A staff training intervention to improve communication between people living with dementia and health-care professionals in hospital: the VOICE mixed-methods development and evaluation study

Article history

The research reported in this issue of the journal was funded by the HS&DR programme or one of its preceding programmes as project number 13/114/93. The contractual start date was in June 2015. The final report began editorial review in December 2017 and was accepted for publication in May 2018. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HS&DR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Rowan Harwood reports that he sat on the National Institute for Health Research Health Technology Assessment Primary Care, Community and Preventative Interventions Topic identification panel from 2014 to 2017. This panel had no relationship to the VOICE study.

Copyright statement

© Queen’s Printer and Controller of HMSO 2018. This work was produced by Harwood et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2018 Queen’s Printer and Controller of HMSO

People living with dementia in hospital

Around 850,000 people live with dementia in the UK, and this is projected to rise to 1 million people by 2025. 1 Dementia is common in acute hospitals, with approximately 25% of beds occupied by a person living with dementia. 2,3 Best practice and policy aim to ensure that older people are treated close to home whenever possible, but hospital admission remains necessary for many acute ailments and crises that commonly affect older people, and is likely to remain so. Patients present to hospital with a range of medical emergencies, such as fractures, urinary tract infection, pneumonia or stroke. Such presentations are frequently complicated by falls, immobility, pain, delirium, dehydration or incontinence. 4 During a hospital admission, patients need health care to cure their acute illness, manage exacerbations of chronic conditions, relieve symptoms, restore function and prevent complications. To do this, health-care professionals carry out a range of health-care tasks or activities, such as information gathering, physical assessments, medical investigations, administering medications and physiotherapy. People living with dementia also need support with other aspects of care, such as eating and drinking, washing and dressing, sleeping and safety, known as the ‘fundamentals of care’. 5 Much of the work of hospital health-care professionals involves these tasks;6 effective communication is a prerequisite of effective care.

People living with dementia are vulnerable and need attention to the psychological and emotional aspects of their care as well as the physical, not least to avoid distress and the challenging behaviours that may result. An acute hospital admission can be a frightening experience for those who do not understand it. There is ample evidence that hospital staff feel ill-equipped to care for, and effectively communicate with, people living with dementia. 7,8 The person living with dementia is usually acutely unwell. The complications of delirium or pain may contribute to distress and disorientation, making assessment and interaction more complex than usual. The environment is busy, unfamiliar and often noisy. The thrust of assessment and treatment is towards rapid evaluation, intervention and discharge, leaving little time for rapport building, giving comfort and nuanced communication with those with communication challenges.

Counting the Cost: Caring for Older People with Dementia on Hospital Wards2 reported that nursing staff and nurse managers found caring for people living with dementia to be challenging. Key areas of concern related to managing difficult or challenging behaviours and maintaining safety and communication.

Communication is not solely the responsibility or role of nursing staff. When admitted to hospital, people living with dementia will encounter, and be cared for by, a wide range of health-care professionals, including doctors, nurses, health-care assistants, pharmacists, social workers and allied health professionals (AHPs), such as physiotherapists, speech and language therapists, occupational therapists and dietitians. They also encounter domestic staff, cleaners, porters and hospital volunteers. Some of the key aspirations set out in the Prime Minister’s Challenge on Dementia 20209 are for all hospitals to become dementia-friendly care settings, and for all NHS staff to have training on dementia appropriate to their role.

Outcomes of hospital care for people living with dementia are worse than for people without cognitive impairment. 10,11 People living with dementia have longer lengths of stay, higher readmission rates and a greater likelihood of dying than people without dementia admitted for the same condition. 12 One-quarter of cognitively impaired patients will have died within 3 months of a hospital admission. 6

One possible contributor to this differential is communication difficulties. These are associated with preventable adverse events in the general hospital population,13 length of stay, poorer functional outcome and institutionalisation among stroke patients. 14–16 Studies in residential care have found evidence that poor staff communication, such as use of ‘elderspeak’ (infantilising communication), may exacerbate problem behaviours, such as resistance to care17 and physical and verbal aggression. 18 Both of these increase costs of care. 19 Relatives of people living with dementia report that ineffective communication can result in the exclusion of patients, and care lacking in dignity and respect. 2 Good communication facilitates person-centred care.

Communication problems in dementia

Dementia presents a particular challenge to communication. People living with dementia may experience deterioration in their communication abilities, as well as problems in memory, disorientation, recognition, reasoning and decision-making. 20 People living with dementia often have impaired comprehension and expression, including word-finding difficulties, lack of coherence and repetition of thoughts. As dementia progresses, communication can deteriorate to a state when no intelligible speech is used. 21

The level of communication disability experienced by a person living with dementia will be influenced by contextual factors external to themselves, such as the environment22 and the communication skills of their ‘communication partners’. 23 Hospitals are difficult environments for people living with dementia and rely on an assessment model based on intensive and repeated questioning. People living with dementia may be unable to communicate their needs (such as pain or need for the toilet), and carers may struggle to understand what the person is trying to convey. Such communication breakdown can lead to unmet need, poor care and distress. 24

Data from Counting the Cost: Caring for People with Dementia On Hospital Wards2 indicated that 72% of nursing staff felt that they lacked particular skills to communicate effectively with people living with dementia and wanted additional training. In one acute hospital, staff reported lacking confidence in providing care to people living with dementia, and having received little or no dementia-specific communication skills training. 25 Staff experience stress and reduced job satisfaction arising from challenging interactions with people living with dementia. 7,26

The Equality Act 201027 obliges public services to make ‘reasonable adjustments’ to ensure that services are accessible to all regardless of ‘protected characteristics’ including disability. Such adjustments can be argued to include the communication skills of staff. Reports into poor care for patients within the NHS have highlighted the need for improved communication between hospital staff and patients to reduce errors and improve care. 28 The National Institute for Health and Care Excellence (NICE) guideline29 on care of people with dementia highlights poor communication between the person living with dementia and staff as a factor associated with emotional and behavioural problems. The Building a Safer NHS for Patients report30 recommended communication skills training for health-care professionals.

The importance of nursing staff regularly engaging with their patients in ‘constructive and friendly interactions’ was emphasised by Francis. 28

The government’s position paper Patients First and Foremost: The Initial Government Response to the Report of Mid Staffordshire NHS Foundation Trust Public Inquiry31 advocated improved education and training on dementia, with a commitment to ‘listen most carefully to those whose voices are weakest and find it hardest to speak for themselves’.

Cowdell et al. 32 observed interactions between health-care professionals and people living with dementia in the acute hospital. Almost all communication was related purely to physical care. Many interactions demonstrated elements of ‘malignant social psychology’,33 such as ignoring, infantilisation, disempowerment, stigmatisation, accusation, imposition and disparagement, despite the health-care professional’s belief that they were being kind. 32 The structured non-participant observation method of dementia care mapping has been used to study care delivery for cognitively impaired older adults. Communication by health-care professionals during routine physical care tasks was frequently brief or absent, with a lack of introductions and courtesies, and even ignoring the patient. Patient-initiated interactions were often deflected by health-care professionals, with promises of attention later. Person-centred care, when it was observed, was time-consuming, particularly if the person living with dementia had a communication problem. 6,34

Communication competencies

The ability to communicate sensitively and achieve meaningful interaction is a core competency for supporting people living with dementia. The National Minimum Training Standards for health-care support workers and adult social care workers in England include ‘effective communication’. 35 There is a wealth of advice on communicating with people living with dementia. This includes eliminating distractions, ensuring that hearing aids are working, taking time, positioning oneself in full view and at the same level, speaking clearly and calmly, and using short, simple sentences. 24 There is also a body of practical expertise among mental health professionals. More abstract components, such as the use of body language, making the person living with dementia feel valued or appropriate turn-taking, can be difficult to describe.

Small et al. 36 identified 10 recurrently recommended strategies, of which they found only three had a positive impact on observed communication breakdowns between family caregivers and people living with dementia (eliminating distractions, simple sentences, yes/no questions). One strategy (slow speaking rate) resulted in more breakdowns, a finding confirmed in other studies. 37,38 A slow speaking rate is disliked by older people,39 but is still recommended in a number of current guidelines (e.g. from the Alzheimer’s Society24). The use of closed (‘yes/no’) questions for successful communication is supported,21 but open questions have been found to be useful for facilitating personal conversations about feelings and concerns. 40 Sentence comprehension can be improved by limiting utterances to one proposition,41 paraphrasing and verbatim repetitions. 37 When presented with vignettes, nurses perceived carers who use simplified language as less patronising, and people living with dementia as more competent. 42 Critical communications from caregivers predict negative behaviours43 and, therefore, positive and affirming communications are recommended. 44

Perceptions about communication may differ from objective evidence from recorded interactions. Recommended communication strategies were thought to be helpful by family caregivers and health-care professionals, but both overestimated effectiveness when audio-recordings of interactions were analysed. Despite this, fewer communication breakdowns were observed when recommended communication strategies were used than when they were not. 36

A systematic review45 of the experiences of communication by people living with dementia during interactions with both family caregivers and health-care professionals identified 15 studies. A single study46 explored the views of the person living with dementia, and 14 studies reported the experiences of family caregivers and health-care professionals. Communication difficulty was a common finding. Wang et al. 47 used content analysis of 15 interviews with nurses to explore these difficulties further, and identified two themes. ‘Different language’ referred to the sense that the health-care professional and the patient spoke different languages and so could not understand each other. ‘Blocked messages’ indicated that health-care professionals struggled to interpret patients’ needs and emotions owing to impaired verbal communication and flat affect. In one study,48 nursing staff deconstructed communication with people living with dementia into ‘being in’ communication, whereby they tried to attune themselves to patients’ feelings and attempted to understand the perspective of the person living with dementia, and ‘doing’ communication, which involved using techniques such as active listening, allowing time to talk and asking questions.

The literature does not identify clear communication strategies that can be used for training to overcome communication barriers for health-care professionals and people living with dementia in the acute hospital setting.

Communication skills training

Research suggests that communication skills cannot be improved through experience alone. 49 Skills can be acquired and retained with appropriate teaching, and this leads to greater confidence in communication. 50–52 For training to be effective it needs to be practical, with opportunities to practice and receive feedback. 53,54 Transferring learnt communication skills to clinical practice happens best when courses contain role play with simulated patients (SPs), structured constructive feedback and discussion led by a trained facilitator. 50–52

Reviews of communication skills training interventions for health-care professionals suggest that communication skills can be improved when communicating with a non-communication impaired patient population,55–57 but evidence for their impact on patient health outcomes is uncertain. 53

A systematic review58 of communication skills training in dementia care identified 12 studies, but none was based in acute hospitals or involved the training of doctors. Four interventions were delivered in the patient’s own home, mostly one to one, with a focus on individualised training of the carer, and were not generalisable to hospital staff. The other eight interventions were delivered in care homes, with marked variability in duration (from 3 hours’ training59 to 15 hours’ training plus 2 weeks’ supervised working). 60 Care home studies that used questionnaires and observational measures showed positive effects on the knowledge, skills and attitudes of trained staff, but recommended communication techniques were not always clearly defined and outcome measures were inconsistent.

A systematic review61 of interventions to improve communication between people living with dementia and nursing staff during daily care reported insufficient published evidence to draw firm conclusions. The review included six studies and focused solely on long-term care facilities. Interventions varied in duration, intensity and type, from a single lecture62 to 4 weeks of work-based training. 63 Five out of the six studies showed significant effects on at least one communication outcome, but interpretation of the clinical relevance of these was limited by methodological quality and inconsistency of outcome definition.

Although the literature gives some guidance on communication skills training competencies, minimal evidence comes directly from the general hospital. Most empirical work is based on family and nurses or carers as communication partners, with no studies of doctors or AHPs. To develop an effective communication skills training intervention for interacting with people living with dementia in acute hospitals, we need a better understanding of what works in this setting through basic research to explore the communication problems and how they can be overcome. 54 Recommended attitudes, techniques and approaches cannot simply be assumed to be effective.

Conversation analysis

Conversation analysis (CA) is a well-established sociolinguistic qualitative method for the analysis of social interaction and communication64–66 that has been used to develop successful communication skills training interventions in fields such as stroke,66 psychosis67 and primary care. 68 For example, in stroke care, the recommended ‘supported conversation’ approach to training health-care staff and volunteers to communicate better with people with aphasia69 was based on empirical work using CA to explore the communication of video-recorded volunteers. 70 The skills needed for successfully communicating with people with aphasia were characterised around the concepts of ‘revealing competence’ and ‘acknowledging competence’. The training emerging from this was found to be effective in several trials. 71,72 CA of outpatient consultations between psychiatrists and clients expressing delusional views has demonstrated how the alternative approaches taken by psychiatrists can lead to a change in client responses and thus to more or less constructive consultations73 and this has also been developed into a tested training intervention. 67 CA has also shown that different communication approaches might be more effective at different times. For example, in conversations about advanced decisions and end of life, CA has shown that a direct approach from health-care professionals is harder for the client to deflect and is necessary when an immediate decision is needed, whereas more easily deflected indirect approaches are more appropriate to encourage patient-led decisions when there is more time and a greater priority on avoiding communication breakdown. 74

The existing literature supports the use of fundamental research using CA to collect evidence about communication between health-care professionals and people living with dementia in hospital, and to use this to develop training.

