An intervention to promote self-management, independence and self-efficacy in people with early-stage dementia: the Journeying through Dementia RCT

Rationale, theory or goal of the elements essential to the intervention

Journeying through Dementia is a manualised psychosocial intervention for people with early-stage dementia. It was designed through consultation with people living with dementia and was intended to equip individuals with the necessary knowledge, skills and understanding to successfully self-manage their dementia symptoms and live well with dementia. 32 It was created in response to earlier diagnosis of dementia by memory services and the consequent requirement for post-diagnostic interventions. A key distinguishing feature of this intervention is that people with dementia do not necessarily need to involve a supporting carer to take part, thereby addressing the potential needs of the estimated one-third of people with this diagnosis who live alone in the community. 33

The intervention comprised the following two elements: (1) self-management and (2) occupational therapy principles.

Underpinning theoretical framework

The JtD intervention is underpinned by social cognitive theory,38 thereby aiming to increase self-efficacy and effective problem-solving (using the knowledge and techniques described above), which in turns leads to positive emotions, ability to self-manage and life satisfaction.

The content of intervention and processes associated with being in a group target a number of key domains that have been shown to be associated with and/or predict quality of life in dementia. 39–42

1. Mood: the intervention leads to improvements in mood either directly through behavioural activation and learning new coping strategies or indirectly by improving sense of self-efficacy.

2. Relationships: the intervention helps the individual to build relationships and enhances their sense of being connected to others and the community.

3. Beliefs: the intervention helps to promote the understanding that life is meaningful despite dementia.

4. Functional skills: the intervention helps the individual retain skills in activities of daily living and cognitive functioning for as long as possible.

The intervention delivery was also informed by Mitchie et al. ’s43 theory of behaviour change, which emphasises the importance of capability, motivation and opportunity in driving change. Within the theoretical framework of the JtD intervention, behaviour change is associated with improved self-management and engagement in meaningful activity. Capability is addressed by imparting knowledge and training in emotional, cognitive and behavioural skills. Motivation is addressed by increasing knowledge and understanding, eliciting feelings about change and experiential learning to demonstrate positive feelings associated with behaviour change. Opportunity is addressed by making both physical and social environmental changes to enable participants to not only experience the behaviour change in situ (i.e. within the groups and one-to-one sessions), but also by incorporating change into their lives.

The Journeying through Dementia intervention

The manualised intervention contains the following menu of topics that map onto the elements described above. Included topics were identified through consultation with people with dementia prior to the study:32

Reproduced with permission from Craig44

The goals of the intervention are achieved in several ways. Participants may adapt the way an activity is performed, make adjustments to the wider environment or use compensatory techniques. The programme supports individuals to recognise, build on and use existing skills, and to develop new interests that might include drawing upon wider resources.

The intervention content and modes of delivery were tested in a feasibility study25 and found to be acceptable to both people with dementia and their supporters/carers. Reported benefits in people with dementia included increased confidence and self-efficacy, participation in activities and re-engagement in fun and friendships.

Procedures

Intervention tailoring

Facilitators were encouraged to work with other topics that groups and individuals identified, as well as those in the intervention manual. However, they were asked to deliver the intervention as described in the manual and training.

Out-of-venue activities were tailored to the expressed needs of the group and the individuals within it, and they could be simple or more elaborate, depending on the group. For example, moving to a public cafe area within the same building so that participants support each other with the activities they find challenging. Within the group, the aims of individuals will differ. For one person it could be ordering a drink, for another it might be handling money and for a third it might be socialising in a public area. Another group could be enabled to practise their individual skills and nurture their interests in a more demanding out-of-venue activity.

The one-to-one sessions were in place to enable participants to work towards meeting individual goals, with or without a supporter/carer depending on wants, needs and circumstances. Goals could be very small and their achievement incorporated some degree of enactment. On occasion, one-to-one sessions were used to discuss issues of concern to individuals to help identify ways of managing the challenge (e.g. using public transport or continuing to drive). As with the group sessions, the overall aim was to enable participants to articulate how they wanted to use these sessions and be supported to identify and achieve realistic goals. Facilitators were advised that the first one-to-one session should be organised before commencement of the group so that they could introduce themselves to the participant and discuss the forthcoming 12 weeks, including how the participant would access the group venue (transport was not provided). We were also aware (from our previous Lifestyle Matters research45) that this initial face-to-face meeting was important for establishing one-to-one sessions in this form of mixed intervention.

Facilitator/supervisor training

People living with dementia played an integral role in the development of the training for the intervention, as well as in intervention development. Initial consultation led to insights that informed the training. This was then informed by consultation with the Scottish Dementia Working Group (a national campaigning group comprising people with dementia), which assisted with the design of the materials for the intervention and added to our understanding of what would make good facilitation.

Facilitator training was designed to mirror the modes of intervention delivery, in that it was group based (i.e. involving facilitators from more than one site at an independent venue), it sought to impart necessary didactic information to facilitators to achieve a shared understanding and it included experiential work, using exercises taken from the manualised intervention. For example, two trainers modelled group facilitation techniques and trainee facilitators were given the opportunity to rehearse and role play responses to potential scenarios that might arise in the groups they were to facilitate. Training took place as near as possible to commencement of intervention delivery.

Taught sessions within the training (didactic information) included the following:

• the background, context and development of the JtD intervention

• how the intervention was underpinned by evidence regarding needs for, benefits of and nature of post-diagnostic support for people affected by dementia

• the underpinning philosophy and approach of the intervention

• information on the models and modes of group work

• information on the role of individual sessions

• information on environment and resource management

• information on methods of record-keeping and reflection through use of reflective diaries.

Discussion topics included the following:

• prior experience, interest and motivations

• understanding the experiences of people living with dementia (with responses to quotes by people with dementia and a film made by people living with dementia)

• a walk through of the manual (identifying the components)

• details on the preferred approaches to group facilitation

• information on how the environment can support and inhibit interaction and self-efficacy

• information on managing groups (including a discussion based around a series of personas).

Experiential sessions included the following:

• what makes a good post-diagnostic support programme (using an interactive activity)

• community resources (including a pictorial and mapping activity)

• guidance on the planning, and facilitation of sessions (with trainees acting as group participants)

• guidance on managing groups and delivering individual sessions (using role play)

• information on methods of record-keeping and completion of other intervention-related documentation for the purposes of delivery and for research.

Facilitators were asked to document all group and individual sessions on a pro forma provided by the research team. They were asked to write records primarily for the benefit of participants and in accordance with Dementia Engagement and Empowerment Project (DEEP) guidance for best practice in providing documentation for people with dementia. 46

Facilitators were asked to write session records as soon after the group as possible and post to all participants before the next session.

Facilitators were also asked to maintain registers of participant attendance and records of the supervision sessions they attended on other pro formas provided to them by the trial manager.

Facilitators were asked to post copies of all these records to the trial manager. These records were requested at the end of the 12 weeks for the first rounds of intervention delivery at all participating sites. For subsequent rounds, intervention delivery facilitators were asked to return records every 3 weeks. The reason for this change was to improve delivery of the intervention in accordance with the manualised intervention on an ongoing basis.

Facilitator supervision (for the purposes of this study)

Once intervention delivery commenced for the trial, facilitators all received weekly supervision from a suitably knowledgeable and experienced clinical professional (AfC band 7 and above) from their place of work, identified and nominated by the research site.

Ideally, people supervising would have themselves previously facilitated the intervention. For the purposes of this study, however, this was not possible, and so a ‘train the trainers’ model was adopted. This involved suitably experienced individuals being nominated as supervisors. Members of the trial team, who were experienced in delivering and supervising psychosocial interventions within trials, then delivered regular supervision to these supervisors.

The supervision protocol identified for the purposes of the study is outlined below (see Report Supplementary Material 1 for the full supervision protocol):

• Suitably qualified professionals who agreed to supervise facilitators at site were invited to take part in the first day of the 2-day facilitator training, ideally with the facilitators they were to supervise, as well as with other supervisors and facilitators from different sites.

• A second half-day session with other supervisors and the trial team member nominated to supervise them was then convened to discuss perceptions and any concerns regarding both intervention delivery and the associated research, as well as to go through practical details, including completion of necessary trial documentation for supervision (see Report Supplementary Material 2 for the supervisor booklet). This second session wase scheduled as close as possible to the initial training.

• On commencement of intervention delivery, each facilitator then received weekly supervision in a format agreed between themselves and the person nominated at site to supervise them [e.g. face to face, Skype™ (Microsoft Corporation, Redmond, WA, USA), FaceTime (Apple Inc., Cupertino, CA, USA) or telephone]. These supervisions included both individual and group supervisions.

• Facilitators and those supervising them maintained records of the nature and duration of each supervision record, using the trial pro formas.

Modifications to facilitator identification and training, intervention delivery and supervision

Slower rates of recruitment at sites than those originally envisaged led to the need for a greater number of delivery sites (n = 13) and, hence, a greater number of facilitators and supervisors. Facilitator and supervisor attrition and absence also led to increased numbers being trained. Overall, we trained 69 facilitators (not all of whom delivered sessions) and 21 supervisors.

Important changes to methods after trial commencement

The most significant study changes after trial commencement are listed below. Study protocol amendments are listed in full in Appendix 2.

Participants and eligibility criteria

Participants in the trial were individuals with a formal diagnosis of dementia and their supporters where these consented to be involved. Supporters included family members, friends and neighbours named as supporters by the person with dementia. Where supporters were trial participants they were referred to as ‘participating’ supporters. Participants and participating supporters were selected based on the eligibility criteria outlined below.

Participant identification and recruitment

Recruiting people living with dementia to clinical trials is known to be challenging and, therefore, a variety of participant identification methods were used to maximise recruitment. The recruitment strategies used varied between sites because of the different configurations of dementia care within NHS trusts (see Report Supplementary Material 4 for the overall participant identification and recruitment process within the JtD intervention).

Regardless of the method of identification, care was taken to cater to the needs of the individual, given the potential for cognitive and communication difficulties. Although there was a standardised identification and recruitment process, it was important that it allowed some degree of flexibility to cater to the specific needs potential participants may have had. Alongside the participant information sheet, a shorter information leaflet, which had been reviewed by the PPI Advisory Group, was used to provide a simplified overview of the trial. This was used before going into more detail with the full-length participant information sheet if the individual was still interested after receiving the initial information. In the first instance, any information about the trial was communicated to the person living with dementia; however, if the person with dementia requested, researchers could communicate with someone they identified as a participating supporter.

Anyone who expressed interest received a face-to-face visit from a researcher who had been offered training on the additional communication needs of people living with dementia.

Researchers were provided with suggestions for recruitment methods, but implementation of these methods was at the discretion of the researchers because of their knowledge of the individual NHS trust. Recruitment was continually kept under review and support was provided from the central team if required. The different methods used are described below.

