Results: impact of changes

Impact of changes: qualitative findings – staff, patients and families

In both forms (full and accelerated), the EBCD process facilitated some observable changes to the nature and use of communal spaces, to structured therapy-led group activity and to increasing activity opportunities at the individual and social group level. Whereas the impact of changes in activity by improving access to more groups or in space by having new or repurposed social areas was tangible and visible, less observable change was seen in communication, and this resonated with the findings of the pre-implementation analysis. This was despite several initiatives developed by the co-design groups, including new codes of conduct, changes to staff induction and renewed emphasis on communication in team meetings. Our data showed that interactions with most staff remained largely ‘task focused’. This raised our awareness of the numerous and sustained approaches needed to make changes in communication. The challenge of changing ‘how’ staff interacted in a prevailing culture of a highly medicalised clinically led environment remained and, findings from our post-implementation interviews and our observations substantiate this.

Beginning with space as the first priority area, we draw on multiple sources of qualitative data to explore the impact of changes. This section also includes an analysis of interviews with patients and carers who had not taken part in the co-design process but were patients on or visitors to the ward during or after the co-design. For greater insight we also explored whether or not these interviewees would report similar (perceived) changes to those reported by the patients, carers and staff who did take part in co-design.

Space

I mean the clock thing was something that really struck me. You know, you kind of assume that everyone has a watch or a phone or something so everyone knows the time but I guess, you know, people don’t and people of a different generation maybe don’t have a phone or whatever or you know.

Changes in the stroke unit ‘space’ included rethinking and repurposing existing spaces (all sites); repainting corridors and bays and creating and introducing artwork in the stroke unit (sites 1 and 3); reclaiming, repainting and equipping day rooms as spaces for activity and patient, family and volunteer use (sites 2, 3 and 4); personalising bed spaces (site 1); enabling greater access to kitchens and bathrooms; and introducing a tool to get to know patients as people: the ‘a little something about me’ board (site 2).

Different priorities did emerge. At site 1, for example, co-design groups prioritised creating the feel of ‘bringing the outside in’, and all of the bays were redecorated with a theme such as lavender or the seaside. Paintings or prints were put up; the walls were painted; digital clocks were put in each room visible to every patient; and hooks for visitors’ coats, photo-hangers and extra shelves to declutter patients’ bedside tables were added (Figure 9). Our interview and observation data support the positive impact that these changes had:

It is a picture of a beach and a cave, similar colours to the photograph from Cornwall that is on F’s wall. He [family member] also shows the photo to F and tells her this is very similar. She smiles.

He also uses his photo-hanger. There are get well cards and photos.

FIGURE 9 One of the new colour schemes in a four-bedded bay at site 1. Clinical items were put on new shelves behind each bed, local art work was put on the wall, and a photo-hanger was put up beside each bed. Reproduced with permission of Fiona Jones, University of London, personal communication, 2020. (Figure not reproduced here; see the original article.)

At site 2, the unit initially agreed that changes to the colour schemes and appearance of the unit should be made and worked with an artist to do this, but the changes were put on hold when estates raised concerns about costs, suitability and access. Corridors in the site remained dark and unwelcoming, in stark contrast to the bright, accessible and very well-used refurbished day room on this unit. Site 3 also worked with a local artist to redesign the colour scheme in the newly refurbished day room, and new communication about use of the space was developed. Site 4 already had light and bright corridors with some artwork in place, and the bedside curtains had local scenes and landmarks imprinted on the materials, so these areas were less of a priority for co-design groups.

All of the wards had emphasised creating a common space for patients and carers. Site 4 had a day room but it was used for wheelchair and seat storage. However, as the project began, the chairs were removed and the space become usable for individual and group activity. At site 4, a seating area (for four to six people) facing floor-to-ceiling windows looking out across the main car park and beyond to the hills surrounding the town was also repurposed. This area had previously been used mainly by staff taking their meal/rest breaks and for a display of Stroke Association materials. During the EBCD process, a magazine and newspaper stand was added, jigsaws were donated and several house plants were placed on an accessible window ledge. This space was used regularly by patients and carers both ad hoc or opportunistically and as part of planned therapy-linked activity, for example getting patients who liked gardening to play a part in watering and caring for the plants and encouraging patients with minor cognitive impairment to work on jigsaws with other patients (Figure 10):

We had a gentleman who was really disengaged, he wouldn’t really engage in therapy, but I gave him the job of watering the plants [in the window area] every day and he started doing that and apparently he did better in therapy after the engagement sessions. One of the physios said ‘I wasn’t up for this woolly hippy stuff [CREATE] that you lot are up to but, I could see how it worked, it worked really well’.

FIGURE 10 New repurposed seating area at site 4. Reproduced with permission of David Clarke, University of Leeds, personal communication, 2020. (Figure not reproduced here; see the original article.)

Another carer mentioned the value of having the window seating area as an alternative to the day room(s):

I think we did one or two [jigsaws] in there but people don’t all go in the dayroom you see, people tend to like the corners [the window seating area] there and people tend to join, you’ve got a nice view over the hills, you can look over the car park and you used to love sitting by the window and looking out and seeing what were going on.

The patients and carers at site 2 appreciated the newly refurbished day room. One patient, who had been on the unit before the room was available but had since seen it as part of a co-design group, felt that it would give patients a choice about what they could do and give them a change of scenery:

They [patients] can go there [to the day room] and there’s people there that will have had the same as they have had, and you can either talk to them or not talk to them . . . they can go to the dayroom and read a newspaper . . . or speak to somebody . . . but it’s just getting out of that bed space and moving. [Previously, when he was on the unit] we had a choice – sit in our bed space or walk to the end of the ward, or walk up and down and that was it.

The room was used for both scheduled and unscheduled activities. Staff reportedly encouraged patients to go in and participate in newly introduced groups. These groups included a regular Sunday lunch club run by volunteers and a breakfast club and a baking club run by the OTs. Other groups were also held there from time to time, such as music groups and bingo (Figure 11).

FIGURE 11 New social space (day room) used for scheduled and unscheduled activities at site 2. Reproduced with permission of David Clarke, University of Leeds, personal communication, 2020. (Figure not reproduced here; see the original article.)

At site 1, observations showed how patients used the new common area. This site previously had had no dedicated space for therapy-related independent practice or social interaction. Prompted by the co-design groups, the end of one corridor, which had been used to store equipment, was transformed into a new open seating area (Figures 12 and 13):

I moved through the ward to the new social corner where I met patient F. who was doing a puzzle. Someone had brought it in for him, it was not his own he told me. He told me he thought the ward had become more inclusive over the time he was there. And he said he liked the social space.

FIGURE 12 End of a corridor at site 1, previously used for storing chairs and hoists. Reproduced with permission of Fiona Jones, University of London, personal communication, 2020. (Figure not reproduced here; see the original article.)

FIGURE 13 New space at site 1, now an area for patients and families to meet and socialise. Reproduced with permission of Fiona Jones, University of London, personal communication, 2020. (Figure not reproduced here; see the original article.)

Redesigning the space also involved taking stock of what was available to support increasing patient activity, and opening up areas previously inaccessible to patients and visitors:

Yeah because there’s patients’ relatives, instead of coming and ask you now, they go there, they make a cup of tea, they can make a cup of tea for their relative so there’s no point in going to the kitchen because they’re not even allowed, or sometimes we’re busy, you say oh I’m busy, but now they just walk straight and do, yeah, and do that which is good.

At all four sites, redecoration involved decluttering and rethinking how the space could be used. At site 3, funds were raised to redecorate the shared day room and add a new kitchen area for making drinks and snacks. A local artist painted the walls to make the space look less like a staff meeting room.

At site 4, the co-design groups explored a range of options for making iPads (Apple Inc., Cupertino, CA, USA) more accessible; towards the end of the project, purpose-built flexible stands on wheels were purchased, which meant that iPads were secure but could be used both in therapy sessions and independently (Figure 14):

I think it’s great. Those rooms, that [ward D day] room is particularly good because, you know, when it was full of chairs we couldn’t even go in there. Now it’s, now it’s a free resource we’re lucky to have it, you know, and anything is better, yeah, I think it’s great. I think those pagers are great, the iPads, you know, they’re great on the stands, wonderful.

FIGURE 14 New mobile iPad stands used at the bedside at site 4. Reproduced with permission of David Clarke, University of Leeds, personal communication, 2020. (Figure not reproduced here; see the original article.)

The changes to how the space was used also had an impact on communication. Having familiar home items in the space to make it more homely and placing a photo-hanger by patients’ bedsides contributed to conversations and to learning about the patient as a person:

I said that it is impressive as a fireman he must have saved many lives, and as an electrician bring light to people’s lives. Patient in bed 3 says ‘and electric shocks’ and he laughs. We talk about their photos in the photo hanger. Patient in bed 4 has a holiday photo from Germany, a river, in his photo hanger. It is black and white. He said he went with his wife and it was a long time ago. Patient in bed 4 is very good with names, I realised this before. He knew the names of members of staff when man in bed 4 didn’t. He asked him again about someone’s name and he knew it. I tell him he is very good with names, the man in bed 3 agrees, says yes, isn’t he. I spent about 20 minutes with them. The patient in bed 3 asks me to pass him his blanket, it is red and cosy. He says it is very nice, he is a bit cold. His granddaughter brought it for him. His grandson used to play rugby and they went to watch him on the weekends. He was very good he says. They say thank you and say it is very nice of me to put the programme on. They don’t want a newspaper to look at whilst waiting for the rugby to start.

At site 2, a therapist commented on how the ‘a little something about me board’, which contained personal information about previous hobbies, activities or work that patients wanted to share with staff, had shaped the therapy activity she had decided to pursue with a patient:

I was planning some upper limb work with this chap but he was a little hard to motivate. I noticed that the board said was a keen domino player, so that’s what I suggested to him, we did the session focused around playing dominoes, something that he liked to do and we still did the upper limb work, so that’s something that’s happened because of the project for me.

The nature of the changes made to space meant that these were most visible and generated the majority of positive comments from patients and staff. We were also able observe the impact that the changes to spaces and the different ways space was being used had on patient activity. Staff members’ comments were orientated around the enabling impact of the changes, as well as the change in aesthetics, which gave a more homely, welcoming feel:

So I think the ward is different, I think definitely things look a lot better and a lot more inviting in the environment it’s engaging in, and I think that, that makes a difference in how we communicate on the ward and how we interact with that patient, because there’s stimulus to talk about, there’s pictures now on the wall to talk about and engage and help patients find their way in colours that help, signpost where they’re going and it’s something to talk about, a neutral that we can talk about that everyone has an opinion on and it doesn’t often require a lot of complex language as a speech therapist it’s nice to have something that people could gesture a thumbs up or a thumbs down, they can nod or shake their heads to say that they like a painting or a photograph. So in that respect I think that’s really helped how the ward, the ward interacts and changes.

Activities

Changes initiated by activity co-design groups included regular visits from therapy dogs (sites 1 and 2); new activity boxes for every bay (sites 1 and 3); increased structured group and individual therapeutic activities (all sites); and increased volunteer and external group supported activities (music, gardening, massage, pet therapy, art and crafts, social Sunday lunches) (all sites) (Figure 15):

We have huge gaps in the day where your patient’s doing nothing, they’re bored, they become institutionalised, so with these extras, like your volunteers coming in, you’ve got various groups, you’ve got your cooking group, your breakfast club, your lunch club, it just makes for a, well it’s a more positive experience isn’t it, well I feel it is.

FIGURE 15 New activities (art groups and therapy dogs) available at site 1. Reproduced with permission of Fiona Jones, University of London, personal communication, 2020. (Figure not reproduced here; see the original article.)

The co-design process often led to discussions about how to connect with community and voluntary groups to a greater extent. At site 1, for example, a family member of one of the patients worked as a local artist. Together with her colleagues, she applied for funding and organised art workshops on the ward at the weekends. At site 2, the patient experience office worked with the volunteer co-ordinator to train new volunteers to provide mealtime and social lunch support following discussion with the co-design group members. In addition, the local university’s music department was contacted, which led to two or three singers coming to the unit every 2 weeks to sing songs both they and the patients had chosen. At sites 3 and 4, co-design groups, aware of some of the activities that had been developed at sites 1 and 2, utilised this learning much earlier by connecting with community organisations, and at site 3 a programme of concerts by local choral and music groups was organised. Observations showed the impact of these activities:

One of the ladies told me she had to cry as she found it really moving and said in a good way. She also walked past the group again later and said that it was really nice and she smiled.

