Effect of caregiver burden on the quality of life of informal caregivers of people with cystic fibrosis in the United Kingdom: a cross-sectional study

This study used an analytical cross-sectional design through an online survey administered to informal carers of PwCF (see Appendix 1) between July–August 2020. The survey began with general questions about the carer. These questions included their relation to the PwCF, some clinical characteristics of the PwCF (including comorbidities and disease severity), their caring responsibilities (including time spent providing care), and demographics questions. The carer was then asked to assess their health using four validated PBMs: EQ-5D, ReQoL-10, CarerQol-7D and ASCOT-Carer. The order in which the PBMs were shown was randomised to mitigate the effects of fatigue on responses to each measure and to minimise biases due to question order effects [17]. Participants were recruited using convenience sampling through CF Voices, an organisation led by carers of PwCF in the UK. Eligibility criteria required participants to be adults who were actively providing care for a PwCF at the time of the study, a criterion inherently met by all members of the organisation. Ethical approval for the study was provided by the Senate Research Ethics Committee at City, University of London (reference: ETH-1920-1907).

To compare the distribution of the responses to the four measures, the responses were converted to utilities using value sets; utility values represent quality of life on a continuous scale from 0 (worst state) to 1 (best state) with different measures measuring different aspects of quality of life. Specifically, the EQ-5D-5L crosswalk value set developed by van Hout et al. [28], the ReQoL-10 value set developed by Keetharuth et al. [29], the CarerQol value set developed by Hoefman et al. [30], and the ASCOT-Carer value set developed by Batchelder et al. [31]. The utilities were analysed using Stata version 13.1® [32] through general mean and distribution analysis and ANOVA.

Severity of the CF experienced by the PwCF was reported by carers as either mild, moderate, or severe. Given the subjective nature of the scale, we sought to validate its use by comparing assessments based on more objective indicators that signal CF severity, such as the % Forced Expiratory Volume (%FEV1) which previous studies have shown to decrease as CF severity increases [33, 34]. Furthermore, we explored whether factors that in theory should not be related to the severity of CF were nevertheless influencing our severity variable, to further test its validity.where \({Y}_{i}\) represents the objective factor being explored and \(Severity\) is a categorical variable equal to 1 for mild, 2 for moderate, and 3 for severe CF. A significant beta coefficient would indicate that there is a relationship between the objective factor and carer-reported severity.

We explored how QoL in the carer differed by CF severity by conducting multivariate ordinary least squares (OLS) regression analyses. We controlled for additional factors, such as the number of other carers (apart from the primary carer filling out the form) and the daily overall care hours spent by the carer (as opposed to daily hours of medical-related care which would likely result in issues with collinearity with the severity scale). Demographic factors such as sex and education were shown to be correlated with the severity scale and were thus excluded to avoid collinearity.

We estimated a regression model for each QoL measure, as set out in Eq. 2:where \({QoL}_{i}\) represents the QoL scores, \({\alpha }_{i}\) is a constant term, \(Severity\) is a categorical variable equal to 1 for mild, 2 for moderate, and 3 for severe CF, Daily Care Hours is a continuous variable representing the number of daily care hours provided by the carer, and No. of Other Carers is the number of other individuals that provide regular informal care to the PwCF (self-reported). The severity variable is split into two dummy variables, indicating whether the carer reported that the PwCF has moderate or severe CF. Thus, this model explores the linearity of the QoL decrements between the severity levels by indicating how QoL scores decrease on average between those self-reporting moderate CF (compared to mild CF) and between those self-reporting severe CF (compared to mild CF).

This study contributes to the literature on the QoL impacts of the provision of informal care for PwCF. We included various validated preference-based instruments to measure different aspects of carer QoL.

Overall, our results suggest that the provision of informal care for PwCF impacts carers’ QoL. These findings align with previous studies on informal caregiving in CF. For example, Daly et al. (2022) [2] found that informal carers of PwCF experience significant health-related burden, particularly in mental well-being. Similarly, Quittner et al. (2014) [6] highlighted that parents of PwCF have elevated levels of anxiety and depression, supporting our findings on mental health impact. More broadly, Our results also complement broader research for other diseases, such as Lacey et al. (2024) [21], which demonstrated that caregiving is associated with worsening physical and mental health outcomes over time. Likewise, Chang et al. (2010) [23] reported that caregiving for patients with dementia led to increased anxiety and physical health deterioration, suggesting a broader trend across caregiving contexts. The effects of caregiver burden on carer QoL has been shown to be significant in many other disease areas recently [35, 36].

