Background
Policymakers have recognized the crucial role of informal carers in supporting older and frail family members, friends, and neighbors and have since implemented benefits to improve their status and well-being [
1]. To assess the carers’ situation, a range of instruments [
2,
3] has been developed, such as the Carer Experience Scale (CES) [
4,
5], the Adult Carers Quality of Life questionnaire (AC-QoL) [
6], and the Adult Social Care Outcomes Toolkit for Carers (ASCOT-Carer), all addressing carers’ quality of life, the Zarit Burden Interview [
7,
8], addressing care burden. In comparison to other tools assessing carers’ quality of life (QoL), ASCOT-Carer, however, is the only one to have a corresponding instrument for measuring the QoL of long-term care service users [
2]. Reflecting both the carers' and service users' situations makes these instruments attractive for a comprehensive assessment of long-term care (LTC) from a QoL perspective.
ASCOT-Carer was developed in England and aims to measure QoL-outcomes of (care) service provision in informal carers [
2]. The instrument consists of seven distinct domains to capture areas of adult informal carers’ (family and non-kin unpaid carers) everyday life that may be affected by LTC services and/or services for informal carers [
3,
9]. The domains address
Occupation, Control over daily life, Social participation and involvement, Self-care (Looking after yourself), Personal safety, Feeling supported and encouraged, and Space and time to be yourself (see Table
1). Taken together, these domains form a concept known as “long-term care related quality of life (LTC-QoL)”, also commonly referred to as “social care related quality of life (SCRQoL)” in the English original [
9]. A survey instrument that captures the effects of LTC service provision on the caregivers’ QoL contributes to improving empirical evidence and thus to a better understanding of how LTC services meet the needs of informal carers in different aspects of their lives.
Table 1
ASCOT-Carer domains.
1. Occupation | Being sufficiently occupied in a range of meaningful, enjoyable activities whether it be formal employment, unpaid work, caring for others or leisure activities |
2. Control over daily life | Choosing what to do and when to do it, and having control over their daily life and activities |
3. Self-care (Looking after yourself) | Feeling able to look after oneself, in terms of eating well and getting enough sleep |
4. Personal safety | Feeling safe and secure, with concerns about safety can include fear of abuse or other physical harm or accidents, which may arise as a result of caring |
5. Social participation and involvement | Being content with their social situation, whereby social situation includes the sustenance of meaningful relationships with friends and family, as well as feeling involved and part of their community |
6. Space and time to be yourself | Having space and time in everyday life. Enough time away from caring to have a life of their own outside of the caring role |
7. Feeling supported and encouraged | Feeling encouraged and supported by professionals, care workers, and others, in their role as a carer |
The ASCOT-Carer instruments consist of an interview version with four response levels (INT4) and a self-completion tool with four response levels (SCT4)
1 which reflect different outcome states:
Ideal state (3)—individual’s preferences are met,
No needs (2)—individual’s needs are met but not to the desired level,
Some needs (1)—there are unmet needs, but no health implications,
High-level needs (0)—needs have an immediate or longer-term health implication [
2,
9]. The total ASCOT-Carer score is the total raw score of the seven domains and ranges from 0 (worst state) to 21 (ideal state). As each domain represents a distinct aspect of the latent construct SCRQoL, weighted combinations of indicators are recommended (Avila et al. 2015). ASCOT-Carer was developed as a preference-weighted measure [
9] to reflect the
value of the care service-induced gain in QoL.
The ASCOT-Carer instrument has gained considerable interest in non-English-speaking countries and has been translated into Japanese, Dutch and Finnish.
2 The availability of instruments in different languages makes it possible to investigate QoL-effects of LTC service provision on informal carers across nations and within a country among speakers of different languages. A culturally valid adaptation of the ASCOT-Carer instrument can contribute to strengthening the evidence base for policy decisions on how to improve LTC-service provision and the living situation of informal carers. For the translated instrument to be used with confidence, there needs to be cross-cultural equivalence between the valid translated and the original questionnaires [
10].
Methodological approaches aiming to establish and evaluate cross-cultural equivalence in questionnaires address both the translation and cultural adaptation process and the assessment of measurement properties of the translated instrument.
Translation and cultural adaptation (together also referred to as “cross-cultural adaptation”) seek to improve equivalence on a semantic, idiomatic, experiential and conceptual level [
11]. For this purpose, systematic multistep guidelines and qualification recommendations for key actors involved in this process have been established (e.g. [
11‐
13]). The assessment of the
measurement model of the construct is essential as it provides the starting point for the choice of methods to
evaluate the measurement properties of the culturally adapted instrument. Reflective and formative measurement models differ in terms of the relationship between the construct and its indicators [
14] with respect to the nature of the construct, the direction of causality and the characteristics of indicators [
15]. In formative models, to which the ASCOT-Carer instruments conceptually belong to, a latent construct (here SCRQoL) is formed by its items (seven ASCOT-Carer domains); variation in the latent construct of SCRQoL is caused by variation in the ASCOT-Carer domains. These seven domains define the construct and are thus not interchangeable, nor do they share a singular common theme. Adding or dropping an ASCOT-Carer domain may change the conceptual meaning of the SCRQoL-construct. In formative measurement models, content and construct validity thus play an important role for assessing validity, defined as the extent to which the interpretation of the results of the measure are warranted [
16], at the score and item level. Methods to assess structural validity, such as confirmative factor analysis, and internal consistency, measured by Cronbach’s alpha, are appropriate for reflective, but not for formative models, in which items may correlate positively, negatively or not at all [
14].