Additionally at site 3, a retired PT and ex-patient had begun to hold fortnightly ‘exercise groups’. The feedback she had received was positive and our observations showed that the groups had almost got ‘too popular’. Many patients wanted to attend but there was limited space available, even in the repurposed day room. At site 4, a new breakfast group was set up and by the end of the co-design group activity it was running 7 days per week. This group was led by therapists and supported by nurses, who continued the activity at weekends when therapists were not present. The co-design group initiated other planned (arts and crafts groups) and ad hoc events (open afternoon, evening meal with other stroke survivors, a small group watching a favoured television programme) and some patients and families noticed the impact of these:

I do see evidence, psychological evidence [of being willing to engage in more activity] . . . and I also think patients are doing more now.

Despite an increase in the opportunity to access activities, our observations showed that there were still long periods of time when patients were inactive. Opportunities to access activities were lost for patients who needed help to get out of bed:

I think certain members of staff are really great at getting patients out of bed and really proactive with that and will look at what patients are signed up to do, whether it be therapy sessions or whether it be something like the reading group which has gone on timetables. And then there are other members of staff who don’t look at the timetables at all.

Co-design groups wanted to focus attention on those patients who were unable to get out of bed and attend groups by designing and providing new activity boxes that could be brought to patients’ bedsides, but we rarely saw anyone using these:

Patient in bed 8 is alone. She is lying in bed. She is not very mobile, but there is nothing else e.g. radio, TV for her. She always smiles when she sees me walking past. The patient, gentleman, 92, is sitting alone. He is saying ‘no no no’ as he quite often does. The doctors are standing around the nurses’ station. 2 of them are looking on their phones, talking about something else, work unrelated.

However, observations and interview data also supported the notion that a small number of patients did not necessarily want to engage even when increased activity opportunities were provided. Their pre-stroke behaviour, activity and leisure preferences were factors in their willingness to participate in what was offered:

There was various things going on that the volunteers organised. There was afternoon teas and stuff like that but you’d no desire to go had you?

Members of staff reported that they were more aware about the activities on the wards post implementation. This was also the case for members of staff who did not take part in the co-design groups:

There, everything, everything is there, everything is there, I’ll go there and give it to them if they are sitting down and there’s nothing to do, give them something to do, even if it is just to practice writing their name or practice signing, I’ll give them a piece of paper and pen so that they can practice and write. [. . .] Yeah, sometimes you think, you can see they want to write, do you want to write, you see them, sort of, practising, oh I don’t know how to write my name again, OK just practice, just keep like that. [. . .] Well before we, yeah now we are doing it but before, as I said, before we were not, no, we are not doing all these sort of things so, but now I don’t know because of the CREATE, the awareness, we’ve got the awareness now so we are trying to encourage them to do all these things.

In addition to being more aware of the need to promote activity, members of staff reported some examples of how their approach to activity had changed:

Well, we just, we’ve got to get the people out [of bed], washed and everything like that, and then we bring them, or they bring them in here [day room] and help with, anybody that’s available, really, help with breakfast . . . because we still do it on a weekend, even when therapy staff aren’t here, the nursing team still do it.

Communication

Co-design groups focused on a variety of ways to improve communication, and the changes had led to training for staff; redesigned ward information with emphasis on bringing items from home to personalise the wards; information on activities and how to access WiFi; common charters of communication agreed by staff – personal and social interaction (London sites); and a new website (London, site 3).

Sites 3 and 4 used many of the ideas developed sites 1 and 2, for example digital clocks, familiar home items list, and a list of activity-related apps that could be used on smartphones or tablets. At site 4 there was interest in adopting the ‘something about me’ board developed at site 2 and increasing volunteer activity. However, the board was not actually taken up by the co-design groups and volunteer activity could not be progressed as the volunteer co-ordinator on site would not engage with the groups. At this site increased volunteer support for and provision of activities occurred following personal contacts between individual members of staff and existing volunteers and volunteer groups. Many co-design groups prioritised updating and expanding the unit information leaflets to highlight how to access free WiFi on the unit and to highlight the activities now provided on and off the unit and how these could be accessed. At site 4, co-design groups also addressed a major concern for all stroke survivors that in leaving the ward for social activity that they may miss important therapy time. The groups sourced telephone ‘bleep’ devices no longer used by medical staff and gave these to patients and carers when leaving the unit so that they could be called back for therapy sessions if required.

Patients and families taking part in the co-design and those on the ward noted the improvements:

I think the information leaflet’s quite good because it says, it tells you things like where the day room is and that you can go into the garden and things like that. That new one, it’s also on the wall outside the ward isn’t it, the new one? Because I remember reading it there and I thought oh this is different from the one that used to be there when Mary was in.

One patient and their carer had been involved in the communication co-design group and were pleased that the ‘my journey’ communication sheets were being used on the ward at site 2. These sheets were for staff to write down information about the patient’s progress and for relatives to leave questions and comments for staff:

No, very happy [that the communication sheets were being used] because some days I’d try and ask [her husband] ‘oh what have you done today’? And sometimes he’d find it hard because when you went in later in the day they were tired and if they’ve had physio and everything else and half the time he’d just sleep, but whereas if you could pick something up and say ‘oh yeah, you did some therapy and how did that go today?’ . . . You can at least make conversation to him about it, without him saying ‘I can’t remember’ . . . Just nice to know he hasn’t just been lying in bed or sat in the chair and just ignored . . . at least if you could read it and say ‘that’s good well done’.

We also observed ways in which the changes to the space, such as the ‘a little something about me’ boards and the refurbished day room at site 2, had a positive impact on interaction and communication. Patients thought that the boards were useful and one patient remarked that they served to ‘break the ice’ with staff who did not know him. Relatives were observed helping staff fill the boards in or even filling them in themselves. Volunteers and other staff who were not based on the unit (e.g. phlebotomists) also appreciated the information provided on the boards and reported that it helped them find topics of conversation to bring up with patients. One phlebotomist said that it was helpful to know something about each patient and ‘see them more as individuals’. Sometimes the boards were not up to date because of bed moves. A health-care assistant reported that she and ‘a few others’ tried very hard to keep the boards up to date but that this could be difficult during busy periods. She also said that not everyone ‘bothers’ to do it, even if they had time.

At site 4, several changes were made to information for patients and visitors about how they could access different areas or be more engaged in their recovery. Staff felt that this was a tangible output that had made a big difference:

So you know the welcome to the ward sheets, we edited them so there’s one for every single ward now, and I know they get handed out at family meetings, I know on [site 4] they do that, the others need encouraging to do it, but there definitely are printed forms for every single ward. They had on what facilities were available, so we changed it to make it a bit more user friendly, it had like a spider diagram that said, there is WiFi, there is a garden, you can go to Costa, and it kind of touched on all the things that we spoke about [in the co-design group], so, actually, families can look at that straight away.

As part of the co-design process at site 3, information leaflets were produced for key workers to give to patients and family members, and the Stroke Association commenced monthly information drop-in sessions. Although those changes may well have facilitated the change in perception about feeling informed, it was difficult to keep track of and report the extent to which those changes were actually being implemented. As notes from an observation at site 3 show, even when new information leaflets and guidance about welcoming patients were put in place, not every patient was made aware of these:

One of the patients in particular is very interested in the project and wants to know more. He asks me to send him an email with some readings so I do. He has never seen any of the leaflets, Welcome to the Ward, Apps list etc. I give it to him. He later tells me he enjoyed seeing the Familiar home items list, but if it wasn’t for me he would have never seen any of it.

Staff from across all sites felt that CREATE had raised awareness that activity was a priority and in some cases said that as a result they had made small changes in how they communicated or interacted with patients. This impression was given by both staff who did and staff who did not take part in the co-design groups:

I think staff are thinking more about how to involve patients and get them talking and for myself I always go in now and make sure that I say hello to everyone, regardless, or introduce myself even if I’m not seeing that particular patient.

One member of staff (site 3, nurse), who did not take part in the co-design groups, described how he now thought more about the patient as a person: ‘this is a person who needs to go out like myself’. He was more open with patients and involved family more, whereas before it had been ‘wash – dress – sit’. He shared an anecdote about two patients who did not like having the same breakfast every day and wanted to go to the canteen to get their own breakfast. Someone said he should take their orders and get it for them, but the patients wanted to go and choose for themselves. He then thought ‘yes, we keep them from making their own choices otherwise’. He took them to the canteen to get a breakfast takeaway, and they have done this every day since. He said that this had happened because of the CREATE study taking place on the unit, and now he would be thinking about alternatives.

Although staff often said that they were enabling in their interactions and communications, this was not always supported by our observations and was especially obvious between routine tasks as discussed previously:

The HCA [health-care assistant] comes back to the gentleman in bed 7. She talks to him. He says ‘sorry’, but I [observer] can’t understand what exactly he says. It seems to me he is apologising to her that she has to do certain things to help him. She goes back to bay 1-4 and talks to the patients in the room, she then grabs continence pads and tissues and goes back to patient in room 7. He says sorry again. She does her job, but doesn’t say anything like, its ok etc. She puts on apron and gloves. She goes in and closes the door.

Although staff reported positive changes to their communication and interactions to promote activity, some supported our impressions that communication was ‘work in progress’:

I am not sure if we have achieved much in terms of the basic communication day to day between the carers of the ward. Not the carers, the staff caring for the patients, because I think that’s where we need an ongoing input. So, that’s I think a big part, that’s difficult to change and I’m not sure if it will change?

Impact of changes: quantitative findings (behavioural mapping and patient-reported outcome measures/patient-reported experience measures)

In this section, we present descriptive results from the quantitative data generated pre and post implementation of EBCD (the intervention): (1) behavioural mapping and (2) PREMs and PROMs. In addition, the results are compared within and between sites.

In summary, behavioural mapping results were inconsistent across sites and showed a mixed pattern of inactivity (social, cognitive and physical) pre and post implementation of co-designed changes. However, taking the broad measure of ‘any activity’, there was some improvement at sites 1 and 2 (full EBCD) but similar levels at both time points at sites 3 and 4 (accelerated EBCD). Our findings also demonstrate the discrepancies that arise when presenting overall activity/inactivity percentages, compared with the number of activities in the three domains (physical, cognitive and social).

Data analysis of PREMs/PROMs also showed a mixed picture across all sites. PREM data analysis showed inconsistencies across items and sites. Despite this, there were consistent changes in relation to item 5, which asked respondents whether ‘there were enough things to do in my free time’, which showed an improvement after the implementation of co-designed activities. No further firm conclusions from PREM data can be drawn. PROM data analysis confirmed that patients treated in the participating stroke units before and during EBCD implementation were not atypical and did not appear to be any more or less likely to be receptive to activity-based interventions than other groups of inpatient stroke survivors in other UK stroke units.

We make further commentary on our chosen quantitative measures, behavioural mapping and PREMs/PROMs, in our methodological consideration in Chapter 6.

Behavioural mapping

As outlined in Chapter 3, behavioural mapping was used to provide an indication of activity levels in each unit on 3 separate days at agreed time points before and after EBCD cycle implementation. The approach used in CREATE recorded episodes of social, cognitive or physical activity. These data were from separate groups of patients in each site; we did not seek to compare ‘before and after’ scores for individual patients but rather used the behavioural mapping data as a broad indicator of activity levels in each site. All patients recruited for behavioural mapping were observed at 10-minute intervals between 08.00 and 17.00 or between 13.00 and 20.00 on 3 separate days, including weekends. This allowed up to 60 observations of each patient per day. During each 10-minute interval, the data for each patient was based on an observation made by the researcher over a period no longer than 5 seconds (see Appendix 7 for the full behavioural mapping protocol).

Of note is that the percentages stated in tables relate to ‘inactivity’, so a reduction is a positive finding.

Within-site comparisons

Site 1 (pre, n = 9; post, n = 7)

Pre implementation, participants in the behavioural mapping at site 1 had a high level of physical inactivity, at 71%; levels of inactivity in the cognitive and the social domains were also high. However, level of inactivity reduced markedly in the physical domain between the two time points, from 71% to 42%; there was a more modest reduction in the level of social activity of 6% and a slight increase in cognitive inactivity of 4% (Table 6).