Baseline and follow-up visits

Outcomes

Sample size

The sample size was calculated assuming a standard deviation (SD) of 11 points and that a difference of ≥ 4 points in mean DEMQOL score 8 months post randomisation was clinically and practically important. 51 The sample size was calculated to have 90% power for detecting this 4-point difference (or more) if it truly exists, which is equivalent to a standardised effect size of 0.36, as statistically significant at the 5% two-sided level. As the JtD intervention was a facilitator-led group intervention, the outcomes of the participants in the same group with the same facilitators may have been clustered. For an individual RCT without adjustment for clustering, the target sample size would be 160 participants per arm (i.e. 320 participants in total). Assuming an average cluster size of eight participants per facilitated group45 and an intracluster correlation of 0.03 this inflated the sample size, by a design effect of 1.21, to 194 participants per group (i.e. 388 participants in total) with valid primary outcome data. We assumed a 20% loss to follow-up and therefore the trial target sample size was to randomise to 243 participants in each arm (i.e. 486 participants in total).

There is no reported or established minimum clinically important difference (MCID) for the DEMQOL measure. Data from the development and validation of the preference/utility-based DEMQOL-Utility (DEMQOL-U) score,51 which compared the performance of the DEMQOL (and DEMQOL-U) with other patient-reported outcome measures for people with dementia (e.g. the MMSE,71 the Bristol Activity of Daily Living Scale72 and the Neuropsychiatric Inventory)73 with established MCIDs, reported changes on the DEMQOL of 5.4, 7.5 and 6.4 points, respectively, for patients who improved by more than the MCID on the anchor measure. This suggests that our proposed 4-point difference, although small, is likely to be of clinical and practical importance.

Internal pilot trial

The following stop/go criteria were reviewed after 8 months of active recruitment to an internal pilot trial:

• recruitment of a minimum of 113 participants across the six pilot sites by the end of the fifth month of active recruitment (i.e. 75% of the 150 participant target)

• recruitment of a minimum of 12 facilitators (i.e. two facilitators identified at each of the six pilot sites by the start of active recruitment to deliver the intervention)

• no more than two of the six planned groups in the internal pilot with fewer than four participants registered for the group by the sixth month of active recruitment.

The TSC assessed whether or not they could recommend that the trial continue in the light of the feasibility results against the stop/go criteria. The results and recommendations were communicated to the funder [see Chapter 3, Internal pilot (stop/go assessment)].

Method used to generate and assign the random allocation

Randomisation was overseen by the Sheffield Clinical Trials Research Unit (CTRU), using a computerised randomisation sequence generated before the start of recruitment. An unblinded member of the research team who did not conduct any outcome assessments entered the participants’ details into the remote web-based randomisation system. Participants were then randomly allocated to receive either the intervention or usual care. Owing to the group-based intervention, participants were randomised after the baseline outcome measures had been collected, which ideally occurred < 2 months before the intervention wave began at each site. In instances when a couple living in the same household both consented, they were randomised as a couple rather than individually. Participating supporters were allocated to the same group as the linked participant.

The participants and, if relevant, their participating supporters were informed of their allocation by letter. Participants allocated to the intervention were told that an intervention facilitator would contact them to discuss the next steps of the trial.

Following allocation of a participant to the intervention group, an unblinded member of the research team informed the appropriate intervention facilitators so that they could contact participants to arrange intervention delivery.

Type of randomisation and details of any restriction

Participants were randomised 1 : 1 to receive either the JtD intervention or usual care. The computer-generated randomisation was stratified by a stratification site and restricted by a fixed block of size 4. The stratification site was either the recruiting site or the group delivery site if the recruiting site was delivering more than one group simultaneously in different locations. The JtD intervention had 16 stratification sites across 13 recruiting sites. A fixed block size was used to ensure that the imbalance between treatment groups at each delivery site was controlled and minimal. Two intervention groups were planned at each site and a maximum number of 13 participants were allowed in each intervention group. The risk of predicting the allocation was deemed minimal, as it was performed by central unblinded members of the trial team and members of the trial team responsible for recruiting participants at site were blinded to treatment allocation.

Blinding and methods used to control bias

Blinding in the context of research studies refers to the concealment from individuals involved in a clinical trial of the allocation to an intervention or control group of participants in the trial. It is used to minimise or prevent differential treatment or outcome assessment of groups in the trial and reduce bias. 74 The trial was single blind as participants were aware of their allocation to either the JtD intervention group or the usual-care control group. Outcome assessments were completed by assessors who were blind to group allocation. The sections below on allocation concealment, blinding and other methods provide further information on the study’s approach to reducing bias.

Analysis populations

The ITT data set includes all participants who consented and were randomised according to randomised treatment assignments (ignoring any occurrences post randomisation, such as protocol or treatment non-compliance and withdrawals) and who had complete DEMQOL data at 8 months post randomisation. This excludes participants who withdrew before randomisation and includes participants found to be ineligible post randomisation. 75

The complier-average causal effect (CACE) data set is a subset of the ITT set. It included a subgroup of participants who were believed to comply with the JtD intervention (i.e. attending at least 10 of 16 sessions), excluded ineligible participants randomised in error and included participants who were randomised to usual care but received and complied with the JtD intervention.

The analysis population for participating supporters included participating supporters who met eligibility criteria, provided informed consent and provided follow-up data.

Statistical analysis

Study oversight and management

Ethics arrangements and regulatory approvals

Data collection and management

The follow-up of participants and data collection was co-ordinated by the CTRU in conjunction with the participating NHS sites and collaborating universities. Data were collected by researchers employed by The University of Sheffield, University of Hull, University of Bradford and University of Nottingham, as well as authorised individuals at participating NHS sites. Information, such as outcome measures, was gathered using the JtD trial CRF, an ad hoc record with both paper and electronic components designed specifically for this trial. The CRFs were made up of paper and electronic parts. During the trial, paper documents were retained in secure locations.

The CTRU’s in-house web-based data management system (Prospect; University of Sheffield, Sheffield, UK, and epiGenesys, Sheffield, UK) was used to capture and store participant data. Data were inputted by trained and authorised members of the trial team whose access was controlled by individual usernames and encrypted passwords. A privilege management feature was used to ensure that users of Prospect had an appropriate level of access to data to allow them to complete their tasks. Participant confidentiality was respected at all times. Access to names and contact details entered into Prospect was restricted to only users with certain privileges. All other data were pseudonymised, identifiable only through a participant identification number.

Validation features of Prospect were used to monitor study data quality. When clarification of data were needed, error reports were generated detailing queries.

Usual care

Information regarding usual care available to participants at intervention sites was recorded with the use of a usual-care template, the format of which was agreed by the study’s health economist. The template was sent to principal investigators at each site, who either completed templates themselves or nominated a person with appropriate knowledge for the task. When a site ran interventions in two or more areas, when usual care differed, usual-care information was requested for each of the intervention areas.

The templates asked for details about care and the care pathway that was usually offered to people who had recently been diagnosed as living with dementia, including where any clinic or group attendance was offered after diagnosis. The template also asked for the method and frequency of review post diagnosis and what action would be taken if a person’s condition worsened. To view analysis of the usual care at intervention sites please see Chapter 3, Usual care.

Additionally, the 8-month HSCRU sections on day service use and community-based service use were analysed and are summarised in Chapter 3, Usual care to provide an insight into the types of services routinely being accessed by people living with dementia over the month prior to being asked.

Rationale

The need to improve and expand on the evidence base for complex interventions has been recognised by health-care providers, commissioners, researchers and the public. 84 Complex interventions present a number of methodological challenges to evaluation, including standardisation of multicomponent interventions, contextual impact and complexity of associating change with the intervention. 85 Trial designs that include mixed methods have been widely recommended to capture such changes and explain the reasons for intervention efficacy or lack of efficacy. 85,86

Researchers need to be able to demonstrate with confidence that their study results are due to the impact of the intervention under investigation and not other extraneous factors. Fidelity assessment uses multiple methods to evaluate validity and reliability of complex interventions. 87,88 For this study, we employed observations and facilitator reports to explore training, delivery and receipt of the group aspect of the intervention.

Aims

The main aim of the fidelity assessment was to assess how well the intervention was delivered in accordance with the protocol and intervention manual, as outlined in the assessment framework presented in Table 4. The framework was designed specifically for this intervention and is based on criteria identified by the Behaviour Change Consortium87 and NICE guidance on behaviour change,89 providing quality assurance parameters for intervention design, training, delivery, receipt and enactment.

Research team and reflexivity

Throughout the fidelity study, two researchers were engaged in data collection and analysis (i.e. the fidelity lead and a second researcher). The fidelity lead remained consistent throughout the duration of the substudy, whereas the second researcher role was undertaken by four successive researchers.

Sampling

A convenience sample of four of the total number of sites participating in the study were invited and agreed to take part in the fidelity assessment and all four sites agreed to participate. A pragmatic approach was taken towards site selection (including local research governance approvals being in place, sites being open and running the intervention and sites having the capacity to participate in the fidelity work). Consideration was also given to geographical location and population size to represent services across the totality of field work sites.

Methods

The fidelity lead and members of the research team developed bespoke fidelity assessment checklists and intervention delivery registers to collect data. The TMG and authors of the manualised intervention reviewed draft versions of several iterations before reaching agreement on the final content (see Report Supplementary Material 5–12 for copies of all registers and checklists).

Analysis

Rationale

Qualitative research can provide a richer understanding of people’s experience, thoughts and motivations in a real-world context. An embedded qualitative substudy was conducted as an adjunct to the main trial to explore insights from those involved in delivering and receiving the intervention to evaluate and inform future development of the intervention.

Aims

The aim of the qualitative study was to enable exploration of intervention facilitation, delivery and receipt, specifically how the intervention produces change in a particular context. 91

Sample

Interviews were conducted with a purposive sample of participants and their participating supporters (when one was present). The participant sampling frame was based on a range of characteristics, including diagnosis, age, ethnicity, gender, MMSE score, living arrangements and extent of participation (including participants who had withdrawn from the intervention) (see Table 32 for further details on the participant sample).

Methods

A critical realist epistemological position was taken, which assumes that from a social science perspective we are able to identify and describe real properties and causal relationships in the world around us. 92 However, it also assumes that not all knowledge is equal when trying to understand the real world. 93 We therefore assumed a level of truth about the effectiveness of the intervention based on participant, participating supporter, facilitator and supervisor perspectives, thoughts and experiences of delivery and receipt of that intervention, and by collecting a range of views that could be deemed an approximation of a wider understanding.

Analysis

The same methods of analysis were applied to all interview transcripts. This involved the trial qualitative lead and a second researcher throughout. The trial qualitative lead and a second researcher first read through all the transcripts to embed themselves in the data. A purposive sample of transcripts were then selected for initial thematic open coding. Selection of the facilitator, participant and participating supporter transcripts for this first open coding included representation from each site and a cross-section of viewpoints. As there were only four supervisors, all transcripts were included in the open coding process.

