A group intervention to improve quality of life for people with advanced dementia living in care homes: the Namaste feasibility cluster RCT

Phase 1: defining the scope of the realist review – concept-mining and theory development

In phase 1, we searched the literature and consulted with stakeholders to develop provisional programme theories about how Namaste Care might work. To identify relevant literature, we searched PubMed and Cumulative Index to Nursing and Allied Health Literature (CINAHL) for all available literature describing the implementation or use of Namaste Care, conducted forwards and backwards citation tracking and hand-searched a book by Joyce Simard, the originator of Namaste Care. 20 We searched for research studies of any design and descriptive items in non-academic journals.

In conjunction with scoping the literature, we conducted face-to-face or telephone interviews with 11 participants involved in delivering Namaste Care, training of care home staff in Namaste Care, and for research within dementia and/or palliative and end-of-life care. Participants were based in the UK, the Netherlands and the USA. Participants were recruited for their known expertise and through snowball sampling. Interviews were conducted either face to face or by telephone or Skype™ (Microsoft Corporation, Redmond, WA, USA) video call. Participants were given a copy of the study information sheet, which provided contact details of the research team, and a consent form that they were asked to read and sign. Interviews were conducted using realist principles36 and were guided by a topic guide. The interview schedules were designed to explore participants’ experiences of Namaste Care for people living with dementia and their views on what they considered to be the essential components of the intervention, and how and on what outcomes the intervention was thought to work. Research Ethics Committee approval was obtained from Lancaster University (reference number 17/wa/0378).

Findings from the literature and interviews were used to develop a preliminary theory in the form of 13 explanatory ‘if–then’ statements. 32 ‘If–then’ statements are the identification of an intervention/activity linked to outcome(s). They contain references to contexts and mechanisms, although these may not be very explicit at this stage. 37 Following this, we held a workshop to review and refine the theory. Participants for the workshop were recruited based on their expertise in Namaste Care and/or in dementia or end-of-life care. The workshop included seven external participants (three of whom had participated in interviews) and six members of the study team (one of whom was a participant and public involvement lead). At the workshop, members of the project team presented the preliminary findings from the scoping, the outcomes identified from the literature and the if–then statements. We adapted nominal group technique to facilitate the discussion of the if–then statements. The purpose was to understand what participants thought was needed for Namaste Care to work, how they thought Namaste Care changed the behaviour of residents and staff, and why/how it worked. Nominal group technique is a process that promotes the generation of ideas to develop a set of priorities and enables the participation of all group members. 38 Participants’ comments were recorded, and statements were ranked by participants in order of importance.

After the workshop, members of the project team who had attended the workshop reviewed the if–then statements, and the rankings, and grouped them into three categories:

1. how Namaste Care is introduced to the care home, including the structure of the intervention, frequency and resources

2. characteristics/approach of the care home staff and characteristics of the Namaste Care programme, for example staff providing person-centred care and engaging in biography work with residents

3. how Namaste Care is delivered, including meaningful activities involving all five senses and adaptation of activities to individual circumstances and preferences.

These categories became the basis for the three preliminary CMOs32 that were taken forward for testing in phase 2.

Phase 2: retrieval, review and synthesis

Description of included evidence

Phase 1

In phase 1 we found 25 papers relating to Namaste Care, 18 of which provided sufficient information for theory development. The majority were descriptive accounts of Namaste Care rather than research studies. Of the seven research studies, three included some before-and-after data,22,44,45 three were qualitative46–48 and one (reported in three papers)27,29,49 used an action research approach. Only five studies21,22,27,44,45 presented data on resident outcomes. The seven research studies and one further Namaste Care study identified during the phase 1 searches50 were taken forward for inclusion in phase 2.

The core elements of Namaste Care, derived from the literature and stakeholder accounts, are:

• the environment (e.g. calm, warm, scented, music, group setting, gentle lighting)

• time (done every day, performed slowly, dedicated time for each person)

• use of loving touch, which might include massage, hair care, skin care, tactile items

• provision of food and drink

• pain assessment.

In phase 2 we included 86 papers. The selection process for studies is summarised in Figure 1. Further details of the Namaste Care studies are listed in additional file 3 of Bunn et al. 32

FIGURE 1.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart detailing the selection of studies for the review.

With respect to research design, the 86 papers reported:

• 17 evidence reviews (not Namaste Care)51–68

• 59 papers reporting 51 primary studies (not Namaste Care)69–129

• 10 papers reporting 8 Namaste Care studies. 22,27,29,44–48,50,130

The designs of the 59 primary studies were:

• 24 randomised controlled trials (reported in 26 papers75–77,82,83,87,95,96,98–102,105,107–115,117,119,123,128,129)

• four non-randomised controlled studies85,91,106,120

• seven before-and-after studies22,44,45,50,79,81,118

• four observational studies (reported in 10 papers69–74,84,86,116,126)

• 13 qualitative studies46–48,88–90,93,94,97,103,121,122,124

• one action research study (reported in three papers27,29,130).

The remaining six studies78,80,92,104,109,127 used a variety of study designs, including retrospective and crossover.

Studies were conducted in a variety of countries including the USA (n = 1720,46,69–74,79,85,87,92,93,101,116,125,126), the UK (n = 1429,44,45,48,50,75–78,89,97,102,103,113), Australia (n = 947,88,99,100,104,114,121,127), Japan (n = 583,95,110,111,120), Spain (n = 396,109,112), Sweden (n = 384,86,106) and Taiwan (n = 3115,118,119). There were two studies in Canada,98,122 Portugal,80,81 Norway90,94 and Italy107,131 and one study in each of the Netherlands,123 France,105 Belgium124 and Ireland. 117 Studies were generally small, with 33 having < 50 participants.

Details of the interventions and how they were delivered

Studies covered a range of sensory and multisensory interventions (Table 1), with multisensory interventions and music therapy being the most common interventions included. Interventions were most commonly delivered by researchers or by outside facilitators such as music therapists. Care home staff were involved in delivering the intervention in only 1381,84,86,103,104,106,111,113,114,117,118,125,127 out of the 59 non-Namaste Care primary studies. Among Namaste Care studies, the programme was delivered by care home staff in six,44–47,50 by Namaste Care workers in two22,27 and by activity co-ordinators in one. 48

The longest and most frequent sessions were reported in the Namaste Care studies, with several22,27,44 reporting that Namaste Care was delivered for 4 hours, 7 days per week.

Context–mechanism–outcome 1: Namaste Care provides structured access to social and physical stimulation

The programme theory is that care home interventions (e.g. Namaste Care) that provide regular and structured access to social and physical stimulation for residents with advanced dementia (C) give staff permission to engage with residents outside task-based care, and trigger responses such as familiarity, reassurance and trust in residents (M), creating a positive impact on resident behaviour and mood (O).

The evidence suggests that one of the most important aspects of programmes is how they enable meaningful relationships to form between providers and residents who are no longer able to communicate easily with speech and have other symptoms consistent with advanced dementia, for example by having the same person provide each session, by incorporating one-to-one interaction into an activity and by providers having the skills to work with people with advanced dementia. 50,64,77,84,92,96,106,121,123,131 By contrast, interventions involving providers who are unfamiliar to residents and/or do not have the appropriate skills99,100 or are unable to engage socially with people with dementia may be less effective. 127 Stakeholders at the workshop suggested that having the same person deliver Namaste Care was not always practical and that rather the aim should be to achieve a consistent approach and attitude towards programme delivery:

What they felt positive about was that they’d managed to create and access what they called a special atmosphere, an environment to practise Namaste Care.

Nam06

Social stimulation appears to be a particularly key component of interventions. In a series of studies, Cohen-Mansfield and colleagues69–74,116,126 evaluated a variety of stimuli for people with dementia living in care homes. They found that social stimulation, especially when it involved one-to-one interaction and the active participation of the resident, had the most dramatic effect on engagement and attention. 73 The importance of one-to-one attention and social stimulation was also highlighted in other studies96,105,113,123 and by stakeholders:

I think when Namaste [Care] really works is when you can create a space where people who are withdrawn can actually come out of that shell and can connect and make eye contact and maybe start to try and talk again.

Nam03

I think most of us are sort of social people, we don’t spend 24 hours a day in a room . . . by bringing them out into a shared space they’re in a room that’s set up for that, they’ve got people around them, and I think the staff there do have good connections with their residents.

Nam04

In Namaste Care, physical stimulation is provided through both the components of the intervention and the environment in which the programme is delivered. For example, scents and soft music were felt to be calming and soothing for residents. 22,27,46 Stakeholders supported this, suggesting that the right space could help overstimulated residents relax and that benefits were also seen outside the Namaste Care space:

We’ve had a lot of residents upstairs that are quite vocal and can be slightly aggressive during personal care and then that’s kind of improved over time with the Namaste and the whole calming atmosphere they’re not shouting out as much.

