2.3.1 Interactional Variation: Introduction of the Memory Assessment

There is variation in the manner in which clinicians transition from the history-taking phase of the consultation to administering the ACE‑III (Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020). History-taking involves between 10 to 20 minutes of clinician-led questions in order to understand the person’s background and their current concerns. There is no guidance on how to introduce the test and each clinician completes this differently. Furthermore, as Jones et al. (Reference Jones, Wilkinson, Jackson and Drew2020) showed, variations in the introduction of the test appear to have implications for patients’ understanding, with some patients displaying uncertainty, as is evident from their embodied or verbal reactions. For example, some clinicians do not mention the test explicitly during the consultation itself, saying, for instance, ‘We’ll just run through a few quick questions, then I’ll examine you’ immediately before the test commences. Utterances of this kind, coming at the end of what is already a substantial period of questioning (during the history-taking), do not clearly convey that the upcoming questions are part of a discrete test, and this can therefore create confusion for patients. Other patients appear to be fully aware of the expectations for the next phase of the consultation when the clinician adopts a different approach: explicitly naming the test, providing information about the questions within it and how the test was validated (see Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020, for a full analysis).

Extract 1 shows another version of test introduction from a different clinician (not explored in Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020). In this version, the clinician introduces the upcoming test by both marking and naming a change of activity, characterizing it as a ‘memory test’ but without furnishing details.

In Extract 1, the clinician tells the patient that they are going to be subject to a memory test (lines 1–2). The turn design conveys the clinician’s high deontic and epistemic authority (see Stevanovic & Peräkylä, Reference Stevanovic and Peräkylä2012; Muntigl, this volume) to decide about what happens next (‘I’m going to…’; notwithstanding the tag at the end of line 2) and expertise to conduct this test ‘on’ the patient (line 2). The test itself is characterized as a memory test, but no further details are given. The patient confirms they are okay to continue – ‘Yeah’ (line 3) – thus claiming (but not displaying) understanding of what is coming next.

Extract 2 shows the moment of transition from history-taking to testing involving the same clinician and another patient. Here, the hierarchical stance is more collaborative with the clinician’s proposal that ‘we’ll do a memory test now’ (line 1) suggesting more of a joint activity for the clinician and the patient.

It is interesting to note that this clinician (Neu1) is the only one (of the five clinicians recorded) to routinely seek consent or permission from the patient to begin the testing phase of the consultation – ‘Is that okay?’ (Extract 1, line 2) and ‘Okay?’ (Extract 2, line 1), thus introducing further variation. Notably, though, in both these cases the clinician does not lift his gaze to the patient when seeking this consent, thus giving it a tokenistic quality. Instead, the clinician is engaged in gathering the test paperwork from a cupboard behind him and attaching the patient’s information to the test paper.

Further to exploring variation between the individual clinicians’ styles of introducing the formal cognitive assessment, we found that clinicians would alter their own practice for different patients. The clinicians in the memory clinic informed us during informal discussions that they generally form an initial impression of the patients’ abilities during the history-taking phase. They reported that within the first five minutes of talking to the patient they have generally formed a ‘working diagnosis’. Clinicians often use an initial ‘gut feeling’ as part of their clinical decision-making (Lindeberg et al., Reference Lindeberg, Samuelsson and Müller2019) and the accuracy of this clinical impression on determining a correct diagnosis has been positively assessed (Pond et al., Reference Pond, Mate, Phillips, Stocks, Magin, Weaver and Brodaty2013). With this in mind, we suggest that clinicians may be adapting their administration of the cognitive assessment tool according to their working diagnosis. That is, in these interactions clinicians can be seen to be (re)designing aspects of the test to suit the patient’s abilities or undertaking extra interactional work to prepare some patients for the potentially difficult nature of the test. This concept of recipient design (Sacks, Reference Sacks and Jefferson1992) relates to the relationship between the tester and recipient of the test. This was particularly evident during the introduction of the test.

The two patients in Extracts 1 and 2 were given a working diagnosis of non-progressive memory problems (which may be summarized under the term functional cognitive disorder (FCD)) after the initial consultation and subsequently received a formal diagnosis of FCD (Pennington, et al., Reference Pennington, Newson, Hayre and Coulthard2015). In contrast, the patients in Extracts 3 and 4) both received a dementia diagnosis from the same clinician (Neu1). Prior to the conversation in Extract 3, the patient had been asked a series of questions requiring her to recall several details regarding her family’s medical history, which she was struggling to remember, and she informed the clinician that he was ‘asking difficult questions’.

Extract 3 begins in a very similar manner to the previous two extracts, with the clinician informing the patient that a memory test will ensue (line 1), and (nominally) seeking consent, ‘Okay’ (line 2). Although the patient confirms readiness/willingness to perform the test – ‘Yeah’ (line 3) – this is said very quietly and is followed by an audible sniff. Sniffing can be interactionally relevant (Hepburn, Reference Hepburn2004; Hoey, Reference Hoey2020), though, to our knowledge, systematic analysis has not been conducted in this sequential environment. Together with the quietly spoken agreement, we speculate that this sniff is perhaps a display of unsureness or apprehension. Rather than commencing the test, as in the previous two examples, the clinician first orients to the patient’s possible ‘unsureness’. ‘So’ (line 4), in turn-initial position, could be seen to be ‘other-attentive’ (Bolden, Reference Bolden2006). In stating that the test is ‘designed to be a bit tricky and stretch your memory’ (lines 5–6), the clinician could be both orienting to his perception of the patient’s cognitive difficulties as well as to the patient’s lack of conviction in her confirmation to proceed with the test. The clinician is working here to absolve the patient should she find the test difficult by characterizing it as designedly difficult. The sense that the clinician is orienting to an already formed working diagnosis of more severe cognitive difficulties is supported further when he states that the questions would be ‘even harder questions unfortunately for you’ (lines 8 and 9). The questions may be designedly tricky, but here the clinician is expressing that the patient (with her particular cognitive deficits) would find them even harder to answer. The clinician here is engaging in additional interactional work to prepare the patient for the task ahead and, in expressing regret about the difficulty of the questions with ‘unfortunately’ (line 9), he is working to build solidarity and set (perhaps low) expectations. This introduction is therefore designed for this particular patient in this particular interaction.

Extract 4, again with the same clinician (Neu1), shows very similar features.

Similar to the collaborative introduction to the test in Extract 2 (‘we’ll do a memory test now’ (line 1)), here the clinician declaratively introduces the test by stating ‘So I’m going to do a memory test with you now’ (lines 1–2). He follows it up again with a tag question seeking consent to proceed, ‘If that’s okay’ (line 2). Apart from the slight change in formation from doing a test ‘on you’ (Extracts 1 and 3) to ‘with you’ (Extract 4), and the tag question being formulated with either ‘is that okay’ (Extract 1), ‘Okay’ (Extracts 2 and 3) and ‘If that’s okay’ (Extract 4), these are all very similar introductions to the test. Unlike in the other extracts, however, here the patient does not confirm that it is okay to proceed. The tag question’s format invites a polar response from the patient, but her response is not type-conforming (Raymond, Reference Raymond2003), suggesting potential disaffiliation. Possible resistance on the part of the patient is first adumbrated by a gap (line 3) followed by an audible sigh (line 4). Hoey (Reference Hoey2013) suggests that sighing has an indexical relationship with emotion – specifically negative emotion. Here the patient could be audibly using the sigh to convey some level of distress. In both these extracts, the patients use of non-lexical tokens the sniff (Extract 3) and the sigh (Extract 4) to (possibly) display some problem with the proposed course of action.

Again, the clinician orients to the patient’s negative stance by displaying his awareness of the disagreeable nature of the memory test – ‘I know these aren’t particularly nice’ (lines 5 and 6) – before reassuring the patient that ‘they’re ok’ (line 9). Here the extra interactional work appears to arise more out of the contingencies of the patient’s response, rather than that clinician’s perception of their ability. However, it is worth noting the contrast between the ways the test is introduced for patients with likely FCD (Extracts 1 and 2) and those with likely dementia (Extracts 3 and 4). Furthermore, when the clinician has a working diagnosis of dementia (Extracts 3 and 4), he gazes up at the patient when seeking consent to complete the memory test. He does not give this level of attention to the patients with FCD.

In summary, we have explored variation in the way that clinicians introduce the formal cognitive examination. We have further shown through analysing cases from a single clinician (Neu1) with different patients how the introduction to the test is often recipient-designed, with the clinician adapting interactional practices when patients have more severe cognitive difficulties. This alteration may either be contingent on the patient’s negative response to the course of action proposed by the clinician or orients to the clinician’s perception of the patient’s ability. In the consultations where the clinician has formed a working diagnosis of dementia during the history-taking phase, they often amended their introduction to include more attention (in their gaze patterns) and convey some of the difficulty the test may pose for the patient. These alterations suggest that there may be good reasons for variation in the introduction to the test.

2.3.2 Interactional Non-standardization and Recipient Question Design

There are other elements of the test that show evidence of recipient design. Different clinicians deviate from the parameters of the test to design questions in ways that function to ‘help’ the patient establish the correct answer. This is evident for both questions that have verbatim instruction, as well as for quasi-scripted questions. Jones et al. (Reference Jones, Wilkinson, Jackson and Drew2020: 465–466) showed how clinicians help the patient with the more basic questions, like identifying the season of the year. Extract 5 demonstrates how the question ordinarily runs off.

This extract shows an unproblematic question–answer sequence which follows the standard administrative procedure for the test, continuing with the next question (line 4). This extract is taken from someone who reported subjective cognitive complaints. In contrast, in Extract 6 the clinician engages in extra interactional work, orienting to the patient’s difficulty in responding to the question and helping the patient to establish the correct answer.