These transcripts were coded independently by the two researchers before a joint review to establish an initial framework of overarching themes and subcategories. Positive, negative and ambivalent cases were considered and the use of two coders ensured that coding did not bias any one point of view.

Framework analysis,94 a theoretical form of thematic analysis, was used to analyse the full data set. 95 NVivo software version 12 (QSR International, Warrington, UK) was used to assist with analysis, which involved the confirmation of the final thematic framework and the development of an index of subcategories that were then applied to all transcripts. Further iteration of the framework was agreed during the coding process for any new subcategories arising from the full data set. All interview data were then charted using the framework, and the resulting data map was used to identify explanations and processes underlying the intervention. Results were also used to identify emergent factors that could influence the uptake and impact of the intervention.

The participant and participating supporter interviews, and the staff interviews were analysed separately, therefore producing two separate frameworks (one for participants and participating supporters, and one for facilitators and supervisors). Finally, the two frameworks were triangulated to identify common themes and divergent cases.

For the purposes of reporting, confidentiality was maintained by using unique participant identifiers and removing identifiable information.

Co-researcher qualitative analysis validation workshop

People living with dementia and their supporters took part in two co-researcher data validation workshops held in February and July 2019 to support the analysis of the participant and supporter interview data. Some of those who were members of the JtD trial PPI Advisory Group took part in the workshops, and others were recruited for these sessions from the University of Bradford Experts by Experience cohort. 96 The first workshop was attended by six co-researchers (two couples, one person living with dementia and one supporter/carer) and the second workshop was attended by 12 co-researchers (five couples, one person living with dementia and one supporter/carer). Both workshops were delivered in a community venue and co-researchers were recompensed for their time and expenses. The purpose of the workshops was to identify similarities and differences between researcher interpretations of the data and the interpretations of people with lived experience of dementia. A total of 20 quotations (nine quotations in workshop 1 and 11 quotations in workshop 2) were selected from the participant and supporter interview data as being examples from the range of themes and subcategories in the framework analysis. These were presented to the workshop attendees, who were then asked to review and respond to the quotations, with particular focus on language, tone and overall meaning of content. The activity was supported by two facilitators who were experienced in delivering workshops and working with people with dementia. Written notes were taken during the workshop and the sessions were also audio-recorded to ensure accuracy of interpretations. The interpretations of people living with dementia were compared with those of the researchers and used to refine the final qualitative analysis.

Rationale

Triangulation enables the comparison of concurrently collected data obtained using different methodologies and from different researchers to be explored for interaction, thereby adding validity to research findings. 97

Methods

Findings from the qualitative study were explored for inter-method convergence, discrepancy or complementary information. 97 The below methods of triangulation were employed.

1. Methodological: results from the two data collection methods were compared (i.e. interviews and checklists).

2. Data source: a range of perspectives were represented, including participants, participating supporters, facilitators and supervisors.

3. Investigator: researchers and a PPI Advisory Group recruited from the University of Bradford Experts by Experience cohort were involved in the interpretation of the interview data.

Analysis

The triangulation protocol for analysis (Table 5) was based on that described by Farmer et al. 98

Participant baseline characteristics

The characteristics of the randomised participants are given in Tables 7 and 8, and are well balanced between the randomised groups. Thirteen sites were recruited to the study and contributed between 12 and 66 participants per site. There were slightly more male participants (58%) and participants were generally white British (95%). The average age of participants was 77 years and 26% of the participants lived alone. Alzheimer’s contributed 60% of the dementia diagnoses and the median length of time since dementia diagnosis was 0.7 years. The baseline MMSE and patient-reported outcome measures are given in Table 9. The participant-reported outcomes were well matched across the randomised groups; the majority of participants (80%) rated their quality of life overall as either good or very good on the DEMQOL measure. See Appendix 8, Tables 37 and 38 for categorical and continuous baseline characteristics by intervention group and missing data status.

Participating supporter baseline characteristics

Table 10 gives the baseline characteristics for the participating supporters. The majority of participating supporters were female (74%) and were the spouse/partner of the participant (79%).

Relationship between group attendance and primary outcome

The relationship between number of JtD sessions attended and DEMQOL change in score from baseline at 8 months for the intervention group is shown in Figure 7. There is considerable variation in DEMQOL change in score across each level of session attendance and there is no clear association between total session attendance and 8-month DEMQOL change in score.

FIGURE 7.

Change from baseline in 8-month DEMQOL score by number of JtD intervention group sessions attended (n = 191).

Participant secondary outcomes

Comparisons of secondary participant-reported outcomes at 8 months are presented in Table 16. Overall, the treatment differences were in favour of the intervention group (excluding GAD-7), but the differences were small (i.e. unlikely to be clinically significant) and only the difference in DFS was statistically significant. There were no differences in secondary outcomes measured at 12 months (Table 17).

Participating supporter secondary outcomes

The comparison of 8-month outcome measures for the participating supporters is presented in Table 18. Mean differences were small and not statistically significant, with the exception of EQ-5D-5L, for which the participating supporters in the control group reported better health (adjusted mean difference –0.06, 95% CI –0.09 to –0.0.2; p = 0.002). This difference was not borne out in the EuroQol-5 Dimensions (EQ-5D) VAS, which showed no difference between the intervention and control groups (mean difference 0.1, 95% CI –2.8 to 3.1).

Training fidelity

As per protocol, a number of formal 2-day face-to-face intervention training sessions were conducted during 2017, prior to sites opening for recruitment. Details of the training are provided in Chapter 1, Intervention: rationale, methods and delivery. The author of the manual led all of the sessions, accompanied by a second trainer with experience working with people with dementia and delivering similar group interventions. Three of these sessions were included in the fidelity assessment (Table 20). Fifty trainees attended. On average, the time from training to delivering the first group was 106 days (range 79–134 days).

Two researchers were present during each intervention training session, with coding shared between three members of the research team, one of whom was present at all three intervention training sessions (i.e. the fidelity lead). Fidelity scored by the researchers on the checklist items was high overall across the three sessions, averaging 95% achievement (range 91–97%) of checklist items. Attendees’ scores supported this, averaging 94% achievement (range 88–97%).

The second training session (on the 23 and 24 January 2017) scored the lowest fidelity. Ninety-one per cent compliance with checklist items was scored by the researchers and 88% compliance was scored by the trainees, which may be explained by the changes made to day 2 of the training. The session had to be significantly modified during delivery because of the large numbers of attendees. On day 1 of the training, supervisors were asked not to attend day 2 to ensure that the facilitators had enough time and opportunity to practise elements of the intervention with their peers. Modifications included reducing time spent on a topic/activity or excluding it altogether. Items identified by the researchers as affected included:

• how the participant and supporter/carer work together on the intervention

• transferring skills from the safety of the group to their wider lives

• reflecting on and sharing their own facilitation style and skills

• exploring the value of out-of-venue activities and how to help group members to plan them

• discussing the value and principles of supervision.

The same pattern was reflected in the trainee checklist scores, with high agreement on items scored as delivered or not delivered overall between trainees, averaging 93% (range 88–96%). There was 100% response rate for trainee checklists.

Eighteen out of 50 trainees felt that at least one item of the trainee checklist (range 1–8 items) had not been delivered at the training. There was agreement between the researchers and trainees on the top three items not delivered. These were:

• Item 7: ‘did the trainer discuss the supporter attended sessions and relationship dynamics?’ (8 out of 18 trainees).

• Item 11: ‘were you able to reflect on and share your own facilitation style and skills?’ (8 out of 18 trainees).

• Item 17: ‘did you discuss the value and principles of supervision?’ (12 out of 18 trainees). This rating related predominantly to the second training session (on 23 and 24 January 2017) and was one of the topics affected by the modifications made to day 2 of the training.

Intervention fidelity

Overall, attendance was good for those who completed the 12-week intervention, with 25 out of 35 (71%) participants from the four sites used to assess fidelity attending 10 out of 16 sessions. Five (14%) participants attended all 16 sessions. This was representative of attendance in the trial (see Journeying through Dementia intervention).

Group meeting fidelity

Of the total 331 available group sessions (i.e. the number of group meetings × number of engaged participants in the four fidelity sites, note that sessions were counted if the participant missed them, but if the participant had withdrawn, then future sessions were not counted), 264 (80%) were attended. Fidelity to the intervention observed during group meetings was very high and achieved a score of between 85% and 95%.

For all four groups combined, a Cohen’s kappa score of 0.68 demonstrated substantial inter-rater reliability (95% CI 0.58 to 0.78) between researchers coding the group meetings. A site breakdown of scores can be found in Appendix 6, Table 36.

One hundred per cent of facilitator group meeting fidelity checklists were completed. Overall, facilitators reported that they had delivered the intervention as intended during the observed group meetings and fidelity averaged 93% (range 84–100%). This reflects the observations of the researchers.

One-to-one session fidelity

Three out of the four fidelity sites returned one-to-one session checklists. Overall, 20 out of a possible 35 checklists were returned. Each participant initially agreed to attend the intervention and completed the first one-to-one session. Eight participants withdrew between session 1 and session 2. Of the remaining 26 participants, only one participant missed a one-to-one session.

Three of the four fidelity sites returned one-to-one session checklists for 20 participants. Each participant had a maximum of four records, one per session. Seven participants had incomplete records and six of these participants were from the same site.

Facilitators recorded that they were able to deliver most items on the checklist during each session, averaging 77% achievement (range 22–100%). The items with the highest rates of non-delivery were item 5 ‘did you help the participant set any goals?’, with 40% achievement, and item 8 ‘did you enable the participant to rehearse skills learned in their everyday life?’, with 22% achievement. This differs from overall adherence to the intervention by all sites, as demonstrated through scrutiny of facilitator records, which revealed that they were not always able to deliver the intervention as intended. One rationale for this disparity may be facilitator awareness of being observed for fidelity assessment.

Supervision fidelity

Usual care at delivery sites

All participants in the study, whether allocated to the intervention or control arm, received treatment as usual from their service providers at the intervention delivery sites (which were all NHS trusts). The nature of treatment as usual varied across intervention delivery sites. There was also variation in the provision of treatment as usual for participants within some delivery sites because of the size of some NHS trusts and recruitment having a large geographical spread. These variations were the same for both the intervention and control arms.

Treatment as usual across delivery sites for people with dementia post diagnosis included:

• prescription of medication for dementia (i.e. cholinesterase inhibitors and/or memantine) or other medication to control symptoms (e.g. depression)

• assessment of needs and, in some instances, referral to various individual and/or group-based sessions, and/or other health and social care services (e.g. Community Mental Health Teams)

• referral to third-sector services or voluntary services (e.g. the Alzheimer’s Society)

• provision of educational material.