Conclusion

This introduction has outlined that communication problems faced by people living with dementia are common in the acute hospital setting and contribute to problems for staff and poorer experiences and outcomes for patients. Staff feel underqualified to communicate effectively with people living with dementia to deliver satisfactory and fulfilling care. We have identified a lack of evidence to support specific communication training interventions for health-care professionals working with people living with dementia in the acute hospital setting. To improve care, and rise to the challenges set by the public and policy-makers around dementia-friendly hospitals, a deeper understanding is required of how health-care professionals in acute hospitals communicate with people living with dementia, which aspects and techniques are good and which cause communication breakdown.

The specific research questions to be answered in this project were:

1. What should we teach? What constitutes good communication skills, including content, linguistics, context and facilitators that overcome communication challenges experienced between health-care professionals and people living with dementia?

2. How should we teach it? What are the components of an effective communication skills training intervention for health-care professionals caring for people living with dementia and how should this training be delivered?

3. Can we teach it? Is this communication skills training intervention feasible, acceptable and effective?

To answer these questions the following empirical research was undertaken:

1. An update of a systematic review on the content and effectiveness of dementia communication skills training courses. 58

2. Conversation analysis of video-recorded encounters, supplemented by observations, to analyse the structure of communication patterns used by health-care professionals to communicate with people living with dementia.

3. Development of a novel communication skills training intervention based on the findings of the CA, systematic review, expert consensus and service user experience. This included a pilot study to test the training course in real time with selected health-care professionals.

4. An evaluation of the effectiveness of the communication skills training intervention on intermediate outcomes using a before-and-after (B–A) design to assess acceptability of the course and changes in self-assessed competence and confidence, dementia communication knowledge and communication behaviours in health-care professionals who completed the training.

5. An interview study of a sample of the health-care professionals who participated in the training and clinical managers, to examine the acceptability and experience of the training and the importance of this training to the skills of the ward-based clinicians.

Introduction

The only systematic review58 of communication skills training in dementia care that we found included papers published up to 2010. None of the 12 studies identified was undertaken in a hospital, the training interventions were varied and the methodological quality of the evaluations was generally poor. This review concluded, however, that communication skills training in dementia care led to an increase in positive interactions and improved quality of life for people living with dementia. It also reported a significant impact on the communication skills, knowledge and competencies of both professional and family caregivers.

The aim of the current systematic review was to update the previous review, in order to inform the development of a new communication skills training course and to identify suitable outcome measures for the evaluation. In doing so, we aimed to identify current knowledge on the content, didactic approach and effectiveness of dementia communication skills training courses in various care settings. Specific questions for the review were:

• What types of communication skills training were evident, taking theory and content into account?

• What didactic methods were used to deliver the training?

• What contextual factors (e.g. location, organisation) have been studied, with what results?

• What is the evidence of the effectiveness of communication skills training, and on what outcomes?

Methods

We developed the search strategy following that described by Eggenberger et al. ,58 in conjunction with a research librarian, and extended it to include online dementia communication skills training. We initially searched for primary research published between January 2010 and August 2015. We updated the review in August 2017 with searches for articles published between August 2015 and August 2017. Electronic bibliographic databases were searched, including MEDLINE, Allied and Complementary Medicine Database (AMED), EMBASE, PsychINFO, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews (CDSR), Web of Science and OpenGrey. Search terms were adapted for use across different databases, including key word and medical subject heading (MeSH) term searches, when appropriate. Box 1 shows the inclusion and exclusion criteria and Figure 1 describes the results of the search and screening process.

FIGURE 1.

Communication skills training systematic review 2010–17, Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram.

As an example, the search strategy for MEDLINE was a keyword search of:

word group 1 communicat* OR interaction* OR behaviour* OR behaviour* AND

word group 2 train* AND

word group 3 dementia OR Alzheimer* OR “cognitive impairment*” OR “behavioral disturbance*” OR “behavioural disturbance”.

Papers were screened by two researchers (RO’B and RA). Disagreement on whether or not texts met inclusion criteria was resolved by a third reviewer (SG or RH). Methodological quality and risk of bias were assessed using standard criteria, based on the Cochrane Effective Practice and Organisation of Care Review Group Data Collection checklist75 and the Quality Assessment Tool for Before–After Studies with No Control Group. 76 Data were extracted from all studies by two reviewers using standardised forms.

Descriptive data were collected on the:

• theory or model underpinning the intervention and method of development

• context for training

• type of participants

• duration and model of delivery

• teaching methods.

The primary outcome data collected were the effectiveness of the training intervention, measured quantitatively, as behavioural changes, or as changes in knowledge, skills, attitudes and well-being, and reported reliability and validity of measures. The systematic review protocol was registered on the PROSPERO database CRD42015023437 (www.crd.york.ac.uk/prospero/display_record.php?RecordID=23437) (accessed 21 August 2018).

Results

We identified 101 studies for full-text review. No full text was identified or accessible for 25 of these. Twenty-one studies were conference abstracts in which no journal paper or report had been published, despite contacting the authors. Two were protocol papers for which the research had not been completed. Two were Doctor of Philosophy (PhD) theses from the USA that could not be obtained and that had not been published. Of the 76 papers with full text available, three required translation into English.

Papers were assessed by two reviewers. Following the full-text review, 49 papers did not meet the inclusion and exclusion criteria. In addition, one study was excluded as it was a duplicate publication under different authorship. This left 26 studies that met the inclusion criteria (see Table 1). Reasons for exclusion included communication training not being the primary aim or a substantial part of the programme, not being specifically aimed at people living with dementia, qualitative studies, protocol papers with no further publications and studies not being training interventions. There was insufficient homogeneity in outcomes for a meta-analysis of the results.

Discussion

This review aimed to identify and evaluate training interventions designed to improve communication in dementia care. Papers published between 2010 and 2017 were evaluated to update the systematic review by Eggenberger et al. ,58 which included papers published to 2010. Communication skills training research for people living with dementia has increased substantially since 2010. Twenty-six studies were identified, mostly B–A designs of variable methodological quality. They used a range of theoretical approaches, and spanned different settings. Few studies were directly applicable to our situation, not being based in acute hospitals or not aiming to improve health-care professionals’ communication with people living with dementia.

Studies demonstrated different teaching approaches, although it was not possible to assess the effectiveness of specific methods. Traditional methods, such as didactic presentations, reading materials and group discussions were popular, as were video-recordings, DVDs and online materials. Role play and simulations were also used. The duration of direct training ranged from a single 45-minute workshop to 120-minute fortnightly workshops for 6 months.

Studies evaluated the effectiveness of the training interventions using a range of outcome measures, including ratings of observed trainee behaviours and subjective ratings of confidence, attitude and knowledge. Several studies developed their own measures or adapted them from previously published measures. Trainees’ communication behaviours showed a variable response to training. Two of the five studies measuring this aspect reported statistically significant improvements in confidence and knowledge after training.

Previous studies indicate that role play and simulations are both viable and acceptable teaching approaches. The review also shows that most interventions used a combination of several approaches to teaching skills. There is evidence that trainee knowledge and confidence improves after training. However, given the heterogeneity of the studies included in this review, it is difficult to draw conclusions about what constitutes optimal communication skills training. The low numbers and poor quality of relevant studies suggest that there is no existing intervention that could be adapted or used in acute care.

Introduction

Conversation analysis is a sociolinguistic method for studying patterns in real-life communication encounters. It analyses what communication partners actually do, rather than what they think or say they do.

To understand how health-care professionals communicate with people living with dementia, and to what effect, we conducted a study using CA to analyse video-recordings of real ward encounters. Patients, family members and health-care professionals often cannot articulate the tacit knowledge they use when communicating, no matter how expert they are, but video-based research can specify such knowledge and skills. CA is a research method that originated in sociology but draws on insights from other disciplines, such as psychology and linguistics. 109 Its aim is to study the structure and order of naturally occurring talk during interactions. The method has been widely used to study health-care interactions. 64,110–112 We focused on identifying the everyday challenges of communicating with people living with dementia in the acute inpatient setting and, importantly, the communication skills that may overcome these issues.

We harnessed the potential of video-based research by using CA to:

• classify verbal and non-verbal practices and patterns within health-care interactions involving experienced clinical communicators

• analyse how the broad recommendations for good practice actually get implemented and operationalised

• analyse episodes in which there are challenges to the operationalisation of interactions and the ways these challenges are managed.

Methods

The study took place on eight acute geriatric medical (health care of older people) wards in a single large teaching hospital. It was approved by the Yorkshire and The Humber – Bradford Leeds NHS Research Ethics Committee (reference number 15/YH/0184). We adapted protocols for recruitment, consent and data collection used by team members during previous studies of dementia6 and CA studies. 54,66 These protocols were developed with patient and public involvement (PPI) input.

Results

Discussion

The study reported here set out to identify effective communication practices that health-care professionals used when interacting with people living with dementia in an acute hospital setting. From analysis of > 6 hours of data from a range of professional groups interacting with people living with dementia in this situation, a flexible phase structure for the encounters was identified. Two areas of interactional ‘trouble’ were identified for detailed analysis, namely how health-care professionals achieved important health-care tasks, particularly in the face of patient refusal, and how health-care professionals closed encounters.

Requests in the data set could be usefully interpreted in terms of the framework of entitlement and contingency, as developed by Curl and Drew. 124 Higher entitlement ways of requesting, which avoided the projection of a ‘no’ response in their requesting, appeared to support co-operation with the health-care professional’s requests. It is possible that by delivering a request in a manner that communicated a confident expert authority, health-care professionals enhanced the patient’s implicit knowledge of the importance of the request. As well as using higher entitlement, health-care professionals referenced the contingencies (or difficulties) for the patient, but explicitly lowered them. In doing so, the health-care professional oriented to the challenges facing the person living with dementia and demonstrated their intent to make the activity as undemanding and straightforward as possible. Offers to help, framing the action as a joint collaborative endeavour, minimising the task size, duration or frequency, and suggesting that the patient ‘try’ all served to lower the contingency. Health-care professionals did not communicate an absolute right to demand the actions of patients, but clearly indicated in their referencing of contingencies that the patient’s needs, abilities and wishes should be considered.

The prevalence of refusal and reluctance in the data prompted us to consider the ensuing predicament that, it appears, health-care professionals regularly face when caring for people living with dementia in hospital. The health-care professional who aims to provide person-centred dementia care will want to value the individual’s personhood and autonomy and respect their opinions and wishes around their health-care choices. 33,128 However, the health-care professional knows that the individual may lack the necessary information about or understanding of the action, or its consequences on their health or welfare, to make a fully informed decision. In contrast, the health-care professional is aware of how failure to complete the task might affect the person’s well-being. Typically, the health-care professional cannot complete such tasks without the active or passive co-operation of the patient. Therefore, the health-care professional needs to balance how they encourage the patient to comply with a course of action with acknowledging their concerns.

‘Person-centred’ care is often contrasted with ‘task-centred’ care,128 but it is our contention that achieving important health-care tasks and person-centred dementia care are not mutually exclusive. The project’s PPI representatives, having cared for people with dementia, attested to facing similar dilemmas, for example when encouraging their relatives to drink or take medications. When an activity is deemed to be in the person living with dementia’s best interest, the supporting person uses a variety of strategies to motivate and encourage the person living with dementia to comply with the request, with a minimum of distress. Our analysis has sought to explicitly identify what these strategies might be, and their relative effectiveness.

No single way of requesting will always lead to an acceptance or agreement, the patient’s agency being primary. However, in identifying what requesting practices ‘do’ in interactions, we aimed to specify this knowledge so as to better inform health-care professionals of communication practices that could enhance their interactions with people living with dementia.

Analysis of closings revealed a common theme of interactional trouble, with the recurrent use of open-ended pre closings, mixed messages and using non-specific and indeterminate future arrangements. These practices are not necessarily inherently interactionally problematic. In settings where patients do not have cognitive impairment, they may not precipitate trouble and, therefore, our recommendations should be taken in context. Our findings emphasise the importance of context in the analysis of health-care delivery, and the limitations of blanket recommendations. Our findings also identify a need to examine best practice guidance as it is actually produced in interaction, using methods that can unpack the interactional detail involved.

Our analysis also highlights the recurring tension in this setting between seeking to treat people living with dementia as full agents who can collaborate in joint communicative projects, and adapting communicative practices to take impairment into account. People living with dementia demonstrate a wide range of communicative abilities and these abilities can vary with time and context, which introduces another level of complexity to any interaction with them. It is feasible that practice could be improved, for example by helping health-care professionals to develop an awareness of the possible implications of using different closing practices with different patient groups, and by explicitly acknowledging the difficulties that an orientation to more generic person-centred practices can create when communicating with people living with dementia.

