of physically disabled patients (Yu, et al., 1993), as well as for lower levels of depression in spouses of demented Hispanic elders compared to spouses of cognitively impaired but not demented ones (Hinton, et al., 2003).
Despite these controversial results, this research area is an important one and deserves further attention: while most caregivers provide care to functionally disabled patients, forecasts for EU countries indicate that the number of people aged 65 and older with dementia disorders will increase substantially in the near future, growing from about 4 million in 2000 to about 4.8 million in 2015 (Berr, et al., 2005; Eurostat, 2003; Lobo, et al., 2000), and a similar trend is expected to affect all main regions of the world (Cleusa, et al., 2005). Therefore, dementia caregiving will increasingly become one of the central facets of the caregiving phenomenon. If resources need to be directed to informal caregivers ­­— not only to improve their own quality of life, but also to slow down deterioration in the cared-for (Leung, et al., 2007) or delay institutionalisation, which is much more frequent among demented people in all age groups (Jagger, et al., 2000), thus contributing to reduce the costs of health care (Bernabei, Landi and Zuccalà 2002; Michel et al., 2001; Stuart et al., 2005) — it is important to know under which circumstances a caregiver reports the worst consequences from his/her care activity, since this has relevant implications in terms of resources allocation.
In the light of the above considerations, the main aim of this study was to explore in depth the hypothesis that dementia caregiving has a stronger impact on caregivers than non dementia caregiving. We focused on caregiver burden, stress, and psychological well being. We reasoned that the inconsistency of previous findings in this area may be due to the existence of influential socio-demographic differences between dementia and non dementia caregivers which have not been appropriately taken into consideration. For example, as discussed earlier, caregiving spouses often show a higher level of distress than other relatives (Pinquart and Sorensen, 2003; Schulz, et al., 1995, 2004); therefore, if significantly more spouses are included among dementia caregivers in comparison to non dementia caregivers, the former will show a higher level of distress that may be unrelated to the patient’s dementia status. The reverse — with consequent results — is also possible. In the same way, evidence exists for the effect of caregiver gender and age, with female and older caregivers reporting higher levels of depression than, respectively, male and younger caregivers (Baumgarten, et al., 1992; Russo, et al., 1995). Care recipients’ characteristics may also play a role in this respect. To give an example, it was found (Miyamoto, et al., 2002) that caregivers of mobile demented patients report a greater burden than those of non mobile demented patients, thus suggesting that, when investigating differences between dementia and non dementia caregivers, patient functional status should be included in the model. Most previous research doesn’t seem to have paid enough attention to all these factors, which may explain why in some studies dementia caregivers reported greater strain than non dementia caregivers, while in others few (if any) differences emerged between the two groups.
On this background, we decided to further explore the comparative impact of dementia caregiving by using a three-step procedure of data analysis. We first developed a detailed description of the dyad caregiver-demented/non demented care recipient, to evaluate the extent to which the two chosen sub-samples were comparable. We then performed multivariate regression analyses, in order to determine whether patient dementia status significantly affected caregiver burden, stress and psychological well-being, once the effect of potentially confounding factors has been taken into account. Finally, by using a multivariate analysis of variance we sought support for a second hypothesis, namely that spouse caregivers would report worse consequences from caregiving than non spouse caregivers. This would bring further evidence for the particularly critical position of this group of carers.

