Caring for partners, dependents or relatives with chronic disease can have a substantial impact on the health of the caregiver [1–3]. This has been recognised by policy makers and the UK Department of Health (DH). A standing commission on caregivers was established in 2009 and the DH has been promoting a national strategy for caregivers since 1999. The physical and emotional impact associated with being a caregiver is also recognised by the National Institute for Health and Clinical Excellence (NICE) in their review of the cost effectiveness of new therapies. The NICE methods guide (2013) for manufacturers’ highlights how the indirect impact of a disease or treatment on caregivers can be considered when assessing the cost effectiveness of therapies. A treatment that maintains or improves a caregivers’ health related quality of life (HRQL) is a relevant benefit to be considered in an economic evaluation. NICE use relevant HRQL data (in the form of a utility) to estimate quality adjusted life years in cost utility analysis and has a preference for the EQ-5D [4].

Despite this, the impact of caregiving on health has infrequently been considered in health technology assessments (HTA) in the UK. Two reviews of outcomes data included in NICE appraisals [5, 6] make no reference to any appraisals that directly included caregiver HRQL data. Two appraisals (for Alzheimer’s disease and Multiple Sclerosis) included adjustments to mean utility values to represent the effect of treatment on carers HRQL, and both utilized the same Alzheimer’s caregiver data. From an economic evaluation perspective this seems to be an under researched area [7–9]. However, given the high prevalence of unpaid caregivers in a range of therapy areas we believe more work is needed. Further discussion of the role and importance of caregiving has recently been called for [10].

The present study aimed to measure the impact on HRQL of being an unpaid caregiver of someone with Multiple Sclerosis (MS) including the EQ-5D as the measure preferred by NICE for economic evaluations. Previous research has demonstrated the negative impact on HRQL of being a caregiver to a person with MS, with added burden associated with increased disease severity [11–14]. However, to the authors’ knowledge, no exploration of the potential for this impact to be captured by preference based utility measures suitable for economic analysis, such as the EQ-5D, has been conducted to date. Thus, the aim of the present study was to explore the overall HRQL of caregivers of people with MS including a utility measure compared to matched controls, and more specifically explore the impact of different levels of functioning in people with MS and how that related to caregiver’s HRQL compared to matched controls. This study was exploratory and was designed as a cross sectional survey of caregivers of people with MS and a matched control sample from the UK general population.

As shown in Figure 1, the MS caregivers’ health related quality of life was lower than the socio-demographically matched controls on all domains of the SF-36. The group comparisons were statistically significant across all domains and summary scores, with the greatest differences seen in relation to mental health and physical and social functioning. Similarly, the HADS results showed that caregivers of people with MS reported significantly higher levels of anxiety and depression than controls. Furthermore in terms of probable clinical ‘caseness’ , the HADS scores demonstrated that, the caregiver population contained a significantly greater proportion of probable anxiety ‘cases’ (χ2 = 38.83, p < .001) and depression cases (χ2 = 25.71, p < .001). The frequency of anxiety and depression probable clinical cases by population is presented in Table 4.

The EQ-5D utility values also demonstrated a significant difference between the controls and caregivers of people with MS: M = 0.82, SD = 0.25 and M = 0.74, SD = 0.28, respectively [t (398) = 2.99, p = .003]. The estimated disutility values associated with being a caregiver relative to the control population, stratified by PDDS MS severity and with all socio-demographic variables considered as covariates, are presented in Figure 2. These show lower caregiver utility compared to controls in line with greater loss of functioning in people with MS. At PDDS 0–1 the difference between caregivers and controls is negligible (− 0.002, SE = 0.053), increasing at PDDS 2 – 3 (− 0.045, SE = 0.057) and showing significant differences between caregivers and controls at PDDS levels 4, 5 and 6, [disutility - 0.142 (SE = 0.062), - 0.160 (SE = 0.055), - 0.173 (SE = 0.054), respectively]. At PDDS level 7 (a level where the person with MS would typically require use of a wheelchair) the caregiver’s quality of life is less affected (PDDS 7: - 0.030, SE = 0.038) and appears in line with disutility at PDDS 2 – 3. This is also true at PDDS 8 where the person with MS would be predominantly bedridden (PDDS 8: - 0.095, SE = 0.075). As illustrated by the error bars in Figure 2 (+ / - 1.96 * SE), only PDDS groups 4, 5 and 6 demonstrated significantly lower utility scores than the control population.