The only physical activity recorded pre intervention for > 1% of the observations was engagement in activities of daily living (ADL) (which could include personal grooming, changing clothing, eating and drinking), at 7%. Changes were evident post intervention, with a 20% increase in engagement with ADLs and evidence of increases in the following activities: mobilising (7%), repositioning (5%), independent practice (usually of therapy-directed tasks) (4%) and transfers (2%). Although cognitive inactivity increased slightly overall, again, the post-implementation observations showed that some cognitive activities were now occurring, including writing and art activity, listening to music or audiobooks (5%) and reading (2%). There was a small reduction in social inactivity (6%), which seems to be accounted for by an increase in talking and communicating with other patients, relatives and staff.

Post intervention, overall level of any activity increased by 15.1% at site 1, despite inactivity being higher in the cognitive domain (Table 7).

Site 2 (pre, n = 10; post, n = 10)

As at site 1, levels of inactivity in all three domains at site 2 were disappointing, with only cognitive inactivity showing a reduction between the two time points (13%) and increases of 1–2% seen in levels of inactivity in the physical and social domains (Table 8).

The most common activities recorded pre intervention were talking with other patients, relatives or staff (35%), engaging in ADL (which could include personal grooming, changing clothing, eating and drinking) (9%), watching television (8%) and reading (6%). Post intervention, some changes were evident, mainly in cognitive activity, with small increases in reading (14%), listening to music/using electronic devices (8%) and participating in group activities, including art and writing (6%). Despite the small decrease in social activity observed, there was in fact a small rise in the percentage of time spent talking/communicating among participants at this site (37% from 35%). When the overall levels of any activity and inactivity recorded are considered for this site, these data show a small increase in activity of 15.8% (Table 9).

Site 3 (pre, n = 12; post, n = 7)

Pre implementation, both physical and cognitive inactivity levels at site 3 were better than at sites 1 and 2 and similar to site 4. This site had lower levels of social inactivity than sites 1, 2 and 4. Between the two time points, levels of social activity remained the same, at just over one-third of the observed participant time. The level of cognitive inactivity fell by 5%, whereas the level of physical inactivity rose by 6% (Table 10).

In common with sites 2 and 4, the most common activity at site 3 pre implementation was talking with other patients, relatives and staff (27%). Use of mobile phones accounted for 10% of observed participant time in social activity. Engaging with ADL accounted for 13% and mobility accounted for 3% of physical activity. Reading was the most common cognitive activity, at 10%, but patients also engaged with music or audiobooks 5% of the time, watched television 4% of the time and used electronic devices 3% of the time. Post-intervention talking with other patients, relatives and staff was unchanged (27%). There were small increases in using mobile phones (up 3%) and electronic devices (7% increase); participating in group activity was evident during 4% of observed time, as was mobilisation at 4%, and listening to music/audiobooks (up by 1%). Engaging with ADL fell slightly to 11% and reading fell to 9%, whereas watching television remained the same. At this site, overall activity levels changed very little (up 2.9%) but it had started with the highest baseline activity level of 82.4% (Table 11).

Site 4 (pre, n = 6; post, n = 6)

Here, too, the data show relatively high levels of inactivity in the three domains, with only social inactivity reducing by 10% between the two time points. As at site 2, post intervention there were small increases in inactivity in the physical domain and, at this site, also in the cognitive domain (Table 12).

The most common activities engaged in pre intervention at site 4 were talking with other patients, relatives or staff (25%), watching television (22%), engaging in ADL (14%), using electronic devices (4%) and reading (3%). Post intervention, some changes were also evident at this site, most notably in talking with other patients/relatives or staff, which increased by 12%, and reading, which increased by 10%. There was a reduction in television viewing and in engaging with ADL. The trend evident at the other sites of changes in overall level of any activity was also seen at site 4, with an increase of 11.5% (Table 13).

Cross-site comparisons

Pre implementation, site 1 had the highest level of physical inactivity and site 3 had the lowest level of social inactivity. However, overall levels of inactivity across the three domains were broadly similar across the sites. The largest change in levels of inactivity occurred at site 1, where physical inactivity levels fell by 29%. At site 2, levels of cognitive inactivity fell by 13%, mostly accounted for by an increase in participants reading independently or with others. Social inactivity fell by 10% at site 4, accounted for by an increase in talking with other patients, relatives and staff. When comparing overall activity, site 1 had the lowest level pre implementation, at 62.1%, and levels at sites 2 and 4 were 72.7% and 80.6%, respectively, with site 3 having the highest overall level of activity, at 82.4%. Post implementation, all sites increased overall levels of activity, with sites 1 and 2 showing increases of 15%, site 4 showing an increase of 11% and site 3 showing an increase of 3% from its relatively high baseline level.

Summary

Although our behavioural mapping data are largely consistent with previously published work in terms of levels of inactivity, they do suggest that the overall time spent inactive on these stroke units was not as much as that seen in previous studies.7,86 The behavioural mapping data, although from a small number of participants at each site, highlighted the differences in the provision of individually focused activity opportunities (which remained limited) and the changes we saw in the space/environment of the units, which led to an increase in structured group and informal social activities not easily captured by behavioural mapping. In addition, the behavioural mapping process highlighted the long periods of time that patients spent largely alone, despite the changes to space and activity opportunities, and was one of the main ways in which we observed the limited and task-oriented staff communication processes that largely did not change in the sites between the pre- and post-EBCD periods.

Overall, we question the relevance and validity of using behavioural mapping as a reporting tool in CREATE in relation to anomalies in reporting; the small number of patients who were able to consent the day before behavioural mapping took place; not including scheduled therapy as part of our reporting; and finally the impact of contextual issues such as staff shortages and the severity of disability of the inpatient caseload. These methodological considerations are discussed in more detail in Chapter 6.

Patient-reported experience measures and patient-reported outcome measures

Participants were sent a PROM/PREM pack containing a survey about their experiences as well as the impact of their stroke. Demographic details of all respondents can be found in Appendix 9.

We provide a summary of pre-implementation and post-implementation PREM and PROM data below and the full data set can be found in Appendices 10–12. First, we report on PREM items that scored particularly high or low, and on broad indicators of the quality of the patients’ experiences in these units. As these data are from different cohorts of stroke survivors at each time point and in each site, they provide an indication of patients’ overall experiences of these stroke services during their inpatient stay.

For reference, Box 1 shows the items referred to in the following commentary and Table 14 shows the PROM/PREM response rates.

1. When I arrived I was given information about the unit and what would happen during my stay.
2. The facilities on the unit were good.
3. There was somewhere secure to keep my belongings.
4. I was able to discuss personal matters in private.
5. There were enough things to do in my free time.
6. There was a friendly atmosphere in the unit.
7. I felt the staff really cared about me.
8. The staff worked well as a team.
9. I felt able to talk to the staff about any problems I had.
10. I was asked what I wanted to achieve during my stay.
11. I felt as though the staff and I were partners in the whole process of care.
12. The staff kept me informed every step of the way.
13. My family or carer was involved in discussions about my treatment if I wanted them to be.
14. I am happy about the amount of therapy I received for . . .
    1. Swallowing problems
    2. Speech and communication
    3. Improving mobility
    4. Independent living
    5. Continence
    6. Other
15. I received enough emotional support.
16. I felt well supported and prepared for my discharge.
17. I am satisfied with the progress I made during my stay.
18. Thinking about how you were treated during your therapy sessions, which of these statements would best describe how you felt: patronised, respected, offended, supported, ignored, listened to, frustrated, motivated?

Pre-implementation patient-reported experience measure data: summary of findings

In terms of the specific focus on the CREATE study, responses to statement 5 were most relevant. Pre implementation, between 30% and 55% of respondents disagreed that there were enough things for them to do in their free time, with only 4% of respondents indicating agreement with the statement at site 1, 27% agreeing at sites 2 and 4 and 44% agreeing at site 3. By contrast, across all sites, ‘facilities’ were considered to be good, with between 74% and 82% agreeing with statement 2.

Across all units, some dissatisfaction was evident in areas of information provision, with < 50% agreement with statement 1 (range 36–48%). Similarly, there were low levels of agreement with statement 10 (range 27–48%), suggesting that limited attention was given to identifying personal goals for recovery. At the same time, these respondents had a higher frequency of agreement with statements 11–13, which focused on feeling as though they and the staff were partners in the process of care (statement 11: range 50–75%), that they had been kept informed during the care process (statement 12: range 41–61%) and that family or carer involvement in treatment discussions could occur if wanted (statement 13: range 67–82%).

There was strong agreement with the statements that there was a friendly atmosphere on the units (statement 6: range 82–91%) and that staff really cared about them (statement 7: range 68–87%), and there were similar levels of agreement that the staff worked well as teams (statement 8: range 67–85%). The respondents largely agreed with statement 9, that they felt that they could talk to staff if they had a problem (range 60–80%). Although the responses to statements 7–9 were positive, up to one-third of the participants disagreed with these, so experiences clearly varied within and across units.

The responses to statement 4 were more mixed, indicating some concerns related to feeling able to discuss personal matters in private (range 51–67%). This reflects to some extent the limited number of private rooms/spaces on the stroke units and echoes concerns raised by former patients in the filmed interviews at sites 1 and 2. At site 3, only 4% of respondents agreed that they received enough emotional support (statement 15), with 33% disagreeing. However, 59% of the respondents at this site indicated that this issue was not applicable to them; allowing for this, the responses still indicated a low level of agreement that enough emotional support was received. At site 1 only 35% agreed with this statement, whereas at sites 2 and 4 more than two-thirds of participants indicated that they had received enough emotional support (statement 15: both 67%).

In terms of respondents feeling happy with the amount of therapy or treatment received in relation to the five areas listed in statement 14, the main area of concern across sites was linked to problems related to continence. Here, agreement levels ranged from 26% at site 3 (although 41% indicated that this was not applicable to them) to 42% at site 2, 42% at site 4 and 57% at site 1. Site 3 had lower levels of agreement for swallowing, speech and mobility, at 33%, 48% and 52%, respectively. These differ quite markedly from those at sites 2 and 4, where levels of agreement relating to swallowing, speech and mobility ranged between 85% and 95%. At site 1, the levels of agreement ranged from 61% for therapy or treatment for swallowing to 80% for both speech and mobility. In respect of treatment or therapy to prepare people for independent living, levels of agreement ranged from 44% at site 3 to 56% at site 1 and 82% at sites 2 and 4, where, with the exception of continence, higher levels of satisfaction were also evident with other areas of therapy or treatment.

Post-implementation patient-reported experience measure data: summary of findings

The post-implementation response rates at sites 1 and 2 were higher than those at sites 3 and 4; the reasons for this were unclear, but the timing of the survey could be one factor, at sites 3 and 4 it was distributed close to the Christmas period. The response rates are detailed in Table 14; in all sites and at each time point when these were distributed, the response rate was < 40%.

We report here on the differences in scoring of PREM items between pre and post intervention. We cannot attribute these differences in responses directly to changes occurring in the stroke units during the EBCD process, but we report on the responses as broad indicators of the quality of the patients’ experiences in these units. The full data set can be found in Appendices 8 and 9.

Among the post-intervention responses, agreement with statement 5 increased at three sites; this was most marked at sites 1 and 2, where the proportion agreeing increased from 4% to 40% at site 1 and from 27% to 50% at site 2. At site 4, the increase was more modest, to 36% from 27%, and, despite these increases, the scores still indicate relatively low levels of opportunity to be independently active. However, the level of agreement with statement 5 at site 3 actually fell quite substantially, from 44% to 9%. These responses are not easy to explain and differ quite substantially from the qualitative data, which record that a range of increased activity opportunities were made available and occurred at this site.

As with the pre-implementation responses to statement 2, there were still quite high levels of agreement that the facilities on the sites were good, but, again, it is interesting to note that although the proportion agreeing with this statement increased at sites 2 and 3 by 4% and 8%, respectively, the level of agreement at sites 1 and 4 fell by 10% and 12%, respectively. In terms of information provision, an area of concern in the pre-implementation period, responses to statement 1 also varied post intervention, with sites 2 and 3 showing a reduction in agreement with this statement of 8% and 10%, respectively, and sites 1 and 4 showing no change.

