The nature and impact of caring for family members with a disability in Australia

Research Report No. 16 – June 2008

7. Literature review

Ben Edwards

Caring for a person with a disability can have considerable psychological and physical impacts upon those providing care; it can also have an impact upon the health of other family members. This chapter provides a review of the literature, examining how caring influences various emotional responses (depression, anxiety and anger, as well as positive responses), and highlighting gaps in our understanding of families' psychological response to caregiving. We then address the physical health of carers. We were unable to identify any studies that have examined the physical health impact of caring for any family member other than the primary carer and this is a major limitation of the research to date.

Mental health

Multiple family members

Very few researchers have examined the mental health of the primary carer and other family members caring for a person with a disability. The few studies that have focused on carers reported significant rates of depression. Edwards and Clarke (2005) reported that the incidence of clinical levels of depression38 among relatives of cancer patients was 20.8% within three months of diagnosis; six weeks later it was 18.6%, dropping to 10.0% at 6 months post-diagnosis. Families of patients with cancer in the palliation phase had even higher rates of depression39 (Kissane et al., 1994): 35.0% of spouses and 28.0% of children had clinical levels of depression, and 15.5% of spouses and 14.5% of children had severe distress levels.

Studies of families caring for a person recently diagnosed with cancer suggest that there was a significant correlation between family members' depression and anxiety (Edwards & Clarke, 2004). Over the course of 6 months, the correlations between family members' depression and anxiety reduced, but were still significant (Edwards & Clarke, 2005). The correlation between family members' depression and anxiety was lower for families caring for a person with cancer who was in palliative care than for families with a person recently diagnosed with cancer (Kissane et al., 2003).

A study of 251 families in which the mother and father cared for their adult children who had either a developmental disability or schizophrenia also showed that mothers' and fathers' psychological distress was very similar. In this study, Pruchno and Patrick (1999) found mothers' and fathers' depression and life satisfaction40 were not significantly different from one another and the correlation between the couples' levels of depression and life satisfaction were also significant. This suggests that families may experience similar levels of mental health problems in response to caring for a sick or disabled relative.

The mental health of family members of a person with a disability also seems to be affected by the amount of responsibility taken for caring (Perlesz et al., 2000). An Australian study of 79 families in which a person suffered acquired brain injury (ABI) showed that primary carers41 had significantly higher mean levels of depression, anxiety and anger42 than secondary and tertiary carers within the same family, as well as when the different types of carers were compared as independent groups (Perlesz et al., 2000). Primary carers also had greater levels of clinical levels of depression (37.7%) and anxiety (44.1%) than both secondary carers, who had a depression rate of 18.6% and anxiety rate of 15%, and tertiary carers, who had a depression rate of 27.3% and anxiety of 31.8%. These data suggest that, although other family members who have care responsibilities were affected by caring, primary carers are the most susceptible to mental health problems.

In summary, the few studies we were able to identify that focused on the mental health of the carer and of other family members who did not have a disability were limited, as they were based on convenience samples of carers for a specific type of disabling condition (Edwards & Clarke, 2004; Kissane et al., 2003; Perlesz et al., 2000; Pruchno & Patrick, 1999). Representative, population-based studies of families who care for a person with a disability do not appear to have been conducted. The available evidence does suggest that the mental health of carers and other family members were associated with one another, although primary carers' mental health was worse than secondary or tertiary carers.

Primary carers

The primary carer has been the focus of the majority of studies investigating the influence of caring for a person with a disability on carers' mental health. A series of meta-analyses of the mental health of carers provides some of the more accurate estimates of the emotional impact of providing care.

