Types of knowledge

Senior ED doctors in ‘virgin’ sites suggested that they had heightened knowledge of delirium and of the consequences of undetected delirium, meaning that they were more attuned to the importance of identifying it. Furthermore, understanding of the delirium trajectory, including its deleterious consequences for patients in both the immediate and the longer term, increased the salience of work to detect it. The combination of clinical knowledge and experience meant that senior staff were able to quickly recognise a particular constellation of factors in the patient that made them think of delirium. This tacit knowledge of delirium, aligned with the clinical presentation, sensitised these professionals to noting and interpreting observational cues to construct a clinical picture:

There comes a point in a senior doctor’s life where the database of thousands and thousands of patients that I have seen is sufficient for me to be able to recognise certain clinical patterns both from history and examination and that’s often brief . . . but it’s quite common to get the key points of the history that will point towards the diagnosis . . . so recommending a comprehensive history and a full physical examination . . . in the first place, and then targeting investigations to those most relevant . . . there’s a number of different things that can cause it.

Consultant, ED, Avonfield 04

Importantly, these different types of knowledge were not shared by all staff in the ED. A common view expressed by senior doctors was that although staff generally were ‘more willing to consider delirium as part of the clinical picture’ they lacked knowledge of the profound negative effects of undetected delirium:

Do they understand it? No, I don’t think they appreciate it. I think that people are getting better at picking it up but do I think that they recognise the significant mortality rate . . . that’s associated with it? No I don’t think so. I think they might vaguely be aware but . . . I think the numbers would startle them.

Consultant, ED, Avonfield 01

Consequently, it was suggested that these staff might be less focused on delirium detection, according it lesser significance in the overall assessment process and in the construction of the differential diagnosis. Lack of clinical experience with patients with delirium was also seen as a barrier to the development of tacit knowledge.

The enhanced understanding of delirium that accompanied knowledge and experience meant that senior staff were often able to quickly recognise a particular constellation of factors in the patient that made them think of delirium. By incorporating knowledge of delirium with the clinical presentation, these professionals noted and interpreted observational cues to construct a clinical picture; furthermore, understanding of the delirium trajectory increased the salience of detecting it.

Less experienced staff, or those with different professional roles, might have only some elements of this delirium knowledge, leaving them less able to recognise its presentation and its meaning for the patient. Introduction of a screening tool might, some suggested, improve recognition of delirium, but only if the pattern of signs presented by the patient offered a signal or trigger to the practitioner to consider delirium as a possibility. This in turn was seen to require at least some knowledge of the circumstances in which delirium typically developed. The corollary was that senior staff might perceive use of a screening tool as less relevant, focusing instead on assessment using a range of clinical tools among those patients who exhibited signs of probable delirium.

Inconsistency of knowledge

Among junior staff, the pattern of awareness and knowledge of delirium was inconsistent, shaped partly by training and partly by previous experience in a setting in which delirium was prevalent. One interviewee described how knowledge of delirium gained through geriatric experience in a care of older people ward, in addition to taught sessions, had helped him to understand delirium better, and to develop a proactive approach towards cognitive screening which carried over into his practice in potentially less delirium-prevalent settings:

I imagine that if I’d not come from a ward, a base ward, where there was very much a focus on cognitive screening, I think I would have had much more of a hit and miss approach to the deployment of cognitive screening and I wouldn’t have easily recognised its importance and prioritised doing it above other things for the patient.

Junior doctor, ED, Avonfield 06

Several others indicated that their knowledge of delirium developed into deeper understanding only when they connected their theoretical learning with practical experience of patients with delirium and its effects. Thus:

At medical school you are taught that delirium is a medical emergency, that it’s an acute medical problem that needs treatment straightaway. But as much as I understand the seriousness of it, I generally appreciate finding out the source of it . . . It’s more that this patient has delirium because they’ve got sepsis. I need to treat the sepsis, so that’s what worries me more to be honest . . . So if I’m worried that . . . the patient is kind of disoriented to time and place, confused, then I’ll go down that avenue in terms of what I should be looking for in the collateral history.

Junior doctor, ED, Avonfield 11

Awareness of the significance of delirium therefore enhanced sensitivity to observational cues that might indicate its presence. Furthermore, it translated into a line of exploration in history-taking and information-gathering as part of developing a differential diagnosis.

The following account further shows how important it is to understand the significance of delirium so that it can be prioritised when making a differential diagnosis. Here, a junior doctor with interests in psychiatry and emergency medicine and a recent rotation in psychiatry described how he would elicit information from an older patient, among those ‘presenting an extra challenge’:

I tend to ask open questions to start off with because you can just gauge whether the patient is completely oriented and knows what’s going on and if I’m worried that I don’t think that patient is kind of oriented to time and places, confused, then I’ll go down a completely different avenue in terms of what I’ll be looking for in the collateral history; if anybody’s with them at the time, and chatting to them about what’s happened. And I also tend to go through the AMT [cognitive assessment tool] with patients like that as well . . . it’s very difficult to know what’s new for them because sometimes you’ll turn around to the relatives and you’ll say, you know, ‘Is this a new amount of confusion do you think?’ and they’ll be like, ‘I’m not really sure to be honest,’ and . . . you have to treat it as it’s new . . . but trying to figure out a source of that confusion as well . . . And then depending on how well that patient is and whether there’s an obvious source and what their home circumstances are like, you know, sometimes you can get them back home with a course of antibiotics, other times you need to admit them because it’s not safe to send them home, given their kind of social support, and the fact that they’re just very confused . . . More often than not I’ve ended up admitting people that have come in acutely confused, because they’ve just not been well enough to go home . . . and there’s been other issues like they’ve been tachycardic or hypotensive, and so they’ve needed to stay in clinically.

