Data analysis

Qualitative data were managed in NVivo V.10.46 Using a process of ongoing integrative analysis,47 themes were identified and reviewed at each site and then discussed by FJ, DC, KG-W and SH in three monthly face-to-face meetings, followed by review of the full data set. Analysis was underpinned by use of NPT’s constructs and subcomponents. Once EBCD/AEBCD activities ceased in all sites, summary memos with researcher reflections were used to construct a single integrated account. Confirmability occurred through independent, then joint and team half-day analysis cycles, followed by discussing emerging findings with SSC members. Credibility and transferability are evident in the use of data extracts to support explanations of observational and interview data in terms of participants’ engagement with EBCD/AEBCD, as well as the facilitators and barriers to developing and implementing increased activity opportunities in sites.

Coherence and cognitive participation: making sense of and engaging with EBCD/AEBCD over time

Prior to research activity, members of each stroke team attended EBCD/AEBCD training led by the Point of Care Foundation.28 These staff, supported by a researcher, were to lead EBCD/AEBCD implementation at their site. Full-day training for sites 1 and 2 introduced the CREATE Study and the theory and practice of EBCD/AEBCD; sites 3 and 4 received half-day training focused on implementation. Participants in AEBCD sites heard firsthand of how sites 1 and 2 were working to increase activity and discussed changes made by these teams.

it was helpful meeting other people involved in the study on their wards, what they did and how it involved patients. (AEBCD training participant feedback, site 3)

Training afforded these staff the opportunity to start making sense of EBCD/AEBCD and to think individually and collectively about sharing their knowledge and understanding with colleagues, about what activity opportunities were possible and what these would entail practically for the day-to-day work of their colleagues and themselves. Training enabled reflection on work with patients and carers using EBCD/AEBCD, and how it differed from written feedback about patients’ and carers’ experiences. However, ensuring other stroke unit staff understood and engaged with the EBCD/AEBCD process was more challenging. There was no requirement for cascade training or other forms of knowledge sharing. In addition, changes to the UK Health Research Authority and Research Ethics Committee approval processes led to a delay of 4 months between training site 1 and 2 staff and commencing EBCD activity, early interest quickly dissipated; understanding of what the EBCD approach would mean for wider staff groups’ day-to-day work was limited.

However, researcher presence during pre-EBCD/AEBCD observation and interview periods enabled wider groups of staff to engage with and make sense of the project. Staff were curious about data generation methods used including behavioural mapping and observations, researchers used these opportunities to explain EBCD/AEBCD. Patients and carers also became aware of the study through the conduct of observations and behavioural mapping; they sought study information and expressed views on activity outside of therapy.

Coherence, cognitive participation and collective action: the facilitated EBCD/AEBCD process

In sites 1 and 2, EBCD’s structured and facilitated approach was a major factor contributing to staff progressing to a more engaged position, readier to commit to considering how change could happen in their site and to thinking through who needed to be involved. Separate and then joint meetings enabled large numbers of staff to participate in EBCD events (figure 1). They viewed trigger films, heard firsthand accounts of patients’ and carers’ experiences, worked in small groups with colleagues, former patients and carers to identify ‘touch points’, and agreed priorities for action. These were high-energy meetings with shared enthusiasm for change and proved a powerful catalyst for larger groups of staff to share the view that change to stroke unit environments, access to resources, and routine-working practices to increase patient activity was possible. These events led to understanding of the role of co-design groups as being to work on ideas for activity opportunities generated by participants in joint meetings. In sites 3 and 4, starting the AEBCD process at the later joint meeting stage meant knowledge and understanding was more limited initially. However, joint meetings still proved important in engaging staff, particularly as information on how sites 1 and 2 had addressed issues identified in trigger films were shared at these meetings.

