Long‐term implementation of the Managing Agitation and Raising QUality of lifE intervention in care homes: A qualitative study

The dementia care home workforce receives little specific training. There are successful interventions, improving care outcomes, but it is unclear whether or how to sustain these effects. The Managing Agitation and Raising Quality of Life (MARQUE) intervention aimed to train care home staff to reduce resident agitation and improve quality of life. It was designed for sustainability, with implementation plans agreed with managers. MARQUE improves quality of life. In this separate study, we aimed to examine implementation around 2 years later.

� Change was seen in the areas of communication, staff respect and understanding of the roles of others, and the ability to try new practices � This adds to broader understanding of the factors that impact upon the ongoing implementation of psychosocial interventions

| INTRODUCTION
Globally, the number of people living with dementia is expected to reach 152 million by 2050 1 with 60% living in low-to middle-income countries. 2 In some countries, many people with dementia live in care homes. In the United Kingdom, for example, around a third of 1.6 million people with dementia are expected to be living in a care home, so that social care costs will triple to £45.4 billion. 3 Nevertheless, in many countries, the workforce often has little training, despite being pivotal to the care of older people. Carers work within a challenging and complex system, caring for people with highly complex health and care needs. 4 -6 There has therefore been considerable focus on designing and evaluating evidence-based interventions, to improve outcomes for residents living with dementia and sustaining them with enabling and reinforcing strategies embedded in the system. 7 Recent randomised controlled trials (RCT) had short-term outcomes, Wellbeing and Health for People with Dementia (WHELD) improved quality of life and agitation, 8 and Managing Agitation and Raising Quality of Life (MARQUE) improved quality of life but not agitation. 9 In contrast, EPIC (Dementia Care Mapping) did not change agitation or quality of life 10 and BEYOND, educational and symptom management in people with young onset dementia did not improve agitation, reduce psychotropic prescription or decrease other neuropsychiatric symptoms. 11 We do not know whether interventions enhance long-term outcomes. Changes in staff practices may not be sustained, 12 barriers include poor communication and organisational constraints in the complex care home setting. 13,14 Encouragingly, qualitative investigation following the WHELD intervention found that staff had continued to use a range of intervention activities and processes 9-12 months later. 15 Factors influencing this sustained change were 'recognising the value' of the approach for all; 'being well practised' with enough support to consolidate skills and 'taking ownership of the approach'. Similarly, during development of a measure to support comprehensive assessment of people living with dementia in care homes, the implementation requirement identified was leadership, through supervision and care planning, thus indicating value. 16 This study follows on from MARQUE, a cluster RCT, involving 20 care homes in London, Cambridgeshire, and Buckinghamshire to examine the effectiveness of a complex manualised staff training intervention. 9 As the agents of change were staff and managers, 17 we explore from the staff perspectives whether and to what extent after the trial staff continued to use what they had learnt or sustain any changes into day-to-day practices.

| Intervention
In the original study, the intervention was delivered by two trained and clinically supervised psychology graduates to all day-shift care staff. MARQUE comprises six manual-based interactive sessions (see Table 1 for details of session content).

| Care homes recruited
The RCT eligibility criteria were registered care home, up to 2-hr public transportation from UCL, no plans to close within 12 months, not currently participating in another intervention study, and care home agreement to make the training sessions mandatory for all direct care day staff. 9 Ten were randomised to intervention and 10 to usual care. Laybourne (AL) followed up with the 10 treatment homes.
We show home characteristics in Table 2.
The average home size was 47 residents (range 26-76); 8 were privately owned, 9 registered dementia homes, and all were rated 'good' by the sector regulator, the Care Quality Commission, at the time they were recruited.

Session 2: Pleasant events
This session focused on the importance of pleasant events for residents. It included a focus on how to plan for and include residents living with severe dementia and how to build activities into day-to-day care. The session introduced the idea that even small interaction could be pleasant events.

Session 3: Improving communication
This session discussed communicating with people living with dementia, with a particular focus on how to respond when residents are distressed. It also included discussion and exercises on effective communication with the team and with relatives.

Session 4: Understanding agitation
This session introduced the DICE approach, 18 focusing on describing and investigating episodes of agitation. The content is framed in terms of recognising and understanding the unmet needs of residents with agitation.

