Participants’ characteristics
Thirty-one patients with dementia participated in the study and were observed for a total of 170 hours at bedside ( and ). A total of 52 health-care staff and four carers were interviewed (). Interviews were supplemented with informal conversations recorded in field notes with staff and carers during observations. Among the health-care staff interviewed across the four case sites were seven HCAs, 31 nurses (staff nurses, sisters and specialists), three doctors in training, five medical consultants, a pharmacist, a physiotherapist and four clinical educators. Ward management was informed that all members of staff were invited to take part in interviews, but there were challenges in finding time when they would be available and researchers tried to minimise disruption to normal work activities.
Data collection at each case site: patients and duration of observation
Observation periods at the four sites
Data collection at each case site: interviews (number of participants)
Patients included in the study had a mean age of 88 years (range 75–99 years). A number of challenges were encountered in recruiting patients for all case sites. Among them was the lack of a documented diagnosis of dementia in ward notes despite strong indications that the condition was present, and many patients had no available carer.
Observations took place at the bedside, and included any HCP interacting with the patient. It did not follow specific members of staff. As a consequence, we do not have data about the ‘current shift’ from the HCP perspective, so data collection activities could not be recorded in terms of shifts.
Our initial research protocol planned for ‘40 shifts per case sites’ (i.e. 20 shifts per hospital ward), but this had to be interpreted as ‘periods of observations’. For the purpose of reporting data collection activities, a period of observation was defined as a period of continuous observation at a patient’s bedside in any 1 day. One or more periods per patient could be done per day of data collection.
Our periods of observations ranged from 30 minutes to more than 6 hours (see ).
Themes emerging from qualitative analysis
The analysis identified four overarching themes which are discussed below: communicating pain with dementia; carer–clinician communication; use of analgesics; and putting a picture together.
Communicating pain with dementia
In assessing and managing a patient’s pain, the guiding principle used by staff was to rely on self-report (the patient vocalising their pain), with the patient being the main information source.
However, for many of the patients observed in the study, communication barriers of various sorts hindered a patient’s ability to provide staff with information on their pain experience. These included issues related to language and cognitive impairment, the impact of patterns of work on a patient’s ability to communicate, and issues of trust and familiarity. We describe and discuss each of these aspects (see Language and cognitive impairment, Patterns of work, time, location and division of labour and Trust and familiarity).
Language and cognitive impairment
Patients with severe dementia showed significant communication difficulties. Interviewees explained that questions about pain should be rephrased to account for this. For example, when questions were asked for the purpose of gathering and recording a pain score in a form structured with the three options ‘mild, moderate and severe’, these were to be translated into words patients could understand, taking into account the ability of each individual:
Somebody might have no concept of what moderate means, for example.
Nurse specialist, H1
Patients appeared to be using various gestures, postures, bodily movements, behavioural prompts, metaphorical expressions and a combination of these in what was interpreted as an expression and communication of pain. Data from interviews suggested that nurses and clinicians also looked at physical and behavioural signs to understand patients’ pain. One interviewee commented that the identification of these non-verbal communication cues depended largely on staff skills, experience, knowledge and perceptions and added that ‘we need to get staff to think differently’ (nurse, H3).
Some patients made use of metaphorical expressions to communicate their experience of pain. For example, a patient explained the pain in her knee as ‘It’s murder, it’s awful. For quite a while it’s alright and then suddenly it’s [sic] murder’ (field notes, H1) and another patient expressed headache as ‘red hot’ (field notes, H3). An interviewee suggested that ‘it is not always what is said, but how it is said’ (nurse, H3) that gives a cue of a patient’s pain.
