“Atypical Interactions” in Healthcare: A State-of-the-Art Literature Review of Conversation-Analytic Research, with Reflections on Equity, Diversity, and Inclusion

ABSTRACT

Our state-of-the-art review examines conversation-analytic (CA) research on healthcare encounters in which the patient has a disability that can affect their communication, using the lens of the social relational model of disability. We report findings via the following themes: where (and why) interactional challenges arise, practitioner barriers and facilitators to effective communication, how behaviors interpreted medically as symptoms can be understood as interactional competencies, and understanding the role of companion. We reflect on equity, diversity, and inclusion in CA research, arguing that labeling interactions as “atypical” can lead to assumptions that findings are circumscribed and have no wider applicability. Existing research has focused on specialist settings created for people with one particular condition, overlooking their healthcare interactions with other healthcare services. We argue that CA research on healthcare and beyond should include a more diverse range of interactants and also discuss why. Data in examples are in multiple languages.

Background

People with disabilities face unfair and avoidable differences in healthcare and outcomes compared to the general population. For example, people with intellectual disabilities are more likely to die from causes that are preventable with sufficient care (Glover et al., 2017), and people with dementia more likely to be admitted to hospital than those without (Prince et al., 2016). These are populations with increased likelihood of comorbidities, and hence they are likely to require more contact with healthcare services. Issues of equity, inclusion, and (respecting and adapting to) diversity are being highlighted worldwide. The United Nations Convention on the Rights of Persons with Disabilities, which 177 countries have signed on to, asserts equal access and opportunity to healthcare for disabled people, including appropriate adaptation of communication (Articles 21 and 25). Conversation analysis (CA), with its emphasis on sequence, turn-taking, and recipient design in both verbal and nonverbal aspects of communication, is uniquely placed to help healthcare services understand communication challenges and what adjustments do actually support interaction for people with conditions that may impact their communication.

“Atypical” interaction research

There is a long-established tradition in CA literature of studies in which one participant’s communication may be affected by a specific condition. Wilkinson’s state-of-the-art review (Wilkinson, 2019) identified the most commonly researched conditions in this field as dementia, intellectual disabilities, aphasia, autism, hearing impairment, and brain injury. Studies of interactions in which participants have one of these conditions are often grouped together as a subcategory of the field: CA of “atypical” interactions, a term that is increasingly common in CA literature. We take this opportunity to examine the wider literature on “atypical” interactions in healthcare settings and examine the notion of “atypicality” in relation to broader issues of inclusion, diversity, and equity within (and beyond) CA. We start by asking, atypical in relation to whom? These conditions are not uncommon, especially in healthcare services. The notion of “atypical” communication of course suggests that there are “typical” ways to communicate. This is perhaps an inevitable consequence of a discipline that was developed to identify universal rules of interaction, with key insights and building blocks of the method derived from “classic data” still frequently drawn upon in contemporary CA research of predominantly “neurotypical” interlocutors.

Second, the ways that these conditions may affect communication, such as a delay in language processing or production or difficulties comprehending complex sentence constructions, are also prevalent in the wider population. Everyone will engage in interactional practices that at one time or another could be deemed “atypical,” just as everyone may all require some kinds of communicative adjustment and support depending on the situated context (Williams et al., 2015). Furthermore, these diagnoses cover a broad spectrum of symptoms, and people with these conditions may not in fact have symptoms that affect their communication or may not impact their communication at all times; individuals with the same diagnostic label may also have different or varying symptoms, and there may be a host of situational, contextual, or cultural factors that impact how interaction unfolds for people with these conditions (e.g., time of day, amount of background noise, lighting, the organization of the spatial environment, mood, how the condition is understood and dealt with by people from different cultural backgrounds). Indeed, we may all interact in ways that may be deemed “atypical” depending on a range of such factors and variables but may not be classed as such without an accompanying diagnosis of a communication disorder. This can make pinning down “atypicality” elusive. There is also a danger that people, or people’s communication, may be categorized as “atypical” a priori based on their condition.

Aim of this review

This is a state-of-the-art review of the CA literature on “atypical” interactions in healthcare. The aim is to draw together the CA research on healthcare in which the patient has a disability that may affect their communication and to reflect on this body of research in relation to issues of equity, diversity, and inclusion (EDI). In order to critically evaluate these issues, we discuss the findings in the context of the social relational model of disability (Thomas, 2004). This model views disability as partly created by the interaction between the individual and the social, cultural, and physical environment (Thomas, 2017). In brief, people can be disabled by their embodied impairments and by the way they are treated by others (Thomas, 2004). This way of understanding disability acknowledges that people are impacted by their impairments and their “impairment effects” but that other factors, such as attitudes, expectations, and encounters with social barriers can also disable people or impact the level of disability they experience. Experiences of impairment and the extent to which they result in disablement may vary according to context and circumstances (Reindal, 2008). The effect of impairment and degree of disability experienced can often be ameliorated by relationships and interactions with others, how we construct our world, and the expectations and assumptions we have for who participates within it. Focusing on these interactions allows the “socially imposed” (Thomas, 2004, p. 33) challenges to take center stage, potentially paving the way for societal improvements. In examining the current literature within this framework, we reflect on issues of EDI both within healthcare communication and in the wider context of research in this area.

