The Person, Interactions and Environment Programme to improve care of people with dementia in hospital: a multisite study

Care of hospitalised older people with dementia

Studies have identified the organisational and cultural features of acute care delivery that have a negative impact on the care of patients with dementia. These include inflexible care routines that privilege the performance of medical and physical care tasks over time spent with patients,27,28 and physical environments that are difficult to navigate. 21,22,25 The nature and quality of staff interactions with patients has been regarded as problematic, with staff encounters that are primarily task focused and involve limited engagement of the person at a personal level;17–19,21,22,24,25,29–33 and communication practices that are poorly responsive to the emotional needs of patients17,21–25,27,30–32 and/or lead to a failure to ‘see’ and act on behavioural cues that might indicate distress or pain. 34,35 Studies have reported stigmatising and negative attitudes towards people with dementia among some staff and that such patients should not be admitted to acute wards. 27,29–32 Often conveyed as ensuring their safety and that of other patients, the consequence is to see people with dementia as ‘other’. Conversely, staff with knowledge of dementia express lower perceived strain in caring for a person with dementia and less negative attitudes. 32 Research with caregivers has reported inadequate support for their relative with nutrition and hydration,17 both risk factors for delirium;36 the person with dementia not being treated with dignity and respect;17,22,24,27,37,38 and limited engagement in decision-making, for both the person with dementia and the caregiver. 17,22,37 A key factor, contributing to poor care delivery from UK studies, is lack of knowledge and skill in dementia care. 20,22,25,38

What is person-centred care?

‘Person-centred’ care has become ubiquitous in UK health and social care policy discourse as synonymous with care quality. Prioritised in the National Service Framework for Older People,39 it is viewed as key to upholding dignity in care40 and forms one of the key Principles of Nursing Practice developed by the Royal College of Nursing (RCN). 41 However, either its meaning in policy is implicit or there is varying emphasis on different domains. It is commonly conveyed as individualised care, holistic care, choice, autonomy and dignity; and the use of the term, and the particular emphasis attached to it, has varied over time and in relation to wider policy interests (see Appendix 1 for a document review).

From research, there is a similar lack of clarity on what is person-centred care. Two systematic reviews examined its meaning in health,42,43 although they adopted different perspectives. Kogan et al. 42 explored research relating to older people, and Sharma et al. ’s43 review of reviews addressed understanding of person-centeredness in nursing and medicine. Both concluded that there was no consensus on meaning, posing difficulties in operationalising the concept. Neither included studies in acute care, and only 3 out of 132 papers in Kogan’s review related to people with dementia, all of which were conducted in long-term care. To explore this further, we consider theoretical conceptions of personhood and person-centred care in dementia and draw on research evidence to explore what quality care might look like for people with dementia on acute wards.

Theoretical conceptions of person-centred care in dementia

Practice of person-centred care on acute wards

Research relating to older people with dementia on acute wards primarily highlights difficulties in care practice, and evidence reviews conclude that a chasm exists between the rhetoric of ‘person-centred care’ and actual delivery. 21,22 Apart from a recent study,24 this research does not draw on a theoretical conception of ‘person-centred’ care in dementia, although several studies offer insight into what person-focused practice might look like.

Studies on the quality of care delivery on acute wards have been mainly small scale, employing ethnographic18–20,30,33 or mixed methods, including interviews with caregivers. 23,24,29,31,32,54 Two20,54 drew out components of good practice from observation, interviews and conversations with patients. Tolson et al. ,54 employing critical incident interviewing with patients (possible with only a small number) and caregivers, identified four themes within a temporal frame around which components of quality practice could be delineated. ‘Settling in’ referred to the process of adjustment involving responsiveness to emotional signifiers and acknowledgement of the person as valued and part of a family/social network. Its significance stemmed from the person’s need for reassurance, given the disorientation and distress resulting from being in an unfamiliar environment. ‘Visible love’ by engaging with the affective dimension of caring, including the connections between the patient and family, generated a sense of the person’s worth. ‘Reaching me, reaching you’ related to the interactional quality of encounters in which sharing and knowing assumed central importance. ‘My condition’ embraced the acute illness, existing health problems and the cognitive impairment. Quality practice meant achieving a ‘balance of care’ based on understanding the inter-relationship between these three. ‘Best’ practice in acute care required attention to the particular needs of people consequent on the cognitive impairment; strategies that promoted ‘settling in’, thereby enabling people to feel valued; and proactive communication that facilitated connections with the person’s biography and significant relationships in providing care.

Norman20 identified two further components pertinent to ‘good’ practice. First, patients were viewed not just as passive ‘recipients’ of care but as active agents, conveying a sense of ‘self’ in their actions and interactions. When staff viewed what patients did as meaningless or problematic, it resulted in a construction of patients as reduced and dependent. Conversely, seeing patients’ actions and interactions as meaningful drew forth a staff response that reinforced the person’s individuality and sense of self. Second, a distinction was drawn between two types of nursing work, ‘practical’ and ‘emotional’, similar to that in recent literature between ‘task-centred’ and ‘person-centred’. ‘Practical’ work was defined by minimal collaboration in which a provider/recipient relationship prevailed; ‘emotional’ work involved higher levels of collaboration, including prioritising patients’ needs. However, this binary conception of forms of work in the context of people with dementia on acute wards appears limited. Instead, employing the conception of Tolson et al. 54 of the ‘balance of care’, it would seem more useful to consider ‘practical’ and ‘emotional’ work not as polar opposites, but as infused, one with the other.

Nilsson et al. ’s32 ethnographic study on a cardiology unit offers negative evidence in support of Tolson’s ‘balance of care’ conception. The overarching category, ‘falling behind’, refers to the way in which patients’ needs associated with dementia went largely unnoticed until they escalated and emerged as problems that disrupted the usual flow or threatened safety issues in the unit. The response generated was reactive and not proactive, with negative consequences for patients, relatives and staff. Clissett et al. 24 employed Kitwood’s44 conception of the basic needs necessary to sustain personhood as an a priori framework to determine what was person-centred care, namely care that offered comfort, attachment, inclusion, occupation and identity. They concluded that although staff promoted attachment and inclusion, there was little evidence that they supported patients’ sense of identity, occupation and comfort; and that practice was poor in these areas. Their conclusions and interpretations do not flow unambiguously from the empirical data presented. First, the content of what was contained within these domains appeared somewhat arbitrary. Second, the way findings were presented (conveying accounts of how these needs were met followed by missed opportunities) meant that the weight given to positive aspects of practice, in relation to lost opportunities, is unclear from the illustrative data. Third, what shaped these positive and negative features of practice were not elucidated, as the focus was on the interpersonal level of staff–patient engagement. Fourth, there is an assumption that needs conceived of as necessary to secure well-being developed in a care homes context are equally relevant to acute care.

Another recent qualitative comparative study of care delivery to people with a cognitive impairment on a specialist medical and mental health unit and standard care wards within a randomised controlled trial33 provides further clarity on what person-centred practice might look like on acute wards. This involved a subsample of 60 (30 in each group) out of 90 trial participants observed using dementia care mapping and unstructured observation captured in fieldnotes. Many aspects of practice were similar across specialist and standard care wards: crowded, noisy physical environments; delivery of care in a routine and task-oriented fashion; communication that was usually brief or absent when meeting physical care needs; and the prioritisation of physical over emotional needs. Some features of person-centred care were observed, described as care that included personal communication during routine care tasks, validating patient concerns, facilitating independence and delivering care at a relaxed pace. The main difference between standard care and the specialist unit related to the work carried out within the activities room, with designated activities co-ordinators in the specialist dementia unit. The authors concluded that the profile of patients on both ward types was similar (high dependency resulting from acute illness, functional dependency, behavioural and psychological problems with increased cognitive impairment). Although the activities programme was very successful for some people, severity of illness and impairment meant that others could not participate. The trial,55 of which this qualitative study was a component, is the only published research to date that has examined the outcomes of care practice on specialist medical and mental health units compared with standard care. The findings reported increased satisfaction among caregivers on the specialist unit but no significant differences in organisational or clinical outcomes.

In summary, although person-centred care is presented as a marker of care quality in health research, there is no consensus on its meaning. Although the humanistic values underpinning person-centeredness offer a guide for practice, further work is needed to concretise what this person-centred practice might look like in an acute ward environment, and what methods and tools might be effective in improving practice.

Developing the observation tool

The Person, Interactions and Environment (PIE) observation tool was developed for use in the first national audit of dementia on NHS hospital wards25 during a 3-year research study (2008–10). The aim was to develop and feasibility test an easy-to-use tool to:

1. enable local observers (staff or volunteers) to view care from the perspective of the person with dementia

2. provide observers with a means (and vocabulary) to report on care quality based on the findings

3. enable the care provided to be compared with that of other hospitals nationally

4. produce data for use by ward staff to develop action plans to improve practice.

The study was overseen by a steering group for the National Audit of Dementia (NAD), comprising professional and ‘experiential’ dementia experts (including individuals with dementia, and caregivers).

Process of Person, Interactions and Environment tool development

The first phase involved iterative scoping reviews to examine the concept of person-centeredness, and the observation tools used to measure it in an acute setting that could potentially be adapted for the audit. The literature review identified several potential candidate tools. However, these were based on nursing models and not dementia specific;56,57 or they were dementia specific but not tailored for use in an acute setting;58–60 or the complexity, level of expertise required and cost involved in using the tool made it impractical. 59,60 It was agreed to develop a new observation tool for the audit.

A subsequent systematic search of the literature was undertaken to explore the dimensions of care valued by hospitalised older people generally and by those with dementia specifically. The findings from this review were discussed through a process of stakeholder consultation via focus groups with staff and patient representatives from different specialties in four hospitals, and a patient and public involvement group. The purpose was to inform the content and structure of the proposed tool and how it might be used in hospital.

The PIE tool has three dimensions: person, interactions and environment. These represent the elements of knowing and using information about the patient as a person in care delivery, the quality of staff–patient interaction, and the effect of the immediate organisational and physical environment. They reflect inter alia the personhood of the individual, including their subjective experience (incorporating biography, values, life preferences and significant relationships); the interactional or relational character of the social world in which the person is embedded and that shapes sense of self and identity; and the nature of the surrounding physical, social and care environment that provides the context in which care is delivered. Drawing on interactionist44 and social positioning45 theories of dementia, it was understood that the type and severity of the cognitive impairment both have an impact on the individual’s experience of the interpersonal, social and physical world, which in turn affects the person with the condition. This dynamic has implications for the care delivery of people living with dementia on acute wards.

The feasibility of the draft tool and guidance was tested in seven hospitals (18 wards) and through further consultation with the advisory group. Key criteria were that it should be acceptable to staff and feasible for use with minimal resources.

Following modifications and refinements, the PIE tool and guidance were presented in the form of a workbook manual and a one-day workshop to train hospital staff in its use. For audit purposes, PIE involved gaining informed consent to observe the care of up to six patients with dementia or problems in a ward area in real time. Observations were conducted by two staff (one external to the ward), who paired up to observe different patients for at least two periods of 2 hours (one in the morning and one over lunchtime). Following reflection and feedback with the ward team, each ward identified one aspect of good practice and one requiring improvement, with achievable short-term and longer-term actions to be undertaken at ward, directorate and trust/health board levels.

Use of Person, Interactions and Environment in the National Audit of Dementia

An enhanced audit component of the NAD was conducted in a subsample of 55 hospitals (145 wards), and included PIE. The findings from PIE observations, conveyed in the published report,25 were that care and communication were generally reactive and based on an organisationally set, task-driven routine rather than being person-focused, flexible and proactive. Although there were ‘pockets’ of positive, individualised care in the practice of individual staff, or as elements of ward practice, only a handful of wards (5/105) reported practices that were consistently focused on the person. There were periods of care-based activity interspersed with inactivity, leading to lack of attention and stimulation. The environment was often impersonal and dementia unfriendly, with a lack of orienting cues, dementia aids or areas for socialising.

Although the findings were not novel, the breadth of coverage underscored the nature and extent of the problem of care delivery to patients with dementia. Furthermore, a comparison between wards offered preliminary evidence of ‘propellers’ or features of staff–patient interaction that supported person-focused communication, suggesting scope to enhance care.

As the primary purpose of the audit was to document current practice, the process of translating action-planning at ward level to effect organisational change, and the systems and mechanisms required to support and sustain it, was not developed. Nor was there an attempt to examine whether action plans derived from observation affected staff practice, patients’ care experience or health outcomes. These conclusions prompted the current study: the development of PIE as a generalisable practice development process in conjunction with the PIE observation tool (PIE Programme), and an evaluation of its effectiveness in improving care for people with dementia on acute wards.

The Person, Interactions and Environment intervention and implementation process

PIE adopts a systematic approach, using a set of steps to implement and embed change in routine care on hospital wards, each step being supported by tools and guidance. Implementation is a cyclical process comprising observation of current practice (observation tool and guidance). Observation is a means to, first, document care as delivered and, second, consider the impact of the care on patients with dementia to promote fresh insights and potentially challenge existing practice. Reflection on observations is the basis for identifying goals and action plans to improve practice (action-planning tool and guidance). Goals for change are prioritised and action plans are formulated. Action plans specify steps to move from current practice towards achieving goals; how these will be taken forward and by whom; and how they will be communicated to the wider staff group to extend the programmes’ reach and engage them in action. A review of progress against planned action, including an appraisal of barriers to and facilitators of change (review tool and guidance), enables an adjustment of action and/or a review of goals. The process is illustrated in Figure 1.

FIGURE 1.

The PIE service improvement process.

Theory of change

In designing the PIE implementation process, we drew on normalisation process theory (NPT)66 as a sensitising framework. From the range of theories available on implementation,67,68 our use of NPT stems from its sociological focus and the utility of the broad constructs that constitute it. In contrast to psychological theories of change that address individual behaviour, NPT focuses on microsocial processes that affect the implementation of a practice (or technique) in an organisation or clinical setting to account for how individuals and organisations understand and make sense of the new practice, engage with it, participate in the work relating to it, and reflect or appraise its effects in relation to valued goals. NPT appeared most apt for PIE implementation. PIE is a whole-ward intervention; it does not simply involve change in individual behaviour, but also requires organisational, cultural and team-level change. Furthermore, the PIE intervention uses and alters roles and patterns of communication among staff and between staff and patients. Changed meanings, practices, roles and relationships are intended to become part of ‘the way we do things here’ (ward culture).

Within NPT, ‘normalisation’ refers to the work of individuals as they engage in activities that become routinely embedded in knowledge and practice. NPT postulates four generative mechanisms that operate individually and collectively to explicate how practices become ‘normalised’ within routine care: coherence, cognitive participation, collective action and reflexive monitoring. New practices, the theory contends, become routinised when:

• the work that defines and organises a practice or intervention is understood as meaningful and invested in, in respect of the knowledge, skills, behaviours and actions required to implement it at an individual and collective level (coherence)

• the work is perceived as worthwhile and appropriate for the commitment of individual time and effort in order to bring about the intended outcome (cognitive participation)

• work practices, and the division of labour through which these are carried out, are modified or adapted to incorporate the change or intervention into the social system of the host organisation (collective action)

• those engaged appraise the effects as attributable to the intervention and congruent with valued goals (reflexive monitoring).

In PIE, the work of generating coherence begins with the introductory workshop, with its content and participatory nature focused on a range of ways for participants to ‘see the world’ from the perspective of the person with dementia. This continues with the first step in the improvement cycle, namely staff’s real-time observation of patients with dementia. Taking a ‘step back’ from work roles to observe ‘persons, practice and context’ for a short time, as an ‘outsider’ looking in, will, it is assumed, help staff to ‘see’ things in a different way (coherence). This is the prerequisite to reflect on, and examine, what may have been taken-for-granted assumptions, activities and practices. In contrast to the use of the PIE observation tool in the NAD, the number of observations, by whom and when, are flexible, to facilitate local buy-in. Collective discussion of observations within the PIE team further reinforces coherence, and by establishing the need for, and possibility of, change, will secure cognitive participation. These reflections form the starting point for identifying goals for change and the necessary steps or action plans to achieve the goals. Action-planning, including who does what and how it will be communicated and discussed within the wider team, is intended to pursue the work of collective action. A review of progress and reappraisal of goals and actions in the light of what works will, it is assumed, act as leverage to continue the process (reflexive monitoring).