Analysis of usual-care pathways within participating sites revealed variance in the approaches taken following diagnosis. Some sites retained patients on their lists, while others referred to alternative services. The economy of care within the trusts was very different, for example, the availability of local services and service providers, and the role of third-sector organisations in providing care and support services.

Immediately post diagnosis, all trusts reported that they conducted some form of assessment of needs as part of usual care, followed by referral to support services (e.g. Community Mental Health Teams or third-sector support). Participating trusts provided limited follow-on support to people with dementia and their supporters. In almost all instances (with the exception of one site) patients were referred back to primary care if there were no complicating factors, or when they completed support groups or therapy provided by memory clinics. Two sites described using a ‘shared care protocol’ with primary care. Referrals back to the memory clinic were made by GPs, a member of a Community Mental Health Team or, if present, a memory support worker when a person with dementia’s condition had deteriorated.

The provision of group-based interventions as usual care was common across the 20 trial sites (Table 24). CST was the most commonly provided group-based intervention (provided at nine sites). CST groups were provided for 7–14 weeks with either one or two sessions per week. CST groups were provided as post-diagnosis support, with one site reporting that this was typically in the first 6–12 months post diagnosis. Several sites reported that CST groups were not available to all people diagnosed with dementia because of a high demand for the service. CST differs from the JtD programme, as it follows a detailed and prescriptive session-by-session plan of group exercises and activities that are facilitator led and takes place wholly within the delivery setting.

Memory or Living Well with Dementia groups were run at five sites (two and three sites, respectively) and a psychosocial intervention was provided by a third-sector organisation (the Alzheimer’s Society) at a further site through referrals from memory clinics. In one of the five sites, the group was not run at the same time as JtD groups, as the usual facilitators were delivering the JTD intervention. The groups provided by sites varied in length from four sessions of 90 minutes to 10 sessions of 2 hours. All were shorter than the JtD programme. Educational groups were additionally provided at three sites. Memory groups may focus on reminiscence, which is not a central theme of the JtD programme. The Living Well with Dementia groups may have had some of the features of the JtD programme; however, they do not generally include enactment of learnt skills in the community or the mix of individual and group sessions that the JtD programme incorporates.

The post-diagnostic review approach varied greatly at sites (Table 25). GPs were most often tasked with the provision of post-diagnostic review of patients (seen at eight sites). Usual post-diagnostic care at seven sites was to increase the frequency of patient reviews, depending on a patient’s need and/or deterioration. Six sites carried out patient reviews at least once in the first 3 months post diagnosis, with four sites conducting annual reviews of patients as part of their provision of usual care. Only two sites provided a path by which patients could refer themselves back to their local memory clinic.

As usual care, appropriate medication was provided to patients depending on the type of dementia, progression, and behavioural and psychological symptoms of dementia. Annual patient reviews were commonly used to assess medication regimes and a limited number of memory clinics assessed medication stability within the first few months of diagnosis. As patients were commonly referred back to primary care after diagnosis, GPs became responsible for monitoring patients’ tolerance of and adherence to medication.

Services used at 8 months post randomisation

A review of the services used at 8 months post randomisation from sections within the HSCRU on day and community-based service use shows that although dementia-specific day services were used, the type of services and groups varied greatly in content and were not often provided by the NHS. Additionally, although one-to-one health-care professional home visits did take place, these were rare.

Review of day service use data indicates that 100 of the 371 (27%) activities reported were clearly dementia-specific activities that varied in type, from voluntary memory cafes (n = 31), dementia groups (n = 29), singing groups (n = 13), Dementia Forward (n = 8), Alzheimer’s Society clubs (n = 6) and other groups (n = 13; e.g. Age UK, Age Concern, exercise and other groups). Five participants reported attending memory groups run through voluntary services and one reported attending an NHS-run Living Well with Dementia group.

Overall, 23 dementia groups were reported relating to 29 of the 371 (8%) reported activities and these ranged from church groups for people with dementia to lunch and reminiscence groups. Of note, three of the groups referred specifically to young-onset groups and one to a young-onset group theatre trip. At most sites there was only one participant attending the groups, but at some sites there were more popular groups. These included the York Minds and Voices99 (a peer support and activism group), the Humber Butterflies group100 (a group that involves speakers, quizzes and music) and activities run by the Essence Service101 in Sunderland (which includes outreach drop-ins and a well-being group). Very few groups of any kind were offered by the NHS (four in total) and the majority were voluntary. These examples of dementia groups appear to be open-ended rather than a course, and they highlight the variation in what is available to people with dementia across different areas.

Review of community-based service use data provided further information about which participants attended clinics to receive CST (n = 4) and which health-care professionals visited the participants at home. The following health-care professionals visited participants at home:

• memory support workers (two participants), a memory nurse (one participant) or others from the memory clinic (four participants)

• mental health nurses (seven participants) and community mental health nurses (three participants)

• occupational therapists (six participants)

• psychiatrists (two participants), community psychiatrists (two participants) and community psychiatric nurses (two participants).

The information also indicates that in one case there was an Age UK home visit and in another a befriender visited the participant’s home. Although these data show that there may be one-to-one appointments with professionals, such services are very rare (n = 31, 6%) compared with the overall number of community-based services reported (n = 536).

Allowance for uncertainty

Bootstrapping was carried out to allow for uncertainty. A total of 5000 bootstrap replicates were run and bias-corrected bootstrap 95% CIs were estimated.

One-way sensitivity analysis was performed (1) on utility estimates, using alternative preference-based weights or measures and (2) using alternative ways of delivering the intervention (e.g. without supervision and with facilitators being band 4 or above) were used. A further sensitivity analysis used multiple imputation to impute missing cost and QALY values. 110 A total of 20 imputations were run.

All analysis was carried out in Stata.

Outpatient visits

Overall, there were slightly more outpatient visits in the usual-care arm than in the JtD arm; however, overall costs are similar in the two arms of the trial.

Accident and emergency attendance

There were a total of 41 visits to accident and emergency. Nineteen participants had one visit and two participants had two visits in the JtD arm. Fourteen participants had one visit and two participants had two visits in the usual-care arm. The costs of accident and emergency visits are slightly higher in the JtD arm.

Hospitalisations

Overall, there were 22 hospital admissions (nine in the JtD arm and 13 in the usual-care arm). Reasons for admissions varied and included stroke, chronic obstructive pulmonary disease, pneumonia, heart attack and breathing problems. The costs between the two groups were similar.

Other hospital services

Overall, there were 15 accesses to other hospital services in the study (seven in the JtD arm and eight in the usual-care arm). These included day services and community hospitals. Costs were slightly higher in the usual-care group, although they were still similar.

Transport

There were 26 emergency journeys (15 in the JtD arm and 11 in the usual-care arm) and 38 non-emergency ambulance journeys (10 in the JtD arm and 28 in the usual-care arm, with one person in the usual-care arm having regular journeys for dialysis). The overall cost of transportation over 12 months was similar in both arms of the study.

Community services

The top community services accessed are presented in Appendix 9, Table 42. The most common services accessed were general practices (GP and practice nurse), memory clinics and chiropodists. The usual-care arm accessed community services slightly more than the JtD arm and their costs were slightly higher.

Day services

The majority of participants in the study accessed day care, lunch clubs, social clubs or other services, although only 8% accessed these services via either the NHS or the local authority. The remainder of these services were accessed through charities or private sources. Given that the aim of the cost-effectiveness analysis is to present the costs from an NHS and social care perspective, services accessed only via the NHS or local authority are presented here. Examples of other day services accessed included exercise classes, gym sessions, dementia carers groups and walking groups, and these activities were assumed to last 1 hour. Table 29 shows the average costs with bootstrap CIs. Participants were more likely to incur higher day care costs at month 12 than month 8 of the study. Participants in the control arm incurred higher day care costs than participants in the JtD arm.

Medications

Medication costs were slightly higher in the JtD arm.

The overall cost of the JtD intervention, including intervention and resource use costs, was £1676, whereas the overall cost of usual care was £1067. Therefore, the JtD intervention was more expensive by an average of £609.

Expectations: participants and participating supporters

Both participants and participating supporters talked about experiencing a range of emotions following diagnosis, including anger, depression and fear, while at the same time giving examples of how they had sought to manage this positively. Various strategies were described, including using diaries to support memory, and looking after pets to give focus and maintain a sense of responsibility. Failing memory was significant for the majority:

Yes, my initial reaction was it’s a life sentence, or a death sentence and I was very, very depressed . . . then I thought to myself well, I’ve got to keep going because they might find a cure . . .

Participant 14, individual

Yes, it sounds a bit silly but it’s like if you’ve had a book and some pages have been torn out . . .

Participant 09, individual (talking about impact on memory)

Those participants who did report a change in their daily routines and living situation discussed that sometimes this was in response to concerns of family members regarding safety and ability to cope:

I used to cook and this sort of thing and now she [sister] certainly won’t let me iron my clothes. And she won’t let me cook because I could have caused accidents with someone . . .

Participant 09, individual

Two dyads described loss of friendships and social activity since diagnosis. For example, one couple had stopped playing bridge, which led to loss of contact with a friendship group. However, most participants and participating supporters described remaining reasonably active and independent. Notably, some participants were driving and/or using public or community transport unaccompanied:

. . . I drive, I do voluntary work on a Monday, I go with my friends for coffee, I can go do the shopping . . . so I’m quite independent.

Participant 11, individual

When asked what expectations they had of taking part in the intervention, a reiterated desire was hoping that it would not just potentially help them, but also help others going through the same experience.

The majority of participants and their supporters talked about wanting to know more about the condition, and wanting to share knowledge and problem-solving with others in the same situation. Several participants wanted to improve their mental and physical well-being:

I’ve got to be honest and say I’m a little bit ashamed . . . but I’ve got into a rut . . . I should maybe be doing a bit more activity in my life than what I am.

Participant 05, individual

Only one supporter indicated that they had not expected to see any benefit for the person they cared for:

I must admit, not a lot. You know, because it’s not something that’s going to improve.

Supporter 06, individual

Some participating supporters described the challenges of understanding dementia and its consequences for them and the person they cared for. This could lead to unrealistically high expectations, as observed by a facilitator:

I think that’s been quite hard for her [supporter] and it became a little bit more about her at times, and that was hard to manage . . . and I think that they think that they’re [participant] just going to get better, often carers can just think that, you know by coming to the group it’s going to improve their memory and make them better . . .

Facilitator 10

Intervention training for facilitators and supervisors

NHS staff who delivered the intervention attended a 2-day training workshop to prepare them. The timing of this was significant, as receiving training too early could be detrimental:

[I]t [training] would have been a few months before we were actually going to be doing our first session. So, I guess lots of what we do when we practise in the training, we’ve kind of forgotten about by the time we think of doing it.