Methodology

We focused on the six conditions on which the largest amount of relevant CA research has been conducted, as presented in Wilkinson’s review (Wilkinson, 2019). A brief overview of these conditions can be found in Table 1 in the supplementary materials.¹ We acknowledge that due to the scale and scope of this review, this is necessarily incomplete and not an exhaustive list of conditions that may impact communication.

Included papers were published in English and used CA to present findings on healthcare interactions in which the patient has the above conditions. Papers that evaluated interventions that featured people with these communication-impacting conditions were excluded, as the focus of this paper is on healthcare as it currently occurs in practice. Additionally, CA studies in condition-specific settings that examined interactions that did not include the person(s) with the condition were excluded, as our focus on the social relational model necessitated their presence. There were no date or geographical limits. Our “condition” search terms were combined with “conversation analysis” in Google Scholar for an initial literature overview, with this search ending when relevant papers stopped appearing. Mindful that this presented a cursory overview, we hand-searched journals found through Google Scholar and their references, those that routinely include CA research, and journals that focus on health communication, disability studies, and specific conditions (see Table 2 for journal list). Experts in specific areas were contacted for reference list suggestions. Citations were tracked both forward and backward for included papers. Two literature searches were conducted: one was completed on June 17, 2022, and a further search for new publications was completed on June 6, 2023.

NVivo and Microsoft Word were used to open-code the findings of included papers. These codes were initially grouped according to condition, and subthemes were generated within the literature for each condition. These subthemes were then considered as a whole and reflected upon in the context of the social relational model of disability to create the themes below.

Findings

Eighty-seven papers were included in the final review, and details of specific papers can be found in Table 3 (supplementary file). We present the findings of our review as follows: (1) understanding where (and why) challenges arise, (2) clinician barriers to effective communication, (3) clinician facilitators to communication, (4) recognizing the competencies of those living with the condition, and (5) understanding the role of companions. These findings will then be reflected upon in the context of EDI principles.

Understanding where challenges arise

CA papers of clinician–patient interactions with patients who have one of the included conditions often focus on instances of interactional breakdown. However, often these CA studies are not highlighting these breakdowns as evidence of “deficits” of the person with the condition. For example, even in their paper that specifically focuses on hard-to-interpret talk in dementia care in hospitals, Pilnick et al. (2021) state that they are not labeling the input of people with dementia as “errors” but instead looking at “breakdowns of shared understanding” (page 4). Another example is Wiklund (2016), who focuses on repair sequences preceded by prosodic and nonverbal features from adolescent autistic patients in neuropsychiatric group therapy sessions. Where these are associated with trouble source turns, misunderstandings can be generated by literal interpretations and topical discontinuities by the autistic patients.

Other studies explore how impairments may impact traditional structures of consultations. Dooley and Barnes (2022) and Solomon et al. (2016) look specifically at problem presentation sequences. Take this following extract from Solomon et al. (2016) in which Leo, an 8-year-old autistic boy, is asked by the primary care doctor (DOC) to point to a place where his stomach hurts.

Extract 1. (Solomon et al., 2016)

Leo gives an embodied response to the doctor’s request on line 4 by inflating his stomach, which the doctor treats as insufficient, as she redirects the question for Leo’s mother to answer (line 8). The authors place this extract in comparison to typically developing children who demonstrate understanding of the medical agenda at a young age. However, the authors then show that when Leo’s mother later places her hand on his stomach, Leo places his hand on top of his mother’s. They argue that this could be Leo joining his mother to indicate the location of his pain or just a reaction to his mother’s touch and that interpreting these nonverbal behaviors is a challenge for clinicians and analysts alike.

Other studies explore challenges specific to the condition or situations that those with these diagnoses may encounter. For example, one topic of study in intellectual disabilities is the provision of accessible information (Chinn & Rudall, 2021; Mander, 2016). Mander (2016) outlines how the person with intellectual disability aligning to the preference structure of a question can have profound implications in a healthcare context. Body et al. (2013) demonstrate how patients with brain injury will routinely use the word “accident” over the clinical terminology of “injury” and that the attribution of agency routinely contrasts between clinician and patient, with patients talking about “the accident” compared to “your injury.” This is an interesting portrayal of how personal processing of a life-changing occurrence may misalign with the clinical description, causing communication breakdown.