Introduction

Having uncovered new evidence about what communication practices might usefully be changed, we next sought to establish how these practices might be changed through a training intervention. The aim was to develop, through a transparent and robust process, a complex intervention that was ready for feasibility testing. 139

‘Intervention development is seldom a fixed prospective linear process’. 139 In common with other intervention development studies (e.g. French et al. 140), the process described in this chapter was complex, time-consuming and resource intensive. A number of intervention development approaches were used to support development, but in practice the process was iterative, messy and unique.

This chapter describes what actually happened, with accounts given for what was done, by whom and why. The aim is not to publish a description of the intervention that would make the intervention completely replicable, but to make the decision-making processes transparent and to justify the educational approaches that were taken. The output is presented in Findings (what was decided and why), using the structure of the Template for Intervention Description and Replication (TIDieR) checklist for intervention description. 141

The objectives for intervention development were to produce a well-described intervention with learning outcomes based on empirical research. It had to be underpinned by relevant theory, but feasible within the practical constraints of the project. We set out to develop the intervention using an explicit process of expert consensus, and to evaluate it as robustly as was practicable.

Methods

Processes (how the decisions got made)

The process of intervention design intended to synthesise existing evidence, new evidence, and educational, clinical and experiential expertise to produce a training intervention, illustrated as a four-stage process (Figure 2).

FIGURE 2.

Process of intervention design.

Inputs were findings from the systematic review of communication skills training in dementia care, the CA study and the interviews with local experts.

The systematic review identified a variety of candidate components, including content (what needed to change) and theories, teaching methods and modes of delivery (possible mechanisms for change). However, the quality of the reviewed studies was (at best) moderate and few of these studies were in acute care. Findings could not necessarily be taken at ‘face value’ and used in a new intervention, but needed further consideration and interpretation. 142 This was done by producing tables of components (duration, theoretical underpinnings, and teaching methods and modalities) for critical appraisal and discussion.

The CA findings identified new empirical knowledge about what effective (and less effective) communication looked like. Findings were summarised by the CA analytic team into a list of ‘potential trainables’ (see Chapter 3), but these were not sufficiently refined for training delivery. For example, they had not been considered in detail for relevance, acceptability and intelligibility by an audience unfamiliar with CA. Findings on ‘requests and refusals’ and ‘closings’ were discussed by the intervention development team. Others had previously described the application of CA findings in a group training context. 143

The interviews with local experts in communication skills training complemented the knowledge of the team. Interviewees shared experiences and opinions on questions about intervention design, including benefits of simulation; CA video methods; practical aspects of simulation and video methods (setting up, duration, feeding back, making ‘safe’, use of video and playback within simulation); training group size and composition; trainer and facilitator expertise required; recruiting to, and administering, training courses; methods of evaluation; and promoting the implementation of learning.

The intervention development team was the mechanism through which consensus was sought and intervention decisions were made. Most decisions required further deliberation or work following an intervention development team meeting. A core team undertook this, comprising three clinical academics (RO’B, SG and RH), who had considerable experience across medicine, nursing and AHP roles, working clinically with people living with dementia in acute settings, and in training and education across professional groups.

The work of ‘operationalising’ training included preparing materials and resources. Particular attention was paid to preparing simulation exercises. This was done through six meetings between the lead simulator (MM) and research speech and language therapist (RO’B), with variable involvement of the rest of the core team. The lead simulator produced provisional scenarios for the SPs, which were developed and amended through iteration in collaboration with the core team to ensure their clinical authenticity.

Findings (what was decided and why)

The intervention development team agreed specific learning objectives for the communication skills training:

1. To enable the health-care professional to reflect on, and analyse, his or her own communication, and that of others, when interacting with people living with dementia in the acute health-care setting.

2. To enable the health-care professional to synthesise new and pre-existing knowledge about communication into his or her own clinical and personal context, in order to create new practices.

3. To be able to identify and deploy flexibly a variety of effective communication practices when interacting with people with dementia in the acute health-care setting.

Following the pilot course we derived specific learning outcomes from these objectives:

1. analyse specific aspects of my communication

2. combine what I already know with new knowledge from the course

3. apply this knowledge creatively to me in my clinical interactions

4. be able to flexibly use a variety of effective communication practices.

The TIDieR checklist encourages clear specification of the core components in order to support implementation and replication. As well as reporting what was decided in the development of the intervention, the checklist makes explicit how and why each decision was made.

What: procedures, activities and processes used in the intervention

The ‘activities’ or ‘processes’ comprise the training approaches used.

Content

The content of the training was designed around the new empirical findings from the CA (Table 8). The issues for discussion were how these potentially complex linguistic findings could be distilled and ‘translated’ for a varied health-care professional audience, and how much other content there should be about communication and dementia, based on other research or approaches.

TABLE 8 - Summary of intervention content development

Sources of evidence	Summary of key considerations for intervention design (content)
CA findings	New CA findings were: the phased structure of patient interactions how to request in the face of reluctance how to close an interaction to everyone’s satisfaction The new CA findings are essential to the training content The CA findings need to be ‘translated’ into a few trainable practices that are comprehensible to a health-care professional audience
Practical considerations	The project funding specified that training would be based on the new empirical findings, although not restricted to this Given the number of ‘trainable practices’ from the CA, we prioritised content that was likely to have the most practical influence
Systematic review	The review showed no consistently agreed content for communication skills training in dementia care, and mixed empirical justifications The Australian MESSAGE intervention specified empirical evidence for each communication strategy. 79,97 Small et al. 36 identified 10 strategies from carer reports One study had content based on person-centred care in dementia99
Expert opinions (Dr Patrick Costello, personal communication and Dr Becky Whittaker, personal communication, 2016)	Training content should include clear learning outcomes Teaching on each phase of the interaction, from ‘openings’ through to ‘closings’ would give a good structure Exploring the patient’s reasons for reluctance could increase empathy but is speculative and contrary to empirical CA Ethics considerations in dementia care (best interests decision-making, coercion vs. persuasion) need to be included All levels of communication difficulty from dementia should be included, but the analysis focused on those with some remaining verbal ability Person-centred dementia care should underpin the training

The CA findings were validated by the clinicians as being highly relevant. PPI members concurred, identifying with the challenges of trying to get important tasks done when a person living with dementia was reluctant, and partings in various contexts. They also agreed that these situations were important to their relatives when unwell in hospital. Clinicians and PPI intervention development team members felt that the findings described communication practices they had not been aware of before and, therefore, regarded them as highly relevant content for the communication skills training.

The intervention development team discussed the number of communication practices that could feasibly be trained in one course. Previous CA-based training has focused on a few practices only. 67,149 We wanted to address two quite independent areas of trouble in the interaction, which involved presenting seven (closings) and 10 (requesting) communication practices to try or avoid. ‘Requesting’ in particular necessitated introducing some complex concepts.

The phase structure of encounters was felt to be helpful for health-care professionals, and there was a desire to present training in a logical ‘openings’ to ‘closings’ structure. Video material on openings was used to orientate trainees to encounters, introduce the experience of learning and encourage self-reflection on communication practices. Simplification was achieved by grouping requesting practices together under the three headings in order to help trainees better identify and remember specific practices: (1) ‘raising entitlement’, (2) ‘lowering contingencies’ and (3) ‘making the task explicit’.

Person-centred care in dementia was presented as an underpinning philosophy,128 but the intervention did not specifically teach on person-centred care. Three openly available e-learning resources relating to dementia,150 person-centred care and communication,151 previously developed by members of the study team, were recommended as pre-training preparation for trainees, who were encouraged to use these resources if they felt it necessary to revise more basic concepts, allowing the intervention to focus on new content. The person-centred care philosophy was emphasised by the facilitators during group discussions and the simulation and video workshops.

Teaching methods

Simulation

Simulated patients are professional actors who represent patients for the training or assessment of health-care staff. This is in contrast to ‘role play’, in which trainees enact roles other than their own. The possibility of professional simulation being used in the new intervention was proposed in the funding application, which included collaborators with expertise in this. The systematic review highlighted simulation and role play as key teaching methods in nine of the 26 studies, including one in which simulation was the sole teaching method,81 with significant gains in the confidence of medical trainees. We decided to use simulation as an experiential learning method, which would give trainees the opportunity to practice skills in ‘real time’ interactions.

The use of simulation in communication skills training has been challenged by a number of authors, however, who have identified ways in which simulated encounters may be ‘inauthentic’ because of systematic differences from naturally occurring interactions. 152–154 Despite potential limitations, simulation was viewed as the best method available for the online practice of new communication skills, particularly for patients with communication and cognitive impairments who could not themselves be easily trained to give feedback. Simulation is reported to be the part of training that trainees remember and value most.

Good-quality simulation has the potential to involve the whole of Kolb’s experiential learning cycle147 (Figure 3). The simulation represents a concrete experience, created as an opportunity from which the trainee can construct their own learning. Opportunities need to be given for the trainee to reflect on their own performance and also to observe and reflect on fellow trainees (reflective observation), which can be particularly valuable for trainees who are reticent or anxious about simulation. Trainees can be given the opportunity to think about and to try to make sense of their experience (abstract conceptualisation), through interpreting their previous knowledge and experience of communication, and through the input from the study findings, in the light of their new experience. Finally, they can be offered an opportunity to actively experiment through rerunning the simulation or discussion with fellow trainees and facilitators, and trying the practices out in their real clinical contexts.

FIGURE 3.

Kolb’s experiential learning cycle.

To address concerns about authenticity, we undertook to develop the simulation in innovative ways using the CA findings and data. 155 The process involved considerable consultation and development work. The research speech and language therapist (RO’B) selected suitable potential participants with dementia from the original video data, and these videos and transcripts were viewed by the researcher and lead simulator. ‘Scenarios’ were developed, the character was given a fictional name and clinical and social history, with descriptions of their retained abilities, appearance, demeanour and manner of speech. Unusually for simulations, we developed additional information for the simulator about the person’s typical interactional patterns, based on close scrutiny of the videos and transcripts and in the light of the CA findings and training content. Some examples are given for the simulation role of ‘Annie’ (Table 9).

TABLE 9 - Information on simulation character ‘Annie’

Response	Speech
You sometimes produce a lot of speech – which doesn’t make sense	. . . they’ve normally got . . . packages in cardboard on end . . . so they don’t break the points off . . . it was just bare like that . . . and it was about that day, Saturday . . . down there . . . and bent over . . . straightened it out . . . display . . .
In response to health-care professional requests:
You question request	What for?
You do not agree at first	Hmmm
You agree eventually if request is clear and direct	OK

Once the final scenarios had been agreed as acceptable by the intervention development team, they were sent to the simulators for comment and familiarisation (see Chapter 5).

Simulation has to be properly facilitated, with appropriate support for trainees. Simulation can provoke anxiety, which can inhibit learning. This was operationalised through:

• building relationships from the beginning of training using participatory exercises (e.g. in pairs, valuing all members and all contributions)

• building group identity for simulation work, keeping groups constant with the same facilitator

• allowing trainees control over aspects of their simulation, such as choosing what task to carry out

• encouraging trainees to pause their simulations (‘time out’) to give control and to allow for advice/ support mid-simulation from the group

• organising feedback using Pendleton’s model,156 in which the trainee gives positive feedback on self, then facilitators and observers reinforce with further positives, before the trainee, facilitator and then observers make suggestions for change

• encouraging trainees to replay simulations to experience their ability to control and change their communication behaviours.

Real video data

Use of video in communication skills training is not a teaching ‘method’ in itself, but is commonly used and was reported in 13 out of 26 studies in the systematic review (see Table 3). Some studies described how videos were used in teaching, referring to the demonstration of ‘good’ and ‘bad’ examples, usually staged with actors. Two studies78,96 reported the use of video of real-life encounters, showing ‘good practice’ either from expert trainers78 or from the everyday encounters of nursing aides. 96

As an alternative, conversation analysts have reported ways in which they have used analyses of real encounters as an educational resource (e.g. Kitzinger143) or in evaluations of interventions designed to change practices. 67,149 These methods have been developed into the training approach known as the ‘Conversation Analytic Role-play Method’ or ‘CARM’. 157

This approach involves pausing recordings at a key point in the interaction (e.g. after the patient’s refusal of a request) and asking trainees to consider, usually in small groups, what they might say or do next. After sharing and discussing suggestions, the recording is replayed to show what actually happened. Trainees then discuss the real response and consider how it led to the desired or undesired outcome. The trainee experiences the unfolding of an authentic interaction without knowing the outcome and is given the opportunity to ‘role play’ their responses. By using examples that play out in the direction of conversational travel that is not desired, and contrasting this with more positive examples, the trainee can experience and analyse for themselves what interactional approaches work best.