The patient in Extract 6 displayed extreme levels of cognitive decline during history-taking (he did not know his age or why he was at the clinic). He subsequently scored only 31 out of 100 on the ACE‑III, which is highly indicative of dementia. In the test immediately before this question about the season the patient said the wrong day, did not know the date, replied with ‘Monday’ when asked the month, and replied with the name of the country when asked what year it was. Here the clinician is again asking, ‘what season of the year are we in’ (lines 1 and 2), and then proceeds to produce candidate options for the patient (line 2). Given the patient had just responded with the name of the country when asked the year, the clinician is here mobilizing recipient design by restricting the category of responses the patient can produce. This kind of anticipatory work – anticipating trouble and explicating possible answers for the patient deviates from the guidance. Trouble in responding is confirmed when the patient utters a marker of hesitation and pauses (line 4). Instead of moving on or reasserting the same question (which is a more typical course of action when a patient’s response is delayed), the clinician states, ‘I know it’s hard to tell at the moment’ (lines 5–6). This implies that the current weather condition, which is visible from the window, is atypical for the season they are in, thus assisting the patient to determine the correct answer (e.g., if it was snowing, and the weather was atypical, one might deduce that it was perhaps spring or summer). It could also work to excuse the patient for his displayed lack of knowledge – helping to save face – placing the blame for his inability to respond on the atypical weather rather than his failing cognition (Goffman, Reference Goffman1955). The clinician further prompts for a reply with ‘what would you say’ (line 8) – this implies that a guess based on this information may be acceptable. Despite this extra interactional work, the patient incorrectly responds with ‘autumn’ (line 9) (when in fact it is spring). The clinician offers a receipt for the incorrect response with ‘Oh okay’ (line 10), a turn in which the oh‑prefacing is the only indicator that the response was in some way unexpected (Heritage, Reference Heritage1998).

Jones et al. (Reference Jones, Wilkinson, Jackson and Drew2020) also showed how clinicians deviate from the administrative guide during the ‘attention–subtraction’ task, which states, ‘Ask the participant to subtract 7 from 100, record the answer, and then ask the participant to keep subtracting 7 from each new number until you ask them to stop’ (Right Decisions Service, 2017). This ordinarily runs off as shown in Extract 7:

In Extract 7 the clinician does not offer any further guidance to the patient and does not attempt to support the patient’s handling of the calculations. However, Jones et al. (Reference Jones, Wilkinson, Jackson and Drew2020) showed that some clinicians alter this sequence by repeating the patient’s response after each subtraction within the design of the next sum, meaning the number of origin is repeated back to the patient: for example, ‘and seven away from ninety-three’ (Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020: 465). We suggested that this showed evidence of co-construction, where the clinician appears to be helping the patients by adding information into the question and thus placing less of a burden on the patient’s attention skills to independently remember the numbers. These different designs demonstrate a divergence from the standardized test requirements given in the guidance and also places differential ‘cognitive load’ (Chandler & Sweller, Reference Chandler and Sweller1991; Majlesi & Plejert, Reference Majlesi and Plejert2018) on patients.

We have also found occasions when the same question is amended, or as in the next case (Extract 8) abandoned by the clinician when it is clear that the patient is unable to complete it accurately. The patient in Extract 8 scored 65 out of 100 on the ACE‑III and was diagnosed with dementia. Again, the patient had displayed interactional signs of dementia during history-taking (Jones et al., Reference Jones, Drew, Elsey, Blackburn, Wakefield, Harkness and Reuber2016; Reuber et al., Reference Reuber, Blackburn, Elsey, Wakefield, Ardern, Harkness, Venneri, Jones, Shaw and Drew2018) and was given a working diagnosis of dementia by the clinician after the initial consultation. The scoring guide states that clinicians should ‘not stop the participant if they make a mistake. Allow them to carry on and check subsequent answers for scoring’ (Right Decisions Service, 2017), as the patients can score up to five points on this question, one for each correct subtraction. For example, in Extract 7, the patient got the first subtraction correct (line 2) but all the subsequent subtractions wrong (lines 6–7), so would have scored 1 for this question (although notably the error arises from the first incorrect answer because the remaining answers are correct in terms of subtracting seven each time). Even though the patient got the second subtraction wrong, the clinician allowed them to continue with the course of the question. In Extract 8 the clinician administers this question quite differently:

The patient projects trouble in answering, with a long gap (line 2) and the turn initial delay, ‘er’ (line 3) before producing the wrong answer ‘Ninety-two’ (line 3) (albeit only incorrect by one number). The clinician then starts to narrate the next course of action he is going to take, which departs from pursuing further answers from the patient (‘Okay what we’ll do is’, line 5). Instead of completing this phrase with something like ‘move to the next question’, the clinician proceeds to actually produce the next question: ‘can you spell me the word “world”’ (lines 5–6). The clinician is here orienting to the patient’s troubles, and perhaps his own working diagnosis of the patient’s abilities, and is choosing not to follow the administrative and scoring guidance by abandoning the question. This is another way in which the clinicians design and administer the assessment for particular patients.

We have shown how clinicians vary the administration of the ACE‑III, both in the manner in which it is introduced within the initial consultation and regarding the design of certain questions by deviating from the scripted or quasi-scripted guidance that is meant to ensure that standard administrative procedures are followed. There is evidence of recipient design, which is often locally occasioned in the interaction itself, for example when patients display some trouble with the course of action or in responding to a particular question. These troubles may also be predicted by clinicians based on their perception of ‘how the patient is doing’ in the consultation more generally and the working diagnosis they have formed. Clinicians do extra interactional work to orient to these troubles and work to help the patients with the questions. We now move to show how this ‘special attention’ appears in different sequential locations within the administration of the ACE‑III, notably in the clinician’s third-turn responses (Sacks, Reference Sacks and Jefferson1992; Schegloff, Reference Schegloff1992; Schegloff, Reference Schegloff2007).

2.3.3 Recipient Design in the Third Turn: FCD

Another place where clinicians demonstrate some additional interactional attention to the patient is in the third turn. Elaborations after a potentially completed sequence of talk has been widely discussed in CA literature (Schegloff, Reference Schegloff2007), identifying different purposes for third-turn utterances in different interactional environments, such as in everyday interaction (Beach, Reference Beach1993), courtrooms (Atkinson and Drew, Reference Atkinson and Drew1979), job interviews (Button, Reference Button1987), news interviews (Clayman, Reference Clayman1988; Clayman & Heritage, Reference Clayman and Heritage2002; Clayman et al., Reference Clayman, Heritage and Hill2020), classrooms (McHoul, Reference McHoul1978; Koole, Reference Koole2010) and survey interviews (Houtkoop-Steenstra, Reference Houtkoop-Steenstra2000; Maynard & Schaeffer, Reference Maynard, Schaeffer, Drew, Raymond and Weinberg2006). Expansion in the third turn might be designed to be either minimal or non-minimal (Schegloff, Reference Schegloff2007). Non-minimal post expansions project further talk, such as repair. Minimal expansions represent one further turn at talk following the second pair part and do not project further talk; hence they are also known as ‘sequence-closing thirds’ (Schegloff, Reference Schegloff2007: 118). Minimal post-expansions accomplish a range of actions, including receipts of information using tokens such as ‘okay’ (Beach, Reference Beach1993), minimal confirmatory/assessing responses such as ‘good’ (Maynard & Marlaire, Reference Maynard and Marlaire1992) and what survey methodologists characterise as feedback (Houtkoop-Steenstra, Reference Houtkoop-Steenstra1997; Maynard & Schaeffer, Reference Maynard, Schaeffer, Drew, Raymond and Weinberg2006). Interviewers can use this third-turn position not only to acknowledge a response, but also to convey evaluations of it for the purpose of providing reassurance and motivation (Swallow and Hillman, Reference Swallow and Hillman2018). However, interviewing protocols often advise administrators to produce ‘neutral’ acknowledgements only, thus refraining from indicating whether a response is correct or incorrect (Mehan et al., Reference Mehan, Hertweck and Meihls1986). Maynard and Marlaire (Reference Maynard and Marlaire1992) demonstrated how acknowledgments (such as smiling and nodding following a correct response) do not necessarily affect an individual answer but may have a cumulative influence on performance.

Importantly, the ACE‑III administrative guide has no advice on how clinicians should respond to answers. In our dataset, there appear to be different approaches adopted by different clinicians and for different patients and their abilities. That is, we see a distinction in the interactional uses of third turns depending on whether the patient has FCD or dementia. Extracts 9–11 exemplify routine practice when a patient has a working diagnosis of FCD and Extracts 12–15 (Section 2.3.4) feature patients with dementia. We found that when there is a working diagnosis of FCD, clinicians generally do not produce third-turn responses or, if they are produced, they are typically minimal – either ‘okay’ or ‘good’ – and appear at the end of a task sequence rather than between individual question–answer sequences. The focal sequence within Extracts 9–14 is taken from an ‘attention–orientation’ task in the ACEIII, which is scored out of 5 (Extract 15 deals with a different question). On the test paper clinicians are told: ‘Ask: which – No./Floor, Street/Hospital, Town, County, Country’. The test then moves to an ‘attention–registration’ task where the participants are asked to repeat and remember three words.

In our first example (Extract 9), the clinician uses no post-expansion in the third turn.

After each response the clinician moves to the next question in the sequence, even when the task is changing (line 12). Looking across the dataset, this lack of minimal post-expansion is typical when the patient has FCD. However, there is often some ambiguity in the initial question in this section: ‘what’s the name of this place?’ (line 1). As Schegloff (Reference Schegloff and Sudnow1972) shows, a range of possibly correct answers are relevant in response to such a question (e.g., the specific room, the building, city and so on). In this case the patient responds with the name of the city (line 2) (other patients provide alternative place terms including more generic responses such as ‘a hospital’). The test is here seeking the name of the specific hospital as the ‘correct’ response and therefore, in the third turn (line 4), the clinician repairs the question to be more specific: ‘the name of the specific building’. Although this is not prefaced by an explicit marker of repair, for example ‘I mean’ (see Schegloff, Reference Schegloff1992), following a gap (line 3) and turn-initial delay the clinician partially repeats her prior turn (‘the name of the’) before replacing the ambiguous term ‘place’ with the more particular locational formulation ‘this building’. This attends to the lack of specificity inherent in the initial question as set out in the test papers. That is, in order to gain a point on the test, the patient needs to produce the name of the building in response to this question. Following the reformulated question, the patient does produce the required response (line 5) and the clinician moves on with no further elaboration or acknowledgment.