Differences between carers and non-carers

A meta-analysis examining the differences between carers and non-carers in terms of their psychological health was provided by Pinquart and Sorenson (2003). Their meta-analysis of 84 studies indicated that carers had significantly higher levels of depression43 and stress,44 and lower levels of general subjective wellbeing45 than non-carers. Spousal caregiving was particularly stressful, as there were large differences in the levels of depression and stress of carers who were spouses compared to non-carers. There were small but consistent and statistically significant differences in the levels of depression and stress between adult carers who were the children of the people with a disability and non-carers. Caring for a person with dementia was particularly challenging. When the differences between carers and non-carers were compared, studies of carers of people with dementia reported the highest levels of stress and greatest impairment in psychological wellbeing.46 Pinquart and Sorenson also noted that differences between carers' and non-carers' levels of depression and subjective wellbeing were less pronounced in population-based studies with representative samples of carers than those with non-representative samples. This finding underscores the importance of taking a representative sample of the general population in obtaining accurate estimates of the impact of caring for a person with a disability on mental health.

Gender differences in the mental health of carers

A meta-analysis of gender differences in the mental health of carers suggested that female carers had higher levels of depression and lower levels of wellbeing than male carers47 (Pinquart & Sorensen, 2006). Although these differences were significant, gender differences were only small in nature (explaining only 2.9% of the variation in scores on depression, carer burden and subjective wellbeing).

Are gender differences in carer mental health caregiving specific?

In the general population, women report higher levels of mental health problems than do men. Pinquart and Sorenson (2006) also investigated whether gender differences in mental health outcomes for carers simply reflected this population difference. A meta-analysis of gender differences in the general population indicated that there was a greater difference between female and male carers with respect to depression than there were between males and females who did not care for a person with a disability. However, gender differences in subjective wellbeing for carers was similar to the general population. Consequently, caregiving seems to have a particular effect on women's depression levels.

Population-based studies of the impact of caregiving on distress

Several epidemiological studies have investigated the impact of caring on primary carers' emotional distress, although none has focused on other family members. Using a representative sample of close to 10,000 people in England, Scotland and Wales, Hirst (2003) concluded that, for the period 1991-2000, male and female carers had higher rates of psychological distress48 than those who were not in a caregiving role. When a diagnostic interview has been employed to detect the prevalence of psychiatric conditions in epidemiological studies, similar findings have also been reported. In a population-based survey of 9,953 people in Ontario, Canada, that used a diagnostic interview,49 carers had higher rates of affective disorders (6.3% versus 4.2%) and anxiety disorders (17.5% versus 10.9%) than non-carers (Cochrane, Goering, & Rogers, 1997).

Some epidemiological studies have also been longitudinal and consequently enable an assessment of the psychological impact of transitioning into and out of the caring role. Longitudinal analysis of a nationally representative sample of 8,826 US adults followed for five years (1987 to 1992) indicated that transitioning into caring for primary kin50 was associated with increased levels of depression51 (Marks, Lambert, & Choi, 2002).

Transitioning to caregiving also led to an increase in the level of negative and positive psychological wellbeing, depending on the gender of the carer and whether the carer was providing support to a child, spouse, another family member or a friend (Marks et al., 2002). For example, the transition to providing care for a child with a disability was associated with higher levels of depression and hostility,52 but was also marked by a greater sense of autonomy53 for female carers than other women not in a caring role. Men who started caring for a child also had higher levels of depression than non-carers, were less happy and had a lower sense of mastery54 than other males who were not providing care for a child with a disability. There are several types of caring that were associated with positive outcomes for the carer, although they did vary between men and women. For instance, men who provided care for their parents-in-law reported more positive relations with others and having a greater purpose in life than men not providing care. Providing care to a friend was particularly positive for women. The transition to this type of care was associated with more autonomy, personal growth, purpose in life and self-acceptance than when not providing care.

One of the limitations of the study by Mark and colleagues (2002) was that it did not take into account the amount of caregiving provided by the carer. Other large-scale longitudinal community studies have addressed this issue and reported that higher levels of care were associated with higher levels of psychological distress. For instance, one US community-based population study that had information on 428 individuals showed that people who transitioned into heavy caregiving55 had significantly higher levels of depression and lower levels of mastery56 over a five-year period than people who were not carers or who had transitioned into a moderate caregiving role (Burton, Zdaniuk, Schulz, Jackson, & Hirsch, 2003). Another longitudinal population-based study, using data from the British Household Panel Survey, reported that, compared to non-carers, the risk for onset of high distress57 increased with the amount of time devoted to caring each week (Hirst, 2005). For instance, compared to non-carers, women who provided care for 10 to 19 hours per week were 1.7 times more likely to have high distress, while the odds of high distress were 2.9 times more likely for women caring 20 hours or more per week. Hirst also noted that ongoing care increased carers' vulnerability to recurring high distress. The increased risk of the onset and recurrence of high distress remained, even after caregiving ceased. For example, when caring for 20 hours or more a week, women had twice the risk for the onset and recurrence of high distress than women who had no caring responsibilities. Another epidemiological study of women also noted that excessive numbers of hours providing care (in this case, 36 or more per week) increased the odds of depressive or anxious symptoms by 5.6 times, even after controlling for a range of other factors, including previous depressive or anxious symptoms58 (Cannuscio et al., 2002).