Junior doctor, ED, Avonfield 07

It is notable here that the use of the terms ‘confusion’ or ‘acute confusion’ did not necessarily denote a lack of knowledge of delirium. The terms in this context were employed as descriptive labels drawn from the patient’s history and observational cues that opened up lines of investigation to probe the source, and that might conclude with a diagnosis of ‘probable delirium’. When such a cause was not evident, this was generally viewed as a reason for admission in both the 4AT ‘virgin sites’:

I think anyone who’s acutely confused and that you haven’t . . . got an easily reversible cause . . . I think they would need admission really . . .

Junior doctor, ED, Cranford 01

And:

I would imagine that a lot of patients with confusion get admitted and it’s left at a later point for people to work out if their confusion is new, the extent to which it’s new . . . I would think it’s difficult to arrange the speedy discharge of an A&E [accident and emergency] attender when you can’t clarify the history of confusion.

Junior doctor, ED, Avonfield 06

A key issue emerging from the findings from these 4AT ‘virgin’ sites (i.e. where there was no systematic, routine delirium screening tool in use, including the 4AT) is that assigning value to delirium detection is at least partly contingent on experiential knowledge of delirium and its impact on patients, as well as learned knowledge. We have characterised professionals with such multilevel knowledge of delirium as ‘delirium knowledgeable’.

The routine use of a delirium screening tool might be expected to enhance delirium awareness, if not knowledge of delirium, which in turn would result in its more consistent detection. The relationship appeared to be less straightforward, as suggested by interviewees from an ED within a 4AT ‘experienced’ site. Here, although the 4AT had been introduced into the ED more than a year previously, it was not currently in routine use. Delirium knowledge was described as patchy:

So I think it’s still under-recognised . . . and you’ll maybe get a handover from triage that the patient’s agitated, or . . . they could be septic, that’s another term. So they might have a high temperature, and it’s only really picked up that they’re delirious I think if someone is familiar with the concept . . . and that’s usually after they’ve seen a doctor or an experienced nurse who’s maybe done some geriatric training. But even then it can slip through the net.

Consultant, ED, Denbury 17

‘Delirium knowledgeable’ staff, however, were likely to ‘see’ behaviours as possible indicators of delirium, requiring further investigation. Similar to ‘delirium knowledgeable’ participants in the 4AT ‘virgin’ sites, it appeared that knowledge of delirium among some ED staff made them more sensitive to indicators of probable delirium when taking a patient’s history. From the account below, it is evident that eliciting understanding involved a combination of strategic questioning, interpretation based on tacit knowledge of risk factors for delirium, and experience of and sensitivity to the meaning of observational cues:

Within a few minutes of speaking to them, I’ll appreciate whether they’ve got an understanding as to where they are, why they’re here . . . their presenting complaint would be one for me . . . the story of their presentation would make me think. Before I go in to see somebody I often have a very quick look in their medical history. So anything there . . . that they’ve been in before with confusional states, they’ve had UTIs [urinary tract infections], they may already have a diagnosis of cognitive impairment. So I may have that information before even going to the room. And also just how I see them. What they look like as I’m approaching them . . . all of those little things would trigger in my head . . . what’s going on here . . . And I would note on the history ‘have you considered . . . [Delirium]; ‘does the patient need this’. But it would be the area the patient goes to for their management that should be the area that does the 4AT.

Senior nurse, ED, Denbury 16

Among ‘delirium knowledgeable’ staff, although the type of work involved in eliciting and noting the presence of delirium in the ED is described in this account, systematic screening of delirium was viewed as problematic in this setting. To understand why, we next consider the specific ED context and purpose.

Pace of work in the emergency department

Across all sites, the pace of work was uniformly perceived as a barrier to delirium screening:

The problem is that it’s such a rapid turnover there’s probably a lot of concern from staff that they don’t want to fill in another score sheet or another tool . . . I’ve had similar problems trying to institute scores for frailty . . . no one really wants to have more work to take on . . . it’s probably quite a short-sighted view . . . the scores we have on the triage sheet at the moment are vastly undercompleted even for such questions as ‘what’s the patient’s early warning score’ and ‘do you think the patient’s septic?’

Senior doctor, ED, Cranford 02

Two specific features were raised that highlighted how critical time was in the ED: (1) the short time available for working with each patient to arrive at a decision about the immediate action to be taken; and (2) the access to information relevant to addressing the key features of probable delirium, namely whether or not the ‘confusion’ was of rapid onset and was fluctuating.

