The process of practice redesign in delirium care for hospitalised older people: A participatory action research study

1.1. Why select PAR? 

One of the authors is an experienced PAR researcher. Based on Koch’s expertise in conducting over 20 separate PAR projects, findings strongly support that collaboration and involvement in key identified areas of concern (in this study it is older people and delirium) leads to sustained reform and/or practice changes (Koch and Kralik, 2001, Koch and Kralik, 2006, Koch et al., 2002). PAR principles guiding this study were voluntary recruitment, collaborative decision making, and recognising that, although researchers enter the field with an identified problem (older people and delirium), the actual process and direction of the research is dictated by participants.

PAR is a process in which ‘we’, researchers and participants, systematically work together in a group in cycles of ‘looking, thinking and acting’. Delirium and the high percentage of older people who succumb to this condition was the main practice problem requiring a response. ‘Looking’ builds a picture and allows the PAR group to gather information about the condition, both from practice and current literatures. ‘Thinking’ occurs when participants receive feedback provided by researchers from a previous PAR session and they are facilitated by researchers to reflect, interpret and explain. A distinguishing feature of this PAR process is a continuous feedback process, with a feedback document being provided by researchers before the next group session. Although it may resemble the minutes of the previous meeting, it differs in that it also provides preliminary analysis of work in progress and identifies the movement toward group action. At each meeting participants read the feedback and discuss what they think is significant to practice. ‘Acting’ usually means participants decide to ‘act’ on reflections and they suggest ways toward resolving the identified practice problem. The researchers systematically track the actions decided by the PAR group and these actions are described as the outcomes of the study.

Theoretically in the PAR cycle, there is a conceptual difference between the ‘look, think and act’ elements. In practice, however, these conceptual differences begin to dissolve and merge. In this paper we will show the PAR group’s movement toward action. As researchers we have tracked participant decisions and actions. Precisely because clinicians set the agenda for practice improvement, the course a study takes cannot be predicted. Learning to live with uncertainties whilst researching is prerequisite for researchers. Let us explore the PAR process as it evolved during this study.

1.2. Review of the literature and evidence-based guidelines 

It was clear that clinicians wanted to improve their practice in caring for older people with delirium. Whilst we were mindful that the clinicians would drive the agenda to improve delirium care, we knew that time restrictions would not allow the development of critical literature review skills necessary to initiate the ‘looking’ phase of the PAR process. As researchers, we were obliged to review current literature prior to the group meetings so resources would be ready if and when the group made a request.

In preparation for the first of thirteen PAR group meetings, researchers reviewed the literature using CINAHL, Medline and Google Advanced Scholar. Excluded from review was literature addressing delirium related to a surgical event, intensive care, palliation or as a result of medication or alcohol withdrawal. Researchers selected only literature that was pertinent to nursing care for delirium in older people and 118 articles were reviewed. We summarized the work of key authors and, on a global scale, identified the work of three major research teams. In the section to follow we summarize these readings.

Key research had been undertaken by Inouye and her colleagues from Yale School of Medicine (Inouye et al., 1990, Inouye et al., 1999a, Inouye et al., 1999b, Inouye et al., 2000, Inouye et al., 2001, Inouye et al., 2005, Inouye and Charpentier, 1996, Inouye, 2004, Inouye, 2006) particularly in relation to implementing multiple preventive interventions, the detection of delirium and the development and testing of the confusion assessment method (CAM) tool. In Canada, McCuster et al.’s medical research team focused on adverse outcomes of delirium, increasing length of stay of older people in hospital and they have drawn attention the to high morbidity and mortality rates following delirium episodes (McCusker et al., 2001, McCusker et al., 2002, McCusker et al., 2003a, McCusker et al., 2003b, McCusker et al., 2004). Prevention of delirium is best practice and Milisen and his research team in Belgium (Milisen et al., 2001, Milisen et al., 2004a, Milisen et al., 2004b, Milisen et al., 2005) explored the role of nurses in the prevention, detection and management of delirium.

Researchers selected the guideline from the United Kingdom for the prevention, diagnosis and management of delirium in older people (British Geriatric Society and Royal College of Physicians, 2006) as a key evidence-based guideline to guide clinicians in considering practice change. This selection was based on guideline validation in the United Kingdom, preventive emphasis, multidisciplinary audience, applicability to the acute medical care setting, user friendliness and, most importantly, the guidelines were evidenced-based.

