The care of dying people in nursing homes and intensive care units: a qualitative mixed-methods study

Nursing homes

A hierarchical ‘top-down’ approach was adopted to the recruitment of nursing homes. Nursing home managers and owners were contacted by telephone to establish LCP-using status, and to ascertain whether or not they would be interested in receiving information about the study. A letter (see Appendix 10) was then sent to all who wished to receive it stating the purpose of the study and the nature of their potential participation. This was followed up by a telephone call to arrange an introductory meeting. If site participation was agreed in principle at this meeting, arrangements were made for the researcher to meet with individuals and groups of staff to give further information about the study.

As previously discussed, the researcher aimed to see at least 50% of the workforce, to give the staff an opportunity to ask questions, express concerns and talk about the proposed research. The researcher would meet the staff in small groups or as individuals, and a record was kept of the names and numbers of staff seen in this way. Visits and meetings were held at different times of the day, including through the night, in order to reach as many staff as possible. Any staff who felt strongly that they did not wish to participate in the study were asked to let the researcher know by e-mail or telephone (to ensure confidentiality), and this information was noted so that these individuals would not be approached for interview or for the observational element of the research.

Intensive care units

A similar recruitment process was adopted in the ICUs. Managers and consultants were approached initially by telephone to check their LCP using status and to gauge initial interest. They were then sent an introductory letter and more information (see Appendix 10). A follow-up telephone call was made, and if there was interest in participation at this stage a face-to-face visit was arranged. If agreement in principle was achieved from ICU managers at this stage, meetings were organised for staff, primarily nurses and doctors, which in the ICU setting were often formal and provided staff with an opportunity to ask questions and/or voice concerns. Again, these meetings were held at various times of the day and night in order to reach as many relevant staff as possible. The researcher kept a record of staff attending these meetings. In some of the larger ICU sites employing over 300 clinical staff, it was only possible to see a small proportion of the workforce. The aim in these large sites was to cascade general written information through the system, using notice boards and by word of mouth. Staff were invited to approach or contact the researcher if they had any questions or concerns about the research.

Nursing homes

Staff turnover in nursing homes was a particular challenge for the recruitment and retention of sites in this study. This was predominantly the case for sites in London and the south-east. Six nursing homes were quite quickly recruited to the study and a significant amount of work had already been undertaken to introduce the research to individual staff in each site. However, three of these subsequently withdrew as a result of a change in the home’s management. Several attempts were made to recruit replacements, but it proved possible to replace only two of the three homes in time for meaningful data collection to take place. An additional problem was experienced with another nursing home in the south-east, where a decision was taken by the management team half-way through the study to opt out of the observational stage of the research. Unfortunately, this happened too late in the course of the study for the site to be replaced.

Similar problems were experienced in the north-west sites. In one, home agreement for the observational phase of the research was withdrawn 5 months after their initial agreement to participate and following a significant amount of data collection (documentary analysis and point 1 interviews had been undertaken). A replacement was found for this nursing home. Another home was taken over by a different management company during the course of the study and, although the home did not formally withdraw, no opportunities for observation were made available to the researcher after this point.

Similar challenges to recruitment and retention have been reported in the literature. For example, Tilden et al. 42 reported a higher rate of turnover of key personnel in ‘non-completer’ sites in their study of the relationship between communication, teamwork and palliative/end-of-life care practices on patient outcomes. They concluded that researchers wishing to undertake large-scale studies in nursing homes should ‘anticipate significant challenges and delays in recruitment and retention of the sample, and should budget accordingly’.

Intensive care units

The speed of recruitment in the ICU setting was slowed by uncertainty about whether or not the LCP was being used. Many of the sites in the north-west had implemented the LCP at some point in some shape or form and it was necessary to recruit from a wider geographical area than originally planned.

In addition, at the time of the recruitment, ICUs were experiencing significant organisational change. Some ICUs were in the process of merging and the allocation of level 3 beds across the sample often changed in response to emergent need; the maximum number of level 3 beds that were available in each site are detailed later (see Table 3). Robust statistics on the proportion of ventilated patients in ICUs at any one time was difficult to locate and the proportion of patients required for inclusion in this study (> 80% mechanically ventilated) was not always reached. Together, these factors made the identification and recruitment of appropriate sites challenging and time-consuming.