Time in the emergency department to achieve purpose

Decisions about the immediate action to be taken for the patient were made quickly, usually within 4 hours, to meet the performance target and manage demand. Even a screening tool described as brief and concise, such as the 4AT, was considered impractical, or at least it was not seen as a priority in this context, as a nurse in the 4AT ‘experienced’ site indicated:

It’s not consistently completed particularly on days like today, when you are extremely busy. It’s probably not one of the first things that you would prioritise . . . because of staffing more than anything . . . it’s not something that people automatically think about.

Staff nurse, ED, Denbury 18

This pattern was corroborated by interviewees working on the MAU and acute wards in this site, who reported that the 4AT was rarely completed for patients who came to them. Typically, documentation from the ED comprised a patient history, and information on acute presentation and differential diagnosis. Even so, staff acknowledged that it was difficult to accurately assess cognitive impairment in the ED environment because of the impact of the admission and the event that gave rise to it (see Impact of the admission process on patients).

A key feature of delirium is its fluctuating course and recent onset. In the ED, time and context were perceived as barriers to accessing relevant information for addressing what was an apparently simple question. Thus:

The acute . . . and fluctuating course . . . that’s so difficult in ED . . . when this is done, within 4 hours of them coming in. Unless we’ve got the family there or we know them well, I think that it’d be really difficult to get a judgement on that . . . unless you’ve got a really good history from someone else I don’t think you’d be able to . . . score that very accurately.

Junior doctor, Care of older people ward, Denbury 13

Impact of the admission process on patients

Features of delirium, such as ‘alertness’, were viewed as affected by time and context, with the busy, dynamic environment making it difficult for patients to concentrate. It was also argued that the trauma of the event giving rise to admission meant that patients required reassurance and support, and ‘tests’ that were aimed at identifying deficits might add to a patient’s anxiety. For ‘delirium knowledgeable’ practitioners, having reassuring conversations with patients, and employing ‘comforting’ strategies to alleviate distress, could enable understanding of cognitive impairment as part of a whole picture:

I think when people come in and they’re either in pain or distressed or, you know, that they’ve had a journey by ambulance, they’re excited, they’re you know, upset by that . . . It’s a busy noisy department and so concentration possibly is an issue for patients. It’s probably not the best time . . . to be asking them to count backwards, or you know ‘tell me the numbers from’ . . . We would all struggle with no cognitive impairment under those circumstances . . . So I do get that . . . just informally, chatting to them can give you quite a lot of information without putting them through the added sort of almost ‘examination’ stress, of doing a formalised test.

Senior nurse, ED, Denbury 16

The approach conveyed here offers a partial explanation of how ‘knowledgeable delirium practitioners’ are sensitive to, and take account of, delirium risk as part of their overall assessment, yet eschew the use of a screening tool.

Purpose of emergency department work

Several participants described the ED as a ‘blank canvas’ insofar as the work was unpredictable. The primary purpose of the work was viewed as obtaining a differential diagnosis to inform decision-making about an immediate course of action within a designated time. The process was conceived of as ‘layers of decision-making’: first, deciding if the patient was seriously ill; second, deciding whether or not to admit the patient, cognoscente of the gatekeeping function of the ED; and third, deciding whether or not a probable diagnosis fitted the overall patient picture. Prioritisation of delirium screening in this setting was influenced by its perceived value in contributing to such decision-making. Even when delirium detection in itself was valued, unless it influenced decision-making it might not be prioritised:

The icing on the cake almost is where a diagnosis fits into that because you may not know precisely what the diagnosis is . . . you might have a list of differential diagnoses . . . so that for the person in front of you, you don’t actually know what the cause of their problem is but you know that they can’t go home . . .

Consultant, ED, Avonfield 01

Across all sites, it was suggested that a screening tool (whether the 4AT or another tool) would be used more consistently if staff understood why it was important to use it, if they could see tangible benefits for patients and if screening was seen to make a difference to patient outcomes in this particular setting:

It’s a good concept. The problem is that trying to change people’s behaviour is very difficult if they don’t understand why they are doing something . . . What we kind of want to know, in the emergency department, is . . . if by doing this score are we making a difference? And, you know, how many, what effect is this having on patients dying or on saving patients from getting unnecessary treatment.

Consultant, ED, Denbury 17

Even so, an assessment of benefit would probably involve consideration of the nature of the ED environment and the challenges of context and time in that setting. Additionally, it was noted that experienced and ‘delirium aware’ ED staff considered delirium risk as a factor in their overall assessment, although this drew on pattern recognition through multiple sources of information, including interpretation of observational cues. Non-‘delirium aware’ staff, by contrast, were unlikely to ‘see’ and interpret this information. Paradoxically, all other things being equal, knowledge of delirium in the ED context might obviate the value attached to the use of a screening tool, although, similarly, a lack of delirium knowledge contributed to non-use of such a tool.

Medical assessment unit

Like their ED colleagues, members of the multidisciplinary team (MDT) in assessment units were required to draw on their knowledge of a wide range of conditions to pursue an investigatory strategy to develop and refine their diagnosis and to implement a management plan for the patient. Importantly, and in contrast to the ED, the lengthier assessment period meant that ascertaining relevant knowledge of the patient’s biography and medical history was more feasible, despite the fact that time was also regarded as a scarce resource.