Co-design groups at all sites had identified communication with patients and families (rather than information provision specifically) as an area for improvement. The post-implementation responses show that providing information early in the inpatient stay remains an area of concern for patients. At sites 3 and 4 the changes introduced (new unit information leaflets) came towards the end of the co-design period and may have occurred too late to be captured in these questionnaires, but at sites 1 and 2 communication-focused changes had been implemented for longer periods of time. At the same time, another statement related to information provision (12: ‘the staff kept me informed every step of the way’) showed improved levels of agreement at each site (range 56–64%, up from 41–61%).

Responses to statement 10 indicated higher levels of agreement at three out of the four sites (i.e. sites 1, 2 and 4) (range 40–64%, up from 27–48%), most notably showing a substantial increase at site 4, from 27% to 64%. At site 3, the level of agreement fell by 8%, from 48% to 40%. So, while acknowledging the lower overall response rates post implementation, some improvement was evident, but responses to these items in the PREM indicate that progression to a more patient-centred approach remains an issue for these sites.

Responses to statement 4 show higher levels of agreement in three out of the four sites (i.e. sites 2–4), with percentage increases of between 4% and 12% (range 55–73%). Site 1 had a 7% reduction to 60% agreement from 67%, so finding somewhere private to discuss personal matters was a concern for some patients. Statement 9 addresses broadly the same area of patients being able to talk to staff about any problems; agreement with this statement fell in three sites out of four from the relatively high pre-implementation levels, which ranged from 62% to 80% to 52% to 76% post implementation, a reduction of between 4% and 10%. Responses to statement 15, which explores receipt of emotional support, showed an increase in agreement in three out of the four sites (i.e. sites 1–3) (range 44–67%), with a substantial improvement at site 3 (up from 4% to 44% and with only 8% indicating that this was not applicable post implementation at this site, down from 59% pre implementation). At the same time as highlighting these areas where services still had room for improvement in communication and support, respondents indicated high levels of agreement with statement 6 (range 80–100% agreement), that there was a friendly atmosphere on the unit, statement 7, that patients felt that staff really cared about them (range 76–93%), and statement 8, that the staff worked well as a team (range 80–97%). Statement 17, relating to patients being satisfied with their progress during the inpatient stay, had higher levels of agreement at all four sites (range 76–86%). Statement 16, addressing patients feeling supported and prepared for discharge, showed small increases in agreement at sites 1 and 2 (to 56% from 54% and to 86% from 81%), whereas site 3 showed a small reduction, from 59% to 56%. The reduction in levels of agreement with statement 16 at site 4 is more marked, with this down 17% from 81% to 61%.

Responses to the five areas listed in statement 14 (respondents feeling happy with the amount of therapy or treatment received) present a mixed picture, as in the pre-implementation period. These post-implementation responses are from a smaller number of patients at all but site 1 and will reflect to some degree the individual post-stroke disabilities and impairments experienced by these respondents. At site 1 there were lower levels of satisfaction with therapy for swallowing problems (20% vs. 67%), but 64% of respondents indicated that this issue was not applicable to them and a further 22% indicated that they were unsure. A similar substantial reduction is seen in responses to the speech and communication statement at site 1 (36% vs. 80%) but again 48% indicated that this issue was not applicable. In addition, quite a large reduction in satisfaction is seen for continence (57% to 28%), but 52% indicated that this was not applicable, and the level of disagreement with this statement among patients at this site fell from 36% pre implementation to 12% post implementation. Improving mobility also showed a reduction but to a lesser extent (68% vs. 80%).

The overall PREM data analysis presented above shows a mixed picture across all sites, with inconsistencies across items and sites. Response rate fell significantly at site 3, which might have been related to the timing of data collection, either side of the Christmas period. Other than the changes in relation to item 5, which we would have expected following the implementation of co-designed activities, it is difficult to draw any firm conclusions from this relatively small sample. We make further commentary on our chosen PREM in our methodological consideration in Chapter 6.

Patient-reported outcome measures

The PROM questionnaire comprised the Subjective Index of Physical and Social Outcome, the EQ-5D and the Oxford Handicap Scale. We did not intend the PROM to be used to derive scores for each of the three subcomponents, but rather to review and describe participants’ perceived levels of disability and adjustment to life after stroke. As previously indicated, these self-completion questionnaires were distributed to two different cohorts of stroke patients, and returned by post. Those who returned questionnaires may not be representative of the stroke patients who received care in the participating stroke units during the time periods identified. Nonetheless, these data provide some insight into the post-stroke characteristics of patients who returned the questionnaires. Below, we describe the broad issues emerging from these data.

Questions 1–10 were from the Subjective Index of Physical and Social Outcome and were about physical functioning/mobility. All respondents had similar difficulties with dressing in the pre- and post-EBCD intervention cohorts; most reported some difficulty with dressing themselves, and on average 17% of respondents could not dress themselves fully post stroke. Across sites, up to 25% of respondents reported either a lot of difficulty with moving around their homes or not being able to move around their homes independently at all. Among the post-EBCD intervention cohorts, respondents at sites 1, 2 and 4 felt marginally more satisfied overall with their ability to perform daily activities around the home than did those at site 3.

With the exception of site 1, where rates were higher in the pre-intervention cohort, those participants responding in the post-EBCD periods reported that they had a lot of difficulty or could not shop for or carry a few items for themselves, indicating relatively high levels of dependency in this area. For question 5, those indicating they were completely dependent or needed assistance most of the time increased in the post-EBCD cohorts at sites 2, 3 and 4, but fell at site 1 from 58% to 32%.

Questions 6–10 related to social/emotional functioning. Except at site 2, where the percentage fell by 9%, response rates indicated that more people in the post-EBCD cohorts were never bored. In all but site 4, there was a small increase in the numbers of respondents reporting a good deal or quite a lot of communication with friends or associates. At site 4 there was a reduction of 30% in the response rate for these two options in the post-EBCD cohort. Again, in relation to levels of satisfaction with activities and interests, respondents were completely or mostly satisfied, whereas respondents at site 4 reported lower levels of satisfaction with responses to these two options in the post-EBCD cohort (down by 30%). There were no differences in response rates to the item relating to visiting friends in either the pre- or post-EBCD cohort.

Our data showed that, across sites, between one-third and half of respondents felt self-conscious to some degree about their appearance, with up to 11% reporting that they tried to avoid going out in public, which is a significant proportion. At the same time, between 45% and 60% of the remaining respondents indicated that they were perfectly happy with their appearance when out in public. There were no major differences in responses between sites in the pre- or post-EBCD cohort.

Questions 11–15 are from the EQ-5D. In this PROM we did not ask participants to rate their perceived level of severity in each of the five domains and so standard scoring is not applicable. We describe below the overall levels of agreement with the statements.

In all sites, > 46% of respondents reported some problems with walking, and between 5% and 18% reported being confined to bed. A similar pattern was evident for washing and dressing, with between 40% and 73% of respondents in the pre-intervention cohort at sites 1–3 reporting having some problems with washing and dressing or being unable to wash or dress themselves. In addition, a high proportion of respondents across all sites reported some problems performing usual activities or being unable to perform usual activities (range 55–84%), with some variation between sites. More than one-third of respondents experienced moderate pain or discomfort, with between 5% and 22% reporting extreme pain or discomfort. Responses also showed that 31% and 46% reported being moderately anxious or depressed. Rates differed slightly between the pre- and post-EBCD cohorts across sites, with a reduction in respondents selecting this response at site 1, no change at site 2 and increases of 5% and 13% at sites 3 and 4, respectively. Rates of those not reporting being anxious or depressed varied by site from 36% to 62%, but there was no consistent pattern between pre- and post-EBCD response rates.

Questions 16–21 are from the Oxford Handicap Scale. The first three items identified little or no impact of disability and impairment on everyday life and the remaining three items indicated moderate to severe impact on everyday life.

Overall, the responses indicate that between 55% and 89% of the respondents had no symptoms, symptoms that did not interfere with everyday life or symptoms that had caused some changes in the respondents’ lives but the respondent was still able to look after themselves. This level of response was largely similar in the pre- and post-EBCD cohorts at sites 1–3 but increased from 77% to 89% at site 4. However, responses to the remaining three items in this section of the questionnaire (questions 19–21) indicated that between 9% (site 3 pre- and site 4 post-EBCD cohorts) and 45% (site 2 pre-EBCD cohort) had moderate to severe symptoms that significantly changed their lives and resulted in their needing help to care for themselves. At the most dependent end of this spectrum, between 3% (site 2) and 27% (site 3) of participants in the post-EBCD cohorts at these sites reported needing constant attention day and night.

Commentary on patient-reported outcome measures

Acknowledging the limitations of the sample size and response rates across sites, these self-report data are largely consistent with the levels of mood disturbance reported in systematic reviews,21,75 and disability and impairment reported by the Stroke Association in the UK and in stroke data internationally.17 There does not appear to be any great difference in levels of disability and impairment between sites and between the pre- and post-EBCD cohorts. There is, however, some indication that respondents in the post-EBCD cohort at site 3 reported higher levels of post-stroke impairment and disability than at other sites; however, the response rate is < 20% for this cohort. Similarly, any signals that respondents in the pre- and post-EBCD cohorts at site 4 reported less impairment and disability and lower levels of mood disturbance have to be viewed in the light of the very low response rate of only 12%. Overall, these data indicate that stroke patients treated in the participating stroke units before and during implementation of the EBCD intervention were not atypical and do not appear to be any more or less likely to have been receptive to activity-based interventions than other groups of inpatient stroke survivors in other UK stroke units.

Our target recruitment number was based on the anticipated numbers of patients likely to be cared for and discharged from the stroke unit in the 3 months prior to the commencement of the EBCD process and then cared for and discharged in the 3-month period during which EBCD activity took place in each unit. We recognised that this number was no more than a crude estimation influenced by external factors such as caseload factors and admission rate. As it quickly became apparent that response rates were low in almost all sites, we extended the period of time for distributing pre and post questionnaires to 6 months. The differences in the numbers of patients sent questionnaires reflects the number discharged from the units during each time point and known to be alive at that time. This partly reflects length of stay, which was often closely linked to stroke severity.

We had no direct contact with these potential participants, but at site 2 (Yorkshire) the consultant physicians encouraged their patients to complete PROM/PREM questionnaires at their 6-week follow-up appointment, which may have increased the pre-EBCD response rate at that site. We did not send repeat questionnaires at this site as the response rates were good; however, the consultants changed at that site so no similar process was in place for post-EBCD questionnaires. We requested that principal investigators use this method of completion at follow-up clinics at site 1 but they did not do so because of concern about participant burden. At site 3 there were no 6-week follow-up appointments and so, if no response had been received to the questionnaire, second copies were sent out 4 weeks later (after checking whether non-respondents were still alive). We judge that the difficulties with recruitment to and completion of the measures, as outlined above, influenced the validity and the contribution of the PROM/PREM data to our overall findings.

Implementation and engagement

The following section describes the experiences of staff, patients and families who took part in the co-design, and synthesises findings from across all post-implementation interviews. For the most part, taking part was viewed positively, even though patients and families had been approached about taking part at a time of great personal disruption in their lives and staff had little or no additional time to dedicate to the process:

I think CREATE has been an absolute breath of fresh air for the stroke unit and I think all the challenges that we still continue to face on the ward have been bearable because of CREATE and the positive influences that they have had, so . . .

The ‘experience’ of taking part (staff, patients and families)

The themes below demonstrate our understanding of some of the more nuanced reasons why participants wanted to take part and found the whole experience positive. The themes demonstrate why they wanted to be involved and the benefits to individuals, as well as the perceptions of the advantages of working together in a group and seeing through the changes.

The added value of giving something back

Family members and patients appeared to be driven to be involved because of their desire to help others if possible, whether because of positive or negative reasons:

Well it was, it seemed good and it seemed hopefully helpful in understanding my experience, so no I’m pleased to be able to give something back more generally across the whole process. I feel I had very good care there, obviously you can always find little things that, and it’s good to have that patient, carer feedback is built in and I’m very happy to contribute to that.