The PIE implementation process is systematic and tools are provided for each step (observation, action plans and reviews). However, the content – how observations are conducted (the number conducted, and for how long, by whom and when they are conducted), the goals identified and the action plans pursued – is flexible and tailored to individual wards. Implementation is conceived of as a process and not a one-off event. 69 A significant dimension of PIE implementation is temporality. Change through successive observation, action-planning, action and review cycles mean that the process is likely to be non-linear and multidirectional,70,71 the success (or failure) of action plans creating the conditions for subsequent review and new/reworked action plans. The commitment and skill of those leading the change and in working through each phase of the improvement cycle will affect implementation and outcomes, so that there is a degree of unpredictability of the process (and the outcomes).

PIE is an exemplar of a complex intervention. 72 It contains multiple interacting components; the levels at which change is directed are several (individual, collective and organisational); and there is flexibility in delivering it. The change process is emergent and non-standardised. 73 The ward setting within which the intervention is being introduced, and with which it interacts, is also dynamic, adding another level of complexity. 74

Process of change

1. How is delivery of PIE achieved, and what is actually delivered?

2. How does PIE produce change, or, conversely, if the intended changes do not occur, is this because of poor/inadequate theory or poor implementation?

3. How and in what ways is PIE implementation (and outcomes) affected by the organisational and resource context of acute care delivery?

Outcomes

1. Does PIE change ward staff practice in ways that are consistent with person-focused care?

2. Is PIE effective in improving care for people with a coincidental dementia on acute wards, as experienced by those patients and/or observed by their relatives/caregivers?

3. Is there preliminary evidence of effectiveness of PIE in improving selected clinical outcomes (days in delirium and falls rate)?

Sampling strategy

Our intention was to recruit eight wards in four acute NHS trusts. Cases were sampled purposively. Wards with a substantial proportion of older people with dementia among their patient intake were identified in trusts (care of older people wards, dementia wards and trauma orthopaedic wards), varying in size and type of catchment locality in three English regions. We employed criteria of ‘readiness’ to implement PIE,80 informed by previous research on delirium prevention:81

• interest expressed among senior acute hospital staff in taking part in research to improve person-centred care for people with dementia

• agreement of senior ward staff to engage in a practice improvement programme over 18 months

• commitment from a clinical lead at directorate level or an individual with responsibility for practice development to assume a leadership role in initiating and facilitating PIE.

Data collection methods

A common set of data was collected for each ward over a 7- to 8-week period at baseline, and at 9 and 18 months following PIE implementation for those wards that pursued PIE.

We collected qualitative data via interviews with staff to examine their views on care delivery and the meaning of person-centred care; explored how routine care was accomplished in time and space using ethnographic observations and informant interviews with staff; conducted in-depth case studies with patients and their relatives or caregivers to understand how people with dementia experienced care through observation of patients over their acute episode and conversations ‘in the moment’ with them, and interviews and conversations with relatives/caregivers; and built up a picture of how their journey into and through the ward unfolded in time, via case notes and observations of handover and multidisciplinary team (MDT) meetings.

Quantitative data relating to ward structure (trust policy, resources and priorities) and physical environment; organisation (staff and patient profiles) and care culture (staff questionnaire), which shape care delivery and patient experience, were collected. Outcome data to examine the impact of PIE were collected at baseline and at 9 and 18 months for selected clinical outcomes (days in delirium and falls), although the systematic collection of these data across all implementation sites proved problematic (see below).

Person, Interactions and Environment process evaluation methods

An embedded process evaluation86 addressed the following research questions derived from NPT concepts, using a similar approach to that we employed in two recent studies. 87,88

• Were the strategies employed to engage staff (interactive workshops and manual) sufficient to engender a sense of purpose around the beliefs and values that inform person-focused care for people with dementia (coherence)?

• Was the process for leading and implementing PIE (team and service improvement cycle) perceived as worthwhile and feasible to invest in to effect change in ward culture and practice (cognitive participation)?

• Was the work involved in pursuing PIE implementation (observation, action-planning and review) conceived of as necessary and effective in engaging staff in implementing and sustaining change (collective action)?

• Did staff individually and collectively appraise PIE as worthwhile for themselves and their patients in benefits or outcomes achieved (appraisal)?

We utilised mixed methods to examine how delivery was achieved, what was actually delivered, the mechanisms of change and the effect of contextual factors on delivery. 89,90

• Observation of PIE action-planning and review meetings to provide a contemporaneous account of the intervention: how it was activated, who was involved, problems encountered and solutions developed. Given the locally determined nature of the content (flexibility) within a systematic process, the question relating to fidelity is couched in terms of how precisely each step in implementation was pursued and proceeded to the next phase. 91

• Conversations/informant interviews with implementation team members to explore how the programme was understood and communicated to the wider ward team; how different programme components were implemented (tools and guidance); and whether they were perceived as necessary, sufficient and of value in facilitating change. These took place at intervals of around 3 weeks within full and partial implementation wards.

• Collection and analysis of PIE documentation: observation, action-planning and review tools to examine the content of the programme as enacted.

• Staff narrative accounts of the process: interviews with a sample of PIE team members at the conclusion of implementation provided retrospective reflection on the process and impact of change. Additionally, we conducted interviews with individuals in sites that had not implemented PIE to explore implementation barriers (29 interviews with 37 staff).

• Events ‘log’: a contemporaneous ‘events’ log was maintained by researchers to capture change in policies, systems and procedures outside PIE implementation that might affect the change process and outcomes. This included initiatives undertaken by the trust aimed at improving the care of people with dementia.

We had intended to organise two cross-site workshops at interim and conclusion of implementation to compare experiences and the contextual factors influencing these. Because of the unevenness of PIE implementation and the wide geographical location of sites, this did not prove feasible. One cross-site meeting occurred of staff in wards in two geographically distant hospitals within Seaford Trust (see Chapter 5).

Impact of Person, Interactions and Environment

Analysis

Qualitative data (interview transcripts and ethnographic field notes) were analysed using grounded theory methods,105 including simultaneous data collection and analysis, constant comparison, searching for negative cases and memo writing. Descriptive and analytic codes were developed into higher-order categories through processes of data reduction and reassembly. First, an analysis was pursued for each data set (interviews, observations, patient case studies) within individual wards (the case study). We then compared analytic themes across these data sets, drawing out, for example, similarities and differences between staff interviews and observations on the understanding and practice of person-centred care. Through regular research team meetings, the emerging data from individual case studies, and reflections on them, were discussed. The similarities and differences between wards were drawn out, and we identified topics for more focused observation and ideas of analytic interest to pursue. Hypotheses were generated to explore and account for variation between cases; these were tested out through a cross-case comparison and search for alternative explanations.

Regarding ward practices with respect to patients with dementia, we proceeded as follows. From initial observations, we focused on aspects of practice relating to the work of communication, routine care tasks, responses to different types of distress, the strategies adopted and the attitudes and knowledge that informed them. For each case study, all incidents and episodes relating to this work were mapped, alongside the environmental, organisational and interactional contexts in which they occurred and the physical and emotional responses of patients. Through a comparison between the incidents and episodes captured in memos, we identified patterns of action and interaction. These patterns and what shaped them were refined through cross-case comparison. For example, the open code ‘distress’ was explored through a perusal of all incidents and episodes pertaining to it. Through constant comparison, different forms of distress were identified, which varied by how the distress was manifested, the perceived need giving rise to it and the persistence of its expression. Staff responses were similarly coded and categorised based on the type of response and the context in which it occurred. The relationship between the varied forms of distress and the responses to them were interrogated to identify patterns, which were further refined through a comparison between cases.

A similar analytic process was pursued in relation to PIE implementation. Qualitative data from interviews, workshop notes, observation of action-planning and review meetings, and examination of completed documentation were drawn together for each ward in chronological order. This provided a within-case descriptive account of implementation over time, the nature of the engagement of staff with each step in the cycle and the barriers encountered, using NPT as a sensitising framework. The emphasis was on delineating the sequence of implementation steps over time and in the context of events within the hospital and trust; the conditions that had an impact on the temporal flow of action, and their consequences, whether persisting with or abandoning PIE. We then examined implementation processes, through a cross-case comparison, to discern generalisable features that might account for variation. An analysis of the qualitative data was conducted manually, moving iteratively between the empirical data, sense-making in relation to the data and reflection on the literature.

Quantitative data were analysed using descriptive statistics.

Research team

Three investigators conducted the case studies with the academic support of the lead author (an experienced qualitative researcher and teacher in qualitative methods). The three attended a 5-day module on the theory and practice of qualitative research at the University of Leeds, which included ethnographic methods and analytic strategies. All four met regularly (at least monthly) to consider emerging findings and problems encountered. Between meetings, telephone conversations and e-mail discussions facilitated dialogue and mutual support, as the observations made heavy demands on researchers both practically and emotionally. The wider research team, including the chief investigator, met every 2 months to discuss progress. This was in addition to receiving support and advice from the Programme Management Group.

Patient consent

The consent procedures adhered to the Mental Capacity Act 2005106 and accompanying Code of Practice. 107 First, when an individual was identified by staff as having a cognitive impairment, we considered how best to approach that person, with advice from ward staff (e.g. times when the person might be more alert, or when their relative was present). Second, the researcher explained the study in clear terms and ascertained whether or not the person understood the information. Third, if the person was deemed unable to make a decision about participating in the research, we sought information regarding a suitable personal consultee (close relative or friend), either from the patient or from a staff member. Every effort was made to secure a suitable personal consultee, or, in their absence, a nominated consultee to give patients without next of kin the opportunity to participate. If the consultee advised that a patient who lacked capacity would be willing to take part in the study, then that person was included, providing that they showed no signs of unwillingness (e.g. becoming distressed, upset or anxious in the presence of the researcher or when discussing the study). Consent was ongoing. The researchers checked that people were willing to continue, and were sensitive to signs of distress, including verbal and non-verbal cues.

General ward observations

The decision not to seek research consent for the general observations of ward routines was carefully considered. As observations were unobtrusive, carried out in public or semipublic spaces, and did not identify individual staff, patients or visitors by name, we considered that it would be impractical, and probably more intrusive, to seek formal written consent from all those present. Instead, we sought informal, verbal consent. Before each observation, the researcher reported to a senior staff member and introduced themselves to any staff they had not previously met, doing this at an opportune moment that did not interfere with the workflow. At the end of each observation period, the researcher ‘signed out’ with the senior staff. It should be emphasised that, throughout the study, observations were conducted with the full knowledge, support and co-operation of ward staff. If anyone appeared unhappy, distressed or uncomfortable by our presence on a particular occasion, we did not include them in the observation. This happened rarely. There were several instances when a patient became annoyed with the researcher when she indicated that she was unable to help the patient leave the ward and return home (call a taxi or get their coat). More often, when a patient was distressed and no staff member was available, the researcher sat with the person and sought to provide reassurance.

In the report, we have used pseudonyms throughout – for trusts, wards, staff, patients and caregivers – and have altered some biographical details to protect anonymity, privacy and confidentiality.

National policy context: dementia

When the study began, in early 2013, the timing seemed propitious. National policy on improving the care of people with dementia had been given impetus with publication of the Prime Minister’s Challenge on Dementia 2015. 9 This coalesced with other policy drivers: the report of the Francis inquiry, published in March 2013,108 on the failings in Mid Staffordshire NHS Foundation Trust, particularly around the care of frail, older patients; and the findings of the first National Audit of Dementia Care on acute wards. 25

Interest in improving hospital care for patients with dementia was reflected across the five trusts. All had in place, or were in the process of developing, dementia strategies, although the content of these varied.

Organisation

City Trust is a long-established foundation trust, being among the first wave to achieve foundation status. It provides services to a core population of around 500,000, and employs > 5000 staff. Acute services are mainly located within a 900-bed hospital, with an 80-bed facility providing rehabilitation and outpatient provision from a separate site. It is a large, busy trust, with > 120,000 accident and emergency (A&E) department attendances annually. Situated in a region with a post-industrial heritage and a diverse culture, its diversity has contributed to an increased demand on health and social care provision at both ends of the age spectrum.

Since attaining foundation status, City Trust had generated an annual surplus. However, annual reports noted the challenge of securing improvements in quality and performance, managing reduced income growth, identifying efficiency savings and responding to the demands of a multiethnic population within one of the most deprived English local authorities.

Up until 2011, the trust was assessed as meeting the performance targets set by the national regulator, Monitor. During 2012 and 2013, performance targets on waiting times for treatment were breached, resulting in a review by the NHS Support Team, the launch of a comprehensive turnaround programme, and a return to compliance in April 2013. In January 2014, the Care Quality Commission (CQC) issued a report on an unannounced inspection the previous October on the main hospital site and assessed that action was needed in several areas, including respecting and involving people who use services and assessing and monitoring the quality of provision. A warning notice was issued on staffing because of staff vacancies carried. Strategic action centred on making improvements in problem areas, including the recruitment of staff, for implementation during 2014. A CQC inspection in October 2014 assessed the trust as ‘requiring improvement’; medical care was ‘good’ at both hospital sites. A change in leadership at board and senior management level at this time created some organisational instability. This was the context in which the strategic priorities around the care of people with dementia were established.

Care of people with dementia

A strategic trust priority from 2011 to 2012 was improving the physical environment for patients with dementia. City Trust was successful in securing external funding for this purpose, augmented by investment through its capital improvement programme. The design and execution of the changes on several older people’s wards was described as ‘exemplary’ in the CQC’s 2014 inspection report. The 2013/14 annual report announced the extension of the improvement programme to the second hospital.

Since 2012, a rolling programme of ‘dementia awareness’ training tailored to different staff groups had been delivered jointly by the dementia lead within the elderly services directorate and the trust’s practice development lead. Both subsequently launched an initiative to recruit ‘dementia champions’. The appointment of a lead nurse for dementia in 2014 instigated a dementia strategy through the formation of a multidisciplinary dementia steering group.

Rivermead: structure and organisation of care delivery

Rivermead is a ‘step-down’ ward for rehabilitation. The hospital in which it is located previously had more inpatient wards but now mainly comprises outpatient services and rehabilitation beds, acute services being centralised in the main hospital. Bed numbers fluctuate according to bed management demands: usually 27 during ‘winter pressures’ and 23 at other times.

Patient profile

There were 123 admissions over 7 weeks at baseline, eight of which were readmissions (approximately 18 admissions per week). All of those admitted had come from the main hospital following an acute episode. The median length of stay was 14 days (range 1–48 days). Table 1 provides a summary of the patients’ sociodemographic characteristics.

TABLE 1 - Rivermead patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
65–69	1 (1)
70–74	0
75–79	17 (14)
80–84	32 (26)
85–89	38 (31)
90–94	30 (24)
≥ 95	5 (4)
Gender
Men	28 (22)
Women	95 (77)
Residence prior to acute admission
Own home	108 (88)
Extra-care housing	2 (2)
Intermediate care	1 (1)
Residential care home	2 (2)
Nursing home	2 (2)
No information	8 (6)

More than half of the patients were in advanced older age (i.e. aged ≥ 85 years), and women dominated. Most patients had been living in their own home before the event that had precipitated their acute admission.

Over one-third (n = 44) of patients had an ongoing cognitive impairment: 36 had a diagnosed dementia and eight had been assessed as having probable dementia. Almost one-fifth had delirium (n = 24), of whom 12 had both delirium and dementia. Thus, just under half of patients (n = 56) on this ward had a cognitive impairment on admission. Just over half of patients had delirium during their stay (n = 64), although for most of them the mean number of days in delirium was 3 (range 1–27 days). Only one-quarter (n = 32) of patients were identified as having no cognitive impairment either on or during admission.