The CARM training approach can assist the trainee in experiencing the ‘disorienting dilemma’, in which their existing knowledge is challenged by something new. 158 Revealing the negative impacts of interactional practices motivates change, for example in communication partners of people with aphasia. 159,160 The importance of communication partners identifying positive alternative communication practices with which to replace the negative has been highlighted. 159 This process should follow naturally from the CARM approach, in which participants observe and analyse for themselves the interactional consequences of a variety of practices selected by the trainer to meet specific learning objectives.

An awareness of authentic practices that work better than others may not, without practice, be sufficient to change behaviour in real conversation. 161 We therefore decided to combine CARM-inspired techniques with simulation to give trainees the awareness of practices that might benefit from change, followed by skills practice and feedback. Practising communication skills with immediate feedback may work by increasing the facility with which the techniques are used, thus strengthening the trainee’s capability to change, whereas awareness building may be necessary for building motivation to change. 159,160

The research speech and language therapist completed ‘CARM’ training and used this approach to support the planning of the training schedule and resources. Short video extracts were selected to explain and illustrate each key learning point. Shorter clips were animated, with the ‘trainable’ words or phrases shown after the extract. This allowed the facilitator to involve the trainees in identifying the useful practices before confirming their findings with the animations (Figure 4).

FIGURE 4.

Sample slide showing animation of key words. Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

Three longer sequences (between 2 and 4 minutes) were used for ‘CARM’-style workshops. A sample slide is shown and stopped at the point for discussion (Figure 5).

FIGURE 5.

Sample slide from CARM-type video workshop showing CA transcript. Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

Reflective learning

Reflection forms part of experiential learning, described in Kolb’s cycle. 147 Simply experiencing something is not enough to learn from it. Reflection and reflective practice, developed from the work of Dewey162 and Schön,163 have been identified as core professional skills for health-care professionals. 164 Reflection is defined as ‘a metacognitive process that creates a greater understanding of both the self and the situation so that future actions can be informed by this understanding’. 165

The process involves ‘noticing’ an event of interest, using a critical reflective stance and applying insights to further situations. A systematic review of reflective learning in the education of health-care professionals found that reflection enabled deeper learning and improved the integration of new learning with existing knowledge and skills. 164

We required health-care professionals to become more aware of their communication practices, critically evaluate where their practices and those of colleagues work well or less well, and then integrate new knowledge in order to develop enhanced communication skills.

Teaching activities that provided opportunities for reflective learning were therefore used. This included the training being split over 2 separate days, with 1 month between sessions. A PPI contributor suggested that a reflective diary might promote the implementation of communication strategies in everyday practice. This was refined into a guided reflection; trainees were asked to reflect on an interaction that had gone well and one that had gone less well. This was thought to be a more realistic request to make of busy health-care professionals than an open-ended ‘diary’ of events. Two reflective models were suggested, a classic descriptive reflection model (what happened, how it felt, what went well and not so well, what else you could have done, and how you might handle it differently next time)166 and a model based on content from the training (‘Did you request any actions of the person living with dementia? Did you try any of the VOICE techniques for requesting? How did they go?’). Trainees were informed that they would be asked to share their reflections when they attended the second day of the course.

The sharing of the reflective diaries took place in groups of up to five trainees, with a facilitator to support, encourage and challenge. Facilitators drew out from trainees’ reflections any learning needs for that individual for the second day of the course, as well as any questions or challenges for the whole group.

Small group discussion

Simulation, annotated video clips and reflection all incorporate elements of small group discussion. All the studies in the systematic review that delivered group-based training mentioned using group ‘discussion’, ‘activities’ or ‘exercises’. Active participation is required for deep learning to take place,167 which may include discussion among peers. The size of the group may support or inhibit involvement; an optimal ‘small group’ is often regarded as between six and eight members. 168

Although early models of experiential learning and adult learning theory emphasised the individual learner,147 social learning theorists highlight the importance of the social context in which learning occurs. 169,170 ‘Supported participation’ and ‘constructive discourse’ are ‘collaborative learning opportunities’ that help learning. 158,171,172

The VOICE training incorporated much small group discussion. Interactive group tasks were designed to encourage active participation in a non-threatening way. Small group facilitators required skills in managing group dynamics and facilitating trainees to participate fully through a combination of listening, questioning and responding. 168

Tasks were also designed as ‘buzz groups’ for pairs of trainees to discuss topics together, with the facilitator’s role being to answer queries and lead feedback in plenary sessions. Discussion activities in pairs involve all trainees and contribution from the whole group, whereas open questions tend to get responses from only a few. Buzz groups use time efficiently, allowing trainees to enter into more detailed talking and thinking together. The energy created by such activities is palpably different from the dynamic that follows the asking of a question to the whole group, but does require skilful management of timing and contributions.

Membership of the small groups for the first simulation was decided by the facilitators, aiming for a variety of perspectives based on factors such as professional group, level of seniority or experience and apparent confidence or anxiety.

E-learning

Internet-based educational approaches in health care have become increasingly popular. 173 A review of internet-based education in health care showed that it was as effective as conventional teaching. 174

Project funding included support for the development of an e-learning resource known as a ‘reusable learning object’ (RLO). A RLO has been defined as ‘an interactive, multimedia web-based resource based on a single learning objective that can be used in multiple contexts’. 175

‘Multimedia’ implies the use of audio, text, images and video in combination, and ‘interactivity’ means the involvement of the learner in exercises related to on-screen content, or interaction with other users or a trainer. Both interactivity and the use of multimedia content improve the effectiveness of online training. 176 Learners accept e-learning technologies best if they are easy to use technically, align with their values and norms, and are perceived to be an advantage over alternatives. The desirability of formative feedback and dialogue with others has been emphasised. 177

Two studies92,93 in the systematic review featured online training only, both describing an interactive, multimedia resource. These two interventions covered broad dementia-related learning objectives (not just communication), and required 2–4 hours’ time investment. They were well received by care home staff trainees, and had positive impacts on measures of knowledge and confidence. Both studies included the use of videos, which trainees evaluated as being helpful in learning new ways to care92 and valuable and ‘believable’. 93 Trainees liked the flexibility of delivery, but in both studies, internet access and information technology caused problems. When asked how to improve the training, some participants suggested that sharing ideas in a group would be more beneficial. 92

The RLO focused on a single learning objective, and aimed to provide around 15 minutes of learning. We intended to use a blended learning approach in which the RLO reinforced learning from the face-to-face training,178 consistent with a behaviourist model of learning, in which repetition and positive reinforcement are key to the retention of new knowledge and behaviours. 179 The opportunity for trainees to review and revise the recommended communication practices was offered through the use of ‘real’ video encounters, which had been consented for potential online use.

The focus of the first RLO was on ‘requesting in the face of reluctance’, as this was felt to present the greatest conceptual challenge (Figures 6 and 7). Relevant information was presented in text, audio and video (‘talking head’) formats to maximise accessibility. After each slide giving information, an interactive activity was used as reinforcement, using video and transcripts. Trainees were given immediate feedback on correct or incorrect answers and the chance to self-correct. In this way, trainees were given a summary and reminder of previous teaching, allowing for multiple repeats of information delivered in their preferred modality. ‘Testing’ and feedback from the activities gave trainees immediate, private feedback on whether or not they understood key learning points.

FIGURE 6.

Screenshot from Re-useable Learning Object (RLO). Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

FIGURE 7.

Screenshot from Re-useable Learning Object (RLO). Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

Photographs reproduced with permission of Nottinghamshire Healthcare NHS Foundation Trust.

The process of developing the RLO was supported by the Health E-Learning and Media Team (HELM) at the University of Nottingham. A ’storyboard’ was written, assembling text, video clips, exercises and illustrations. The text was audio and video-recorded, allowing for background commentary and ‘talking head’ sequences.

Summary

Evidence supporting use of each teaching method modality is given in Table 10.

TABLE 10 - Evidence supporting each teaching modality

Teaching method	Sources of evidence	Considerations
		Key for intervention design	Practical
Simulation	Systematic review	7/26 studies in the systematic review included simulation, role play or ‘live skills’ practice One study81 reported confidence gains from a simulation intervention for medical students	Funding for professional simulation included within project grant
	Individual interview (Dr Patrick Costello)	Simulation is the part of communication skills training that health-care professionals report remembering positively in the long term Good simulation requires skilled facilitators to create a safe learning environment and, with expert knowledge and experience of the relevant clinical field, draw out appropriate learning Good simulation is personalised by the learner, to allow them some control and increased safety
	Expert panels (September 2016–February 2017)	Simulation offers the opportunity for ‘real-time’ practice of communication skills, trying out new skills and decision-making Simulation allows for reflection, feedback and retrying in ways that clinical practice experiences cannot Simulation has been criticised because of demonstrable interactional differences between ‘real’ and ‘simulated’ encounters, so work is required to increase authenticity ‘Role-play’ between trainees or with volunteers would allow more trainees per course, but lack of authenticity in ‘performing’ dementia makes this undesirable
Real video data	Systematic review	Many published studies (12/36) used video data, the second most consistently used technique The videos included both good and bad practice examples Videos were mostly of actors but occasionally were from real encounters (e.g. trainers demonstrating their expert skills) None used video based on CA or the ‘stop–start’ technique used in CARM152
	CA literature	Using real video data avoids potential inauthenticity of simulation by showing what happened in a real encounter Real video can be stopped and started to allow trainees to reflect on what was said and what they might say next The ‘best’ examples to use of the trainable practices were taken from video from which those practices were identified
	Individual interviews (Ms Becky Whittaker)	Significant technical challenges in video-recording and playing back simulations within a training day; takes a lot longer per trainee Video playback of real encounters used in CA can be used in training Real video data allow trainees to assess the process of their interactions (how did we get there?) as well as the outcome (did we get there?) Technical aspects must be smooth – good play back, stop–start at will, subtitling to help with audibility
	Expert panels (September 2016–February 2017)	Stopping unwanted communication behaviours is easier than starting new ones; video examples of interactional ‘trouble’ needed as well as ‘positive’ examples Video methods balance disadvantages of simulation Reflections on ‘real’ videos need to be facilitated to maintain respect for interactants and avoid digressions into negative judgments or unhelpful speculation on motivations
Small group discussion	Systematic review	All papers that described a group training intervention (15/26) reported using group ‘discussion’, ‘activities’ or ‘exercises’
	Expert panels	Simulation and CA-based video techniques involve group discussions as part of their reflective, learner-centred, experiential approach Active participation, facilitated by small groups, needed for deeper learning to occur
E-learning	Systematic review	Two exclusively online training studies92,93 Both used video examples (probably staged) Care home staff needed support with technology Technology needs to work well and be easily accessible Flexibility of delivery was appreciated by care home staff	Study had funding for development of e-learning resource and an academic specialist included as collaborator Only about half of the videos had consent to share online
	Expert panels	A variety of online tasks helps to maintain interest Clear learning outcomes needed Better to keep short and specific (separate ‘requests’ from ‘closings’) Need to avoid using same clips for online resource as for face-to-face training E-learning resource primarily to reinforce learning for trainees
Reflective practice	Systematic review	8/26 papers described using self-reflection	Training needed to be split over 2 days to give time for practice in the clinical setting
	Expert panel/s	Need to link classroom learning with ‘real’ clinical experiences to support implementation Trainees need to individualise their learning to their differing contexts A ‘reflective diary’ supports linking learning to practice Ensuring participation in written reflections can be difficult Reflecting on one positive and one negative experience would be a realistic expectation

What: materials

Various resources were required to prepare training, but these were the items most often omitted from descriptions in an analysis79 of interventions used in randomised trials. 180

The learning content, including the interactive exercises and presentation of the video materials, was prepared in Microsoft PowerPoint^® 2016 (Microsoft Corporation, Redmond, WA, USA) (see Figures 5 and 6). PowerPoint allows video/audio-recordings to be animated with transcript appearing as the words are spoken, thus enabling a CARM-type training approach in video workshops. 152 Online video resources were embedded to illustrate points.

Paper-based resources were also used. A ‘Summary of Recommendations’ card, listing the key training content on a two-sided A4 sheet, to be used as a reference for trainees both in the training and afterwards was produced following experience in the pilot course. One side showed the practices to ‘try’ and to ‘avoid’ in relation to requesting and closings, and the other side showed a diagram of the structure of the encounter, and a summary of person-centred care (Figures 8 and 9). 33,128

FIGURE 8.

Summary of recommendation on requesting and closing. Copyright Nottinghamshire Healthcare NHS Foundation Trust. Reproduced with permission.

Copyright Notinghamshire Healthcare NHS Trust. Reproduced with permission.

FIGURE 9.

Summary of recommendations on structure of the encounter and person-centred dementia care. Copyright Nottinghamshire Healthcare NHS Foundation Trust. Reproduced with permission.

Copyright Notinghamshire Healthcare NHS Trust. Reproduced with permission.