Post hoc analyses were conducted to explore the reduced disutility at PDDS 7 and 8 by evaluating the impact of receiving external caregiver support including professional assistance from social services. No difference in utility values were found for caregivers recei ving versus not receiving support (M = 0.76, SD = 0.23 and M = 0.77, SD = 0.22, respectively).

This study aimed to explore the HRQL of caregivers of people with MS compared to matched controls. There was a particular interest in the impact of disease severity on the disutility associated with being a caregiver compared to controls, due to the paucity of caregiver data currently being considered in economic analyses. The results demonstrated that caregivers of people with MS reported significantly lower HRQL, as measured by the SF-36, HADS and EQ-5D, compared to matched controls. There was also a significantly greater proportion of HADS anxiety and depression clinical cases amongst caregivers compared to matched controls. Furthermore, lower levels of functioning in people with MS was mirrored by lower levels of caregiver’s HRQL up until the point at which mobility problems required wheelchair use. However, once caregiver reported MS severity had progressed to this point the difference in health utility scores between caregivers and controls reduced, indicative of better caregiver HRQL at the most advanced levels of MS severity. The reason for this is unclear and while post hoc analyses suggested that this was not due to reduced burden from external caregiver support, this and other potential explanations could be usefully further researched in a longitudinal study.

The significant impact of providing care on HRQL demonstrated here provides support for the need to consider caregivers’ HRQL in economic assessments of new health technologies. Furthermore, contrary to concerns expressed by Al-Janabi et al. [7], the results also suggest that in line with the NICE reference case the EQ-5D may be an appropriate measure to capture caregiver HRQL (2013). However, we also believe that there may be merit in conducting further research into the development of a caregiver specific measure to capture their specific concerns. The high level of variability seen in the EQ-5D utility scores at each PDDS level suggests that core concepts associated with being a caregiver to a person with MS may not currently be captured by the EQ-5D, and this could be explored through qualitative research. For example, the pattern of disutility captured by the EQ-5D here suggests the potential physical and psychological impact associated with caring for an individual who is bedridden may not be captured by the EQ-5D. Another potential explanation for the reduced disutility observed with increased disease severity found here may be the development of coping strategies or changes in health/disease perception over time. A different study design would be needed to address this hypothesis properly.

The results from the current study reflect a different pattern of HRQL impact than the values used in the NICE HTA submission in MS that utilised caregiver data [6]. That assessment assumed significant disutility only occurred at the very severe stages of MS; an assumption based on data from caregivers of people with Alzheimer’s disease. While the current findings are exploratory and need to be considered in light of the study limitations and the potential limitations of the EQ-5D to assess HRQL in caregivers, they do suggest that caregiver HRQL is impacted earlier in disease progression than previously indicated. Thus, caregiver utility data may be an important consideration in the evaluation of new MS treatments associated with slowing MS progression from relatively early stages of disease.

This study has provided evidence of the significant negative impact caregiving can have on HRQL and supports the potential consideration of caregiver HRQL in economic assessments of new treatments. The results also suggest that the EQ-5D is an appropriate measure of caregiver HRQL, but given the exploratory nature of the study and the pattern of disutility associated with MS severity reported here; further research into the measurement of caregiver HRQL is also warranted. In relation to MS, the findings reported here suggest the impact of disease progression on caregiver HRQL should be considered relatively early when evaluating new treatments. Issues relating to exactly how caregiver disutility should be incorporated into economic assessments and consideration of other pertinent issues, such as cultural differences related to caregiving, also require further consideration.