The reasons for admission are shown in Figure 2, although for most patients multiple factors (medical, functional and cognitive) were implicated.

FIGURE 2.

Rivermead: reasons for admission. COPD, chronic obstructive pulmonary disease; SOB, shortness of breath.

A further indication of a patient’s degree of frailty is their discharge destination (Figure 3). Ten per cent of patient discharges involved a new long-term care admission; a further 10% of patients were discharged to intermediate care for additional rehabilitation; and 2% died in hospital.

FIGURE 3.

Rivermead: discharge destination.

Rivermead: care culture

At the start of fieldwork (February 2014), Rivermead had a stable ward leadership and purpose. The staff group had worked together for a long time (median 4 years and 3 months, and up to 25 years among those completing the Climate of Care questionnaire; n = 15), reflecting the broader profile of a stable team.

The responses to the Climate of Care questionnaire indicated that staff shared an explicit philosophy of care and culture of caring, were respectful of each other’s skills and knowledge, and valued patient involvement in decision-making. The ward manager was viewed as someone who inspired confidence, consulted the team about daily problems, acted in a caring manner towards the team, was explicit about the care standards expected and set an example in hands-on care, actively coaching individuals. At ward level, staff felt supported and engaged, a picture reinforced from observation and interviews. The ward manager was a key driver in shaping the care culture: an experienced nurse and manager, she perceived her role in priority order as ‘looking after my patients’; ‘looking after my staff’ (emotionally, developing knowledge and skills, and encouraging them to participate and further their careers); and ‘as a manager’, ensuring compliance with trust policies and procedures. She had undertaken advanced training in dementia, and conveyed, in values and practice, a commitment to providing individualised care to patients.

There were mixed views regarding staff relationship vis-à-vis the trust. The downside of being located at a distance from the main hospital site was poor communication and links with senior or specialist colleagues. There was a sense that colleagues in the main hospital lacked understanding of the ward’s purpose (rehabilitation), and had little appreciation of their particular skills and the complexity of the patients with whom they worked. Positively, they valued the sense of camaraderie engendered among staff in the smaller site and the greater autonomy and predictability of the work. External pressures on day-to-day-work were viewed as constraining their ability to deliver the quality of care they aspired to provide. Among some staff, a source of stress was the lack of time to reflect on ways to improve care delivery. As the bridge between the hospital and ward, senior staff, particularly the manager, sought to contain the tension between the priorities and requirements of the trust and their espoused values on providing individualised care and supporting staff. This, along with their physical separation from the main hospital, meant that the ‘troubles’ of the trust around meeting performance targets, and the changes at trust and senior management team level, were less salient for the ward team. Furthermore, the stability of the staff team meant that although there was recourse to Bank and agency staff to cover some shifts, hospital staff vacancy levels had less of an impact on Rivermead than on other wards with an unstable staff group.

Cedar: structure and organisation of care delivery

This orthopaedic trauma ward had moved to a temporary location at the end of 2013 and there was uncertainty surrounding its future purpose. Subsequently, following bed closures and ward remodelling, a new ward was created, involving half of the beds on a former older people’s ward and half for older trauma patients. This ward came into existence in August 2014 and agreed to participate in the study, with fieldwork commencing in September 2014. Given the late entry of this ward into the research, it was agreed that PIE implementation would occur over 9 months rather than the 18 months originally envisaged.

Of the 28 beds on this ward, 14 were for patients aged ≥ 60 years with fractured neck of femur; the remainder were for older, medical patients.

Patient profile

During baseline, there were 190 admissions to Cedar over 7 weeks, four of which were readmissions (approximately 27 per week). The median length of stay was 7 days (range 1–47 days), facilitated by access to ‘step-down’ beds (Rivermead or similar). Table 2 provides a summary of the patients’ sociodemographic characteristics.

TABLE 2 - Cedar patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
29–64	14 (7)
65–69	5 (3)
70–74	15 (8)
75–79	25 (13)
80–84	47 (25)
85–89	49 (26)
90–94	27 (14)
≥ 95	8 (4)
Gender
Men	77 (41)
Women	113 (59)
Residence prior to acute admission
Own home	58 (31)
Residential care home	4 (2)
Nursing home	10 (5)
No information	118 (62)

Just under half of patients were in advanced older age (44% were aged ≥ 85 years). The majority for whom data were available had lived in their own homes before being admitted to hospital.

On admission, at least one-third (n = 62; missing data, n = 96) of patients had an ongoing cognitive impairment: 41 with a diagnosed dementia and 21 assessed with probable dementia. Eight patients were identified with delirium, of whom seven also had dementia (missing data, n = 120). For those patients for whom data were available (approximately half, n = 94), just over half (n = 50) had a cognitive impairment on admission. Of the data collected on delirium during admission (missing data, n = 157), 18 patients had delirium (median of 3 days in delirium; range 1–12 days).

The reasons for admission to Cedar are shown in Figure 4. The pattern reveals a more acute profile than in Rivermead, suggested also by the relatively high proportion of patients who died during their admission (15%).

FIGURE 4.

Cedar: reasons for admission. COPD, chronic obstructive pulmonary disease; SOB, shortness of breath; UTI, urinary tract infection.

The data on discharge destination (Figure 5) reflected the fact that a significant minority of patients on this ward were discharged to the ‘step-down’ ward when they were medically stable but still insufficiently recovered to return to their usual residence. The symbiotic relationship between the acute older people’s wards and rehabilitation wards such as Rivermead contributed to the very short length of stay on Cedar, the ‘step-down’ wards providing additional therapy for patients deemed not safe to be discharged.

FIGURE 5.

Cedar: discharge destination.

Cedar: care culture

The ward manager had spent most of her working life nursing older people and, like her colleague on Rivermead, was an experienced ward manager. Although a number of team members had come with her from a care of older people’s ward, melding a staff group into a team with a dual focus on complex medical conditions and orthopaedic trauma was still a work in progress when fieldwork commenced. The level of patient dependency was regarded as high, and the physical demands of providing care weighed heavily on staff, as many patients required two people to assist them with personal care.

In interviews, the ward manager conveyed a strong ethos of individualised care, tempered with an acknowledgement of the reality of the multiple and simultaneous demands on staff. The work of a ward manager was presented as balancing different and competing interests: meeting hospital targets regarding flow and spending time with patients, reducing falls risk and undertaking rehabilitation work towards functional improvement.

From their responses to the Culture of Care questionnaire (n = 14), staff echoed the priority that their manager attached to individualised care, treating patients with dignity, putting effort into providing good-quality care, meeting patients’ care needs, and involving patients and caregivers in decision-making. They valued the skills, competence and support of colleagues; had trust and confidence in each other; communicated and worked well as part of the MDT; and felt able to bring up issues and challenge practice. The manager was perceived as inspiring confidence, consulting with the team about daily problems, acting in a caring manner towards staff, being explicit about the care standards expected and setting an example in hands-on care.

Similar to colleagues on Rivermead, staff on Cedar perceived a gap between their aspiration to provide individualised care and the inadequate support and resources for spending time with patients and developing new skills. In comparison with Rivermead staff, they were less critical of the resources available to them at the hospital level, particularly valuing the access they had to expertise on aspects of care. This probably reflected their physical proximity on the main hospital site, being closer to decision-making structures. For senior ward staff, particularly the manager, hospital priorities on patient flow, reducing length of stay and cutting beds imposed a heavy burden, as they were closer to the everyday impact of these pressures.

Organisation

Valley NHS Foundation Trust provides acute, rehabilitation and community services in hospital and community settings for a geographically dispersed rural and urban population of > 200,000. It covers an area of > 700 square miles in north-west England and acquired foundation status in 2010.

When the study began, the main hospital had 395 beds for acute, elective and specialist care, and some 55,000 A&E department attendances annually. It had a strong local presence and community connection, reflected in its large volunteer base, with around 400 volunteers carrying out over 30 different roles across wards and departments. The trust also managed three community hospitals in different locations across its wide area.

Annual reports between 2010/11 and 2013/14 revealed an organisation consistently meeting national targets on finance and governance, clinical safety and effectiveness and which had been operating with a relatively stable trust leadership. Similar to City Trust, Valley had invested heavily in the modernisation of its built environment, including improving the ward’s physical environment.

As in other study trusts, NHS reorganisation posed considerable uncertainty for forward planning. Simultaneously with organisational upheaval, the following challenges were noted: delivering safe and quality care, and managing a 3–5% year-on-year increase in volume demand on services alongside the requirement on providers to find efficiencies of 4–5% per annum to absorb the growth within a fixed overall real-term budget. The dual strategy pursued in meeting these competing demands was to cut costs and develop new services to grow income.

Care of people with dementia

Local strategic priorities included reducing ‘slips, trips and falls’, which were viewed as a particular risk for people with dementia. Initiatives included the development of a system to more accurately monitor falls; the introduction of screening, assessment and evaluation tools to improve risk management processes; the purchase of monitoring equipment to reduce falls incidence; the provision of floor-level beds for patients at high risk of falling out of bed; and the inclusion of falls prevention and bed rail use in mandatory training.

For people with dementia, priorities were improving nutritional care and implementing the ‘Butterfly Scheme’. The former involved the expansion of an established ‘feeding buddy’ system with trained volunteers to assist vulnerable patients to eat and drink. The ‘Butterfly Scheme’, introduced in 2011, used a butterfly symbol to highlight the needs of patients with a dementia diagnosis; provided staff with practical guidance to meet the needs of these patients; and alerted staff to the collection and use of information from relatives to better engage with these patients and to inform care planning. The scheme was supported through mandatory training for the clinical and support staff who directly provided care.

Increasing staff awareness of dementia was prioritised through a programme of ‘dementia champions’, and the creation of more ‘dementia-friendly’ environments was pursued through ward refurbishment in 2013/14, supported by external funding.

Oak: structure and organisation of care delivery

Oak ward was a 28-bed, acute orthopaedic trauma ward. Around 30 months before the study, there had been a change in medical leadership: orthogeriatric consultants had been introduced alongside orthopaedic surgeons to provide comprehensive geriatric assessment and care planning for older patients. This involved separate MDT meetings on a twice-weekly basis.

Patient profile

There were 174 people admitted during baseline, of whom one was a readmission. This was a mixed-age profile (range 16–101 years). There were two distinct admission patterns: younger people with limb fractures typically sustained in motoring accidents, and people in advanced older age with fractured neck of femur and comorbid chronic health problems. The median length of stay was 9 days (range 1–78 days). Table 3 provides a summary of patients’ sociodemographic characteristics.

TABLE 3 - Oak patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
< 40	23 (13)
40–49	16 (9)
50–64	21 (12)
65–69	5 (3)
70–74	5 (3)
75–79	24 (13)
80–84	23 (13)
85–89	35 (20)
90–94	19 (11)
≥ 95	3 (2)
Gender
Men	77 (44)
Women	97 (56)
Residence prior to acute admission
Own home	148 (85)
Residential care home	4 (2)
Nursing home	19 (11)
Dementia unit	1 (0.5)
No information	2 (1)

Whereas around one-quarter of patients overall had a diagnosis of dementia or were assessed as having probable dementia on admission, this applied to 40% of those ≥ 80 years. Furthermore, 9% had a delirium on admission, the majority of whom (11/15) had delirium on dementia. Most people had been admitted from their own home; 13% were residents in a nursing or residential care home. Just over one-third of older people (aged ≥ 65 years) had delirium during their inpatient stay (36%), and the median number of days in delirium was 4 (range 1–28 days).

The reasons for admission to Oak ward are shown in Figure 6. Given the nature of the ward, admissions were dominated by orthopaedic trauma; among older people, the largest single reason was a fractured hip.

FIGURE 6.

Oak: reasons for admission.

Patient destinations on discharge are shown in Figure 7; 5% died in hospital and, of the rest, 9% went to a new permanent nursing or residential care home (although this applied to just under one-fifth of older patients).

FIGURE 7.

Oak: discharge destination.

Oak: care culture

During baseline (end of November 2013 to early February 2014), the ward team was undergoing change. Until then, it had been relatively stable, with a strong, supportive team ethos. The manager had been in post for 6 years, first as a junior sister and then as manager, and many of the nursing and care staff had worked together for ≥ 5 years. However, several nurses had left or were in the process of leaving when fieldwork commenced; difficulties and delays in recruiting new staff resulted in the use of temporary staff, which had an impact on workload and ward morale.

The manager placed high value on supporting her staff, considering that a ‘stressed’ staff member affected everyone, both colleagues and patients. Observation revealed a leader who was a visible and cheerful presence on the ward and who conveyed to staff that their work was appreciated. Delivering ‘quality care’ was seen to require attention to patients’ emotional as well as physical and care needs; furthermore, spending time with patients was conveyed as a necessary and valued part of nursing and care work. The current staffing difficulties were regarded as compromising the realisation of this vision. In her interview, the manager envisaged that an aspect of her role was to absorb pressures from senior management so as to sustain a calm and ‘happy’ care environment. This conception of the ward leader as ‘buffer’ between the demands of senior management and the day-to-day work of providing direct patient care was felt as isolating and stressful.

The change in the staff team was evident among those completing the Climate for Care questionnaire (n = 14). Respondents were equally divided between those who had worked on the ward for between 1 and 6 months and those who had been there for ≥ 5 years. Staff reported that they shared an explicit philosophy of care, that there existed a culture of caring for patients and that the psychological aspects of care were highly valued, as was the involvement of patients and their caregivers. Staff were generally negative about the adequacy of support services to enable them to spend time with patients and also considered that there were insufficient numbers of staff who had the knowledge and skills to provide quality patient care. This was felt across the board: among both long-standing staff members and those who had been in post for ≤ 6 months.

Respondents considered that the quality of care provided by the MDT met patients’ individualised needs and was responsive to the needs of caregivers; staff were viewed as treating patients with dignity and respect; there was good communication between team members; and people worked well together and were supportive of each other. The general view was that there was insufficient time to reflect on performance and ways of improving care delivery. This view was supported by observation: staff members were in continuous motion, a combination of the physical demands of patients who were immobile and staff shortages. External pressures acted as a constraint on delivering the quality of care that they aspired to provide. The team viewed the ward manager as supportive, a visible presence and a role model in delivering hands-on care. In contrast, their perception of the hospital management was negative: they considered that staff views were not listened or responded to, and that support for staff at ward level, when needed, was not forthcoming.

Ambridge: structure and organisation of care delivery

Ambridge was a 30-bed care of older people ward, although bed numbers were reduced to 29 during baseline because of staffing shortages. The ward was designated for patients with ‘acute and complex needs’.

Patient profile

There were 171 people admitted to Ambridge over 9 weeks (around 17 per week), of whom two were readmissions. (Baseline fieldwork extended a further 2 weeks here, compared with other sites, to take account of the disruption caused by the move to the new space.) The median length of stay was 11 days (range 1–89 days). Table 4 shows the patient profile.

TABLE 4 - Ambridge patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
65–69	1 (1)
70–74	6 (3)
75–79	18 (10)
80–84	51 (30)
85–89	50 (29)
90–94	37 (22)
≥ 95	8 (5)
Gender
Men	80 (47)
Women	90 (53)
No information	1 (–)
Residence prior to acute admission
Own home	144 (84)
Residential care home	19 (11)
Nursing home	7 (4)
Intermediate care	1 (–)

Over half of the patients were in advanced older age (56% were aged ≥ 85 years). One-quarter had a diagnosis of, or probable, dementia. From fieldwork (handover, MDT meetings and observation), this appeared an underestimate and inconsistent with findings from other care of older people study wards. Just under one-fifth of patients (n = 30) had delirium on admission, of whom most (21/30) had a delirium on dementia. Data on delirium incidence were not collected.