2. Subjects and Methods

2.1 Sample

The overall sample was constituted by means of a random extraction from the lists of users of the ‘Scheduled Home Care’ (SHC) service provided in six National Health System districts of Central and North Eastern Italy (Ancona, Camerino, Fabriano, Senigallia, Bologna-Sud and Ferrara), which all granted ethical approval to the study. The SHC is a service targeted at patients who have been assessed by the local Health District’s Multidimensional Assessment Unit and officially classified as “unable to reach the doctor’s general practice”, so that the general practitioner is granted a supplementary allowance to visit them at home, up to four times per month. This criterion of SHC eligibility allowed therefore the inclusion in the overall sample — represented by 413 older people (mean age: 84.44 years; range: 65-105) and their primary family carers (mean age: 60.07 years; range: 20-87 years) — of only caregivers of older people presenting at least a moderate degree of functional disability.
In this article we focused on two subgroups of this general sample: dementia and non dementia dyads. We identified dementia dyads as those in which the care recipient both suffered from one of the following illnesses (identified by means of dichotomous ‘yes/no’ items in the care recipient’s questionnaire): Alzheimer disease, Parkinson disease, dementia or senile dementia and reported a severe impairment (i.e., highest cognitive disability) as scored by the Short Portable Mental Status Questionnaire (SPMSQ) (Pfeiffer, 1975). We considered as non dementia dyads only those in which no formal diagnosis of any of the above mentioned illnesses was reported, and in which the care recipient showed no impairment (intact cognitive abilities) at the SPMSQ. We adopted this classification of the dyads, which is more stringent than that used in other studies (e.g. Ory, et al., 1999), to avoid misclassification of patients with mild and moderate cognitive disability, who might actually be affected by dementia or pre-dementia despite the lack of a formal diagnosis (Galluzzi, et al., 2001; Honig and Mayeux, 2001; Ravaglia, et al., 2003), whose disclosure is by many carers often not wished (Corner and Bond, 2004; Fahy, et al., 2003). By means of this procedure, we obtained two clearly distinct groups of dyads, to better show possible differences between dementia and non dementia caregivers. Moreover, since 22 care recipients attended a local Alzheimer day care centre on a more or less regular basis, we decided to exclude them and their caregivers from our analyses, in order to avoid possible bias due to the potentially positive effect due to the support received by using such service. This left data available for 84 non dementia dyads and 53 dementia dyads.