Extract 10 is very similar, in that the initial question requires some revision. However, here the clinician acknowledges the end of the task sequence before moving to a new activity (‘Good’ – line 13).

The clinician’s initial question here is a little more constrained than in Extract 9, asking ‘what building are we in?’ (Extract 10, line 1). Despite needing to identify a ‘building’ (Extract 10) instead of a ‘place’ (Extract 9), the patient does not respond with the name of the hospital, but labels the type of building, that is a ‘hospital’ (line 2). Again, the clinician attends to the test requirements for a more specific locational formulation by asking the patient ‘What’s the name of the hospital’ (line 3). Although this may not be interactionally salient to the patient, analysing the data in relation to the external requirements of a standardized test reveals the clinician’s orientation to be more specific. Reformulating the question enables the patient to produce the ‘correct’ locational formulation to gain a point on the test. As noted, this question appears to be a source of trouble in many of the interactions we have examined, suggesting a need for further advice to clinicians regarding the specificities of this question. In this case, the patient goes on to produce the required response (line 4) and the clinician moves forward, notably using and-prefaced questions (Heritage & Sorjonen, Reference Heritage and Sorjonen1994), which was not a feature of the revised question at line 3. This further suggests that the question at line 3 was not the next in a sequence of questions, but rather a clarification of the earlier question. There is no overt acknowledgment of the patient’s responses until the end of the task sequence – ‘Good’ (line 13). This minimal post-expansion is designed to close the prior sequence before initiating the next (Beach, Reference Beach1993; Schegloff, Reference Schegloff2007). It could also be audibly assessing the prior responses as being accurate, hence providing the patient with reassurance (Swallow & Hillman, Reference Swallow and Hillman2018).

We see something similar in Extract 11, where ‘Okay’ (line 12) is used to receipt the answer and to close the sequence before moving to the next task.

Merritt (Reference Merritt, Shuy and Shnukal1980: 144) suggests that ‘Okay’ acts as a ‘bridge, a linking device between two stages or phases of the [service] encounter’. Furthermore, as Maynard and Schaeffer (Reference Maynard, Schaeffer, Drew, Raymond and Weinberg2006) suggest, ‘okay’ signifies the boundaries of related questions in that interviewers withhold third-turn acknowledgments when a subsequent question links to the topic of its predecessor, and produce acknowledgments when the next question shifts topic. In Extract 11, ‘Okay’ (line 12) bridges topics and shifts between two discrete tasks on the ACE‑III. Clinicians also do not appear to confirm the responses of patients with FCD, even if patients seek confirmation. For example, the patient in Extract 11 is uncertain of the county but nevertheless gives the correct response, followed by a confirmation check ‘isn’t it’ (line 9). The clinician does not confirm that the patient has given the correct response and instead continues with the next question.

In sum, when there is a working diagnosis of FCD, clinicians generally do not produce third-turn responses, or if they are produced, they are typically minimal, either ‘okay’ or ‘good’, and appear at the end of a task sequence rather than between individual question–answer sequences. The third turn can also be used to address a lack of specificity within the design of the initial question as required by the test. Clinicians use this turn to restrict the range of place terms to a specific locational formulation to enable a patient to score a point on the test. Third-turn utterances often look different when clinicians interact with people with a working diagnosis of dementia, as discussed in Section 2.3.4.

2.3.4 Recipient Design in the Third Turn: Dementia

When clinicians have established a working diagnosis of dementia based on a patient’s ‘performance’ during history-taking, they often conduct the ACE‑III differently. Some clinicians do not require these patients to be as specific in their responses (as compared with Extracts 9 and 10). Also, they do not use the third turn to seek a more specific, and thus correct response. For example, in Extract 12 the patient responds to the question about the ‘name of the building you’re in’ (lines 1 and 2) with ‘Hospital’ (line 4). As we have seen above (Extract 10), this locational formulation is not accurate and requires further revision to receive a point on the test. Here, the patient has a working diagnosis of dementia and the clinician does not alter the question in the third turn; instead, they appear to accept this more general locational formulation and move to the next question (although it is not clear if the patient received a point for this question on the test paper).

Extract 13 also demonstrates the clinician’s lack of orientation to eliciting a full and correct response to this question. For the purpose of the transcript in Extract 13 we have used the pseudonym ‘Sandington’ for the hospital name.

Despite the patient only partially and somewhat incorrectly establishing the first sounds of the hospital name, ‘Sandi‑’ and ‘Sadding’ (lines 6–7), the clinician confirms this as being correct – ‘that’s right. Yeah’ (line 8) – and provides the correct name for the patient – ‘The Sandington’ (line 8). Clinicians thus seem to offer more leeway in what counts as accurate response when a person has dementia; in the recording of this session the clinician can be seen to be ticking the box on the sheet following this response.

Clinicians also appear to ‘help’ and reassure patients by using the third turn to explicitly confirm responses as being correct, for example ‘that’s right’ (Extract 13, line 8), or to offer a reassuring assessment of their performance when they get a question right, for example ‘very good’ (Extract 13, line 15). The extract ends with the clinician closing the task sequence with ‘okay’ (Extract 13, line 27). Maynard and Marlaire (Reference Maynard and Marlaire1992) demonstrate that administrators typically use ‘good’ when an answer is correct and ‘okay’ when it is incorrect. In Extract 13, ‘good’ (line 10) and ‘very good’ (line 15) are used when the individual question is (accepted as) correct, and ‘okay’ (line 27) is used following an incorrect response. Although further systematic analysis is required, it appears to be the case, certainly with people with dementia, that ‘good’ is often used when a response is correct and ‘okay’ when it is incorrect (also see Extract 14).

Extract 14 demonstrates this pattern of clinicians using ‘okay’ more frequently throughout each task sequence following incorrect answers. On these occasions clinicians are more likely to excuse the patient’s inability and reassure them, for example ‘Okay. Not to worry’ (Extract 14, line 11).

Interestingly in this extract the clinician does reformulate the first question about the name of the building. The patient’s response ‘your job’ (line 5) is not ‘close enough’ to being correct to enable the clinician to accept it as an adequate response and therefore he reissues it – ‘what’s this building’ (lines 6–7) – then reformulates it – ‘do you know what it’s called this building’ (lines 7–8). The patient then answers ‘no’ (line 10) and continues to suggest that not knowing is temporary – ‘not at the moment’ (line 10). The clinician expands this sequence with ‘Okay’ (line 11) and then offers the patient some reassurance that he should not worry about not knowing – ‘Not to worry’ (line 11). The third-turn utterances across this extract – including ‘okay’ + reassurance, for example ‘not to worry’ (line 11), ‘no worries’ (line 15), ‘don’t worry’ (line 22) – occur after all three of the incorrect responses in the task sequence. Following the patient’s two correct responses in this task sequence, when correctly identifying the county (line 28) and country (line 32), the clinician uses the third turn to confirm the response and positively and emphatically assess it: ‘Yes. ABSolutely. Good’ (line 29) and ‘Yeah’ (line 33). This illustrates how clinicians are far more likely to use the third turn in these question–answer sequences when the patient has dementia.

Furthermore, clinicians sometimes use post-expansion sequences to do extra interactional work to account for a patient’s inability to answer correctly. The patient in Extract 14 is unable to identify the town or city, and further audibly implies that he should be able to do so: ‘I should be‑’ (line 23) (although he stops before completing this phrase, which could be heading for ‘I should be able to…’). The clinician again offers a reassuring response – ‘No it’s alright’ (line 24) – and then continues, working to save face, with a reasonable account of why the patient legitimately cannot respond: ‘You’ve been moving around a bit. That’s okay’ (line 26). In Extract 15, this extra interactional work is also evident when the patient is unable to name the president of America who was assassinated in the 1960s.

In the third turn the clinician accounts for the patient’s inability to answer, mitigating the ‘fault’ or inability from the patient and instead suggesting it is legitimate to not know the answer given the time period since the indexed event’s occurrence, that is ‘it was a long time ago’ (line 9). It is interesting to note here that the clinician asks the patient if he can remember the event (line 7). The patient resists the implication that his not answering is because of his cognitive inability to remember, but instead imbues this with an epistemic dimension, asserting that he did not possess the knowledge in the first place – ‘I didn’t even know’ (line 8). Nevertheless, clinicians can work to account for an incorrect response and reassure the patient in more-than-minimal expansion of the sequence. This type of expansion and inability account is not generally seen in relation to people with FCD.

3.4.1 Meeting A

An elicitation of patient orientation to the clinic purpose has not been described in diagnostic appointments in other conditions, suggesting that this is a result of clinicians expecting an initial misalignment about the meeting purpose (Dooley et al., Reference Dooley, Bass and McCabe2018).

The doctor in this meeting directly asks Maggie what she was expecting from the clinic (line 1).

After a pause, Maggie provides an answer aligning with the memory clinic agenda, illustrating awareness of her memory problems through expressing hope that she can receive ‘help’ (lines 2–3). The doctor’s elongated and emphasized ‘oh right’ (line 4) indicates a change of state, registering Maggie’s response as new knowledge and thus unexpected (Heritage, Reference Heritage, Atkinson and Heritage1984). This may be because the doctor did not expect the patient to be aware of the clinic purpose (an expectation which is justified, as will be discussed below). This also could be the reason for their extended confirmation (‘yes well that’s what we do’ lines 4–5), to fill a gap where they expected to be informing the patient. The consultation then progresses to the next stage, which involves feeding back test results.

As in diagnostic delivery in other conditions, test feedback in memory clinics is designed to achieve intersubjective understanding of the nature and extent of Maggie’s memory difficulties. This allows the doctor to forecast the diagnostic news and thus calibrate Maggie’s expectations in relation to the forthcoming diagnosis (Maynard, Reference Maynard2003). Extract 2 illustrates how the doctor approaches the test feedback gradually, starting with positive assessments of where Maggie performed well on the tests (lines 1–9).