Similar to in the ED, some senior doctors in this setting conveyed a breadth of clinical experience through which they had acquired knowledge and understanding of delirium: they were ‘delirium knowledgeable’. Additionally, specific local initiatives directed at enhancing awareness of delirium among staff, as in Cranford, had, it was suggested, improved delirium knowledge among MAU professionals:

Nurses here are fairly switched on as far as delirium goes. We don’t have so many patients here as we do on elderly care who have fluctuating confusion, but we do get it here and nurses deal with it. I trust them.

Senior doctor, MAU, Cranford 02

In Denbury, implementation of the 4AT in the MAU was regarded as having enhanced awareness of delirium. This was symbolised in the language employed, with the term ‘delirium’ explicitly used instead of ‘cognitively impaired’ or ‘not quite right’. The 4AT’s introduction had been supported by a short information session on its use as well as on the significance of delirium. Nursing staff were designated as responsible for screening for delirium and completing the tool, with the 4AT being incorporated into the nursing assessment. Even so, interviews showed that multilevel delirium knowledge among MAU staff in this site was variable.

‘Delirium knowledgeable’ interviewees in Denbury MAU described the process of identifying patients with probable delirium in a similar way to ‘delirium knowledgeable’ participants in 4AT ‘virgin’ sites, namely that it was seen as involving more than completing a score on the tool. It involved ‘empathic connection’ with a patient’s distress and anxiety as a result of the event giving rise to the admission, history-taking in a ‘comforting’ conversational style to help the patient relax, and eliciting information ‘almost like a chat’ to ascertain whether or not the patient might have cognitive impairment. Discrepancy in the accounts from the patients were described as ‘red flags’, triggering questions such as ‘What is going on here?’ and ‘Is this normal?’. These also prompted further information-gathering. Completion of the tool by these nurses in this context formalised what might in the past have been a comment to the doctor that the patient was ‘really confused but it isn’t normal for her’. They regarded screening for delirium as a necessary and legitimate part of the habitual assessment work in the MAU. A positive screen also directed and prescribed action:

I would check what bloods have been done in accident and emergency and probably add a few more tests because we’re looking at a confusion screen . . . widening the screen . . . and should help get a diagnosis quicker [of possible infection] and work out the appropriate antibiotic.

Staff nurse, MAU, Denbury 04

Whereas delirium knowledge and its management enabled nursing staff to take the initiative and act on the results of screening, for others the perceived value of screening for patient care was not self-evident. This had a knowledge dimension: if staff did not understand the significance of delirium, they were less motivated to use the tool. This also had an action dimension: if the pathway from detection to action was unclear, sustaining engagement with screening could be compromised.

I think there’s an assumption that we’ll just carry on with this person who has got known cognitive impairment and we’ll carry on observing it, which I suppose is what the 4AT is for . . . But what are we doing with it? . . . Are we doing anything about a known cognitive impairment when they’re scoring one to three? I don’t know.

Staff nurse, MAU, Denbury 08

Several nurses indicated that, when under time constraints, they would prioritise specific assessments over the 4AT; skin/pressure assessments were specifically mentioned as more visible indicators of performance and care quality. Although noting that they had used the delirium screening tool on only a few occasions, they also acknowledged that ‘others do it more often’. Yet others prefaced their accounts of use by referring to what ‘should be done’: it was what was expected of them.

Variable knowledge of delirium and lack of clarity on the pathway from detection could affect consistency in use of the screening tool.

Surgical assessment unit

We conducted the study in only one SAU; therefore, the picture that emerged from comments of interviewees, of a low level of understanding and awareness of delirium, may not be typical of these units. Even so, this echoes the accounts of the care of older people liaison team on orthopaedic wards in Denbury.

In Cranford’s SAU, senior nursing staff acknowledged the delirium knowledge deficit, and viewed this as problematic in the light of the high level of risk among one group of their patients, namely older people with hip fracture. Even though information-gathering did not include use of a screening tool for delirium, the nursing assessment pro forma contained four questions relating to risk factors for delirium, based on NICE guidelines (having current hip fracture, being aged ≥ 65 years, having cognitive impairment and having severe illness). For those at risk of delirium, a pro forma delirium prevention care plan was to be completed (e.g. action on hydration, nutrition, pain, medication review). Both were introduced on wards as part of a drive to increase awareness of delirium and delirium prevention (albeit without explanation, training or support in use). Interviewees here were unclear about how these risk factors were established and about the content and purpose of the delirium prevention plan. Unsurprisingly, they reported that the pro forma was not completed on the unit. A senior nurse, who was supportive of the idea of delirium knowledge as enabling staff to ‘see’ the value of identifying delirium, welcomed the introduction of a short, easy-to-use tool such as the 4AT, but went on to suggest that, to trigger use in this setting, staff needed to understand why they should invest in it and what action would be taken as a consequence:

It looks pretty user-friendly and that it won’t take too long to do which is the key, you know, ‘cos otherwise people just won’t do it. But then obviously we need to know why we’re doing it and what to do once they trigger a certain score as well.