The reasons for admission to Ambridge ward are shown in Figure 8.

FIGURE 8.

Ambridge: reasons for admission. AF, atrial fibrillation; COPD, chronic obstructive pulmonary disease; SOB, shortness of breath; TIA, transient ischaemic attack; UTI, urinary tract infection.

Discharge destinations are shown in Figure 9. Two-thirds of patients returned to their own homes on discharge, a smaller proportion than in Oak, reflecting a higher number who died during admission (9%) and a higher proportion of new admissions to long-term care (13%) or to hospice/palliative care (2%).

FIGURE 9.

Ambridge: discharge destination.

Ambridge: care culture

Similar to Oak, Ambridge was a ward in transition. This had a spatial dimension (being parked in a temporary location while awaiting the move to its permanent, refurbished home) and a staff dimension (change in ward leadership and departure of experienced staff). Both contributed to disruption.

During most of baseline, the junior sister managed the ward. An experienced nurse, this was her first management post. She expressed strong commitment to providing care that treated ‘people as individuals’ and therefore placed store on finding out their concerns and wishes, including people unable to communicate through language. The combination of few experienced staff on the team, others in post only a short time, and reliance on temporary staff to cover shortages, exacerbated the difficulty in creating and sustaining a coherent ward culture. Whereas some on the ward had chosen to work with older people, and had knowledge and understanding of dementia, others had not.

There were only four respondents to the Culture of Care questionnaire: two had worked on the ward for 12 and 10 years, and the others had done so for 5 and 6 months. Their responses echoed the views of staff from informant conversations during observations and formal interviews. They conveyed common values: placing emphasis on caring for and supporting patients, and seeing the involvement of patients and caregivers as important. Yet they considered that there was an absence of resources and support to allow them to spend time with patients, and insufficient staff to provide quality care; and as a team, they did not review how to improve care delivery. Although supportive of each other, they reported an inability to challenge poor practice or to raise difficult issues with colleagues. They were most negative about the hospital, perceiving that their concerns and opinions were neither listened to nor responded to, and that assistance was not available when needed.

Organisation

Ironbridge Trust, similar to City, is a long-established foundation trust. It provides acute, elective, specialist and some community services to a core population of around 650,000. The location of a major trauma centre, it has > 150,000 A&E attendances annually. With around 2000 inpatient beds and employing some 16,000 staff, it is one of the largest trusts in the country and one of the biggest local employers.

Annual trust reports between 2010/11 and 2013/14 presented an organisation that, in leadership, efficiency, safety and caring, was successful and consistently meeting national and local targets. Similar to other study trusts, it presented as being challenged by the multiple and competing political, organisational, performance, financial and regulatory demands.

A major transformation in service delivery was initiated by the integration of adult community and acute services from spring 2011. A further integrative step from autumn 2014 involved combining the Directorate of Geriatric and Stroke Medicine and the Primary and Community Services Care Group, a union with a direct impact on our study ward.

Care of people with dementia

From 2011/12, the trust’s strategic priorities included improving the care of people with dementia and creating ‘dementia-friendly’ physical environments on care of older people wards.

In 2011, a new ward was established that was dedicated for older patients with dementia undergoing hip fracture surgery; and, in 2012, an integrated multiprofessional care pathway for patients with both dementia and delirium was implemented, supported by training. A discreet symbol was developed (as in Valley wards) to enable staff to recognise people with dementia; this was aligned with a personal information booklet, ‘All About Me’, launched in spring 2014, for use by patients with confusion/dementia and their caregivers.

Netherton: structure and organisation of care delivery

Netherton was an acute medical ward for those aged ≥ 70 years with comorbid dementia or delirium/acute confusion. It was not a specialist dementia ward:59 it did not employ specialist mental health staff, therapy input was similar to that in other care of older people wards, and it had no additional staff providing activities. During baseline, Netherton comprised 28 beds (including one bed for ‘winter pressures’, which was consistently in use during fieldwork).

Patient profile

There were 75 people admitted during baseline (an average of eight per week), and no readmissions. Most patients came directly from A&E or the assessment unit (70%; 52/75); just under one-third had been admitted from other hospital wards. The median length of stay was 41 days (range 1–126 days); extended lengths of stay were attributed to changes in the system for financially assessing patients, resulting in delays to some patients being discharged. Table 5 summarises the patients’ sociodemographic profile.

TABLE 5 - Netherton patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
60–64	1 (1)
65–69	0 (0)
70–74	3 (4)
75–79	9 (12)
80–84	14 (19)
85–89	30 (40)
90–94	16 (21)
≥ 95	2 (2)
Gender
Men	18 (23)
Women	57 (76)
Residence prior to acute admission
Own home (including sheltered housing)	66 (88)
Residential care home	3 (4)
Nursing home	5 (6)
Assisted living	1 (1)

Just under two-thirds of patients were in advanced older age, a slightly older profile than in other participating wards, and women dominated. Most had been admitted from their own home.

Most patients (n = 70, 93%) had dementia (the remaining patients had delirium but not dementia). Nearly two-thirds (60%; 31/52) of new acute admissions had delirium. Overall, 70% (54/75) had delirium during their stay, of whom nine (13%) were in delirium for ≥ 30 days. The median number of days in delirium was 16 (range 1–57 days). The reasons for admission are shown in Figure 10.

FIGURE 10.

Netherton: reasons for admission. UTI, urinary tract infection.

Discharge destination data were available for 44 patients (the remaining 25 still being in the ward at the end of data collection). These are presented in Figure 11.

FIGURE 11.

Netherton: discharge destination.

Of those for whom information was available, 25 (57%) were discharged to their own home; one-fifth involved new admissions to long-term, mainly nursing home, care.

Netherton: care culture

An established, stable team, half of them had worked together for at least the 8 years in which the current ward model had operated (acute, dementia care). This included senior staff (ward manager, senior sister and charge nurse) who had a wealth of knowledge and skill in caring for people with dementia, acquired through formal self-directed learning, personal and professional experience. They conveyed, through formal and informal mechanisms, a consistent and coherent team ethos and a commitment to the ongoing appraisal of work practices. Twice-yearly away days were held for staff to examine practice and identify problem areas to work on. Everyone was expected to attend (the away days were organised as two separate days to facilitate participation). Through these various mechanisms, new staff were enculturated into a style of team working that emphasised continuous improvement as ‘how we do things on this ward’. The ethos of person-focused care was reflected in the responses to the Culture of Care questionnaire, completed by just over half of the team (n = 18).

Consensus existed among respondents that the team shared an explicit philosophy of care; the psychological needs of patients were considered; involving patients and their carers was an important part of the work; and these values and expectations were communicated to new staff. There was agreement that patients were treated with dignity and respect; they experienced individualised care, and patients’ and families’ care needs were met (reflected in research observations).

Most respondents agreed that there were sufficient ward resources to deliver good care, including staff with the requisite knowledge and skills to provide it. Just under half regarded staffing levels as insufficient to allow them to spend enough time with patients.

They felt supported as a team, could rely on each other, were relatively comfortable about bringing up problems with colleagues and felt that they could influence ward decision-making. In agreement with staff on other study wards, there was too much work to do in too little time, but they particularly valued the learning and development opportunities available.

Strong, positive feedback was received about the ward’s leadership and management style, especially the support provided and clarity of expectations around achieving care excellence. Perceptions of the wider hospital were mixed, although views were more neutral to positive about the availability of training opportunities and access to resources.

Organisation

This foundation trust is one of the largest in England, employing around 11,000 staff (one of the top five employers in the region). It treats more than 1.2 million people and has 250,000 A&E attendances annually. Acute services are primarily located in one main hospital, with additional beds in two smaller hospitals. The study wards were located in one of these, comprising around 229 beds and providing a range of community, outpatient, inpatient and emergency care services. This hospital was situated approximately 6 miles from the main hospital, within a self-defined, organisational community (it has its own local authority) of 250,000 people.

Trust board annual reports from 2011/12 reveal an organisation in which balancing finances was a persistent and significant challenge. In 2013, the trust’s contingency fund was under pressure; the research site hospital was in overspend by 15% and had not delivered on its cost improvement programme. A plan to reduce costs by closing flexi beds was set in train; and the financial position of the hospital was in focus, including weekly rectification update meetings. The trust was also reviewing staffing models as a cost-saving measure.

By November 2013, the A&E waiting time performance target had been missed for five consecutive quarters, creating concern that Monitor would intervene and prompting micro-management to regain control. All elective inpatient surgery was temporarily suspended, which resulted in failure to meet the 18-week referral-to-treatment target and the creation of a backlog in excess of 1000. The research site hospital met A&E waiting time targets, but, with significant pressure on the trust affecting all sites, work was conducted to review capacity and demand.

A CQC inspection in autumn 2013, the results of which were published in January 2014, deemed the trust as requiring improvement on all criteria apart from ‘effective’. The trust board and senior management were under considerable pressure to balance development and redesign against significant challenges, including winter pressures, increased activity and securing financial recovery.

Care of people with dementia

Local strategic priorities during 2011–14 included a programme to deliver improvements to the care of people with dementia and delirium, including policy, ward environment and drug treatment. The dementia strategy and steering group selected key areas to monitor, including education, improved pathways, metrics, and the embedding of personalised ‘all about me’ documentation into practice. A supported integrated discharge team was established to facilitate the timely discharge of frail elders, and joint old-age medical and old-age psychiatry were to be available on every site. The trust dementia steering group placed emphasis on improving dementia screening rates, which were well below target. Of the trust hospitals, the research site had best compliance with screening: from 3% in September 2013 to 33% in December 2013 (the target was 90%).

A trust-wide ongoing programme of ‘dementia awareness’ training had been delivered before the start of the research, tailored to different staff groups, with the intention of utilising a group of ‘dementia champions.’ In recognition of the increasing number of patients with dementia, 30 mental health nurses were recruited to Bank staff, and some volunteers received specialist training in dementia in the main hospital, with a rollout plan to other sites.

Denton: structure and organisation of care delivery

Denton comprised 13 beds. A designated ‘enhanced recovery ward’, it was intended to provide dedicated support to people with dementia who were medically fit and mobile and would benefit from the expertise and calmer environment offered there. It was also part of the trust’s winter initiatives to manage patient flow. When fieldwork began, the ward was relatively new, having opened in January 2013. Early in the data collection phase, it emerged that long-term funding to the ward was insecure and that bed numbers were not included in the hospital total. The ward underwent waves of imminent closure threats. The impetus for keeping the ward open came from clinicians who were passionate about having access to a resource for people with dementia and from positive feedback from patients and relatives.

Patient profile

During baseline, there were 39 people admitted over 9 weeks (approximately four per week), of whom one was a readmission. Patients were admitted from other wards in this hospital, following an acute stay. The median length of stay was 25 days (range 4–79 days). Table 6 summarises the patients’ sociodemographic characteristics.

TABLE 6 - Denton patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
65–69	2 (5)
70–74	1 (3)
75–79	5 (13)
80–84	6 (15)
85–89	12 (31)
90–94	9 (23)
≥ 95	3 (8)
Gender
Men	15 (38)
Women	24 (62)
Residence prior to acute admission
Own home	34 (87)
Residential care home	4 (10)
Nursing home	1 (3)

More than half of patients were in advanced older age (median 88 years), and the majority were women. Most had lived at home before their acute admission. All (n = 39) had dementia; one-quarter (n = 10) had delirium (of whom seven had delirium on dementia). During their admission, most patients had delirium (n = 33); the median number of days in delirium was 11 (range 1–36 days). The reasons for admission are shown in Figure 12.

FIGURE 12.

Denton: reasons for admission. UTI, urinary tract infection.

A further indicator of the frailty of the patient profile is the discharge destination pattern (Figure 13), which shows that 16% of patients died and 38% went to long-term care.

FIGURE 13.

Denton: discharge destination.

Denton: care culture

Most staff had worked together since the ward opened, some 7 months prior to baseline. All had chosen to work there because they had an interest in dementia. Many of the resources on the ward – for example games and reminiscence materials – had been purchased through fundraising initiatives by ward staff. There was heightened goodwill among staff, patients and relatives in joint working towards such initiatives.

Ten Culture of Care questionnaires were completed: eight by nursing and care staff, and one each by a doctor and a housekeeper.

There was consensus that the team had a caring approach, shared an explicit philosophy of care, placed high value on the psychological care of patients and the provision of enabling support, and regarded the involvement of patients and caregivers as very important. There was agreement that the team had access to adequate resources and that they routinely engaged in reflection and discussion aimed at improving care delivery.

The perceptions of the MDT were almost unanimously positive, with good working relationships. There was a consensus that patients were treated with dignity and respect, that they experienced good individualised care and that patients’ and families’ care needs were usually met. Observation supported this picture, attesting to a style of collaborative work with families that embraced sensitivity and responsiveness to their anxieties and concerns. There was very positive feedback about the emotional and practical support offered within the team, with respondents unanimous about the trust, competence and confidence they shared, and that they felt comfortable about discussing difficult issues together.

There was a slightly mixed response with regard to balancing resources and demands. Although some considered that there was too much work to do in too little time, most agreed that they could follow best practice within existing resources. They were generally very positive about ward learning and development opportunities.

Universal, strong, positive support about all aspects of ward leadership and management was conveyed: the ward manager was reported to consult daily with the team and to be an ongoing presence who was involved in hands-on patient care.

The most variable responses were provided to questions about the hospital, indicating mixed views about hospital resources, access to expert assistance and the responsiveness of hospital management to staff concerns. It was largely agreed that the hospital provided adequate training opportunities and treated staff with dignity, and that staff had the authority to make decisions.

Comparing all wards in the study, Denton respondents were most universally positive, summed up in the comment of a HCA:

I have worked on [Denton] since it opened . . . I have never been as happy since I started working for the trust. It is a lovely well-run ward . . . patients get great care here.

Beech: structure and organisation of care delivery

This ward provides care to a mix of older people with complex medical needs and those who suffered orthopaedic trauma, primarily fractured neck of femur. During baseline, the ward usually had 24 beds: 12 trauma and 12 medical. There was an additional bay, open as part of the hospital’s capacity-management strategy, making a total of 31 beds (over five bays and three side rooms). Historically, the ward had undergone many changes over the years (of purpose, size, patient group and physical location).

Patient profile

During baseline, admission data were recorded for approximately 5 weeks, during which there were 60 admissions (12 per week). The median length of stay was 20 days (range 4–153 days). The physical and cognitive abilities of patients varied. For example, ability to mobilise spanned those who walked around the ward independently to people who were mostly in bed. Many patients were physically frail, requiring assistance to transfer or to mobilise. Table 7 summarises the patients’ sociodemographic characteristics.

TABLE 7 - Beech patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
65–69	1 (2)
70–74	6 (10)
75–79	7 (12)
80–84	11 (18)
85–89	10 (17)
90–94	17 (28)
≥ 95	8 (13)
Gender
Men	13 (22)
Women	47 (78)
Residence prior to acute admission
Private residence	50 (83)
Residential care home	4 (7)
Nursing home	1 (2)
Respite care	1 (2)

More than half of patients were in advanced older age (median 87 years; range 67–99 years). The majority had lived at home (including many who had caregivers) prior to the event that had brought them into hospital. Just over one-third had dementia (n = 21). Given the age profile of these patients, and based on observation, this probably underestimated the dementia prevalence (data were incomplete; data were missing for 27 patients. On admission, at least one-fifth had delirium (n = 12; missing data, n = 27). Of those who were assessed for delirium during their admission (n = 18; missing data, n = 42), almost half were identified as having a delirium (n = 8) for a median of 3 days (range 1–15 days). The reasons for admission are shown in Figure 14.

FIGURE 14.

Beech: reasons for admission. AF, atrial fibrillation; TIA, transient ischaemic attack; UTI, urinary tract infection.