Nam02

Although some non-Namaste Care studies referred to the importance of environment (e.g. having a private space or a quiet room), in many studies the space was not described. However, studies92,97 did identify practical as well as therapeutic benefits to having a designated space, suggesting that sessions were less likely to be cancelled because of competing priorities and that activities could take place as and when needed by the residents. However, although space might be an important context, it is unlikely to trigger staff engagement without additional resources such as the allocation of time and management support. A study127 evaluating the use of a sensory room for people with dementia (Snoezelen^®; www.snoezelen.info) found that staff missed sessions because they did not see them as a priority. Stakeholders suggested that for Namaste Care to be achievable and adopted as a core part of the work of the care home, it was important that staff were given permission, through the appropriate allocation of time and resources, to engage with it:

I think that’s a really big thing that Namaste adds is that there’s like structured time to really pay attention to the residents and yeah, and give them that extra time, and also the opportunity to make contact with the residents.

Nam09

I think it encourages them to work together and it encourages them and gives them permission to find space for their residents . . . and the fact that everybody’s doing it makes it acceptable within the care home, it’s not like somebody’s taken an hour out of their day to try and spend time with a resident and everybody else is saying you should be working, everybody’s doing a similar thing.

Nam04

The originator of Namaste Care suggests that it should be delivered twice per day, 7 days per week. 20 Three non-randomised studies22,27,44 reported that Namaste Care was delivered in this way, and this was endorsed by some stakeholders who saw it as important in normalising the approach within the day-to-day work of the care home:

When you see proper results is when it’s a programme, when it happens 7 days a week and before and after lunch and that involves a huge change in the culture of the care home.

Nam03

However, some stakeholders suggested that this intensity was unlikely to be feasible in most care homes in the UK. From the non-Namaste Care studies we found little empirical evidence on the optimal ‘dose’ of sensory interventions such as Namaste Care, although a meta-analysis53 of music therapy found that sessions provided twice per week had a more statistically significant impact on disruptive behaviours, anxiety and mood than weekly sessions.

There was little evidence on the benefits of group versus one-to-one delivery. A meta-analysis53 of music therapy for people with dementia found that group therapy had more positive effects on disruptive behaviours and anxiety than individual therapy. However, this analysis did not distinguish between those with and those without severe dementia. One randomised controlled trial115 found that group music therapy was more effective for residents with mild and moderate dementia than for those with severe dementia, and another123 suggested that it was more difficult to achieve therapeutic goals if the ratio of participants to therapist was too high (e.g. five residents to one therapist).

Context–mechanism–outcome 2: equipping staff to cope effectively with complex behaviours and variable responses

Programme theory: interventions that include a ‘toolkit’ of multisensory activities equip staff to work effectively with residents with complex behaviours and variable responses, leading to improvements in resident outcomes (e.g. reduced agitation) through triggering responses such as engagement and connection between residents and carers.

The use of ‘loving touch’ is perceived to be key to Namaste Care, with touch thought to evoke an emotional response that leads to physical engagement. 47 We found some evidence to suggest that touch (such as hand massage) can have a calming effect,95 reduce behavioural symptoms,69,120 improve sleep87 and increase engagement. 93 Hand massage may be more effective than simulated social intervention (e.g. holding a doll) because the intervention combines one-to-one social interaction with sensory stimulation. 116

Stakeholders also reported that Namaste Care had a positive impact on relationships and resident behaviour:

Our residents are calmer and eating more, we’ve got residents talking that never used to talk, we’ve got really anxious residents that would be constantly calling for somebody that will actually sit in there for an hour and not call out once . . .

Nam01

Music also appears to trigger emotional responses in people with advanced dementia. There is evidence that receptivity to music can remain until the late stages of dementia. 58 Primary studies73,86,97,107 and reviews53,58 reported that music therapy improved communication and connection, increased engagement and reduced agitation. 84,86,97,131

There is some evidence to suggest that the most effective interventions are those that equip care staff to cope effectively with the complexity of caring for people with advanced dementia. A systematic review64 of interventions to reduce agitation in people with dementia found that the complexity of behaviour associated with dementia required a multifaceted response that could be tailored to the needs of individuals. Stimulating a range of senses may be particularly important for people no longer able to verbalise,50 and as cognitive function deteriorates people with dementia can become very sensitive to sensory experiences. 78,81 Stakeholders talked about the impact of Namaste Care on care home staff and the way in which the staff perceived people with advanced dementia:

I think watching staff I think what you see is that they realise that this person that may be end of life, they may have really quite advanced dementia but we’re still reaching them . . . they’re still living, they’ve still got all the things that we have, we just need to find it in a different way and I think that changed people’s perceptions.

Nam01

The multisensory nature of an intervention such as Namaste Care means that staff have a range of activities to draw on, giving residents a choice of what is delivered and access to different stimulation (touch, auditory, olfactory and visual). 50,64,75,78,79 The assumption from this evidence is that it is the combined effect of being able to use a range of activities that triggers staff capacity and ability to respond to residents’ symptoms and behaviours, and not individual activities.

Context–mechanism–outcome 3: providing a framework for person-centred care

Programme theory: multisensory interventions that focus attention on residents’ individual biographies and that attempt to connect with residents’ reality make staff more responsive to residents’ needs and lead to improvements in resident outcomes (such as increased responsiveness).

Many studies20,60,61,64,100 suggested that, because people respond differently to the same stimuli, interventions need to be tailored to the needs of the individual. Aspects of interventions that needed to be tailored included the environment,133 the music played,84,110,115,119 the aromas used,82,83,111,124 the way someone was touched99,100,106,113 and how someone was spoken to. 93,121 It was also seen as important to consider people’s known habits and preferences, the stage of dementia that people presented with and whether or not their current preferences may be different from their previous habits. 93,97 How this might work in practice, where the intervention creates a heightened awareness of the individual’s preferences, is illustrated in this quotation from a stakeholder:

While you’re doing it and while you’re observing it you notice things that actually that might not work so well for that person . . . so it’s about thinking about your residents and how they’re changing and what we need to do to keep people involved.

Nam01

Using past and current preferences to tailor interventions was reported to reduce agitation64,69,114 and increase alertness or engagement. 72 There was also evidence (although this was largely qualitative or anecdotal) that personalised interventions made staff notice more about residents and their abilities,27,29,50,93 leading to improved communication between staff and residents,61,68,77,84,127 the development of trusting relationships between residents and caregivers89,92,106 and a shift towards a more person-centred culture of care:29,46

You’ve got to have fundamentally good nursing care and the staff need to have good dementia care training as well . . . but what I think Namaste [Care] does is to make it real for them, you know, it makes the person-centred care real for them and it then feeds into the basic care that they’re giving.

Nam03

In our original programme theory, we hypothesised that Namaste Care would have benefits for family members, either through better connection with their family member with dementia or through improved communication with staff. Few studies measured outcomes for family members, although there was some anecdotal and qualitative evidence that Namaste Care improved connections between family members and residents44,48,50,85,93,94 and between relatives and care home staff,44,46,50 for example as described by this stakeholder:

Very often I think relatives don’t see the positive relationship that carers have because when you’re visiting somebody the staff step back, it’s a visit and they just turn up and say, we need to change your mum or it’s lunchtime, will you feed her or shall I or it’s time we changed her for bed or whatever and so it’s very tasky but in Namaste the family actually see the efforts to get somebody to respond and the positive stuff.

Nam03

People with dementia

Residents were included in the trial if they lived permanently in the nursing home (and were not present to receive respite or day care). The intervention addresses the needs of people with advanced dementia, so participants needed to have had a formal assessment of advanced dementia based on having a FAST score of 6 or 7. 2 A FAST score of 6 or 7 indicates a need for assistance with personal care and urinary and faecal incontinence, and a higher score reflects reduced mobility and a reduced ability to speak. This was assessed by the nursing home manager or another experienced member of staff. However, residents who were bedbound and unable to leave their room to join the group were not eligible to participate. All participants lacked mental capacity (as assessed and documented with the capacity assessment process in use within each site). As the study relied on proxy data collection, each resident needed to have a key worker member of staff available who was willing to provide proxy outcome data.

Informal carers

Informal carer participants were recruited if they were > 18 years, and self-identified as the relative or friend who acted as an informal carer for a participant included in the study. The informal carer could be, but was not necessarily, the person acting as personal consultee.

Nursing home staff

All nursing home staff, including managers, nurses, care assistants and activity co-ordinators, who were paid to provide care were eligible to participate.

Nursing homes

Recruitment of the cluster nursing homes was undertaken between August 2017 and November 2017. Commitment to participation from one small provider chain prior to the award of funding led to two nursing homes from this group participating. A wider search for participating nursing homes was undertaken using CQC databases and local ENRICH (Enabling Research in Care Homes) network contacts.

People with dementia

Senior staff in the nursing home were asked to identify residents who met the inclusion criteria. As all participants lacked mental capacity, the process of recruitment and consent involved personal consultees or, if no personal consultee was available or responded, a nominated consultee was identified.

Informal carers

Eligible informal carers of residents participating in the trial were identified by the nursing home manager or a senior staff member. This person was not necessarily the informal carer who had acted as a personal consultee for the resident.

Nursing home staff

Information about the study was distributed to all nursing home staff in information packs and at staff meetings.