2.2 Methods and Statistical Analysis

Questionnaires: data have been collected by means of two semi-structured questionnaires, administered separated to the care recipient and to the caregiver by trained interviewers. Wherever possible, data were collected in a single meeting with the members of the dyad, during which the care recipient and caregiver were successively interviewed. In many cases, the caregiver attended the care recipient interview and, if appropriate, gave her/his support whenever the latter was not able to respond personally.
Care recipient questionnaire:
• In addition to basic socio-demographic measures, a module of 17 items from the EASY-Care questionnaire (Philp, 2000; Richardson, 2001) assessing patient functional disability was included. These items concern the ability to carry out activities of daily living (ADL) and instrumental activities of daily living (IADL), assigning a score to each response according to the relative weight of the explored activity in determining the overall functional disability. A total score of disability is obtained, which ranges from 0 to 100, with higher scores indicating higher functional disability. Cronbach’s alpha for this scale was .89 in the present study.
Patient cognitive impairment was measured by administering the Short Portable Mental Status Questionnaire (Pfeiffer, 1975). We chose this tool because of its brevity (which prevented a possible drop out by the least cognitively fit), for it can be correctly administered by persons without formal neurological training (Welch and West, 1999), and because its scores are less easily affected by education, age and social class than for instance the Mini Mental Status Examination (MMSE), even though the MMSE has better sensitivity and specificity (McKenzie, et al., 1996; Yeh and Liu, 2003). The SPSMQ is made up by ten items assessing patient orientation and general/personal information. Its total score is determined by counting the number of errors made by the patient (with scoring taking into account the patient’s educational level), and this is used to distinguish between four categories of cognitive functioning: intact, mild impairment, moderate impairment and severe impairment. Cronbach’s alpha for this tool was .93 in this study.
In addition, a measure of patient chronic conditions was used as an indicator of care recipient health status. This was accomplished by administering a checklist of 20 of the most relevant chronic pathologies reportedly diagnosed in the Italian population (heart diseases, diabetes, cancer, arthritis etc.), for each of which a ‘yes/no’ response was requested. The illnesses reported by the patient were added up in a new variable indicating the total number of chronic conditions affecting the patient. Alzheimer disease, dementia and senile dementia were excluded from the computation, since these were employed as inclusion criteria for the categorization dementia/non dementia dyads. On the other hand, Parkinson’s disease was included in the computation, since not all of the patients affected by this disease are dementia patients.
• Caregiver questionnaire:
Caregiving circumstances. In addition to standard socio-demographic characteristics, the caregiver was asked about his/her relationship to the cared-for relative, and his/her place of residence in relation to the cared-for. Caregiving involvement was estimated in terms of duration (‘How long have you been caring for your relative?’) and of intensity (‘How many hours per week do you dedicate to the different caregiving activities?’). An indication of the social support available to the caregiver was also obtained, by asking him/her the following question: ‘How many persons may you rely upon, in case of necessity (please indicate up to a maximum of ten)?’. The total number of persons indicated by the caregiver was the measure of social support used in this study.
Health condition. Caregiver health status was estimated by asking the ‘yes/no’ question: ‘Are you currently suffering from any particular illness?’. The total number of illnesses the caregiver was currently affected by was used as a health indicator, which was integrated by items indicating whether the caregiver had been hospitalised in the last year (‘yes/no’), as well as the total number of medications currently used by him/her. Two further questions explored caregiver smoking and drinking habits, while a final, summarising item asked: ‘On the whole, how do you judge your current health condition?’ (1 = very good; 5 = very bad). This last question may be a powerful predictor of patient health, since self-reported health questions have been found to predict morbidity and mortality even better than physicians’ ratings (Ferraro, et al., 1999; Spiers et al., 2003).
Caregiver Burden. The caregiver was administered the Montgomery Burden Inventories, which investigate the two constructs of objective and subjective burden (Montgomery et al., 1985). We used seven items from the objective burden subscale, to explore how the caregiving activity has influenced seven different spheres of caregiver personal life (time to oneself, privacy, personal freedom, etc.). Responses for this tool are given on a five-point scale (-2 = improved a lot, +2 = worsened a lot, in this study). A Cronbach’s alpha of .81 was obtained here. With regard to the subjective burden measure, we used 11 items assessing how often the caregiver experiences particular feelings (usefulness, depression, guilt, etc.) in the relationship with the assisted relative. Responses are given on a five-point scale (1 = almost always, 5 = rarely or never). Cronbach’s alpha was .58 for this tool.
Stress. Caregiver stress was measured by the Relative’s Stress Scale (Greene, et al., 1982), which consists of fifteen items assessing the amount of stress and upset experienced by family caregivers of elderly relatives. Responses to items are given on a five-point scale (0 = not at all, 4 = considerably). Cronbach’s alpha was .88 for this instrument.
Anxiety. The Zung Self-rating Anxiety Scale (Zung, 1971) was used to obtain a measure of caregiver anxiety. This scale is made of twenty items and responses are given on a four-point scale (1 = none or a little, 4 = most or all of the time, for 15 items; the reverse for the others). Cronbach’s alpha was .86 in the present study.
Depression. The Zung Self-rating Depression Scale (Zung, 1965) was used to measure caregiver depression. Twenty items make up this psychometric tool and responses are given on a four-point scale (1 = rarely, 4 = almost always, for 10 items; the reverse for the others). In this study, a Cronbach’s alpha of .85 was obtained for this measure.

Caring for a Family Member with Dementia: Evidence from a Cross Sectional Comparative Study on Caregiver Burden and Psychological Well-being: This study is based on data collected for a project co-funded by the Italian National Research Council (CNR grants n. 96.01583.CT10, 97.01296.CT10 and 98.01504.CT10) and the European Union (contract n. BMH4-98-3374).

Cristian Balducci: Department of Cognitive Science and Education, University of Trento, Via Matteo del Ben, 5, I-38068 Rovereto (TN), Italy,
Maria Gabriella Melchiorre: Corresponding author: Maria Gabriella Melchiorre, Department of Gerontological Research, I.N.R.C.A. (Italian National Research Centre on Aging:, Via S. Margherita 5, 60124 Ancona, Italy, Tel: +39-071-8004887; Fax: +39-071-35941; e-mail:
Sabrina Quattrini, Giovanni Lamura: Department of Gerontological Research, I.N.R.C.A. (Italian National Research Centre on Aging: Via S. Margherita 5, 60124 Ancona (AN) – Italy.

Pages: 1 2 3

Tags: ,