Data on discharge destination are presented in Figure 15. Just under three-quarters of patients returned to their usual residence on discharge; a further quarter were equally divided between those who died and those who were newly admitted to long-term care.

FIGURE 15.

Beech: discharge destination.

Beech: care culture

The ward model was relatively new, and there was considerable ongoing work necessary, from the ward manager’s perspective, to construct a team and forge a common care ethos; this was made more difficult by a shortage of permanent staff and a reliance on temporary workers. This was reflected in the responses to the Culture of Care questionnaire, completed by five staff only (two nurses, one HCA, one geriatrician and one who did not provide his or her job role). Interpretations of the questionnaire data are limited, given the paucity of data; however, there are clear patterns in the responses that echo those derived from observation and interviews. Of all of the wards, the respondents here were most negative.

Generally, the respondents conveyed that among the team there was low value placed on involving patients and caregivers, insufficient basic ward equipment, inadequate support to allow time to be spent with patients, and no routine mechanism for improving care delivery. Regarding multidisciplinary working, there were mixed views: although communication was perceived as good, it was felt that patients did not experience individualised care (a view shared by the ward manager during an informant interview). With regard to decision-making, respondents considered that they had little influence in how the work was managed, and that there was too much pressure on resources and demands, such that there was insufficient time to deliver best care and pursue appropriate training. Although the manager was perceived as available and supportive, she was not visible in direct care provision. Respondents indicated that they did not feel well supported by the hospital, were not fairly rewarded for their work and were not treated with respect, and that management did not listen to their concerns.

Rose: structure and organisation of care delivery

Rose ward was for patients recovering from stroke who required medical care and rehabilitation. Non-stroke patients with medical needs were also admitted, including those with palliative care needs, particularly into the four single rooms, of which there was a shortage in this hospital. The ward had 23 beds (sometimes 24, with an additional bed in the male bay). It seemed, from the patient profile, that the ward was evolving from a specialist unit into a mixed medical ward.

Patient profile

During baseline there were 167 people admitted over 8 weeks (approximately 21 per week). The median length of stay was 5 days (range 1–91 days). Table 8 provides a summary of the patients’ sociodemographic characteristics.

TABLE 8 - Rose patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
18–64	48 (29)
65–69	12 (7)
70–74	19 (11)
75–79	14 (8)
80–84	36 (22)
85–89	21 (13)
90–94	11 (7)
≥ 95	6 (4)
Gender
Men	81 (49)
Women	86 (51)
Residence prior to acute admission
Own home	157 (94)
Residential care home	4 (2)
Nursing home	4 (2)
Care home other	1 (1)

The age (median 77 years, range 18–99 years) and physical and cognitive abilities of patients varied considerably. Given the age range and variability of recovery, there was potential for added complexity. Some patients had life-changing, newly acquired disabilities and others aimed to resume paid work following their recovery. Many of those currently on the ward required staff assistance to transfer or to mobilise.

The majority had lived at home before their acute admission. On admission, 13 patients (8%) had dementia and three had delirium. Among those assessed for delirium during their admission (n = 19; missing data, n = 148), over half were identified as having delirium (n = 11), for a median of 4 days (range 1–7 days). The reasons for admission are shown in Figure 16. Just under half of patients had suffered a stroke; the remainder had been admitted for various acute medical conditions.

FIGURE 16.

Rose: reasons for admission. CVA, cerebrovascular accident; FAST, Facial drooping, Arm weakness, Speech difficulties, Time to call emergency services; SOB, shortness of breath; TIA, transient ischaemic attack; UTI, urinary tract infection.

Data on discharge destination are shown in Figure 17.

FIGURE 17.

Rose: discharge destination.

The proportion of patients discharged to their usual residence was reflective of a population who, before admission, was slightly younger and healthier than that of other wards in this study. Additionally, patients here benefited from access to the Early Supported Discharge and Hospital at Home services, both of which facilitated the continuation of monitoring, nursing support and rehabilitation on acute discharge.

Rose: care culture

Based on observation and staff interviews, the manager had ward presence but limited patient contact. She was newly promoted and appeared stretched. She was supported by the sister and the charge nurse, both of whom were visible and active in providing direct patient care and had long experience of working on the ward. Several HCAs had been based on the ward for many years. The existence of a stable group of staff at different levels over time provided continuity and a strong sense of collegiate working. Morale was affected by staffing difficulties, as was working in a physical environment regarded as ‘tired’, cluttered and dark. The manager considered that her staff team were overburdened and was protective of them, for example being careful to ensure that feedback on performance and critical incidents was conveyed and reflected on constructively.

The responses to the Culture of Care questionnaire (12/30 questionnaires returned from nursing, medical and therapy staff) indicated mixed views about the care and ward working environment. Positively, it was reported that the team shared an explicit philosophy of care and that the involvement of patients and caregivers was valued. However, staff considered that there was inadequate support to allow them to spend time with patients. They conveyed a positive sense of providing good care and working as a team, for example indicating that staff worked well within the MDT, provided individualised care and treated patients with dignity and respect. Respondents were largely neutral about team support, raising difficult issues and being able to participate in ward decision-making. There was much disagreement about the extent to which resources and demands negatively affected work and care, although there was agreement on the availability of, and access to, training. Ward leadership was perceived as positive, although, in comparison with City, Ironbridge and Seaford wards, this was expressed in less enthusiastic terms. The perception of support from the hospital was mixed; respondents were generally positive about being able to obtain expert assistance when needed and being treated with dignity and respect.

Organisation

Seaford Trust is one of the largest in England, with five hospitals, spread over a wide geographical, socioeconomically diverse area and serving a population of approximately 759,000. The three acute hospitals (two of which were involved in PIE) employ > 7500 staff and serve a bed base of 1107. It was awarded foundation trust status in 2009.

Trust annual reports from 2010–11 to 2013–14 present a picture of high performance and future uncertainty that was similar to that in other participating trusts. In the recent past, it had received the accolade of Trust of the Year, was the top performing hospital in the Healthcare, Excellence and Leadership Awards and met all of Monitor’s governance and financial targets. In the 2 years from 2012 to 2013, the pattern of high performance was seen to persist, although the foreshadowed difficulties included an increasing demand on services, particularly as a result of ‘winter pressures’.

A major strategic initiative pursued was the development of a Shared Purpose Framework and set of values that would guide practice, which was launched in 2013, when the study began. This was intended as the starting point for a programme of cultural change and was aimed at engaging staff at all levels around four key purposes: person-centred care, safe care, effective care, and creating an effective workplace culture to sustain and enable quality improvement.

Care of people with dementia

Dementia per se was not identified as a key priority in trust annual reports until the 2012/13 report. This announced two key initiatives: the establishment of a dementia team and a new Enhancing Quality Programme pathway for dementia, both launched in early 2013. The dementia team comprised a matron, two experienced nurses and a consultant working closely with them to lead improvements and support the implementation of the dementia strategy. The matron and specialist nurses were attached to each of three district general hospitals in the trust, including the two study wards.

The dementia team was in post for < 1 year when the research began but had already instigated several initiatives that were quickly and widely acknowledged. In 2013, they won the trust’s Caring Award for their work, which included a trust pathway for dementia care, staff dementia training, a ‘confusion pathway’ and a pocket guide to preventing falls and managing confusion, in collaboration with the falls matron.

Poplar: structure and organisation of care delivery

The easternmost site, where Poplar ward is located, is an acute hospital with 388 beds providing a range of emergency and elective services and comprehensive trauma, orthopaedic, obstetrics, general surgery and paediatric services. The hospital dates back to the 1930s, when the original building was constructed. Between 1996 and 1998, most services were relocated and expanded into a new hospital building linked to the original facilities.

The hospital has several specialist units and new staff accommodation. Co-located on the site are mental health facilities for working-age adults and older people run by the local NHS and Social Care Partnership (mental health) trust.

Poplar is a 22-bed orthopaedic ward for older people. Most patients are admitted as emergencies following a fall and the sustaining of a fractured neck of femur.

As baseline data collection started, the ward clerk, who had agreed to assist with accessing the ward profile data, sustained a fall and was hospitalised with bilateral ankle fractures. She was off sick for the whole period, which made data collection a challenge, as the researcher did not have access to the electronic patient data.

Patient profile

At baseline there were some 90 admissions, all emergencies, having a median length of stay of 15 days (range 1–36 days). For the reasons above, there are many missing data. Table 9 presents the Poplar patient profile.

TABLE 9 - Poplar patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
65–69	3 (3)
70–74	5 (6)
75–79	12 (13)
80–84	21 (23)
85–89	21 (23)
90–94	20 (22)
≥ 95	6 (7)
No information	2 (2)
Gender
Men	27 (30)
Women	61 (68)
No information	2 (2)
Residence prior to acute admission
Own home	46 (51)
Extra-care housing	1 (1)
Intermediate care	0 (0)
Residential care home	10 (11)
Nursing home	0 (0)
No information	33 (37)

Patients were mostly aged ≥ 80 years and female. The majority were admitted from home following a fall. According to documentation, approximately 50% of patients had a degree of cognitive impairment at any one time. This is supported by staff interview data, although the numbers appeared to fluctuate upwards. Data on delirium were not collected. Figure 18 reports the reasons for admission.

FIGURE 18.

Poplar: reasons for admission. ‘Other’ includes dislocated hip, hip fracture, wrist fracture, displaced neck of femur, cellulitis.

Data on discharge destination are presented in Figure 19.

FIGURE 19.

Poplar: discharge destination.

For more than half of patients discharged at baseline, their destination was not recorded, reflecting difficulties with data collection. Nor is there any information on discharge for those patients remaining on the ward at baseline. Of those for whom information is available, some three-quarters were discharged to their usual residence, and the rest were divided between intermediate care placements, transfers to another hospital and long-term care.

Poplar: care culture

Twelve questionnaire responses were received from the 20 distributed (mainly nurses and HCAs, with one volunteer and one student nurse). Although most had not worked for long on the ward (seven had done so for < 2 years), most had worked in the hospital for between 5 and 21 years and were experienced in the care of older people.

Responses showed strong agreement on the existence of a shared philosophy of care that valued psychological support and involving patients and caregivers. It was considered that the ward had sufficient equipment and skills among the team. The interviews further suggested a strong team who could rely on each other, as well as shared information and skills. Team members considered that they had influence on decision-making and that there was time to share task-related information. There was strong support for the leadership and management style of the ward manager in all areas. In particular, she was seen as an ongoing ward ‘presence’, supportive and protective of her staff and keen to ensure that everyone received appropriate training. She was also viewed as keen to appoint staff with a particular interest in dementia care. The dementia nurse linked to the ward was highly respected and much appreciated. She had worked on Poplar prior to her role change and was well known.

Most staff agreed that there was too much work to do in too little time and without adequate resources. They generally felt that they were not listened to by senior managers, and nor were they rewarded for their work. This is largely supported by comments from the 2014 CQC inspection regarding leadership (which included evidence of bullying and harassment). Moreover, there was some resentment about a recent ward move from an environment that staff perceived as more conducive to the well-being of patients, and a sense that other wards, less skilled in dementia care, tended to send people with dementia to their ward.

Crane: structure and organisation of care delivery

Crane is a 26-bed ward based in the westernmost hospital. Opened in 1979, the hospital was part of a plan to expand the town and surrounding area in the late 1960s by relocating people from London. It employs about 2500 people and has 476 beds. It offers a wide range of specialist facilities and services, including a shop run by Friends of the Hospital, a volunteer-run society, providing basic goods and services to patients and visitors, which raised £2.5M in donations.

Patient profile

There are a large number of days towards the end of the baseline for which there is no documentation, as the acting ward manager, who had been maintaining records, left around this time.

During baseline there were 121 admissions. The median length of stay was 8.5 days (range 1–64 days). The sociodemographic patient profile is presented in Table 10.

TABLE 10 - Crane patient profile: sociodemographic characteristics

Characteristic	n (%)
Age range (years)
< 65	1 (< 1)
65–69	1 (< 1)
70–74	4 (3)
75–79	23 (19 )
80–84	27 (22)
85–89	37 (30)
90–94	20 (16)
≥ 95	8 (7)
Gender
Men	69 (57)
Women	52 (43)
Residence prior to acute admission
Own home	92 (75)
Intermediate care	1 (< 1)
Residential care home	15 (12)
Nursing home	7 (6)
No information	6 (5)

The majority of patients were in their eighties, and there were slightly more men than women. Three-quarters had been admitted from their own homes. From documentation and staff interviews, at least half of the patients had dementia, but this fluctuated upwards to over three-quarters at times.

The reasons for admission are shown in Figure 20. Many included multiple pathologies, and those categorised as ‘other’ included undifferentiated diagnoses, various cancers, and heart and autoimmune diseases, plus long-term conditions or infections.

FIGURE 20.

Crane: reasons for admission. ‘Other’ includes cellulitis, heart failure, shortness of breath, cancer and urinary tract infection.

The data on destination at discharge (Figure 21) indicate that most patients returned to their usual residence, although there is considerable missing information, including in respect of those who remained on the ward at the end of baseline data collection. There were also a number of deaths on the ward during this time.

FIGURE 21.

Crane: discharge destination. IC, intermediate care.

Crane: care culture

Five responses were received from the 13 Climate for Care questionnaires distributed. This is a small and unrepresentative sample, limiting interpretation of the findings. Two of the respondents were experienced, one having spent 12 years on the ward; the others had been there for < 1 year.

The questionnaire responses and interviews suggested that there was support for a shared philosophy of care; psychological aspects were highly valued, expectations were communicated and staff were knowledgeable. There was support for the effectiveness of the MDT, with occupational therapists being particularly valued as a fount of information. Staff reported trying to shift from a task-orientated approach towards supporting patients’ families. Their approach to dementia care was shaped by the lead dementia specialist nurse based at the site and considered an enormous support through the provision of education and advice. She saw her role as offering a resource for practice-based learning and modelling care. Her work was acknowledged at senior management level, including by the deputy nurse for quality and the senior consultant. As with Poplar, there was support for the leadership style of the ward manager, which, at baseline, was an ‘acting’ one. Overall, the picture depicted was of an enthusiastic, motivated and optimistic workforce.

Less positively, there was ambivalence about the availability of wider trust support. Respondents considered there to be too much work for the time available and that resources were inadequate. In particular, there was insufficient time to sit with individual patients, it was difficult to take staff from the wards for them to undergo training, and the physical environment was regarded as noisy and cramped. Similar to Poplar, ward staff complained of a lack of social space and the loss of a bathroom. They also expressed frustration at the paucity of step-down facilities, which meant that patients remained on the ward longer than was ideal for their well-being.

‘Individualised care’

On all wards, staff typically conveyed person-centred as ‘individualised care’, ‘seeing the person and not the medical condition’ and ‘tailoring care around the person’. The term was often presented as a short-hand descriptor of ‘good practice’ or ‘quality care’ that staff aspired to deliver:

All my patients are elderly and very, very vulnerable, and they might not have dementia as their diagnosis; a little bit forgetful . . . scared . . . muddled; they’re out of their own environment . . . I firmly believe you give [person-centred care] to everybody that you nurse.

Manager, Rivermead Ward, City Trust

Multiple meanings were ascribed to the term ‘person-centred’ that have been variously referred to as according respect, acknowledging people’s essential human dignity, treating people as individuals and finding out what is important to them relevant to their treatment and care. Staff who had been in practice or had undertaken professional training prior to the policy use of the term ‘person-centred’ tended to refer instead to ‘individualised patient care’. Yet the meaning they attached to ‘patient-centred’ was no different to what others referred to as ‘person-centred’.

‘Person-centred’ care on acute wards

Pressed to elaborate on the meaning of ‘person-centred’, interviewees offered examples that varied from addressing patients by their preferred name, being flexible about such care routines as washing and serving breakfast, providing a choice of meals, respecting privacy, personalising bed space (family photographs) and involving patients in decision-making on treatment and care. Thus, the content varied from the courteous and respectful to meaningful engagement based on knowledge of the person. Different professionals on different wards placed more or less emphases on these various dimensions of ‘person-centeredness’.