The Australian evidence

In the last decade, there have been three major large-scale community-based epidemiological studies of the emotional impact of caring for a relative with a disability: the Victorian Carers study; the 2003 Australian Bureau of Statistics Survey of Disability, Ageing and Carers; and the Carer Health and Wellbeing Survey.

The Victorian Carers study was a large-scale, longitudinal, population-based survey of 976 carers,59 and 219 non-carers60 who formed a comparison group (Schofield et al., 1998). Female carers61 reported significantly lower levels of life satisfaction62 and positive affect, as well as higher levels of negative mood, compared to female non-carers. Over the course of 15 months, the life satisfaction of both female carers who continued to care and non-carers decreased, but at the same rate for carers and non-carers alike.63 However, at both measurement periods, female carers had lower levels of life satisfaction than their non-caring counterparts. The Victorian Carers study also examined differences in life satisfaction and positive and negative mood between continuing carers and those who were no longer caring.64 Between the first and third interview (30 months later), life satisfaction had decreased for those still caring, but had increased for those no longer doing so. For positive or negative mood, the rate of change over time was the same for continuing carers and those who were no longer caring. There were some subtle demographic differences in positive mood, with spouses who were no longer caring having significantly lower levels of positive mood than spouses who were continuing caring at both follow-up interviews.

Several differences in the emotional outcomes of carers at the first interview period were also noted for selected demographic characteristics, such as the type of disabling condition of the care recipient and the carer's gender. When care recipients were divided into those with physical impairment, undiagnosed memory loss, and Alzheimer's disease or dementia,65 and the carers' emotional outcomes compared based on these groups, carers of people with a physical impairment had significantly higher levels of life satisfaction and lower levels of negative affect than carers in the other two groups. Although there were no gender differences in levels of life satisfaction or positive affect, Schofield and colleagues did report that female carers had significantly higher levels of negative affect and a greater sense of role overload than male carers (Schofield et al., 1998).

The second major Australian study was the 2003 Survey of Disability, Ageing and Carers (SDAC; ABS, 2004). Carers were asked: "What physical or emotional effects have you experienced as a result of your caring role?" They could say yes or no to the following options: "Your physical or emotional wellbeing has changed", "You feel satisfied", "You feel weary or lack energy", "You frequently feel worried or depressed", "You frequently feel angry or resentful", and "You have been diagnosed as having a stress-related illness". These questions were not standardised measures of emotional wellbeing, which limits the validity of the findings. Carers' views of the emotional effects of caring will be the focus of discussion in this instance and further examination of carers' views of the physical effects of caring will be discussed in Chapter 10. As can be seen from Figure 7.1, approximately 1 in 3 carers frequently felt worried or depressed, while 1 in 7 frequently felt angry or resentful. One in 10 carers had also been diagnosed as having a stress-related illness. Despite these results, about 1 in 4 carers reported a sense of satisfaction from their caring role.

