The original is unfaithful to the translation.
—Jorge Luis Borges, On William Beckford’s Vathek
The excellent paper by Stringer et al. (Reference Stringer, Leroi, Sikkes, Montaldi and Brown2020) in this issue of International Psychogeriatrics reminds me of the enigmatic pronouncement quoted above by the famous Argentinian author Jorge Luis Borges (Campbell, Reference Campbell2018). The context of this quote is that Borges is musing on an observation that the English translation of the gothic French novel Vathek (1786) is superior to the original French because English language as a medium appeared to be better suited than 18th-century French to communicate the central message of that specific novel. A translation is a reincarnation that can be truer to the spirit of the original than the original itself (Campbell,
Reference Campbell2018).
It strikes me to be the case that in an analogous way, Stringer et al.’s cultural adaptation of Amsterdam Instrumental Activities of Daily Living Questionnaire (A-IADL-Q) (Sikkes et al., Reference Sikkes2012) for use in the UK (A-IADL-Q-UK) is a reincarnation that, in some ways, is superior to the original! Through a process of multistep iterative modification, they do not simply adapt the instrument for use in another country; they substantially refine the language and content of the instrument for contemporary technologically advanced British life.
Functional assessment is a vital task in cognitive assessment. It forms the boundary between mild cognitive impairment (MCI) and dementia, and therefore, tools such as the A-IADL-Q play a very important role. This study sets out to accomplish three objectives:
-
1.
Use a committee of lay and professional stakeholders to assess face and content validity of a translated version of A-IADL-Q and to generate new candidate items for inclusion in the questionnaire.
-
2. Assess whether the questionnaire items are of relevance to the lives of older adults in the U.K and to adapt these items to ensure relevance if needed.
-
3. Pilot-test the questionnaire in older individuals with subjective cognitive decline and MCI.
They were successful in achieving all three. Modifications to ensure cultural adaptation and on-going relevance to the lives of older British adults resulted in a 55-item version (the translated scale had 69 items; 20 items were added, 34 items were removed, and language was adjusted for 33 items), which tested well in the pilot phase, and demonstrated convergent validity with cognitive and functional measures.
An examination of the appendix document titled “Details of changes made to all IADL items for each step” in Supplementary Materials is particularly revelatory in understanding the process of cultural adaptation. Consider the test item “Did he/she prepare cold meals?” in the original questionnaire. During the review process by stakeholders, concerns are raised that it is unclear what the question is asking about. Developer review leads to the decision to change the language to “making a sandwich or snack” so that it clarifies what the question is referring to more specifically. While reviewing the item “Did he/she operate the coffee maker?,” concern is raised that it is “Not thought to be culturally relevant to people in the UK. Most people do not own coffee makers” which leads to a change in wording to “making a cup of tea or coffee.” One of the items that gets removed is “Does he/she use cheques?” as it was felt that this task is increasingly not done anymore. This reminded me of my time at the inpatient geropsychiatry rotation during my residency training where a trained nurse would administer a battery of functional assessments to patients suspected of suffering from cognitive impairments, and one of the tasks in the battery was to balance the cheque book. I would invariably think that I have never balanced a cheque book in my life, and I wondered how I would feel if I were subjected to such an obsolete task as part of a medical assessment.
A further aim of the study was to develop a self-report version of the informant-based questionnaire. Self-report versions are important in both clinical as well as research settings. Clinically, patients often do not have informants or informants are not always accessible. Secondly, self-report measures can be utilized far more widely and conveniently in research studies with larger samples, especially population-level studies, where it may not be feasible to individually identify and contact the appropriate informants. For these reasons, the development of the self-report version by Stringer et al. (Reference Stringer, Leroi, Sikkes, Montaldi and Brown2020) is a welcome addition to the field and will facilitate the inclusion of assessment of instrumental activities of daily living in future research studies.
Instruments are sensitive to small differences in wording. For instance, in a study by Marshall et al., responses to remembering to turn off the stove were sensitive to functional status (Marshall et al., Reference Marshall2015), but in the present study, responses to “using the hob” were not. These are similar items but not identical, and this minor difference may be all that is needed to capture that sensitivity to functional impairment.
A set of guidelines by Beaton et al. published in 2000 have provided invaluable guidance to researchers engaged in the task of cross-cultural adaptation over the past two decades (Beaton et al., Reference Beaton, Bombardier, Guillemin and Ferraz2000). The guidelines describe the five stages of cross-cultural adaptation as (i) initial translations; (ii) synthesis of translations; (iii) back translations; (iv) expert committee review; (v) testing the prefinal version; and (vi) review of the adaptation process by the developers/coordinating committee. In the Stringer et al. publication, we are essentially looking at steps (iv)–(vi), since the starting point is an already translated version of the A-IADL in British English.