Throughout her positive feedback, the doctor encourages agreement from Maggie with multiple tag questions (lines 4, 5, and 9). Maggie responds to these with a single quiet acknowledgement token (line 10). This minimal response may be a result of the qualifications in this description (‘some areas’), projecting that a negative assessment will follow. The doctor continues with a general statement that there were ‘a couple of things’ Maggie ‘struggled’ with (lines 11–13). The rising intonation invites a response, but Maggie does not react either verbally or non-verbally. On the doctor’s pursuit of this (line 15), Maggie furrows her brow and looks down, and the doctor persists (line 17), this time outlining specific difficulties with the address-recall task. Maggie responds with a qualified agreement (‘quite possibly’ lines 18–19), and the doctor continues to describe Maggie’s ‘retrieval problem’ (lines 24–26). Again, Maggie makes neither a verbal nor a non-verbal response (line 27).

This withheld response is important because it demonstrates a lack of affiliation with the interactional project that the doctor is trying to achieve – that of forecasting the upcoming diagnosis. A comparison with the discussion of the brain scan results shows how differently patients can react to test feedback (lines 28–36). There is a notable change in Maggie’s involvement: she nods vigorously, sits forward in her chair and agrees in overlap (line 38). Comparing Maggie’s engagement with discussing the physical problem with her heart, that is both known to Maggie and less stigmatized as a condition, with the results of the cognitive testing shows how hard it is to face these results, even when she knows she needs help with her memory (Extract 1).

The diagnosis delivery follows, with the doctor directly telling Maggie she has Alzheimer’s disease in lines 1–2 of Extract 3.

Maggie has an immediate negative reaction to the diagnosis. Her ‘I have’ is characteristic of ‘ritualized disbelief’ in ‘surprise’ sequences (Wilkinson & Kitzinger, Reference Wilkinson and Kitzinger2006). This is accompanied by an exaggerated open-mouthed, wide-eyed expression (line 3), before Maggie tuts and breathes out loudly while closing her eyes and tilting her head back and to the side (line 5). Stoic, unmarked responses to diagnosis (e.g. ‘mm’) before quick progression to treatment discussions are the norm (Maynard, Reference Maynard2003), so this strong reaction is unusual in diagnosis settings. In the pause (line 8), Maggie is not moving and the doctor nods. Maggie disaffiliates by voicing her displeasure: ‘I’m not very pleased’ (line 9). After demonstrating understanding (line 10), the doctor challenges Maggie’s surprise, suggesting that this diagnosis may have been anticipated (lines 13–21). Maggie sighs and looks away with a downturned mouth (lines 22–23). Her positive assessment of her abilities demonstrates misalignment with the diagnosis (line 24). This misalignment requires attention from the doctor before they can progress to treatment and management, which they do by responding with emphasized agreement, realigning and affiliating with the patient’s viewpoint (lines 25–27).

It is not typical for people receiving the diagnosis to initiate treatment discussions in these meetings. In doing so, Maggie is showing alignment with the interactional project of getting treatment that was also evident in Extract 1 where she asks for ‘help with her memory’. In Extract 4, the doctor confirms that Maggie can have tablets but refrains from strongly endorsing them, only offering them to Maggie ‘if you would like’ (line 7). Maggie immediately accepts the doctor’s recommendation (line 8). However, in neglecting to go into more detail on the medication, the doctor demonstrates that their focus is on Maggie receiving help with living at home (lines 13–22). Maggie’s response to this is noticeably different: she withholds any verbal or non-verbal reply. She thus appears to align with and support further talk about treatment or medication but refrains from aligning with or supporting talk focused on other issues. The need for home support ‘to make sure you’re safe’ is problematic for Maggie’s self-assessment that she ‘does things well’ (Extract 3), and not aligning with this may be a mechanism to save face.

In sum, Maggie comes to her diagnostic feedback wanting help with her memory. She does not engage with her negative results in the cognitive testing, which is common throughout the dataset. Her explicit demonstration of surprise when receiving a diagnosis is less common, with many people receiving dementia diagnoses adhering to the norm of reacting stoically to bad diagnostic news (Dooley et al., Reference Dooley, Bass and McCabe2018). While Maggie is keen to receive medication, she does not engage with the suggestion of support in the home. This demonstrates how a person with dementia’s alignment with the doctor’s interactional project of diagnosis and treatment is connected to their self-presentation: Maggie demonstrates understanding that she has memory difficulties and medication can help, but disaffiliates with the diagnosis and idea that she has support needs.

The misalignment evident in Meeting B follows a different pattern, but also reflects the role of self-identity for the person receiving the diagnosis in these meetings.

3.4.2 Meeting B

As in Extract 1, the doctor in this meeting starts with a direct question to elicit orientation (line 1).

After a delay, Judith responds to the doctor’s question by stating that the doctor is going to check her medication (lines 3–4). It is worth noting that the referral letter does ask the patient to bring her medication, so Judith is partly correct. The doctor offers qualified agreement, with ‘well’ projecting a more complicated response (line 8) (Heritage, Reference Heritage2015). In describing the meeting as a ‘follow up’ and highlighting ‘the problem that you talked with the doctor about before’, it appears that Judith has already spoken with a clinician about her potential dementia (lines 11–15). The doctor clarifies the purpose of the current meeting, and the topic of the previous meeting, as ‘memory’, which prompts only a minimal acknowledgement from Judith (lines 16–17). This meeting is from Site B, where the cognitive testing appointment would have happened weeks previously, and it may be the case that Judith does not remember this. However, whatever the reason, this interaction starts with a mismatch between what the doctor and Judith expect to happen in the consultation.

Doctor B does not discuss the test results in as much detail as Doctor A did, instead simply forecasting a diagnosis, as shown in Extract 6 (lines 1–3). The ‘that’ in line 1 refers to the cognitive testing.

While this bad news is mitigated (‘a little bit worse’), it is asserted directly with little facework. Judith’s delayed, well-prefaced response, alongside her laughter, demonstrates discomfort (Haakana, Reference Haakana2001; Heritage, Reference Heritage2015). Her statement ‘I suppose you notice’, with the ‘you’ indicating what other people as opposed to what Judith herself notices, demonstrates misalignment and a possible differing perspective.

Judith, similar to Maggie in Extract 2, is more willing to engage in the results of the brain scan.

The doctor states that the blood vessels are not ‘working perfectly’ (lines 1–3). Judith aligns with emphatic agreement in overlap (line 4). She displays existing knowledge of her vascular problems, downplaying them with a dismissive description ‘and everything else’ (lines 7–8). While aligning with the immediate action of discussing the results of the brain scan, this downplay of known symptoms indicates misalignment with the doctor’s interactional project of a forecasted ‘new diagnosis’. Her laughter is also likely to be a reflection of the delicacy of this misalignment (Haakana, Reference Haakana2001). The doctor’s response is a direct reaction to this: with the ‘well’ projecting disagreement as they recharacterize the blood pressure issues as one of the ‘causes’ of her dementia symptoms (lines 10–11).

The test results are followed by delivery of the diagnosis as a ‘mild form of a condition called dementia’, which Judith responds to with a minimal acknowledgement (not shown). The doctor then elaborates, pursuing further response from Judith by explaining the different types of dementia. He finishes this at the start of Extract 8, defining Judith’s dementia as Alzheimer’s disease.

Judith misaligns with the diagnosis delivery by topicalising her hearing impairment (line 5). This could be read as Judith accounting for her test results as a basis for the diagnosis by claiming mitigating circumstances because of her deafness. Her laughter frames this as a joke, which the doctor acknowledges with a smile and an upgrade to include her ‘visual problems’ (lines 9–11), while the companion joins in the laughter. The doctor upgrades this further with a smile and a comment about Judith’s ‘selective’ deafness (lines 17–18), appearing to make explicit Judith’s lack of acknowledgement of the diagnosis. This co-constructed, light-hearted interchange works both to mitigate Judith’s potential complaint and to align with her self-identity.

The discussion moves to treatment, where Judith challenges the doctor’s recommendation to start medication.

Judith’s reaction to the suggestion of medication and follow up demonstrates misalignment with both the interactional project of medication prescription and also the diagnosis itself in her questioning of the need for treatment (line 9). This question is directed towards her companion. The companion’s response invokes the doctor’s expertise, and she looks towards the doctor as she answers (lines 11–13). The doctor begins to explain further, but the companion interrupts with a strong, persuasive endorsement (lines 17–18). Judith’s reluctance is evident in her agreement on the basis ‘it can’t do any harm’ (line 20), indirectly displaying her opinion that medication is not necessary. Her laughter again illustrates the delicacy of disagreeing with the doctor’s recommendation. However, this time, the doctor and the companion do not join in; their agenda in encouraging Judith to take the medication is prioritized.

To summarize Meeting B, Judith demonstrates misalignment with the interactional project of diagnosis feedback throughout. She sees the meeting purpose as a discussion of her medications, does not appear to remember her assessment, suggests a different perspective of her abilities than is indicated in the testing, responds to the diagnosis with a joke, and questions the need for medication. However, this misalignment again appears to reflect a project to preserve self-identity. She highlights her deafness and vascular difficulties as sources of her problems, which, given the stigma surrounding dementia, may be more socially acceptable diagnoses. She also often laughs and jokes in serious moments, which is a common coping mechanism for Judith to maintain interactional competency in challenging situations (Saunders, Reference Saunders1998).

3.4.3 Meeting C

Meeting C differs from the others in that it does not start with the doctor eliciting Bill’s awareness of the reason for the meeting. They instead describe their agenda (lines 1–5).

The doctor does not start with an expectation question as in the other examples, instead stating the meeting purpose and allowing time for a reaction, which may be a method to elicit orientation without directly threatening patient face. The statement does however forecast a potential diagnosis in stating that something ‘might be going on’ and that the clinic can ‘help’ (line 4). Bill’s upward intoned, emphasized ‘oh’ demonstrates surprise (line 7) (Heritage, Reference Heritage, Atkinson and Heritage1984). He has attended a clinic in Site A, so has just completed the cognitive testing and had a brain scan in the few hours before this meeting. Therefore, this indicates some disorientation. The doctor responds by asking for a ‘go ahead’ from Bill to continue (line 8). Bill’s well-prefaced response in overlap again misaligns in that it does not give express permission, instead demonstrating concern regarding the need to take further medication (lines 9–10).