Figure 7.1 Carer's reports of emotional and physical effects of caring

Source: ABS (2003)

The final and most recent large-scale Australian study has been the Carers' Health and Wellbeing study, a survey conducted by Carers Australia, Australian Unity and Deakin University (Cummins et al., 2007). The survey consisted of 4,107 carers66 who were members of each of the state/territory Carers Associations, and they were compared with data from about 30,000 adults from the general population who had completed the bi-annual Australian Unity Wellbeing Index survey. Cummins and colleagues reported that carers had the lowest levels of personal wellbeing - a measure of life satisfaction - of any of the demographic groups that had been studied previously (e.g., people living alone and unemployed; people earning less than $15,000 and unemployed; people unemployed). When depression was considered, 19% of carers had extremely severe levels of depression, with a further 18% exhibiting severe levels and 19% moderate levels of depression, according to the clinical cut-offs developed for the Depression subscale of the Depression, Anxiety and Stress Scale (Lovibond & Lovibond, 1995).

There were also gender differences in relation to carers' personal wellbeing, depression and stress.67 Male carers had higher levels of personal wellbeing than female carers - the reverse of findings in the general population, where females had higher levels of personal wellbeing than males. Both male and female carers had significantly lower levels of personal wellbeing than their non-caring counterparts. Female carers also had significantly higher levels of depression and stress than male carers. Another notable finding was that the gap in personal wellbeing between carers and the general population was smaller when carers were older. Depression and stress followed a similar pattern to personal wellbeing, with carers' depression and stress significantly lower when carers were aged 55 years or older, compared to carers at younger ages.

Examination of differences in carers' mental health and wellbeing by household composition suggested that caring and having children was particularly stressful (Cummins et al., 2007). Carers with children (either sole parents or those with partners) had lower levels of wellbeing than carers without children. Sole parents had the highest levels of depression and partnered parents with a child had the highest levels of stress.

Factors associated with carers' mental health

There is a wealth of literature on factors associated with carers' mental health and psychological wellbeing. The reader is referred to several excellent reviews that focus on this topic (Cummins, 2001; Savage & Bailey, 2004). Rather than conduct yet another selective review of the factors associated with the mental health of carers, we focus here on meta-analyses of factors associated with carers' depression. Of the carer stressors that Pinquart and Sorenson (2004) examined in their meta-analysis, behavioural problems of the person with a disability had the strongest association with carers' depression. More hours of care per week had the next strongest association with carers' depression, followed by the person with a disability having cognitive impairments, the person with a disability having physical impairments, and the number of caregiving tasks. The years spent caring for the person with a disability had the weakest association with carers' depression in all of the studies analysed. When carers' resources were considered, informal support provided by friends or relatives had the strongest association with carers' depression, followed by carers' level of education, formal supports from professionals, and income (Pinquart & Sorensen, 2007). Behavioural problems also had the strongest association with carers' psychological wellbeing, followed by the person with a disability having physical impairments or cognitive impairments, and the hours of care provided. There was no statistically significant association between the number of years spent caring for the person with a disability and the carers' psychological wellbeing (Pinquart & Sorensen, 2007).

Limitations and future directions

Despite evidence suggesting that other family members are affected by caring for a person with a disability, most studies have focused on primary carers. Consequently, we were unable to locate any large-scale epidemiological studies that focused on the primary carer and other family carers. One suspects that this neglect has not been as a result of the view that other family members are not important, but rather that the additional resources to interview other family members may be too prohibitive. One method to obtain some data on this may be to ask the primary carer a series of questions about the influence that caring has on family members' mental health. Although there are some limitations to such an approach, the use of proxy reports of the mental health and wellbeing of people with a disability (Pierre, Wood-Dauphinee, Korner-Bitensky, Gayton, & Hanley, 1998) and the general population (Ellis et al., 2003) have been used in previous studies. The main limitation to such an approach is that proxy reporters (such as a relative) tend to under-report mental health problems, so in this context we would expect that a proxy report of the mental health impact of caring to be an underestimate.

The Victorian Carers study and Carers' Health and Wellbeing study have been the only Australian large-scale epidemiological studies to assess the mental health of carers using standardised measures (Cummins et al., 2007; Schofield et al., 1998). Both have limitations. The Victorian Carers study was a sample of carers residing in Victoria and the data collection was conducted in 1993. While more recent, the Carer Health and Wellbeing study only surveyed carers who were members of the state/territory Carers Associations, and only 37% of these members completed the survey. Representative recent national estimates of the mental health problems of carers would be important for policy-makers.