Trainees were provided with paper-based resources to support exercises, including a ‘reflective diary’, a written transcript for a video workshops and a sheet of questions to consider during a final, small-group, ‘implementation’ exercise. Simulation workshops were supported by descriptions of the scenarios, lists of tasks for health-care professionals to choose from, observation feedback sheets to guide trainees in their role as observers of simulation and guidance on simulation for facilitators.

Simulations also required ‘props’ to support authenticity, including hospital beds, chairs, bedside tables, blankets and pillows. Other resources supported the carrying out of selected tasks; these included a choice of drinks and biscuits and wash things, with access to water, cups and bowls. Simulation experts were clear that trainees should not be ‘pretending’ that something was there that was not (i.e. miming), because the health-care professionals should not be seen as ‘acting’ but rather as being themselves in their professional roles; only the simulator is truly ‘acting’.

The e-learning RLO forms part of the ‘materials’ of the course and are available online (lead author Rebecca O’Brien; see www.nottingham.ac.uk/helm/dev-test/dementia_and_request).

Pilot course

A pilot of the VOICE for Dementia communication skills training course was run with two primary objectives:

1. to optimise the intervention

2. to optimise the evaluation of the intervention.

These aims were consistent with the modelling of process as part of the ‘development’ phase of the Medical Research Council (MRC)’s complex intervention framework. 183 Although definitions of pilot and feasibility studies vary,184 this pilot was designed to be a trial version of the communication skills training course to test whether or not the processes could all work together and allow trainers, simulators and a sample of trainees to experience the intervention and provide feedback on it. 185

Summary of intervention design

This chapter has presented the ‘who, what, why and when’ of the intervention design processes. The development and evaluation of an evidence-based intervention is a complex, non-linear process. In this chapter, we have described how the VOICE training was coproduced with a wide group of informants. The resulting approach, based on original findings from the CA analysis, has been described in detail. We have presented findings from piloting the training course and its evaluation. Throughout this chapter, attention has been paid to the implications for the delivery of the intervention beyond the context in which it was developed.

Introduction

This chapter describes the training of actors who play the part of patients for training or assessment purposes (known as SPs). ‘Training’ includes pre-training preparation and resources, the pre-course training day and ongoing feedback and coaching during the courses. We needed SPs for both the training elements and the effectiveness evaluation study. This involved creating scenarios and training a team of SPs to bring scenarios to life.

Six SPs were recruited on the basis of expertise, experience and ‘looking the part’, that is, age and gender were appropriate for the scenarios. Their experience spanned:

• work in formative and summative settings

• work in a variety of health-care contexts

• simulating a range of acute, chronic, mental and physical conditions

• meeting diverse educational agendas (breaking bad news, end-of-life consultations, exploring patient centredness)

• being facilitated, self-facilitating and giving feedback

• knowledge and experience of confidentiality, simulation delivery (consistency, authenticity, adherence to learning outcomes) and teamwork.

Establishing training requirements

The six SPs recruited had to simulate six scenarios. Four scenarios were required for the training sessions (Jack, Maureen, Tom and Alice) and two scenarios for the evaluation assessment (Stan and Annie). Two training scenarios were used on day 1 of the course (Jack and Maureen) and two on day 2 of the course (Tom and Alice). The level of communicative impairment on day 2 was greater than on day 1, in order to present the trainees with more challenge and provide them with more scope to demonstrate their acquisition of communication skills (Table 11).

The training approach

Our approach was based on four out of the five domains from the practice guidelines from the Association of Standardized Patient Educators Standards of Best Practice:186

1. safe work environment

2. scenario development

3. SP training

4. programme management (including communication/feedback processes and channels).

The fifth domain (professional development) was not relevant to course set-up.

Conclusion

Authentically simulating a person living with dementia for the purposes of communication skills training is difficult, but experienced SPs were able to successfully learn and deliver simulations following a careful process of scenario development, training on specific aspects of dementia and the educational objectives of the course, and active feedback and support.

Aim

We aimed to evaluate the effectiveness and acceptability of the communication skills training intervention. We used the first three levels of Kirkpatrick’s four-steps evaluation model: reaction, learning and behaviour. 188 The aims of the VOICE communication skills training course were that health-care professionals would increase their confidence in caring for people living with dementia, increase their knowledge of dementia communication and change their communication behaviours. The communication skills training had to be acceptable and useful to the health-care professionals, and feasible to run. Line managers had to be willing to release staff from clinical practice to attend the course and see the benefits of dementia communication skills training.

Study design

Methods

Results

We delivered the course six times between January and May 2017. We recruited 45 health-care professionals who attended one of the courses, and 44 out of 45 participants attended the second day. For many course dates, the course was oversubscribed, although cancellations at late notice meant that numbers attending each course ranged from six to nine participants. Participants comprised a mixture of doctors, nurses and therapists attending each course, with 8 out of 45 (18%) doctors, 19 out of 45 (42%) nurses and 17 out of 45 (38%) AHPs (occupational therapists, physiotherapists, speech and language therapists and one orthotist). One activities co-ordinator also participated. Forty (89%) of the participants were female, 40 (89%) were white, four (9%) were Asian and one identified as mixed race. They had a median 5 years’ experience working with people living with dementia (range 0.3–33 years) (Table 12). Twenty-nine participants (64%; four courses) attended the training at site one, and the rest attended at site two (two courses).

The baseline questionnaires for one participant were not returned, despite repeated requests. One participant did not attend day 2 of the course nor complete outcome questionnaires. Analysis was therefore confined to 43 participants. Five participants missed at least one question on the Dementia Communication Knowledge Test.

Participants increased their confidence in dementia care and knowledge of dementia communication following communication skills training. Confidence improved in all categories, and overall on the Confidence in Dementia Scale (32.8/45 vs. 38.3/45) and course-specific confidence questions (Tables 13 and 14).

Participants improved their knowledge on the course-specific Dementia Communication Knowledge Test (mean 7.2/10 vs. 8.8/10), with a mean improvement in total score of 1.5 (95% CI 1.0 to 2.0) (Table 15).

The course was acceptable to participants, with 95% reporting that the course met their expectations, and 98% reporting that they would recommend it to other health-care professionals. The course was evaluated highly in all the categories investigated. At the end of the course, high scores were given to the question asking the participants if they remembered the skills, were using them in practice, were finding them useful and were confident in awareness and use of communication skills (Table 16).

The response rate to the e-mail follow-up 1 month after the second day of the course was 31 out of 44 (70%). Participants gave a mean score of 8.6/10 to the question ‘do you remember the skills you learnt in the training course?’; 8.4/10 for the question ‘are you performing the skills you have learnt in the training course?’ and 9.3/10 for the question ‘are these skills helpful in your role as a healthcare professional?’ There was a small increase in the proportion of participants remembering what was taught (mean 8.2/10 at the end of the course vs. 8.6/10 1 month later; p = 0.02), no change in whether or not the health-care professional was performing the skills learnt (mean 8.7/10 vs. 8.4/10; p = 0.05), and a small reduction in whether or not the health-care professional felt that the skills were helpful (9.7/10 vs. 9.3/10; p = 0.003).

Communication behaviours in the evaluation simulated encounters were considered present or absent only when both speech and language therapist raters agreed. Inter-rater reliability for each communication behaviour was mostly fair or moderate (kappa range 0–0.79; Tables 17 and 18).

The impact of training on communication behaviours displayed in the evaluation simulations was variable. Results showed that following training, when closing an interaction, participants were less likely to make a vague arrangement (56% before vs. 16% after), more likely to be specific about closing the conversation (51% vs. 79%) and more likely to announce completion of the task (0% vs. 14%) (Table 19).

There were no significant changes in behaviour on the communication techniques related to requests (Table 20). Eighty-six per cent of participants did not make the initial request explicit either before or after training and 79% did not make a subsequent request explicit before or after training. Eighty-eight per cent did not soften the initial request by checking for agreement (‘. . . is that OK?’) before or after training. However, many participants already used some of the recommended communication techniques prior to training. For example, prior to training, 74% of health-care professionals were highly entitled when making a ‘subsequent’ request (i.e. not the first request) before training; 98% of health-care professionals reduced contingencies for subsequent requests.

In the evaluation of emotional tone in the evaluation scenarios, the PPI raters showed poor inter-rater reliability on ETRS items (kappa 0.01–0.10). The communication of the health-care professionals was thought to be slightly more controlling (2.2/5 vs. 2.8/5; p = 0.002), bossy (1.9/5 vs. 2.3/5; p = 0.02) and dominating (1.9/5 vs. 2.5/5; p = 0.006). There were no differences in the other categories of emotional tone (warm, nurturing, directive, affirming, respectful, patronising, supportive, polite, and caring).

Free-text feedback identified that the most helpful parts of the course were the simulation workshops, including the immediate feedback, and being able to practice the skills learnt (mentioned by 27 participants):

The simulation exercise. We were able to take part in a small formative groups where we were open and honest with each other. The feedback from the ’patient’ was very helpful.

Participant 17

I wouldn’t say I enjoyed it as such, but the simulation part was really helpful. Being able to stop and replay was particularly good and getting feedback/watching others.

Participant 18

The reflective exercise between the two days was mentioned by five participants:

Reflection of my interactions. Discussion with colleagues, to learn from their experiences and realise that we all feel the same challenges.

Participant 12

Specific techniques/skills learnt were mentioned by eight participants. Being able to watch others undertake communication tasks, interdisciplinary learning and small group sizes were also valued.

The participants were asked how they thought that the course would help them care for patients. A number of participants responded that they felt more confident in their own skills:

Given me the confidence that what I do is correct and works and that I have a high entitlement to do task, and lower the contingency to ensure important aspects of care are achieved.

Participant 25

Much more confidence with persisting with/approaching patients with dementia.

Participant 36

Discussion

We evaluated a dementia communication skills training course using a B–A design. The course was acceptable to participants who reported using the communication techniques taught 1 month after the training. Participants increased their knowledge of dementia communication, were more confident in communicating with people living with dementia and showed some changes in communication behaviours in a simulated encounter. Participants found the simulation workshops, the reflective exercises and the teaching on the specific communication behaviours particularly useful. They felt that increased confidence would improve their care of people living with dementia.

The evaluation of educational interventions is less well developed than for therapeutic interventions in health care. The acquisition of relevant knowledge and skill is generally helpful to health-care professionals, but is cumulative. Individuals will integrate the output of any given teaching intervention with their prior experience, expertise and attitudes. Large-scale randomised controlled trials with ‘hard’ patient-related outcomes (such as mortality) are logistically difficult (or impossible). In any case, there are many more influences on patient-related outcomes than communication alone. Evaluation of education therefore typically focuses on intermediate outcomes, usually self-reported by trainees. We used the long-established theoretic evaluation framework of Kirkpatrick to demonstrate improvement in both confidence and knowledge. 188

A B–A design has disadvantages, not least ‘social desirability bias’; trainees may, subconsciously, report what they think course providers or educational researchers want to hear. If they have enjoyed a course or activity, they are likely to be well disposed towards it, regardless of any real benefit. We undertook a feasibility study to see if we could run a practical course, including innovative use of SPs, within funding, practical and logistical constraints. The evaluation of outcomes was statistically underpowered. Our trainees were volunteers, who by the very act of taking part were displaying enthusiasm for the subject, were well disposed towards learning and almost all had better than average knowledge, skills and confidence before training. We used a mix of established and (unvalidated) bespoke measurement scales that we mostly analysed by item (and which had face validity at minimum).

We attempted to measure changes in communication behaviours, using a video-recorded simulation that we blind-rated according to a checklist of behaviours. We demonstrated some changes in behaviour, especially in relation to closings, but none relating to requests, which perhaps formed the greater part of the training. This was partly because of the high baseline prevalence of some behaviours, and underpowering. We achieved good inter-rater reliability when training our independent speech and language therapist-raters, using video material from the pilot course, but this was not so apparent in rating the evaluation simulations. This will have reduced power to detect real differences, but also illustrates how complex communication behaviours are and how difficult it can be to objectively ascertain them. A communication encounter involves multiple elements: assessing a situation and communication level or ability, creating a rapport with the communication partner, assessing the practical problems and solutions for task completion, and undertaking a negotiation. This is dynamic. For example, a health-care professional may ‘test’ the situation by making a polite, non-threatening, low-entitled, indirect request, maybe with some explanation or rationale (‘I’m wondering if you’ll let me take your blood pressure?’). If the person is reluctant or refuses, different approaches may be tried sequentially, amid possible diversions or distractions, gradually introducing higher-entitled requests and lowered contingencies, until acceptance or abandonment (as part of a ‘leave and return’ strategy). Different techniques may be tried at the next attempt.

We tried to capture this in the communication behaviours checklist, for example by differentiating between first and subsequent requests. Even so, the raters (who were speech and language therapists and specialist health-care communication clinicians, with a grounding in both practical communication problems and linguistics) struggled to reach agreement on whether or not a behaviour was displayed. The task was perhaps easier for closings, when the action was more defined and concrete. The health-care professional trainees themselves considered the course to be successful on their self-assessments of reaction (whether or not the learning was a valuable experience), confidence (whether or not it enabled individuals to know if they were doing the right thing) and learning (whether or not the participant’s knowledge increased after the course). ‘Confidence in competence’ is an important professional attribute. 25 An unmeasured outcome reported by trainees was that we gave them a language to articulate what they already did, helping then to teach or guide members of staff they are managing or mentoring.