Level 1: background, contextual data about views, attitudes, and documented practices surrounding end-of-life care in each setting are explored through data types 1–3

Level 2: case study approach, centred on the observations of patients to allow a detailed contextual analysis of a limited number of events from a variety of relevant perspectives, explored through data types 4–8

The following five ‘types’ of data are all inter-related and driven by the observations of care in the last hours or days of life. The central component of this study was the collection of observational data, alongside the accounts of staff who had delivered care during the observed period, the case notes reporting the care provided and the views of bereaved relatives about the care during the observed period.

Patient choice at the end of life

One important element of a ‘good death’ identified by senior staff was the importance of taking the patient’s stated wishes and choices into account. The promotion of patient choice at the end of life was reported to be a key objective in ACP.

It’s all about . . . anticipating someone’s death and giving them the opportunity to have the death they wanted. Discussing it and not being afraid to broach the subject with someone . . . we’ve got to take away the stigma of death because it’s inevitable. We’re all going to die someday and it should be part of your own choice. I think we’re trying to include it in the patient’s choice.

NNB4S133

All of these senior nurse/managers reported that they had a process in place to facilitate engagement in such discussions and there appeared to be a high degree of correspondence across the sample in the stated purpose, nature and content of these discussions. Offering patients some degree of choice over particular aspects of their own end of life required managers and nurses to have the confidence and skills to open up discussions with patients (where they had capacity) inviting them to contemplate their own death and with relatives to contemplate the death of their loved one. These more experienced respondents broadly felt that they had the skills to meet this challenge.

I think with being experienced and being mature I am well able for that. I think I can do it sympathetically and in a language that they understand and I think I’m approachable enough for them to . . . come to me if they have any questions or anything like that. I think that when you’re running a shift in nursing homes you have to be prepared for that. It’s an unfortunate part of the job. It’s not the nicest part of the job but still in all it’s somebody’s job and that’s ours isn’t it.

NNB4S133

Do not attempt cardiopulmonary resuscitation, hospitalisation and preferred place of death

A major focus of these ACP discussions was to document important decisions about life-saving interventions and the patient’s preferred place of death.

We try and discuss sort of what they want, whether they want to be resuscitated or not, if they want to go to hospital and their preferred place of death really . . . so we discuss that with the resident . . . especially if somebody wants to be resuscitated or not, we try and get that ascertained straight away.

SNA2S136

Although the above example suggests that patients can choose to be resuscitated, this was recognised by others as being a medical decision.

We discuss resuscitation status you know but it’s the GP’s decision at the very end of it.

NNB6S154

One respondent also spoke about the importance of asking patients and relatives to consider other issues that might be part of an advance decision to refuse treatment to ensure that the patient received the ‘best treatment’. While she made it clear that she did not take such decisions on behalf of patients, she illustrated the strong personal beliefs that underpinned her drive to offer these choices:

We would never assist dying but you know we have to look into . . . do we need an advance directive, are they choosing not to have antibiotics for their next chest infection . . . that’s where I would come in. I think people have you know the right to have the best treatment. To prolong someone’s life if they’re not having very good quality of life does that make a lot of sense to keep giving them antibiotics if they have absolutely no quality of life? That’s something they and the family need to decide. I don’t take that decision at all. I just make sure that that it’s legally written properly . . . I would hope . . . in a few months, that we could be sure that everybody who is, wants to be DNR [do not attempt cardiopulmonary resuscitation], it’s signed and it’s written in their chart and it’s on this Coordinate my Care.

SNB1S166

Although the majority of the managers felt that elderly people should be enabled to die in the home, in familiar surroundings, there was recognition that if a patient decided that they did want to go into hospital at the very end of life, their choice should be respected.

We have to respect their wishes. If somebody came to you in the home and said ‘oh I’m not well I want to go to hospital’ we have to record it. We cannot force them, this is their wish.

SNA3S136

At the same time as recognising that patients should be able to make informed decisions about hospitalisation at the very end of life and understand when death is imminent, some respondents were clear that hospitalisation did not offer anything that was not available in the nursing home.

But you have the discussion with them that is there anything in hospital that they can do that we can’t do? Which theoretically there isn’t.

NNA5S166

But then you go and explain to them what would the hospital do that you won’t get in the home? What is the difference? Then you explain to them that ‘you are better cared for here because you have got your own room here. Anything you want we can give it to you here. Your family can be with you here’. So we explain to them afterwards and they will agree. We don’t force them, we don’t pressure them, but we explain clearly to them.

SNA3S136

An additional component of these discussions involved eliciting those elements of care that patients might consider to be important to them in the future.