For others, the sense of ‘giving back’ was an incentive to make the experience better for other patients in the future:

Because we had enough sort of concerns and situations she was in, so that’s why we decided to do it, and hopefully, [. . .], the little bit, the contribution we’ve made will help other people who are coming to that situation afterwards. So that was the reasoning out of it.

Patients and families felt that they had been taken seriously at the joint events, noting how members of staff came together and took time out of their working day to discuss how the stroke unit could improve. They felt that the groups provided an authentic way to share ideas and to give time to digest and discuss issues of importance:

Somebody said something about ‘well that’s all well and good, you know, but you have to bear in mind the schedule of the nurses’ or, and I thought ‘oh OK, this is like a real thing, this is not just like, or we could put up some new curtains in the wards, this is like people arguing about stuff’, and I thought ‘OK, cool, I’ll get stuck in, I’ll do my bit’, so yeah, I’ve enjoyed the process, it’s been very good.

The added value of seeing previous patients was highlighted by staff, who rarely had a chance to see those for whom they had previously cared. They viewed the experience as positive as a result of the learning they could gain from hearing the patients’ experiences:

So it’s actually nice to see them come back and see how they’re getting on, and actually not only see how they’ve improved in terms of their cognitive functioning, or their speech, or their mobility and stuff like that, but seeing them using their experience to then make a difference for others in a similar situation.

Similarly, staff felt that they had learned something from working together with the patients that would be of benefit to them in their professions:

I think I’ve learnt that when people sit together from different areas or different groups or different levels, it’s very powerful, very, very powerful because it’s so much easier to resolve a problem. And I think just to persist, just that just to continue by today, even sometimes when you like really you feel like I can’t go on anymore, just to continue, continue, continue, eventually it will happen.

The group dynamics

There is an ongoing debate about equity and power-sharing in co-design/co-production and we were concerned about the impact that aspects such as the setting (the ward), the members of the group (staff in uniform) and the facilitator (which in CREATE was the researcher) might have on equity. However, participants generally felt that they were regarded as valued members of the group and able to express their opinions to staff, and referred to examples of everyone contributing:

I think that worked well and there wasn’t a hierarchy, do you know, it wasn’t just a case . . . there wasn’t really a sort of an us and them sort of attitude, and I think that, you know, the staff that attended, and the support people, and the HCA [health-care assistant] people, they were very interactive, and I think that was good as well, because I don’t think anybody felt that they couldn’t say what they thought, which is important, yeah.

I thought once we’d got over the sort of initial, you know, start-up of it and everything, because you’re always a bit ‘what’s going on here’, but I think, yeah, 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, that’s, yeah, very happy to say that we didn’t feel intimidated in any way.

With regard to staff being naturally more dominant, we anticipated that because the co-designed changes were to happen in the hospital it would be the staff who had the authority and power to change, and so they would take on the tasks. This would put them in a somewhat more ‘powerful’ position. Again, from the interviews, the patients did not comment negatively on this apparent power shift towards staff during the co-design process. In fact, at site 4, for example, one of the carers commented on the necessity of staff taking on those tasks and sustaining them:

I said at the first meeting, the staff were going to have to take the brunt of whatever the changes were, and could they fit it in? And that’s, the only thing I can say about going forward, it’s got to be kept, the momentum’s got to be kept up by the staff and that’s the only people you can ask to do it.

Staff members repeatedly stated that they viewed patients as equal members of the co-design groups and found their contributions invaluable. However, there was an ongoing balance to be struck between being creative with the changes and doing something plausible within the time frame:

I mean what is difficult with the nature of, and I think of the design where you had the patients and the staff within the co-design group is, the patients have fantastic ideas and really great ideas that we would be very keen to see, but almost, from sitting on a staff perspective, when these ideas are being suggested we straightaway know what’s plausible and what’s not plausible within economic constraints or within space constraints, or within the general sort of health-care system, and sometimes that can be quite a challenge in terms of relaying that back to the relatives or the patients in the group to say, actually it’s a great idea but it’s not going to be possible for X, Y and Z. So then it’s really up to us to then try and find a happy medium almost and try and suggest something that would be plausible within resources that we have.

The value of being part of a group

Being part of the group itself and not just seeing the changes take place was invaluable for some of the patients. They felt part of something that helped them to ‘forget’ about the stroke for a short time, while at the same time being able to process what had happened together with people who had had similar experiences:

It’s [CREATE] helped me through me life because the experiences I’ve gained from it, you know, not just talking about strokes, but actually being part of a group, I think that helps one heck of a . . . [. . .] forgetting all that, at the end of the day it’s just nice to be with a group of people that you can talk to, and without realising it, [. . .] they’ve had the same problem as me and we’re all talking about what we’ve had, [. . .] I know that it could happen, there’s no point in thinking negative [. . .] But the one good thing I know now, and that is, if I do have another stroke I know a few people at [site 1] that can help me through it, you know what I mean, so . . .

Hearing everybody else’s part of what they had gone through . . . it’s been really good, it opened my eyes quite a lot, because you sit there [on the unit] and you see things going on, so yes I enjoyed being part of making something better.

One member of staff who had not taken part in the initial events said that morale was higher on the ward now and that there were always new milestones to focus on and look forward to, as things were still being put in place. Before, she had thought that ‘things will never happen’, but after CREATE she knew that actions could actually occur as she was surprised at how quickly things had changed. In addition, in terms of teamwork, CREATE had had a positive impact:

I take from the process that getting people to work together can get a lot more done than one person trying to fight alone, which is sometimes the way it feels, so being able to quite easily get these things together, and also I’m not the only one that has ownership of it, [. . .], but there’s a whole team of us that sort of have ownership a bit of this, I think we’re all equal, so I think if, if something like doesn’t happen I can at least try and speak to those other people and see their thoughts, so I still feel that we’re a team even though we’re not meeting, we might not see each very often, but that, that team’s still there for, for future ways to work.

Staff in particular noted the way in which the project had enriched working life and given them a new purpose and meaning. Some felt that taking part in EBCD had helped staff morale through bonding and orientating around a common purpose:

Staff seemed to be enthusiastic. It felt like it had knitted them together, perhaps more than just ordinary work does.

Another member of staff said that everyone should feel part of it: ‘we are all in it’. This member of staff had not taken part in the co-design groups but also said that EBCD was for the staff as well as the patients. Members of staff commented that the ward was also now a nicer working environment for them. Staff also felt that they had personally learned from CREATE for their own practice and were more aware of how to improve activities, space and communication:

It’s really nice to get a perspective from them [stroke survivors and caregivers] as well, because I think they brought a few different things to the table, from an outside perspective of how relatives want to be involved and the things that they missed and the things they didn’t know about, because, again, you totally take for granted what we know because you’re doing it every day.

Although interviews generated a lot of positive feedback about the experiences of taking part in the co-design groups, staff highlighted several caveats relating to the resulting additional workload, concerns about sustainability beyond the project and the need for buy-in and support from all staff, especially senior management:

I would, maybe like I say, just get more staff involved which is hard, or the staff that are involved to communicate with the other staff to tell them to pass that on so that things, some people didn’t know about the volunteers but . . . I think we should have more staff meetings within our groups just to say, especially after what’s happened on the wards ‘oh this and this and that’.

The recognition senior leadership teams gave was inconsistent at some sites and the following quotation illustrates how some managers did not prioritise CREATE:

It’s not the first time they’ve disappointed me but that’s disappointing, because, you know, a lot of work was done, I mean I invited my boss, I did invite him quite last minute to be fair but, you know, they didn’t go, you know, ‘oh no, we’ve got other things on’, but they’re always saying ‘oh we’ll come and support you at things’ and then when you want them, they’re not there.

For context and richness to further interpret the use of EBCD in these stroke units, we now describe our process evaluation.

Process evaluation

The process evaluation was informed by NPT. We draw on data generated on the impact of developing and implementing co-produced interventions on the quality and amount of independent and supervised activity (social, cognitive and physical activity) occurring outside formal therapy on the four stroke units. Table 5 shows the range of data sources used for analysis and interpretation.

Normalisation process theory is a middle-range theory concerned with understanding how complex interventions are implemented in and integrated into existing health-care systems.72 As outlined in Chapter 3, NPT proposes four constructs that represent different kinds of work that people do around implementing a new practice: coherence, cognitive participation, collective action and reflexive monitoring (see Table 4). In our analysis we used NPT in two ways. First, we used it as a sensitising device in our review of the qualitative data generated through observation, interviews and researcher process notes and reflections on EBCD activity. NPT’s constructs were useful in identifying and thinking through factors that emerged as barriers to and facilitators of using EBCD and to introducing change at the four sites. This was particularly the case in relation to exploring what occurred at different stages of the implementation process within and across the participating sites. Second, we used NPT as a structuring device as we progressed in the analysis from identifying barriers to and facilitators of linking these, where appropriate, to NPT’s constructs. We used the constructs as part of interpreting how being part of the study and using EBCD had an impact on the day-to-day work of staff on the stroke units and, over time, how changes to spaces and to the availability of activity opportunities appeared to have an impact on stroke patients’ experiences.

To examine the factors and organisational processes acting as barriers to or facilitators of successfully implementing, embedding and sustaining co-produced quality improvements in these settings, we identified and considered the ways in which study participants engaged with the EBCD approach, and the CREATE researchers’ roles in supporting the participants at each site. We sought to understand how staff, former patients and carers participated in the process of identifying and agreeing priorities for change, designing solutions to realise changes and finding ways to introduce these changes into existing working practices in the stroke units. We return briefly to the use of NPT in Chapters 5 and 6.

Making sense of CREATE and the experience-based co-design process (coherence and cognitive participation over time)

Four members of each stroke unit team were invited to attend CREATE and EBCD training in London. Sites nominated staff; our only stipulations were that those attending should include both therapists and nurses, and that these staff should be aware they would be taking a lead role in the EBCD process at their site. During the full-day training for staff from sites 1 and 2, participants met with researchers and heard about the aims and design of the CREATE study. They also heard directly from one of the originators of the EBCD approach (GR) as well as from an experienced clinician who had led an early EBCD project. We involved a designer (AM) in this training. The day addressed the theory of (a.m.) and practical applications of EBCD (p.m.). Feedback suggested that participants found the day to be lengthy and that the more theoretical overview of the methods was less engaging and less useful than the practical applied material. Based on this, we reduced the training time for staff from sites 3 and 4 to a half-day; we retained the input from the experienced former clinician and focused on practical applications of EBCD by reporting how sites 1 and 2 had worked with the EBCD process and the changes they were putting in place.

Accepting the comments regarding the preference for the more applied approach to training, most participants evaluated the training sessions positively. Staff from sites 3 and 4 heard first-hand accounts of how those at sites 1 and 2 were working to increase activity and discussed examples of changes already made by these teams. The groups also generated a series of practical implementation questions that were discussed during the training session (Figure 16).

FIGURE 16 The EBCD training event for staff from sites 3 and 4. Reproduced with permission of Fiona Jones, University of London, personal communication, 2020. (Figure not reproduced here; see the original article.)

It was helpful meeting other people involved in the study on their wards. What they did and how it involved patients.

From researchers’ reflections on participating in the training sessions and from informal contacts with staff attending both sets of training, enthusiasm levels varied even at this early stage of the study. In both training sessions, staff from one site were much more engaged than staff from the other. In session 1, staff from site 2 (Yorkshire) were more engaged, whereas in session 2 staff from site 3 were more engaged (London). We explored whether this variation could be explained by whether staff had chosen to attend the training or had been selected; however, this was evident for nursing staff in one site only (who had been selected).

Attending training afforded these core staff groups the opportunity to start making sense of the EBCD approach and to think both individually and collectively about how to share this knowledge and understanding with their stroke unit colleagues, about what changes could occur in these settings and what such changes would entail practically for their day-to-day work and that of their colleagues. The training also challenged these staff to think about how working with patients and carers as part of the participatory EBCD approach would differ from their experience of using mainly written feedback on patients’ and carers’ experiences of the stroke units (e.g. through the Friends and Family Test and through surveys).