2.1. Organisational commitment 

Researchers were invited to conduct this study in an acute medical ward. Delirium had been identified as a clinical practice problem requiring urgent research attention by clinicians and the nursing unit manager. Our experience is that when administrators sanction, support and promote every aspect of the PAR process, the actions decided by clinicians are likely to take hold, be implemented and sustained in practice. Organisational structures, contract agreements and communication strategies were negotiated and agreed upon at the commencement of the study. Negotiated organisational support included allocated rostered time for participant clinicians to attend the weekly 1h PAR groups. A study manager (researcher and first author) was assigned to be the overall coordinator, and her role included negotiating, facilitating the introductory sessions, designing communicating strategies and devising timelines within the ward and the wider hospital.

2.2. Ethical approval 

Facilitation of the PAR group had been approved by the relevant Human Research Ethics Committee but it transpired that the proposal required two variations. First, we recruited a high-speed typist to record proceedings concurrent to actual PAR group meetings and permission was sought to include her to the group. Second, when we realised that a profile of the ward’s patient population was not available, we sought permission for the development and testing of a data collection tool using a retrospective chart audit. Both variations were approved.

2.3. Recruitment of volunteer clinicians to the study 

The study manager gave five introductory sessions to the ward’s multidisciplinary staff across all shifts. Introductory information packs containing a sample delirium guideline, study information letter/consent/return envelope and introductory session content were distributed. A locked yellow staff response box was left on the ward. Potential clinician participants were asked to complete their expression of interest and consent to join the PAR group and insert this information into the yellow box. In this way we recruited eight volunteer clinicians including nurses and one physiotherapist. The research support person (typist) was included in the group, making membership twelve people, a workable sized group (Koch and Kralik, 2006).

2.4. Researcher roles 

The three researchers had separate roles in the PAR group: (1) group facilitator, (2) observer of group dynamics and (3) a time and management coordinator. We were united in our planned response to the group: our task was to listen. One of the most difficult tasks in facilitation is to relax and trust the PAR process, and to let participants decide the direction and the flow of the conversations. We gave ourselves permission to contribute to the group discussion when invited only. The PAR group was to be in charge of its own agenda and participant voices were to be privileged in subsequent feedback cycles. Our role was to be backstage whilst the PAR group set the agenda, looked, thought and decided upon their actions. Making sure that the group ‘moved on’ toward action was the researcher’s second facilitation task.

2.5. Articulating values and interests 

PAR research is openly driven by the values and interest of all those participating including the researchers. Before we commenced the study researchers openly discussed their values and interests.

The research team’s philosophy of care was acknowledged as being driven by primary health care principles: appeal to social justice, equity and person centred care. In terms of person centred care, we believed that inclusion of a person’s unique mental, social and health history was essential for establishing an understanding of how the person had been functioning prior to admission and how this may have changed (Rockwood, 1997). Rockwood’s (1997) provocative remark, that staff may know the current laboratory values but it is rare for them to know whether the same person can brush their teeth now or whether they was able to do so 2 weeks ago, reminds us that we are dealing with a person rather than test values and that changes to everyday function are predictors of adverse hospital outcomes. Engagement with the patient and their families was, in our mind, a key to person centred delirium care.

As researchers we became well informed about current delirium literature and we realised that we held a strong position which we might be able to share if invited by the PAR group. Our position was that actions tied to the prevention of delirium would be the most desirable outcome. We were entranced by evidence in the literature which supported prevention of delirium in older patients, particularly work by Inouye et al. (2000) in the Hospital Elder Life Program. Sensitivity to the influence we might exert as researchers was made apparent. As discussed, we intended to offer our suggestions only when requested to do so. Regular researcher group debriefing and reflection meant we would be positioned to carefully monitor our influence on the research process and group decisions.

On the basis of reviewed literature it was also clear to the researchers that delirium remains underreported and often undetected (Inouye et al., 2001). Even if the PAR group did not select a delirium prevention focus we had identified a potential, powerful and pivotal role nurses could play in early detection and management of delirium. Ideally situated near the patient and providing around the clock observation (Gaudreau et al., 2005), we believed that nurses could assess and observe patients at risk, carefully document their mental and functional state, involve families and/or others who know the patient, but most crucially, they would be on hand to observe delirium’s acute development and fluctuating nature. Early detection of the patient’s abrupt and transient change in mental status could be noted and addressed. Would the group see this potential role as clearly as the researchers?