It felt quite exciting—it will be interesting to see how it develops, keen to be involved and contribute. (Staff feedback after joint event, site 3)

Other features of EBCD/AEBCD that facilitated cognitive participation and collective action, mainly in co-design group members, were the defined and time-limited nature of the EBCD/AEBCD process. Clinical leads and service managers understood EBCD/AEBCD was a tried-and-tested service improvement model, this legitimised staff time and resource allocation committed to EBCD/AEBCD and associated improvements. Staff in all sites noted EBCD/AEBCD contrasted with previous attempts to introduce change, which were often ‘poorly defined’ in terms of timescales, roles and responsibilities, and operated without additional resources. EBCD/AEBCD’s participatory approach appeared to add a sense of responsibility for staff to deliver on agreed actions, and not to ‘let down’ patients and carers they worked with in co-design groups. As commitment to increasing activity opportunities grew, small groups of staff, not all of whom were stroke unit based, worked together to progress actions agreed by co-design groups.

Cognitive participation and collective action: patient and family member participation in EBCD/AEBCD

EBCD/AEBCD ensures patients and carers express their priorities for change and engages them as equal partners in designing and implementing solutions. In sites 1 and 2, separate patient and carer meetings provided opportunities to explore experiences of stroke and activity/inactivity after stroke. Trigger films demonstrated commonality and difference in experiences and began the process of identification by these former patients and carers as a group, with a shared belief in changing patients’ activity experiences in their local stroke units. Although former patients and carers in sites 3 and 4 did not have local trigger film participation in common, they recognised long periods of inactivity described in the site 1 and 2 trigger films and lack of activity opportunities as similar to their own; trigger films validated important issues raised. EBCD/AEBCD provided former patients and carers with the means and confidence to give voice to their experiences and appeared to help these participants understand how they could work with staff as partners in bringing about improvements.

We felt able to say what we wanted to say and what we wanted to say has turned out to have a valuable effect, so, yeah, very happy to say that we didn’t feel intimidated in any way. (Carer, site 4 post)

Participants in EBCD/AEBCD shared a common experience with their peers and with the staff with whom they exchanged ideas and problem solved in co-design groups. This also appeared true for managers, professional leads and volunteers who were ‘external’ to the day-to-day work of stroke units but involved in EBCD/AEBCD events. Patients’ voices and the authentic examples depicted in trigger films were critical in all staff reorientating their focus to patients’ experiences. These events contributed to participants’ progression from shared understanding of the importance of change to working together to define ways to implement ideas into existing working practices. This included thinking through how changes to increase activity would be actioned, who would take responsibility and who would routinely deliver these actions. These mechanisms represent progression through cognitive participation and collective action over time. Clinical staff facilitated all co-design groups, patients or carers did not seek to lead; but their contributions were actively pursued, and observational and interview data findings confirm these were influential in improvements that occurred. Examples included designing and painting murals (sites 1 and 3) and redesign of patient and carer facing documentation (sites 2 and 4), actions designed to facilitate increased activity through providing stimulating spaces and information about the purpose of new spaces and activity opportunities.

Cognitive participation and collective action: wider staff groups have less involvement in planning and actioning change

Co-design groups brought together those who believed in and were committed to making change happen. These small, mainly self-selected groups provided a stimulus and workspace for staff, former patients, family members/carers and volunteers to engage with and commit to planning and actioning changes to increase activity opportunities and drive new working practices. However, in the early months of the study, it was difficult for staff not involved in co-design groups to envisage whether or how increasing patient activity may affect their roles and working practice. Coherence and cognitive participation, linked to commitment to change and comprehension of the possible benefits of changes in individual and collective practice, developed at different rates across wider staff groups at all sites.

