This study is one of the first on HRQoL and caregiver strain in an Australian sample of people caring for family members with PD. By comparison with the literature, both the caregivers and the care-recipients in our sample demonstrated relatively high levels of HRQoL, although that of the people with PD was, as expected, somewhat lower than their caregivers. Previous studies have reported reduced HRQoL in groups of people with PD of similar age and disease duration [29, 30] when using the EQ-5D VAS. Others have noted greater emotional distress  and increased rates of depression in caregivers of people with PD [5, 32], findings that suggest caregiver HRQoL may also be reduced. In the current study, however, the mean EQ-5D VAS scores in our cohort of caregivers were similar to non-caregiving Australians of comparable age . Further, it appeared that caregivers in our sample had higher scores for HRQoL than cohorts in the Netherlands [15, 34], Spain [10, 35] and the UK [36, 37]. However, as the Australian value was within one standard deviation of other results, this difference should be interpreted with caution.
Unlike previous studies [4, 38], we failed to find an association between HRQoL in the caregiver and care-recipient. Although Martinez-Martin et al.’s sample was similar to ours in age and mean Hoehn and Yahr status, they used different assessment tools for caregiver HRQoL (the caregiver-specific SQLC: Scale of Quality of Life of Caregivers ) and care-recipient HRQoL (EQ-5D  and PDQ-8 ). The slightly different domains covered potentially impact on attempted comparisons across groups. Differences in culture or facilities between Spain and Australia may also affect the HRQoL of caregiver or of the people with PD. Peters et al.  found a moderate correlation in HRQoL between the mental component score of this measure in people with PD and their caregivers in their UK sample of > 700 dyads using the SF12 . Despite using the same instrument for both groups, differences between their cohort and ours may explain why they found a relationship between the two. The Peters et al. cohort was similar in age but may have had more long-standing disease; 35% of patients had been diagnosed for five to ten years and 32% for longer than 10 years, whereas the mean disease duration for our participants was 7.5 years. Another potential difference lies in the fact that the EQ-5D captures information about how people perceive their HRQoL at that time, whereas the SF-12 and PDQ-39 ask about perceptions over the past four weeks. This may lead to under-reporting from the EQ-5D.
Contrary to expectations, no relationship was found between caregiver HRQoL and caregiver strain, perhaps related to the relatively high HRQoL and low levels of strain reported by our cohort. As more than half of our caregivers reported optimal health, the impact of a ceiling effect to the EQ-5D cannot be discounted. An Australian study found a strong tendency of the EQ-5D to allocate an excess of responses a 1.00 value and therefore reduce the capacity of the instrument to discriminate at the upper end of the range . Although MCSI scores were more evenly distributed, some 80% of caregivers reported scores in the lower half of the scale, meaning relatively low levels of strain. At the individual level however, some caregivers reported near-maximal strain, reinforcing the potential of extreme strain in family caregivers.
In comparison with the literature there were no studies found examining the relationship between caregiver strain using the MCSI with caregiver HRQOL. While there have been several which have examined its precursor, the Caregiver Strain Index (CSI), correlations were not reported [37, 41]. However several studies have found a statistically significant negative correlation between the Caregiver Burden Index (CBI) and caregiver HRQOL (SQLC) [5, 42], whilst another has found a significant negative association with the Zarit Caregiver Burden Inventory (ZCBI) .
Our findings of reasonable HRQoL and fairly low caregiver strain may relate to the nature of our sample. Depression is the most commonly cited predictor of reduced HRQoL in people with PD  and known to have a negative impact on caregiver HRQoL  and stress, strain or burden [31, 44]. People with depression, whether patient or caregiver, may be less willing to participate in a research project, especially when the major study required 14 months commitment from participants. Further, dementia was an exclusion criterion (MMSE score of less than 24), potentially contributing to selection bias because cognitive issues in the care-recipient are known to increase caregiver burden  and distress .
Our group of caregivers was under less strain than has been reported in other papers, perhaps because of relatively higher HRQoL. For example, Peters et al.  reported a mean of 11.89 (SD 6.39) (maximum possible for maximal strain is 26 points) on the MCSI, whereas the median for our caregivers was only 5. This may reflect a longer disease duration and considerably poorer HRQoL of their Parkinson’s group (mean PDQ-39 index score of 41 (SD 18), compared to 23 (SD 12) for the participants with PD in our study).
In support of the idea that better HRQoL in care-recipients reduces strain on the caregiver, a significant correlation existed between the HRQoL in the individuals with PD and the level of strain experienced by their caregivers. Greater correlations were found using the disease-specific PDQ-39 than the generic EQ-5D measure. The added strength of the PDQ-39 correlation is not surprising, given that it is disease-specific, comprises 39 detailed items, and evaluates HRQoL over the past four weeks, rather than a single day as in the EQ-5D. Peters et al  similarly reported higher correlations between the PDQ-39 and MCSI (rs = 0.56, p < .001) than the generic SF-12 in care-recipients and the MCSI (physical component: rs = −0.33; p < .001; mental component: rs = −0.41, p < .001 [note that SF-12 and MCSI are scored in different directions, leading to negative correlations]).
Understanding the direct impact of patient HRQoL on the levels of caregiver strain is an important result. It has been suggested that families are “the most valuable and also the most vulnerable resource” in the treatment of PD . Recognition of this has led to the development of educational and support programs for family caregivers of people with PD [15, 34]. It also suggests that HRQoL should be a routine part of the assessment of patients with PD by health professionals, in conjunction with caregiver strain. Results from the PDQ-39 can inform the treating staff about how important various problems are to the individual, enabling management to be directed towards what is important to the patient, rather than most obvious to the therapist. PDQ-39 items such as difficulties dressing, carrying shopping or getting around the home, can be improved with targeted intervention by physical therapists, occupational therapists or nurses. Prioritising treatment by patient perceptions of difficulty provides emotional as well as physical benefits as the people with PD feels “listened to” and their opinion valued. Improved HRQoL of the care recipients may then flow on to help minimise caregiver strain.
The cross sectional design, appropriate for the aims of the study, meant that issues of causality between variables could not be investigated. A further limitation of this study is that depression was not measured. Depression is a notable factor that contributes to the HRQOL and caregiver strain [9, 28, 31, 32, 45] although it was beyond the scope of this study to include measures of depression. Similarly, other pertinent caregiver personal factors that may influence strain and HRQOL such as sense of coherence, coping styles and resilience were not examined. The above factors could be explored in future research on caregiver quality of life and caregiver strain.
This sample cannot be taken to be completely representative of all caregivers. The sample was one of convenience with ambulant clients who had volunteered. It is possible that volunteers with PD and their caregivers who were not overly burdened by their condition could have been more likely to volunteer, or the responses of caregivers may have been in a more positive direction than for the broader community of caregivers of people with PD. Nevertheless, the sampling methods were similar to most other studies which examined the HRQOL of caregivers throughout the world.