We agreed that it was important for clinicians to recognise delirium superimposed on dementia (DSD). Research by Fick et al. (2007) revealed that nurses, whilst having a high level of general gerontology and psychiatric nursing knowledge, had difficulty in making the distinction between delirium, dementia and DSD. Would the PAR group self recognise this potential knowledge deficit? And if not, how could we draw attention to this potential deficit without taking the direction and agenda away from the group?

In the reviewed literature only 25 papers focussed on delirium prevention, the majority addressing management. We speculated that delirium is drawn to the attention of clinicians when it has gone awry; when the patient had already developed delirium and ‘the horse has bolted’ (Rigney, 2006). We also speculated that attention is particularly focused when delirium is observed to be hyperactive. This would explain the literature’s preoccupation with management and relative neglect of prevention.

We agreed that Rockwood and Bhat’s (2004) definition of delirium was appealing. Delirium as viewed by Rockwood is a high-order failure of a complex system and the body’s response to noxious insults which the frail older person cannot tolerate, e.g., infection, the environment, temperature change. Inouye (2006) discusses vulnerable or frail older people as being those who have factors at admission which predispose them to a failure in their body systems and which, when they are further exposed to precipitating factors or noxious insults, their system is overwhelmed and delirium develops. Some predisposing factors identified were age over 65 years, dementia, sensory impairment, malnutrition and dehydration; some precipitating factors identified were medications, inter-current illnesses, surgery and prolonged sleep deprivation. With the average age of patients admitted to this ward being 80 years, most patients had the potential to develop delirium based on increased age and we suspected most would have active intercurrent illnesses. We were ready with this understanding, but would we be asked to share?

Researchers considered instruments used to assess confusion and/or cognitive status in patients with delirium. The literature reviewed identified the CAM (Inouye et al., 1990) as the gold standard. With no consistent tool being used on the ward to assess and document confusion it would certainly be an advantage if the PAR group were to agree to use this tool for this study to provide robustness to data collection (Lemiengre et al., 2006). But would the PAR group come to this decision in their own right? It would be important to observe how our values and preferences operated throughout the research process.

Reflection on our values and interests were woven into the data generation strategies. Concurrent recording of the PAR process meant we could reflect on values and interests as they appeared in the verbatim script. In addition to close scrutiny of group data/text, we decided to hold researcher debriefing meetings straight after each weekly PAR group session. These verbal reflection sessions would be recorded, transcribed verbatim and analysed. In this way we would be able to see the ways in which our horizons were operating in the context of the research process. Armed with the literature, guidelines and our values and interests, we were ready to meet with participant clinicians.

2.6. Looking 

In the first group meetings we were engaged in the ‘looking’ phase of the PAR cycle and were concerned with gathering information with clinician participants. We used story telling to initiate information gathering within the PAR group. Each participant was invited to talk about an experience with delirium. As discussed, stories were transcribed verbatim and returned to participants the following week with analysis against the guidelines (British Geriatric Society and Royal College of Physicians, 2006). In the following group meetings, we sat in a circle to read each story, noting the evidence and discussing what could be learned to advance practice. Education was mutual; the research team became familiar with ward practice as described by PAR group members, whilst the participant clinicians reflected on their practice with the help of the stories, evidence from the guidelines and current literature. Participants were looking at aspects of delirium care that were significant to their practice, with describing delirium emerging as one the first items on the PAR group’s agenda.

2.7. Thinking 

Stories told by PAR participants were predominantly about patients who appeared to be already in an advanced stage of delirium, whether they arrived at hospital with delirium or developed delirium during their hospital stay was not clear, but hyperactivity had alerted staff to pay attention to these patients. Everyone could recall a distressing event when a patient had lost ‘the plot’. Dealing with the management of an acute situation, where a patient had become loud, disruptive and aggressive, was the most common story line.

Ten of the generated stories were about an event where a patient was in a highly agitated, frantic mood. Clearly dramatic memories were dominant. It was recognised that fear was experienced by staff, the patient with delirium and their family members. Dramatic events with patients in hyperactive delirium were often cited as learning experiences for participants and we were reminded that health care professionals are required to deal with their own reactions at the same time as meeting patient care requirements and addressing family/visitor concerns.