Demographic and clinical characteristics

For residents, data on their age, sex, ethnicity, existing medical conditions and stage of dementia (assessed using the FAST2) were collected at baseline.

Potential outcomes for main trial

We considered two contender primary outcomes: (1) quality of dying (dementia) [using the Comfort Assessment in Dying – End of Life Care in Dementia (CAD-EOLD)]16,135 and (2) quality of life [using the Quality of Life in Late Stage Dementia (QUALID)]136 (Table 2). For people with advanced dementia living in nursing homes, both quality of life and quality of dying are important outcomes. It is also not always clear which is the most appropriate outcome to measure for people living and dying with advanced dementia. Although this was designed as an end-of-life study, it was not clear if the population with advanced dementia who were eligible to receive the Namaste Care intervention would die during the study. Consequently, one of the feasibility aims was to see whether or not an outcome measure about quality of dying would be appropriate for a full trial.

We chose 4 weeks as a primary end point because we wanted to include as many participants as possible in the analysis and recognised that, in this frail and ill participant group, intervention effects needed to be rapid to be meaningful. Attrition due to death is a limiting factor in the successful completion of studies that involve participants with advanced disease. 137 Hence, using early end points is recommended. 138 In previous work evaluating Namaste Care, early deaths (< 2 months) were not uncommon. 139 Benefits from the intervention have been reported within days,139 so by recording an early assessment at 2 weeks and 4 weeks some record of temporal change was made. Missing data and attrition were also likely to be an issue and so having an early measure of 2 weeks could be used to impute missing data.

Secondary outcome measures

A number of secondary outcome measures were also included, chosen because of their potential to measures changes in outcomes that reflect different dimensions of the Namaste Care intervention (see Table 2).

Informal carer interviews

Interviews with informal carers were undertaken to assess the informal carers’ perceptions of Namaste Care (intervention arm) or informal carers’ perceptions of the activities offered in usual care (control arm). Interviews were conducted approximately 16–24 weeks after the first resident was recruited at the nursing home. The informal carer could also be interviewed if their relative died, but to reduce distress this would be at least 8 weeks after the resident’s death, as specified in the ethics application.

Manager interviews

The managers of all eight intervention and control nursing homes were invited to take part in an interview before the first resident was recruited in their nursing home. The aim was to explore their organisation’s readiness for change and the context of care. Demographic contextual data about each nursing home were also collected.

All eight managers agreed to take part in an interview, and the interviews took place in the study nursing homes. The interviews were audio-recorded and transcribed; the length of the interviews ranged from 15 to 40 minutes. The topics included in the intervention arm interview were based on the criteria for readiness for change identified by Goodman et al. 163 that addressed aspects of the care home’s readiness to take on new interventions and the fit between current care approaches and Namaste Care. The interview in the control arm covered the same topics but questions were asked in a more general way without reference to Namaste Care.

Nursing home staff interviews

Group interviews were held in each participating nursing home at the end of the study to assess staff members’ perceptions of Namaste Care (intervention arm) or perceptions of the usual care activities (control arm).

Structured observation

To assess the fidelity, acceptability and appropriateness of Namaste Care, observation of the Namaste Care sessions was undertaken at 2, 4 and 24 weeks after the first resident was recruited in the intervention homes. Observation was conducted in the control homes at 2 and 4 weeks to assess the delivery of ‘usual’ care. In control homes, the researcher observed the residents in communal spaces where ‘usual’ care activities were taking place. A structured schedule that reflected the core components of the intervention was used in both control and intervention homes, as were field notes. Observation was undertaken for up to 20 minutes at the beginning, middle and end of the sessions, and audio-recordings were made to ensure that all verbal data were captured.

Session activity log

An activity log was completed by the staff delivering the Namaste Care session at each session to assess the fidelity and appropriateness of the Namaste Care (intervention arm only).

Interviews

All interviews (manager, informal carer, nursing home staff) were recorded, transcribed and anonymised. Framework analysis was used to analyse the transcripts,167,168 aided by using the qualitative analysis package NVivo 11 (QSR International, Warrington, UK).

Observation analysis and session activity log analysis

Definitions were developed for the core components of the Namaste Care session to ensure that researchers applied the level of agreement scores consistently during analysis. The data from the observation forms and recordings were compared with those from the relevant Namaste Care activity logs to analyse the extent of agreement between the different types of data. This comparison was initially carried out independently by two researchers and then cross-checked to ensure that definitions had been applied consistently across all of the nursing homes. Any discrepancies were discussed and advice sought, as appropriate, from a senior researcher in order to reach agreement.

Learning for a future trial

1. Ensure that enough time is required for the necessary agreements to be in place so that randomisation can occur promptly and this should be factored into study planning.

2. The study design should allow the study to commence as nursing homes consent, rather than waiting for all sites to be ready (randomisation in blocks of four nursing homes).

3. Research support is required for nursing home managers agreeing to participate to facilitate contracting and governance.

Cluster: nursing home

Six out of the eight clusters (nursing homes) (four in the intervention group and two in the control group) remained in the study for full analysis. The withdrawal of two nursing homes after randomisation and before the start of resident recruitment was a result of staffing issues (the resignation of a manager) and start delays (that led a manager to decide that the intervention would not work for the residents in that care home). Once the trial had commenced, no nursing homes withdrew.

Nursing homes

Facilities varied in size from 37 to 60 beds, but the number of nursing home beds provided varied from 24 to 60 (Table 9). Two nursing homes were dual registered and also provided residential care. Mean number of nursing beds was 48.5 (control) and 42.67 (intervention), slightly higher than the national average. Provider status was four private independent (intervention) and two private independent (control) and two not-for-profit (intervention). All nursing homes had undertaken either palliative care programmes [Gold Standards Framework (four sites) or Six Steps to Success in End of Life Care (five facilities)] or bespoke training with a local hospice (one site). Two nursing homes used both the Gold Standards Framework and Six Steps.

Volunteers were present in only three of the intervention nursing homes. In four intervention nursing homes and one control nursing home, a designated general practitioner (GP) worked with the facility, as opposed to individual residents keeping their own GP once they had been admitted to the nursing home. This may explain the relatively small number of GP practices that most facilities worked with, ranging from one to eight. There are no national data regarding the number of GP practices that care homes work with, but studies identify a range from 1 to 50). 173,174 It is recognised that the number will reflect the size of the facility, and the number at the higher end of the range is from older data in 2002. A lack of national data means that comprehensive comparisons cannot be made to examine the representativeness of this sample in this trial and in any future trial.

Participant recruitment

Recruitment was undertaken over 6 months for residents between (14 February–8 August 2018), over just under 6 months for informal carers (21 February–1 August 2018) and over 10 months for nursing home staff (18 January–9 November 2018). The longer duration for staff recruitment reflects the need to continue to recruit staff for proxy data collection for residents throughout the study.

People with dementia

Two hundred and forty-three people with dementia were assessed for eligibility from the six participating nursing homes (see Figure 3). In the first instance, all facilities recruited four residents per facility, and then after 1 month they were able to recruit further participants. In the intervention sites this number was shaped by the availability of space within the Namaste Care area, so in only two intervention homes was a further resident recruited. The control homes were able to recruit a further two and four residents.

The time taken to recruit residents to be able to undertake the baseline visit using the consultee process varied from 29 to 79 days across the sites (Figure 4). The mean number of days between giving out the first consultee pack and consenting the fourth resident (defined as speaking to last consultee) in the six remaining homes was 32 days (range 11–44 days). Even when consent for the fourth resident had been obtained, there was a delay in the commencement of data collection, with an average of 26 days (range 2–43 days) between the fourth resident being consented and the baseline visit as a result of nursing home staffing issues. One personal consultee refused to give consent for the person with dementia to participate because of the amount of paperwork.

FIGURE 4.

Time taken to consent first four residents, by nursing home.

Informal carers

Informal carers for the 32 recruited residents were invited to participate. Twenty (63%) informal carers consented to participate in the study, but only 12 (38%) provided any data at the point of data collection; the eight who provided no data did not provide reasons for this (Table 11). One informal carer withdrew on receiving the baseline questionnaire, stating that there was too much paperwork and he or she was now housebound because of health issues. This recruitment rate for informal carers is comparable with other care home research. 177,178

Informal carers were nearly equally male and female, with no difference in mean age across the two groups, of 69.88 years. Again, this was predominantly a group with white ethnicity, with only one African Caribbean relative. Apart from one, all informal carers were family members. There were no major differences between the two groups.

Nursing home staff

Data from nursing home staff at baseline (n = 97) were collected from all eight of the facilities recruited to the study. The purpose was to establish if staff data could be collected and to what extent this could be done, and if any data were not provided when requested. The findings would be used in a full trial to provide a context for the intervention implementation. After two sites withdrew, data from 67 staff at ≥ 2 weeks were analysed for the six remaining sites (four intervention and two control). This is a predominantly female workforce, with approximately 54 (80.6%) staff identifying as female (Table 12). The staff ranged in age from < 20 years to > 60 years. There was a difference in ethnicity between the groups, with the intervention group (which included homes in large urban areas) being more ethnically diverse. The majority of staff worked full-time. The workforce was relatively stable, with a mean time worked in the unit of 5.89 years and a range of time working in the facility from ‘just started’ to > 30 years. The majority of staff in both intervention and control homes held a diploma or certificate-level qualification.