Alternatively, some communication behaviours may simply be difficult to change, or our methods were inadequate to do so.

We did not formally measure whether or not the course changed patient outcomes, but health-care professionals reported that they were still using the knowledge and skills 1 month after the course, and had started disseminating it to colleagues.

We concluded that the VOICE communication skills training course was feasible to run, and defined conditions for it to do so successfully, including the use of simulation and video excerpts of real-life communication encounters. The evaluation of educational benefit, based on intermediate outcomes, strongly suggested that it had been successful. However, we studied only a relatively small group of health-care professionals who were experienced and interested and we cannot extrapolate to the general health-care workforce. A cadre of highly trained practitioners might, however, be useful in front-line practice, in role-modelling and in the case-management of difficult cases and teaching.

Introduction

We developed and evaluated a communication skills training intervention using a B–A design and quantitative measures of course perception, knowledge, confidence and behaviours. The course included innovative features, and the use of simulation in training was unfamiliar to health-care professionals other than doctors. We wanted to understand and explore these features further, in order to help validate, or refine, the intervention choices made. We were also interested in whether or not, and how, the communication strategies that we taught were useful in practice, and wanted to understand practical and contextual factors in real hospital settings that might enable the use and dissemination of the findings, or provide barriers to implementation. We were aware that hospital clinical settings are busy and hard-pressed, and that resources and time for staff training are limited. We wanted to understand the value placed on communication skills training by clinical managers.

Methods

Results

Eighteen themes and 11 subthemes were identified, describing the experience and effectiveness of the communication skills training. These themes and subthemes were organised into categories derived from the study aims (Table 22).

Discussion

The interviews with health-care professionals attending the VOICE communication skills training course and their managers have demonstrated that the participants found the course acceptable and useful, and were remembering and using the skills in practice. The use of simulation was particularly valued as an opportunity to practise skills in real time. Learning from other professional colleagues was also useful and valued. There were challenges to using the skills in practice, and focusing on a critical mass of health-care professionals on a ward attending the course might be beneficial.

Interviews allow an in-depth exploration of issues, and may reveal things that cannot be anticipated in advance. This study provided an independent and overwhelmingly positive description of what trainees thought about, and took from, the course, some months after it had taken place. However, interviews also represent a ‘public voice’, with the risk that participants report what they think they ‘should’ be thinking or doing, or what they think the interviewer wants to hear.

Some of the findings betrayed misunderstanding, for example that requesting in a highly entitled way was a mechanism for making requests, not a separate category from it, or that the communication techniques taught would take more time, rather than to save it, as was intended. However, overall, it appears that course participants had understood and retained what was taught, were using it in practice and were finding it useful.

Developing the study proposal

Our monthly meetings with the PPI group meant that we already knew that communication between health-care professionals and people living with dementia on hospital wards was an important area that needed improving.

The methods for researching this area were suggested by a speech and language therapist (RA). We felt that we needed to agree the methods, involving video-recording people living with dementia on the ward, as acceptable with the PPI group. Insights from the PPI group were collected through two meetings and a survey. The PPI groups were attended by 12 carers or lay people and one person living with dementia. These identified the importance of the topic of communication between staff and people living with dementia, the need for staff training in appropriate skills and the group provided examples of helpful staff communication behaviours. The group discussed the acceptability of video-recording interactions at length, agreed that video was important to capture non-verbal communication and suggested that mealtimes and discharge discussions were potentially important occasions to video. They also highlighted the possible need for multiple cameras or a wide-angle lens to capture all participants in an interaction (staff, patients and carers).

One of the group, Kate Sartain, agreed to be a coapplicant on the grant. As coapplicant, she attended all the project management group meetings. She has provided detailed feedback on the application and helped to write the lay summary.

Patient and public involvement and governance

We recruited a further two PPI members to attend our PMG meetings. These members both had recent lived experience of close family members with dementia who had experienced hospital care. We recruited another PPI member to the SSC and asked the Alzheimer’s Society to provide a further representative at this meeting. These PPI members at the PMG and SSC meetings support our research with constructive suggestions and challenge our assumptions at times.

Development of study documentation prior to ethics submission

Our PPI coapplicant reviewed all study documents and the lay summary submitted to the NHS Research Ethics Committee, as did a second carer of a person living with dementia. The review ensured that the language used in the participant information sheets was acceptable to someone with dementia or their carer. One of the PPI reviewers, following her review of documentation, asked whether or not family members would be involved in the video-recorded conversations. Her concern was that if the conversation was about a sensitive matter (e.g. a conversation about discharging the patient to a care home), the carer should be present and she would not want the study to interfere with this. This resulted in us changing our procedures slightly to allow an informal carer to be present during the video-recorded conversation if they wished. We also introduced carer participant information sheets and consent forms for them to be included in the video-recordings (two carers were included in the video-recorded conversations).

Intervention development

Three PPI members, including Kate Sartain, were members of the intervention development group (which also included health-care professionals, educational experts and experts in including simulation in training, conversation analysts and academics). The team met four times over a period of 5 months for whole-day meetings. The intervention development group discussed the duration, content and structure of the training, including the SP role profiles and the content of the reusable learning object. The group was shown video-recordings to be used in the training to get their views on the acceptability of them. The PPI members made the following recommendations on the training:

• The course should be 2 days rather than 1. It was felt that 1 day was insufficient for health-care professionals to grasp the content and change their approach behaviour.

• There should be a reflective diary between day 1 and 2 of the course. This was an innovative idea that proved very successful on the course and was developed into a reflective workshop on day 2.

• The PPI members questioned how person-centred some video-recorded health-care professionals were. The videos chosen illustrated well the communication techniques we were to teach on the course. However, these comments changed the focus a little. The focus of the training became how health-care professionals use the taught communication skills techniques alongside person-centred care.

• The PPI members considered that our initial method for rating communication behaviours shown in the simulation assessments would be too difficult for service users to use (as they would have found the rating form difficult to use themselves). This issue, together with the issues raised by the PPI members that the techniques we were teaching health-care professionals might result in them being less person-centred, resulted in us choosing the ETRS as a tool for service users to rate the simulated encounters before and after training. All members of the intervention development team practised using the ETRS on a pilot video-recorded simulation assessment. It was found to be acceptable by the PPI members, although agreement between the group members on rating scores was low.

Delivering the training intervention

Kate Sartain attended two of the 2-day training courses. Her role was to support the participants (health-care professionals), to help with the administration of consent and evaluation measurement scales, and to report back to the team on the fidelity of the intervention and the acceptability of the training from a service-user perspective. She reported that the course was acceptable, well run and delivered what was planned. Kate Sartain considered that the simulated workshops were done in a very supportive way, but raised a question about whether or not there was more we could do to support participants who are very anxious about simulation. This question was raised after Kate Sartain noticed how one participant appeared ‘out of her depth’ and did not return to day 2 of the course. This situation occurred on the last training course, but we are considering how to make it clear what the course involves and how to provide additional support to health-care professionals who find simulation very challenging when we put on future VOICE training courses. Kate Sartain commented that it was very clear that the interprofessional mixed training groups were obviously of value to the participants and the ambience of the day allowed for supportive conversations. Kate Sartain also raised the issue of health-care professionals who do not have English as their first language. Consideration of further research into this matter is vital if this training is to be of value to the diverse NHS workforce. Kate Sartain also has suggested that further research will be required to design a training package suitable for the workforce in the community, particularly in care homes.

Evaluation of the training

We are aware of recently published guidance from Alzheimer’s Europe that it is no longer acceptable not to include people living with dementia in research. 200 We accept this recommendation, while being aware of the challenges this represents. However, we felt going forward that we need to include people with dementia into our research and did this for the final stage of our research, namely the evaluation of the training. We wanted to know whether or not health-care professionals would remain as person-centred after training as before. To do this, we convened a group of seven service users. These included two people with early dementia and four carers. We organised five half-day sessions (with refreshments and lunch provided) and asked the service users, following training, to rate the B–A evaluation simulations using the ETRS. 193 Feedback from the group at the end was that they found the exercise stimulating and interesting and that they very much enjoyed being included in the research.

Dissemination

Our dissemination plans are ongoing. However, our PPI coapplicant Kate Sartain has presented a poster at the Alzheimer’s Europe conference (Berlin, Germany, October 2017) on service-user involvement in research. Kate Sartain made the opening address at our VOICE dissemination conference (Nottingham, UK, October 2017). She is supporting the work we are doing to develop future VOICE study courses.

Patient and public involvement values highly the positive effect this training will have on the ability of health-care professionals to provide skilled care to people living with dementia in an acute setting, removing much frustration and anxiety. PPI representatives believe that dissemination is essential for the well-being of the patient, their carers and health-care professionals.

Summary of findings

We video-recorded 41 encounters between health-care professionals and people living with dementia in the acute hospital. We used CA to understand what worked and what did not work, in real, practical settings. Encounters followed a recognisable phase structure: opening, purpose, information gathering, business and closing, although not all encounters contained all elements. Most of these phases were trouble-free in interactional terms. Two phases were consistently, and strikingly, associated with problems: requests (and consequent refusals) and closings. The manner in which things were said had a major influence on acceptance or refusal. Unusually for health-care communication, requests were often met with an unmitigated refusal (‘no’). Skilled health-care professionals used several devices in order to gain the agreement of the patients living with dementia: they asked more directly, raised entitlement (authority to ask) and lowered contingency (reduced the difficulty), by making the task sound smaller or shorter, asking the person living with dementia to ‘try’, offering to help or proposing collaborative action (do it together). Closings were often prolonged, with the person living with dementia not recognising the usual verbal or body language cues that the encounter was coming to an end and often reopening the conversation. More satisfactory closings resulted when the end of the task was declared, a specific arrangement was made for what was to happen next, and body language that was congruent with the message or ‘closing idioms’ was used.

These original, and ‘teachable’, findings, together with evidence from a systematic review, were used as the basis for a new communication training course for experienced health-care professionals. An intervention development group was convened that included researchers, clinicians, educationalists (and, in particular, people with expertise in simulation) and PPI. The course comprised 2 days, 1 month apart, and was grounded in experiential learning theory. It used didactic learning, video clips and transcripts from real life, simulation and reflection on practice. We were concerned to draw on, and integrate, health-care professionals’ prior knowledge and experience, and to ensure that the principles of person-centred care were adhered to. As preparation, we asked trainees to complete three brief electronic-learning packages (RLOs) on dementia, basic communication and person-centred care. In addition, we developed a new RLO on requests and refusals that we asked trainees to undertake before the second session as revision. A RLO on closings is in preparation.

A training programme was devised, and manualised, and actors were trained to credibly simulate people living with dementia, with particular regard to refusal and extended closings. Courses comprised six to nine participants, with two trainers/facilitators. Simulations took place in groups of three to five, and allowed for the action to be stopped in order to ask advice or try different strategies or rerun. Peers, facilitators and simulators all fed back on performance. A pilot course was run with experienced health-care professionals, all of whom had an interest in education, and adjustments were made based on the experience.

We ran six courses in two hospitals, involving 45 participants, 44 of whom returned for the second day. Trainees were interdisciplinary, with nurses, doctors and AHPs taking part alongside each other. We undertook a rigorous analysis of the education, including three of Kirkpatrick’s four levels of educational effectiveness. These were feedback on the course, its usefulness and the methods employed; tests of knowledge and confidence in a B–A design; and an assessment of whether or not the course changed communication behaviours. This was done by asking trainees to perform an assessment task with a simulator before and after the training, which was video-recorded, in a crossover design. Videos were blind rated by two independent speech and language therapists against a checklist of behaviours. A panel of PPI representatives, including two people living with dementia, rated the test videos for ‘emotional tone’ as a measure of person centredness. Trainees were asked if they were using the techniques taught and if they were useful in practice 1 month after the course. An independent occupational psychologist interviewed a sample of trainees and managers to investigate facilitators of, barriers to and value placed on the training using a thematic analysis.

The course rated very highly. Knowledge and confidence both improved, statistically significantly, despite fairly high baseline scores. Some aspects of communication behaviour were more commonly observed in the test videos after the training than before. Emotional tone was mostly unchanged (i.e. communication had not become more, or less, person centred), although videos were rated as being slightly more dominating, bossy and controlling after training. Techniques were remembered, used and found useful 1 month after the training. The interviews found that the course was very well received, validated all the decisions made during intervention development, emphasised the value of simulation, interdisciplinary learning, reflection and the 2-day structure. Some participants found simulation uncomfortable, but almost all recognised its educational value. Learning was regularly used in daily practice and was, to an extent, cascaded to other staff, or used as a framework in teaching other staff. Length, cost and lack of consideration of communication during aggressive episodes were considered weaknesses. Wider dissemination was supported.