Clinician barriers to effective communication

Many papers highlight how clinician communication can close opportunities for patients with these conditions to engage in interaction. For example, Pilnick et al. (2021) show how clinicians’ noncommittal responses to hard-to-interpret talk can cause patient frustration. Helasvuo (2004) found that the more open-ended the preceding question by the speech and language therapist, the more difficulty the aphasic speaker had in finding the words to answer it. Similarly, compound questions can cause confusion in people with dementia (Elsey et al., 2015; Jones et al., 2016). In residential care for people with brain injury, Klemmensen (2018) demonstrates how an occupational therapist (OT) appears to anticipate potential troubles in an interaction with residents with brain injury and withholds uptake or highlights potential non-sensical talk, which can lead to residents being marginalised. Plejart et al. (2015) also show how the coordination between the clinician and interpreter, including nonverbal factors such as gaze as well as the interpreter choosing what needs to be translated, will affect patient understanding.

Several studies have noted distinct interactional differences between people with aphasia and their everyday communication partners versus speech and language therapist (SLT) clinical interactions. Beeke et al. (2003) demonstrated that a person with aphasia uses subject-verb-object constructions in clinical settings but much more varied constructions in everyday conversation. Lindsay and Wilkinson (1999) found that SLTs were reticent to refer to an aphasic speaker’s troubles in talk, whereas everyday conversation partners employed direct correction of errors and open class repair initiators. Spouses were also more likely to engage in “hint and guess” sequences. This was echoed by Laakso (2014), who found that everyday conversation partners were more likely to engage in collaborative completion during “word searches” initiated by the person with aphasia, whereas SLTs would turn word searches into a therapeutic exercise. These papers demonstrate that clinical and nonclinical interactants have markedly different communication strategies. Given that clinicians in these contexts will be using communication to diagnose and make treatment plans, these differences may have a marked impact on patient outcomes.

A group of studies in audiology examined a different type of clinician barrier to effective communication. These papers address a clinical challenge: that the uptake of hearing aids is low among older adults, despite the positive impact they may have on their lives. The papers display how clinicians’ talk can hinder these discussions, for example, by using closed questions (Coleman et al., 2018) or presenting generalized information (Heinemann et al., 2012). At the center of these analyses of clinical communication is how patients often try to talk about the psychosocial aspects of hearing loss and that the clinical response to these concerns is minimal. Ekberg et al. (2014) describe the “negative emotional response” that patients have when first hearing the cost of hearing aids, which is in the tens of thousands for those who are not eligible for support. They show how audiologists (AUD) focus on the technical aspects of the discussion, without responding to emotional concerns. (For further examination of audiology interactions, see Ekberg et al., 2024).

Extract 2. (Ekberg et al., 2014)

They then demonstrate how this leads to these emotional concerns being raised again later in in the consultation and an eventual patient decision to not try hearing aids. Similarly, Brouwer and Day (2012) discuss how noncompliance in hearing aid use manifests in the presentation of problems in audiology clinics. Patients use the problem presentation stage of the consultations to describe a fault with their hearing aid and a solution (rather than a symptom), positioning them in a place of expertise and authority level with the audiologist. However, this then negatively impacts the patient, as the audiologist leads with this technical discussion and thus only engages minimally when the patient does address the impact the hearing loss has on their social life. Many of these papers describe how avoiding talking about the emotional side of hearing loss can delay the progression of the clinical agenda, as patients often resist treatment decisions (Brouwer & Day, 2012; Ekberg, Meyer, et al., 2020; Heinemann et al., 2012).

Clinician facilitators of effective communication

Studies also highlight how clinicians can modify their talk to make things easier for people with communication impairments to participate in achieving institutional goals. For example, Dooley and Barnes’ (2022) analysis of problem presentations in home visits to people with dementia demonstrates how healthcare practitioners (HCPs) bypass the traditional solicitation of patients’ problems to instead state their reason for visiting. This “request for confirmation” turn design was found in all the consultations recorded as part of this study, but only 16% of those in Robinson and Heritage’s study of general primary care consultations (Heritage & Robinson, 2006).

Extract 3. (Dooley & Barnes, 2022)

The complaint of the person with dementia in line 4 demonstrates a lack of prior knowledge of the doctor’s visit. Again, there was evidence of patients with dementia being unclear as to why the doctor is visiting across the dataset in the study. The authors thus argue that the reason clinicians begin consultations in this way is to avoid embarrassment or confusion on part of the patient, especially as it was always family or carers who had called the doctor. It allows, as in lines 5–9, clinicians to instead offer reassurance and clear explanations of the task at hand. In this instance, the clinician’s reformulation of “temperature” to “fever” elicits agreement from the patient, and they are able to continue with the consultation.