The doctor’s response is empathic in tone, their upward intonation aligning with Bill’s assessment that eight tablets a day is a lot (Heritage, Reference Heritage, Stivers, Mondada and Steensig2011). Their affiliation, laughter, and indication of prior knowledge of Bill (line 12) could be performing alliance work before addressing the issue of diagnosis, which Bill’s surprise suggests might be difficult. Also potentially resulting from this surprise, they give a generalized clinic description (‘we see people’ line 15) and minimize the symptom profile of clinic attendees (‘with possible mild memory problems’ line 16). They then add specificity by naming the potential diagnosis as ‘something like a dementia’ (line 20). It is unusual for doctors to explicitly mention dementia at this point in the meeting, with references to ‘memory’ being more common. This again could be in response to the fact Bill’s wife has dementia: his detailed understanding of the condition in its more severe stages adds particular delicacy to breaking the diagnostic news. Bill replies with a topic shift in beginning to talk about his wife, not acknowledging the doctor’s description of the meeting purpose (lines 22–24).

After some discussion of Bill’s experiences leading up to the memory clinic appointment, the doctor gives a brief description of the brain scan results and a lengthy discussion of Bill’s experience of the cognitive testing. Extract 11 begins with Bill accounting for his difficulties.

The doctor again shows empathy within their own epistemic status as the ‘tester’, that they ‘appreciate it’s quite stressful’ (lines 3–4). Bill is demonstrating awareness of and accounting for his forgetting the UK prime ministers (lines 5–17). The doctor begins to lay groundwork for the diagnosis, saying those answers were correct, ‘but’ Bill’s score was 58/100 (lines 19–24). Bill’s reception of this news and his description of the score as ‘average’ (line 25) indicates a lack of knowledge of what this score entails, appearing to imply that this would be in line with other people. The doctor thus follows this with an explicit diagnostic forecast in an effort to ensure understanding (‘there is a problem with your memory’ lines 27–28). Bill’s response, similar to Judith’s in Meeting B, misaligns with the forecasting of bad news by making a joke about having to open his brain up. Similar to Judith, his joking could be a method of controlling an emotional response and saving face, which is common in dementia (Hedman et al., Reference Hedman, Hellström, Ternestedt, Hansebo and Norberg2014; Saunders, Reference Saunders1998). Similar to Doctor A, this doctor affiliates with and upgrades Bill’s joke in saying that they ‘clearly’ won’t be doing a brain biopsy and there is joint laughter.

The doctor continues by explaining how all the testing is considered to make a diagnostic decision, starting the diagnosis delivery in Extract 12.

Bill has a strong negative reaction to the diagnosis, produced in overlap. Similar to Maggie, he demonstrates surprise and disappointment with a rhetorical interrogative, latched expletive, and an explicit expression of understanding in stating that he does not want to be like his wife (lines 1–3). His laughter (line 6) again could be a method of holding back further emotional reaction, particularly given that typical reactions to diagnoses tend to be stoic and restrained (Dooley et al., Reference Dooley, Bass and McCabe2018; Maynard, Reference Maynard2003). He is smiling, and as the doctor reformulates the diagnosis as a ‘dementia’ (line 7), his smile drops. The doctor then downplays the certainty of the diagnosis, potentially in response to the strong negative reaction, saying ‘there’s no hard and fast test’, but that it ‘looks like’ a dementia (lines 10–15). Bill aligns with this explanation in overlap, demonstrating knowledge (‘no right one, yeah’ line 12). Bill then leans forward to look at the brain scan on the doctor’s computer screen (line 16).

The doctor explicitly differentiates Bill’s dementia to his wife’s, likely as a resource to soften the bad news in reaction to Bill’s negative response (Stivers & Timmermans, Reference Stivers and Timmermans2017). Bill again aligns with this, displaying in-depth understanding of the diagnosis in reply – ‘well it’s arterial isn’t it’ (line 22) – before appearing to search for the correct term and showing recognition of the doctor’s supply of ‘vascular’ (lines 27–30).

Bill therefore aligns with the doctor’s interactional project of diagnosis, showing understanding and an emotional reaction. The doctor reassures Bill and then moves on to talk about medication.

The doctor’s treatment recommendation is in the form of a suggestion (Stivers et al., Reference Stivers and Barnes2018), a recommendation format that gives Bill agency over his decision. This format is likely to be a direct response to Bill’s previous resistance to taking more medication in Extract C1. There is clear misalignment here: Bill does not respond directly, instead laughing and joking that the medication should be for his wife (lines 4–7). The doctor continues with their explanation (lines 9–16), but Bill explicitly questions his need for memory tablets, entering in overlap. He formulates his diagnosis (‘thing for me’) as ‘to do with the blood pressure’, and expresses concern that the recommended medication will make this worse (lines 17–19). The doctor answers indirectly, stating that Bill is on blood pressure medication, insinuating that he need not have any concerns in that regard. They add further explanation that the tablet increases ‘one of the chemicals; in the brain (lines 26–27). Similar to Maggie and Judith, Bill is demonstrating more active engagement with his physical health issues than his memory problems, and thus it may be that this is detailed, physical explanation of ‘choline’ is an attempt to engage Bill in discussion of the memory tablets. However, Bill’s response cry (‘oh god’ line 28) demonstrates displeasure at this description. The doctor, likely in reaction to this strong response and continued misalignment, offers Bill the choice not to take the medication, citing Bill’s previous concerns about taking too many tablets (lines 33–35). Bill aligns with this and the doctor’s closing remark (‘then that’s fine’ line 40) suggests the matter is closed.

While often providing strong and occasionally contrary reactions to what the doctor is saying, Bill does appear to align with the doctor’s agenda throughout. His test results and disorientation to the meeting purpose indicate a moderate dementia, but he is engaged in and displays understanding of the diagnostic descriptions. His strong reaction to his diagnosis and his resistance to medication is embedded in a comparison between his own circumstances and those of his wife, who is in a care home with dementia. Hence, his preference to engage with the vascular description of his condition may serve to maintain his identity as the ‘healthy’ member of his marriage, differentiating himself from his wife’s experience. Having seen first-hand what dementia is like and having been a carer of a person with dementia, receiving a dementia diagnosis is likely to be particularly emotionally challenging for Bill.

4.1.1 Overview of the Chapter

First, some central concepts of applied CA are presented (Sections 4.2.1 and 4.2.2). This is followed by some major challenges related to minority ethnic persons with dementia (Section 4.2.3). Analysis of an episode from an interpreter-mediated dementia assessment demonstrates micro-level circumstances that lead to macro-level problems in equity in care for the group in question (Section 4.4.2). Ethnographic accounts of clinicians and interpreters are also provided (Section 4.4.1). The chapter ends with an overview of some interventions that are connected to and were developed as a result of challenges such as those revealed in the analytical Section (Sections 4.5 and 4.6). The aim of this chapter is thus not to be only an original, empirical study. Rather, the text is organized so that the analytical section is used as a basis for discussing problems that occur every day in memory clinics for minority ethnic people, interpreters, and the medical professionals responsible for diagnosis, and how these problems can be resolved.

4.2.1 Prerequisites for Applied CA

Application and implementation of CA-derived findings in relation to the field of ethnicity and dementia will be described as consisting of several factors/steps,Footnote ² in particular:

1. (1) Identification of a problem area, and presentation and discussion of observations and preliminary results together with relevant stakeholders;

2. (2) Multidisciplinary research collaborations and networks, in which results from CA are discussed alongside findings on the same topic from studies utilizing different methods and theoretical frameworks, qualitative as well as quantitative;

3. (3) Collaboration and dialogue with national and international interest organizations;

4. (4) Raising funds for research that is devoted to an overall problem area in need of development: for example, inequities in dementia care in relation to ethnicity, language background, and education;

5. (5) Presentation of results to politicians and policymakers, nationally and internationally, preferably in the shape of reports and information texts on different levels of complexity that are empirically sound, multidisciplinary, and ecologically valid, but not too long. This, however, is only possible if steps 1–4 have been carried out successfully.

4.2.2 Social Problem-Oriented and Interventionist CA

Antaki (Reference Antaki2011) provides an overview of different forms of applied CA: foundational, communicational, diagnostic, institutional, social problem-oriented, and interventionist. A brief description will be given of the social problem-oriented and interventionist approaches, since they are key for the topic of this chapter. When CA is social problem-oriented, it is used to address an issue that is perceived to be challenging for a specific group or groups of people, and tied to some kind of societal context. Interventionist CA is:

Needless to say, an intervention cannot be meaningfully developed unless a problem has been identified. The first step is thus to explore what it is that needs intervention.

4.2.3 What Social Problem?

As part of a large program devoted to Living with Dementia, conducted at the Centre for Dementia Research (CEDER) at Linköping University, a subgroup of scholars formed in 2011 to investigate living with dementia in relation to ethnic and cultural diversity and multilingualism. The overall program was, by and large, concerned with the fairly culturally and linguistically homogenous setting of Sweden. As ethnicity and dementia were explored, it rapidly became clear that not only was this area surprisingly under-researched, but that the societal need (nationally as well as internationally) for increased knowledge about the impact of dementia on minority ethnic people was huge. Reviewing existing studies, several things stood out as potentially unjust in comparison to the provision of dementia care and support for majority ethnic populations in Western societies. Some main observations could be made based on existing international research:

1. (1) Persons from minority ethnic groups sought help for their symptoms later than persons with a majority ethnic background, and sometimes did not seek help at all (Alzheimer Europe, 2018; Gove et al., Reference Gove, Nielsen, Smits, Plejert, Rauf, Parveen, Jaakson, Golan Shemesh, Lahav, Kaur, Herz, Monses, Thyrian and Georges2021). This could be partly explained by differences in conceptions of dementia, and also cultural and religious habits, which may differ significantly from (biomedical) perspectives applied in health care in Europe and the WestFootnote ³ (Alzheimer Europe, 2018).