Generally, HCAs and less experienced staff appeared unfamiliar with the term:

Person-centred care – do you mean looking after a certain person?

HCA, Ambridge Ward, Valley Trust

And:

Is that where you . . . set out a care plan for that person . . . I think that’s what it means, setting out a care plan for that person that has that condition.

HCA, Poplar Ward, Seaford Trust

Unfamiliarity with the term or an inability to articulate what it meant did not imply a lack of attention to the person or denote depersonalising care practices. On all wards, individual care staff accorded value to spending time with patients, building trust and engendering a sense of safety. Despite a lack of formal dementia training, these workers explained how they drew on personal knowledge to make a connection with how the patient might be feeling. From her own experience of being an inpatient, one HCA emphasised the emotional and practical work necessary to assuage patients’ anxieties:

When I’ve been to patients here, and you said you’d be back in a minute and you weren’t, I didn’t understand how frustrating that was until I experienced that myself. For somebody to say to you: ‘I’ll be back in a minute’ and then you never see them again and then all of a sudden this person’s gone off shift. [And] not being told your diagnosis, what’s going on . . . well they don’t get angry. But I was angry that I was just left there . . . So now I say with their permission, ‘do you want me to go and see what’s going on for you’? I will then come and read the Kardex (the patient’s case notes) and if I can give them any information . . . because a lot of them think that they’re being shoved into homes . . . Just to know that they’re still in control because to lose control must be awful.

HCA, Rivermead Ward, City Trust

An empathic connection with the patient’s experience informed an approach to action aimed at reducing uncertainty and perceived loss of control.

‘Seeing the person, not the dementia’

Most participants did not distinguish between ‘person-centeredness’ for patients generally and for people with dementia, for whom ‘person-centred’ translated into ‘seeing the person, not the dementia’:

. . . it’s about . . . the individual as a human being before they are a condition or a set of circumstances, and before you attach the labels and treatments you acknowledge that they are a person . . . I’m not sure that I would particularly say that the person with dementia is different . . . although it might be more challenging to identify how you respond . . . you might have to . . . get information from a wider range of sources.

Practice development co-ordinator, Ironbridge Trust

The way staff translated the espoused values of ‘individualised care’ into practice for patients with dementia revealed important differences between those who were knowledgeable about the condition, whether derived from experience or from training, and those with poor understanding of it. First, we describe staff views on what was necessary to provide individualised care and the conditions that shaped practice; and, second, we draw out the specific knowledge and skill required, and challenges posed, in delivering individualised care to people with dementia.

‘Knowing the person’

There was consensus that ‘knowing the person’ was crucial to delivering ‘individualised care’. Relevant knowledge, as a minimum, included aspects of the person’s life history, significant relationships, and likes and dislikes pertinent to care delivery (e.g. food preferences and their preferred name).

Routines

The daily routine and the tasks that constituted it had a similar shape and content on all acute wards, although the timing and degree of flexibility varied.

On some wards, the day for patients started early, at around 06.30, as night staff began washing and dressing the few patients who were awake. Activity in corridors and by the nurses’ station began with the arrival of the early morning shift at 07.00, although bays and rooms remained in semidarkness. Nursing handover from the night shift signalled the start of the new day. Each patient was briefly discussed: whether or not they were settled during the night, adverse events that transpired, action to be taken during the day. Snippets of information gleaned from patients during the shift were shared (e.g. ‘whenever we talk about home, she becomes teary and doesn’t want a care home’).

Staff were allocated a working area, with often one HCA for each bay and one nurse for two bays. After handover, the dispersal of staff to their work locations signalled the washing and dressing of patients and the preparation of breakfast. Most wards allowed a degree of flexibility in the order of these activities, but for all wards this was a pressured, busy time, the aim being to get patients ready before the arrival of the rest of the MDT. On some wards, the lights were turned on in the bays straight away; on others, the ‘lighting up’ of patient areas was pursued at a more leisurely pace. Staff moved from patient to patient, seeing who was awake and who needed help with toileting and washing. Domestic or cleaning staff were often present from this time.

Breakfast was typically served from 08.00 onwards: the clatter of trolleys added to the chatter of staff as they greeted patients, moving quickly in and out of bays and rooms with trays. On all wards, there were those who required either direct help or coaxing and encouragement to eat, tasks that could occupy a HCA with a single individual for 30 minutes. With one or, at most, two staff in a bay of five or six patients, several of whom required assistance, patients who just needed encouragement might be overlooked or left waiting. Although personal care tasks were the primary domain of HCAs, a strict division of labour between nurses and HCAs was atypical. On most wards, once it seemed that sufficient progress had been made with getting patients up, and breakfast was under way, at around 08.30 or 09.00, nurses would begin medication rounds. This was another pressured task, as different patients might need medication before or after eating, or pain relief in advance of seeing the therapist.

From 09.00, the pace of work shifted up a gear with the arrival of other members of the MDT – therapists and doctors, phlebotomists, pharmacists, porters, the ward clerk – adding to the cacophony of sounds: telephones ringing, buzzers buzzing, trolleys squeaking and patients calling out. To the external observer, the picture conveyed was of ordered chaos as each professional went about their specific tasks, working through with patients: doctors doctoring, therapists mobilising, nurses medicating and care staff moving quickly in and out of bays assisting with care tasks. The multiple and competing demands on staff during this period meant that calls for assistance were not always responded to speedily, particularly when a patient required the help of more than one staff member with personal care. Shouts of ‘too late; I’ve done it . . . the bed’s a mess’ were not uncommon, as assistance with toileting was not responded to quickly enough, adding to the pressure on staff and the distress of patients.

By late morning, patients were dozing or propped up with pillows, and those well enough to get up were dressed and sitting on a chair by the bed; the ward round was under way and therapists were occupied in one-to-one work with individuals. On several wards, the placement of tables and chairs in spaces in or between bays facilitated conversations between people; however, with the exception of Netherton and Denton (dementia wards), on most wards there was limited or no collective space for activities or interaction. Where communal rooms existed, they were rarely used by patients; more often they were used by staff. Wards varied in the extent to which patients used the ward space beyond their immediate location around the bed and in the bay. On some wards, it was unusual to see patients moving about the bays and in the corridors. Although partly due to lack of space and people being unable to move because of frailty, pain or hip replacement, it also reflected the care environment. On some wards, unsupervised movement was discouraged as it was seen to be symptomatic of ‘wandersome’ behaviour and, therefore, placing patients at risk of falling. On others, staff were accepting of patients walking about and they kept an eye out for, and were ready to provide supportive assistance to, individuals whose mobility was poor.

At midday, the round of mealtimes, toileting and dispensing of medication resumed. It was common for staff to have spent most of the morning into early afternoon (from 07.00 to 14.00) on their feet, apart from a short break mid-morning.

On several wards, open visiting from late morning saw a trickle of visitors, who were often encouraged to assist their relative/friend with eating. On most wards, visiting times were scheduled for up to 2 hours from 14.00, and up to 2 hours from 18.00. In practice, there was considerable flexibility: when patients were very ill, or when visitors had come a distance or had simply arrived early. The steady stream of visitors in the afternoon altered the ward rhythm as movement and noise emanated from patients and visitors. Their departure heralded the repetition of routine tasks: meals, clinical observations, toileting and then visiting again. The exiting of visitors in the evening was the signal for commencing night work. Formal handover from day to night staff reversed the sequence of the early morning, and another busy period began of putting patients to bed, toileting and administering medications. For patients, the dimming of the central lights in the bays signalled the day’s end, although movement within bays and between beds, bays and toilets continued into the night.

The apparent orderliness of the routine was deceptive; it was punctured by the moaning and shouting out of people in distress, a flurry of activity with the arrival of a new patient on a trolley, the pacing of a patient by the entrance to the ward anxious to leave and being remonstrated with by staff, or a sudden crisis event such as a patient collapsing or falling. For several people with a cognitive impairment on each ward, the period of late afternoon/early evening often presaged extreme agitation and distress. This manifested in shouting out or pacing up and down the bay and corridor. Such behaviour was disruptive to other patients. It was also difficult for staff to manage as many of them were near the end of a 12-hour shift. For staff, too, the semblance of orderly transfer from the day to the night shift could be disrupted and protracted: crises or staff shortages during the day meant that the ‘day’ jobs were not completed to time. Although the ‘burden’ of organising and managing completion of the shift ‘work’ fell heavily on senior ward staff, the unfinished tasks were also assumed by nurses overextending the day shift and adding to the ‘night’ work.

Practices

From a comparative analysis of fieldnotes within and across sites, we could discern a continuum of practices supportive of, or barriers to, sustaining personhood, and the organisational, spatial and care environmental factors that shaped them. We report on three inter-related illustrative features of practice: communication, toileting and continence (exemplar of personal care), and responding to distress. How staff communicate with patients may either contribute to patient distress or ameliorate it; and staff responsiveness to requests for assistance with toileting will be shaped by the ability of patients to convey their needs and by ward organisational and environmental factors. Our intention here is to draw out what might be the dimensions of ‘person-centeredness’ in practice in the specific context of the acute ward.

Engagement with Person, Interactions and Environment to programme installation

Both Seaford wards implemented PIE to timescale over 20 months (February 2014–October 2015, Poplar) and 19 months (May 2014–December 2015, Crane). Each was located around the middle of the continuum in their responsiveness to patients with dementia and being person-focused in their practice (see Chapter 4). The first step in programme installation occurred with the introductory workshops. The recent establishment of a specialist dementia team in the trust opened up the possibility of developing reciprocal linkages of mutual benefit to improve practice. Over several meetings with the PIE researcher, the dementia team lead expressed interest in the team becoming actively involved in PIE.

From installation to initial implementation: a critical stage

The first PIE cycle on Poplar began after the introductory workshop. Five team members, working in pairs, conducted 5 hours’ observation over different times of the day, including a weekend, in time spans of around 1 hour. For both the observers and the observed, it was perceived as ‘out of their comfort zone’ at the beginning. ‘Observers’ found it difficult to look without acting, and the objects of observation, the ‘observed’, were wary:

’Why are you doing that?’ . . . Now no-one takes any notice of us [observing] do they? . . . they see it as supporting them . . . and it’s made everyone aware of dementia.

Senior nurse, Poplar Ward

In this cycle, there was a gap of around 6 weeks between the observations and the development of action plans during which very little PIE work occurred. Local contextual and situational factors were implicated. The ward was temporarily re-sited because of refurbishment creating disruption for staff; this was the second move, and it came not long after the ward had moved to what had been intended to be its permanent home. An imminent CQC site visit also absorbed the energies of senior staff. Both contributed to difficulties for the team in identifying a mutually convenient time to meet, reflecting on their observations and developing action plans.

Informal discussions between individual team members helped to sustain momentum, as the practice development facilitator, supported by the dementia nurse, acted as a PIE ‘driver’. Together they reviewed the observations, and agreed a process for communicating findings and for extending discussion to the wider staff group. All observations were organised as claims (positive or favourable), concerns (negative or unfavourable) and issues (areas of potential disagreement). This method, from Fourth Generation Evaluation,117 was being employed with staff across the trust to engage different stakeholders on a common vision to improve care culture. Observations under each heading were captured on paper and displayed in the staff room, inviting comment.

The journey to initial implementation on Crane was considerably slower than on Poplar. Only one set of observations had taken place by the end of May, and no work on action-planning had occurred. Partly, this reflected situational factors at ward level: the acting ward manager, who had been supportive of the research, retired as implementation started. Ward leadership, for a period, was held by two acting-up sisters, one of whom was on leave when the initial implementation began.

A new ward manager with a practice development background was appointed in June 2014. The research team proposed a second workshop to focus on action-planning and to re-energise the PIE process. This was well attended, including by the incoming manager. It generated renewed enthusiasm and resulted in a planned meeting to organise further observations. Implementation resumed in August 2014, some 5 months after the first workshop, and the lead specialist dementia nurse assumed responsibility for steering PIE. Team members included a ward sister, a senior OT who attended the workshop, and several of her staff (an OT and a therapy assistant).

Although both wards experienced some disruption (physical relocation and transition to a new ward manager), which extended the timeline from installation to initial implementation, such changes were temporary. Critical at this juncture in moving PIE forward was the proactive role assumed by external ‘drivers’, and the role enacted first by the Poplar ward manager and then by the new ward manager on Crane, as a facilitator, supporting and encouraging staff to get involved.

From initial implementation to full adoption

The first action plan developed on Poplar established several broad areas to work on, with short- and medium- to longer-term action around communication, nutrition and activities. An important feature of action-planning was the way in which observations generated creative solutions to address multiple issues simultaneously. Box 2 presents an action plan around mealtimes as a social event, starting in a single bay. Addressing several goals, it was initially directed at communication. It set in train a stream of work beginning with small steps, and subsequently was expanded to all bays through the process of appraisal and review (including new observations). An additional goal to which the initiative contributed was patient mobilisation; getting people up and moving between bed and tables.

On Crane, implementation began in earnest following the second workshop and continued thereafter. For the first cycle there were four sets of observations, typically conducted in pairs, each 1 hour long, undertaken at different times and in different spaces. Plans arising from these included immediate action on nutrition: eliciting patients’ food preferences and ensuring the consistent use of food charts, initiating work to increase activities with patients, and including the engagement of ‘dementia visitors’ in the work (a recent initiative via the dementia team, these were volunteers who came to the ward once weekly) (Box 3).

This example is illustrative of the emergent nature of implementation and change. Action plans evolved through the dynamic of action and appraisal of progress; this in turn opened up further possibilities that had not been anticipated at the outset. Observation about noise from the radio resulted in a plan to elicit patient preferences about the music played on the ward; moved on to establishing music as a collective and purposeful activity at lunchtime; and then extended to plans for a regular monthly singing session, as patients appeared to derive pleasure from joining in. Similar to Poplar, a small step at one level could lead to a significant change in how things were usually done, with space and momentum created through the action-planning and review process.

Full adoption towards sustainability

On Poplar, by the conclusion of the study, ‘mealtime as a social event’ had not only been extended to all bays but had been incorporated into the ward routine. There were aspects of the initiative that required ongoing work based on experience: men appeared more reluctant to engage than women, and staff needed to consider which patients would get on with each other.

PIE observations continued periodically throughout the research. New issues that were identified and that resulted in immediate action plans to pursue included ensuring that patients did not feel isolated, ignored or disconnected (e.g. leaving curtains closed around patients post care delivery, being mindful not to exclude patients from discussions when within earshot, and regularly checking that the clocks in the bays were accurate in time and date).

A persistent focus of action-planning was the provision of stimulating activities for patients who were well enough to take part but insufficiently recovered to be discharged, including puzzles, newspapers, games and reminiscence resources (RemPod, Stroud, UK). Plans required identifying staff time and sources of funding to purchase materials. By the end of the third improvement cycle, reviews indicated some positive change: staff sitting with, and encouraging, patients to read and using pictures as prompts for conversation. Overall, these action plans were evaluated as ‘partially met’. Engagement was constrained by staff availability, which waxed and waned depending on demand, patient flow and level of patient acuity, which was unpredictable.

Over time, the composition of the PIE team changed as some members moved on and were replaced by new staff. Although never a formal member of the team, the ward manager continued to play a crucial enabling role in supporting and encouraging staff to get involved in PIE and in facilitating the communication of action plans to the wider staff group. Her commitment to PIE, the authority and legitimacy this gave to implementation work, and the support provided to enable staff to take time out to do observations were critical to the process. Her inclusive management style was a factor in involving HCAs directly in the PIE team. One HCA, when asked by the researcher why she felt able to take the initiative with PIE, commented, ‘the ward manager treats all her staff as equals’ and ‘there is no hierarchy here’. Another HCA, who took special interest in nutrition (which dovetailed with PIE action plans), also reported that her enthusiasm was a result of the ward manager’s encouragement. Apart from a period of absence (as a result of injury), the ward manager was present throughout the research but retired just as it came to a close.