Senior nurse practitioner, SAU, Cranford 06

Elderly assessment unit/frailty unit

In EAUs/frailty units, staff presented as experienced in and knowledgeable about the multiple needs of older people, including those arising from a cognitive impairment. Among their patients, delirium was common. Identifying, assessing and developing a management plan for those with delirium was seen as a routine part of their work, although how this was accomplished differed between the two 4AT ‘virgin’ sites.

In Cranford EAU, although the patient assessment form did not include a delirium screening tool, it was commented that experience in caring for older people meant that ‘acute confusion’ was at the forefront of thinking and that any indication of inattention would sensitise staff to look for ‘reversible causes’. In this site, the staff’s perceived skill in noting and making sense of changes in the patient contributed to the low value attached to screening tools; confidence in observing and reporting changes meant that medical staff could proceed to assessing and investigating factors contributing to delirium. Moreover, as the purpose of the setting was to develop a management plan, tools were indicative of a problem that would prompt formation of an action plan, but the view was that they did not in themselves inform such a plan:

As a geriatrician what you are doing is looking at the whole picture to inform the most appropriate management plan . . . you are using your observational skills as a doctor; you are also drawing on cues from staff from their knowledge of the patient. Nursing staff will also know – but they may not use the word ‘delirium’. They’ve observed that the patient has become more muddled, more confused as they’ve seen them on the ward; or the relative has said they are more confused now than they were at home.

Consultant, EAU, Cranford 05

In Avonfield, the AMT10 was conducted with all patients as a case-finding tool for dementia, with a flow chart to delineate resulting action. Aligned with the collateral history, this was also used to identify new, acute confusion. The rationale for this approach was the co-occurrence of delirium and dementia and to engage staff in thinking about the link between them.

The pattern observed previously in ED – that understanding and knowledge of delirium meant that staff were more sensitive to picking up on observational cues that might, for example, indicate change in alertness and perceptual disturbances – was evident in both units. An understanding of the risk factors and a direct, routine involvement with patients meant that nursing and care staff were seen to play a crucial role in ascertaining if a patient did not seem to be ‘quite right’. Both doctors and nurses interviewed emphasised their joint and collaborative endeavour in identifying delirium: ‘they appreciate that it means something’s going on that needs investigating’. Medical staff were seen to note and act on the information communicated by nursing and care staff; the latter in turn felt valued and confident about communicating their observations in MDT meetings, in ward rounds and informally with medical colleagues.

Pace of work

A feature of the MAU/EAU was that, in comparison with the ED, staff had a longer and more sustained period in which to observe, collect collateral information and review a patients’ progress. As patients were admitted at all times of the day and night, the constant busyness and noise of the unit disturbed those trying to rest or sleep. Among some staff, the continuous hum of activity and the merging of day and night meant that it was a ‘bad environment’ in which to screen for delirium. Furthermore, it was considered that patients were likely to be disoriented and distressed, particularly those admitted during the night, on account of both the medical event that precipitated hospital admission and the admission itself. Generally, although the nature of the care environment in the MAU might make it difficult to obtain an accurate baseline picture from which to determine delirium, the consensus was that it did not create an insurmountable barrier.

Systems for integrating delirium detection into routine practice

In the 4AT ‘experienced’ MAU setting, participants described the work undertaken to implement the 4AT into ward routines. The 4AT was incorporated into the unitary patient record (UPR), which comprised multiple screening and assessment tools and was intended to be completed by nursing staff for all new admissions along with documentation of ‘care rounding’. The documentation involved daily interval review and documentation by nurses of skin/pressure areas, mobilisation, toileting, continence and catheter care hygiene, analgesia, pain, fluid balance and care plans. Direct, ongoing engagement of nursing (and care) staff with patients meant that they were seen as well placed to observe change and fluctuations. A division of labour was in place such that nurses were responsible for completing the delirium screening tool and junior doctors conducted the cognitive assessments (AMT and MMSE) as part of the clerking process.

Some aspects of the system to implement delirium screening in Denbury were not universally regarded as meaningful. A senior nurse questioned the value of screening everyone who met the specific age criteria; she felt that screening should be more flexible and draw on tacit knowledge about and expertise in delirium risk: ‘use your head’. Thus, it was argued that observation of and conversations with patients should inform which patients were ‘at risk’ of delirium and, therefore, should be screened. Such an approach was, however, predicated on having an in-depth knowledge of delirium and delirium risk:

In our paperwork it says over 65 years . . . we’re trying to make it ‘use your head’, basically. If they’re 55 but they’re ill enough that you’re concerned that they’ve got some sort of delirium/sepsis, use the tool . . . But the lady in bed [bed number], who is over 65, so technically we should be doing the 4AT. She’s all with it. She’s in for cardiac reasons. She’s no side-symptom of infection. And I just explained it to her by ‘it’s a screening tool we have for anyone over 65, do you mind?’, ‘no’, but she laughed all through it . . . I was a bit embarrassed about her but . . . that’s the first time I kind of thought ‘why am I doing this?’