2. (2) Persons from minority ethnic groups received an unspecified dementia diagnosis significantly more often in comparison to the majority ethnic population, were younger when receiving their diagnosis, had lower scores on cognitive screening tests, and were prescribed less dementia-specific medication but more anti-psychotic drugs in comparison to the majority ethnic population (Nielsen et al., Reference Nielsen, Vogel, Phung, Gade and Waldemar2011a, Reference Nielsen, Vogel, Riepe, de Mendonca, Rodriguez, Nobili, Fade and Waldemar2011b, Reference Nielsen, Andersen, Kastrup, Phung and Waldemar2011c).

3. (3) Tests of cognitive functioning used in memory clinics were rarely adapted to minority ethnic persons, and not available in all languages needed, neither were they adapted to people with limited education and literacy. A lack of adequately trained interpreters also made cognitive screening a particularly demanding endeavor in the overall diagnostic process (Torkpoor et al., Reference Torkpoor, Fioretos, Essén and Londos2022).

4. (4) When living in residential care and perhaps not being proficient in the language spoken by other residents and members of staff, communication was often limited, and opportunities to interact in the person’s best (or only available) language were not always available (e.g., Jansson, Reference Jansson2014; Small et al., Reference Small, Chan, Drance, Globerman, Hulko, O’Connor, Perry, Stern and Ho2015; Söderman & Rosendahl, Reference Söderman and Rosendahl2016; Alzheimer Europe, 2018, inter alia).

It should be stressed that the social problem was not the minority ethnic persons themselves, but the unpreparedness of the majority ethnic society to be able to offer the same quality in dementia care as that provided to the majority ethnic population.

Based on the above findings from existing research, memory clinics and residential care were identified as key environments for further scrutiny, and several studies were designed in CEDER to investigate these settings, primarily (but not only) using anthropology, ethnography, and CA (Antelius & Kiwi, Reference Antelius and Kiwi2015; Majlesi & Plejert, Reference Majlesi and Plejert2018; Jansson et al., Reference Jansson, Plejert and Lindholm2019; Plejert, Reference Plejert, Gavioli and Wadensjö2022). This chapter, however, focuses on interpreter-mediated dementia assessments in memory clinics.

4.2.4 Prior Work on Interaction in Dementia Assessments

Analyses of video-recorded interactions during interpreter-mediated history-taking and formal clinical dementia assessments (Van De Mieroop et al., Reference Van De Mieroop, Bevilacqua and Van Hove2012; Plejert et al., Reference Plejert, Antelius, Yazdanpanah and Nielsen2015; Majlesi & Plejert, Reference Majlesi and Plejert2018; Plejert, Reference Plejert, Gavioli and Wadensjö2022) have shed further light on already documented obstacles identified by studies within related fields such as neuropsychiatry (see Nielsen et al., Reference Nielsen, Vogel, Phung, Gade and Waldemar2011a, Reference Nielsen, Vogel, Riepe, de Mendonca, Rodriguez, Nobili, Fade and Waldemar2011b, Reference Nielsen, Andersen, Kastrup, Phung and Waldemar2011c), clinical memory research (Torkpoor et al., Reference Torkpoor, Fioretos, Essén and Londos2022), and CA work on monolingual dementia assessments (e.g. Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020), as well as interpretation in other health care settings (e.g. Bolden, Reference Bolden2000; Raymond, Reference Raymond2014). Of particular interest for the present chapter is work that has investigated the lack of adequately adapted tests – linguistically, culturally, and educationally – to minority ethnic patients with limited literacy (Nielsen & Jørgensen, Reference Nielsen and Jørgensen2013; Gove et al., Reference Gove, Nielsen, Smits, Plejert, Rauf, Parveen, Jaakson, Golan Shemesh, Lahav, Kaur, Herz, Monses, Thyrian and Georges2021; Torkpoor et al., Reference Torkpoor, Fioretos, Essén and Londos2022), and investigations of interaction during dementia assessment administration (Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020; Jones et al., Chapter 2 this volume). Prior work on interpreter-mediated dementia assessments has stressed how the interaction between participants may be characterized by lengthy sequences of repair, due to issues such as the interpreter’s unfamiliarity with the test, or the patient’s difficulty in understanding a task. In monolingual settings, it has also been highlighted that tests of cognitive functioning are often not carried out in standardized ways and according to formal instructions (Jones et al., Reference Jones, Wilkinson, Jackson and Drew2020; Jones et al., this volume). Exactly what that means for diagnosis remains to be investigated further. However, the training of the professionals involved in the cognitive assessment appears to be a key issue, irrespective of setting or occupation (clinician or interpreter). A common denominator of monolingual and multilingual assessments is that the way a test is administered apparently has consequences for how it is understood by the patient (see Jones et al., Chapter 2 this volume for a wider discussion). This issue will be returned to in Sections 4.6 and 4.7 in this chapter.

4.3.1 Ethical Approval

The CEDER program, including the subproject on ethnicity and dementia, was ethically approved by a Swedish board for ethical vetting. The specific ethical challenges for research involving people living with dementia are elaborated on further in the introductory chapter of this volume.

4.3.2 Setting, Participants, and Data

When the project started, contact was established with a selection of memory clinics. Often, before being able to conduct any work in the field, several informal visits were made, and meetings and discussions were held with members of staff and managers, to ensure that everyone was aware of what the study was about, was able to have their say, and was willing to participate. Eventually, collaboration was established with two Swedish memory clinics, in which video recordings were made of activities such as history-taking, advice concerning daily activities, and tests of cognitive functioning, involving four patients, four interpreters, and five clinicians. Fieldwork also comprised ethnographic observations and recorded interviews with clinicians, as well as informal conversations with interpreters and clinicians. The resulting video recordings were approximately 15 hours in length, and recordings of interviews and fieldwork were approximately 20 hours in length.

The data in Section 4.4 comes from a video recording of a Turkish-speaking 80-year-old woman (PA) during the performance of a test of cognitive functioning. The woman had lived in Sweden for almost 20 years and had been able to speak some Swedish, but had successively lost parts of this ability. At the time of the assessment she could still understand some Swedish, and managed to produce Swedish words and phrases as responses to questions. The assessment took place in her home; the visit was recorded from beginning to end and consisted of three parts: history-taking, a test of cognitive functioning, and an assessment of the woman’s ability to perform everyday tasks, such as making coffee. An occupational therapist (OT) responsible for the assessment had chosen to conduct the entire event in the patient’s home, since OT wanted to visit PA in her everyday environment. Also present during the event was the OT, a professional interpreter (IN) speaking Turkish as her first language and Swedish as a second language, the woman’s granddaughter (GD), and the granddaughter’s boyfriend.

The recording lasts for an hour and a half and has been divided into eight parts: Extracts 1–8. In Section 4.4 they are transcribed verbatim in Swedish and Turkish, and an idiomatic translation into English is also provided. Transcriptions were provided by native speakers of Swedish and Turkish. Translation into idiomatic English was done by the author and is provided below each line of original language, in bold face.

The video recordings were scrutinized using CA, with multimodal notations where deemed relevant. All recordings were watched several times, and it should be noted that at the first viewing, a lot of the interaction in Turkish was not comprehensible to the researchers. This issue, was resolved once the recordings were translated and transcribed by CA-informed research assistants proficient in Turkish and Swedish, and any uncertainties were subsequently double-checked by a second native speaker of Turkish, speaking Swedish as a second language.

In contrast to Swedish and English, which are both Germanic languages with a subject–verb–object word order, Turkish ordinarily has a subject–object–verb word order, and there are also other linguistic and semantic features that may make the structure of utterances different to that found in Germanic languages. Therefore, in the extracts cited in Section 4.4, for turns that run over more than one line, the English translation is not placed below each line of the original language, but rather displayed as chunks, so that the content of what is being said is not misinterpreted as word-for-word translation. For consistency, this also applies to longer turns in Swedish as well as Turkish.

4.4.1 Ethnographic Accounts of Challenges in Interpreter-Mediated Dementia Assessments

In systematic ethnographic interviews, as well as in informal conversations, clinicians attested that interpreter-mediated assessments, and lack of adequately adapted screening materials, were great concerns. Clinicians often felt anxiety concerning assessment accuracy in relation to tests. Even if they stressed that cognitive screening was just one part of a larger assessment process, it was their viewpoint that the pathway towards certainty was more challenging and took a lot longer for minority ethnic patients in comparison to native Swedish ones. The quality of interpreting and access to proficient interpreters were worrisome: “Nuances are missing,” “I sometimes have no idea what the interpreter and patient are talking about,” and “I doubt everything they say is rendered to me,” were frequent statements by clinicians in interviews. Some clinicians also stated that they had been taught “never to interfere or interrupt the interaction between patient and interpreter,” even if they were uncertain about what was being discussed. This account also conformed to some interpreters’ self-reports that they kept strictly to their professional oath of conduct (Kammarkollegiet, 2020), for example that they only rendered as exactly as possible what was said by the patient or clinician, with no omissions or additions. Readers familiar with interpreting research know that such beliefs are rarely accurate (Wadensjö, Reference Wadensjö1998; Karliner et al., Reference Karliner, Jacobs, Chen and Mutha2007; Hsieh & Kramer, Reference Hsieh and Kramer2012; ).

In the informal conversations with interpreters (unfortunately not recorded) during fieldwork, several other challenges were revealed, not least in terms of the interpreters’ work conditions. Often, interpreters were not given the test materials in advance, and they were asked by clinicians to translate texts and test materials that were not available in the patient’s language – so called prima vista interpreting – on the fly. This latter practice requires specific training and should be asked for in advance before an interpreter is appointed. In fact, in Sweden an interpreter not trained in prima vista need not do it if they feel that they cannot perform at a professional level. To support this impression from our fieldwork, it should be noted that results of a survey of 209 interpreters asking about their perspectives on mediating within logopaedic activities in Sweden showed that, despite feeling uneasy about it, they often thought that refusing a prima vista task would be disloyal to the patient and clinician, and they therefore often performed the task nonetheless (Aburto Maldonado & Eklind, Reference Aburto Maldonado and Eklind2021).

