How well do the adult social care outcomes toolkit for carers, carer experience scale and care-related quality of life capture aspects of quality of life important to informal carers in Australia?

Internationally, there is a growing call to include informal carer costs and benefits in economic evaluations to evaluate the broader impacts of health and social care services [1, 2]. In this context, informal carers provide care beyond normal expectations within a pre-existing relationship (family member, relation, friend, neighbour) such as assistance with personal care, household activities or practical support, and generally do not receive payment for the care they provide [3, 4]. Over the last 15 years, three carer-specific preference-based instruments have been developed to measure outcomes for economic evaluations, the Adult Social Care Outcomes Toolkit for Carers (ASCOT-Carer) [5], Care-related Quality of Life instrument (CarerQol) [6] and Carer Experience Scale (CES) [7]. Whilst the number of economic evaluations including carer effects has grown in recent years, few have been conducted in the Australian setting [2, 8-10].

In Australia, there are about 2.8 million informal carers with over a third acting as the primary carer, i.e. the person who provides the majority of care [11]. In 2020, informal carers provided an estimated 2.2 billion hours of care, on average 786 h per year or 15 h per week (35.2 h for primary carers) [11, 12]. If services were purchased from formal care providers, the replacement costs would be $77.9 billion, almost 40% of the total spending on health in Australia in the same year [11, 13]. Given societies implicit reliance on carers’ willingness to fulfil this role and the economic consequences if this situation should adversely change, it is crucial that carer costs and effects are considered in economic evaluations [11]. In turn, this would also better inform healthcare decision-makers on actual societal costs, increasing the chance that welfare optimising decisions are made.

Two recent studies have investigated the relative construct and discriminative validity, test–retest reliability and responsiveness of the European-developed ASCOT-Carer, CarerQol and CES in a survey of Australian carers [14, 15]. Studies in England have also compared ASCOT-Carer, CES, CarerQol and EuroQol-5 Dimension-5 level (EQ5D-5L) [14, 16]. These studies indicate that the instruments tap into different constructs of carer-related QoL and caring experiences, reflecting the original purpose of the instruments and suggests the ASCOT-Carer, CES and CarerQol cannot be used interchangeably [16, 17]. The ASCOT-Carer was developed to measure social care-related QoL and support of carers in the setting of policy and formal support interventions [5, 18]. Whereas the CarerQol measures the impact of informal care on carers’ QoL, combining the burden of caring and valuation of their well-being (happiness) in the context of an evaluation in health care [6]. The CES captures the caring experience rather than carer’s QoL per se [19]. Validation has been investigated for the constructs in each of the instruments: the ASCOT-Carer with carers in England [5]; the CES with carers of older people in England [19] and the CarerQol with carers in eight European countries [6] [20]. Each of the instruments has preference weights, allowing the calculation of a summary score which reflects carers’ preferences for difference aspects of carer-related QoL [7, 21-23].

An Australian population was only included in one of the CarerQoL validation studies [22], with the general adult population using hypothetical carer scenarios. Content validation of instruments assesses relevance, comprehensiveness, and comprehensibility of the questions and the overall instrument; ensuring interpretation is as intended, all aspects important to the specific population are included and that the instrument’s constructs/domains measured as proposed [24]. Cross-cultural adaptations of instruments are important to capture differences in linguistics, colloquialisms, context and culture, even if translation is not required [25]. Qualitative research is ideally placed to validate the face value and content of instruments by exploring these social and cultural variables that may differ between informal carers in different countries [26]. Given the differences in health and social care support between countries [27-29], it is important to evaluate the applicability of these instruments in an Australian setting [25].

Consequently, the aims of this analysis were to identify aspects of carer QoL important to Australian informal carers and explore how well the ASCOT-Carer, CarerQol and CES constructs capture these aspects in the Australian context.

This analysis identified aspects of carer QoL important to Australian informal carers and explored how applicable the constructs of ASCOT-Carer, CarerQol and CES were to this population. Behaviour-mood of care recipient, Caring Responsibilities, Finances, Health, Own Life, Perceptions of carers, Relationship with care recipient and Support were identified as aspects of caring that affect carer QoL in Australian.

Comparing domains most importance in our sample with previous studies that developed preference weights and tariffs for the instruments showed mixed results. Occupation and control over daily life for the ASCOT-Carer where the most preferred among English carers [21], whereas our Australian sample found self-care and time and space to be yourself to be the most important. Suggesting that, with further investigation, Australian preference weights for the ASCOT-Carer may be in need of development. In our sample the importance of domains in the CarerQol (most important, mental health; least important, combining care and other activities) and CES (most important, getting on with care recipient and activities outside of caring; least important, control over caring) were in line with instrument tariffs [7]. CarerQol tariffs, developed for Australia, indicated mental health and combining caregiving with other activities as the most and least preferred [22, 23]. Whilst in a sample from United Kingdom, the CES found activities outside of caring and getting on with the care recipient as most preferred and control over caring the least [7].