We have shown that it is possible to collect demographic data on staff, which in a future trial will enable inferences to be drawn about the representativeness of the participating nursing homes with respect to their staffing profile.

Primary outcomes: QUALID and CAD-EOLD

Completion rates of the potential primary outcome variables were high, with only one questionnaire not completed at 4 weeks (see Appendix 1, Table 35). Three residents had missing CAD-EOLD values at baseline as a result of partially completed questionnaires, but all were completed at 4-week follow-up. The second potential primary outcome, QUALID, had an incomplete measure for one resident at 4-week follow-up, but was otherwise complete.

Secondary outcomes

Among the secondary outcomes, completion rates were also high for the CMAI short form (complete at baseline, two missing at 4 weeks) and PAIN-AD scales (all complete at both baseline and 4 weeks). The NPI-Q had greater numbers of missing data as a result of non-completion of questions about specific symptoms (see Appendix 2, Table 36). For example, staff who worked only day shifts reported that they could not comment on sleep symptoms.

Session log summaries by nursing home

We were also able to record how many residents attended each session; one facility (n04) provided Namaste Care for only two residents per session (Table 16; see Figure 7), whereas other facilities maintained delivery to four residents for the majority of the time (n01, n02). Where eight residents are shown (see Figure 6, site n02), this indicates that sessions were being run twice per day.

It was also possible to illustrate the delivery of the intervention over time (Figures 5–8). There were different seasonal and participant patterns in delivery of the intervention. During the summer there were periods of time when Namaste Care was not delivered (n04, n07), with a shorter break in September at site n01. Site n02 delivered Namaste Care consistently across the week over the time period. The number of residents attending sessions also varied over time, with some reduction in the summer months (n01, n02, n07). The facility offering more resident sessions per day was able to consistently offer these to the same residents.

FIGURE 5.

Number of residents attending sessions in nursing home 01 over time in 2018.

FIGURE 6.

Number of residents attending sessions in nursing home 02 over time in 2018.

FIGURE 7.

Number of residents attending sessions in nursing home 04 over time in 2018.

FIGURE 8.

Number of residents attending sessions in nursing home 07 over time in 2018.

Primary outcomes: QUALID and CAD-EOLD

Higher values of CAD-EOLD indicate greater comfort, and lower values of QUALID indicate higher quality of life. In the Namaste Care group the mean value of CAD-EOLD increases slightly after 4 weeks and then decreases, but remains higher than at baseline at 24 weeks. The QUALID decreases slightly after 4 weeks in both groups. At 24 weeks, QUALID continues to decrease for the Namaste Care group, and increases again in the control group (Table 17; see Appendix 3, Tables 37–39, and Appendix 4, Table 40). No statistical comparisons have been made, as the uncertainty of estimates would be very high because of the small sample and the number of clusters.

Figures 9 and 10 are box plots of the distributions of the contender primary outcomes by nursing home. The patterns of change between baseline and 4 weeks in both measures varies between nursing homes, and, owing to the small numbers, we did not seek to establish definitive conclusions regarding effect. In terms of relevance of the two measures, although participants were identified as living with advanced dementia, they were not likely to die imminently. Only four people with dementia died during the study and the study inclusion criteria required people to be well enough to leave their rooms and enter a group space, which excluded people who were towards the end of their lives and considered too unwell to join a group activity.

FIGURE 9.

Box plot of primary outcome measure QUALID at baseline and week 4.

FIGURE 10.

Box plot of primary outcome measure CAD-EOLD at baseline and week 4.

These findings raise questions about the suitability of an end-of-life/dying measure to assess the impact of Namaste Care for individuals in a future trial. The usefulness of the CAD-EOLD as an outcome measure for the end of life is called into question by our findings.

Secondary outcomes

All secondary outcomes had lower observed mean values after 4 weeks in the Namaste Care group (see Table 17), although no statistical comparisons were carried out. Variation between nursing homes is present. Appendices 5 (see Tables 41 and 42), 6 (see Table 43) and 7 (see Table 44) present detailed data for each measure.

The NPI-Q (see Appendix 5, Tables 41 and 42) provides two types of data: the presence of symptoms, and then, if present, the severity of a symptom and the distress this caused. The symptoms of most interest were agitation and apathy (Table 18; data on all symptoms assessed are in Appendix 5).

Apathy was present for only one resident in the control sites at baseline and week 4, whereas in the intervention sites it was reported for six residents at baseline and it then decreased to being reported for two residents at 4 weeks. Agitation was more frequently reported in both arms, for 12 residents in the control sites and for 11 residents in the intervention nursing homes at baseline, with this being reduced in the intervention sites to being present in only seven residents. At a cluster level the severity scores varied between nursing homes, with both a decrease in severity (n01, n02, n04, n03) and increases in severity (n07 and n06) reported (Figure 11).

FIGURE 11.

Box plot of secondary outcome measure NPI-Q total score at baseline and week 4.

The distress scores decreased in all clusters between baseline and 4 weeks (Figure 12). The measure of agitation from the CMAI tool showed that there was an overall decrease in all intervention sites and in one control site (n03) (Figure 13).

FIGURE 12.

Box plot of secondary outcome measure: NPI-Q distress total score at baseline and week 4.

FIGURE 13.

Box plot of secondary outcome measure: CMAI total score at baseline and week 4.

Pain scores, as measured with the PAIN-AD, decreased in all sites from baseline to 4 weeks, except in one control site (n06) (Figure 14).

FIGURE 14.

Box plot of secondary outcome measure: PAIN-AD at baseline and week 4.

Actigraphy

On average, data for analyses were available for 22 days (SD 3.4 days, range 13–25 days) after 04R001 and 06R001 were excluded (Table 19). The reasons for exclusion were that one resident was temporarily admitted to hospital and so their device was removed, and one device failed. Variability at the participant level reflected different factors, such as practical decisions (e.g. the device being removed before the end of data collection) and technical aspects (e.g. the device stopped collecting data because the battery was drained).

Using all available data (n = 31) (Table 20; see Appendix 8, Table 45), we observed that participants were generally inactive: the sample mean Euclidean norm minus one (a measure of wrist acceleration and a proxy for physical activity) was 8.6 (SD 4.3) (moderate physical activity is ≈ 100). The estimated mean sleep duration was 7.6 hours (SD 2.05 hours) and sleep efficiency was 68% (SD 14%), lower than the 78.8% sleep efficiency reported among people with earlier-stage dementia. 179 The average sleep–wake ratio was 4.08 hours (SD 3.22 hours) and the average time of wake after sleep onset was 3.5 hours (SD 1.6 hours).

In terms of the participants’ circadian rhythm fragmentation, the average intradaily variability was 0.7 (SD 0.2) (high intradaily variability values of close to 1 or higher can indicate the occurrence of daytime naps and/or nocturnal activity episodes). Regarding participants’ circadian rhythm synchronisation, the average interdaily stability was 0.08 (SD 0.05) (with values close to 0 reflecting a lack of synchronisation).

Data on activity demonstrated that this population was inactive, with an indication that their sleep and circadian rhythms were not as healthy as they could be, that this population had more fragmented sleep throughout the day and night and that this population had a less consistent pattern of sleep over time. This supports similar results observed in previous studies with similar populations. 170–172

Informal carer secondary outcomes

The data from the informal carers on their satisfaction with care (Table 21) show a slight decline in satisfaction with care over time, but, with small numbers and missing data, this cannot be interpreted in any meaningful way.

In conclusion, our learning of how to best sample and recruit nursing homes into a cluster randomised controlled trial of Namaste Care is that this requires time and the use of wider clinical research and ENRICH networks to ensure engagement at all levels of the organisation (if part of a larger chain). However, senior management support does not necessarily ensure that individual nursing home managers feel able or are willing to participate in research.

The recruitment of residents with dementia through a proxy personal or nominated consent process also takes time and requires investment from nursing home staff and researchers. Once recruited to the study, residents with dementia had a low attrition rate, although the reliance on proxy completion for data collection did reduce the number of data collected. Informal carer recruitment rates were lower than those for other groups, but they were comparable with those of other UK studies in this setting.

Nursing home staff data are considered in Chapter 5.

Learning for a future trial

If an ActiGraph device is used, it should be located on a wrist, not on an ankle. This will ensure that it is more visible to staff who can monitor for rubbing or bruising.