Making sense of the study and experience-based co-design in the wider stroke team (coherence and cognitive participation over time)

Stroke units are staffed by large teams of therapists and nurses, two or three physicians and support staff. Ensuring that all staff working in the study sites had an understanding of the study and of the EBCD approach remained challenging; one factor influencing this was the time between researchers introducing the study at sites and the data collection and EBCD activity commencing. At sites 1 and 2, major changes to the Health Research Authority and Research Ethics Committee approvals process, which occurred soon after approval for the study was sought, introduced a delay of almost 4 months between training the core groups and data generation and EBCD activity getting under way. Information about the study and the EBCD approach was disseminated through study posters and newsletters and by researchers explaining them at multidisciplinary team meetings and staff handovers at each site. However, the lengthy delay meant that early interest in the study shared by EBCD-trained core groups members quickly dissipated and the spread of understanding of what the study and the EBCD approach would mean for these units was minimal. We had anticipated that, after the training, the time before any study activity took place would have allowed core staff to make their own observations of patient activity/inactivity in their units and to identify and consider opportunities for increasing activity. However, as changes to working practices were not yet required and researchers’ access to sites was limited to planning meetings, only core team members at site 1 did this. The delay meant that in both sites the core staff and wider teams focused on routine work demands; the CREATE study and the EBCD process were not really ‘visible’ and, prior to data generation through observations and behavioural mapping, did not make demands on any staff group at this time.

Once the study was under way at sites, maintaining awareness of EBCD activity remained a challenge as, after the initial enthusiasm generated by the joint meetings dissipated, often some weeks would pass before the actions planned by co-design groups came to fruition and the changes became visible. Early ‘wins’ such as reclaiming the day room at site 4 and opening the new ‘Social Corner’ at site 1 ensured that participants and wider staff and inpatient groups became aware that change was happening. Researchers circulated newsletters on several occasions in each site to raise awareness among staff who were not directly involved in co-design group activity (see Report Supplementary Material 4).

Infection control restrictions on what materials could be displayed in public and staff circulation areas in some sites sometimes negated this strategy, and it was not uncommon in post-intervention interviews to hear some staff say that they had never read a newsletter and that they did not know what was happening in the study. Therapists at sites 2 and 4 reported routinely sharing information about the EBCD-driven changes during their daily handovers or weekly team meetings, so lack of awareness in these smaller therapy teams was not reported:

Because there’s been a couple of therapy staff involved from physiotherapy and occupational therapy mainly, we meet in a morning for a board meeting and we’ve handed over what the progressions have been from the CREATE study, so if they see anything on the wards this is what it is. So say for example we’re telling them about the updating of the garden and if we need to take anybody to the garden.

Nursing work patterns on stroke units are different from those of therapists, with larger numbers of nurses and health-care support workers in each team. Communication across nursing teams working early, late and at weekends, and on extended day and night duty, is more difficult to achieve, but it was notable that CREATE information was not included routinely in nursing handovers at any of the sites.

However, a factor that increased staff awareness at all sites was researchers commencing data generation; once this occurred, staff interest gradually increased. Some awareness of the study’s purpose also became evident among the patients being cared for on the units at that time through informal conversations and patient/carer questions during the ethnographic observations and through patients’ willingness to participate in behavioural mapping.

Raising awareness and engaging staff, patients and carers through researcher presence on units (coherence and cognitive participation)

At sites 3 and 4 there was no delay between training core staff groups and regular researcher presence on the units as data generation got under way. In all sites this regular researcher presence during the pre-EBCD observation and interview periods was a factor in enabling a wider group of staff to start to make sense of the study. Staff were curious about behavioural mapping and general observational processes, and researchers used this opportunity to provide brief verbal explanations of the study and EBCD to staff and patients on every occasion on which they were generating data. For staff, researchers supplemented verbal explanations with written study summaries, posters and newsletters indicating that the study was under way and explaining how it would proceed. In some units the presence of the researcher prompted staff trained in EBCD to discuss the study during routine team meetings, and this, in turn, reportedly encouraged team members to generate ideas for changes to increase activity. When researchers were conducting observations and behavioural mapping, staff commonly expressed their views about the low levels of current patient activity outside therapy and who could do what differently. Informal comments were very similar to those expressed in the semistructured interviews conducted after the observations:

The presence of researchers on the ward during ethnographic observations seems to have led to staff discussing the CREATE project and the subject of patient activity on the ward. Informally staff tell me they are aware that patients can become very bored and experience low mood if they do not engage in any activity for long periods. This, they report is especially true for patients who are on the ward for a long time; most staff say they know the worst times for inactivity were weekends and evenings. They comment [registered nurses, health-care support workers and therapists] that the study is necessary and that they hope that the situation can be improved. Some staff expressed an interest in being involved in the planned EBCD activities. There were different thoughts about whether this was something HCSWs [health-care support workers] would have more time to do or whether this should be something all staff contributed to. It’s possible that this increased awareness of inactivity led to the observed small number of efforts to increase activity on the ward during the weeks of formal observation.

Although the initial researcher presence was helpful in raising awareness of the issue of inactivity and prompting wider staff groups to think about what would need to happen to change this, informal and pre-EBCD interview comments suggested that some staff were sceptical because they had experiences of good ideas for change being identified but not progressing. It was also difficult for most staff at this stage, namely prior to any EBCD activity, to envisage whether or not, or how, increasing patient activity may have an impact on their existing roles and working practices. Despite some enthusiasm for increasing activity levels, even at this early stage staff routinely expressed the view that their existing workloads and staffing levels, which were often below those recommended in the National Clinical Guideline for Stroke, would not allow them to spend time completing activities with patients. Nurses and health-care support workers most commonly expressed this view. In NPT terms, at this stage, for most staff, the ‘intervention’ was still undefined and not widely understood; commitment to change and comprehension of the possible benefits of changes were not widely evident. In turn, at all sites, there was no expectation or direction from senior therapists or ward managers at this stage that individual staff members or groups should review or change their working practices.

Patients and carers at each site also became aware of the study as a result of the observations and behavioural mapping and, as with staff, they sought information about the study and expressed their views on their current levels of activity outside therapy. This early contact and initial relationship with researchers appears to have influenced several inpatients later (after they had been discharged home) in consenting to participate in pre-EBCD interviews and, subsequently, in EBCD activities. In contrast to some EBCD studies identified in our rapid evidence synthesis, in which the recruitment and retention of service users was problematic, researcher presence proved beneficial in terms of both recruitment to these parts of our study and sustaining patient/carer involvement in EBCD activity throughout the study.16

The influence of participation in experience-based co-design elements (coherence, cognitive participation and collective action)

At sites 1 and 2, the structured and facilitated approach of EBCD was a major factor in staff making sense of CREATE and the EBCD approach and progressing to a more engaged position, readier to commit to thinking about how change could happen in their site, to revised or new working practices and to thinking through who would need to be involved (cognitive participation). At these sites the initial (separate) staff meetings and then joint meetings enabled quite large groups of staff to experience participation in some of the EBCD cycle stages (see Table 2). They viewed edited trigger films, heard first-hand accounts of patients’ and carers’ experiences, and then jointly identified ‘touch points’, discussed experiences that patients said mattered most to them and considered some solutions to these issues. These were high-energy meetings at which there were strong levels of enthusiasm for change; they were a powerful catalyst for larger groups of staff beginning to share the view that changes to stroke unit environments, access to resources and routine working practices to increase patient activity were not only desirable but also possible (Figures 17 and 18):

It felt quite exciting – it will be interesting to see how it develops, keen to be involved and contribute.

FIGURE 17 Staff event at site 2: thinking differently about activities for different patient groups and about how activity could occur in existing spaces. Reproduced with permission of David Clarke, University of Leeds, personal communication, 2020. (Figure not reproduced here; see the original article.)

FIGURE 18 Joint event at site 2. Reproduced with permission of David Clarke, University of Leeds, personal communication, 2020. (Figure not reproduced here; see the original article.)

Other features of EBCD that facilitated the development and implementation of changes were the defined and time-limited nature of the EBCD process and its use as a previously tested patient-centred service improvement model. This seemed to legitimise the staff time and resource allocation committed to participation in the study and the associated intervention. Staff noted that the EBCD approach contrasted with previous attempts to introduce change in working practices, which were referred to as ‘poorly defined’ in terms of timescales, roles and responsibilities and as operating without additional resources. The patient-centred and participatory approach inherent in EBCD also appeared to have added a sense of responsibility for the core staff groups to deliver on agreed actions, and not to ‘let down’ the patients and carers with whom they worked in the co-design groups.

For patients and carers at sites 1 and 2, separate meetings provided the opportunity to explore their shared experience of stroke and stroke unit care. The trigger films demonstrated commonality of as well as differences in experiences, and watching the films began the process of identification by these former patients and their carers as a group of not only the shared experience of stroke, but also a shared belief in the importance of improving the experiences of patients and carers in their local stroke units. There were some differences in how the trigger films were received; at site 2 staff showed visible emotion at the experiences highlighted by patients, whereas at site 1 there was a mixed reaction:

During the film itself, there was intense attention among the participants and signs of recognition of what patients and family members were saying – for example I noticed some staff members nodding, laughter at some of the observations made by participants. I had expected the film to raise visible emotions but I didn’t see any [tears, sighs, shaking of heads]. There seemed to be recognition and some laughter. On the table of mostly staff where I was sitting there was quiet attention but no visible signs of emotion. I don’t know if on other tables there were more visibly emotional responses.

Former patients and carers at sites 3 and 4 did not have the trigger film participation in common but they recognised the issues described in the films by site 1 and 2 participants as similar to their own. It seemed that the shared experience of stroke and a desire to improve the experiences of others in these units were sufficient to form a common bond.

Researcher reflections on the joint meetings and co-design group meetings highlighted the importance of former patients articulating their stroke experience and the impact that their impairments had on their daily lives. This seemed to inform an ongoing consideration of why activity outside therapy mattered to patients and how, practically, it could be increased. This common sense of purpose and shared experience is often referred to in discussion of the concept of a ‘community of practice’. 77 A community of practice has been defined as: ‘a process of social learning that occurs when people who have a common interest in a subject or area collaborate over an extended period of time, sharing ideas and strategies, determine solutions, and build innovations’.77 Although not ‘practitioners’ in the health-care sense, patients and carers participating in the EBCD activities in all of our sites shared a common experience not only with their peers, but also, to a greater or lesser extent in the different units, with the staff with whom they developed an ongoing relationship, shared ideas about activity post stroke and problem-solved in the co-design group meetings.

The structured approach of EBCD and the facilitated events and activities of the EBCD cycle provided former patients and carers with a way to express their experiences and involved them in a group where they worked as partners in bringing about improvements to the stroke units that would increase patient activity:

[. . .] I think, yeah, 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.

The same was true of participation by those who were ‘external’ to the day-to-day work of the stroke units. These opportunities contributed to participants’ progression in thinking from a shared understanding of the importance of change (coherence) to working together to define ways to implement, and ideas generated into existing working practices (cognitive participation). This progression includes thinking through how the changes designed to increase activity could be actioned, who would take responsibility for action and, in the longer term, who in everyday practice would deliver or support the actions. These mechanisms represent collective action in NPT terms, and, although the work of implementation on the stroke units commonly fell to staff, the findings section of this report identifies actions completed by carers and former patients to support change. These included designing and painting wall murals (sites 1 and 3) and redesigning patient- and carer-facing documentation (sites 2 and 4), actions designed to increase activity by providing stimulating spaces and ensuring that inpatients and their families were aware of spaces for and opportunities to undertake joint and independent activity.

The above may suggest that EBCD elements, including co-design group activity, progressed without incident. In terms of core group members and participation in co-design groups, this was not the case at most sites. At site 1, for various reasons, three of the four team members who received the EBCD training were not able to participate in the co-design work. However, other team members expressed an interest and volunteered to take on these roles, so staff commitment to the process evolved more naturally from the outset. At site 2, a core team member leading one of the co-design groups experienced problems with taking time away from their ward-based role and then had an extended period of sickness. This had an impact on and slowed the progress of this co-design group, but, as at site 1, other staff members, in this case supported by the ward manager and site researcher, maintained the activity and the agreed actions were progressed. Actions included securing funding for one of the health-care support workers to provide regular hairdressing and personal grooming sessions outside their normal ward role; this realised both person-centred and social activity goals.