Strengths and limitations

The teaching on the VOICE training course was grounded in empirical research. In the field of communication training this is, perhaps surprisingly, uncommon. We used a rigorous sociolinguistic method, CA, applied to real encounters between health-care professionals and people living with dementia, to identify the structure of interactions, when problems arose, how skilled practitioners tried to overcome these, what worked in practice and what we considered to be ‘teachable’. The analysis uncovered original and interesting new linguistic findings, but was fundamentally directed at what might be taught to fellow practitioners, and practised both with simulators and in everyday care.

Dementia communication has rarely been studied in the challenging environment of the acute hospital. CA has been increasingly used to understand health-care consultation, but most communication teaching is based on experience, custom and practice. The overriding strength of using CA is that it studies what participants do in practice, not what they think or report that they do. This can also mean identifying and making explicit behaviours that the individual does not necessarily consciously know they are doing. By studying skilled practitioners we could identify both difficulties, and successes, and how difficulties or breakdowns were overcome.

Intervention development was multidisciplinary and interprofessional, including experienced clinical educators, PPI, clinicians and experts in simulation. Simulation has been used in teaching both consultation and practical skills, including scenarios involving people living with dementia,81 and difficult conversations at the end of life. Clinical practice is characterised by the need to ‘think on your feet’, responding in real time to a variety of information of unknown veracity, coloured by emotions and reactions both in the patient and the practitioner. Although some approaches, techniques and ‘tricks of the trade’ can be learnt or refined through experience, the opportunity to practice new skills and gain feedback is reported as invaluable by trainees.

Simulators, actors trained to work in clinical education, can provide consistency and challenge, and feedback either in or out of role. 201,202 In some settings, people living with a condition can take part in education, such as aphasia after stroke. However, this is difficult for people living with dementia, especially those with moderate or severe impairment when communication problems are most troublesome. Portraying a person living with dementia is not easy, with a risk of stereotyping or caricaturing, or simply producing chaotic responses. In this study, we carefully developed training for simulators, based on real cases we had observed, to enable a credible simulation experience.

We used mixed training methods, including didactic information giving (lectures, PowerPoint presentation), and made extensive use of video clips, or transcripts, of real encounters. We also used reflections on practice, and considered how to incorporate previously mastered skills and attributes, especially paying attention to understanding and maintaining person-centredness. We refined the training course based on a pilot course, to which we invited experienced practitioners who themselves had a role or interest in clinical education, allowing an informed educationalist’s view.

Education and training initiatives are often evaluated quite crudely. The opportunity to do true experimental studies is unusual. Training can be evaluated at the levels of reaction, knowledge, behaviour and impact on outcomes. 188 We used questionnaires to study trainees’ perceptions of training methods, the role of simulation and the usefulness of the knowledge gained. We used questionnaires delivered before and after training to assess changes in knowledge and confidence. Innovatively, we used video-recorded simulations before and after the training to assess changes in communication behaviour. Two specialist speech and language therapist raters, blind to whether the simulation was before or after training, used a checklist to identify the use of objectively identifiable communication behaviours, using a crossover design to control for differences in the nature and difficulty of the set task.

The interactions between people living with dementia and health-care professionals that we video-recorded had some limitations, largely determined by the need to gain consent or agreement from participants and to set up recording equipment in advance, but were otherwise unstaged. All were initiated by the health-care professional; we excluded interactions initiated by the patient. Health-care professionals were all willing to be video-recorded, and we targeted health-care professionals whom peers thought were good communicators or ‘good with people living with dementia’. This was appropriate because we were looking to see what worked in real-life practice. Less confident health-care professionals, including some with English as a second language, were reluctant to take part. Interactions were typically brief (2–30 minutes). Analysis could not take into account what previous relationship the health-care professional had with the patient, nor what previous knowledge they were working with. It is possible that health-care professionals changed their behaviour as they were being studied, although it is generally considered impossible to ‘fake’ the sorts of behaviour that are ascertained in CA. Families and other carers are especially important in the support of people living with dementia in hospitals; for the most part we did not include ‘triadic’ conversations, as these had a very different dynamic and were difficult to film in a way suitable for CA (i.e. to include all participants’ body language and expression). Many communication problems arise during personal and intimate care, but we did not film these to respect patients’ privacy and dignity.

Conversation analysis is detailed and time-consuming. Within the resources and time frame available to the study, not all themes or foci of interest could be fully analysed. CA, like any qualitative analysis, is to some extent subjective, or at risk of preconception or bias. To overcome this, regular supervision and group data meetings were held that included experienced conversation analysts. Data and proposed interpretations were also presented and discussed at regional and national data-sharing meetings, a common practice among conversation analysts.

The structure of the training intervention was influenced by our previous experiences of learning and teaching communication skills, especially in aphasia after stroke and end-of-life care. We were also influenced by our experiences researching and teaching person-centred dementia care. 6,34

Use of simulation can be controversial, particularly the issue of authenticity, and relatively expensive. We paid particular attention to authenticity in the training of actors.

The evaluation was based on six repetitions of the course. We invited ‘experienced’ practitioners to take part, and in practice this was self-defined. Participants included senior nurses with leadership roles in dementia education and service development; staff nurses from older persons and surgical wards; and AHPs, including occupational therapists, physiotherapists, speech and language therapists, an orthotist, and junior and more senior doctors from geriatric and general medical specialties. The course was free and, therefore, attractive to staff for whom access to advanced education was limited, but participants were enthusiastic volunteers. These enthusiasts or ‘champions’ are an important training target: they will be role models, will direct or supervise difficult situations in clinical areas, and will teach, informally or formally. We previously found evidence that even experienced practitioners lacked confidence in working with people living with dementia, despite this being an important part of their jobs. 25

Only three participants had English as a second language. Although doctors were well used to simulation as a training medium, this was unfamiliar for other disciplines and was seen as challenging or threatening for some. Participants were told that they were expected to undertake the evaluation tasks (questionnaires and video-recorded simulation) in order to be accepted on to the course, and this may have been off-putting. Almost the first activity undertaken by participants was a fairly challenging video-recorded assessment simulation without feedback. The emotional tone rating of assessment videos was administered by a group of PPI contributors, including two people living with dementia. The scale is simple but used words that are open to interpretation. Inter-rater reliability was poor (there was poor agreement between different raters about whether or not a feature was evident in the interaction).

We have evidence (from the assessment simulations) that the VOICE training changed participants’ strategies for closing encounters with people living with dementia, but not their requesting behaviours. We do not have objective evidence of changed behaviours in real-life clinical practice, nor of any impact on well-being of patients. This requires further research with an implementation focus, involving the systematic observation of trained staff carrying out routine health-care encounters. Our trainees were a self-selecting group, who demonstrated an interest in communication, which may explain why they appeared to have high levels of skill in some domains at the outset. This limits the generalisability of our findings and may explain why requesting behaviours remained unchanged. Trainees were using the strategies of raising entitlement and lowering contingency already, and, although the course gave them a new vocabulary with which to reflect on this, it did not result in significantly changed behaviour as they were intuitively doing it routinely. For trainees finding requesting and subsequent refusals from people living with dementia a challenge, our methods may have resulted in objective change in this regard as well. The interviews also pointed to the fact that some trainees found the concepts of entitlement and contingency confusing, which may have also been a factor in the lack of objective change in this trained behaviour. Direct observation of taught strategies being used in real-life clinical practice will also allow us to contextualise positive reports from trainees when evaluating the course and reflecting on strategy use. We acknowledge that trainees may have been subject to social desirability bias when reporting their views to the team, mitigated to an extent through the use of an independent occupational psychologist in undertaking interviews.

The use of simulation in evaluation and testing has been criticised (e.g. in employment procedures, assessing competencies and examinations). 152,203 The main problem is the tendency to perform to the teaching, learning goals or expectations in a way that would not happen in real clinical practice; this is analogous to exaggerated looking in the mirror during a driving test. For example, simulators will be given brief background information and limited key information. The assessment ‘game’ becomes for the trainee to ‘extract’ this information, and verbal devices for enabling this soon become common knowledge, thereby diminishing the validity of the assessment. In a learning situation this is not necessarily a problem; a skill is practised with a reactive human partner, enabling the interaction to be experienced, rather than just contemplated or imagined, and feedback given. In our assessment of behaviour this could have occurred: before the training; the task was undertaken without knowing specifically what we were looking for; after the training; the behaviours we were teaching had been made clear. In educational practice, the ideal assessment is clearly mapped to learning goals, making the argument somewhat tautologous. To overcome this we prompted an overt link to what health-care professionals did in clinical practice, and asked for reflection on this in writing, then in a discussion group.

We did not study the influence of communication training on health outcomes for patients. This would require a very large-scale trial, possibly cluster randomised, with large-scale training of involved staff members. This would be very difficult to do logistically, and is rarely reported for communication skills training. Instead, we interviewed trainees and managers about the usefulness of what they had learnt, whether or not they had used it, the barriers to and facilitators of use in practice, and the priority given to training in communication with people living with dementia by service leaders.

Context

In 2011 Tadd et al. 204 published their Dignity in Practice report and stated that: ‘a key message echoed by staff at all levels in the organisations involved in this study was that the acute hospital is not the “right place” for older people. The prevalence of this view has resulted in the physical environment, staff skills and education and organisational processes acting as barriers to delivering dignified care to older people’.

People living with dementia, and other vulnerable frail older people, comprise core NHS acute and general hospital users. Two-thirds of hospital users are aged > 70 years. Half of emergency admissions of people aged > 70 years have cognitive impairment (dementia, delirium or, most commonly, delirium complicating dementia), and 40% of them have dementia. 10,205 Almost half of people who break their hips have dementia, whereas others have delirium, or develop it postoperatively. 206 People living with dementia are complex and are disproportionately represented among those with very prolonged hospital stays. Health policy rightly promotes ongoing attempts to minimise the need for hospital admission, and to expedite discharge for those who are admitted. However, most admissions are for legitimate medical conditions or injuries, and the delivery of necessary assessment, treatment or future care planning. 4 Caring for people living with dementia is, and will remain, an important part of what acute general hospitals do.

Hospitals are well known to be difficult and challenging environments for people living with dementia. This is partly because of the need to focus on the efficient and safe delivery of effective physical health care, but also reflects a failure to make the ‘reasonable adjustments’ to environment, staffing, training and processes required to make services as good as they can be for people living with dementia. 25,204 Staff often recognise, and are frustrated by, a lack of appropriate knowledge and skills, and identify communication as a key topic requiring further training. 2,25

Communication difficulties are well recognised as a problem for people living with dementia. This includes the specific language skills of understanding and expression, which are compounded by poor memory, impaired mental processing or reasoning, and problems in recognition, planning, initiation and social control. In addition, comorbid problems with hearing or vision, mouth or teeth problems, delirium, insomnia and pain also make communication more difficult. A noisy and busy environment can be overstimulating, and assessment processes involving multiple new and unfamiliar faces and locations, and repeated questioning, can be overwhelming.

Attempts to improve staff training and hospital experience for people living with dementia emphasise communication. Individualising care, seeing the perspective of the person living with dementia, building relationships, promoting inclusion and providing purposeful activities are key components of person-centred dementia care, and implicitly require good communication. Misunderstanding or misinterpretation of instructions or actions, especially when delivering personal or intimate care, or ensuring safety, are important drivers of distress and behaviours indicating distress. Some advice is uncontentious: optimising hearing (e.g. by ensuring that hearing aids are working), introducing yourself and saying what you are doing. Skilled practitioners, especially from mental health and palliative care professional backgrounds, have developed considerable expertise, although they sometimes struggle to articulate exactly what they are doing, making teaching or sharing skills difficult. Little of what is promoted has derived from research using rigorous methods, although much clearly ‘works’. 190

Most published evaluations of communication skills training for people living with dementia have taken place in care homes and have targeted nurses and unregistered care workers. Brief medical student teaching, using simulation, has been reported and was successful, although in some cases it only made students more aware of their limited skills. 81 Attitudes towards communication with people living with dementia have been studied, and a framework for communication has been published based on empirical research,207 but the effectiveness of implementation has not been reported.

Interpretation

Our results draw on CA findings from other settings; the structure of encounters was similar to that previously reported,114 and the roles of increasing ‘entitlement’ and reducing ‘contingencies’ to gain agreement have been described before. 127 The value of ‘direct imperatives’ has also been reported. 208 The findings ‘made sense’ to experienced practitioners, who had not previously had the concepts or language to describe what they were doing.

Conversation analysis is strictly an empirical methodology: it describes and makes explicit what was done and what the response was, and avoids speculating about motivation, or mechanisms of action. Interpretation therefore necessarily goes beyond CA.