Patterns of work, time, location and division of labour
The challenges raised by cognitive impairment in patients were compounded by the organisational context and time frames of staff interactions with patients. Organisational routines and staffing numbers meant that the majority of patients’ encounters with staff during their hospital stay were brief, sometimes extremely brief, and encounters with senior staff members were less frequent. One patient said, ‘they are always dashing’ (field notes, H1), or in the words of a carer, ‘it’s just a fleeting glance, they talk to them and off they go’ (carer, H2):
As an FY1 [foundation year 1 doctor], if I’m only seeing the patients for a few minutes at a time then the rest of the time looking at their investigations or trying to organise things, I don’t usually, I wouldn’t say that I actually get to know the patient as a person [. . .]
Foundation doctor, H2
Furthermore, given the organisation of staff work over rotas and shifts, a patient’s communication about their pain was not always with the same member of staff but may have involved several different HCAs, nurses and doctors. There was awareness among staff that patients with dementia would need more time than usual to communicate pain. For example, a local pain dementia care plan guidance (H1) did include instructions for nurses and HCAs to give enough time to the patient to communicate, but in practice this was a matter of minutes. Indeed, in interviews the lack of time was often voiced by the nurses as a concern:
There’s not enough of us, and we just haven’t that time.
Staff nurse, H1
Patients shouldn’t go a long period of time without their pain being reassessed [. . .] certainly you’ve got to go back, it does recommend half an hour, but as I said, it’s quite difficult to get back within that period of time.
Senior charge nurse, H2
These brief encounters required the patient to be ready to answer questions and to recall their pain experience with little or no forewarning. Moreover, these opportunities at times occurred while patients were otherwise engaged in eating or sleeping, or when they were not prepared to discuss pain.
Patients were directed to use a call button at bedside (a ‘buzzer’) to request assistance. Patients with more severe dementia appeared not to recognise the purpose of the buzzer or forgot it was there, thus their ability to communicate pain was severely limited. In addition, calls for help, including those using a buzzer, could not always be answered immediately, leading to distress and confusion for the patients concerned. Some patients seemed to have no memory of having used the buzzer when a staff member arrived. In other cases, patients expressed a disinclination to use the buzzer and disturb busy staff, or they did not know what the buzzer was ‘for’, thus rendering it unhelpful to the person with dementia. Cases were also recorded where patients verbally reported pain, but at a time when there were no staff members present. During bedside observations it became apparent that it was necessary to be in close proximity to the patient in order to communicate, and in the same bay or at bedside, as patients rarely left their bed or chair. At times, patients’ needs surfaced out of background noise, or while the staff were in the room with other patients, and patients were able to attract their attention through verbal communication or behaviour. However, this was not always the case, and some patients were alone and without interaction for relatively long periods. Many of the ward routines, such as note-keeping and handovers, took place away from the bedside, thus minimising a clinician’s opportunity to communicate with the patient during periods of alertness, or to detect subtle changes in expression and engagement.
Trust and familiarity
Clinicians were aware that relationships of trust and familiarity were important for a patient to communicate their pain. One Elderly Medicine Consultant explained: ‘if you feel that somebody cares about you then I’m sure it makes it easier to express it if you’ve got pain’ (medical consultant, H1), and a student nurse with a HCA background stated, ‘if you are connected with your patient, if you, they know you and they trust you, if you build that rapport, that itself will allow you better access to how their pain is’ (student nurse, H4). However, it seemed to the researchers that establishing these relationships was not facilitated by the brief time frames available for communication and patterns of interaction.
Carer–clinician communication
Relatives, visitors and carers represented an important information source in the recognition, assessment and management of pain. One ward sister explained the reasons for involving carers in the process of pain recognition, assessment and management in terms of ‘how well’ the staff know a patient compared with carers, carers’ ability to communicate on behalf of the patient, and how HCPs would be guided by carers in the management of their loved one’s pain.
[carers] would know the patient much better than we would and they would be able to assess whether [the patients with dementia] they’re in pain or how they’re feeling, and communicate with us. So yes we often ask for input from the carers or the family members to guide us in how we’re managing the patients.