Whilst most stories told by participants emphasised a dramatic event possibly resembling a hyperactive delirium, it was striking that the actual event was a surprise when it occurred. Agitation and disruptive behaviour appeared ‘out of the blue’. The PAR group talked about patients at high risk, but we still had quite a journey before they would consider potential prevention strategies as part of a patient’s care plan. We believed that this was partly because delirium prevention goes unrecognised by health care professionals in the acute care environment. The research team introduced prevention-based literature including the prevention argument used by Inouye in the Hospital Elder Life Program (Inouye et al., 2000) which discussed the impact of delirium in terms of high mortality, institutionalisation, and longer length of stay, risk/precipitating factors including prevention interventions.

During research team debriefing we speculated that the concepts of prevention belonged to the paradigm of primary health care, and was not a philosophy held by many in acute care. Hence we interpreted clinician reluctance to consider prevention and their preoccupation with diagnosis and management of hyperactive delirium in older people as consistent with critical or acute care philosophies and an organisational drive for best bed utilisation. Moreover, the stories were examples of vulnerable older people who were already experiencing delirium, and participant clinician responses were reactive management strategies such as shifting the patient out of their cubicle into a public space for closer observation and chemical or physical restraint.

Researchers strongly agreed that most of the incidents described could have been prevented in the first place, or else early intervention used to curb the extreme symptoms associated with the delirium. It seemed that delirium was detected only when it had reached crisis stage, and identified only because the behaviour of the patient demanded an acute management response. What about observing those patients with hypoactive delirium?

Upon request, PAR group members were given additional readings as their interests in management of delirium were stimulated. It was fascinating to witness the variety of language used by PAR group members to describe their experience of delirium:

This language suggests that clinicians had acknowledged the importance of being proactive. Eventually this came to mean identifying the cause of delirium from the patient’s history (obtained from the person themselves and/or relatives/carers), managing the presenting behaviours and investigations. What is interesting is that whilst management of acute phases of delirium were discussed in the PAR group, language attributed to the major signs for delirium, for instance those described by the CAM tool, were noticeably absent. When would participants close the knowledge gap between practice and evidence-based practice?

Meanwhile researchers, in their debriefing process, were puzzled about ways to breach the clinicians’ knowledge gap. There was a turning point. As clinicians read around definitions of delirium, selected British evidenced-based guidelines and recent literatures that researchers provided on request, we noted that learning had occurred. Once clinicians recognised that their knowledge had been expanded to include evidence-based ‘new’ knowledge, they were keen to share this information in the ward.

The PAR group noted that dementia is a leading risk factor for delirium; two thirds of delirium cases occur in patients with dementia (Inouye, 2006). Agreeing that routine hospital admission processes may identify details that indicate risk of delirium, the PAR group thought that by keeping this information alive and acted upon, health care professionals have an opportunity to ‘catch them quickly’. Clinicians established that ward staff confused delirium with dementia. It was agreed that making this distinction was important in the delivery of good quality care. Clinicians claimed that whilst they could identify the need for ward staff education programs and instigate local recommendations suggested by the evidence-based literature, they felt powerless to make major changes within the hospital. In fact, these perceived constraints imposed by the hospital curtailed actions clinicians might want to take as a group.

Meanwhile 6 weeks of PAR proceedings had transpired and there was little discussion about prevention of delirium. Although researchers had agreed not to direct the agenda, we were concerned, and we made a decision to voice our concerns to ask the group why management of delirium was foremost on the agenda, when prevention and early detection of delirium seemed a logical first step. The confrontation created tension in the group; participants became defensive about their care and their acute care philosophy. One of the participant clinicians told us that we (as academics) “did not really understand what was going on in practice”, and that we “should place ourselves in the ward and in the ‘real’ world, perhaps even complete a night shift on the ward”. Clinicians argued that the delirium guidelines did not speak to actual nursing practice.

Researchers decided to explore perceived constraints, hoping that their identification might liberate the group to move on. A secondary analysis of the same PAR group data was undertaken to facilitate progress. Here we used a concerns, claims and issues (CCIs) fourth generation evaluation framework from Koch (2000). For the purpose of this paper, an issue is an identified problem that requires resources, services and/or polices outside the immediate ward. Whilst these issues affect practice they were outside the group’s sphere of influence, creating much frustration and inertia.