Early ‘wins’ such as a new ‘social corner’ at site 1 raised awareness among all staff, inpatients and visiting family/carers that change was happening. In terms of collective action, some staff incorporated the new space in therapeutic activity or prompted its use independently or with family. Similarly, changing a room at site 4 from a wheelchair storage area to a day room was visible evidence of change and led to its routine use for independent and supervised activity. Such ‘public’ examples enabled wider understanding of how EBCD/AEBCD led to implementation of environmental and practice changes; this stimulated additional discussion among some staff groups about ways to incorporate increasing activity into day-to-day practice:

There’s been staff involved from physiotherapy and occupational therapy, we meet in a morning and we’ve handed over what progressions have been from the CREATE study. So say for example we’re telling them about the updating of the garden and if we need to take anybody to the garden. (Co-design group member, site 4 post)

Communication and engagement across nursing teams working early, late, on weekends, and night duty was more difficult to achieve; observations indicated project information and changes made were not routinely included in nursing handovers at any site; this initially limited coherence and cognitive participation among nursing staff.

Collective action: challenges in implementing co-designed improvements

One challenge of EBCD/AEBCD is the relatively long time period over which change is planned, implemented and evaluated. Managers largely encouraged staff to participate in EBCD/AEBCD but did not allocate time for participation; staff were encouraged to ‘work flexibly’. For nursing staff and HCSWs, attending co-design meetings during shifts was problematic. In site 1 staff sickness and workload challenges meant that three of four EBCD-trained staff did not participate in co-design work. At site 4 two AEBCD-trained staff, both nurses, attended the joint meeting but did not participate directly in co-design groups thereafter. Staff-led co-design groups in all sites experienced similar challenges but other staff volunteered to take responsibility. Enthusiasm for the project across all sites meant some staff attended joint meetings on their days off; and in sites 1–3, some staff participated in co-design meetings on days off or before shifts. Most staff in co-design groups reported completing activities in their own time. Where implementation did not require substantial change in roles and was perceived as enhancing patients’ experience consistent with rehabilitation, staff indicated the extra effort was worthwhile.

I think it’s given me a massive workload, I think it’s doubled it…to be fair. But, I was committed, I mean I took it on, but I've enjoyed that, I'm glad for the changes. [……] I knew we needed changes, so I was happy to help bring the changes. (Staff, site 4 post)

As changes became visible to wider staff groups and involved patients more regularly, staff appeared more receptive to involving external partners. Examples include complementary therapy from a local health network (site 1), fortnightly singing with local university students (site 2) and musicians from a community arts group (site 3). Changes elicited positive and negative responses. At site 2 some staff complained the ‘Something About Me’ (patient information) board mounted behind patients’ beds appeared without consultation. However, these staff also said they liked what the board was designed to do. Such comments highlight how difficult it is to secure cognitive participation and commitment to collective action in staff who have limited engagement with implementation of co-designed interventions. Despite some of the challenges outlined, as activity opportunities became more evident shifts in staff perceptions and behaviours suggested collective action and reflexive monitoring were becoming established in most units. Overall team members’ perceived changes and acknowledged the positive impact on patients’ independence, completion of personal care tasks or engagement in therapy.

Cognitive participation and collective action: influence of managerial authority

Managerial support for service improvements generated through EBCD/AEBCD36 is important. In sites 1 and 2, researchers set up oversight groups including senior managers, matrons/senior nurses and staff with cross-organisational roles. The commitments of these managers meant interaction was largely through email updates or one-to-one meetings. Oversight groups could not be established at site 3 or 4 despite invitations to attend AEBCD activity. Oversight group members in sites 1 and 2 supported project activity, helped navigate complex National Health Service (NHS) Trust organisational structures and, in specific situations, provided resources to enable co-design group ideas to progress. Having identified senior managerial contacts ensured unit-based staff could activate these lines of support and translate ideas into collective action. This was more difficult in sites 3 and 4 where only the chief executive and a therapy lead engaged with AEBCD groups. Site 3 staff completed a sponsored run and the chief executive matched the sum raised. Funds (£8000) were used to redesign a day room previously only used by staff and partly to install a kitchenette for patient and family member/carer use.