Deputy sister, H4
Carers were observed as acting as messengers on behalf of the patient, and helped recognise and interpret pain cues. An example of this process is described in a quote from a carer interview:
My mum has a terrible habit, even though we know as a family, when she needs to go to the bathroom she starts shaking her leg and she was doing that in hospital and she was getting in a panic, she actually was crying, she went in with a really bad urine infection which was causing a lot of pain at the time [. . .]
Carer, H2
However, the observations conducted in this study showed that the majority of staff communications with relatives were concerning medicolegal reasons of consent or about discharge arrangements, and less about ‘the needs of how [patients] are and how we can help them here’ (staff nurse, H1). This observation is strengthened by the lack of documented communication with carers found in patient records. In part, this was also because of the unavailability of family members in a number of cases (something also supported by findings of an internal audit in case site H3 reported by an interviewee, where more than half of respondents said they did not ask a relative ‘because the family were not there’). Several carers perceived staff to be occupied and were therefore reluctant to initiate a conversation, whereas others expressed that they did not perceive themselves as experts in the knowledge of the person they care about. Some were elderly and had dementia themselves. Family conflicts, domestic violence, poverty and deprivation also seemed to be complicating factors in some cases.
Staff expressed the belief that there is a need for clinicians to improve communication skills with carers. They explained this as having the ability to elicit the right information from the right relative/carer, and to assess the trustworthiness of these information sources. One nurse described ‘difficult conversations’ (H1), which required assertiveness and a degree of certainty to be acting on behalf of the patient:
I hear a conversation at the nurses’ station between nurses about another patient. A nurse spoke with the niece, says the patient is OK at home. The other nurse says the niece may not know, may only see her once a week, to talk with (social worker? assistance?), that they may have a completely different picture.
Field notes, H1
The skill that the juniors need to have is in digging out, ferreting out the information that is relevant to a person [. . .] in order for us to make an informed decision [. . .] it’s not just, it’s not just asking the question, ‘How was your relative before they came into hospital?’ It’s really understanding the nitty-gritty of the details, [. . .] of course I don’t have the time, unfortunately, to do all of that for every patient, so [. . .]
Medical consultant, H1
Use of analgesics
The most common pain treatment used in the study sites was analgesic medication. Indeed pain management and pharmacological management of pain often appeared to be one and the same. A limited number of non-pharmacological pain management strategies were used, such as patient repositioning, but these were not frequently observed. Clinicians considered potential side effects of medications, including confusion, when making treatment decisions.
In certain wards, depending on the ‘type of patients’, pain medications were prescribed routinely to all patients. For example, in a ward in case site H4, an orthopaedic surgery ward, patients received ‘pain relief already regularly on their charts for their surgical procedures’ and this prescription was administered ‘every 6 hours’ even if patients did not report pain (staff nurses, H4). This standard acute pain medication prescription followed an established protocol. The prescription was provided by the anaesthetists at the time of surgery. More complex cases, or when pain was clearly not under control, required escalating out of the routine prescription through contact with a specialist pain team and trialled combinations of different analgesic drugs. In other sites, pain medication was administered on an ad hoc basis depending on the patient’s medical condition. In these cases, pain management was commonly approached in a trial-and-error mode, titrating the dose gradually and assessing the patient’s response.
Importantly, the process of (re)assessment for the purpose of establishing the most appropriate medication encountered similar communication challenges as described above, as not all patients were effectively communicating changes in their pain. Judging the level of titration, or the appropriate step in the analgesic ladder, relied on clinicians’ knowledge or ‘sense’ of what the expected pain medication for a given medical condition would be:
Obviously a knowledge of the reason why they’re in hospital and if they’d had a particular surgery, of knowledge of what is happening within the body. [. . .] And how much pain relief somebody would need for that.
Deputy sister, H4
Pain relief medication prescribed to be administered as and when needed [pro re nata (PRN)] was usually considered part of this process of titration, but could not be used effectively with patients with dementia who would not request additional pain relief.