There are many similar analyses of the ways clinicians do interactional work to save the face of patients at moments of communication breakdown in the hearing impairment literature. Ekberg, Hickson, et al.’s (2020) analysis of repair sequences illustrates that patients and audiologists negotiate accountability for hearing problems and in doing so place blame on the audiologist for not adapting their talk appropriately. This is also shown in SLT consultations with children with cochlear implants to encourage learning (Ronkainen et al., 2017). Ekberg et al. (2018) showed how clinicians’ talk directed at parents in child hearing assessments was designed to provide reassurance and minimize mistakes, thus demonstrating clinician orientation to emphasizing capacity over impairment. Skelt (2012) also shows how audiologists will allow patients to close sequences where misunderstandings have occurred, even where there is no shared understanding achieved, allowing the patient to have agency over the conversation.

Other examples of clinician facilitators include the how OTs can summarize, demonstrate their thought processes, and use open-ended questions to help with problem identification in goal-setting meetings for people with brain injury (Hunt et al., 2015). Similarly, Ekberg et al. (2017) show how the use of shared decision-making tactics by audiologists, for example, presenting all the cost options at the start of discussions, can encourage more opportunities for patients to air their views on wearing hearing aids Much of the literature in aphasia focuses on identifying communicative practices that SLTs use to facilitate therapy sessions. For example, Simmons-Mackie et al. (1999) showed how SLT sessions were structured around “Request, Response, and Evaluation” sequences. When the person with aphasia was unable to answer correctly, progressivity was suspended while the clinician encouraged an acceptable response. CA work on these activities has also revealed the degree of cooperation and interactional skills of both SLTs and people with aphasia in managing and collaborating on institutional goals (Merlino, 2018).

Competencies of the person living with the condition

Several CA studies reveal the behaviors of patients that may seem or have been treated as inapposite have logic and interactional functions. For example, Azios and Archer’s (2018) case study of a client with brain injury who frequently sings examines where the singing occurs sequentially in the interaction to see what action it may be performing. The authors describe how all the elements of the song, from delay to the lyrics used, show the client disagreeing with the clinician’s questioning, as in extract 4 below, where the client, Tim, is not engaging with goal-setting activities.

Extract 4. (Azios & Archer, 2018)

Similarly, Damico and Nelson (2005) showed that two “problematic” behaviors produced by an autistic teenager (hand-flapping and nonlinguistic utterances), could be methods to communicate despite a lack of verbal language. Stiegler (2007) analyzed interactions between an SLT and a nonspeaking autistic child, highlighting underlying competencies, an awareness of conversational structure and sequence, and the ability to initiate interactions.

A subset of these studies examines autism diagnosis and assessment. This research foregrounds diagnosis as a social process by highlighting the way autistic patients’ interactions are represented by clinicians and how the impact of clinicians’ conduct and of the test instrument may be transformed into clinical facts about the autistic child (Turowetz, 2015a, 2015b; Turowetz & Maynard, 2018, 2019). Several studies have also highlighted behaviors that can be pathologized as symptomatic of autism but that may instead display communicative competencies (Korkiakangas et al., 2016; Muskett et al., 2012; Turowetz, 2015a, 2015b). Turowetz (2015b) uses the example of a child who, in the course of assessment, attempts to drink from a picture of a cup. This was interpreted by the assessor as an example of his confusion between reality and the pictorial representation, while Turowetz posits that this could also be pretend play: evidence against a diagnosis of autism (Hollin & Pilnick, 2018).

A subset of studies highlighted the competency populations with certain conditions displayed in understanding and aligning to the function of the clinical practice. Chinn and Rudall (2021) found that when clinicians selected patients with intellectual disabilities to answer questions in primary care, the patients generally answer appropriately and without significant gaps or dysfluencies. Likewise, in family therapy sessions, Pote et al. (2011) found that participants with intellectual disabilities managed sensitive talk around therapeutically relevant subjects such as vulnerability and protection by using interactional strategies such as topic shifts to reframe the talk toward more positive therapeutic dimensions.

Augmentative embodied interactional competencies were also demonstrated. For example, Klippi (2014) showed how pointing could be an interactional resource utilized by aphasic speakers for a range of uses. In one example, Klippi demonstrated that an aphasic patient used pointing in three different spaces on the table in front of him to signify his three dogs. These spaces then functioned as semiotic spaces, where subsequent pointing to one of the three spaces would indicate a specific dog. Similarly, Laakso (2014) demonstrated how aphasic patients in speech and language therapy sessions used frowning and laughter to display recognition of linguistic difficulties, thereby managing potential linguistic problems by recognizing, signaling, and managing the trouble source in an embodied display of interactional competence.