Facilitator 02

Two facilitators described how the training had not met all of their needs, resulting in them undertaking ‘top-up’ training:

[A researcher] came out to our base and talked us through it a little bit more again [after the training], and we understood it a bit better after that and I don’t know whether it was just because there was quite a lot of information over the 2 days . . . we came away [from the training] just a little bit confused.

Facilitator 10

The opportunity within the 2-day training to practise components with colleagues and share learning with others was highlighted by all interviewed staff as being particularly helpful:

[D]oing the practical side of it, where they make you get up and almost run a group in front of the other people, they said you will hate it but it will be the most vital part of the training . . . and I did hate it, but it was the most vital part of the training.

Facilitator 10

Most facilitators and supervisors did not feel that a professional qualification was required to deliver the intervention, but experience in leading groups and of working with people with dementia was necessary.

Resources (staffing, manual and materials)

Most facilitators and supervisors found that the manualised intervention and accompanying online resources added to their understanding of the intervention and were suitable for the wide range of staff who would deliver it. Some facilitators found the manual to be a helpful ‘toolbox’ to dip in to, whereas others told us that the manual was overly comprehensive:

[T]here’s so many different topics and you’re trying to encourage the group to think for themselves what they want to do. And there’s so much ‘how do we pick from all of these?’ So, it’s got an advantage and a disadvantage I guess having a lot of choice . . . but then it can also be a bit overwhelming for them.

Facilitator 02

All of the people with dementia who were interviewed talked about being involved, to some extent, in selecting topics for their group. Facilitators also recounted how participants had identified additional topics (e.g. ‘confidence’ and ‘getting to know you’).

Facilitators described how their use of the manual changed over time and, in particular, with improved understanding of the participants they were working with. Several facilitators considered that some activities and concepts within the manual were not accessible for people with more advanced impairments:

I think the conceptual stuff around the importance of engaging with community . . . not sure that was taken on board by a lot of people.

Facilitator 08

[W]e had a couple [of participants] who were quite further on in their [dementia] journey and so I guess lots of the materials and lots of the discussion points and activities were really extremely difficult for them to do and engage in.

Facilitator 02

Most of the facilitators and supervisors suggested that some of the topics and supporting activities presented in the manualised intervention were repetitive or ‘dated’, which they felt made planning and preparing for sessions difficult. However, most described being able to adapt the resources:

[W]e found they [participants] wanted to do things more, they didn’t want to sit and discuss, they wanted to do things and in the manual there wasn’t a lot of doing for some of the topics.

Facilitator 05

Time commitment and workload management

All interviewed facilitators told us that the estimated time to undertake the intervention (i.e. 1 day per week) was inadequate and some were given extra time by their employers, particularly for the first intervention group. The time required reduced with experience of delivering the intervention, but it still remained in excess of 1 day:

[I]t’s been a good 2 days a week if not 2 and a bit days a week. With all the one to ones, the preparation, the actions required . . .

Facilitator 06

. . . [W]e had a bit of insight into what had worked for that first group and what had not. So we could be a little bit more canny about what we were trying to deliver . . . So yes it took less time second time around definitely. However, it still took a lot of time . . .

Facilitator 08

The intervention protocol stipulated that two facilitators were required to deliver the intervention. However, several facilitators said that the number of staff required was related to the size of the group, the needs of the individuals within each group and the type of activities being undertaken. They described accessing additional support (e.g. administrative support for preparing and printing materials, and extra staff/students to assist with out-of-venue activities):

Three [staff] as a minimum, but equally sometimes three was difficult particularly when we were going on outings. When we went on outings I felt that we needed probably at least four depending on the outing really.

Facilitator 06

Supervision

All facilitators appreciated having regular, flexible supervision sessions tailored to their experience and individual needs. Supervision was described as a safe space for reflection to raise issues or concerns and to ask questions:

I felt she [the supervisor] led us in the right direction especially in the beginning . . . so she was very helpful and I think that is what you just need, somebody who is not in the group, not immersed in the group but is there from an outside point?

Facilitator 05

Both facilitators and supervisors described how some groups were ‘complex’ and required more supervision. Three of the four interviewed supervisors described a ‘hands-on’ approach, which included attending a group session to meet participants to understand their needs and the group dynamics. The same three supervisors also speculated that participants may have felt reassured that facilitators had this senior support:

[I]t’s being aware of the individuals who are participating in the group and some of the issues that might come up in supervision about how that person presents . . . so getting a sense of their personalities.

Supervisor 04

The opportunity to have a mix of individual and group supervision was considered useful by all facilitators and supervisors. One facilitator said that individual supervision sessions were important early on to allow time and space to explore personal concerns. At other times, group supervision was preferred, as facilitators wanted to talk through the same issues. All facilitators and supervisors described finding the most beneficial ways of managing supervision.

Co-location of the supervisor and facilitator could have an impact on relationship building and easier access to ad hoc support:

I can’t think of anything unhelpful other than [we] worked part time and we work at different sites . . . So more logistical than anything.

Facilitator 08

In contrast to the benefits described by other interviewed staff, one facilitator did not feel that their supervisor understood their concerns because of differing attitudes to risk:

First I had a discussion with [supervisor] and she couldn’t understand why I was concerned, she just could not understand . . . she felt that the NHS was too risk averse and that was the main problem and it was a problem that we had to manage, basically.

Facilitator 06

Eligibility

All facilitators and supervisors discussed concerns that the abilities and impairments that some participants presented with were different from what they had anticipated:

[A] lot of the training, I think had given us a picture of people who perhaps were more able and more at an earlier stage of dementia . . .

Facilitator 02

[I]n my head I thought we were going to end up with people who were newly diagnosed and quite cognitively intact in those with the mild dementia. That’s not how it was.

Facilitator 08

Challenges to participation recounted by facilitators included the impact of poor memory and cognitive capacity on ability to understand the intervention and enact and take forward learning:

The people who were too cognitively impaired to get anything much out of it other than it was a social activity thing.

Facilitator 08

One participant described observing this in fellow group members:

That is why I was a little bit disappointed with the group I think . . . I think it [dementia] was more advanced with the vast majority of them than it was with me, shall I say, and I found that there was quite a big gap.

Participant 03, individual

It was also observed by several of the participating supporters:

I believe in the moment she [participant] probably got a lot but she’s not taken it from the moment . . . I’ve heard the facilitators trying to enable it to be heard . . . I’ve observed it and they couldn’t have done a better job . . . but it still got missed.

Supporter 01, individual

However, interviewed participants described a range of positive experiences and outcomes, suggesting that they had some engagement.

All interviewed facilitators and supervisors raised how differing participant abilities made planning, organising and delivering the intervention complex. Adhering to the intervention ethos of self-management was particularly challenging. One facilitator described how attempting to meet the needs of the whole group could lessen the experience of individuals:

Because we were always holding back for those [participant’s] that couldn’t [participate] in other respects? Yes, the outings would have been better, the actual topics that came up would have been better and more fully explored for people.

Facilitator 08

Venue, transport and locality

Some facilitators described the challenges of identifying suitable, affordable community venues (as was intended) for the group sessions. Therefore, some interventions were hosted on NHS premises. One participant thought that this was a better approach:

[Y]ou also need a setting that is private because of the sensitive nature of the subject . . . personalised to the people who are taking part it in the group it has to be in a place where we all knew it was safer, rather than public.

Participant 11, individual

However, the same person also thought that use of NHS premises could be stigmatising. One facilitator talked about the benefits of using NHS premises for them, as well as for the participants:

. . . the groups ideally weren’t to be in a hospital setting but we’ve used one of our day rooms [community hospital] . . . people feel comfortable, they feel safe there . . . it’s been nice to know that we’ve got everything we need . . . if we were in the community renting somewhere, you feel a little bit lost because it’s not your familiar environment.

Facilitator 10

Some participants drove or were driven to the venue, whereas others used community or public transport. Some staff described providing and paying for taxis. Everyone interviewed talked, to some extent, about the importance of access to reliable and suitable transport:

Oh if I hadn’t [have] been driving I think I should’ve had to have pulled out altogether.

Participant 01, Individual

Transport has been an issue, there has been a few people . . . decided they would withdraw before the group started because of transport problems.

Facilitator 08

Participants who used community or public transport described several limitations, including poor availability, lack of ‘dementia-friendly’ trained drivers and poor service coverage in rural areas:

[N]ow I try and get the access bus to pick me up . . . if I want it for Friday, [I have to] ring up Monday morning or else you will not get it. I tried to get them to take me to [place not far away] ‘out the area’.

Participant 10, individual

This was echoed by facilitators:

The [participants] who were randomised to the group lived in smaller villages quite far out so that took some organising which we ended up doing through our voluntary services but there were a few times when they couldn’t provide the support.

Facilitator 05

Several facilitators described feeling responsible for assisting with transport, either by driving participants or helping to organise transport. The view was expressed that expecting participants to organise their own transport was difficult:

I’ve filled in some gaps where there’s been transport problems . . . which I know goes against the philosophy behind Journeying through Dementia, but I think in terms of helping people to maintain their attendance, I’ve seen no other option but to fill in the gaps.

Facilitator 08

One supervisor also acknowledged that attending the group with their relative could present logistical challenges for the carer, especially if they were working:

[T]hey’re having to give up a day’s work or to book an annual leave day or shift their working around to be able to be involved in the programme.

Supervisor 04

Several participants, participating supporters and facilitators expressed how the locality of the venue for group sessions was important. Local venues would have improved attendance at groups and out-of-venue activities, and also supported continuation post intervention. However, this was not possible for the majority because of the catchment area for recruitment:

The logistics were difficult because of where mum lived, it would have been easier if she had been closer as you said, but I still think it was important that she did it rather than not.

Supporter 01, individual

Several facilitators highlighted the benefits of developing knowledge of community resources for planning sessions and signposting:

[N]etworking has been brilliant. We’ve been able to get involved with different organisations . . . we’ve developed quite a few links by doing this group that we perhaps didn’t have before.

Facilitator 10

Group dynamics

All the interviewees spoke about the impact of group dynamics on their experience. Influencing factors included group size, gender balance, age range and stage of dementia of participants:

[A] smaller group for me it’s important, if you have a larger group, people may not share so much but if it’s a smaller group.

Participant 11, individual

[I]t was useful because in fact although there were one or two men there similar to myself . . . you do meet people that are not within your same sphere.

Participant 02, individual

[T]hey were a group that just naturally gelled, they all had lots in common, they all lived nearby . . . in the second group it was a complete contrast . . . they were from all over and they didn’t have as much in common.

Supervisor 02

Having common ground was important, but many participants also recognised that meeting people with different interests and experiences was valuable for sharing and learning:

[I]t did a terrific amount of good to you, to find that there was other people, you were not the only one . . . it wasn’t a frightening or sad thing when you found that there’s other people with you with sort of similar problems.