The EQ-5D-5L results suggest the most impacted dimension is anxiety/depression. The ReQoL-10 results provide further insight into the mental health of carers, suggesting that many feel lonely, constrained in their ability to do the things that they want to do, as well as struggling to trust others and feel confident in themselves. A small but significant proportion of carers report occasionally thinking that their life is not worth living, signifying the severity of this mental impact. The carer-specific measures included in the study provide some indication around what may cause these mental health impacts. The CarerQol-7D responses illustrate that most carers have some, or a lot of, problems combining care tasks with their daily activities. Many also report financial problems due to care tasks, and very few respondents felt that they are supported in their care tasks. These general insights are reiterated in the ASCOT-Carer responses. Most respondents are the primary carer, with the PwCF being their child. With the reported number of hours per day for providing care (> 10 on average), coupled with the severity of the CF (> 70% being moderate or severe) and the frequency of comorbidities in the PwCF, it is evident that this sample of carers experience a substantial caring burden. A smaller proportion of carers (7%) were caring for an adult with CF, either as a spouse, sibling, or other relative. These carers may experience different challenges, including navigating adult healthcare services and government processes on behalf of the PwCF, balancing employment with caregiving, and dealing with age-related comorbidities and disease progression [3, 21].

Nevertheless, the CarerQol-7D responses indicate strongly that the overwhelming majority of carers do get fulfilment from carrying out care tasks despite experiencing negative health impacts themselves.

We also explored the impact of disease severity on carer QoL. In CF, there is not a universally accepted severity scale. We opted to collect carer-reported data on CF severity as our primary measure of severity, using three categories: mild, moderate, and severe. To test the validity of this scale, we also collected data on a range of CF characteristics and explored how sensitive the scale was to these characteristics. We found the scale to be robust with significant correlations with measures of CF progression such as %FEV1 and previous hospital admissions. This suggests that carers of PwCF have an accurate insight into the severity of CF in the person that they care for.

Our results suggest that CF severity does have a significant negative impact on carer QoL according to nearly all measures, as might be expected. Although the utilities from each measure are not directly comparable, it is notable that the magnitude of the impact, which is relatively small, is similar for each measure. The level of sensitivity varied by measure. It appears that the impact of severity on the carer-specific measures is largely driven by those reporting severe CF, whereas in contrast, the generic measures appear to be more sensitive to the different levels of severity. A possible explanation is that the experience and provision of care is not substantially different for carers of people with mild or moderate CF, but that the mental health impact of caring for a PwCF with moderate CF is greater than with a PwCF with mild CF, perhaps due to the progressive nature of the disease. Alternatively, it may simply be the case that the carer-specific measures are less sensitive to milder care burdens, perhaps due to the wording of the dimensions and response levels, or perhaps even due to the lower number of response levels (five for EQ-5D-5L and ReQoL-10 vs. three and four for CarerQol-7D and ASCOT-Carer respectively).

This study has several strengths. Firstly, our sample size is large considering the relative rarity of the disease. Also, we have collected data using four validated measures that cover different constructs. This range of measures enabled us to obtain a broader picture of carer QoL, and the detrimental effects that caring duties can have on QoL. Furthermore, we intentionally focused on collecting data using preference-based measures, as these are more relevant for health technology assessment, and relatively few studies have used these measures in this context to date.

This study has several limitations. First, we used convenience sampling via CF Voices, a parent-centric organization, limiting generalisability to the broader CF carer population. Second, data were collected in July–August 2020, when COVID-19 restrictions were still in place. This may have exacerbated reporting of limited activities, loneliness, and mental health issues, potentially biasing results and making QoL appear worse than in a ‘normal’ period. Meaningful comparisons with ‘carer norms’ would require data from the same time. Additionally, carer age was not robustly collected, preventing age adjustment in our models. However, we have no reason to believe carer age varied significantly by CF severity, given the parent-focused recruitment.

Further research is advised now that more time has passed since the UK’s COVID-19 lockdowns. We recommend using multiple measures to assess carer QoL, including EQ-5D-5L for general health assessments and ReQoL-10 for mental health. CarerQol-7D appears more sensitive to disease severity than ASCOT-Carer in our study, though both may be useful. The recently developed EQ-HWB [37] could also provide valuable insights. Longitudinal research would further clarify the impact of care provision and the performance of QoL measures especially with new CF treatments likely shifting the landscape of caregiving duties.