In all sites, staff working outside of stroke units, including therapy service managers, matrons, patient experience managers, patient safety officers, volunteer coordinator and estates managers joined co-design groups. Although not involved in every meeting, their participation was often significant in terms of cognitive participation and collective action. NHS staff are acutely aware of resource constraints affecting their services, and most have experienced frustration at organisational barriers to improving services. Barriers identified included infection control and patient safety requirements, delays or inaction when estates work is requested and bureaucratic processes associated with including volunteers in unit-based activity. In early co-design meetings at all sites, staff that were otherwise dynamic and enthusiastic advocates for increasing patient activity often expressed the view that such barriers were fixed and would limit what could be achieved. The perception that ‘the NHS Trust would not allow’ painting murals, adding shelving in patient bays, adding hot drinks facilities for patients and carers to use independently or having volunteers to support patients with social eating was pervasive. In most sites these perceptions proved to be largely inaccurate; wider hospital staff, patient and volunteer service manager typically explained how changes could be realised, and importantly, provided examples of where such changes were already operating in the same hospital.

Coherence, cognitive participant and collective action: facilitating EBCD/AEBCD

Researchers organised staff, patient and joint events in consultation with EBCD/AEBCD-trained staff, recruited former stroke patients and carers, and ensured they could attend co-design groups. They booked accessible meeting rooms, arranged reimbursement for patients’ and carers’ expenses or arranged transport to and from meetings; without this level of support key EBCD/AEBCD events may not have occurred. Researchers co-facilitated staff and joint meetings; although often initiated by researchers, the joint approach built confidence in EBCD/AEBCD-trained staff. In sites 1 and 2, this increased engagement of key staff in EBCD activities. For sites 3 and 4, AEBCD-trained staff had less opportunity to work with researchers prior to joint meetings. In site 3, this appeared to have little impact on co-design meetings; whereas in site 4, team members were less confident in leading co-design meetings. At all sites, newsletters were produced by researchers or core team member to share work of co-design groups to wider staff and patient groups (online supplemental file 3).

Collective action and reflexive monitoring: implementing and embedding co-designed changes into stroke unit practice

It took three to four co-design groups for staff to recognise potential for changes in spaces for activity and to facilitate supervised and independent activity. For most therapists and therapy assistants (TAs) increasing activity was conceptualised not only as reducing boredom and occupying patients’ time but as a therapeutic opportunity. Social activity focused on lunch and breakfast groups at sites 2 and 4 provided opportunities to work on therapy goals including cognitive challenge and functional task practice. These were consistent with therapists’ aims in rehabilitation, did not require working in new or different ways, and were embraced. In some sites, staff reported changes such as the ‘Something About Me’ board (site 2) or the ‘home in the ward’ personalisation of bed spaces (sites 1 and 3) made therapy more relevant as staff could draw on information made available to them through these methods. Over time, the work of implementation became more focused on embedding regularly occurring group or individually focused activity into therapists’ and TAs’ work although the rate at which this occurred varied across sites.

However, post-EBCD/AEBCD observations and interviews suggested limited interdisciplinary consideration of activity promotion. Observational data indicated that initially many nurses did not develop an individual or collective understanding of EBCD/AEBCD as something they needed to participate in. This suggested that some nurse team members did not engage with the view that increasing participation in social, cognitive and physical activity was consistent with their view of what constitutes nursing work in stroke units. However, nursing staff overall expressed support for the intentions of the co-design groups but the challenges involved in nurses (in all sites) being able to attend these groups are likely to have impacted on the nurses’ collective understanding and engagement in the EBCD/AEBCD process. Ward managers were aware of the challenges registered nurses (RNs) faced in taking time out of nursing care provision during 8-hour or 12-hour shifts and allocated time for RNs to attend and participate in co-design meetings. RN and HCSW attendance occurred at site 2 but not in other sites. Observations highlighted nursing staff in each site had high workloads and were frequently affected by staff shortages. Overall, there was limited evidence of RNs encouraging patient-focused social or cognitive activity. There were exceptions, two RNs and an HCSW at site 2 actively engaged in co-design work, and toward the end of AEBCD activity, nurses at site 4 routinely worked with therapists in daily breakfast groups. It is possible that as therapist and family-led activity increases in these units, greater interdisciplinary participation will follow.