When is cultural adaptation of an instrument needed? Ideally, it is required when an instrument is to be used in another country with another language, in another country with the same language, or in the same country in which the instrument was developed but in a new population segment such as immigrants. The goal of cultural adaptation is to achieve equivalence in four areas (Beaton et al., Reference Beaton, Bombardier, Guillemin and Ferraz2000): semantic equivalence (literal meaning of words is the same); idiomatic equivalence (idioms are replaced with an equivalent expression); experiential equivalence (if a particular experience does not exist in the target culture, it needs to be replaced by a similar item that is experienced); and conceptual equivalence (ensuring that words have the same conceptual meaning).
Cross-cultural adaptation recognizes that adapting an instrument for use in a different culture, country, or ethnic population requires more than a simple translation of the text, although a good translation remains a necessary part of it. What is needed is not primary linguistic equivalence but more importantly conceptual and cultural equivalence. Furthermore, it cannot be simply assumed that if an instrument is translated and adapted, then this will invariably result in retention of psychometric properties such as validity and reliability at item or scale level. The psychometric properties of the adapted instrument need to be empirically demonstrated. What are ways of demonstrating that the new instrument performs psychometrically similar to the original instrument? One can look at distribution of item responses, item to scale correlations, internal consistency, and comparison of the adapted instrument against similar tests performed in the original setting (Beaton et al., Reference Beaton, Bombardier, Guillemin and Ferraz2000).
Despite the existence of cross-cultural adaptation guidelines, fidelity to these guidelines in research literature when using assessment instruments in other cultures remains low. In a systematic review of neuropsychological tests for the assessment of dementia in non-western, low-educated, or illiterate populations, out of 44 studies, 15 studies did not describe any translation and adaptation procedures, and only 5 studies using an existing test described the steps of a complete adaptation procedure, which include translation, back translation, conceptual changes, review by an expert committee, and pretesting (Franzen et al., Reference Franzen2019). This is also true for measures of general health in migrant populations (Uysal-Bozkir et al., Reference Uysal-Bozkir, Parlevliet and de Rooij2013).
Another important aspect of the Stringer et al. study is the effective involvement of dementia experts, patients, and their families. Advisory boards and committees which include lay and professional stakeholders are increasingly being utilized in research these days, even outside the context of cross-cultural adaptation, and is an admirable trend. Martha Sajatovic and her research team (Sajatovic et al., Reference Sajatovic2016; Reference Sajatovic2018), to cite one example, have successfully utilized community advisory boards consisting of patients, family members, clinical providers, and community health care administrators to develop and refine self-management interventions for individuals with epilepsy and serious mental illness.
The realization that challenges in cross-cultural adaptation cannot be anticipated without engaging with stakeholders is illustrated well by considering the cross-cultural adaptations of Montreal Cognitive Assessment (MOCA): in the Korean version, “velvet” and “daisy” were changed to “silk” and “azalea,” respectively, because the original items were unfamiliar to Korean elderly people (Lee et al., Reference Lee2008). Velvet and daisy also have no Filipino translation, so in the Filipino adaptation they were replaced by “silk” and “rose” (Dominguez et al., Reference Dominguez2013). Elderly Filipinos during the adaptation were unable to name the rhinoceros correctly “because it seemed that the animal was not within the realm of their reality and vocabulary” (Dominguez et al., Reference Dominguez2013), so it was replaced by Owl. The researchers also noted: “In the abstraction subtest, the consensus was to change clock to weighing scale (‘timbangan’). This was done because the elderly pilot participants did not consider a clock to be a measuring instrument like a ruler. For these participants, a clock tells the time, but does not measure time” (Dominguez et al., Reference Dominguez2013).
I wonder if we pay enough attention to the question of ensuring relevance of an instrument within the same country and linguistic group which has experienced significant cultural changes with the passage of time. Conceptual and experiential relevance is an ever-evolving target. Societies around the world are experiencing rapid technological changes and associated lifestyle changes. Use of cheques, commonplace enough in 2010 to be included in the original A-IADL-Q, is now uncommon enough that the item had to be removed in the UK adaptation due to irrelevance. I suspect that would be applicable even in Amsterdam. It was in this sense I said that the U.K. adaptation is not merely an adaptation but an improvement to the original.
Instrumental activities of daily living are clearly more influenced by technological changes, but what about core psychiatric constructs such as depression and cognition? Do people experience depression in the same way that they did in 1960 when Hamilton Depression Rating Scale (HDRS) was developed? What if HDRS or MOCA underwent a process of refinement every 20–30 years through a committee of lay and professional stakeholders to ensure they remain relevant to the clinical and research concerns at hand? I am not suggesting at this stage that such steps be undertaken but I am contemplating the radical consequences of taking conceptual and experiential relevance seriously. We can thank Stringer
et al. for bringing such thought-provoking questions to the forefront.