Senior nurse, MAU, Denbury 04

In the context of a MAU, responding to the heterogeneity of the needs of patients across age groups and with multiple and varied presentations posed particular issues for some staff in using the 4AT delirium screening tool. Highlighted was the way in which staff with knowledge of particular conditions employed that specific knowledge and skill as a lens to make sense of, or contextualise, observational cues and behaviours. The extended narrative account below about toxicology illustrates how the overall pattern of presentation confers meaning. It also suggests that practitioners may judge that routinely screening patients is unnecessary without considering the delirium risk, and that they may have an alternative explanation for symptoms because of their knowledge and skills in a different field of medicine:

For the younger age group you . . . look for the obvious first. If they’ve just drunk a bottle of vodka you’re looking at intoxication. What’s the chance of this person . . . having a delirious state before coming in? If there’s no history of it look at the obvious and deal with it from there . . . Or . . . if there’s a history of an ongoing affective disorder or something else going on from a psychiatry point of view I would think psychiatry rather than a delirium. And if you get the history then it’s going to show up . . . So the history taking is going to give you the indication. The 4AT is a bit of a snapshot when they arrive; you need a bit more of a history. So regarding the criterion: ‘has there been a kind of sudden onset?’ Well your history taking is going to show you that. So it’s part of it from there. But if you . . . get a 3 out of 4 on 4AT . . . is it going to change? And it’s just another number where you think ‘fine’, but that could be explained by the fact that they’ve just drank a huge amount of alcohol 4 hours ago. You’re giving yourself a reason why there’s a 4. You don’t necessarily need to be think that . . . we’re looking at delirium; no, we’re looking at an intoxication.

Senior nurse, MAU, Denbury 08

Legitimising the contribution of different staff groups

Among some nurses, the value of the standardised tool was twofold. It gave credence to nursing observations and enabled nurses to exercise professional autonomy to pursue investigatory action, without having to defer to medical colleagues to endorse it.

The understanding of delirium in the context of a perceived valued professional role in detection and action contributed to ownership of the process and sustained ongoing use. From this perspective, screening was not simply a ‘tick-box’ exercise. It involved knowledge-informed interpretation and decision-making on the part of nursing staff.

Others questioned whether or not medical colleagues accorded value to the nurses’ work in detecting delirium. If their work within the overall process of detection and action on delirium was neither understood nor accorded value, this could also inhibit sustained participation:

Doctors aren’t coming to say ‘4AT, we’re acting on it, we’ve done this . . .’ And there’s not enough of: ‘what we are doing next, are we responding to a sepsis or are we responding to . . .’ It’s a multidisciplinary team thing at the end of the day . . .

Staff nurse, MAU, Denbury 07

Overall, delirium screening and assessment was viewed as consistent with the purpose of assessment units, namely to develop a patient management plan within a more extended timescale than was possible in the ED. Nevertheless, as evident in all sites, knowledge of delirium affected the significance attached to detecting and assessing it when formulating a plan of care.

From the accounts provided by the ward staff interviewed, the MAU had relatively high 4AT completion rates (reported as 60–70%). Factors contributing to variability included variable knowledge of delirium, which negatively affected the value attached to identifying it; and uncertainty about whether or not screening was followed by action and, therefore, accorded value by members of the MDT. Moreover, although the delirium screening tool appeared a simple and ‘objective’ indicator of probable delirium, the interpretation of some items, such as ‘recent’ and ‘fluctuating’, were not regarded as straightforward; rather, they made sense as part of an overall pattern in context. Similarly, presenting features that might indicate probable delirium could, depending on the context, be interpreted as meaning something else.

Care of older people wards

As noted earlier, delirium and dementia often co-occur in older patients and the prevalence of delirium on care of older people wards is high. In the study, most of the wards were care of older people wards. Although medical and nursing staff on these wards had experience of ‘seeing’ and responding to patients with both delirium and delirium on dementia, knowledge of delirium embraced the spectrum from ‘limited’ or ‘general awareness’ to ‘delirium knowledgeable’ as characterised above, namely in-depth understanding of its heterogeneous presentation, trajectory, distress caused to patients and anxiety of relatives.

The 4AT ‘virgin’ care of older people wards in this study was, in some respects, atypical of care of older people wards generally in their knowledge of delirium and in the value attached to identifying and managing it. The Avonfield ward was designated for patients with dementia and delirium; and the Cranford ward had been involved in research on delirium prevention.

On both wards, having knowledge of delirium and the significance attached to it was viewed as sensitising staff to observable signs of change. Thus:

The wards I work on staff understand the value of identifying delirium . . . they appreciate that it means something’s going on that needs assessing . . . and even hypoactive delirium, you know, they notice if someone’s not waking up for them and taking their medication . . . they do pick it up and I feel they do know to act on it . . .

Consultant, frailty assessment unit and care of older people ward, Avonfield 02

The close working environment and frequent interaction between staff, patients and caregivers were seen as enabling staff to develop a shared understanding of what was ‘normal’ for the patient and to work with the patient as a person:

I think perhaps in terms of a general experience point of view I suppose it is good to have because I don’t know where else you’d get this experience because [delirium] is a common problem, especially in older people . . . you are treating the patient as a whole rather than just the illness that they’ve come in with, which is good.