‘Putting a picture together’
Overall, understanding a person’s pain in these acute hospital wards involved investigative work and ‘putting a picture together’ of an individual’s pain (‘we’re trying to build a picture’, staff nurse, H1). This process required time and availability of information from various sources, including carers, the MDT assessment, administration of medication and the patient’s response.
The observation of the context of care and document analysis revealed that patient information was shared through face-to-face encounters and written documents such as patient records, medical and nursing notes, transfer reports, checklists, care plans and drug charts. The drug chart was frequently referred to by the majority of team members, and a number of staff respondents in the study stated that they used information available on the drug chart to assess, reassess and review both medication and care plans.
Importantly, however, paper-based documentation was fragmented, not easily accessible or poorly organised. The various documents retained and used by different HCPs were kept ‘in silos’:
So quite a lot of nursing work has to come down in silos so we have nutrition, tissue viability, falls, dementia, and it doesn’t necessarily speak to each other on paper, which I think we’ve quite siloed risk assessments that it’s then difficult to put together holistically.
Nurse manager, H2
In one ward, the intentional rounding forms (for routine comfort checks) were filled in every 2 hours, in what appeared to have become an administrative, rather than investigative, exercise. Staff respondents in the study raised concerns over the large amount of paperwork, some of which they considered redundant:
When we fill in care plans, we’ve got the specialist assessment [forms] and they say the same things as your care plans.
Sister, H1
We also identified ambiguity in documenting the absence of pain. The interviewees reported the tendency to assume that the patient is not in pain if patients’ pain is not recorded in the documents:
If there’s nothing written, I suppose I would assume the question hasn’t. . . has the question been asked? I don’t know, I probably from a personal point of view [. . .] I would have asked it but probably not documented that there was no pain. If I don’t see anything written I would assume that the patient hasn’t complained of pain but I suppose what I can’t say is that they’ve been asked if they’ve got any pain.
Doctor, H1
No use of decision support tools was observed in any of the settings studied for pain assessment or management. In one site (H1), the Abbey Pain Scale101 was recommended in the local set of documentation, but was not available, or appeared not to be known, to the staff. Instead the site used a pain care plan, which was written anew for each patient in a loosely structured form, with narrative entries at assessment. This was used with all patients, with or without dementia, and with bare information recorded regarding patient experience of pain and/or what intervention was used.
One site utilised electronic documentation and at the time of the study it was in the early stages of implementing the Abbey Pain Scale in electronic form. However, no data could be collected regarding this and no staff were observed using it. A manager reported how the tool had been trialled, but ‘not well used on the ward’, that the criteria for using it were unclear and that it ended up being used as a ‘tick-box exercise’ (ward manager, H3). Wards in case site H2 had recently implemented a generic pain assessment (GPA) form, developed locally, to be used alongside the PACSLAC tool.102 In the period of our study, for the patients we observed in this site, the GPA was often with a patients’ drug chart, but left blank or showing only initial entries without follow-up:
Nurses have so many assessments now to do that it’s, they’ve kind of lost their credibility a bit, [the GPA form] it’s just seen as a form and a tick-box exercise [. . .] it’s another thing to do and yet they have a hugely frantic day [. . .]
Clinical educator, H2
Conclusion
Pain assessment and management are complex activities that embrace physiological, emotional, cognitive and social dimensions. Pain is often described as a private experience, but in reality it regularly requires public expression in order to obtain relief. When caring for people with dementia, the complexity of communication is exacerbated by the challenges that the progressive cognitive and functional decline present. These complex social processes occur within hospital contexts and routines organised mainly around the needs of the organisation rather than those of patients. This complexity may, at least in part, explain why simple tools available to assess pain in patients with dementia are not well used in practice. Future decision support interventions need to take this complexity into account during their development, reorganising frames and quality of time for communication with patients, making a more varied range of pain management interventions routine, and devising tools that bring together information to provide a ‘picture of a patient’s pain’ accessible to all involved, within an overall framework of person-centred care.