You have
Access
The original is unfaithful to the translation.
—Jorge Luis Borges, On William Beckford’s VathekThe excellent paper by Stringer et al. (Reference Stringer, Leroi, Sikkes, Montaldi and Brown2020) in this issue of International Psychogeriatrics reminds me of the enigmatic pronouncement quoted above by the famous Argentinian author Jorge Luis Borges (Campbell, Reference Campbell2018). The context of this quote is that Borges is musing on an observation that the English translation of the gothic French novel Vathek (1786) is superior to the original French because English language as a medium appeared to be better suited than 18th-century French to communicate the central message of that specific novel. A translation is a reincarnation that can be truer to the spirit of the original than the original itself (Campbell, Reference Campbell2018).
It strikes me to be the case that in an analogous way, Stringer et al.’s cultural adaptation of Amsterdam Instrumental Activities of Daily Living Questionnaire (A-IADL-Q) (Sikkes et al., Reference Sikkes2012) for use in the UK (A-IADL-Q-UK) is a reincarnation that, in some ways, is superior to the original! Through a process of multistep iterative modification, they do not simply adapt the instrument for use in another country; they substantially refine the language and content of the instrument for contemporary technologically advanced British life.
Functional assessment is a vital task in cognitive assessment. It forms the boundary between mild cognitive impairment (MCI) and dementia, and therefore, tools such as the A-IADL-Q play a very important role. This study sets out to accomplish three objectives:
1. Use a committee of lay and professional stakeholders to assess face and content validity of a translated version of A-IADL-Q and to generate new candidate items for inclusion in the questionnaire.
2. Assess whether the questionnaire items are of relevance to the lives of older adults in the U.K and to adapt these items to ensure relevance if needed.
3. Pilot-test the questionnaire in older individuals with subjective cognitive decline and MCI.
They were successful in achieving all three. Modifications to ensure cultural adaptation and on-going relevance to the lives of older British adults resulted in a 55-item version (the translated scale had 69 items; 20 items were added, 34 items were removed, and language was adjusted for 33 items), which tested well in the pilot phase, and demonstrated convergent validity with cognitive and functional measures.
An examination of the appendix document titled “Details of changes made to all IADL items for each step” in Supplementary Materials is particularly revelatory in understanding the process of cultural adaptation. Consider the test item “Did he/she prepare cold meals?” in the original questionnaire. During the review process by stakeholders, concerns are raised that it is unclear what the question is asking about. Developer review leads to the decision to change the language to “making a sandwich or snack” so that it clarifies what the question is referring to more specifically. While reviewing the item “Did he/she operate the coffee maker?,” concern is raised that it is “Not thought to be culturally relevant to people in the UK. Most people do not own coffee makers” which leads to a change in wording to “making a cup of tea or coffee.” One of the items that gets removed is “Does he/she use cheques?” as it was felt that this task is increasingly not done anymore. This reminded me of my time at the inpatient geropsychiatry rotation during my residency training where a trained nurse would administer a battery of functional assessments to patients suspected of suffering from cognitive impairments, and one of the tasks in the battery was to balance the cheque book. I would invariably think that I have never balanced a cheque book in my life, and I wondered how I would feel if I were subjected to such an obsolete task as part of a medical assessment.
A further aim of the study was to develop a self-report version of the informant-based questionnaire. Self-report versions are important in both clinical as well as research settings. Clinically, patients often do not have informants or informants are not always accessible. Secondly, self-report measures can be utilized far more widely and conveniently in research studies with larger samples, especially population-level studies, where it may not be feasible to individually identify and contact the appropriate informants. For these reasons, the development of the self-report version by Stringer et al. (Reference Stringer, Leroi, Sikkes, Montaldi and Brown2020) is a welcome addition to the field and will facilitate the inclusion of assessment of instrumental activities of daily living in future research studies.
Instruments are sensitive to small differences in wording. For instance, in a study by Marshall et al., responses to remembering to turn off the stove were sensitive to functional status (Marshall et al., Reference Marshall2015), but in the present study, responses to “using the hob” were not. These are similar items but not identical, and this minor difference may be all that is needed to capture that sensitivity to functional impairment.
A set of guidelines by Beaton et al. published in 2000 have provided invaluable guidance to researchers engaged in the task of cross-cultural adaptation over the past two decades (Beaton et al., Reference Beaton, Bombardier, Guillemin and Ferraz2000). The guidelines describe the five stages of cross-cultural adaptation as (i) initial translations; (ii) synthesis of translations; (iii) back translations; (iv) expert committee review; (v) testing the prefinal version; and (vi) review of the adaptation process by the developers/coordinating committee. In the Stringer et al. publication, we are essentially looking at steps (iv)–(vi), since the starting point is an already translated version of the A-IADL in British English.