At site 3, a champion emerged quickly who led the process at site level who encouraged staff to attend training and sign up for events and attended most if not all co-design groups. Site 3 had the most co-design groups and the most attendees, which we believe was largely because of the enthusiasm and persistence of the local champion (see Table 2). At site 4, two EBCD-trained core team members, both nurses, attended only the joint meeting and did not participate directly in any of the work of the co-design groups they had agreed to. Despite having support from their ward manager, both maintained that they could not take time out of their routine work to participate in the co-design meetings; other members of the co-design groups agreed to conduct the review of priorities and action-planning. Progress was slow in the three co-design groups and few actions were completed at this site until a combined meeting of the groups led to rapid realisation of several agreed activity support actions. The additional contribution of the site researchers as facilitators in supporting and sustaining co-design work is explored later. In our view, the development of communities of practice to a greater or lesser extent in the co-design groups at each site provided a support mechanism that sustained the work and progress of these groups when they were faced with the challenges outlined above.

Leadership, managerial authority and support (cognitive participation and collective action)

Previous EBCD studies have identified the central importance of senior staff support and leadership in service improvements generated through participation in EBCD cycles. Closely allied to this is the presence or absence of managerial authority, resource allocation and support for proposed changes. Based on our review of facilitators for co-production and EBCD-like projects, we aimed to set up project oversight groups with membership including directorate-level managers, senior nurses and those with cross-organisational roles.16 At sites 1 and 2, we set up these groups, but the commitments of these often very senior managers meant that interaction was largely through e-mail updates or one-to-one meetings with researchers. We could not engage equivalent staff and establish an oversight group at site 3 or 4. The same processes of offering to meet with these staff individually and repeated invitations to attend EBCD activity were adopted, but these were largely not responded to or the promised attendance did not actually occur. One impact of the accelerated EBCD process was that researcher time at sites 3 and 4 was also reduced. Researchers enlisted the support of unit-based EBCD champions in seeking to involve more senior managers, but this did not lead to the involvement in oversight of the project that we sought. Although oversight group members at sites 1 and 2 were often not active participants, they were important in expressing their support for project activity, in helping to navigate often complex trust organisational structures and in specific situations for providing financial resources to enable co-design group ideas. For example, at site 3, staff raised £8000 from participating in a sponsored 10-km run. The chief executive at the site matched the sum raised and these funds were used to redecorate the day room and install a kitchenette. In our view, the oversight groups at sites 1 and 2 provided direct and indirect support for unit-based staff. Having defined senior managerial contacts in these organisations ensured that unit-based staff could activate these potential lines of support. This was not impossible to achieve, but it was made more difficult at sites 3 and 4, as only the chief executive and the therapy lead, respectively, engaged with the EBCD groups and the researchers.

In all sites, a range of staff working in the trust in roles including therapy service managers, matrons, patient experience managers, patient safety officers, volunteer co-ordination and estates management joined patients and carers in different co-design groups. Typically, these individuals were not involved in every co-design meeting held but their involvement was often significant (see below) and had an impact in terms of both cognitive participation and collective action. NHS staff are acutely aware of the resource constraints affecting their services, and often have experienced frustration at barriers they have encountered when seeking to improve or change services. These include infection control and patient safety requirements, lengthy delays or inaction when estates work is requested and lengthy and bureaucratic processes to work through when looking to include volunteers in unit-based activity. Collectively, we found that staff who were otherwise dynamic and enthusiastic advocates for increasing patient activity often shared the view that such barriers were fixed and would, inevitably, limit what could be achieved. The perceptions that ‘the trust’ would not ‘allow’, for example, murals to be painted on walls, shelving to be added in patient bays, hot drinks facilities to be added for patients and carers to use independently or volunteers supporting patients with social eating were pervasive and initially were shared by some of the clinical managers who attended joint meetings at site 1. These perceptions proved to be largely inaccurate.

The contributions of participants from the wider hospital, patient services and volunteer services were significant and these participants typically explained how changes could in fact be realised, and, importantly, provided examples of such changes already operating in the same trust. These participants were used to working with departments and services that were often criticised in the routine discourse of ‘infection control/estates will not allow it’ and understood what was required to introduce, generate support for and bring about changes involving these departments and services. Sometimes these participants undertook to make things happen themselves; for example, at site 2 the patient experience officer worked with the volunteer co-ordinator to identify new volunteers, develop a mealtime support training programme and follow this through to the provision of a regular volunteer-run social eating group on Sundays. At site 3, the volunteer co-ordinator attended the joint event and many of the co-design groups to provide a rapid link to possible volunteers for activities such as reading support and exercise groups. At site 4, the therapy services manager supported the idea of a breakfast club to increase social contact between patients and provide an opportunity for meaningful functional rehabilitation activity. This support was evidenced by changing the working times and early-morning responsibilities of two therapy assistants who were assigned to establish and run the breakfast group (collective action). Therapists and nurses supported the group, quickly establishing it as a 7-day-a-week activity. In this case the work of implementing change was consistent with staff members’ beliefs about the unit as a rehabilitation space, and the work required of the therapy assistants was defined as rehabilitation support integral to their job purpose and consistent with the kind of one-to-one kitchen activity interventions they already completed with patients prior to discharge (cognitive participation and collective action).

However, support from wider trust services was not always evident. At sites 1 and 4, volunteer co-ordinators either focused on the obstacles to volunteer involvement or ignored repeated requests to discuss routine volunteer involvement on the stroke unit. Co-design groups responded differently to such challenges; some decided not to pursue volunteer activity, whereas others, such as at site 4, made direct contact with existing volunteers to request support for arts and crafts sessions and contacted the volunteer programme of a local college, which already had permission to work in the trust, to set up their involvement in regular sessions of activity support on the unit.

At each site, unit-based leaders emerged; often, but not always, these were core group staff who had been through the EBCD training. They were not always the most senior or most experienced members of staff in the units. For example, at site 2, two members of the co-design groups who were particularly active in increasing patients’ social, cognitive and physical activity were a health-care support worker and a stroke rehabilitation assistant. A similar situation was apparent at site 4, where a therapy assistant worked across co-design groups and formed an alliance with the ward manager (who was not EBCD trained). Together, these two staff members drove the initial changes, which included reclaiming and equipping the day room, repurposing a large window seating area and encouraging therapists and nurses to use these. At site 1, one member of staff who had not been to the initial training emerged as the EBCD lead for the ward after the first joint meeting. A general manager and a hospital matron at site 1 also helped expedite changes such as new curtains, clocks and redecoration; their involvement was prompted entirely by a meeting at which unit staff presented the aspirations for the ward and asked them for help with the barriers being experienced. The meeting was organised by the lead facilitator (a dietitian) and the researcher for site 1. Direct actions related to the requests made by the EBCD champion suggested that these staff (a matron and a general manager) also developed some commitment to the project’s anticipated benefits, evidenced by their allocation of resources and continued involvement in seeing these planned changes through to completion. Although these staff were not necessarily part of the community of practice developed through the co-design groups, they made a defined contribution to the collective action needed to improve patients’ experiences at this site.

In terms of leadership style, perhaps the predominant approach evident across sites was an action-centred approach,78 with most staff (but, again, not all) leading co-design groups focusing on and reviewing progress against agreed tasks in each meeting but at the same time actively including, as far as they were able to, patients and carers as partners in the planning and decision-making of the groups. There were no instances in which patients or carers were invited to lead the co-design groups in any of the sites, although, as indicated above, their contributions were actively sought, and our findings confirm that these were influential in the changes that occurred across the four sites. Medical staff largely did not participate in the co-design meetings, but they continued to show interest in the EBCD approach and supported the work in practical ways. At site 4, the stroke consultant sourced and purchased secure, mobile iPad stands and accessed unused hospital bleepers. At site 1, the medical consultant helped facilitate and set up the staff and joint events and secured funds for equipment from his research budget. In recognition of this, the co-design groups chose to name the new social corner after him when he left the unit.

Experience-based co-design activity was welcomed but added to the workload of core group staff (cognitive participation, collective action and reflexive monitoring)

In common with findings from our rapid evidence synthesis of co-production and EBCD studies,55 core staff group members in the stroke units in CREATE were not expressly allocated any time out of their usual working day to participate in and action co-design activity. Instead, staff were encouraged to work flexibly, and their line managers largely encouraged and supported their staff to participate in the EBCD work. This was somewhat easier for therapists at each site than it was for nursing and health-care support workers, for whom taking time out to attend meetings in the course of a shift was often problematic. Ward managers at sites 2 and 4 tried increasing the number of nurses or health-care support workers on shift on days when EBCD meetings occurred but, on some occasions, the additional staff were moved to other wards perceived to be understaffed on those days, so this strategy was not repeated. These ward managers felt that this represented a lack of understanding in the wider organisation of the evidence about the staffing levels required to provide safe and effective care for stroke patients and was an understandable organisational response to recurring staff shortages in different areas of the trusts. However, the enthusiasm for the project at sites 2 and 3 was such that some staff came to the joint meeting on their days off. This kind of commitment was also evident later in the study at these sites, with staff coming in to work on their days off or before a shift commenced to participate in the co-design meetings. Core team members differed in the amount of time they were able to allocate to EBCD activity and progressing the actions agreed in co-design meetings, but most identified that they needed to complete some activities in their own time. These staff typically explained their willingness to undertake this extra work by referring to patient benefits, which were becoming more obvious as the co-design activity progressed in each site (collective action). Where the work of implementing increased patient activity was perceived to be enhancing the patient experience in ways consistent with staff members’ rehabilitation ethos, and where implementation was not requiring substantial change in staff members’ roles, core group members indicated that the extra effort was worthwhile (reflexive monitoring). Other, less involved, team members’ feedback on changes, and on how these were having a positive impact on patients being more independent, completing personal care tasks or engaging in therapy, increased the likelihood that the changes would be sustained:

No, 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, I was committed, but I’ve enjoyed that, I’m glad for the changes. [. . .] it’s what I wanted for the ward, from the minute I got there I knew we needed changes, so I was happy to help bring the changes.

At each site there was some delay before the ideas and actions agreed in co-design meetings came to fruition. In most sites, changes began to happen, sometimes rapidly after the second co-design meetings and were then visible to the wider staff groups. Changes to the environment and use of space in each of the units were positively received. These spaces then began to be used for structured group activities. As these changes became visible to wider staff groups and involved more patients more regularly, these staff in turn appeared more receptive to further changes involving external partners or providers, for example complementary therapy from a local health network (site 1), singers from a local university (site 2) and young musicians from a community arts group (site 3). When changes became visible and change gained momentum across sites, this elicited both positive and negative responses. For example, at site 2 there were complaints that the ‘something about me board just appeared’, with no real consultation with the wider staff group about why it was needed or who would complete the information required. However, the same staff also said that they liked what the board was designed to do. Staff experiences and comments such as these highlight that it is difficult to secure cognitive participation and commitment to collective action from those stroke team members who have had limited engagement with the implementation of EBCD-driven co-designed interventions.

Overall, however, shifts in staff members’ perceptions and behaviours suggested that collective action and reflexive monitoring were becoming more widespread in most units as activity-focused space changes or opportunities became more evident. However, at sites 1 and 4, continuing the work to improve patients’ activity opportunities and engagement relied largely on core group members or those who had taken on these roles as other staff left or did not participate as expected in EBCD activity.

Researcher facilitation role

Our rapid evidence synthesis highlighted that the facilitator role in EBCD projects was an important factor influencing successful engagement in improvement work focused on service user experience. In CREATE, researcher presence on the units was less frequent than that reported for full- or part-time project facilitators in other EBCD projects, but it served a similar function in prompting, encouraging and supporting core team members to maintain their involvement in EBCD activities.16,60 An important element of that was the role the site-based researchers played in getting the EBCD activity under way and in keeping the process moving at each site. Researchers agreed with core group members that they would organise the staff, patient and joint events in consultation with core team members, and researchers recruited former stroke patients and carers and took responsibility for ensuring that they could attend on the dates agreed. They booked accessible rooms for meetings and ensured that patients and carers were reimbursed for travel expenses or they arranged transport for former patients to get to and from meetings.