Participant 01, individual

I think that is important because people from different walks of life bring different experiences, everybody has got a different lifestyle and different life experience and we all react to things in a different way. So I think it was a good group.

Participant 11, individual

I think there was a few people within the group reflected on peoples’ different ways of approaching things. And if there hadn’t been the group they would have never seen life in that different perspective.

Supervisor 01

However, perceptions of two participants were that different personalities could lead to negative experiences:

[T]here were a couple of them that were ‘blah, blah, blah’ . . . I mean there was nothing wrong but they thought they were prime ministers sort of thing.

Participant 09, individual

[W]hen I went in through the door some of them would snub you, I mean there’s two sides to every story isn’t there with people, but I always talk to people whether they talk to me or not.

Participant 07, dyad

For one facilitator, participant attitudes were recalled as being a significant factor affecting group dynamics:

[T]he second group . . . were quite negative, like ‘how long have I got’ . . . and we wanted to address those things but not bring the whole group down . . . it was quite difficult to manage that . . . whereas the first group were ‘yes let’s do this’ quite proactive.

Facilitator 04

Several facilitators described the impact of carers taking part in groups:

Yes and no. Yes, because the supporter [knows] how they can be better at caring for them [participants] or if there’s anything that they’re doing that they can do differently at home. And no because I think they need that independence when they come to the group sessions. Sometimes having the carers there can stop them from saying things because they probably don’t want to upset them. It would be difficult for them to be more open about their experiences.

Facilitator 03

Out-of-venue activities

Most facilitators and participants talked positively about out-of-venue activities:

Oh it was better being out and about and seeing different things.

Participant 03, individual

[T]o be able to have the freedom to go out and about into the community with people, again that is something that we don’t ordinarily get to do within our service. So, I think that has been a real privilege.

Facilitator 09

However, there were caveats. For one participant, some of the out-of-venue activities lacked purpose:

We met there [garden centre] . . . and I thought, well what are we doing here? . . . there was nothing, I didn’t think organised for what it was . . . and I thought what a waste of an afternoon, to my knowledge, I couldn’t see the point of it.

Participant 06, individual

For one supervisor, the out-of-venue activities were the most challenging aspect of the intervention and they thought that training for this had been inadequate:

[W]hat wasn’t made clear at the outset were the outings, the expectations, how many you would have to do and the circumstances . . . and considerations before you carry them out. So it’s been those that have been probably the most challenging aspects of it.

Supervisor 03

Some facilitators and supervisors were concerned about the risks involved in accompanying people in the community and, in particular, organisational accountability for this. There was tension between duty of care to patients and the ‘self-management’ community ethos of the intervention:

We are professionals, we are entering into this with service users. We have a responsibility and a duty of care. We need to do things properly. We do need to risk assess because we’re in paid employment and we work for an organisation that would expect that of us . . . it was felt that the response [from the research team] was a little bit woolly.

Supervisor 03

One facilitator described how they investigated potential activities prior to suggesting them to identify risks and make decisions about whether or not to proceed. Using the first outing as a test of abilities of the group, as well as a learning tool, was raised by several facilitators. Facilitators from one site related undertaking formal risk assessments for out-of-venue activities.

Selection of and expectations of, out-of-venue activities had to be managed because of the physical as well as cognitive abilities of group members:

Because of the level of physical problems . . . it limited the range of the outings . . . until we got people out of there we didn’t realise how impaired they were both cognitively and physically.

Facilitator 08

There was still a couple of participants who were restless because they wanted to do more but we were almost having to tailor it for the slowest person.

Facilitator 06

Several facilitators perceived out-of-venue activities as a positive experience for participants:

But the outings have been absolutely, really vital parts of the group actually, a bonding session . . . after [the first outing], it just felt much more bonded because I suppose you’re in a social setting aren’t you and it’s a little bit more relaxed.

Facilitator 10

Out-of-venue activities held in dementia-friendly venues or as part of dementia-friendly services that included trained community staff were particularly well received by participants and improved the experience of the out-of-venue activity. Several supporters also expressed being reassured by this:

Oh that was great that was . . . again you were treated like somebody, you’re not patronised, you’re not treated like a cripple.

Participant 14, dyad (Singing in the Rain at the cinema)

And like at the leisure centre you’ve got somebody there [trainer] . . . who knows the limitations . . . of the dementia.

Supporter 09, dyad

One facilitator noted that some participants and some participating supporters preferred the safe routine of the weekly group meetings compared with the novelty of the out-of-venue activities:

[M]ainly it was the sort of anxiety of changing the routine. Even going on the out-of-venue activities it seemed like ‘do we have to go on an outing’ . . . I think they would have been quite happy to meet for 12 weeks in the room and do things in the room.

Facilitator 06

One-to-one sessions

Facilitators talked about how they had used the one-to-one sessions:

We used the individual sessions so, for the first group there were clearly some dynamics and we used the individual sessions to try and talk through how to support those particular individuals who are having difficulties and try to find a way forward.

Facilitator 09

All facilitators and supervisors described the challenges of delivering these sessions, with two facilitators expressing lack of clarity regarding purpose. The same two facilitators also said that, in their opinion, some participants did not understand the purpose of the sessions either:

I think when it was a matter of us going out to initiate something practical with them there was something tangible . . . whether they had any suggestions about the [goal-setting] . . . that was a little bit more abstract So difficult for them.

Facilitator 08

One participating supporter considered that the person they supported would have benefited from a direct approach, with clear expectations of actions between one-to-one sessions:

[I]t’s almost as if there is no expectation set . . . but I think they [participant] forget so quickly . . . if it’s an agreement that’s on the table, this is what you said you were going to do.

Supporter 01, individual

Two facilitators expressed feeling awkward when pursuing one-to-one sessions:

[A]nd it was so difficult because this gentleman did not want to do anything [in the one to ones], he was more than happy with the way things are and it got to the point where he was getting quite irritated with me trying to help him.

Facilitator 10

[S]ometimes it felt like I was trying to force a goal on them.

Facilitator 04

One participant suggested an alternative schedule of one-to-one sessions:

I still feel that the group finished and there’s a cut-off point, there is no, well look there’ll be a one-to-one session at the end after 3 months or 4 months to see how you’re doing, now that would have been more beneficial.

Participant 11, individual

All participants who recalled one-to-one sessions during interviews described them as an opportunity for relationship building and provision of support to meet their needs:

I valued those one-to-one sessions quite a lot because that was ‘me’ time with [facilitator] . . . that was really good, because I knew that was just me and her talking about the things that I felt uneasy about or something that I felt strongly about, yes so to me they were quite valuable sessions.

Participant 11, individual

Only to be almost like a normal person having a normal life and going out for a cup of coffee [with the facilitator] and not having to think of it as just Alzheimer’s.

Participant 13, individual

Several facilitators felt that one-to-one sessions were used by participants as an opportunity for conversation and socialisation rather than to pursue specific goals, and they considered this a positive outcome. One facilitator referred to the lack of resources within health care to provide time to listen:

I think a lot of them just enjoyed chatting to you . . . just having company which isn’t really what it’s meant for but they enjoy just having that chat, talking about themselves, having someone from the NHS listen to what they’re thinking. Because often they say ‘oh I go to my GP and I’ve got 10 minutes and nothing came of it’.

Facilitator 04

Several supervisors and facilitators described using the one-to-one sessions to consolidate observations from the group sessions or to signpost participants to new activities. In addition, they also talked about the ways in which one-to-one sessions could be used for managing the end of the intervention for themselves, as well as for the participant:

[I]t’s almost like a big thing for us in a way that the group is coming to an end, because you develop professional relationships with people . . . and obviously we want the best for them so, we’ve been conscious that things are going to draw to a close and used the individual sessions as well to talk about how they feel.

Facilitator 09

Although the majority of interviewed facilitators recognised the importance of goal-setting, this was frequently described as being challenging for participants to understand and undertake:

I didn’t get a goal from anyone, I mean most of them said I just want to see what it’s all about, which isn’t a goal . . . nobody actually said, ‘oh, this is really what I’d like to get out of this group’.

Facilitator 06

However, several facilitators also described how goals could be identified and explored:

[Q]uite a few participants have struggled to know what goals they would like to achieve, particularly in the first session . . . however, I have just reassured them . . . and on subsequent sessions, we’ve been able to nurture those ideas into tangible goals.

Facilitator 09

Facilitators and supervisors all talked about the impact that a supporter being present could have:

[S]ometimes we’ve looked at issues where the carer has been really helpful and allowed the individual to speak for themselves and other times where they’ve found it quite difficult to manage with the carer there because they’re tending to speak on behalf of the participant.

Supervisor 04

Intervention facilitation

Good relationships between co-facilitators, participants, facilitators and participants, and facilitators and supervisors, were described by all those interviewed as being vital for successful intervention delivery:

[W]e’ve [co-facilitators] got a good working relationship so we’re able to put our hands up to do certain bits and we recognise each other’s skill set and what’s worked well in previous groups. So we will bring those skills together.

Facilitator 09

One facilitator of a lower staffing grade acknowledged how they had needed more support than their professional trained colleagues to gain the confidence and skills to deliver the intervention.

All those interviewed, both staff and those in receipt of the intervention, provided illustrations of how the facilitator–participant relationship was a key factor in promoting participant engagement:

They [facilitators] were brilliant, they let you talk, they shut you up when they had to . . . they didn’t tell you what to do, they guided you, signposted you, they were brilliant. Yes I didn’t feel as if they were patronising me in any way.

Participant 14, dyad

They made the group feel safe to talk . . . Yes and a couple of people were able to express something about themselves because they must have felt safe.

Participant 11, individual

However, one participant described how, for them, poor facilitation negatively affected their experience:

I thought they [some group meetings] were rather boring because I don’t think the leaders put in any effort to run them.

Participant 13, individual

All the facilitators and supervisors talked about how they had broadened their skills or learned something new from their involvement (e.g. building confidence, expanding skills and challenging perceptions of people living with dementia). All facilitators proposed that this enabled them to deliver the intervention as intended.

Several facilitators and supervisors reflected on existing interventions commonly provided to people post dementia diagnosis. They described the JtD intervention as a more proactive and interactive intervention that offers professional input early, and supports enactment through out-of-venue activities and being more community focused:

I’m working in a situation we haven’t got OTs [occupational therapists] in at that earlier stage of someone’s journey and I really see the significance and the value. So for me it’s really reminded me of what difference some of this work can make on peoples’ journey in life and to be able to take some control over their own situation and what benefit that is.

Supervisor 01

So I think the group activities, to be able to have the freedom to go out and about into the community with people, again that is something that we don’t ordinarily get to do within our service.