4.4.2 Analysis of a Video-Recorded Episode from an Interpreter-Mediated Dementia Assessment

The test material used in the episode investigated here was the Montreal Cognitive Assessment (MoCA; Nasreddine 2003–Reference Nasreddine2017), which is a short screening instrument consisting of several subparts. In Extracts 1–8 the participants are engaged in a test of memory and recall, in which the patient is to repeat five words twice, try to keep them in memory, and then recall them when asked to at the end of the test. The OT had chosen to use the Swedish version of the MoCA, although a Turkish version did exist at the time. The words in Swedish were “stol” chair, “plånbok” wallet, “tång” pliers, “munspel” harmonica, and “sax” scissors. While the Swedish version of the test deviates in interesting ways from the English original and the Turkish version, this does not form part of the analysis considered in this chapter.Footnote ⁴

At least two explanations are plausible for the OT choosing to use the Swedish version of the test. First, at the time of the recording in 2015, the Turkish version of the MoCA was in the process of being validated for different populations and diagnoses (e.g., Yildiz et al., Reference Yildiz, Aki, Can, Derle, Kibaroglu and Barak2014). Secondly, had the Turkish version of the test been used, the OT would have had to ask the IN to carry out the test in full, which would have given the IN a medical responsibility that was not allowed by both medical praxis and the interpreters’ professional oath of conduct. However, not all clinicians take this into consideration when they naively treat an interpreter as an information channel (see Wadensjö, Reference Wadensjö1998): clinicians might think that they are in charge of the test when, in fact, the interpreter is performing actions well beyond their professional role (Bolden, Reference Bolden2000). As will also be shown, there are elements of the interpreter being assigned responsibility for the subtest, too. Administering a formal test such as the MoCA requires specific training.

Extract 1 begins at the point of transition from a picture-naming task to a memory task.

As already mentioned, there had been no pre-appointment briefing with the IN and she was not familiar with the MoCA. The procedure for carrying out the memory task was therefore unknown to her, and explanation-work was needed, resulting in some confusion about who was being addressed, when, and about what. The instruction provided by the OT (lines 01, 02) is initially understood by the IN as directed toward the PA, since the OT is not clearly looking straight at anyone, but rather slightly downwards toward the test sheet. The IN (line 04) thus directs her rendition towards the PA. Immediately, however, the OT addresses the IN, asking her to take a look at the words to be repeated and recalled in the test material in front of all of them (lines 05–07), pointing with her pen at them. Since the IN does not know that the PA is only to listen, repeat, and recall the words, she puts the test sheet further into the visual field of the PA and asks her to read and check if she understands the words (line 07, multimodal). Straight away there is thus a misunderstanding between the OT and the IN about how to conduct the test, where the IN in fact is doing exactly what she, in her professional role, is supposed to do, that is render what the OT is saying to the PA. She misunderstands that the question (lines 05, 06) was addressed to her, and not to the PA. The OT uses the second person singular pronoun “du” (you), but this is not helpful here for turn-assignment, because clinicians who are trained in how to work with an interpreter are supposed to address patients using this form, speaking to the patients directly, which is also in accordance with interpreters’ professional oath of conduct. It is therefore plausible that the IN’s first-hand understanding at this point is that the “you” is meant for the PA. All of this happens rapidly, and since the OT does not understand Turkish, it is the IN’s non-verbal conduct of handing over the material to the PA (line 10, multimodal), who reads out “stol” (chair) in Swedish (line 11), that makes the OT understand that the IN mistook the question as being directed towards the PA and as a part of the test procedure. To amend this mistake, the OT initiates repair (line 15), correcting the action by the IN, adding that the PA is not to read, but just listen (lines 15, 17). This explanation is acknowledged by the IN, also addressing the patient in Turkish (line 18) before asking the OT for further information about what to do next (line 19).

Subsequently, the OT initiates a topic that indicates that she is interested in knowing whether the Swedish words to be repeated and recalled are common in Turkish (and thereby easily recognizable for the PA: see lines 05, 06 in Extract 1). The fact that confusion arises due to the language of the test material also surfaces in Extract 2. The OT acknowledges that the PA has read aloud “stol” (chair, Extract 1, line 11) when she saw the words in the Swedish test material.

After acknowledging the PA’s reading out of “stol” (chair) in Swedish, the OT asks the IN about the next two words in the test, which are “plånbok” (wallet) and “tång” (pliers) (lines 20, 21). Wallet is mentioned in passing with falling intonation (line 21), whereas the OT asks explicitly what the situation is regarding pliers. The IN begins to say something about pliers (line 23) but is interrupted by the GD (line 24) as well as by the PA herself, who asks if “tång” means something to do with weight (line 25), which is denied by the IN (line 26). The suggestion by the PA (line 25), however, makes sense from a language-learner perspective, considering the similarity in pronunciation (as well as spelling) of the word for pliers in Swedish “tång”/tɔŋ/ and the word for heavy “tung”/tɵŋ/. It is noteworthy that the IN does not render to the OT what the PA is saying in line 25, and the episode continues as a collaborative, syntactic, co-constructed unit produced by the IN, GD, and OT (Lerner, Reference Lerner and Lerner2004; Bockgård, Reference Bockgård2004) about “tång" being a kind of tool (lines 23, 24, 27, 28). The IN, OT, and GD are at this point thus in agreement about the meaning of the Swedish word (see GD’s acknowledgment token in line 29). The IN then checks for the Turkish translation on an online resource on her phone.

The IN checks her phone at the same time as the GD suggests that “tång” may also mean “havstång,” that is seaweed (lines 31, 32). “Tång” happens to be a homonymous word in Swedish that may refer to either a tool or a plant – it is pronounced and spelled in the exact same way. The OT acknowledges the GD’s comment, but does not say anything further, and the IN overlaps (line 34), objecting to the suggestion by the GD as the non-intended meaning, supposedly since they have just reached agreement (Extract 2) about “tång” with the meaning of a tool being the target word in the test. She code-switches into Turkish, providing information about what can be done with this tool (like to pull a nail=) (lines 34, 35), which is accepted by the GD (line 36). The IN continues to scrutinize her mobile phone, stating, with a quiet and slightly frustrated tone of voice, that she knows (line 37), and she also happens to access the word seaweed on her device before, eventually, she finds the Turkish word for the tool (lines 38, 39), which is “kerpenten.” For the OT, however, it is important that the PA knows what it is, so she persists in her project of making sure that the words to be tested are common in Turkish (line 40).

Ten lines have been omitted, during which the IN checks that the PA knows what pliers are. The OT checks with the IN that harmonica and scissors are also common words in Turkish, which is confirmed by the IN. After this, an attempt to restart the memory task is made.

As can be observed in Extract 4, which language to use is once more a concern for the IN (line 54) when she is given the instruction from the OT (lines 52, 53). The language is confirmed by the OT (line 55) as well as provided at the end of her instruction (line 58). It may be the case that the latter instance is not heard, since it is produced in overlap when the IN requests further information about how to carry out the task (line 57). The lack of hearing may also explain why the IN requests a confirmation of choice of language again in line 59, confirmed by the OT in line 60.

What can be observed in Extract 4 is that not only does the IN’s uncertainty of what language to use need to be clarified, but more precision on how the task is to be carried out and what is expected from the PA is also required. An issue so far not addressed in detail in this analysis is the potential impact that such negotiations might have for the PA’s understanding and execution of the task, since she does have some knowledge of Swedish. The PA is thus exposed to a large amount of linguistic input in two languages and visual access to the test material throughout this negotiation, which may affect her ability to remember the target words.

In the test instructions, the task is supposed to be carried out in a specific way, with the words produced at a certain pace by the test leader and then repeated by the patient; the same procedure should then be conducted a second time. After this, the patient is to receive the instruction to try to store the words in memory, and is then asked to recall them at the end of the test. As has been demonstrated, rather than providing the IN with this full explanation from the start, it is given to her bit by bit, which requires repair several times, stalling the progression of the test, as already evident from Extracts 1–4. A further example of the OT’s explanation-work and the IN’s requests for clarification is demonstrated in Extract 5.

As can be observed in lines 62–63, the IN is positioned as test administrator by the OT, by means of a repair in which the pronoun “vi” (we) is changed into the second person singular “du” (you), concerning who is to read out the list of words once more. Another salient feature in Extract 5 is the query from the IN about the way the PA is to repeat the words: if the IN is to read the words one by one (lines 64, 66, and 76), supposedly meaning that the PA should repeat each word right after it has been read out aloud. The OT responds quickly in a confirmatory way (line 67) to the first part of the question before the IN has finished her turn (line 66). The OT then provides instructions concerning the steady pace at which the reading should be conducted, and why (not so fast cause she must kind of have time to put it in memory) (lines 67, 68). She then returns to describing the second round of repetition (lines 70–72), although the IN once more tries to find out whether to read each item one at a time (line 76). Again, it appears that the OT assumes that the IN understands that all words should be read out at a steady pace by the IN first (as stated in the instructions) before the PA repeats them. As will be shown in Extract 6, there is still a mismatch between the participants’ understanding of this procedure, since the PA repeats each word after the IN, instead of waiting, listening, and then repeating.

The pattern of the PA repeating right after the IN occurs between lines 89 and 106. The IN is still positioned as a test administrator who is responsible for marking the PA’s answers on the test form. Interestingly, the IN does not object to this role, despite the fact that it goes against the interpreters’ professional oath of conduct. It is hard to tell if this positioning contributes to her not rendering the PA’s verbal repetitions to the OT, or if she judges her marking correctness on the test sheet as sufficient for the OT, since this action is taking place within the visual field of all participants present, and the target words are written in Swedish on the test form.

Translating text items without preparation has already resulted in negotiations about the word for “tång,” when the IN had to look it up on her mobile device (Extract 3, and in talk about the other words being easily understood in Turkish). In Extract 6 the IN’s mind goes blank when she is to translate the Swedish word “sax” (scissors) (line 103). Her actions of putting down the pen and looking blankly towards the right with a thinking face (Goodwin & Goodwin, Reference Goodwin and Goodwin1986) are followed by a soft giggle before she turns toward the PA and produces the Turkish word for scissors (line 104). When the PA has repeated the word, the IN tries to mitigate her delay in translating by stating that she also sometimes faces memory or word-finding problems (h.hhh it gets stuck with me too sometimes ha ha) with laughter which, together with her facial expression, indicates embarrassment (Wilkinson, Reference Wilkinson2007; Lindholm, Reference Lindholm2008) (line 107). Saying that she, too, faces difficulties coming up with a certain word, signals that problems with recall could happen to anyone, even an interpreter, which in a sense normalizes the PA’s potential dementia condition currently being assessed. It might also be an excuse for the IN’s display of a sudden lack of competence in her professional role as an interpreter.