Recruitment

Two main factors appeared to influence the selection of residents across both intervention and control nursing homes: residents’ health and the perceived likelihood that the informal carer would consent to the resident participating in the study. Residents considered for participation were those with advanced dementia whose cognitive abilities were most affected:

I thought it was really nice that some people that can’t do other activities were getting so much attention and the intimacy of it as well, ‘cause when you think some people get up and probably don’t have a cuddle, don’t have their arms touched.

n07 (intervention), staff member S017, bolding added to denote theme

Usually, those informal carers who were known to be co-operative, and with whom staff members had a good relationship, were approached for consent:

I’ll be quite honest, family played a part in that as well . . . And we’ll be perfectly frank with that because there were some people that we thought oh they’d be good but no that son would be horrendous.

n03 (control), staff member S005, bolding added to denote theme

Co-operation of relatives and next of kin was a big thing, because there was a lady that I really think that she would have benefited and her brother who’s the next of kin said no. And that was sad . . .

n02 (intervention), staff member S005, bolding added to denote theme

Even though clear inclusion criteria for residents were identified for the study, it was apparent that nursing home staff applied their own informal criteria based on their judgement about family support for such an intervention, about whether a resident would enjoy the intervention, and about whether the resident’s likely deterioration or complex needs might affect the delivery of Namaste Care in a group setting. In the control sites, staff made similar judgements about family members’ interest in being involved in research.

Data collection tools and process

Some informal carer participants found the questionnaires relatively easy to complete:

I think it was fairly straightforward, I think with the forms . . . they were quite easy to follow.

n01 (intervention), carer C002, bolding added to denote theme

Some participants struggled with the closed-question format of the questionnaires, feeling that qualitative methods of collecting some data may be more appropriate:

I mean some of the questions were really hard to answer because there’s no like some of them wanted either a black and white answer wasn’t it and there was no area, grey area.

n04 (intervention), staff member S003, bolding added to denote theme

I think the paperwork was a bit bureaucratic and the discussion was better . . . I don’t think questionnaires really, because they’re you couldn’t put the discussion we had this morning in a questionnaire, you couldn’t frame it, yes, no, was it positive, mark it on a scale.

n01 (intervention), carer C001, bolding added to denote theme

Proxy measures were also challenging in this situation, given the difficulties of understanding how someone was feeling on a day-to-day basis, or assessing how much time a staff member had spent with that resident:

I don’t know how I am meant to scale somebody on a scale of 1 to 100 on how they are feeling . . . I can’t tell you what his health’s like, because as well as we know our residents, personally we could say they’re having a really good day, they’re about 70 . . . but in reality if they were them looking at them they’d go oh well I’ve got dementia and I’m sat in a chair, I can’t walk, I’m zero . . .

n03 (control), staff member S011

Sometimes [name of staff member S005] might have only been there for the length of the session but it asks about, the last week . . . And also [name of staff member S005] doesn’t provide direct care, she is our co-ordinator for activity, so some of the questions would have been a challenge for somebody in that role.

n01 (intervention), staff member S001

This mirrors the experiences of those who undertook the think-aloud interviews reported in Chapter 6.

Although the research training emphasised that it was necessary for the measures to be completed consistently by the same key worker, this was not always possible. To ensure reliability in a future trial, the same members of staff need to complete the outcome measures at both baseline and 4-week follow-up.

Readiness for change

To contextualise the findings from the study, we sought to understand the nursing home context and each site’s organisational readiness for change. This was measured using the Alberta Context Tool, which was given to staff at the start of the study. This tool has 10 domains (Table 22) and measures the culture, leadership and evaluation aspects of an organisational context that may explain why change does or does not happen. Data were available for all eight nursing homes that agreed to participate in the study.

The small numbers of staff responses from each site meant that it was not possible to compare sites statistically. There were few differences across the sites, or between the two facilities that withdrew and those that stayed in the study, or between the intervention and control nursing homes. There were slightly higher scores for formal interactions in n01, indicating that patient-centred discussions were taking place more often with other professionals within the facility. Lower scores for structural/electronic resources were seen for two nursing homes (n07 and n08) and for organisational slack factors related to staff, time and resources to do their care work. The usefulness of this tool for a future trial is not established, and a larger data set is required to enable more meaningful conclusions to be drawn. The Canadian terminology may need to be adapted to more closely align the concept with the roles and terms used in the UK, which would invalidate the tool.

Nursing home values

Initiating the Namaste Care programme

The decision about where Namaste Care would take place, who would provide it and how frequently it would be carried out was made by the nursing home managers at the start of the study. Only one nursing home (n08) planned to carry out Namaste Care twice per day, 7 days per week. The managers of nursing homes 05 and 07 expressed their concerns about providing Namaste Care with this frequency without extra staffing, given the complex needs of the other residents:

Personally I think that 2 hours in the morning and 2 hours in the afternoon every day is quite difficult for us to protect that every day because of the other needs of residents, because of the client group we’ve got we need that constant attention in the lounge all the time.

n07 (intervention), manager

Although enthusiastic about taking part in the Namaste Care study, nursing homes 05 and 08 withdrew from the study before resident data collection commenced. Nursing home 05 was unable to continue in the study because of unforeseen staffing changes implemented within the nursing home by the wider care home group. This example highlights the external factors that can influence a manager’s ability to implement an intervention within their nursing home. Nursing home 08 withdrew because of delays to the study starting, which illustrates the challenges of maintaining the motivation of nursing home managers who are keen to commence an intervention within their facility.

Training and preparation to deliver Namaste Care

Staff members talked about feeling inspired by the Namaste Care guide and the training that they received, which had given them ideas about how to run the sessions and whom they would be most relevant for:

I couldn’t fault it, it taught us everything we knew and in enough detail as well.

n02 (intervention), staff member S005

Staff providing Namaste Care used the manuals, more so at the start of the study, to help them tailor the Namaste Care programme. They adapted it flexibly to meet their own needs and those of the residents, and the nursing home setting:

Maybe to start with I’ve looked at it [guide] but not as time went on.

n04 (intervention), staff member S002

I think we’ve kind of used all the guidelines and carved it into our own to suit the people, yeah I think so, I don’t know.

n01 (intervention), staff member S006

Staff members in two of the nursing homes also mentioned the value of meeting and sharing ideas with staff from other nursing homes:

It would actually have been really nice to have got back together again in the middle to find out how other people were doing.

n04 (intervention), staff member S008

There was something about coming away from the building, coming away to learn and also coming away to experience different homes and different people, ‘cause we don’t meet people from other homes do we.

n07 (intervention), staff member S004

Creating the Namaste Care space

Each nursing home made the decision about where it would carry out Namaste Care during the trial. Three nursing homes had a separate room that could be used specifically for Namaste Care, and one used a communal area that was set aside for Namaste Care at certain times of the day. The size and availability of space affected what could be done in the sessions, how many people could take part and how often sessions could be delivered:

Well it’s just a small conservatory, very cluttered, so by the time you got four people in there in big chairs . . . and by the time you got two staff in there and maybe me, that was seven people in a very small space.

n07 (intervention), carer C001

The challenge of trying to provide Namaste Care in a room that had multiple uses was discussed by staff members and was observed during the 4-week observation in nursing home 02. This particular home had to use its dining room at certain times for Namaste Care:

That was a big restriction and there was times when we planned to do a session at say 11 o’clock and then we’d only find out there’s a meeting with social workers and things, and this is the only room to have a meeting in.

n02 (intervention), staff member S005

Namaste Care very short and feeling that rushed, sign taken down at 11.50. Sense in room that staff conscious towards the end that nearly lunchtime and have to clear up the room quickly. Had to stop someone wheeling trolley in at 11.50 before residents taken away from Namaste Care space. 2 residents left sitting in the room at the end of the session while space cleared and room prepared for lunch.

Observation field note at 4 weeks, n02 (intervention)

Implementing the Namaste Care programme

Two elements of data are presented here: first, qualitative data on perceptions of Namaste Care implementation, and, second, information on the association between the staff reports of Namaste Care delivery and the intermittent observation by researchers. This enables an understanding of fidelity to the planned intervention, and an appreciation of how best to collect data on intervention delivery in any future study.

Challenges associated with the implementation of Namaste Care

Assessing the content of the Namaste Care or usual care sessions

Staff were asked to complete a session log each time a Namaste Care session occurred; this was to capture details of the residents attending, the time taken and the activities completed. Namaste Care sessions were observed intermittently during the study to assess the correspondence between what was observed to occur during the session and the activities recorded in the session log. There was high agreement between the Namaste Care session logs and the observation data. In a few instances the task had been completed by nursing home staff but had not been documented in the Namaste Care session log (Table 24).

There was low agreement between the activity session logs and observation data for the resident being ‘individually greeted’ and also low agreement for ‘farewells’. This was explored during the staff interviews along with ‘comfort assessed’, as this measure could be assessed only partly by observation.

Whether or not nursing home staff in the control nursing homes carried out any of the Namaste Care preparation activities as part of their ‘usual care’ was documented during observation (Table 25). Table 25 lists the Namaste Care preparation activities that took place in the control (usual care) nursing homes. Like Namaste Care, the session in control nursing home 06 was carried out in a certain space and at a regular time, but it occurred only weekly for 1 hour. Although similar activities might have been carried out, the frequency of sessions differentiates this care from Namaste Care.