Dementia (and its related complication, delirium) causes cognitive (or neuropsychological) impairments, including language, information processing and reasoning. Health-care professionals strive to be empathetic and polite. Many are aware of the power imbalance between patient and professional, and the disempowering effect of the unfamiliar hospital environment, care being delivered by strangers and the unusual or threatening nature of many health-care assessments and procedures (including personal and intimate care). Staff adapt their language to mitigate this, often becoming deferential (showing ‘low entitlement’) and offering choices that imply the possibility of refusal. This is also common in everyday English language and culture. In closing an interaction we routinely rely on the giving and registration of cues that the conversation is coming to an end, in order to try not to give offence, but these can be subtle.

In order to decipher the ‘message’ from among the social and cultural etiquette requires understanding, processing, perception and insight, processes with which a person living with dementia is likely to struggle. Person-centred dementia care philosophy holds that people living with dementia require an ‘enriched’ or ‘enhanced’ social environment, in which the health-care professional takes greater responsibility for making the relationship, even in the face of reluctance or resistance. Adapting communication to make it easier for the person living with dementia can be seen as a central part of this. The person living with dementia does not benefit from social etiquette if meaning is unclear, ambiguous, open to misinterpretation, invites refusal or results in a necessary medical or personal care task being neglected or argued over. The person living with dementia needs to feel satisfied with communication, avoiding, when possible, contradiction or argument. Reaching swift and unambiguous agreement is a virtue.

The risk is that language can become unduly coercive, or can fail to respect the identity or vulnerability of the person living with dementia. In many ways all language carries this risk; the lines between agreement, persuasion and deception are subtle. Rhetoric, marketing, propaganda and political messaging all deliberately attempt to persuade or change opinion. 209 CA has an overtly ethical dimension: findings can be used to promote good, or can be misused and result in harm. 143 The importance of professional, ethical and person-centred practice is undiminished by learning what language can be used to gain agreement or end an encounter.

We were aware of this potential problem, emphasised it in teaching and used an independent rating of ‘emotional tone’ of assessment videos before and after training (as a proxy for person-centeredness). This revealed some tension between effective communication and person-centredness, in that independent ratings suggested that communication was slightly more controlling, bossy and dominating after training. However, we do not believe that our findings reflect an incompatibility between person-centred care and effective communication. Several caveats are worth noting in this regard. How a conversation ‘sounds’ (the basis for rating emotional tone) is not necessarily a reflection of its person centredness. For example, a highly entitled direct request may promote inclusion and occupation, and does not necessarily diminish identity. Second, brief video clips offer little information about the context or necessity of a request, nor what occurred before or after. Thirdly, the inter-relater reliability of assessments was poor, suggesting that different people see different things in an interaction.

We undertook this CA-based study of dementia care communication precisely because most best practice guidelines in this area are not underpinned by objective research. Furthermore, it is unclear how person-centred care is operationalised in terms of communication behaviours during health encounters with a person living with dementia (or during health-care encounters generally). We have found evidence that some of the more common strategies for enacting person-centred care (like asking ‘is there anything else I can help you with?’) may be inappropriate for people living with dementia as a result of their communicative and cognitive difficulties.

We (and our trainees) concluded that our training changed knowledge, skills and behaviour and was useful to staff in diverse roles in everyday front-line clinical practice (and indeed may be useful for patients who are cognitively intact). Health-care professionals already have considerable knowledge and skills, technical and discipline specific, but also generic and interpersonal. Many experienced in working with people living with dementia are at least familiar with the ideas behind good communication and person-centred care. The key elements of our educational endeavour were the provision of new knowledge, a framework for understanding why communication can break down, the integration of prior skills and attributes, the opportunity to rehearse, practice and have feedback on communication behaviours, the opportunity reflect on communication encounters between the 2 days of the course and progression to more challenging simulation on the second day.

One feature that was commented on by trainees was the value of interdisciplinary learning. Different disciplines may not regularly observe how others communicate; watching peers and colleagues communicate in simulations proved as valuable as direct experiential learning. Another particular feature valued by trainees was the use of real-life video clips or transcripts illustrating learning points, both positive and negative. These carried especial validity (they were, after all, real) and were often memorable, but also illustrated real-life complexity, difficulties, failures and the sense of negotiation often required to gain agreement, which is difficult to encapsulate in writing.

A further effect of making explicit good communication practice is the engendering of ‘confidence in competence’. Griffiths et al. 25 identified that even when health-care professionals were doing their best, and delivering care well in difficult circumstances, they were often unsure, or frustrated that they were not doing well enough: they lacked ‘confidence in competence’. Health-care professionals’ knowledge that they are doing the right thing is important for job satisfaction and the avoidance of stress and burnout.

The VOICE course was mapped onto the Skills for Health – Dementia Core Skills Education and Training Framework Tier 3 (expert level) for communication, interaction and behaviour in dementia care and for person-centred dementia care. 35 This level defines the expectations of expert practitioners. There is little current provision in the UK for this level of training, and identifying such training is a current Health Education England priority (C Surr, Leeds Beckett University, 2018, personal communication). We have assessed the course using the Dementia Training Design and Delivery Audit Tool (DeTDAT), which assesses how well dementia training and education packages for hospital staff meet evidence-based good practice criteria. The VOICE course met all the requirements of this tool. 190,210

Implications

Hospitals and other care settings should make further ‘reasonable adjustments’ to ensure that staff are prepared to look after people living with dementia. 27 Many factors influence quality of care: the UK Care Quality Commission has characterised this as requiring leadership, attitudes, skills and resources. 12 Staff skills alone are not enough if staff numbers and time, the physical (and auditory) environment, processes and priorities represent barriers to dignified and person-centred care. Leadership requires both a commitment to training and the application of skills and knowledge in practice. Training can influence attitudes: showing that things can be done, and done well, helps to avoid a tendency to nihilism. Good communication helps to support identity, inclusion and occupation, which is more satisfying and defends against objectification and infantilisation.

Teaching adequate staff skills, however, remains central to the provision of good care. Care of people with dementia is complex and can be difficult. Poor communication results in missed therapeutic opportunities, mistakes, distress, denial of choice or autonomy and poor decisions. Distress, or unexpressed need, can result in difficult behaviours. Creating comforting relationships is the key to enhancing well-being and improving satisfaction. Unless senior staff understand and can model best practice, less experienced staff and students will not be adequately supported.

The main barriers to widespread implementation are expense and the need to train actors. The cost is modest in commercial training terms (full economic cost of about £300–350 per person, 2017 value), but health-care professionals often have little or no access to funds for training. Incorporation into undergraduate or postgraduate training structures (such as Foundation Programme or Higher Specialist Medical Training for doctors, or Learning Beyond Registration for other health-care professionals) would provide another avenue. The most likely niche will be as a fairly centralised resource for specialist practitioners. Given the importance of older people with dementia in hospitals, however, the numbers of people requiring the skills that our course teaches is very considerable.

In order to train actors in credible and effective simulation of people living with dementia we are preparing a manual and supporting materials (including video clips and the 22-minute documentary Today is Monday about people living with dementia and the staff who care for them in an acute hospital). 187

In addition, we have developed two brief electronic-learning multimedia packages (RLOs)175 that support the training but cannot replace the face-to-face content.

We are also exploring shorter packages, to minimise cost.

Further research can be done from within the corpus of video data collected for this study, and to explore related questions (including the enactment of person-centred care) and settings (such as care homes). The CA-based methodology is powerful and generates highly applicable practical output, but is labour-intensive and requires careful consideration of consent, data security and reuse of material. 211 However, the methodology provides great opportunities for further understanding communication in health care.

We undertook a feasibility or proof-of-concept study. In the face of known and acknowledged problems, it is likely a priori that teaching communication will be worthwhile for this patient group in this setting. A large RCT to demonstrate benefit in terms of patient-level health-care outcomes would be unfeasibly large and expensive, and would be unprecedented in the field of communication training. Our evaluation study may be considered sufficiently ‘positive’ to support implementation and roll-out without further large-scale evaluation. However, further research should be done to adapt or develop training for a wider body of staff, and to evaluate its effectiveness. This might include unregistered practitioners (such as health-care, therapy or nursing assistants), all registered staff who work with people living with dementia (who may be less skilled initially, more unwilling to learn and engage less well in training than the volunteers we studied) or staff with English as a second or additional language.

The development of communication training and its evaluation is also required for staff who work in care homes, domiciliary care staff and for family carers. The interplay between communication skills and person-centred care requires further exploration. A wider range of communication encounters (beyond health-care practitioner-initiated requests), and how they are managed, might be studied, although the practicalities of CA might make this difficult (e.g. gaining agreement in advance and setting up a camera).

The ultimate goal of staff training is to improve the quality or efficacy of care. Research methods to determine the impact of communication practices on patient outcomes, such as health status, well-being or distress, or health-care-related metrics, such as safety, discharge destination or length of stay, are poorly developed and require attention. Non-participant observation may be required. Similarly, from an organisational perspective, enablers of and barriers to implementation require investigation, including features such as ‘critical mass’ of trained staff, leadership and culture, and competing priorities, and how such conflicts or trade-offs are best managed.

Traditional methods of teaching communication skills for people living with dementia in hospital have been inadequate. We have drawn on multiple different pedagogic approaches to develop an innovative and effective training course, teaching evidence-based key practical knowledge.

Contributions of authors

Professor Rowan H Harwood contributed to the study conception, supervised pilot work, wrote the funding application, supervised ethics and governance application, designed the communication skills training intervention, delivered the communication skills training course, designed the evaluation, drafted the discussion chapter and edited the complete text, had general oversight and leadership, sat on the management group, approved the final report and undertook dissemination.

Mrs Rebecca O’Brien contributed to the study conception, undertook pilot work, wrote the funding application, undertook the systematic review, undertook the ethics application, recruited participants, collected and analysed the video data, designed the communication skills training intervention, designed the simulations and trained the simulators, delivered the communication skills training course, designed the evaluation (including the rating of simulation video data), drafted the interventions development chapter, drafted the CA chapter, sat on the management group, approved the final report and undertook dissemination.

Dr Sarah E Goldberg contributed to the study conception, supervised pilot work, wrote the funding application, undertook the systematic review, supervised the ethics and governance application, undertook the project management, designed the communication skills training intervention, designed the simulations and trained the simulators, delivered the communication skills training course, designed the evaluation (including the rating of simulation video data), analysed training evaluation findings, drafted the evaluation chapter and PPI chapter, edited the final report, chaired the management group, approved the final report and undertook dissemination.

Mrs Rebecca Allwood contributed to the study conception, undertook pilot work, wrote the funding application, undertook the ethics application, recruited participants, collected and analysed the video data, designed the communication skills training intervention, drafted the CA chapter, sat on the management group, approved the final report and undertook dissemination.

Professor Alison Pilnick supervised the CA, advised on the use of CA findings in professional training, designed the communication skills training intervention, drafted the CA chapter, sat on the management group, approved the final report, undertook dissemination and led the linked Economic and Social Research Council impact accelerator award.

Dr Suzanne Beeke wrote the funding application, supervised the CA, advised on use of CA findings in professional training, designed the communication skills training intervention, designed the evaluation (including the rating of simulation video data), drafted the CA chapter, sat on the management group, approved the final report and undertook dissemination.

Dr Louise Thomson wrote the funding application, designed the evaluation, undertook and analysed the interviews, drafted the evaluation chapter, sat on the management group and approved the final report.

Ms Megan Murray designed the communication skills training intervention, designed the simulations and trained the simulators, delivered the communication skills training course, drafted the chapter on training simulators, sat on the management group and approved the final report.

Dr Ruth Parry wrote the funding application, advised on video-recording methods and ethics, advised on the use of CA findings in professional training, sat on the management group and approved the final report.

Dr Fiona Kearney drafted the introduction, checked and edited the references, checked and edited the complete text and approved the final report.

Professor Bryn Baxendale contributed to the study design, designed the communication skills training intervention, facilitated the delivery of training courses, advised on interpretation and dissemination, sat on the management group, approved the final report and undertook dissemination.

Mrs Kate Sartain advised on the study design, video-recording methods and ethics, designed the communication skills training intervention, delivered the communication skills training course, advised in interpretation, was a PPI representative on the management group, approved the final report and undertook dissemination.

Professor Justine Schneider designed the communication skills training intervention, edited references and the final report, compiled the references, sat on the management group, approved the final report and undertook dissemination.

Publications

Allwood R, Pilnick A, O’Brien R, Goldberg S, Harwood RH, Beeke S. Should I stay or should I go? How healthcare professionals close encounters with people with dementia in the acute hospital setting. Soc Sci Med 2017;191:212–25.

O’Brien R, Goldberg SE, Pilnick A, Beeke A, Schneider J, Sartain K, et al. The VOICE study – a before and after study of a dementia communication skills training course. PLOS ONE 2018;13:e0198567.

Data-sharing statement

Owing to the nature of the data collected and conditions attached to ethics approval, there are limited data available for wider use. All queries should be submitted to the corresponding author in the first instance.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health and Social Care.

Work package 3 intervention testing: interview guide