Facilitator 09

Participant-facing communication

There were mixed views from both participants and participating supporters regarding the usefulness and accessibility of the intervention folder and paperwork given to participants throughout their attendance. Some participants remembered the intervention folder and said that they found it helpful, whereas others had no memory of it at all and/or had not used it. Most participating supporters said that they had at least looked at the folder contents to understand what the participant was doing and one participant described how the communications had kept them informed when they had missed sessions:

I missed two because we booked a holiday . . . I still got what we’d done each week, we got these little things [summary of each session] back saying what we’ve done and I even got the ones [session summaries] through when I’d been on holiday . . . so I missed nothing really.

Participant 09, individual

Two facilitators reflected on feedback they had received from participating supporters that suggested that at least some of the supporters were not as informed as they would like to be:

I think that the carers feel that they need to know more about the groups and what they actually did. Because I think the group members come out of the session and then they just don’t talk about it or what they’ve learnt from it.

Facilitator 03

A supporter and several participants proposed the value of shared information:

I could have shared it with [sister] and my brother-in-law and then there’s three of us that have the opportunity to say so, what about this, what about that.

Supporter 01, individual

But it was also what you [participating supporter] could hear there [at the group] you wouldn’t have known if you hadn’t been there, because me not remembering . . . that then gave us a reason to talk about it [the group] when we got home . . .

Participant 12, dyad

Confidence through knowledge and gaining perspective

Most participants talked about how they had gained confidence because of their involvement:

It [the group] brings things into perspective about the reality of life. See it wasn’t all about how bad my situation is, I’m going to lose my memory, I’m just going to lose it – no it was just a good thing, it was kind of building confidence not destroying people, but building them up. That there is life after this.

Participant 11, individual

This extended to reports from facilitators about participating supporters seeing the participant developing confidence and maintaining independence:

He actually had the confidence to go out and join new things by himself. He actually started doing a walking football group and going to a day centre every week. So I think she [supporter] was quite amazed at how he’d developed this confidence and been able to make relationships with people. So, I guess she was a bit more encouraged as well to support him in going out and trying new things.

Facilitator 02

The majority of participants described how they valued the opportunity to share experiences and knowledge, and learn about dementia and how other people live with dementia:

I think I got more information about the difference between Alzheimer’s and dementia . . . and it was interesting listening to the other people as well, how they felt about their diagnoses and how people reacted to it.

Participant 11, individual

However, one facilitator recounted how, for one participant, seeing others at different stages of the dementia journey was a negative experience:

I think the guy that withdrew he felt like he was too mild for the group because he had very recently been diagnosed and he was incredibly independent . . . I think he maybe found it a bit negative because he thought ‘oh wow am I going to get like this’.

Facilitator 04

Enjoyment, feeling valued or empowered

Participants talked about enjoying various activities, sharing this with others and bonding. Examples included ‘bat and chat’ (i.e. social table tennis), bowls, dominoes, quizzes, games and role play. Out-of-venue activities that were recounted included visits to local libraries, cinemas, theatres, leisure centres, steam railways, garden centres, National Trust properties, cafes and markets.

Several participants recalled taking part in out-of-venue activities that they would not have considered otherwise:

[T]he [National Trust property] I thought was something else. Somewhere I would never have gone to in my life because I thought ‘what do I want to go there for’. It was out of this world.

Participant 03, individual

Two interviewed participants expressed disappointment that the intervention had not been as they had expected:

I thought it was quite a flippant intervention and I expected it to be a bit more digging deep and it didn’t do that.

Participant 13, individual

I didn’t think that you could get anything really concrete out of what the questions came through [group meeting activities], we weren’t getting it, I didn’t think there was anything there at the end of it.

Participant 06, dyad

Intervention topics were generally selected week by week. Facilitators talked about how the time to forward plan for speakers to attend or to organise activities was very limited and could lead to disappointment:

The odd meeting she [participant] went to, I’d pick her up and she’d say ‘waste of time today’. They were supposed to do so and so and it didn’t happen . . . And that can be a bit disappointing.

Supporter 07, individual

Several participants talked about how the intervention had enabled them to contribute to their community. For example, one participant started volunteering at the local Alzheimer’s Society, undertaking accessibility checks of public venues to see if they were dementia friendly.

Several participants talked about how taking part in group discussions and one-to-one sessions had empowered them:

It’s making me a bit more doing it myself not just sitting in and letting everybody else do it.

Participant 09, individual

One participating supporter also reflected on this in the person they supported:

[T]he real positive thing is that having done that group he then decided he would go to respite care for a day . . . it made him decide that he would go somewhere, that was the really positive thing that came out of it. Having gone out [to JtD], you know we continued the conversation.

Supporter 02, individual

Activity, hobbies and skills

Most participants said that they had found out about and explored various groups or started new activities during their involvement in the intervention. Examples included Singing for the Brain (Alzeimer’s Society, London, UK), working with the Alzheimer’s Society, going swimming, bowls, walking, meditation, attending dementia-friendly theatre and cinema performances, and joining local support groups:

It’s quite interesting really, they woke me up in some ways. And there was people who used to read a lot, I used to do, I don’t read much now, I’ve been trying to start reading again.

Participant 09, individual

Several participating supporters also identified increased engagement in meaningful activity by the person with dementia they were caring for.

Learning ways to adapt activities was described as helpful by some. For example, one participant started using digital devices to do puzzles and jigsaws after being introduced to this during a group. Another shared how they used rowing machines at a local gym (that supports people living with dementia) as an alternative to open water rowing.

One participant described learning from seeing other people living with dementia being active:

I’ve learned how to swim again. I didn’t have the confidence to swim . . . Well it feels like you’ve achieved something again, something worthwhile. I thought these groups was going to be all sort of bingo and nostalgia . . . but when I went to the group and I met these people and they weren’t just talking . . . they were doing these things.

Participant 14, individual

One facilitator described participants continuing with activities post intervention:

[W]e know that a couple of people have kept up with the activity, you know one person has continued . . . to go swimming, table tennis . . . Another person has started something called Men in Sheds and . . . as far as I’m aware at the moment they’ve continued to do that so it has had I feel a positive impact on the participants.

Facilitator 10

However, successfully delivering an activity within the constraints of a group could be challenging, as discussed by one participant:

[T]hey had a ping pong set in the room and they only had it for a very limited time . . . you hardly had a game not a serious game . . . I would have thought more time for something like that, was very popular, could have been given to that.

Participant 13, individual

Social contact and friendships

Nearly all of the participants interviewed talked about friendships made during group sessions and how these relationships were important for attendance, group cohesion and bonding. Nearly all the participating supporters and several facilitators interviewed had similar observations:

You felt you were getting support there from each individual who was in the room.

Participant 06, individual

[S]he was looking forward to it, she enjoyed the company and the people. They got to become friends as such.

Supporter 03, individual

[T]he biggest thing that they’ve taken away from it is the social side . . . they feel a bit more comfortable and a bit more confident in themselves, with how they are because they’ve been around other people in the same situation?

Facilitator 01

One facilitator described how strong friendships between participants and participating supporters who had met at the group had led to them continuing to meet:

I do know however that so many couples from the first group are still meeting up once a week. They’ve made quite a strong friendship group and they now go to some of the Alzheimer’s groups together and meet up and then they meet up once a week at a little cafe in the town. So I’ve seen them out and about at different other venues.

Facilitator 05

Most participants described the immediate sense of loss of social contact or friendships once the intervention had finished, with two participants expressing disappointment that their expectations of continuing friendship had not been met:

I mean you know there was one or two that, you felt that they wanted to be a friend sort of thing . . . but that’s what made me not upset because we all have our own things, but that we haven’t kept in touch.

Participant 12, dyad

For one facilitator, this sense of loss was related to the group being focused on the ‘here and now’ and less so on the future. Participants described receiving no further contact from fellow group members, despite expressed intentions to continue meeting post intervention and having been given the tools to keep in touch with their group by the facilitators:

They [the group] all wanted to see each other again so, if they don’t I won’t believe it . . . because they were all so positive, they wanted it to happen and I think it will. It’s just a waiting game isn’t it? We’ve all got each other’s numbers.

Participant 03, individual

Fear of rejection, logistics or someone taking responsibility for facilitating the group or activity were all cited as reasons for possible lack of continuation:

I thought well I’m sitting here saying nobody’s done it to me [contacted], but I haven’t done it to anybody else either so . . . it’s a different environment though isn’t it, I think you feel safe if you’re in a group, rather than [contacting] somebody and perhaps get a negative [response].

Participant 12, dyad

One dyad (a participant and their participating supporter) considered that without the structure of a facilitated group then people would not be motivated to attend:

[P]ersonally I think that people would just drop out and not bother.

Participant 14, dyad

Several facilitators considered participating supporters or other carers to be a potential significant factor in participants being able to continue activities beyond the intervention, for example by organising or taking a lead:

I think it’s the ones where the supporters are keen in this group, they might keep in touch but if the supporters aren’t then they might not.

Facilitator 04

One participating supporter indicated how they had considered offering to manage future meetings post intervention. However, concerns over safety and taking responsibility had prevented them doing so:

I think leaving it to the last session, to really make people aware that it’s the last one is, I mean and I’m sure it was very clear that there were only so many but I really felt the people’s loss at the time.

Supporter 01, individual

One facilitator speculated that the length of the intervention may have created a dependency among participants and perhaps contributed to the sense of loss they perceived that group members had felt:

[P]eople have said in the group towards the end is a sense of loss and grief . . . and that doesn’t normally happen with our own groups which are about 6 weeks long. We wondered whether 12 weeks was building a dependency for these people?

Facilitator 08

Supporter engagement

Most facilitators and supervisors spoke of the potential benefit to participants of participating supporters or other carers attending the three identified group sessions. One facilitator reflected that they had received feedback from several participating supporters that although their relative was enjoying being involved, often they could not recall what they had done and therefore could not continue at home unaided. However, they also acknowledged how carer attendance could change the dynamics of a session:

So [participant] said I want to be able to use the bus, I didn’t know any of the background . . . I had to do all the groundwork and I came away thinking, is this really practical and that’s when I talked to the husband and he said you’re not doing it.

Facilitator 06

Several participants, their participating supporters and facilitators also suggested that participating supporter involvement could be instrumental in the participant attending the group or not. This was a challenge for those participants without a supporter:

Where they [participant] are by themselves with no supporter it’s a very different picture both on the one to ones and . . . actual attendance to the sessions.

Facilitator 08

Several participating supporters and facilitators described how attendance could offer tangible benefits to carers:

[I]f I hadn’t got friendly with these women and networked with them I wouldn’t have known anything about what we were entitled to . . . talking to them you learned something every week.

Supporter 09, dyad

One supervisor recognised the role that the participating supporters play as a carer, often to their spouse or parent, and the stress that this can cause the individual. They considered that this could present challenges for the facilitators when trying to meet the needs of the participant and the participating supporter or other carer in sessions when carers attended:

[Q]uite often I guess a carer will be a spouse and probably have their own health problems and trying to facilitate their needs as well in those carer sessions can sometimes prove quite difficult.