Length of Namaste Care sessions

The length of the Namaste Care session documented in the session log was compared with the observation data to assess the agreement between the two sets of data (Table 26). The data show that the nursing homes sometimes overestimated the length of the session in the activity session logs, although they had been asked during the study training to document the time accurately. This may indicate that nursing home staff were concerned that they might have been viewed in a negative light if it was shown that they were not carrying out the length of session agreed with the research team.

Perceived benefits of Namaste Care

Both informal carers and staff commented on the impact that Namaste Care had had on the quality of life of the residents who took part in the study, especially in terms of being more aware of and engaged with family, staff and their surroundings:

She seemed to have more eye contact rather than just kind of staring, and she would smile more occasionally you know when we’d go in and perhaps give her a kiss and a hug . . . Mum seemed to react quite a lot then with her facial expressions.

n01 (intervention), carer C002

She’s responding more as well with proper words we can understand, ‘cause I asked her the other day are you all right? She said ‘Yeah I’m fine’ and I like did a double take, I’ve never heard her talk clearly.

n02 (intervention), staff member S005

Residents were generally found to be much more relaxed and less agitated during and after a Namaste Care session:

I think he’s more relaxed, more sort of settled and looks contented.

n04 (intervention), carer C003

The people who are in there, in that time they are much calmer than while they are sitting in the communal lounge, there seems to be much relaxed.

n07 (intervention), staff member S003

This helped informal carers spend more quality time with the residents:

Well we’ve got a lady that’s quite stiff, and when we do the touch therapy she loosens up and the family have noticed it and they’ve started massaging and stuff like that . . . they think she’s better after a massage now so they do it . . .

n01 (intervention), staff member S005

Staff members’ view of Namaste Care was similar to that of informal carers, explaining that it was different from their day-to-day practice, which was of great benefit to the quality of care and life of the residents:

For me, what I see is it’s a one to one . . . something different away from the client group, away from the communal lounge, somewhere in a quiet area.

n07 (intervention), staff member S003

Many of the staff members who were involved in delivering the Namaste Care sessions appreciated the one-to-one time they spent with the residents, feeling that it helped them build a rapport with residents and provide a better quality of care:

It’s bonding as well with them in’t it I suppose, in a different way, bonding with them without actually having to do the talking and that, just the touch therapy and that . . . It’s like a softer bond, it’s very mellow and soft and gentle.

n01 (intervention), staff member S006

As a result, some staff members talked about adopting elements of Namaste Care into their day-to-day practice:

I find myself more doing little bits and bobs with other residents, you know if they’re distressed and that I think right I’ll just use that now and I’ll do a hand massage and whatever.

n04 (intervention), staff member S002

Although informal carers and staff members talked about residents being able to close their eyes and relax/sleep during the Namaste Care sessions, or being more alert after the sessions, they did not comment on residents' quality of sleep at night:

It’s also stimulation in a soothing, calming way, ‘cause some of ours went to sleep.

n02 (intervention), staff member S007

With regard to level of activity, one resident was reported to have become more mobile following a few Namaste Care sessions, while a female resident in another nursing home who was known to undress in public was reported to have stopped this after attending the sessions. Some of the benefits were perceived to be subtle:

A lot of people would assume that that sort of care is what people do anyway, and the subtlety of the difference between the intensity and the frequency, they wouldn’t see that.

n01 (intervention), carer C001

Informal carers appreciated the one-to-one attention received by the residents:

A little bit of like I say one-to-one attention and a bit of extra care . . . anything that’s going to try and make life easier for them.

n04 (intervention), carer C003

Additionally, among those informal carers who were able to attend the Namaste Care sessions, the sessions were considered an opportunity to be involved in the care of the residents and connect with them in a more intimate way:

One of the biggest impacts was on us as a family, to think more about how mum could be more involved and sort of reached somehow . . .

n01 (intervention), carer C002

Challenges in attributing benefits to Namaste Care

Although informal carers and staff reported that Namaste Care had a positive impact on residents overall, some remained cautious about drawing a direct link between Namaste Care sessions and the changes observed in the residents:

I don’t know if it was because of the Namaste or something happened in his brain, but it’s happened during this time, the man suddenly started to walk or to recognise his daughter when she’s come to visit and when she tried to leave the building he gone to the window and tried to wave at her.

n07 (intervention), staff member S003

They noted the difficulty in determining whether changes in residents’ mood and behaviour (e.g. feeling calmer and acting less agitated) were the result of the sessions or the gradual progression of their dementia:

They’re so ill with dementia, I think it was hard to gauge how much benefit they actually got from it. I think they did benefit from it, but as to say how much, it would be very hard to say because they change so much, they change every day. The men, you know you can go in one day and they’ll be quite settled and they’ll talk to you and smile and stuff, and then the next day they don’t want to know.

n02 (intervention), carer C002

But it’s hard to assess whether it is Namaste or whether it’s a deterioration of her dementia . . . at the end part of that dementia spectrum, if they were presenting with noisy behaviour before, they go really, really calm, and that’s the deterioration, and we could be viewing it as a positive thing but it’s not.

n02 (intervention), staff member S005

They also highlighted that, in many cases, this impact was not permanent or long term, as residents’ agitation continued to fluctuate:

Well he just seemed to be not as agitated as he normally is, but this can change as you know, I mean every day is different and some days even now when I go, he can be quite agitated or sleepy, or on the other hand he can be talking.

n02 (intervention), carer C002

Without being able to ask the residents directly about their experience of Namaste Care, informal carers were also wary of overstating the positive experience and impact of the sessions:

It’s difficult ‘cause you don’t get a lot of response from him . . . I don’t know whether, I mean [resident’s name] doesn’t open his eyes really so he’s not and I don’t know how much they hear, but he must be hearing it because he’s nice and relaxed.

n04 (intervention), carer C003

Perceived impact of usual care

There was recognition of the person behind the dementia in the usual care and activities that staff members provided in the control nursing homes:

We’ve gone through a bit of a change process, due to changes in staff really, we’ve just brought on board this week a new full timer who is just looking like she’s going to be absolutely fantastic, and she’s really tailoring the activities specifically to the individual in terms of their level of disability and interaction, so she’s doing lots of sensory activities, lots of one to one . . . There’s a lot of trial and error, but what [S008] does is she does a ‘this is me’ on admission to get some idea of good ideas that we can theme activities around.

n03 (control), staff member S004

However, a staff member in one of the control nursing homes reported:

I think we do struggle with finding activities for people with advanced dementia. At the moment the activities co-ordinator in particular provides a relaxation morning so that’s quite nice with a darkened room and lights and music, which is quite nice, and there’s some hand massage, it’s difficult for interactive group work.

n06 (control), staff member S003

Satisfaction with care was also reported in control sites. The two informal carers interviewed in the control nursing homes were very appreciative of the care provided and felt that the staff were doing everything they could to engage with residents:

They’re still trying with them, they’re still trying. I’m trying to think who they’ve had a reaction out of, you don’t see much reaction out of any of them. I think they’re doing everything they possibly can to engage them really.

n06 (control), carer C003

This particular informal carer spoke about some of the activities that she had seen when visiting the nursing home:

I’ve been there when they’ve done the light activities . . . when they’ve played the nice classical music and it’s soothing . . . when they’ve had the tambourines out and that type of thing, so there’s a lot of that. And they tried to do like craft stuff with them as well . . . We went when they did the dogs as well, they bring the dogs in, the husky dogs, they do all that type of thing.

n06 (control), carer C003

Feasibility of resource use data collection and valuation

Feasibility and acceptability of economic outcome assessment

The study was designed to assess the feasibility of using a number of outcome measures. The EQ-5D is the standard measure used in economic evaluation, although there is currently some uncertainty about which specific measure to choose of the new EuroQol-5 Dimensions, five-level version (EQ-5D-5L) and the older, more established, EuroQol-5 Dimensions, three-level version (EQ-5D-3L). 184–187 The EQ-5D has been used in nursing home settings,188–191 but there is some concern about whether or not health measures such as the EQ-5D capture the outcomes important to those coming towards the end of life. 192,193 However, alternative generic measures for those in the later stages of the life-course194 are relatively new and untested, although they are included in the National Institute for Health and Care Excellence (NICE) recommendations for assessing social care interventions195 and recommended in other settings for long-term care. 196 This work explores proxy completion rates for EQ-5D and two alternative ICECAP (ICEpop CAPability) measures for those nearing the end of the life-course: the ICECAP measure for Older people (ICECAP-O)197,198 and the ICECAP Supportive Care Measure (ICECAP-SCM) for those at end of life. 199,200 Given the much smaller evidence base for these alternative capability measures, the research focuses mainly on exploring the feasibility and acceptability of these two generic capability measures.