Junior doctor, Care of older people ward, Avonfield 05

Health-care assistants, nurses, doctors and therapists conveyed a deep understanding of people with a cognitive impairment. There was an expectation on nursing and care staff to spend time with patients and acquire personally meaningful knowledge of those patients as individuals:

You get to know your patients when they’re not right: ‘oh Elsie’s not right today . . . she’s sleeping more than usual . . . not communicating as she would normally . . .’ we spend a lot of time with patients so we notice changes when you’re washing or bed-bathing the person . . . and because everybody’s different . . . it’s really important that you know the person as an individual and what they’re like on a daily basis . . . we talk about those changes at handover.

Health support worker, Care of older people ward, Avonfield 09

Personally meaningful knowledge acquired by health-care assistants through their direct contact with patients was valued, communicated, listened to and shared in multidisciplinary fora.

In care of older people 4AT ‘experienced’ wards, nursing staff reported that the introduction of local guidelines for delirium, aligned with organisational impetus on screening with 4AT implementation, had brought delirium further to the front of awareness; and that this was in the context of a pre-existing knowledge gap. Thus:

Interviewer:

. . . what sort of training have you had in the past on delirium?

Interviewee:

[shakes head] . . . None [then goes on] . . . I have looked at information websites for myself . . . and there has been a big drive on delirium just now in the hospital as well as a sepsis drive, to increase detection of delirium and get us to distinguish between delirium and dementia . . .

Staff nurse, Care of older people ward, Denbury 02

Knowledge of the significance of delirium conferred priority on detection and action:

. . . the recognition of delirium is crucial ‘cos it’s like any acute medical condition it needs to be addressed as importantly as you would do the rest, the heart attack or anything else . . . we’re all getting better at identifying it earlier . . . and there’s quite a focus on it in care of older people . . . now it should impact their medical care, their nursing care and whether they will have the ability to engage with physio[therapist] or OT [occupational therapist] so it has to be part of reviewing their progress or their medical condition . . .

Clinical fellow, Care of older people ward, Denbury 09

Awareness of delirium, it was suggested, meant that ‘problem’ behaviour of, for example, patients ‘kicking, spitting and aggressive’ could be understood in context of their condition, facilitating empathic connection with the patient:

It makes it, in your head, more acceptable, because you know that [normally] they would never act that way, they wouldn’t . . .

Staff nurse, Care of older people ward, Denbury 21

The level of awareness and attention about delirium on care of older people wards was in marked contrast to that reported among staff in other medical and surgical wards. Junior doctors in particular, from their experience of multiple rotations in various surgical and medical specialties, offered a wide comparative perspective. The prevalence of delirium (or cognitive impairment) on specific surgical wards, such as cardiothoracic, was viewed as very low in comparison with that on care of older people wards. The profile of patients (younger) and the clinical purpose (mainly elective procedures) suggested low risk, which was seen to explain why neither cognitive nor delirium screening appeared on clerk-in sheets, also justified on the basis of very low prevalence. On orthopaedic wards, the high prevalence of delirium among subgroups of patients, particularly older people with a hip fracture, had been addressed by having a liaison care team for older people to provide support for these patients. Although working closely with nurses and junior doctors on these wards, a division of labour existed such that senior surgical staff were focused on surgical aspects of care and had little involvement with nurses and junior doctors in relation to other aspects of the work.

Therefore, from a comparative perspective, staff on care of older people wards had a heightened awareness of delirium, given their exposure to patients at risk of, and experiencing, delirium. Moreover, nursing and care staff were in a powerful position to note changes in patients because they had more direct and prolonged contact with them. However, the descriptions of practice revealed that some staff were uncertain about what to do if they identified delirium in this setting.

Systems for integrating delirium identification into care management in routine practice

Our study was focused on delirium detection. What emerged from interviews with staff on care of older people acute wards was that there was an intimate relationship between knowledge of delirium, the value attached to detection and the processes in place for the subsequent management of patients with delirium.

In Denbury, study participants were from three different care of older people wards. Staff interviews in each ward showed some ambiguity about which professionals were responsible for detecting new cases of delirium and the use of the 4AT in that process; the impact that detection had on the management of patients identified with delirium, including those for whom a positive screen had been completed in the MAU; and what systems were in place to review recovery from delirium. It is possible that the disparate findings reflect the fact that interviews took place while considerable movement of staff was happening: junior medical staff had just begun their new rotation, which also coincided with the recruitment and induction of nurses into the department. Thus, the findings may pertain to local, temporal features of care delivery. Even so, they pose organisational issues relevant to embedding new practices on delirium into work routines.

Delirium detection: division of labour or disconnection?

Among the junior doctors interviewed from Denbury care of older people wards, there was consensus that, as part of the clerking-in process, it was their responsibility to ensure that a cognitive assessment was conducted, typically using the AMT10. The 4AT, on the other hand, was for nurses to complete, consistent with the system operating on the MAU.

From the doctors’ perspective, their cognitive assessments informed a plan of treatment that was then reviewed by a senior doctor in which ‘it’s the AMT we refer to’. The 4AT, on the other hand, was in the nursing profile (on their assumption that it was completed by nursing staff) and not something to which they would refer or look at:

Whereas the AMT because it’s in our . . . booklet . . . that is what we look at . . . the medical team don’t use the 4AT very much . . . I don’t think many people would go into the nursing notes to find the 4AT score . . . I’ve never had one of the nurses flag the 4AT score to me. They might flag that they think the patient is confused . . . I’ve never had anybody say: ‘oh we’re worried about . . . and this was their score’.