Role of companions

Studies of the role of companions, for example, family members or paid carers, in clinical interactions often explore the tension between companions being in some way responsible for the person living with the condition alongside the need to support and recognize their independence. How companions manage this role and how others treat companions and their rights and interactional responsibilities can have profound implications for the people they support. For example, people living with dementia may not provide clinicians with what companions see as an accurate depiction of their situation, which causes delicate situations for clinicians as they seek to clarify the patient’s history and presenting problem (Dooley & Barnes, 2022; Dooley et al., 2018; Elsey et al., 2015). The issue of who has epistemic primacy over the patient’s health is often explored by examining speaker selection. Chinn and Rudall (2021) found that patients with intellectual disabilities visiting primary care services were overwhelmingly selected to speak by the healthcare practitioner. Companions were only selected to speak in 9.3% of the total question–answer sequences and typically in specific contexts, such as when the patient’s answer was treated as lacking sufficient information. They also found that family members were more likely to be chosen to speak than social care staff and were more likely to self-select/intervene to answer questions directed at the patient. Antaki and Chinn (2019) focused on instances where the clinician asks the patient a question, and the companion intervenes to provide the answer. They describe a gradient of intervention, from hints and tips to naked imposition of the patient’s epistemic terrain. When companions did impose, they generally did so in a way that was carefully designed to give the turn, and epistemic primacy, back to the patient. In the following extract, a primary healthcare clinician (NUR) has asked the patient with learning disabilities about his mobility. He then responds by detailing the guidance he has received from the OT. We join as he reaches the end of his extended turn.

Extract 5. (Antaki & Chinn, 2019)

In response to the nurses question that seeks information on the patient’s mobility, the patient gives a positive assessment of how he has been managing his mobility issues (So far >t’ch wood< ((taps table)) “s bin going good). Perhaps mindful of directly contradicting the patient with their own assessment, the companion skillfully draws on the indirect reported speech of the patient as evidence that his mobility may not be being managed quite as unproblematically as he has portrayed, recalling a recent conversation in which the patient had spoken of difficulties getting up the stairs. This highlights the precarious position that companions may be in in these settings; to intervene risks stepping into the patient’s epistemic space, while to not intervene may leave the clinician with an overly positive view of the patients’ independence or their management of health issues.

Companions can also intervene to fill gaps in knowledge and restore or aid intersubjectivity (Chinn, 2022). This can take the form of repairing patient talk to aid the comprehension of the clinician or through recasting or reformulating the clinician’s talk to facilitate understanding and response from patients. However, these interventions can also be seen as de-skilling both clinicians and patients by relieving the pressure of both parties developing and practicing communicative strategies to achieve mutual understanding as well as implicitly casting doubts on the other parties as competent social actors (Chinn, 2022).

Applications

This review demonstrates that the level of disablement a person may experience in communicational terms may be lessened or ameliorated though skilled interaction from clinicians and companions. Hence, there is great scope for the use and insight of CA in both policy and communication training for clinicians and medical professionals in this context. Broadly, disability-relevant health and social care policies and training programs in the Global North are centered around core principles of choice, independence, control, and personalization (Carey et al., 2019; Christensen & Pilling, 2014; Lin & Lewis, 2015; Power et al., 2022). However, the papers in this review demonstrate how these concepts are talked into being and are never in an agenda-free environment. CA can uncover whether and how these values become interactionally enacted and/or relevant forces in practice and the issues interlocuters face in so doing, which in turn can be fed into policy and training.

In this review, we have found that the interactional competency of interactants living with these conditions (or the recognition of this competency) is often linked to, or partly contingent upon, the communication skills of the clinician(s) they interact with. While staff training based on interactional evidence cannot eliminate the very real communication issues people may live with as a result of their condition(s), being attuned to the idiosyncrasies of how a person communicates is likely to lessen some negative factors associated with having a condition that impacts communication, such as struggling to be understood or feeling isolated. Put simply, communication is a joint endeavor, the success of which rests on all parties. To conversation analysts, this is not surprising since the discipline is built on an understanding that our talk is shaped by prior utterances, which in turn shapes future talk (Schegloff, 2007). This adds further weight to a social relational perspective of understanding disability in its relational context and environment. For example, we can see how our understanding and interpretation of disability, or rather behavioral “atypical” traits associated with certain conditions, can pathologize and transpose behavior as being indicative of certain conditions, whereas a CA approach can demonstrate that these actions may have been prompted by professionals themselves, rather than emanating from individuals (Maynard & Turowetz, 2019).

Likewise, we have shown that CA is uniquely well placed to uncover competency and orderliness in individuals’ communication that may have been missed (or misinterpreted) in the to-and-fro of the interaction itself (Korkiakangas et al., 2016; Muskett et al., 2012; Turowetz, 2015a, 2015b). This brings to mind a quote from Weismer et al. (1992, p. 68): “[i]ntelligibility is as much in the ear of the listener as it is in the mouth of the speaker.” However, it is important not to miss people’s vulnerabilities and dependencies in the rush to empower them through recognizing their competencies (Redley & Weinberg, 2007). To ensure that the balance is correct, interventions with CA should be built from the ground up with collaboration and input from people living with these conditions or those who have expertise by experience of supporting them (such as carers), as is already occurring in the Better Conversations program (Volkmer et al., 2021), the VOICE study (O’brien et al., 2018), and the Forget Me Not group (Dooley & Webb, 2023).