Physical health

As stated at the beginning of the chapter, we were unable to identify any studies of the physical health impact of caring for a family member with a disability for people other than the primary carer. Therefore, the focus of this review is on primary carers' physical health.

Primary carers

Differences between carers and non-carers

Meta-analyses have indicated that carers have poorer self-reported health than non-carers (Pinquart & Sorensen, 2003; Vitaliano, Zhang, & Scanlan, 2003). In addition to self-reported health, Vitaliano and colleagues' meta-analysis of the health consequences of caring for a family member with dementia indicated that carers also had a 23% higher level of stress hormones and a 15% lower level of antibody responses than did non-carers.

Gender differences in the physical health of carers

Meta-analyses of gender differences in the physical health of carers have suggested that female carers report worse physical health than male carers (Pinquart & Sorensen, 2006; Vitaliano et al., 2003). Moreover, the gender difference in self-reported physical health was greater than that observed in the general population (Pinquart & Sorensen, 2006). Although female carers had worse levels of physical health than male carers, Vitaliano et al. reported that male carers had slightly (but not significantly) higher levels of stress hormones and worse cardiovascular and metabolic outcomes than female carers. They speculated that these contrary findings may be explained by a tendency for women to be more aware of their health problems, or by the fact that self-reported health is related to distress and distress is higher in female carers than their male counterparts.

International population-based studies of the impact of caregiving on health

As was the case with mental health, there are several epidemiological studies of the physical health of primary carers.68 Perhaps the most well-known and influential epidemiological study of these is the US Carer Health Effects Study (CHES; Schulz & Beach, 1999). The Carer Health Effects Study was an ancillary study of the Cardiovascular Health Study, a large population-based study of the elderly. The CHES consisted of 392 carers, and 427 non-carers who served as a comparison group. Analysis of mortality risks after four years indicated that, after controlling for confounding factors,69 carers who were providing care and experiencing caregiving strain70 were 63% more likely to die than non-carers in the comparison group. Longitudinal analyses from the CHES also indicated that those who had transitioned directly into heavy caregiving had poorer self-reported health71 and health behaviours - such as missing doctors appointments, forgetting to take medications or not getting enough exercise - than non-carers (Burton et al., 2003). Data from 1,219 informal carers (15% of the total sample) in the Ontario Health Survey also suggested that physical health was affected in those providing care. Carers were 30% more likely than non-carers to have at least one physical health problem, even when controlling for participants' age, gender, employment status and economic disadvantage (Cochrane et al., 1997).

The Australian evidence

The Victorian Carers study, the 2003 Australian Bureau of Statistics Survey of Disability, Ageing and Carers, and the Carers' Health and Wellbeing study have been the three major epidemiological studies that have examined carers' physical health in Australia.

In the Victorian Carers study, female carers (n = 757) were compared to a group of 219 females who did not have caring responsibilities. Female carers were 43% more likely to have a major health problem, 33% more likely to use medication and 39% more likely to have poor health than females with no caring responsibilities. Longitudinal comparisons of carers and non-carers over 15 months indicated that carers were 33% more likely to have a continuing health problem, but not to continue to take medication. Although carers' and non-carers' self-rated physical health changed at the same rate over the course of 15 months, carers rated their self-rated physical health as significantly poorer than non-carers (Schofield et al., 1998).

The Australian Bureau of Statistics 2003 Survey of Disability, Ageing and Carers collected some information about the physical health of carers; however, these were combined with emotional wellbeing measures (ABS, 2004). Almost 30% of carers indicated that their physical or emotional wellbeing had been affected by caring, while one in three carers felt weary or lacked energy because of caring (see Figure 7.1).

The Carers' Health and Wellbeing study (Cummins et al., 2007) focused on carers' pain, injury caused by caring, and the medical or psychological conditions of the carer.72 Only 13.5% of carers did not experience any physical pain, compared to 36.0% of the general population. Almost two-thirds of carers (65.7%) reported they never carried an injury caused by their caring; of these carers, 21.3% reported they were more likely than not to be carrying an injury. Sixty-nine per cent of carers reported that they had a medical or psychological condition that warranted visiting a doctor; however, only 76% of these carers were receiving treatment for their condition. Many who were not receiving medical treatment said that this was because they had no time (27%) or for "multiple reasons" (23%).73 The majority of carers with a medical condition reported that they had had it for less than one year.