Junior doctor, Care of older people ward, Denbury 03

Among doctors, the rationale for using the AMT10 was twofold: it was the preferred tool on the wards and on their clerking-in checklist; and it was familiar: ‘I know it by heart; I don’t need to refer to it’. The 4AT, on the other hand, was unfamiliar and the meaning of the score as a representation of the patient’s presentation was not obvious.

Generally, the picture that nursing staff conveyed was that if the 4AT was completed on the MAU, it was not repeated on admission to the ward; only in its absence would it be done on ward admission.

‘Delirium knowledgeable’ nurses emphasised their reliance on behavioural cues that might indicate a change and a probable new episode of delirium; the formal system for delirium detection using the 4AT did not feature:

I think the nurses would go to the doctors . . . and say ‘we think’ or during a ward round . . . or one of the rapid rundowns we would say: ‘overnight this . . . [happened] . . . ’what do you think, we’ve sent urine . . . blood’. A lot of the time we do take the initiative . . . if we think there’s an infective source.

Staff nurse, Care of older people ward, Denbury 02

These nurses would take the initiative in investigating the source of the observed changes, drawing on tacit knowledge – ‘deep down you can distinguish delirium from dementia’ – and on learned and experience-based knowledge: they did not wait for medical endorsement of the decision before acting.

Among staff generally, it was reported that serial use of the 4AT did not occur, or even that it was appropriate:

It’s almost by the time [the patient gets to the ward] I suspect we’ve already got all the information that a 4AT would capture and so we should be doing more than just a 4AT . . . it gives us a baseline . . . then it’s not just screening . . . it becomes an investigation doesn’t it? . . .

Occupational therapist, Care of older people ward, Denbury 22

Among some senior doctors, the interface between the detection and management of patients with delirium posed challenges that required further work to address:

I think everybody should be screened, but then it’s knowing what to do after the screening tests, and how to do it . . . it’s then the follow-up, and I think that’s what we struggle with . . . that’s a general problem [not unique to us]. We should reassess things . . . [some patients with delirium] might well have an undiagnosed dementia that we can’t diagnose yet because they’ve got delirium . . . and for [new cases] is a different approach needed?

Consultant, Care of older people ward, Denbury 15

Integrating detection with the work of management

In the MAU, the purpose of delirium detection using the 4AT was broadly understood by staff, and there were systems in place to include the 4AT within the nursing UPR, thereby establishing clear lines of accountability for completing it as part of the nursing assessment. By contrast, ward nursing staff conveyed considerable uncertainty about the purpose of the 4AT in this setting. Although it was included in the nursing assessment, medical staff assumed the lead in assessing cognitive impairment and forming an action plan in relation to it and nursing staff seen to provide a supporting role in feeding back observational cues that might be indicative of delirium.

Senior staff on Denbury wards referred to routine MDT meetings as arenas in which the progress of recovery from an episode of delirium was discussed and new cases of delirium were identified. This included decision-making on whether a patient was safe enough to move to another ward or to engage with rehabilitation or for the team to assess the person’s progress with recovery. They saw nursing staff as playing a critical role in this decision-making as a consequence of their more direct and ongoing interaction with patients. Even so, for nurses, the question posed was where the tool fitted into routine practice.

Aspects of the ways in which delirium management was accomplished on Avonfield’s care of older people ward are also illustrative of integration of delirium detection and management and co-ordination of interdisciplinary work is practised. Similar to on Denbury wards, delirium was initially identified in the EAU, using the AMT10. Patients with delirium then underwent routine investigations to examine aetiology. As part of the admission to the ward, a joint treatment plan was discussed at the MDT meeting. The ward care culture prioritised spending time with patients, and also drawing on the families’ own knowledge of the person and engaging in practices that were regarded as the ‘good nursing aspects of delirium management’. These included non-pharmacological strategies, such as attention to hydration and nutrition, sleep, hygiene, mobilising patients and stimulation. Nursing and care staff reported that knowledge of patients, aligned with understanding of delirium, meant that they were sensitive to what was ‘normal’ for the person. Observations were communicated formally and informally through forums such as MDT meetings and handovers. Person-knowledge informed how staff worked with patients to effect a calm and supportive environment:

Some of our patients with delirium we have to provide one-to-one nursing which makes it difficult for the rest of the ward ‘cos that takes one whole person off the ward team . . . so a patient we had recently, he just wanted you to walk around and hold his hand all the time, otherwise, you know, he’d get distressed or fall or something, so it was doing that. Or it’s observing somebody from a bit of a distance ‘cos some patients get overwhelmed if somebody’s with them all the time. It’s just like I say, it’s just assessing each patient individually.

Senior nurse, Care of older people ward, Avonfield 08

The ongoing review of progress with treatment plans, supported through the systematic observation of patients’ emotional, behavioural and relational action and interaction and involvement of families, contributed to the evaluation of progress. The work was conceived of as a joint and shared enterprise involving all members of the MDT. It was the consistency in the way in which joint work was pursued that reinforced the sense of value in each other’s particular disciplinary expertise.