Additionally, CA-informed healthcare training materials should be freely accessible and designed for use by the populations with conditions outlined here as well as companions who frequently support them in healthcare settings (e.g., friends, family, paid support workers). This is vital because companions are key interactants whose contributions can facilitate or impede healthcare interactions (Chinn, 2022).

Last, CA can be used to better understand the interactional enactment of specific institutional activities to be adjusted to better fit the needs of all parties within the consultation. For example, work by Deborah Chinn and colleagues into how annual health checks are delivered for patients with learning disabilities can, and has already begun to, play a vital role in exploring the interactional contingencies of such practices (Antaki & Chinn, 2019; Chinn, 2022; Chinn & Rudall, 2021).

Issues for reflection

Examining the literature within the framework of the social relational model of disability has allowed us to highlight similarities between widely disparate conditions that render clinical communication “atypical” and provide generalized observations on the implications of these studies on healthcare communication research more widely. We now plan to reflect on these using the categories of “equity,” “diversity,” and “inclusion.” This enables us to meditate on the work that still needs to be done, both in healthcare communication and in CA research, to improve the experiences for these often-underserved populations.

Equity

While equality entails that everyone is treated fairly (Brewis, 2017) and with respect and has equal access and opportunity for healthcare provision, health equity refers to inequalities that may be unfair or as a result of injustice, whether intentional or not (Kawachi et al., 2002). Health inequities relate to the difference in opportunities groups have to achieve optimal health, leading to avoidable differences in health outcomes (Braveman, 2006). It is important to consider disability and the focus of previous CA research on populations with communication impairments within this lens since such health inequities can stem from social, environmental, and interpersonal aspects of healthcare delivery (Baciu et al., 2017).

CA highlights successful (or unsuccessful) communication as a shared endeavor and hence, in essence, has an egalitarian spirit. However, given that we were able to find only 87 articles in the vast CA literature on healthcare communication that focused on populations with these conditions demonstrates that there is still much work to be done. We know that interventions based on CA evidence can improve healthcare through both experience and outcomes (Kronman et al., 2020; Mccabe et al., 2016). We also know the vast health inequities that many of these populations face, and thus a focus on “able” populations in CA research, will increase preexisting health inequities.

Our theme “competencies of the person living with the condition” demonstrates how CA findings can emphasize the abilities of those with considered “impairments.” This is especially important given the pathologizing of communication difficulties as “symptoms,” which is likely to occur in healthcare and is specifically highlighted in the literature exploring autism diagnosis.

However, it is worth noting that only 37 articles in our review highlight patient competencies, with many focusing on challenging situations (such as communication breakdown) where clinician talk can help or hinder the task at hand. This is very different from the literature in non-healthcare settings, where it is much more common to take a competency-based approach and focus on each interactant equally (e.g., Worrall, 2022). In sharing the focus more equally between the person with a particular condition and their interactional partners, rather than a focus on the behaviors of the “typical” communicator, these studies can provide a more complete picture of interaction, which may be more fruitful in improving communication for these patients. This should be an aim for future CA research in clinical settings.

A reconceptualization of the populations in this review as representing the diverse spectrum of human communication, and not “other,” is needed if we want to meet the communication needs of these populations fairly and equally. One way of doing this may be through comparative studies, similar to the study of Dingemanse et al. (2013) across languages, where we can start to pinpoint where differences actually do occur, rather than assuming them due to someone’s diagnosis. Indeed, we have highlighted some examples of such studies here (Elsey et al., 2015; Jones et al., 2016; Solomon et al., 2016).

Diversity

Diversity involves engaging with “atypicality” while recognizing that certain groups experience more inequality and inequity than others (Brewis, 2017). As discussed above, CA research including people who are diagnosed with these included conditions is in the minority despite the need, repeatedly emphasized by both clinicians and patients, for evidence-based guidance to improve communication (Nicolaidis et al., 2015; Perry et al., 2014; Ziviani et al., 2009). This highlights the lack of diversity in the populations sampled for CA research.

Even where there are relatively extensive findings in certain clinical populations, these are in condition-specific services for that condition. For example, the clinical literature on aphasia is mostly based in speech and language therapy, and that on hearing impairment is mostly based in audiology. That only one study in autism focuses on a nonspecialist healthcare setting (Solomon et al., 2016) is a notable gap in the literature given that autistic people are more likely to use a variety of health services (Zerbo et al., 2019), have lower satisfaction with patient–provider communication, and significantly more unmet needs compared to the general population (Nicolaidis et al., 2014).