Limitations and future directions

With the exception of the Victorian Carers study, very little information has been collected on the self-reported physical health of carers in Australia; the Carers' Health and Wellbeing study was the only large-scale Australian study to focus on carers' pain, injuries and medical conditions. Moreover, even less information is available from large-scale epidemiological studies on other aspects of physical health, such as health behaviours, stress hormones or antibody responses. While the cost of collection of physiological information (such as stress hormones or antibody responses) is prohibitive for population-based studies, information about self-reported physical health, pain, injuries, medical conditions and health behaviours can easily be obtained through interview. Future studies could easily collect self-reported health information from carers; however, significant funds would be required to collect physiological information from a large group of carers.

Another major limitation of the research to date is that data have not been collected on the physical health of other family members, either in Australia or internationally. Again, the costs associated with a thorough examination of the physical health of family members in a family caring for a person with a disability may have prohibited such studies.

Conclusion

In summary, very few studies have examined the mental health of the carer and other family members who did not have a disability and we did not find any studies on the physical health of carers and other family members. The available evidence does suggest that the mental health of carers and other family members are correlated with one another; however, the studies were limited, as they were based on convenience samples of carers for a specific type of disabling condition (Edwards & Clarke, 2004; Kissane et al., 2003; Perlesz et al., 2000; Pruchno & Patrick, 1999).

International and Australian large-scale studies that focused on the mental health of primary carers suggest that:

  • carers have higher levels of depression and stress, and lower levels of subjective wellbeing than people with no caring responsibilities;
  • female carers have higher levels of depression and stress, and lower levels of wellbeing than male carers;
  • greater care needs (e.g., behavioural problems, cognitive and physical impairments of the person with a disability, and number of caregiving tasks) and more hours of care are associated with higher levels of carer depression; and
  • lower levels of depression are associated with informal support from friends or relatives, formal supports from professionals, and income.

The literature on primary carers' physical health suggest that:

  • carers have poorer self-reported physical health, higher levels of stress hormones and lower levels of antibody responses than non-carers; and
  • carers are more likely to have a major health problem, use medication and experience physical pain than non-carers.

Footnotes

38. The Beck Depression Inventory - Fastscreen (Beck, Steer, Ball, Ciervo, & Kabat, 1997) was used to measure depression.

39. Kissane et al. (1994) used the cognitive items from the Beck Depression Inventory to measure depression.

40. Depression was measured using the Centre for Epidemiological Studies - Depression scale (CES-D; Radloff, 1977) and life satisfaction was assessed using the Life Satisfaction Index (Neugarten, Havighurst, & Tobin, 1961).

41. Primary carers were considered by the family or the researcher to bear the most responsibility for caring for the person with acquired brain injury. Using the same criteria, secondary carers bore the next most responsibility and tertiary carers the next most responsibility.

42. Depression was measured using the Beck Depression Inventory, anxiety using the State Anxiety scale of the State-Trait Anxiety Inventory (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983) and anger from the Profile of Mood States (McNair, Lorr, & Droppleman, 1971).

43. Depression was mainly assessed with the Hamilton Depression Rating Scale (Hamilton, 1967; 15 studies), the Center for Epidemiological Studies - Depression scale (CES-D; Radloff, 1977; 13 studies), the Beck Depression Inventory (Beck et al., 1997; 11 studies), clinical interviews (4 studies), and other scales (13 studies).

44. Stress was most often assessed with the Perceived Stress Scale (Cohen, Kamarck, & Mermelstein, 1983; 6 studies), single-item indicators (2 studies), and other scales (14 studies).

45. Carers' subjective wellbeing was mainly assessed with life satisfaction scales (e.g., Life Satisfaction Index; Neugarten et al., 1961; 11 studies), positive affect scales (e.g., Positive and Negative Affect Schedule; Watson, Clark, & Tellegen, 1988; 9 studies), single-item indicators on happiness (5 studies), and other scales (9 studies).