Session 5: Practical responses and making a plan
This session focused on creating strategies to manage agitation, including practical and environmental changes and when to ask for additional help. The session also introduced the importance of building these strategies into a plan that can be evaluated.

| Analysis
We anonymised transcribed data and used NVivo 12 to manage data and code. We used thematic content analysis 19

| RESULTS
Six homes agreed to follow-up, two declined because they lacked the time and two did not respond. Non-participating homes did not differ from others in planned action, size or provider type. We interviewed or senior carer (n = 6). In all homes, staff reported that at least one MARQUE component was sustained (Table 3).
MARQUE implementation was often patchy, and this was thought to be because of staff turnover affecting communication of new learning and process and the ability to try new things. In one home, a senior carer said that many staff trained by the MARQUE team had left and thought this was why only one action plan was now implemented. We found three main themes: (i) communication, (ii) respect and understanding of roles and (iii) ability to try new things. Table 4 has been split into parts a-c, presenting illustrative quotations by theme.

| Communication
The creation of a common language and a shared action plan meant

| Respect and understanding
There was increased mutual understanding of workforce roles and respect and acknowledgement of skills, knowledge, and experiences of colleagues. One manager talked of how some staff at higher grades were surprised at how much junior staff knew about the residents and their care. A nurse-lead, reported to be at times impatient and formidable to junior staff, was now more patient and seen to understand that carers can find some situations of care intimidating. This complemented the activity coordinator activities which tended to be whole-home, complex events. Our intervention was co-produced building on evidence of what works for agitated behaviours 21 with input from experts-by-experience as well as interviews with care home staff. 22 The new practices of working together and supporting each other meant residents lived in a more pleasant environment, leading to improved well-being.

| Ability to try new things
With the progression of dementia and variation from person to person, high-quality, detailed observational care based on strong interpersonal relationships is essential to dementia care. 23 Personcentred care is central to policy and practice in the United Kingdom; the Care Act (2014) explicitly places the individual at the centre of care processes. 24 We found that MARQUE strategies, including concrete tools, provided structure for all staff, irrespective of grade, to develop more person-centred interactions with residents. Indeed, engagement of the wider staff team during complex interventions is an important implementation facilitator. 25,26 Training which aims for little changes, not large home-wide 'blanket' activities offers a practical way to personalise care to individuals living with dementia. Planned action including that which focused on emotional and relationship aspects of care was sanctioned by senior management and therefore became legitimate tasksimportant in realist review of people living with dementia in residential care and hearing-related communication care. 27 This echoes process evaluations where leadership in the form of valuing the intended changes was seen to be an important factor in implementation. 16 One of the MARQUE aims was to improve communication between staff and with residents and we are encouraged by the staff reporting this, including improving communication, which staff had emphasised was needed 22 and the evidence is that communication reduces agitated behaviour. 21 Here, teams communicated about shared action plans, common goals developed by staff groups and these underpinned the broader common goal of improving care and coping with residents' distress. This may have been enhanced by a better understanding of roles within the team, central to successful inter-professional working. 28 An important finding is that staff felt better able to cope and look after residents, in-house, without usual levels of input from external care professionals. They tried new strategies which sometimes succeeded, which was both rewarding for staff and good for residents.
Through DICE, staff were explicitly given permission to fail, if an intervention was evaluated as unsuccessful, they could 'throw the DICE again'. This supports process evaluation findings where individuals' openness to change was important. 29 Here, staff demonstrated a confidence and openness to creativity. These changes may partly explain our trial findings; in the short-term MARQUE cost money in training but saved money from reduced use of health services. 9 If staff feel more competent with residents and often have a deep relationship and connection, rather than external professionals completely unfamiliar to them, this may be best for the resident, as well as having economic benefits. Our findings are like the WHELD study, but more long-term showing changes were still in place after the intervention and suggest that embedding changes in practice in care homes is possible years later.

| Strengths and limitations
The qualitative interviews enabled us to collect wide-ranging perspectives from those who participated in the intervention. The interviewer was unknown to the participants, apart from the home managers, allowing some distance from the trial and the inter-

APPENDIX 1
Action plans themed by the main MARQUE sessions for care homes taking part in study.

Call-To-Mind DICE Pleasant events Communication Relaxation
Activities