The ICECAP-O is a generic capability-based outcome measure for older people. It has five attributes (attachment, security, role, enjoyment and control), each with four levels. The descriptive system was generated using in-depth qualitative methods198 and valuations were obtained using best–worst scaling. 197 The measure has been validated in a number of settings; its use with dementia patients, however, has been limited to translated versions,201–203 although these suggest favourable validity assessment,201 with the measure showing good convergent and discriminant validity for people with dementia. 202,203

The ICECAP-SCM is a generic capability-based outcome measure for adults at the end of life. It has seven attributes (love and friendship, choice, physical suffering, emotional suffering, dignity, support and preparation). As with the ICECAP-O, the descriptive system was generated using in-depth qualitative methods200 and valuations were obtained using best–worst scaling. 199 The measure is still very new, and published evidence so far has been limited to its use in palliative care patients in a hospice setting. 204

Relationship between economic outcome measures and key measures likely to be taken forward in a full trial

Pearson’s correlations were used to explore the strength of the relationship between overall scores for each of EQ-5D-5L, ICECAP-O and ICECAP-SCM and each of QUALID and the CMAI. The strength of relationship was assessed based on Cohen’s criteria of interpreting correlations of 0.10–0.30 as weak, 0.30–0.50 as moderate and ≥ 0.50 as strong. 206

Think-aloud assessment of feasibility and acceptability for ICECAP-O and ICECAP-SCM

The think-aloud technique was used to assess the feasibility of proxy respondents completing the ICECAP-O and ICECAP-SCM. This involves verbalising thoughts while completing questionnaires.

Feasibility of resource use data collection and valuation

All costs are presented in 2017/18 Great British pounds.

Feasibility and acceptability of economic outcome assessment

Think-aloud assessment of feasibility and acceptability of ICECAP-O and ICECAP-SCM

Twelve proxy staff respondents were included in the think-aloud study, providing think-aloud assessments of questionnaire completion and data from semistructured interviews. Residents for whom the respondent was acting as proxy were, according to the proxy respondents, aged in their early 70s to their 90s; nine were male and three were female. Residents had lived in the nursing home for periods of time that ranged from relatively short (5–6 months) to longer (4–5 years); most had lived in the nursing home for between 1 and 2 years. All proxy respondents had substantial care home experience, having worked in care homes in all cases for at least 2 years. Most had known the resident since the resident had arrived in the nursing home and most spent considerable amounts of time with the resident for whom they were proxying, usually at least 1 hour each day but often substantially more.

The numbers of errors for each of the ICECAP-O and ICECAP-SCM are given in Table 31. The overall error (struggle) rate for ICECAP-O was 6.7% (5%) and for ICECAP-SCM was 8.3% (6%). (Definitions are provided in Appendix 9.) More insight and understanding about the feasibility of completing these measures as a proxy respondent in this context came from the qualitative analysis of the data.

Co-applicants/co-researchers

The two representatives from the Alzheimer’s Society Research Network who had helped design the original bid stayed on as co-applicants in the core study management group. They joined monthly teleconferences, main trial meetings and the protocol development meeting at the end of the study. This helped us to maintain a focus on making our research understandable by challenging the language we used and our underlying concepts.

Public involvement panel

We set up a public involvement panel. We recruited PPI partners from the Crewe area, as this was where our PPI co-applicants were located and where they had additional contacts through their local networks. In addition, one PPI partner lived in an assisted living facility where people with advanced dementia are cared for and an additional member was identified through this. We recruited five PPI partners including our two co-applicants. We met at the assisted living facility to ease travel for PPI partners.

The public involvement panel worked alongside the research team. It was facilitated by Nancy Preston and supported by the research associate. The panel helped to review study documentation, to support information given in care homes (in person), and to analyse and develop a further study. The group met six times during the study, as well as communicating by e-mail. At the initial meeting, the group agreed that, in their experience, current care was variable – at times excellent, but at other times inadequate – and that there is a need to improve the care for people with advanced dementia at the end of life in nursing homes.

Representation on strategic oversight groups

We ensured that there was PPI representation on both the Trial Steering Committee and the International Research Advisory Group. Members attended the relevant meetings in person or by teleconference.

Co-applicants/co-researchers

The two PPI co-applicants have been invaluable to the research. They have maintained the voice for people with dementia and families caring for people with dementia. They immersed themselves in all aspects of the research by attending training events about the Namaste Care intervention but also by taking part in consultation exercises related to phase 1 of the study (realist review). They were integral to the writing of the report and to discussing the findings and take-home messages.

Public involvement panel

The public involvement panel was very successful and we have replicated this method for engaging PPI partners in subsequent research bids. We started with six members, but over time two members dropped out. These were the members with perhaps less current involvement in the direct care of someone with dementia.

The initial meeting focused on getting to know one another and sharing expectations of what the panel could focus on. The initial meetings involved explaining our research methods and giving an account of the Namaste Care intervention. We gave plenty of time for questions and these in turn helped us to develop our methods of communicating about the trial.

A key concern for us was convincing nursing homes of the value of conducting regular Namaste Care sessions, ideally twice per day every day. One of the members of the public involvement panel was able to share how important regular activity was in the care of her husband and explained that, when this had failed, it had major repercussions. This was supported by the experience of other members of the public involvement panel. This information was then shared at the training sessions with nursing home staff to reinforce the message.

The panel reviewed all of the outcome measures; these were predominantly the questionnaires but also included the ActiGraph device. All of the questionnaires were provided in one booklet and the panel gave detailed feedback about the wording used to introduce the booklet and each questionnaire. In addition, they helped us remove wording that could cause anxiety for the carers, such as references to end of life. When reviewing the booklet of questionnaires, one PPI member became slightly distressed. He assumed that the inclusion of questions about certain symptoms meant that his partner would ultimately experience these symptoms, and this made him anxious about the future. We were able to discuss this at the meeting, but it was an important lesson for the researchers about how best to phrase requests of PPI partners and to give more explanations about the tools we use, which cover a broad range of possible symptoms and do not necessarily mean that everyone will experience these. With respect to the ActiGraph device, the members of the public involvement panel stressed the importance of knowing someone’s life story in relation to wearing this, for example, if the person wore a watch and, if so, on which wrist they wore it. In addition, they suggested clear processes for staff to follow if a person did not want or no longer wished to wear the ActiGraph device.

The public involvement panel became involved in recruitment at the nursing homes. One or two members joined researchers at all our site initiation visits. Their role was to support the researchers when discussing the trial both with care home staff and with carers of residents. This also meant that family members could speak to someone, other than the researchers, who understood not only the trial but also the experience of caring for a relative with dementia. The panel members were able to answer questions about the study using more accessible language.

Finally, the public involvement panel members were involved in looking at our analysis. Not only did they assist in ensuring the clarity of results but they also asked pertinent questions from their perspectives about what would be interesting for family carers to know. They were particularly interested in the results from the ActiGraph device and were involved in discussions about how the data could be shared in a meaningful way. We agreed that an overview of the results was most appropriate. The panel also attended the dissemination activities after the trial.

The public involvement panel members wrote their own piece for the Alzheimer’s Society research network newsletter about their involvement in the study as part of a longer article about involvement in research. 209 The relevant extract is presented below:

Our PPI brief is to comment on the research as it progresses and ask for any clarification we feel is needed, advise on communications with people with dementia and their carers, and assist with dissemination and evaluation. We have current knowledge of practice in care settings and can ground the academic theory of care research in the reality of care practice. We are well placed to suggest what is likely to be possible, which can save time, effort and money. And as we’re not intimately engaged in the research, we can stand back and call attention to the more general but equally important aspects of the work, such as making sure the findings are shared widely.

Reproduced with permission from Alzheimer’s Society, 2019. Care and Cure Magazine209

The public involvement panel members felt that, should the study lead to a main trial, an explanatory leaflet for relatives would be helpful.

Representation on strategic oversight groups

One panel member sat on the Trial Steering Committee and one PPI member sat on the International Research Advisory Group. They were able to attend three Trial Steering Committee meetings and one International Research Advisory Group meeting, respectively.

Sampling and recruitment of nursing homes

It was possible to recruit both nursing homes and residents, but this took several months. The length of time from initial contact to the baseline visits at the start of the study was, on average, 34 weeks, reflecting the time needed to identify facilities and then address contractual arrangements. In some sites the recruitment of the initial four residents in each nursing home took up to 2.5 months.

Recruitment, retention and attrition rates

Of the eight nursing homes recruited, six remained in the trial (two withdrew before the trial started). Only one resident withdrew after the trial began, for health reasons; four other residents died during the study. The rate of attrition among informal carers was low.

Selection, timing and administration of primary and secondary outcome measures

The choice of two primary outcome measures was to inform a future trial. The quality-of-life tool (QUALID) was appropriate, but the quality-of-dying measure (CAD-EOLD) was not, because those well enough to attend Namaste Care in a group context were not imminently dying, and their deterioration was not fast. One secondary outcome measure (CMAI) was found to be useful in this population. Data from both the QUALID and the CMAI, although not analysed for effect, did show a change in scores at 4 weeks, and in QUALID at 24 weeks, indicating an improvement for residents. For QUALID, this was sustained until measurement at 24 weeks. Measuring the primary outcome at 4 weeks was appropriate given the slow deterioration in this population. The ActiGraph device was acceptable to wear, and data were collected with minimal device failure. This mode of data collection addresses the limitations of proxy outcome measures in offering objective measurement, but there are still questions to be answered about the clinical relevance of the findings. All economic outcome measures (EQ-5D-5L, ICECAP-O and ICECAP-SCM) were feasible for proxy completion, with high completion rates and a weak to strong correlation with QUALID recorded at the different time points. Staff proxy respondents noted some difficulties in completing the ICECAP capability measures in a think-aloud study.