When is cultural adaptation of an instrument needed? Ideally, it is required when an instrument is to be used in another country with another language, in another country with the same language, or in the same country in which the instrument was developed but in a new population segment such as immigrants. The goal of cultural adaptation is to achieve equivalence in four areas (Beaton et al., Reference Beaton, Bombardier, Guillemin and Ferraz2000): semantic equivalence (literal meaning of words is the same); idiomatic equivalence (idioms are replaced with an equivalent expression); experiential equivalence (if a particular experience does not exist in the target culture, it needs to be replaced by a similar item that is experienced); and conceptual equivalence (ensuring that words have the same conceptual meaning).
Cross-cultural adaptation recognizes that adapting an instrument for use in a different culture, country, or ethnic population requires more than a simple translation of the text, although a good translation remains a necessary part of it. What is needed is not primary linguistic equivalence but more importantly conceptual and cultural equivalence. Furthermore, it cannot be simply assumed that if an instrument is translated and adapted, then this will invariably result in retention of psychometric properties such as validity and reliability at item or scale level. The psychometric properties of the adapted instrument need to be empirically demonstrated. What are ways of demonstrating that the new instrument performs psychometrically similar to the original instrument? One can look at distribution of item responses, item to scale correlations, internal consistency, and comparison of the adapted instrument against similar tests performed in the original setting (Beaton et al., Reference Beaton, Bombardier, Guillemin and Ferraz2000).
Despite the existence of cross-cultural adaptation guidelines, fidelity to these guidelines in research literature when using assessment instruments in other cultures remains low. In a systematic review of neuropsychological tests for the assessment of dementia in non-western, low-educated, or illiterate populations, out of 44 studies, 15 studies did not describe any translation and adaptation procedures, and only 5 studies using an existing test described the steps of a complete adaptation procedure, which include translation, back translation, conceptual changes, review by an expert committee, and pretesting (Franzen et al., Reference Franzen2019). This is also true for measures of general health in migrant populations (Uysal-Bozkir et al., Reference Uysal-Bozkir, Parlevliet and de Rooij2013).
Another important aspect of the Stringer et al. study is the effective involvement of dementia experts, patients, and their families. Advisory boards and committees which include lay and professional stakeholders are increasingly being utilized in research these days, even outside the context of cross-cultural adaptation, and is an admirable trend. Martha Sajatovic and her research team (Sajatovic et al., Reference Sajatovic2016; Reference Sajatovic2018), to cite one example, have successfully utilized community advisory boards consisting of patients, family members, clinical providers, and community health care administrators to develop and refine self-management interventions for individuals with epilepsy and serious mental illness.
The realization that challenges in cross-cultural adaptation cannot be anticipated without engaging with stakeholders is illustrated well by considering the cross-cultural adaptations of Montreal Cognitive Assessment (MOCA): in the Korean version, “velvet” and “daisy” were changed to “silk” and “azalea,” respectively, because the original items were unfamiliar to Korean elderly people (Lee et al., Reference Lee2008). Velvet and daisy also have no Filipino translation, so in the Filipino adaptation they were replaced by “silk” and “rose” (Dominguez et al., Reference Dominguez2013). Elderly Filipinos during the adaptation were unable to name the rhinoceros correctly “because it seemed that the animal was not within the realm of their reality and vocabulary” (Dominguez et al., Reference Dominguez2013), so it was replaced by Owl. The researchers also noted: “In the abstraction subtest, the consensus was to change clock to weighing scale (‘timbangan’). This was done because the elderly pilot participants did not consider a clock to be a measuring instrument like a ruler. For these participants, a clock tells the time, but does not measure time” (Dominguez et al., Reference Dominguez2013).
I wonder if we pay enough attention to the question of ensuring relevance of an instrument within the same country and linguistic group which has experienced significant cultural changes with the passage of time. Conceptual and experiential relevance is an ever-evolving target. Societies around the world are experiencing rapid technological changes and associated lifestyle changes. Use of cheques, commonplace enough in 2010 to be included in the original A-IADL-Q, is now uncommon enough that the item had to be removed in the UK adaptation due to irrelevance. I suspect that would be applicable even in Amsterdam. It was in this sense I said that the U.K. adaptation is not merely an adaptation but an improvement to the original.
Instrumental activities of daily living are clearly more influenced by technological changes, but what about core psychiatric constructs such as depression and cognition? Do people experience depression in the same way that they did in 1960 when Hamilton Depression Rating Scale (HDRS) was developed? What if HDRS or MOCA underwent a process of refinement every 20–30 years through a committee of lay and professional stakeholders to ensure they remain relevant to the clinical and research concerns at hand? I am not suggesting at this stage that such steps be undertaken but I am contemplating the radical consequences of taking conceptual and experiential relevance seriously. We can thank Stringer et al. for bringing such thought-provoking questions to the forefront.
Conflicts of Interest
None.