Feedback to clinicians included a concurrent CCIs analysis. The following issues relating to best practice were identified: delayed transfer of patients from the Emergency Department; routine ward activities were not conducive to provision of rest and sleep; assisting with the patient’s orientation was not possible as relatives were not able to accompany and/or stay with the older patient. Underreporting of delirium and attributing confusion to dementia was viewed as an education deficit across disciplines.

A wide range of assessment skills was identified as prerequisites for working in this acute care ward, with older people and delirium. Clinicians perceived that management driven by length of a patient’s stay, was incongruent with best practice delirium care which required more time for older patients to recover from delirium. Once these constraints to practice had been identified, communicated and most importantly, acknowledged, the PAR group appeared to be ready to explore what could be done at the ward level to improve delirium care.

3.1. Chaos to calm 

When the research team returned to the ward 9 months after the study commenced there was a sense of calm amongst clinicians and in the ward environment. There was evidence that practice had changed. Physical and chemical restraints had not been used in the last 3 months. The nurse manager reported that early detection strategies had prevented episodes of acute hyperactive delirium. Whilst there continued to be older people admitted with a diagnosis of delirium, there were fewer incidences of delirium developing on the ward and there was less disruption to other patients, especially at night.

The strategy of raising the awareness of delirium in older people was successful. Delirium, or the ‘D’ word, is discussed during verbal handover and written communication has changed. The unit manager reported that the workload had decreased—“the ward is not as heavy” and most importantly, in terms of resource allocation, ‘there has been no need to recruit additional staff’.

Ward clinicians had claimed ownership of the DAP. Champions had been self-selected from PAR participants and fellow staff members were educated about the DAP. Staff claimed ownership of the protocol, its attributes were user friendliness, accessible language, clinical relevance and most importantly, no formal documentation had been required as ‘staff hate additional paper work’.

3.2. Innovation and sustainability 

Of significance to the research reported here is that the use of PAR in this study yielded a health service innovation that is congruent with many of the factors outlined in Greenhalgh et al.’s (2005) conceptual model for the determinants of diffusion, dissemination and implementation of innovations in health service delivery. It is already clear that aspects of the pilot study are congruent with success indicators for change shown by Greenhalgh et al. For example, practitioners noted that prevention of delirium is preferable to its management and in this way the innovation has a clear unambiguous advantage in effectiveness, both in reduction of patient suffering and reduction in cost of their care when episodes are less frequent. That the DAP was developed by the clinicians themselves ensured it was compatible with their values, norms and needs. The DAP, developed as a simple memory prompt, is perceived by their peers as simple to use, and was therefore more easily adopted. There were no risks associated with its use. Indeed, a feature of the deliberations during PAR was that the innovation did not increase their work (see page 22). The benefits of the DAP were visible as the roster showed fewer staff members were required because the incidents of delirium had decreased in the ward (Greenhalgh et al., 2004, p. 596–597). The introduction of the DAP improved performance; practitioners were confident and more satisfied with their role in prevention of delirium. In addition, the knowledge required to use the DAP is explicit in its simplicity which also means it can be transferred from one context to another. The actual PAR process and the introduction of the DAP can be augmented as participants are also champions, they are excited and enthusiastic about their new development and use the train the trainer strategy to infiltrate the ideas throughout the organisation. This also promotes the diffusion (Rogers, 1995) or passive spread of the innovative ideas and practices generated are already reverberating within the organisation. The PAR process subscribes to a partnership and community development model as advocated by Potvin et al. (2001). Partnership models encourage a shared and organic understanding of the meaning and value of the innovation in use and should work toward a shared language for describing the innovation and its impact.

In conclusion, this paper has described a PAR practice redesign study. Based on the findings, there is incentive to trial this participative process in other acute care settings. If so much is achieved in just 6 months, we are encouraged about what can be achieved in 3 years. Of course, the research team cannot predict the course of practice redesign, or what agenda and action the PAR groups will take but we are confident that working collaboratively with practitioners is the way to bring evidence to practice in delirium care for older people in acute care settings. An evaluation of the uptake and utility of the delirium alert protocol is currently underway (involving the authors Day, Higgins in conjunction with clinicians Li and Dumont) at the time of writing this paper. Whilst the findings of this study will be published elsewhere, early findings suggest that awareness of prevention of delirium has improved and that there have been better outcomes for older patients.