Intervention acceptability, fidelity and sustainability

The Namaste Care intervention was acceptable to staff delivering it, to the family members of people with dementia receiving it, and to the people with dementia receiving it, in the opinion of those staff and family members. The person-centred nature of the intervention allows residents to change their response to what is offered, allows residents to not participate on particular days or to stop participating during a session, and allows for changes in activity to better meet in-the-moment needs. This is supported by the realist review. In our programme theory (see CMO2), we concluded that because Namaste Care includes a range of activities it is easier for staff to find activities that meet individual preferences and responses.

The aspiration of the intervention’s originator was for Namaste Care to be delivered in two sessions per day, each 2 hours long. This was not consistently possible in any of the facilities. When Namaste Care was delivered twice per day, the sessions were shorter. The delivery format reflected the space available, staffing levels and routines in each facility. We understand that a forthcoming version of the originator’s book will acknowledge the challenges in delivering 14 2-hour sessions per week and will support a more flexible delivery model to suit different nursing home contexts (Joyce Simard, March 2019, personal communication).

In terms of sustainability, delivery fidelity changed over time, partly reflecting wider seasonal staffing issues. No longer taking part in a trial has allowed sites to continue to deliver Namaste Care in a way better suited to their resources and environment. We conclude that the implementation and delivery of a person-centred intervention needs to be nursing home centred.

The health economics costing analysis found that homes differed in the extent to which they incurred additional costs from providing Namaste Care, with some incurring no additional costs and others incurring costs equivalent to an average of £222 per resident across a 4-week period.

Validated interview measures

Of the two contender primary outcome measures, quality of life was judged to be the more appropriate for evaluating this intervention in this population. Although people with advanced dementia were recruited, they were not near enough to the end of life for a quality-of-dying measure to be appropriate. The secondary outcome measure for agitation (CMAI) was also judged appropriate to evaluate this intervention. However, as identified in the interviews with staff about data collection and in the think-aloud interviews, staff found it challenging to complete the proxy tools in terms of having the confidence to assess the experiences of people with dementia. Caution therefore needs to be used when interpreting such results.

Actigraphy

As a device for data collection, the ActiGraph device was acceptable to and tolerated by all participants. In the light of the one adverse event identified, we would recommend that the watch be used only on the wrist in this population to ensure that it can be checked for any irritation more easily and more often. Further consideration is needed of how best to use these data to support the findings from the proxy outcome measures.

Informal carer outcomes

Data were collected from informal carers who participated. The SWC-EOLD, which focuses on care at the very end of life, was therefore not the most appropriate informal carer measure. The numbers were too small to make meaningful statements about completion rates.

Nursing home staff outcomes

Measures were less well completed than proxy measures on residents.

Additional data

Data on the organisation’s readiness for change were collected to provide context for implementation. Some items are not necessarily transferable to a UK context, particularly with respect to resources. We would therefore recommend not including this in a full trial.

Review of full trial indicators

We have reviewed the full trial indicators as articulated in the funding bid (Table 32). The feasibility trial indicators were partially met. The acceptability (to staff and family) and appropriateness of the intervention in a UK context was met, but not the dose indicator with respect to the frequency of session or length. Recruitment of residents was possible and was limited by the Namaste Care space rather than by an inability to recruit appropriate residents. Once recruited, attrition to receive the intervention was low, and only four deaths were recorded among the sample over 6 months. There was an almost 100% completion rate for the primary outcome measures and little attrition, as expected in an end-of-life study. The numbers of deaths are more in line with those in a palliative population. 210

On this basis, we propose that a full trial would be possible given the following recommendations regarding the trial design, changing the intervention specification to make it more person centred and changing the implementation process so that it can be used flexibly in different nursing homes.

Sample size calculation for a definitive trial

We consider two potential primary outcome variables for a definitive trial, QUALID (quality of life) and CMAI (agitation). Analysis of the primary outcome at 4-week follow-up would use a mixed analysis of covariance model, with nursing home as a random effect and baseline value of the outcome measure as a covariate. To calculate the sample size, we need estimates of the following parameters.

Sampling and recruitment

Outcomes: selection, timing and administration

The identification in the realist review of the importance of the moments of connection in the Namaste Care intervention has led to the person with dementia’s level of social engagement being considered as an outcome of interest. However, because the direction of change in response to Namaste Care can be either greater social engagement or becoming calmer, it may be necessary to identify two primary outcomes or to identify an outcome measure that would allow for measurement in both directions. The CMAI (agitation measure) as an outcome, alongside a secondary measure of apathy, is proposed. A focus on using only the QUALID, CMAI and one further measure for apathy, supported by actigraphy, or another objective measure, would also reduce the administrative burden on proxy respondents.

The limitations of proxy data collection outlined in this study for the health and economic measures will continue through to a full trial, as there is currently no better alternative. However, there is no reason to believe that the effects of proxy measurement would differ between the trial arms as long as proxy respondents are selected in a similar manner in both. The collection of actigraphy data as an objective alternative that overcomes these limitations remains a possibility. If the primary outcome measure is agitation, it may be possible to support the use of these data to corroborate proxy reports in this domain. Consequently, we would propose that the proxy outcome measures still be used, but that caution is applied when interpreting their results.

The lack of blinding to the status of sites among the individuals completing outcome measures cannot be avoided because the tools require knowledge of the person being assessed and so it would not be possible to bring in external data collectors. This limitation will carry into a full trial.

Use of observation alongside the activity log of Namaste Care is not warranted in a future trial as fidelity was good. The selection of economic outcome measure(s) should follow policy guidelines when a full trial is commissioned. Currently, NICE social care guidelines195 recommend a parallel evaluation based on capability measures, such as ICECAP, meaning that all three economic outcomes collected here (EQ-5D-5L, ICECAP-O and ICECAP-SCM) could be useful for aiding decision-making.

Intervention content and implementation

We recommend that this person-centred intervention be implemented in a nursing-centred way, or in a way that allows context-level adaptation to occur. 216 Table 34 presents the learning from this trial with respect to the intervention specification and also the implications for the implementation processes to support a future trial.

The focus of the intervention varied according to the needs of the recipient and staff confidence. Staff members reported flexibly using the intervention components as ‘intuitive’, reflecting some staff members’ backgrounds in touch therapies. Hand massage was frequently used in this study, possibly reflecting what staff members had been taught during training. However, the use of ‘loving touch’ is perceived as key to Namaste Care, and the realist review found evidence to suggest that hand massage has a positive impact on resident outcomes. Such an activity may be particularly beneficial because it combines social interaction with sensory stimulation, and it is a relatively easy way to introduce this concept and practice. Future training needs to incorporate a wider range of potential activities to promote their inclusion in Namaste Care sessions.

Identifying an appropriate person-centred framework to structure activities would strengthen the way in which the person-centred focus in Namaste Care is provided. For example, McCormack and McCance’s person-centred framework217 offers five dimensions (working with a person’s beliefs and values, engagement, shared decision-making, having sympathetic presence, and providing holistic care) that map on to Namaste Care practice, but offers a theoretical and empirical basis for including these activities.

Discussions about introducing Namaste Care are important at the start of the study, even though staff members may feel that they are familiar with the concepts. Family and informal carer involvement is important not only in maintaining a person-centred focus on which activities are offered and to which residents, but also potentially as a wider resource for its delivery. The development of family-focused materials, as suggested by the public involvement panel, would support this engagement.

The setting used to deliver the intervention in a group context varied, which influenced how many people could be involved, what could be provided and for how long it could be provided. In multiuse rooms, the sessions were generally shorter. This is supported by the realist review. Studies suggested that there were practical benefits to having a designated space, as sessions were less likely to be cancelled because of competing priorities and because activities could take place as and when needed by residents. However, in our study, the site in which the Namaste Care room was used daily for other purposes delivered sessions most consistently. Flexible criteria for the Namaste Care space and an exploration of the use of volunteers and family members may create more options to ensure that the intervention can be delivered. Learning from other ongoing international studies in this area could inform recommendations. 31

In terms of implementation processes, the resources created for the intervention delivery were well received, but not greatly referred to following the initial training. Further review of the materials, including the guide, would be undertaken prior to a full trial (see below) at which time consideration of family and volunteer roles could be addressed. The 1-day training with an external facilitator and follow-up visit was well evaluated, but a need for ongoing support was identified; for example, regular meetings for care home staff providing the Namaste Care to learn and share experiences would be beneficial.

The ongoing interest in Namaste Care has led to an Alzheimer’s Society-funded implementation study of the intervention currently running. 218 As part of this study, a community of practice has been established (www.adscommunities.ning.com) that could act as a wider resource and ongoing support in a future trial. The Worcester study team has also developed its own manual and training package, which would need to be reviewed before a main trial to ensure a coherent message about Namaste Care and its delivery.