46. The other groups studied were carers of people who did not have dementia or mixed disability groups.

47. Details of the measures of depression and subjective wellbeing are provided in footnotes 43, 44 and 45. Also, subsequent analyses that controlled for caring stressors, such as hours of caring and social resources (e.g., informal support) accounted for much of these differences.

48. Mental health was assessed using the 12-item General Health Questionnaire (GHQ; Goldberg & Williams, 1991).

49. Diagnoses from the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, revised, were generated on the basis of a structured diagnostic interview.

50. "Primary kin" was defined as a child, spouse or biological parent in need of assistance because of a physical or mental condition, illness or disability.

51. As measured by a 12-item version of the CES-D (Radloff, 1977).

52. Hostility was measured by 3 items: "On how many days of the past week did you feel irritable, or likely to fight or argue?"

53. Autonomy, personal growth, purpose in life, and self-acceptance were measured by three-item versions of Ryff and Keyes' (1995) psychological wellness scales.

54. Mastery was measured by a five-item version of the Pearlin Mastery Scale (Pearlin, Lieberman, Menaghan, Menaghan, & Mullan, 1981).

55. People were considered to be providing heavy caregiving if they helped another person with one of the following: eating, dressing, bathing, transferring, toileting and walking. Moderate caregiving referred to the provision of help to someone with one or more of the following: using the telephone, preparing meals, doing light or heavy housekeeping, shopping or managing money.

56. Depression was measured using a 10-item version of the CES-D (Radloff, 1977) and mastery was assessed using a seven-item version of the Pearlin Mastery Scale (Pearlin et al., 1981).

57. As measured by the General Health Questionnaire (GHQ; Goldberg & Williams, 1991).

58. The Mental Health Inventory of the SF-36 (Ware et al., 2002) was used to assess depressive or anxious symptoms.

59. Carers were identified by asking the following question: "Do you or does anyone in your household take the main responsibility in caring for someone who is aged or has a long-term illness, disability or other problem?"

60. This group were all adult females.

61. Female carers comprised 78% of the 976 carers.

62. Schofield et al. (1998) provided details of the 6 items used to measure life satisfaction. Positive and negative mood was measured using the Positive and Negative Affect Scales (PANAS; Watson et al., 1988).

63. Comparisons with carers and non-carers were not possible for negative and positive affect because non-carers' negative and positive affect were not assessed.

64. Carers who were no longer caring because the person whom they were caring for had died were excluded from this analysis, as information about their emotional wellbeing was not gathered in follow-up interviews. It should be noted that there were two follow-up interviews: at 15 months and 30 months after the first interview.

65. The physical impairment group included 186 carers reporting their relatives as being moderately or severely physically impaired, but with no intellectual impairment. The undiagnosed memory loss group consisted of 182 carers who described their relative as having memory loss, but did not mention Alzheimer's disease or dementia. The Alzheimer's disease or dementia group described their relative as having Alzheimer's disease, dementia or senile dementia. Most care recipients in the last two groups also had some physical impairment.

66. Questionnaires were sent to 11,200 members of the state/territory Carers Associations, so the response rate to the survey was 37%.

67. The Stress subscale of the Depression, Anxiety and Stress Scale was used to measure stress (Lovibond & Lovibond, 1995). The Personal Wellbeing Index (International Wellbeing Group, 2006) was used to measure personal wellbeing.

68. We were unable to identify any epidemiological study that focused on the physical health of multiple family members.

69. Age, gender, race (white cf. other), years of education, stressful life events, and prior physical health were controlled for.

70. "Caregiving strain" was defined as the carer reporting physical or emotional strain in providing help with one or more activities of daily living to the person with a disability.

71. Self-reported health was rated "excellent", "very good", "good", "fair" and "poor".

72. Information on carers' body mass index was also collected; however, the data was not presented, so the percentage of overweight or obese carers was not able to be determined by the authors, nor could comparisons be made with the general population.

73. In addition to "not enough time", the multiple reasons included: "appropriate treatment unavailable where I live", "difficulty with transport", "cannot afford treatment" and "other".