Researchers liaised with core team members to co-facilitate the staff and joint meetings, with staff presenting feedback on their agreed priorities and co-facilitating the small-group work at each of these meetings. This joint approach, although often initiated by researchers, helped to build the confidence of core team members, and at sites 1 and 2 it gradually increased the engagement of these key staff members in the EBCD activities, supporting cognitive participation and collective action. At sites 3 and 4, because the separate staff meetings did not occur, core team staff had less time and fewer opportunities to work with researchers before the joint meetings took place. At site 3 this appeared to have had little impact on the progression of the co-design meetings, whereas at site 4 core team members either were less confident in chairing and leading the co-design meetings or did not participate in these meetings, citing workload pressures. At this site, the researcher was asked to chair the co-design meetings and did so while ensuring that responsibility for action to bring about change was allocated to staff participants, sometimes in conjunction with former patients and carers, for example in redesigning the ward information leaflet and developing the WiFi access leaflet.

In common with reports from facilitators of other EBCD projects,79 researchers spent considerable time trying to make contact with core staff members and others participating in EBCD to set up meetings or to check on progress with agreed actions. Co-design meetings were commonly arranged to run late morning to early afternoon to maximise staff participation. Core staff were asked to record action plans for each meeting; this worked with some staff in some sites (most commonly when therapists or therapy assistants led the groups) but not in others. In consultation with the core team members, it was agreed that the researcher or one of the core team members would take responsibility for communicating a summary of the discussion and actions agreed at each meeting. The researchers prepared newsletters every 4–6 weeks to report on the work of the co-design group; these were circulated to wider staff and patient groups. Researchers at each site became aware of a degree of reliance on them to co-ordinate EBCD activities; this was more evident at sites 3 and 4, where researchers sometimes felt the need to go beyond their regular research role. For example, at site 3 they participated in a fundraising 10-km run, collected and transported paint supplies across the city so that a mural could be completed on the unit and attended a weekend music concert provided by children at Christmas. At site 4, the researchers attended and co-facilitated an open day set up by core group staff to showcase the work of the co-design groups.

This illustrates one of the main challenges faced by the researcher, and with the staff in the co-design groups in both phases was a practical and predicable response in that researchers agreed their role would include co-ordination of EBCD and co-design meetings. However, at sites 3 and 4, staff reliance sometimes went beyond researcher co-ordination to, on occasion, expecting that researchers would take on the responsibility for some of the actions agreed by the co-design group. It is possible that one reason for this in accelerated EBCD sites (i.e. sites 3 and 4) was that there was less time for staff members supporting the co-design groups to progress from cognitive participation, that is committing to a set of actions to bring about change, to a position of collective action and undertaking the work required to actually implement changes in practices. This may also be an unintended consequence of researchers’ regular presence on the units and their willingness to support staff in their EBCD-generated planning for change.

Increasing patient activity opportunities in routine work (cognitive participation, collective action and reflexive monitoring)

As the co-design groups completed their sets of three or four meetings and patient activity outside therapy was increasing, core staff group members in each site, particularly therapists, began to recognise that the changes to space and opportunity for activity could be incorporated into the daily routine of the unit. For most therapists and therapy assistants, the work of increasing activity was conceptualised not only as reducing boredom and occupying time for patients but also typically as a therapeutic opportunity. For example, in lunch and breakfast groups at sites 2 and 4, patients’ activity provided opportunities to work on therapy goals, including both cognitive challenge and functional task practice. These areas of work are consistent with the goals of rehabilitation and of therapists; they did not require therapists to work in particularly new or different ways and, as a result, were likely to be embraced and not resisted or rejected. Importantly, increasing patient activity did not mean that patients spent less time in formal therapy, which would have had a direct impact on the continuous SSNAP clinical audit of therapy provision. In some sites, staff reported that changes such as the ‘a little something about me’ board (site 2) or the ‘home in the ward’ personalisation of bed spaces (sites 1 and 3) could make therapy more interesting and relevant to patients if, during therapy sessions, staff were able to draw on patients’ personal information through these methods. In NPT terms, these staff had moved from committing to increasing activity (cognitive participation); they understood the work required to establish regular activities and they had identified staff who would lead on or routinely support activities (collective action). In these areas at sites 1 and 2, the work of implementation became more focused on embedding regularly occurring group or individually focused activity into the daily work patterns of therapists and therapy assistants (reflexive monitoring). Post-EBCD observations and interviews suggested that there was limited interprofessional consideration of activity promotion outside the initial joint and then regular co-design meetings. This was the case at all four sites, with dialogue about activity promotion being more likely to occur during uni-professional and non-professional contacts than in multi- or interdisciplinary forums, the latter tending to focus on discharge planning rather than on providing opportunities for reflexive monitoring of the changes brought about by EBCD activity.

Our observations suggested that nursing staff at each site had high workloads and were frequently affected by staff shortages. However, at all sites we also identified periods of up to 30 minutes every day and at the weekends when most health-care support workers would congregate at the shared workstation area. Observations indicated that conversations at these times were often, but not always, social rather than focused on patient planning or organisation. Registered nurses were more often engaged with activities that facilitated patient discharge or supported the organisational functions of the hospital, for example reporting bed status and ordering supplies. Nonetheless, there was limited evidence of registered nurses in charge of the ward encouraging or directing health-care support workers or other registered nurses to participate in patient-focused social or cognitive activity. This differed from some instances when we observed more senior therapists encouraging or directing stroke rehabilitation assistants or therapy assistants to prompt or support social and cognitive activities. In some cases, for example at sites 2 and 4, we observed health-care support workers trying to involve other nursing staff in activities they had set up with patients during these periods, but there were few observed instances when they were able to do so.

In our view, the observational and interview data suggested that, in NPT terms, many of the members of the nursing staff groups in each site did not develop an individual or a collective sense (coherence) of the CREATE project or the EBCD approach as something they needed to participate in (cognitive participation). Nursing staff typically expressed verbal support for the aims of the study and the intentions of the co-design groups when they became more aware of these, often after several co-design meetings had taken place and changes were more visible or evident, but they often did not seek to become part of these groups. Ward managers at sites 2 and 4 were very aware of the challenges that registered nurses face in taking time out of nursing care provision during typically very busy 8- or 12-hour shifts and were prepared to allocate time for the nurses to attend and participate in co-design meetings. Registered nurse and health-care support worker attendance occurred to some extent at site 2 but not at site 4. Although they did not obstruct or resist changes emerging from the co-design groups, nursing staff largely excluded themselves from this work. This suggested that, despite professing a commitment to facilitating rehabilitation and enabling stroke survivors to regain their independence, many members of the nursing teams in the participating sites did not appear to feel that increasing patients’ participation in social, cognitive and physical activity was consistent with their established view of what constituted legitimate and required nursing work in stroke units. There were some exceptions, particularly two registered nurses and a health-care support worker at site 2 who actively engaged in co-design work. Similarly, at site 4 the commitment of the nursing team, prompted by the ward manager, to the 7-day-a-week breakfast group was a specific example of nursing staff supporting a structured group activity. However, overall, across sites, nursing staff were less engaged than other staff groups and volunteers with increasing patients’ activity opportunities.

Did the reduced elements/accelerated approach impact on the implementation process (coherence, cognitive participation and collective action)?

In an earlier study of accelerated EBCD, Locock et al.60 focused on the impact of using an edited film of patients’ experiences compiled from an existing national archive. In our study, at sites 3 and 4, we used existing films from our first two sites. However, we also removed the separate patient and staff meetings. As a result, all participants in these sites had less opportunity to explore experiences and to prioritise areas of change in working practices or resource availability. The other important impact of removing the separate staff meeting was that the opportunity for a larger group of staff members to discuss the project and gain insight into the EBCD process was lost. This appeared to have had an impact on the numbers of attendees at the joint meeting at site 4, which was much smaller than that at the meeting at site 2. The result was that the number of staff, particularly nursing staff, who knew of and sought to participate in study-related activities remained small. However, this was not the case at site 3, where the accelerated EBCD also occurred; here, staff participation was equal to that at site 1. In fact, members of staff asked for a separate staff meeting prior to the joint event so that they could brainstorm ideas for change beforehand, and thus somewhat initiated a replication of the staff event themselves, although this was much more brief. Their ideas were captured and formed a guide for their priorities at the joint event (Figure 19).

FIGURE 19 Ideas captured at pre-meeting: staff at site 3. Reproduced with permission of Fiona Jones, University of London, personal communication, 2020. (Figure not reproduced here; see the original article.)

The absence of the edited film and separate patient meeting did not appear to have an impact on the commitment and engagement of former patients, carers and volunteers who participated in accelerated EBCD. Most of these participants had been recruited after being interviewed about their experiences in the stroke unit, so a basic relationship had been established with the researcher at that site. These former patients and carers participated in the majority of the EBCD activities and thus repeatedly engaged with dialogue about, action-planning for and, later, evaluation of changes to the environment and, to some extent, changes to the working practices of stroke unit staff.

Co-production approaches such as EBCD enable participants to focus on and make the changes they care most about in terms of their experience of a service;61 in the EBCD work in the CREATE study, former patients and carers spoke repeatedly of wanting to improve the experience of people who, like them, would need to spend an often extended period in a stroke unit. In this sense, the coherence and cognitive participation of stroke survivors, carers and volunteers was enabled by their continued participation in the work of co-design groups. Because former patients, carers and volunteers in each site were active contributors to designing changes, they played a key role in ensuring that EBCD champions and other staff involved in the work of the co-design groups maintained the momentum and that they saw through the work planned. Former patients and carers became an integral part of co-design groups in each site as they believed in and were committed to making changes happen. As indicated above, these participants were part of the communities of practice that developed following the joint events and through the regular co-design meetings. Opportunities for former patient and carer involvement in reflexive monitoring occurred to some extent as co-design meetings progressed. However, as these participants largely stopped being directly involved after the last co-design meeting, the celebration event was the main opportunity they had to engage with staff they worked with on the EBCD and to review what had been achieved and what was required to sustain these achievements.

Reviewing the qualitative data generated in our study using the lens of NPT ensured that we adopted a systematic approach to exploring the factors and organisational processes that acted as either barriers to or facilitators of successfully implementing, embedding and sustaining EBCD-driven co-produced quality improvement in the form of increases in social, cognitive and physical activity in these stroke units. NPT proved valuable for reviewing our interpretation of the findings from this large data set. In relation to barriers to and facilitators of using full or accelerated EBCD, and implementing complex interventions in health-care settings more generally, using NPT helped identify the challenge of ensuring not only that staff have the time and opportunity to comprehend the nature of the intended change in practice, but also that the mechanisms of change, in this case the facilitated, structured participatory co-production methods inherent in EBCD, are communicated to and involve as many staff in the change process as is feasible. The process evaluation highlighted that, despite its participatory approach, EBCD is subject to the same implementation challenges as other complex health-care interventions in terms of variable organisational buy-in, differences in the commitment of professional leaders, and differences in the engagement of staff whose primary concern is ensuring that their responsibility for care and treatment is discharged and for whom existing routine working practices are familiar and get the required work done.

However, moving beyond NPT’s focus on implementation and the work this brings, it has also been important to reflect on EBCD as a process and the ways in which many stroke unit staff engaged with this process. We found that, rather than being conceptualised by stroke unit staff as an additional burden to those already having an impact on their practice, such as the continuous clinical audit process of SSNAP and the continuing challenges associated with high workloads and staff shortages, the facilitated and structured approach inherent in EBCD provided a legitimate and, to some extent, protected space in which to jointly plan and co-design changes in unit spaces, unit processes and some staff practices. These changes were focused on issues that former stroke patients identified as important and that staff recognised and endorsed as areas that would improve patients’ levels of activity and overall experiences in stroke units. In the context of the CREATE study, although a significant number of changes became evident in all sites post EBCD, differences in enacting what was often referred to as the rehabilitation ethos in the participating stroke units also influenced the involvement of different staff members in promoting and facilitating increases in social, cognitive and physical activity. These, together with the cessation of the time-limited, facilitated and structured EBCD process and with the movement of staff, are factors that will influence the sustainability of the changes now evident in the participating units. Changes involving structured therapy-led activity, such as breakfast clubs and other group activities, are more likely to become embedded in staff members’ routine work and depend less on individual staff members. Activities that depend on volunteer and external group engagement are more at risk of cessation when key individuals move on. We have highlighted the challenge of sustaining changes that occur in co-production projects previously.55