Almost all participants reported aspects of carer QoL not captured by the carer-related instruments. Many of these aspects that were perceived as not covered by the instruments, could have been included in the domains. This was also the case in a study looking at patient, self-reported, QoL aspects not captured by EQ-5D-5L [39]. As caring experiences are subjective and responsibilities and challenges vary greatly between carers, they may have felt the domain did not completely encompass their experience of carer QoL. This reflects how some aspects of QoL, which can be important to individuals, cannot necessarily be translated into a question for a QoL instrument. Particularly when required to be applicable to a broad range of carers (e.g. caring for partner, child, parent), align with the construct of the instrument, and also fit with other considerations (e.g., timeframe).

Alternatively, respondents may have interpreted the questions differently or focused only on certain portions of the question (e.g. heading, examples or explanatory text). A content analysis of the end-of-life patient-reported outcome measure showed that interpretation of questions is related to individual circumstances, where in financial matters varying themes of money, investments, funeral arrangements and wills emerged [40]. This could also explain why similar domains across instruments were treated differently. The domain of support is present in all three instruments, however, respondents reported different types of support were missing in each instrument (i.e. formal and informal support for the carer and/or care recipient). Comparable results were also found in more detailed studies of the exploratory factor analysis using this same dataset, where only a moderate correlation was found between CarerQol and CES support items and also for relational problems [17].

Content and/or face validation of the three instruments has not been performed with Australian carers, so detailed information of how each instrument's questions are interpreted and understood by this population is not known. The broader project performed a content comparison of the three instruments showing they each perform well in measuring their relevant domains with Australian carers [15, 17]. However, the qualitative component of this study suggests that some differences in question interpretation may exist.

The majority of sub-themes (over 50%) related specifically to the care recipient. However, only two domains include aspects of carer QoL that are influenced by the care recipient (CarerQol’s relational problems and CES’s getting on with the care recipient). Consideration was given to a similar domain in the development of the ASCOT-Carer, however, it was omitted because it did not fit with the construct of the instrument (social care-related QoL/impact of care services on carer QoL) [18].

This relationship between the care recipient and carer has previously been proposed as an advantage of the CES in capturing broader aspects of caring [7, 15, 16]. Given the possible interdependence of care recipient and carer QoL [41-43], instruments capturing both could be included in economic evaluations of carer and patient interventions to fully capture the effects of an intervention [44]. However, consideration also needs to be given to the type of evaluation being performed, the perspective taken and the possibility of double counting which could overestimate the benefits of an intervention [2, 45]. Keeping these factors and participant burden in mind, an appropriate combination of instruments may be used in measuring carer QoL in studies focused on informal carers.

Some of the missing aspects of carer QoL in the instruments, as reported by our sample, are intentionally not covered by the instruments as they each have been developed with different intentions and measure different constructs of CrQoL. The CarerQol was developed and intended to measure the impact/burden of caregiving on QoL and so, appropriately, does not include any themes specifically about the care recipient [6]. Similarly, the CES missing themes of finance and health are reasonably missing as the instrument’s purpose is to measure the experience of caregiving. Health problems are not directly measured by CES, as qualitative research indicated that this was linked to other attributes included in the instrument [19]. And the ASCOT-Carer does not measure finance and health, as the instrument was developed as a measure of social care and support services on carer QoL. Financial hardship due to caring and health were considered in the early development of ASCOT-Carer [31], however, it was excluded as it was outside the scope of the instrument’s purpose. Although health was not considered as a separate domain in the ASCOT-Carer it is captured by the lowest QoL (high-level needs) response option for each item and indicates that the carer has high-level needs that, if unmet over time, put the carer at risk of poor physical and/or mental health.

Open-ended responses and quantitative data collected from a sample of Australian informal carers, suggest there are multiple aspects of caring that impact carer QoL. Consideration should be given to measurement of care recipient health and well-being and spillover effects affecting carer QoL, with thought to the risk of double counting.

The ASCOT-Carer, CarerQol and CES appear to be relevant for an Australian informal carer population and include most of the aspects of quality of life important to them. The interpretation of questions may differ in Australian informal carers which requires confirmation with a content and/or face validity assessment.

The findings support previous research that the selection of an instrument should take into account the aim, purpose and constructs of the instrument.