What happens after this episode is that the OT instructs the IN to ask the PA to repeat all the words once more, which she does, and again the PA repeats straight after every word spoken by the IN. This time, the OT interrupts and provides the IN with instructions that she is to read out all the words first, before the PA repeats them.

The IN is ticking the boxes in the test form for each word, and does not confirm the answers in speech to the OT this round either, and the OT follows the progress and results of the PA by watching the marks being made on the sheet by the IN. At last, they manage to conduct the task in the way described in the test manual. However, this process has required a fair amount of repair work and negotiation in two languages, both of which the PA is proficient to various degrees. All participants have had a lot of visual access to the test sheet, despite the fact that this subtask on memory should be performed without visual cues.

The task is finished when the OT asks the IN to tell the PA to try to keep the words in her memory (Extract 7, line 125, 126). The PA’s understanding of this instruction is displayed in line 130, as she asks for confirmation, and is positively responded to by the IN (lines 131, 133).

The administration of the MoCA continues with a couple of further subtasks before it is time, as the penultimate assignment, to recall sandalye, cüzdan, kerpeten, mızıka, and makas.

As can be observed, the IN gives the PA the instruction to try to remember the words that they had previously discussed (lines 504–507), and before the PA answers, she once more jokes about needing to be reminded herself (line 508), which is rendered by the IN with laughter (lines 509, 510). The PA is very much oriented towards the task and produces the first three target words at a fairly regular pace (lines 511–516). The PA then looks straight ahead, thinking for some time, before she comes up with the Turkish word for scissors, followed by a long silence and an account about being finished (line 518). In accordance with the test instructions, it is allowed to provide brief semantic cues at this point, which the OT instructs the IN to do (lines 520–523). When provided with this input (lines 524, 525), the PA quickly manages to remember “kerpenten,” which is acknowledged as correct by the IN (line 527). The task is closed by the OT assessing the task as successful (line 528), also rendered in Turkish by the IN (line 529). Largely, the PA has completed the memory and recall task successfully. To what extent she would have scored in a similar way if conducting the test in Turkish only will never be known.

4.5.1 Feedback to Memory Clinic Staff

During the project, the participating memory clinic teams were offered feedback using anonymous video clips and rough transcription extracts as a basis for discussion. Whereas clips like Extracts 1–8 would generate discussions about choice of test material, the role of prima vista interpreting, and how to better prepare interpreters for an assessment, other episodes dealt with challenges associated with patients’ level of education, and cultural bias in relation to a test (see Nielsen & Jørgensen, Reference Nielsen and Jørgensen2013; Plejert et al., Reference Plejert, Antelius, Yazdanpanah and Nielsen2015). By and large, video analyses seemed to corroborate the experiences and “suspicions” that clinicians had expressed in the ethnographic interviews and informal conversations carried out before and during video-data collection. One benefit of the feedback sessions was that they appeared to be an eye-opener for many clinicians to explicitly show how challenging the situation was for interpreters, too, not just for themselves (see also Torkpoor et al., Reference Torkpoor, Fioretos, Essén and Londos2022).

Overall, even if clinicians were often aware of many of the issues that may influence an intercultural encounter, access to translated video and/or sound recordings of interpreter-mediated patient encounters made them aware of the more precise details of linguistic, cultural, and educational matters that were causing problems. Despite the small scale of this intervention with the participating teams, the sessions made clinicians consider changing certain practices, for example concerning choice of test material, to try to more systematically brief interpreters before an appointment, and to interrupt in the interpreting process when they felt uncertainty about what was going on in talk between patient and interpreter when this was not referred back to them. Based on video episodes from other activities (not considered in this chapter), there was also discussion on how different activities might be more or less challenging from the interpreter’s point of view, with history-taking being a lot less demanding than being asked to translate and administer a formal test.

4.5.2 Presentations at Training Days

Analyses highlighting challenges like the ones presented in this text, published in articles as well as in more popular sources directed towards clinicians and interpreters, resulted in several invitations to speak on “training days,” that is one- or two-day events for professionals involved in dementia health care. A challenge at such events was to strike a balance between criticism of existing practices and not threatening face, while at the same time suggesting easy-to-implement interventions, taking into account financial and practical constraints on the various stakeholders. Relating to the episode analyzed in this chapter that displays the struggle of the clinician and the interpreter, with several restarts of a test task before it was carried out in accordance with formal instructions and expectations, may of course be intimidating. When talking about such episodes to those concerned, it was therefore important to view and discuss them in light of circumstances at the time, that is the fact that clinicians in 2015–2017 did not have access to more suitable tests. Being an interpreter, unless as an accredited interpreter within law, is a badly paid, low-status job in most countries, including Sweden. Many interpreters in Sweden have only a very basic training. When presenting research to this group, it was therefore important to highlight differences between activities in memory clinics (e.g., history-taking being easier to interpret), and to encourage them to ask clinicians for information about tasks and materials in advance of an appointment, particularly for the more demanding tasks such as cognitive screening.

One quite straightforward intervention on these training days was to encourage implementation of the briefing–interaction–debriefing process (BID, see Langdon & Saenz, Reference Langdon and Saenz2016). BID is a rather simple routine (and there are other ones similar to it), in which clinicians meet interpreters a little in advance of an appointment and inform them about the aims of the encounter, what materials are to be used, whether text translations might be needed, and so on. A debriefing also takes place right after the appointment, in which the interpreter and clinician discuss potential areas of confusion. The BID procedure is thus not derived from CA per se, but is an intervention that might positively affect some of the challenges revealed by CA and the problems reported in ethnographic accounts. For the session examined in this chapter, a briefing session could have prevented side-sequences during the patient encounter, for example as to whether an item to be translated was a common word in the patient’s language and so on, consecutive instructions and meta-talk on how to carry out a task, or could even have provided a chance for the interpreter to say that she was not skilled in prima vista translation. A pre-appointment briefing in clinics, with information about tasks and materials, has proven to be high on interpreters’ wish lists (Aburto Maldonado & Eklind, Reference Aburto Maldonado and Eklind2021). This was also expressed by interpreters at these training days.

It should be mentioned that staff in some memory clinics claimed to already work according to routines resembling BID, but practical issues, such as increased costs (for booking longer sessions with interpreters), and the personal needs of interpreters (for example, the need to pick up their child from nursery school, catch a bus, and so on) very often hampered the execution of a BID routine.

4.5.3 Research Network Outcomes

Substantial research funding enabled the establishment of several networks, the three most influential of which, where CA scholars had an impact, were the Research Network on Ethnicity and the Dementias, the Nordic Thematic Network on Ethnic Minorities and Dementia, and the Alzheimer Europe expert group on intercultural care and support (Gove et al., Reference Gove, Nielsen, Smits, Plejert, Rauf, Parveen, Jaakson, Golan Shemesh, Lahav, Kaur, Herz, Monses, Thyrian and Georges2021). Many of the participants in these networks and groups were the same, but whereas the first network consisted exclusively of researchers conducting empirical work, the other two also comprised representatives from interest organizations, professionals from health care institutions and services, and persons assigned to report research findings to policymakers at national and international levels (Gove et al., Reference Gove, Nielsen, Smits, Plejert, Rauf, Parveen, Jaakson, Golan Shemesh, Lahav, Kaur, Herz, Monses, Thyrian and Georges2021). The research network fed into the thematic network and upwards regarding everything from cultural conceptions of dementia and their consequences for help-seeking patterns, attitudes towards daycare activities, intercultural encounters, and multilingual interaction in residential care, to challenges associated with diagnostic instruments and (lack of) adequately trained interpreters, highlighting similar difficulties as those exposed in Extracts 1–8. By and large, the aim of these networks was to address the overall social problem of inequity in care for minority ethnic people with dementia, proposing recommendations and solutions on the basis of results from empirical studies using different methods and theoretical frameworks, and in dialogue with a wide range of stakeholders. Assisted by the Nordic Welfare Centre,Footnote ⁵ ideas and recommendations were communicated to the national boards of health and welfare of the respective Nordic countries, and to the Nordic Ministry of Health.

The Alzheimer Europe report was published in 2018 in English, French, and German, and has subsequently been modified into several shorter and more popular guides in a number of European languages, directed toward staff in residential care and health care clinics. The essence of the original report has been turned into a commentary, and also a policy proposal for improving equity in care for minority ethnic persons with dementia in Europe that was presented to the European Commission in August in 2020 (see Gove et al., Reference Gove, Nielsen, Smits, Plejert, Rauf, Parveen, Jaakson, Golan Shemesh, Lahav, Kaur, Herz, Monses, Thyrian and Georges2021). In that policy proposal, factors related to analyses of language, culture, and interaction may be summarized as set out in the list below. This hopefully resonates with what can be observed in Extracts 1–8 in this chapter when it comes to the training of clinicians (Gove et al., Reference Gove, Nielsen, Smits, Plejert, Rauf, Parveen, Jaakson, Golan Shemesh, Lahav, Kaur, Herz, Monses, Thyrian and Georges2021:4), who need to:

• Undergo training in cultural awareness, sensitivity, and competence so as to improve communication and build a relationship with people from minority ethnic groups, thereby helping to ensure timely, accurate, and differential diagnoses and treatment of dementia amongst members of minority ethnic groups;

• Use culturally sensitive/fair and appropriately validated screening, assessment, and diagnostic tools for people from minority ethnic groups, which are administered with the help of qualified/trained interpreters (preferably with medical accreditation) when needed – mere translations are not sufficient;

• Make every effort to enable every person to be assessed and diagnosed in their best/preferred language at no extra charge, and for people to be informed of this possibility;

• Not ask relatives and friends to act as interpreters, except for emergencies or exceptional circumstances, but to consult them during the assessment process if required and subject to the agreement of the person being assessed.