The aim of this study was to translate and culturally adapt the original English-language ASCOT-Carer instruments (INT4 and SCT4) into German and to examine aspects of validity of the translated instrument. We report on the results of assessing linguistic and content validity as part of the translation and adaptation process. In addition, we investigate construct validity of the translated measure using survey data for testing expected relationships of the German ASCOT-Carer instrument with comparator outcome measures (convergent validity) and with selected subgroup characteristics of informal carers (discriminative or known-groups validity). As the ASCOT-Carer instrument follows a formative model, each domain is relevant for defining the latent construct. Thus, we assess construct validity on both the score and item level of the translated instrument. We shed light on challenges for cultural adaption and validation and discuss solutions to these challenges that may inform future studies on adapting the English ASCOT-Carer instrument into other languages. A valid German version of the ASCOT-Carer instruments will be a useful tool for national and cross-national surveys on the effects of LTC services on the QoL of German speaking informal carers.
The rest of the paper is organized as follows: First, the methods section describes the approaches applied for cultural adaption, the methods for assessing linguistic, content and construct validity. Then, results of the current study are presented, followed by a discussion of the main findings and comparison to previous work on validation of the original English ASCOT-Carer.
Results part 1
Cultural adaptation during the translation process
Some sections of the translation of the ASCOT-Carer instruments into German required cultural adaptation to appropriately reflect the meaning for the German-speaking target group. First, a few English expressions challenged the forward translation into German (e.g. “control over daily life”, “to feel clean and presentable”, “thinking about myself”, “to be yourself in your daily life”) as there was no meaningful direct translation. Where no appropriate final wording was found in the translation process, alternative expressions were tested in the professional review and cognitive debriefing phases (e.g. for “control over daily life”). Second, we found that some translations into German resulted into back translations that did not literally match the original because of idiomatic expressions which were preferred to literal translations (e.g. ‘role’ as a carer was changed to ‘task’ of a carer). Third, contemporary German language seeks to account for the gender of the actors, e.g. ‘der Interviewer’ (male interviewer), ‘die Interviewerin’ (female interviewer). As English usually has no grammatical gender, the gender of German nouns (masculine, feminine or neutral) added complexity to sentences referring to actors. Last but not least, cultural adaptation was needed for the terms referring to specific actors (e.g. carers), services and institutions (e.g. the British National Health Service) in the English and Austrian long-term care systems.
Linguistic and content validity of the German ASCOT-Carer
Linguistic and content validity were also evaluated by cognitive debriefing interviews. Five female informal carers, aged 46–72, caring for their relatives between one and 20 years, were involved in the cognitive debriefing to assess how well questions and response options were understood and whether the wording was appropriate.
The cognitive interviews showed that the respondents’ comprehension of the German translation of the ASCOT-Carer instrument was satisfactory, as they were able to adequately explain their responses with respect to each of the ASCOT-Carer domains. Minor changes to the wording were made in cases where at least one interviewee expressed confusion or ambiguity or misunderstood a part of the item (e.g. the preferred translation for ‘support’ was ‘Unterstützung’ (assistance), not ‘Hilfe’ (help)). These changes to the translation were made in accordance with the English ASCOT development team. Respondents generally understood the four response options as intended and were able to distinguish between them without difficulties. We used different German expressions (‘wie ich es will/möchte/mir vorstelle’, i.e. ‘the way I want/would like’) for the ideal state phrased ‘as I want’ in the original English tool as they seemed more suited to the respective specific contexts.
The Control over daily life domain could not easily be translated over. Thus, two options for an adaptation were tested with informal carers. The cognitive interviews for this domain revealed that option 1 ‘nach eigenem Ermessen’ (at their own discretion) was perceived as too broad a concept whereas option 2 ‘selbstbestimmt gestalten’ (being able to influence) was viewed as more specific and tangible. The second option was better understood and, therefore, used in the final questionnaire for the quantitative data collection.
The domains Self-care, Social participation and involvement, and Occupation were understood as intended. As interviewees preferred short names for ASCOT-Carer domains, the final wording for the Social participation and involvement domain was ‘Sozialleben’ (social life) instead of ‘Soziale Kontakte und Engagement im gesellschaftlichen Lebens’ (social contacts and social involvement). One of the response options for the Occupation domain was changed to ‘Ich verbringe etwas, aber nicht ausreichend Zeit’ instead of ‘Ich verbringe einige, aber nicht ausreichend Zeit’ (‘I do some of the things I value or enjoy with my time, but not enough’), as the German term ‘einige’ was understood to mean ‘a lot of’. The German translation of the questions about Feeling supported and encouraged was associated, as intended, with the feeling of appreciation and empathy from others, but also with financial support. Some respondents thought about having support in general (the fact that a care worker comes) while others reflected more on the actual support they were receiving by particular care workers or services. Informal carers, who were not in direct contact with care workers, seem to be less likely to feel supported by them and tended to think about support in more general terms. For one respondent, the idea of feeling supported as an informal carer did not seem relevant (this might be due to the fact that the respondent was also employed in the care sector and did not feel the need to rely on getting support).