A focus on certain specialist settings, that is, those that are created solely for people diagnosed or being assessed for specific conditions, undoubtably leads to less diversity with the wider CA and healthcare literature. Due to the focus on assessment/testing, most of the autism literature is on children or adolescents, revealing a significant gap in terms of autistic adults or people who are diagnosed later in life (Hand et al., 2021). This is the same in the dementia literature, where it is only findings from nonspecialist settings, hospitals (O’Brien et al., 2018), and GP home visits (Dooley & Barnes, 2022) that include patients with advanced dementia. It also limits the type of findings that we can draw on. For example, Chinn’s dataset underpins most of the papers examining intellectual disabilities, and these focus on annual health checks. These appointments impose a particular kind of interactional structure, focusing on a semi-standardized check which is not responsive to immediate health concerns or new/current symptoms. There is still a paucity of research examining more complex interactional demands on patients with intellectual disabilities that may occur in unscheduled or emergency healthcare (e.g., describing pain levels or shared decision-making about treatment) or in primary care visits that are not planned in advance to check general health. Given the health inequalities people with learning disabilities face (Heslop et al., 2014), this is a potentially important future area of study for CA research.

There is also a clear lack of diversity on a broader scale. There is no CA study of healthcare interactions using sign language, for example. There is very little CA research focusing on brain injury. There is little work in emergency, acute, or primary care settings. Another issue of key import is the ethnic and cultural differences within these populations.

Another diversity issue in this population relates to capacity to consent and how it affects populations taking part in research. It was often difficult to know whether the studies were only including people who had capacity to take part, as information regarding informed consent and capacity was often elided. This echoes work by Penn et al. (2008), who found that informed consent was rarely touched upon in the descriptions of studies recruiting people with aphasia. Where capacity was mentioned, this is sometimes to be clear that people could be excluded from the study if they had significant cognitive difficulties (e.g., Horton, 2006).

Inclusion

Inclusion means that we find a space where everyone is represented. To promote inclusion, we must first understand how people with certain conditions may become excluded (Williams et al., 2020). One way inclusion has started to be broached in CA more broadly is through working with members of the public as co-researchers, both in planning research projects/questions and in the analytic process, so that they can bring lived experience to bear on guiding the analytic focus (Dooley, 2020; Webb et al., 2020). This is particularly important with disabled people and those with communication or sensory impairments, as they are more likely to be marginalized. Appropriate time and funds need to be provided within research teams to do this effectively (Webb et al., 2020).

Another issue related to inclusion is recruitment. One reason much of work studying those with these conditions focuses on social care or specialist clinical settings is that this is the easiest place to recruit participants. These are congregate settings where everyone may be a relevant research participant, and as researchers we thus do not need to rely on the relevant people happening to “pop up” in opportunistic sampling, which is time- and resource-intensive. Data collection windows are often short and need to be focused on where researchers know they can get data. However, without broader research into how people in these populations interact with general health services/clinical environments and how those clinicians interact with them, CA will sadly underrepresent these populations in the broad variety of clinical CA work that has developed.

The challenge is not only that people with these conditions represent a minority of the general population but that they represent an even smaller minority of the population that access healthcare. Many of the people within the populations we are discussing have barriers to going to their doctor with a health problem, through negative previous experiences, accessibility and travel issues, or trouble identifying and explaining a health issues (e.g., Ali et al., 2013). This adds further difficulty in researchers recruiting these participants in nonspecialist settings. However, we as a research field need to increase the involvement of those with conditions that affect communication in studies of nonspecialist settings. In the next section, we discuss how we can make this happen.

The future of CA and “atypical” communication in healthcare settings

So, how may we counter these issues? We have four suggestions.

First, we need to include the discussed clinical populations in general studies of all healthcare interactions, especially gateways like primary care. Researchers in general healthcare settings should expect, and plan for, the entire spectrum of the population being part of their cohort. Ethical and methodological challenges result in people who may lack capacity frequently being excluded from research (Shepherd et al., 2022). For instance, researchers may lack knowledge about assessing capacity or how to produce accessible project information. Remedying this may require participant information targeted at specific groups of people (e.g., producing “easy read” or accessible information or short films or working with family members) and extra familiarization by researchers with the policy regarding capacity and consent. There are many useful, in-depth resources available on how to do this (e.g., www.capacityconsentresearch.com). These steps will require additional resources and an increase in planning, data collection, and analysis time, which of course must be recognized by researchers and funders alike; however, it is our role as researchers who care about these issues to push for this extra resource. Further proposals on improving equity and diversity with regard to CA research are discussed in the Editorial of this issue (Parry & Barnes, 2024).