Supervisor 04

The content and formatting of study newsletters intended for study participants

One suggestion was to produce the newsletter in a narrative rather than newspaper (i.e. columned) format, as people with dementia find reading across the page (rather than up and down) to be more accessible.

Formatting and content of study materials intended for participants

Feedback was sought on drafts of the recording sheets facilitators would be asked to complete and send to participants after each group session. They were also asked about the formatting and content of information sheets for potential recruits to the study.

How to retain participants at the 8- and 12-month follow-up phases of the study

Group discussions were stimulated through a small number of alternative scenarios of participant involvement in the study. Various suggestions were provided, including emphasising the contributions made by participants in both arms of the study in the participant newsletter. Members also suggested booking researcher visits in advance and face to face at a previous visit, so that follow-up visits would just need to be confirmed. This was included in study follow-up researcher guidance. The idea of sending a reminder card (rather than letter or sheet) before booking follow-up appointments was also proposed, as it was considered that a card would be bright, more visually appealing and less anxiety-provoking than formal letters or information sheets. A reminder card was subsequently designed, approved and used.

The accessibility of questions for qualitative interviews with participants

Members were asked for their feedback on how researchers could make the questions intended for the qualitative substudy with participants more relatable and accessible. As a consequence, the language and wording were changed to make them simpler and clearer. It was also suggested that giving participants prompts and reminders during the interview about what had taken place in group intervention sessions might help with recall. As a consequence, the researchers who observed group intervention sessions as part of the fidelity assessment noted down some of the group activities so that they might prompt recall at the later interviews.

Strengths

Limitations

Strengths and limitations of the qualitative approach

The researchers who conducted the interviews observed a number of group meetings as part of the fidelity assessment, which were the same meetings attended by the participants interviewed, thereby providing consistency in approach. The interviewers were, therefore, better placed to probe during interviews to maximise engagement, using knowledge of activities from these observed groups. This was also the case when supporters were present during dyad interviews. Some of the participants and supporters recognised the researchers from the observations, which also created a more relaxed environment for the interviews as the researchers were already known to them. When possible, we encouraged separate interviews with participants and supporters, and the majority seemed happy to do this. When we did have dyads interviewed, this was at the request of the participant and was based on saving the interviewer time, the participant wanting reassurance or couples expressing that they were a ‘partnership’.

Significant challenges in planning and conducting interviews with people living with dementia are the lack of abstract thought, difficulty following sentences and recall due to memory problems, both short and long term. The JtD PPI Advisory Group (i.e. people with lived experience of dementia) was consulted on the language and content of the interviews and the approaches that might be used to aid recall, and we followed its advice to make the interviews as accessible as possible for people living with dementia. Interviews were planned to be around 30 minutes in duration, but some interviews took ≥ 45 minutes. Although none of the participants asked to stop an interview, fatigue could have been an issue for some, which can further impair memory. The time lag between attendance at the intervention and taking part in the subsequent interviews may also have contributed to poor recall. Therefore, the findings should be read and interpreted in the light of this.

Important components of the intervention were the out-of-venue activities and one-to-one sessions. It was not possible to observe these components because of practicalities, methods and the comfort of participants. Therefore, we were not able to prompt and aid recall of these aspects in the way that we were able for the group sessions. This may have limited the discussion of the receipt of these elements of the intervention.

We involved people living with dementia as co-researchers in the analysis of the qualitative data to help interpret findings from the participant and supporter interviews and thereby validate and enhance the analyses conducted by researchers. This was helpful when people living with dementia were still cognitively able to understand and engage with the task presented, which involved commenting on individual quotations taken from the interviews. However, when people had more severe capacity issues, it proved more difficult for them to engage and contribute in a meaningful way.

Strengths and limitations of the fidelity assessment approach

To support reliability and validity of our findings regarding delivery of the intervention, we incorporated data from multiple sources, including independent observers (i.e. researchers) and from a novel perspective (i.e. that of self-report by facilitators and supervisors at each of the sites selected for fidelity assessment). However, presence of the observers at group meetings and the direct involvement of facilitators and supervisors in the fidelity assessment by asking them to self report may have led to heightened awareness and wanting to please the research team. Whether or not the delivery at the four fidelity sites was reflective of all the sites that took part in the trial might, therefore, be questioned.

Evaluation of the integrity of the intervention required a comprehensive and detailed understanding of the manualised intervention and facilitator training programme to maximise impact in delivery and receipt of the intervention for the fidelity assessment, which also had to be objective. Members of the research team (CC and GM) had extensive experience through developing and testing the intervention prior to the trial but did not participate in any aspect of the fidelity assessment.

The observation checklists used for fidelity assessment were formulated based on the underpinning ethos of the intervention, as well as the materials designed to deliver the intervention (e.g. training materials, the manual and other resources). Whether or not such tools are sensitive enough to identify the complex constructs within this type of intervention and the existence, or alternatively absence, of mechanisms for behaviour change needs further exploration.

Although observations were made for one group at two different time points at each of the four fidelity sites, observing only one group at each site may have resulted in the researchers missing potential facilitator drift. Observing multiple groups would have addressed but would have required additional resources.

Recruitment and retention of participants

The challenges of recruiting to dementia research studies are beginning to be documented. 120 Overall, recruitment to the JtD trial was successful in that the number of required participants was achieved. However, the need to recruit to a group-based intervention and slower than expected recruitment meant that the recruitment period and number of recruiting sites was extended. A mixed toolkit of methods was used to recruit participants, for example, recruitment via secondary care, primary care, the JDR database, third-sector organisations and general promotion at sites. The majority of recruits came from referrals from secondary care, identifying participants via the JDR database and visiting groups such as dementia cafes run by third-sector organisations. Mailouts from general practices were also run successfully at many sites in conjunction with the NIHR Clinical Research Network. The lessons learnt from recruiting to the JtD trial will be disseminated to add to the existing body of knowledge.

In terms of retention, during the of summer 2018, the study team identified that the active withdrawal rate was higher than anticipated, at > 20% of participants at the primary outcome measure time point of 8 months. The trial sample size calculation was based on an expected withdrawal rate of 20%, and so it was critical that the active withdrawal rate be brought below this figure. In consultation with study sites, the TMG, TSC and PPI representatives, the study put in place the following four measures to attempt to reduce the withdrawal rate.

1. Continuity with who was contacting and visiting participants (e.g. those who recruited the participant should also attend further visits, if possible).

2. Purposely designed participant correspondence that was more visually appealing and less anxiety-provoking compared with formal letters or information sheets (e.g. a new document was an appointment card, rather than a letter).

3. An updated script was produced for researchers to follow when calling participants. The script was less formal and emphasised the value of participant contribution to the overall study.

4. A limited number of measures were approved to be collected over the telephone, as a last resort, to capture outcome information when the participant seemed likely to withdraw.

After putting these measures into place, the study team observed a continuous drop in the withdrawal rate to 19.2% by the end of the study.

Delivery of the intervention

The study provided strong support to facilitators and supervisors to equip and support them to deliver the intervention. This included training, ongoing feedback on intervention delivery and provision of timely responses to any issues raised. Some difficulties did arise because of an unavoidable gap between training and intervention delivery for some facilitators. However, this was, in many cases, addressed with refresher training provided to sites before intervention delivery commenced.

Resultant findings regarding intervention delivery demonstrate how well this complex intervention can be delivered in practice and, alternatively, the ways in which it is readily eroded, as illustrated by the facilitator records of intervention delivery. It raises questions about workforce readiness to work with people with dementia post diagnosis in a meaningful way. Issues that might militate against optimum delivery included lack of experience of working with people with dementia, an overmedicalised approach and facilitator perceptions of risk.

The difficulty that facilitators appeared to have in creating meaningful written materials for people with dementia is significant and worthy of concern. The findings raise questions about the nature of dementia education that the workforce might benefit from, both those who were more experienced and those without previous experience of working with people with dementia. The findings also question the ability of the existing workforce to deliver all aspects of the intervention as intended and, in particular, the more innovative components that involved taking action in partnership with the person with dementia, rather than limiting sessions to discussion and verbal support. This is important given the proliferation of such programmes for people with dementia and also for those with other long-term conditions.

For clinicians and practice

This multicomponent intervention, which was designed taking into account what people with dementia said would be helpful post diagnosis, proved too demanding for existing services to deliver. Some of the complexity expressed by intervention facilitators and supervisors could be attributed to the challenges of responding to the diversity of needs that people with dementia present with, which can be amplified in a group setting. Furthermore, the complexity of dementia, the diversity of populations living with dementia and the additional comorbidities that people often experience means that it is unlikely that there will ever be one intervention that is suitable for all, and tailoring to individual needs must be a key component. Despite this intervention being designed primarily for people with the diagnosis, participating supporters also articulated receiving benefit from being involved and, in particular, from meeting others in the same situation.

For policy-makers

How to provide the best treatments, care and support for people diagnosed with dementia is not resolved and requires continuing investment in services and in research, with this being supported by robust, evidence informed policy frameworks.

For society

Societal awareness of people living with dementia is increasing and positive messages about how to live as well as possible with the condition are helpful. The JtD intervention aimed to enable people to live well through practical strategies to support self-management and continued participation in life. Even if the intervention cannot be delivered as intended through services, society should seek to support people in these ways so that they can live a meaningful life.

For researchers and academics

It is important for researchers involved in all forms of dementia research to be transparent when reporting the processes and findings of studies, for example, the challenges of recruitment and how they might be overcome; whether or not designed interventions meet requirements and can be delivered as intended; and how we can develop the science of outcome measurement so that instruments to measure positive effects are validated for use with people with dementia and do not rely solely upon recall. As a priority, the research community needs to interrogate the reasons for almost all studies of interventions designed for people with dementia (both pharmacological and psychosocial) reporting similar neutral results to the findings produced by this trial (see Recommendations for future research).

Economic implications

The economic results are clear. The intervention delivered in its entirety is unlikely to be cost-effective. This finding is set against a backdrop of uncertain cost-effectiveness for the majority, if not all, psychosocial interventions delivered through existing services.

On services and usual care

We recommend that elements of the JtD intervention and, in particular, opportunities for people with dementia to practise new and neglected skills should be adopted by clinical practice and provided to individuals who might benefit. Given the additional demands of the group aspect of the intervention, rather than trying to manage the challenges of delivering this intervention in a group setting, certain elements intended to be delivered through the group could be provided on a one-to-one basis. However, it must be acknowledged that participants described benefiting from the friendships and camaraderie provided through the group sessions. Social care and third-sector organisations should be encouraged to provide this form of ongoing support for people in the earlier stages of dementia and separate groups for those supporting them (i.e. their carers).