Personal safety turned out not always to be understood in the same way. Some respondents did not find it easy to define or delimit the concept of personal safety and included financial security. In addition, some reflected on safety in connection with formal help. Feelings of better safety seemed to result from receiving regular formal help and from having trust in the care workers. On the other hand, some characteristics of service provision, such as frequent changes of care workers, seemed to negatively affect feelings of safety. While some respondents refer to injury risk and physical strain due to caregiving, others had difficulties to understand how safety is related to their caregiving and had a different idea of safety in mind than being safe from accidents and abuse. These different ways to interpret ‘personal safety’ stress the importance of using the interviewer prompt with this domain (‘with ‘feeling safe’ we mean feeling safe from fear of abuse, being attacked or other physical harm, such as accidents, which are a result of your caring role’).
The German translation of the questions about Space and time to be yourself captured, as intended, the carers’ ability to have enough time away from caring and to have a life of their own outside of the caring role. One respondent interpreted ‘space’ literally as having a separate apartment, not shared with the cared-for person, and thus having a life outside of the caring role.
Based on these findings, a few recommendations for amendments to the questionnaire were sent to the translation agency, which incorporated changes in consultation with the Austrian researchers and the ASCOT development team. The translated version (following proofreading and final checks) was then taken forward in the field phase (quantitative data collection).
Discussion
The translation and cultural adaptation according to ISPOR TCA guidelines aimed to develop a German version of the ASCOT-Carer instruments (INT4, SCT4) that has sufficient linguistic, content and construct validity and can be employed in German-speaking surveys as a measure of LTC-QoL of informal carers.
The analysis of cognitive debriefing interviews provided evidence for linguistic and content validity. The interviews showed no major issues with comprehension of the German translation of ASCOT-Carer, except for the Personal safety domain, where the restriction to ‘as a result of caring role’ was not always picked up by the informal carers. Therefore, to ensure the understanding of the Personal safety domain, we recommend briefing interviewers to pay extra attention when addressing the domain and to stress the importance of the prompt included in the question.
In addition, we found solid evidence to support convergent validity as part of construct validity of the translated German ASCOT-Carer. The ASCOT-Carer score was significantly correlated with other measures of conceptually-related constructs, particularly to scales also capturing carers’ quality of life (such as EQ-5D index and measures of carers’ experience and burden). As to be expected, the weakest association was found with the EQ-5D index that seeks to capture health-care instead of long-term care-related QoL.
For most of the ASCOT-Carer domains, significant associations with conceptually related constructs were found, except for
Personal Safety, with no corresponding alternate measure in the data, and for the
Control over daily life domain and the CES item ‘control over caring’, which seem to measure different aspects of having control. The ASCOT-Carer domain
Control over daily life was intended to reflect a broader concept that may also be influenced by other areas of life, not only caring [
34], while the CES item has a narrower, more specific focus.
The comparison between subgroups of informal carers to explore discriminative or known-group validity focused on characteristics that are well supported by previous studies. As expected, informal carers with high care intensity or no opportunity to take a break from caring showed lower LTC-QoL, as informal carers who cared for service users with low cognitive skills and challenging behavior.
While the analyses presented in this paper supported the construct validity of a culturally adapted German version of ASCOT-Carer and may be useful as a reference for assessing the validity of ASCOT-Carer translated to other languages, there are some limitations to this study. To begin with, we did not match English and Austrian samples of informal carers to investigate cross-cultural validity as defined by COSMIN study design checklist [
13]. Second, we did not investigate measurement properties that require standardized interview data at two time points (e.g. test–retest reliability and responsiveness, such as the sensitivity of the instrument to changes of LTC service receipt over time) as this was beyond the financial means of this project. We thus encourage future research to assess these measurement properties.
The findings of this study provided good evidence for a culturally adapted German version of the ASCOT-Carer instrument. The cognitive debriefing interviews support its linguistic and content validity. Since almost all related constructs were significantly associated with the German ASCOT-Carer (score and individual domains) and in the expected direction (convergent validity), and since the same holds true sub-groups of informal carers (discriminative or known-group validity), there is good evidence for its construct validity. Furthermore, the analysis of qualitative as well as quantitative data comes to similar conclusions as reported for the original English ASCOT-Carer instrument [
2,
3] and therefore support the construct validity of the German translation. These findings support the use of the German ASCOT-Carer instrument to capture LTC-related QoL for informal carers in Austria and other German speaking countries and can thus be utilized for national evaluations of LTC outcomes and comparative studies. Research is encouraged to assess further measurement properties of the translated instrument.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.