On Crane, alongside ‘music as stimulating activity’, which expanded in form and content over time, new action plans identified and pursued included encouraging patients who were well enough and near to discharge to dress in their own clothes during the day, something that was not usual on this ward. This was viewed as supporting the normalisation of the transition from hospital to home and encouraging independence. It was also aimed at nursing, care and therapy staff to increase person-focused interaction while they helped the person choose their clothes and assisted them to get dressed. Making this happen involved negotiating with relatives to bring in clothing and ensuring that staff went about the work with individual patients.

Four PIE cycles were completed on Crane by the end of the study. Subsequent observations were shorter – around 30 minutes’ duration – to make the process manageable given the demands on staff time. Ward staff who were not part of the PIE team conducted some observations, extending the programme’s reach. An explicit objective of the manager, integrated into action plans, was to engage the wider staff team in discussion about change and provide opportunities for them to put forward ideas through team meetings and handovers. Inclusivity was valued as a principle of teamworking and a strategy to ensure that staff ‘buy-in’ to change. For example, the use of food charts was systematised through flagging up, on handover sheets, which patients needed them.

Trying out different ideas around activities to increase stimulation for patients was an ongoing topic of action plans; this included encouraging the use of reminiscence materials secured for the ward via the dementia team and creating themed ‘memory boxes’ to add to them. This work tended to proceed in fits and starts. Initially, it was very successful and had the support of the ‘dementia visitor’, a volunteer who came to the ward once-weekly, and two nurses who were keen to take part. As on Poplar, success depended on demands, and the investment of time and staff was an ongoing issue, particularly during winter pressures, when the team helped to cover an additional six beds. The cramped nature of the environment and lack of separate space for patients outside the bays was an additional constraint. In addition, on this ward, engagement in activities was assessed as being ‘partially’ implemented. The long-term plan, advocated by the specialist dementia team through the dementia strategy, was the recruitment of an activities organiser to work with a number of wards, including Crane, to enhance activities for patients.

The composition of the PIE team for a long time did not extend beyond senior nurses, OTs and therapy assistants. As therapists rotated between wards, this meant that individuals changed, although ongoing therapy involvement was maintained. More than 12 months into implementation, a staff nurse was ‘recruited’ to the PIE team. Even so, success in accomplishing specific action plans required the active engagement of front-line staff, including nurses and HCAs (e.g. music at lunchtimes, observational monitoring of nutrition intake and encouraging patients to dress in their own clothes). Strategies to enable ‘buy-in’ of the changes were both essential and contributed to translating plans into action on the ward.

Did implementation work as intended?

Person, Interactions and Environment change process

The full adoption and routinisation of PIE in Poplar and Crane enables us to draw out how and in what ways the change process as envisaged was carried through in practice, and, conversely, what was different from partial and non-implementer wards.

Implementation spiral

We envisaged implementation proceeding in cyclical and sequential fashion: through observation, reflection, action-planning, action and review. Interest was in whether or not the functions reflected in each step were necessary. In practice, the process took the form of a spiral, with simultaneous reviewing, amending action plans and actions in pursuit of them, alongside identifying new action plans, all at the same meeting. As actions were reviewed, some were abandoned as unworkable or an unnecessary addition; others became fully embedded in ward routines, becoming part of ward culture; still others were extended beyond the areas in which they were tried out. This was most pronounced on Crane ward (Figures 22 and 23). Thus, although each step was a necessary component of implementation, change occurred as an interactive spiral movement.

FIGURE 22.

Person, Interactions and Environment implementation on Crane, Seaford Trust. Dark green, researcher- and/or PIE leader-led actions and documentation; blue, PIE team members’ documentation; light green, PIE team leaders’ documentation.

FIGURE 23.

Person, Interactions and Environment implementation on Poplar, Seaford Trust. Dark green, researcher- and/or PIE leader-led actions and documentation; blue, PIE team members’ documentation; light green, PIE team leaders’ documentation.

Person, Interactions and Environment and normalisation process theory

On both wards, although implementation was neither smooth nor linear, each of the change steps was necessary and facilitated movement forward. Observation and reflection sensitised staff to aspects of routines and practices that they had previously taken for granted; once observers resolved their initial wariness about whether or not they were doing it ‘right’, observation was viewed as feasible and enjoyable and offered new insight. Thus, the process generated coherence among PIE team members and increased interest and motivation to move to the next step, namely action-planning (cognitive participation). Collective reflection and discussion on observations gave rise to creative action to effect change that was perceived as meaningful and worthwhile (collective action). The work of negotiating step changes, with different team members taking responsibility, reinforced coherence, participation and action. Actions involved different PIE members in ways that provided opportunities to reach beyond their existing roles (e.g. HCAs on Poplar). Simultaneous review of progress (appraisal), modifying goals and action plans, and extending existing work into new areas generated by experience of plans in action further enhanced coherence and cognitive participation. PIE was seen as a useful and effective means of making change happen.

Several aspects of PIE implementation on both wards add further insight into the mechanisms of change and, therefore, how the NPT framework could be further elaborated on in particularly complex and dynamic environments, such as acute wards.

First, to achieve ‘normalisation’ of new ways of doing things, strategies to generate mechanisms of change (coherence, cognitive participation, collective action and appraisal) needed to occur at multiple levels. It was not sufficient for the PIE team to make sense of the intervention to distinguish it from their current ways of working, and to buy into the work as a worthwhile investment; front-line staff also had to be engaged. Without their active participation, the depth and reach of action plans to achieve coverage across the ward would not happen. In this, the ward manager as ‘facilitator’ assumed a key role in enabling this to occur. On both wards, strategies to secure engagement in the change process beyond PIE teams were actively pursued through established, routine fora, for example handovers and board rounds as well as staff meetings and supervision.

Second, there was the dynamic environment of acute wards. Staff turnover was one aspect of this; another was the unpredictability of the pattern of ward work. Staff turnover affected both PIE teams: a HCA on the Poplar team moved on from the ward within a couple of months of the first cycle; and the rotation of therapists on Crane meant that individual team members changed regularly. There were no difficulties in recruiting new people to replace those who left, itself an indication that PIE was valued and worthwhile.

Unpredictability in the work pattern derived from the particular conjunction of patient need and profile at any one point in time, and the variable weight of demand on routines and practices. ‘Winter pressures’ were often cited as contributing to high demand, but in actuality duration was elongated as the additional beds extended from November to June. Sustaining understanding and ‘buy-in’ to PIE (coherence and cognitive participation) required processes for renewal. For example, involving an interested individual to carry out observation, paired with a more experienced team member, proved useful. Time and situational factors pertaining to work patterns and flow meant that coherence and cognitive participation might require reinvigorating prior to securing collective action. Maintaining morale and belief in the possibility of change during periods when demand on staff surged required flexibility in devising and implementing action plans.

Evaluating progress through appraising action, and delegating aspects of the work to different team members with appropriate support, had the potential to reinforce coherence, participation and collective action, and to extend and deepen reach of the action. Critical appraisal and reflection on why actions were not pursued as intended (i.e. whether alternative strategies to reach the same goal needed to be considered, or whether the goal should be changed) also improved action-planning. The generative mechanisms then operated interactively, between levels and over time, and not in a linear fashion. Sustaining momentum in a dynamic environment meant that the generative mechanisms theorised as necessary to ‘normalise’ new innovations in practice were not one-off achievements but had to be reworked until they were no longer recognised as new but were simply part of ‘the way we do things’.

Organisational context

At the trust level, PIE implementation was bookended by a negative CQC inspection report and an outcome that resulted in the trust being placed in ‘special measures’, although care on medical wards, including on older people’s wards, was judged to be good. Implementation took place between the first visit by the regulator in March 2014 and ended (for research purposes) with the results of the second visit in November 2015.

Although work relating to the CQC inspection, subsequent outcome and responsive action dominated the business of the trust board and senior managers, it did not appear to have a direct impact on PIE progress or the day-to-day work of the wards. Related issues, including changes in personnel at board and senior management, were similarly not seen as directly affecting ward organisation and practices. What appeared decisive was that, despite these events, trust strategic priorities remained unchanged. Action taken by the board in response to the findings of the inspection included intensifying the drive to engage staff at all levels in joint work to translate shared values into improving services. For ward staff, the factors that had most resonance for organisation and care delivery were demand on beds and the availability of resources to facilitate the timely flow of patients in and through the ward and transitioning beyond it. The systemic pressures staff on Poplar and Crane experienced in relation to demand and flow were a constant. They occurred before and after the trust was placed in special measures and were not affected by the remedial action taken by the board in addressing the CQC recommendations. We return to this in comparing the experience here with that of City Trust, which also received a poor CQC rating at about the same time as Seaford, with very different consequences for PIE.

In addition to the general problem of demand and flow, there were specific contextual features of Crane that influenced PIE implementation in practice: ward design and layout, and the cramped nature of the space. The latter in particular made PIE observations by staff difficult, as there was no obvious place from which to discreetly observe patients and staff. The lack of a day room, much wanted by staff, meant that, as with Poplar, finding a meeting venue was problematic. In the later stages, the OTs offered one of their rooms off-ward, when it was not in use. After the creation of hubs at each site (resource rooms where events were held and material on continuing professional development was freely available), meetings were held here. Ward factors also included the recurrence of norovirus, which twice meant ward closure and a halt to PIE activity (and delayed fieldwork).

Engagement with Person, Interactions and Environment

Rivermead and Netherton were initially ‘slow’ to get going (not dissimilar to Seaford wards) and then moved forward with initial implementation, albeit in fits and starts. Neither, for various reasons, proceeded to full PIE adoption.

In terms of practice with patients living with dementia (see Chapter 4), staff on Netherton were observed as being at the high end of the quality continuum; those on Rivermead were in the middle, similar to Seaford wards.

More so than in other PIE wards, Netherton in Ironbridge Trust had in place multiple facilitating factors that augured positive engagement with PIE. The introductory meeting with staff revealed a passionate commitment to continuous improvement in care delivery for their patients, directorate-level support to facilitate the research, and interest among the training and practice development team to assist with PIE. Senior ward staff had worked together over a long period and had created a cohesive team.

The ward team in Rivermead was similarly long-established and cohesive. The manager was very supportive of improving the care of people with dementia and had worked closely with the practice development lead (who was identified as the external ‘driver’); both had a particular interest in dementia care.

Senior staff on both wards had an interest in delirium, understood the value of delirium screening and worked to facilitate the collection of DOS data as part of routine nursing observations.

Initial implementation: a sticking point?

Following the workshop, at the instigation of the Netherton PIE team, the site researcher met with them to discuss the process and tasks. The ward manager suggested that she would lead PIE and that they would carry out practice observations to build their confidence during December. The first PIE cycle began in earnest with observations from mid-January 2014. These were conducted over a 2-hour period in pairs on two occasions, at different times of the day (mid-morning and afternoon). Staff initially felt inhibited and anxious, concerned that colleagues might act differently knowing that they were being observed, similar to Seaford. Afterwards, they met to share reflections.

Observations indicated positive features of practice (staff being responsive to and reassuring anxious patients; encouraging interaction between patients; offering choice of food and drink and going about it straight away; ward being clean; patients in beds appearing content; other patients being up and about in the day room and corridor). Feedback was provided to staff on shift and welcomed as acknowledging the strength of teamworking. Immediate action was initiated from these observations: the introduction of short, staggered staff breaks in the morning to ensure responsiveness during a period when staff were relatively invisible in the bays (observed through buzzers going unanswered, and patients being left waiting longer than usual for assistance), and attention to the temperature of the ward for patients who were inactive by offering blankets (staff who were constantly in motion did not notice this).

Over the following 2 months (February–March 2014), further observations occurred involving all seven PIE team members and practice development co-ordinator. The additional support that had been anticipated from staff at directorate level did not materialise. Apart from individual examples of a particular patient being ignored or a dozing patient not being offered a drink, observation elicited a picture of responsive practice, an image with which the ward ‘outsider’ – the practice development co-ordinator – concurred. A facilitation meeting between members of the research team, the PIE team lead and the senior charge nurse was held in early April 2014 to support the translation of observations into action plans. It was reported that plans on reducing noise in the morning and increasing activities were being developed from the second set of observations.

By summer 2014, 6 months after initial implementation began, progress had stalled. An examination of action plans from previous observations by the research team revealed difficulties in moving beyond a general ‘wish’ list to identifying specific steps and the devolution of responsibilities to achieve them. For example, although the use of volunteers to support activities such as reminiscence, exercise groups and games was identified as an action, as was increasing activities provided by staff to patients, concrete steps to take these forward were vague (‘manager to pursue’). Indeed, from informant conversations with PIE team members, although they concurred with the objective of doing more stimulating activities with patients, they felt that this could not be done with their existing staff complement. Similarly, engaging volunteers was viewed as a good idea, but the steps to make it happen had not been worked through.

The research team offered a further half-day workshop to consider action-planning and review, held in August 2014, when interim data collection was to start. Five PIE team members attended, including two senior staff. Anonymous ethnographic observations were used to explore practice in its organisational context. These were intended to provoke a discussion of factors shaping practice that might be a focus for change without ‘blaming’ individuals and to work through action-planning around concrete examples. Evident from discussion was the openness of the team to debate aspects of practice that were regarded as contentious (use of language and terms of endearment with patients). The workshop appeared to generate renewed enthusiasm as had previously occurred with Crane. It was agreed that they would develop an action plan for discussion with ward staff at the forthcoming 6-monthly staff away day in November (to which the researcher was invited). Thus, as the PIE implementation phase approached 9-month interim data collection, there was the prospect of resumption.

In Rivermead, similar to Seaford wards, nearly 3 months elapsed between the introductory workshop and the first set of observations. In May 2014, three pairs of staff (three HCAs, a housekeeper, a nurse and a ward manager) conducted observations at different times of the day: early afternoon, teatime and late evening, over 2 hours on each occasion. Reflections on observations occurred informally afterwards. Despite the number, range and length of observations, documentation revealed a relatively superficial portrayal of practice and goals for improvement, largely related to professional ‘visitors’ to the ward (pharmacists, phlebotomists, porters and medical staff). For example, a positive feature of practice was greeting every patient on entering the bay; this contrasted with what was regarded as poor practice, that ‘visiting’ professionals did not do likewise. Observers here were mainly care staff, posing the question whether or not their social location made it more difficult to raise issues of practice; in other sites, observations were undertaken by a more hierarchically diverse group. Yet feedback with the researcher revealed that observers had ‘seen’ and noted little interaction between staff and patients for long periods in the evening, but as it was interpreted as requiring additional staff to resolve it did not feature as a goal for change. ‘Knowledge’ of the problem was uncontested; their judgement that they were powerless to act on it had an impact on its utility as a focus of action-planning.

Informal conversations between virtual PIE members occurred to discuss observations, but no formal meeting to pursue action-planning took place. The practice development co-ordinator, who was to provide a steer for PIE implementation, had taken up the post of matron shortly after the PIE workshop. Although she had negotiated time for PIE, her working life was absorbed by organisational demands: staffing shortages and a wide-ranging reorganisation of ward models and bed closures following a poor CQC inspection (inadequate staffing being a particular focus of concern). The ward manager was also on sickness leave in the month following observations. Similar to Netherton, the PIE implementation record had stuck on observation.

Starting again

In Netherton, the away day in November 2014 was attended by half of the ward team (a second day was to involve remaining staff). The site researcher and PIE team facilitated discussion about PIE. Ward observations were considered in groups charged with identifying priorities and action plans arising from them. Deliberations (recorded for the PIE team to work on afterwards) were wide-ranging, generating lots of ideas, varying from staff training in working with patients deemed ‘challenging’ to reducing bed numbers to allow for more time with individuals. From subsequent discussions on the day, senior staff appeared preoccupied with changes occurring at ward level: what they perceived as increased medical acuity and dependency of their patient profile, a push from managers to speed patient flow (a recent managerial directive to all wards required reduction in length of inpatient stay), and concerns around staffing, particularly the loss of a qualified member of staff ‘taken away’ to cover neighbouring wards as the hospital entered ‘winter pressures’.