Second, and this is crucial for step one to have any broader impact, we should actively seek to advertise and share our data once collected. This could be an important vehicle for reducing health inequalities experienced by people who, at least in some circumstances, interact in “atypical” ways and enabling diversity of focus in CA research (Joyce et al., 2022). The populations we discuss above are minorities, and therefore any research, even with an inclusive approach to recruitment in general clinical settings, will perhaps not recruit many people from groups featured here. But if, for example, many researchers shared the one or two isolated clinical interactions from people with intellectual disabilities they have collected in primary care, then it would be possible to amass a dataset.

We need to escape from academic silos and competitive natures to improve our research outcomes. There are ethical barriers to data-sharing, but these are not insurmountable. For example, the One in a Million dataset (Jepson et al., 2017) amassed around 300 video recordings of medical consultations alongside transcripts and linked demographic, survey data, and medical record entries, all of which had participant consent to reuse for future research. This was possible with forethought about what strategies should be in place to maximize the acceptability of reuse, manage access to the data, and ensure that the dataset can only be used by academic researchers (i.e., not for commercial use) and through each researcher seeking access to the dataset applying for ethical approval to obtain access. Data-sharing would not only reduce data collection costs and participant burden, it could also help build large collections of interactions including minoritized populations in general settings. This would be an invaluable resource in better understanding how to tailor support for populations with conditions that impact the way the communicate, which could in turn help to reduce the health inequalities they face.

The third area for future research is to increase our focus on multimodal communication. No multimodal intellectual disability research in clinical settings exists, despite the calls for a better understanding (Nicholson et al., 2021). CA on sign language is nonexistent in clinical settings but also sparse in non-healthcare settings, with some notable, innovative papers demonstrating the similarities and differences in turn-taking, overlapping speech, and repair (e.g., Manrique & Enfield, 2015; Mccleary & De Arantes Leite, 2013). Even where the majority of talk is not conducted through gesture, this is still important lest methodological choices (i.e., a focus on the verbal only or at least as a priority) can ascribe interactional incompetence where in fact a closer focus on gesture could help reframe this view (Dindar et al., 2017). This is particularly important when thinking about using CA research for training purposes. A focus on these aspects may also serve to highlight how behavior that may at first seem difficult to interpret may, with the benefit of closer focus and analysis, be revealed to be strategies used to save face and retain control (e.g., Azios & Archer, 2018; Pajo & Klippi, 2013).

Last, the CA study of interactions including people whose communication may be affected by a clinical diagnosis, both in healthcare settings and more widely, needs to not be considered niche or “other” by CA researchers. Understanding more about the variety and breadth of the way we communicate can only enrich our knowledge of the ecology and fundamental nature of communication. A large proportion of the findings discussed above are equally relevant to general settings. For example, findings related to companion roles mirror findings in primary care consultations with children (Stivers, 2001) or palliative care (Pino et al., 2021). Equally, the inclusion of interpreters (Plejert et al., 2015), the issues faced in dementia diagnosis or in hearing aid discussions (Dooley, 2020; Ekberg & Barr, 2020), navigating triadic consultations (Antaki & Chinn, 2019), or how clinicians balance task-based talk with therapeutic or social talk in healthcare settings (Gordon et al., 2009) are all clinical challenges that occur in other contexts. Instead, we should look for similarities to other clinical contexts and realize that the addition of research with different populations will provide richer diversity and thus applicability to CA as a whole. As highlighted above, this can mirror work from the Max Planck Institute, which has demonstrated that in-depth comparative studies of different languages can actually further our understanding of the fundamentals of CA (e.g., Dingemanse et al., 2015, 2013).

Conclusion

The social relational model of disability reframes disability not as a characteristic of an individual but rather as arising as an interaction between their embodied impairments and through socially produced, socially constructed barriers (Thomas, 2004). Communication can constitute such a barrier. While communicational problems may not be “cured” through staff improvement or communication training, the level of disablement a person experiences may be lessened or ameliorated though skilled interaction from clinicians and companions. CA that underpins healthcare communication skills training may be a way to support those living with conditions that may affect their communication, those who accompany them, and clinicians and could contribute toward a reduction in health inequities. We hope that this overview of the literature and the accompanying reflections on EDI can inspire researchers and clinicians to take on this challenge.

Acknowledgment

We would like to acknowledge the help and support of Suzanne Beeke, Charlotta Plejert, and the members of DemiQual for advice and guidance on relevant literature to include. Also, we would like to acknowledge our colleagues at the Atypical Interaction conference in Newcastle for a spirited debate on “atypical interaction,” which helped us in our reflections. Our thanks to Rebecca Barnes, Ruth Parry, and Charles Antaki for their careful reading of multiple drafts of this article and suggestions for improvement. Last, we would like to thank the reviewers for their genuinely useful and constructive feedback.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08351813.2024.2305047

Notes

¹ Supplementary materials for this article can be found via its online version.