Following the away day, action-planning stalled again; the length of time since observations had taken place and the changes in the ward patient profile since then were arguments proffered by senior staff for carrying out a further set of observations. These were organised with external support from practice development/clinical education colleagues. This would be a new start.

A second PIE cycle was begun in February 2015 with ‘external’ ward observers. Observations were conducted over 1 week at different times and locations. As before, feedback was very positive, reinforcing the picture of person-focused practice. As a means of reinforcing and celebrating the quality of care delivered on the ward, and to their professional peers beyond, PIE observation was successful. The OT who had conducted some observations with colleagues took the initiative in writing up action plans from them but was reluctant to assume a leadership role in PIE, viewing her position as a ‘part-time’ member of the ward team (with nurses and HCAs perceived as the ‘core’).

In the months that followed, prompts by the site researcher revealed little further progress: the PIE team had not met. This remained the position at the end of December 2015. Their difficulty in developing action plans (recognised by senior ward staff) suggested that a contributory factor was a ceiling effect around identifying areas for improvement at least in the context of existing resources. The staff group (and senior staff) had many ideas for desired changes, but they were seen to require additional staffing.

Final data collection was delayed, as the timing coincided with a CQC inspection; interim collection had also been delayed because of infection and ward closure. The timescale for study completion meant that a further round of data collection was not feasible.

In Rivermead, informant conversations between the researcher and PIE team members between July and September 2014 revealed that ward staff were working day to day; there was neither the time nor the ‘headroom’ to do more than get through the daily routine. ‘Winter pressures’ beds (an additional five) were still open on the ward: they closed and reopened during the same week in September. The matron (external PIE lead) continued to be enmeshed in work around staffing, bed closures and ward remodelling. The research team offered a workshop to help with moving PIE on; this was not taken up.

With autumn behind and winter approaching, the anticipation of increased demand made the prospect of resuming PIE remote. Senior staff indicated that although the team had made positive changes over the years ‘in the small things’ (respectful of patients’ dignity and privacy), moving to the next step required a more fundamental examination of their practice and review of resources. Yet the daily pressures on staff made this ‘out of time’.

Late December 2014, the matron held an action-planning meeting with the PIE team. Action plans were discussed and the team was to develop and consider them with the wider staff group. The confluence of elapsed time since the observations, demand pressures on the ward and long-term sickness among established staff resulted in no further work on PIE.

Interim data collection began in January 2015, a heightened period of ‘winter pressures’. Observation indicated no discernible improvement in practice that might be attributable to PIE, unsurprising given that implementation had not been pursued.

Partial implementers: barriers and facilitators

Creating local engagement with Person, Interactions and Environment

Based on their response to PIE, we characterised non-implementers as ‘tentative’: wards that expressed continued interest in PIE but timing for change was wrong; and ‘disengaged’, wards that did not engage with PIE after reassessment of the decision to proceed (i.e. ‘true’ non-implementers).

Staff perceptions of Person, Interactions and Environment

Interviews with staff at the conclusion of the study showed that PIE had harnessed enthusiasm and engendered confidence, collaboration, a sense of empowerment and agency, including among unqualified staff, in trying out new ways of working. They also talked about the time pressures on them from increased throughput and higher patient dependency, both of which affected the time that they were able to spend with individual patients, thereby reinforcing the criticality of staff outside the ward in driving change. Given the timing (winter 2015), this probably reflected seasonal variation in demand, although the lengthening period during which ‘winter pressures’ beds remained open on most sites suggested ongoing high demand:

Because we’ve been really busy, we’ve had a lot more patients with higher needs and I think we’ve struggled . . . to achieve what we had been achieving . . . We come up with the ideas, but it needs someone [to say] ‘yes that’s OK let’s do that’ . . . Whereas if you’ve not got that support I think some of it might flounder.

HCA, Poplar Ward

Some staff were impatient at what they regarded as the slow speed of change. The lack of involvement of front-line staff in the Crane PIE team was a disappointment to some senior staff. Yet without the engagement of nurses and HCAs in pursuing action plans, PIE adoption would not have occurred and nor would its reach have gone beyond the PIE team. The Crane manager considered that PIE had stimulated reflection and action across the team as a whole, but that sustaining the process would necessitate renewal through the involvement of new staff in PIE. One proposal being considered at the conclusion of the study was to initiate new staff through the use of the observation tool alongside an experienced ‘observer’.

On both wards, plans were in train to continue PIE beyond the study conclusion. A further year on, PIE had been sustained in Crane, with the increased participation of front-line staff. The merger of Poplar with another ward encompassed plans to extend PIE to the joint ward.

Experience of care

Without exception, patients and relatives at baseline described their experience of care in positive terms: staff were kind and responsive when patients needed assistance; and relatives indicated that staff made themselves available to discuss their concerns, kept them informed about what was happening and involved them in decisions about the future. In the detail, we get a more nuanced picture of both experiences in context of what was viewed as ‘good’ and variability in responding to preferences and needs.

Did Person, Interactions and Environment make a difference to patient experience?

The five patient case studies following PIE implementation were purposively selected to include those who were frailer as critical exemplars of patients perceived to be more challenging and therefore particularly testing of progress made. All had limited ability to communicate verbally, although this varied depending on the time of day, and they deteriorated over time.

Days in delirium

Before the research, none of the wards had routinely used a screening tool for delirium; although City Trust was considering introducing the SQiD,119 a single question asked of relatives to assess onset of increased confusion. The collection of DOS data was inconsistent. Unfortunately, although there were some data on days in delirium for both Seaford wards at baseline, none was collected at subsequent time points. Thus, on both Seaford wards at baseline, one-fifth of patients for whom DOS data were collected screened positive for delirium, although, for most of them, the number of days in delirium was small (median = 4 days). Mid-way during baseline, both wards stopped collecting DOS data, deeming this too onerous to pursue. On Crane (care of older people ward), most patients who had screened positive for delirium at baseline had dementia (75%); the proportion was small (30%) on Poplar (orthopaedic trauma ward).

Both partial implementer wards – Netherton and Rivermead – collected consistent DOS data at baseline and interim data collection points. Although neither proceeded far with PIE implementation, it is of interest to examine delirium rates on both wards in order to highlight some challenges in collecting and interpreting results. Table 12 reports delirium rates at both time points.

Delirium occurrence (on admission and during inpatient stay) was considerably higher on Netherton than on Rivermead, which was unsurprising as most patients admitted had dementia (the few who did not had been assessed as having delirium). The co-occurrence of delirium and dementia was evident from patient profiles for all participating wards (see Chapter 3). Further people with dementia are at particular risk of developing delirium. 36,94,95 However, from multiple sources of data (observation, patient case studies), we found that care quality for people with dementia on Netherton was consistently high. Most people during their admission experienced delirium at some point, including the small number who were not admitted with it. The number of patients experiencing extended delirium (≥ 30 days) was very small and was similar across time points.

Although the overall number of patients with incident delirium across time points is too small to draw inferences about the relationship with care quality, the results highlight the scale of the problem of delirium among this ‘at-risk’ group, given the level of distress caused to patients120–122 and the challenges in providing appropriate care as a consequence. To make meaningful inferences about the relationship between delirium and care delivery would require the collection of delirium data over a longer period of time pre and post intervention with a sufficient sample size to make the necessary adjustments to reflect varying levels of risk.

Falls

Similar to delirium, falls data were not collected consistently in Seaford wards. Table 13 provides the number and rate of reported falls per 1000 occupied bed-days for Netherton and Rivermead wards at baseline and interim.

As discussed (see Chapter 2), risk factors for falls not only relate to the actions of staff, but patient-specific factors are also implicated, including advanced older age, dementia, delirium and frailty, as are practices regarding rehabilitation. 101 Furthermore, the pattern of falls fluctuates, and it is argued that pre- and post-intervention data need to be collected over ≥ 9 months to deal with potential confounders resulting from changes in staffing, case mix, activity or recording practices. 101 The threefold increase in the falls rate in Netherton, for example, between baseline and interim is likely to reflect the more dependent patient profile than at baseline. Field notes during interim did not indicate that the quality of care delivered had deteriorated. Similarly, the reported reduction in the rate of falls on Rivermead at interim does not, from observational data, imply that the quality of care had improved; rather, staffing difficulties that placed increased pressures on ward staff suggested that staff were able to spend less time with patients with dementia at interim. The multiple and interacting risk factors for falls (as for delirium) present considerable challenges in drawing inferences about the impact of an intervention to improve care delivery on falls rate. Here, also, larger samples to deal with confounders, and a longer timespan for data collection to take account of seasonal and case-mix variation, would be necessary to detect meaningful change.

Patients with dementia on acute wards: the challenge for care providers

The patient profile on wards dominated by older people (dementia, care of older people and orthogeriatric wards) revealed the scale of the challenge in acute hospitals to improve the care of people with dementia. People with a cognitive impairment, including delirium, constituted at least half of the patient profile. Many had moderate to severe impairment; they lacked the ability to communicate verbally, and required assistance with personal care and active support with eating and drinking. On every ward there were patients who called out repeatedly, seemingly in deep distress. Acute care for people living with dementia, therefore, cannot be conceived of as peripheral to, or disruptive of, the routine work of ward staff, but is a central feature of it. 23,24,29,30 As well as the health event that precipitated acute admission, these patients were in advanced older age and had multiple health problems; they were physically frail and had care networks that were often fragile. Achieving a ‘balance of care’54 in an acute setting required the delivery of clinical, therapeutic and personal care based on an understanding of the interaction between these multiple and complex needs, including cognitive impairment, and how they mediated the process of recovery. Although all participating wards had introduced training in caring for people with dementia, the level of skill required to respond to the complexity of need of their patient profile was not generally addressed on these programmes.

On all wards, medical, nursing and care work was generally perceived by staff as rewarding, physically hard, emotionally taxing and demanding of specific skills, expertise and knowledge of dementia and delirium. Some staff on most general wards (i.e. not the dementia wards) felt ill-equipped in terms of training and resources to provide appropriate support. They conveyed a gap between a broadly supportive and engaging staff culture at ward level and ambiguity and tensions between competing priorities at the level of the hospital that affected their day-to-day working practices. The width of the gap varied considerably between trusts.

Three out of 10 wards did not attain the RCN staff-to-patient ratio for safe working on older people’s wards,123 namely eight staff for 28 beds, a ratio of 1 : 3.5. Most did not meet the recommended ratio of registered nurse to HCA of 65 : 35. The picture was bleaker than this suggests in that, on several wards, the staff complement was maintained only through the extensive use of Bank and agency staff, with consequences for staff morale, the sustainability of a coherent care culture and the availability of ‘headroom’ to engage in service improvement. Of the two wards that successfully implemented PIE, both had a relatively generous staff complement, a ratio of registered staff to patients of 1 : 5.6 (Poplar) and 1 : 4.4 (Crane), although here, too, staff were often moved to support other wards when numbers were depleted.

Across most wards, environmental changes to create more ‘dementia-friendly’ spaces had occurred, making for more attractive environments that were easier to navigate. However, communal or patient space was extremely limited, and the layout of some wards meant that the scope for movement of patients to recover or sustain mobility was restricted.

‘Person-centred’ care: meaning, content and practice

Implementing Person, Interactions and Environment: interaction of micro-, meso- and macro-level factors

Despite the gap between the quality of health care that patients should receive and what actually happens, programmes to secure improvements in care quality have demonstrated inconsistent, variable and patchy results. 116,127,128 This in part reflects the fact that such interventions are more appropriately construed as events in multilayered and interacting systems, meaning that the organisational, political and social contexts in which the intervention is located are not simply background noise; they are part of the programme of change74,89,90,116,129–131 and interact with it. Other features of complexity include the multiplicity of programme components and their dynamic and emergent nature such that different groups interact with programmes in different ways.

PIE was fully adopted in only two out of 10 study wards, and had proceeded to innovation and sustainability in those two. To explore why some wards adopted PIE and others did so only partially or not at all, we examined the relationship between what was intended to happen and what actually happened, and the interface with the organisational context over time, using a stages-of-implementation framework. 69

Normalisation process theory

Normalisation process theory offered a useful framework to consider what might be the components of the PIE change process. However, even on successful implementer wards, strategies to generate mechanisms of change (coherence, cognitive participation, collective action and reflexive monitoring) needed to occur at multiple levels. Here, ward managers as change facilitators played a crucial role in engaging staff beyond the implementation teams in PIE, using routine communication mechanisms (handover and team meetings) to ensure their ‘buy-in’. Furthermore, ‘buy-in’ at all levels was not achieved once and for all. The dynamic nature of ward environments meant that, for PIE adoption to proceed, PIE teams were confronted with the task of developing systems for the renewal of staff engagement in sustaining the programme. The experience of ‘partial’ implementers also attests to the fact that the generative mechanisms hypothesised in NPT require ‘receptive’ organisational contexts to effect routinisation (i.e. they do not work in isolation).

Impact and outcomes of Person, Interactions and Environment

Evidence from observation and staff interviews on both full adoption wards suggests that the PIE Programme had a significant impact on practice. Specific changes, as for example, communal meals, increased staff knowledge of patients, staff–patient interaction, patient–patient interaction and nutrition intake, were introduced. Observation and interviews indicated that providing opportunities for patients to socialise facilitated patients conversing with each other as well as with staff. Although the content of patient–patient conversation was not always meaningful to the observer, it conveyed a sense of warmth, conviviality and sociability that was absent on other wards, and similar to what occurred on the dementia wards through communal dining. It reinforced the ‘friendliness’ and ‘looking out’ for each other expressed by some patients with their fellows, and was of demonstrable benefit. 141,142

There was also evidence that practice change had a positive impact on the experience of patients and caregivers, although the heterogeneity of need and severity of impairment meant that some of the more visible changes did not affect everyone equally. Indeed, findings throw into relief the varied needs of patients on acute wards living with dementia: opportunities for social interaction and activities may be valued by some but not all; for others, the use of touch and emotional connection with the person in ‘their world’ involving time spent in one-to-one contact may be more appropriate.

We are unable to draw any conclusions about the effect of PIE on clinical outcomes, as data on delirium and falls were not collected on these wards.

Learning from Person, Interactions and Environment implementation

Within the full adoption wards, PIE was effective in identifying a need for change and providing a systematic method to pursue action-planning and appraisal where organisational supports were in place. Observations generated creative solutions to address multiple issues simultaneously; action plans established a stream of work from small steps through the appraisal process; and the bottom-up/top-down approach enhanced sustainability and reach. 138

There are aspects of learning from the research and experience of PIE implementation to incorporate into refinements of the guidance and tools. First, the dimensions of person-focused practice from the research relating to communication, tasks and managing distress may provide an explicit focus for service improvement, albeit without constraining the flexibility and creativity of implementers. Second, the content of some of the PIE documentation may be simplified without undermining its function in the process. Thus, the distinction made in the action-planning documentation between short-, medium- and long-term action plans appeared unnecessarily complicated and unhelpful. A key factor in teams getting going and keeping going was identifying goals and priorities around which staff could be mobilised to act. Indeed, the difference between short-, medium- and long-term action plans is likely to be in the number of steps required to achieve longer-term goals. At the same time, progress in working through the steps to achieve a longer-term goal may involve a reassessment/review of the goal and/or how to pursue it. Third, while some simplification of the documentation is suggested, we consider that completion of observation, action-planning and review tools through initial and full implementation is helpful in systematising engagement with the process, as others have argued. 74 Fourth, it became evident early on that additional support and guidance to translate observations and reflections into action plans was required and provided, requiring the expansion of this part of the manual.