We find out what they like so even down to . . . there’s patients that would love a hug, there’s patients that would not want a hug, there’s patients who would like to hold their hand, there’s patients that would like their face stroked, there’s patients who wouldn’t. It’s finding out . . . those finer things so it’s there at the end. And also all the spiritual bits, you know, because spirituality is very different, very different to everybody really. So it could be that you want the priest there, you know, every day saying prayers, it could be that you don’t want that, that you want to make friends with somebody that you’ve, you know, fell out with years ago. It could be anything erm it could just be listening to a piece of music, I don’t know. But finding out all that beforehand is really important so that it’s what their wishes are, it’s what we follow out.

NNA2S116

Timing of advance care planning discussions

The GSF promotes early ACP discussions with all residents. It was reported that, for those patients new to the nursing home, these conversations took place soon after admission.

They [manager and other trained staff] make sure that when they come in . . . we don’t do it on the day but within a week we have to have finished the plan. We have to complete an identification form and an advance care plan.

OK so within a week of arrival you have to have completed the advance care form and the DNAR [do not attempt resuscitation] form?

Yes and the care plan as well.

Right OK so with regards to end-of-life decisions, what kind of things are in the advance care plan? What are you asking them at that point?

Actually asking them if they want to be sedated, the first choice . . . of place where they want to . . . be you know when they are passing and if they have any special requests or any person they want to be with them at end of life. Also if they want a vicar or priest to come and give them the last rites.

SNA1S166

Speed in gaining this information was reported to be of particular importance for continuing care patients who had been sent from the hospital setting specifically for end-of-life care. These discussions, which were reported to cover arrangements for cremation or burial, would sometimes happen within 1 or 2 days of the patient’s admission to the nursing home. In the context of a short relationship and from a position of vulnerability, it is not difficult to anticipate that some residents might find this troubling.

I say ‘where would you like to be cared for in the last days of your life, and your end of life care?’ . . . So we have to prepare them for that . . . get all the questions answered before and also whether they are for cremation, for burial, what are your likes and dislikes, who would you like to be with you when you are at end of life care. We have to get all these clear before [they deteriorate] . . . we do that at the start, as soon as they come in . . . I mean . . . at an appropriate time obviously. Maybe after one or two . . . or 48 hours but after that it may not be possible to do that.

SNA3S136

There were examples, however, which illustrated a more sensitive approach, where the timing and pacing of these conversations took into account the perceived disposition of patients and families at the point of admission. Some nurses reported visiting the patient and family prior to admission to try to gauge how appropriate or otherwise it would be to engage in such discussions and when the best time to do this might be.

I usually go out and do the assessments so you get to know the people, OK not that well but you sit and chat with people in the hospital and with families when you meet them and you know whether you can broach the subject of end of life or you shouldn’t broach it just yet. Some people don’t like you to you know, it’s bad enough having to come to a nursing home, they are all distraught, devastated, feel so guilty and they don’t want me sitting there going ‘so end of life what do you think I should do, what would you like to happen to your mum when she dies’. I don’t think it’s appropriate. Once you get to know them a bit better you can sort of broach the subject and I tend to, if you are getting doctor in a few times and they’ve got lots of things going on with them, then you sort of say ‘look what are we doing’.

NNB1S155

In some sites (e.g. NNB6S154 and NNA5S166) the opportunity to involve patients directly in such conversations was reported to be limited as a result of issues of capacity through conditions such as dementia.

The drive to include all patients in the nursing home in such discussions sometimes meant that patients who were not expected to die within, for example, 12 months were participating in ACP discussions. Undertaking such discussions so far ahead, particularly where they included advance decisions to refuse treatment, raised issues about the appropriateness of the timing of these conversations.

Now because we are doing some advance care plans quite well in advance the GPs aren’t as happy to be filling out DNAR forms which I can understand because obviously they need to have a discussion with the relatives and everything but sort of for the DNARs it doesn’t make sense . . . it’s a bit mixed up . . . and I also think with the Harold Shipman thing the doctors aren’t too, not all of them are really happy to be doing a DNAR form for somebody who might die in 3 or 4 years.

NNB6S154

Refusal to participate in advance care planning discussions

Two instances were reported by staff of a patient and their relative refusing to take part in ACP discussions. Both patients were male and in their sixties. In SNA3, although there was recognition that the patient and his brother were finding the discussion difficult, a sense of frustration with this situation was also clearly evident.

So we have a gentleman who is 63 years old . . . he has prostate cancer and we are not able to deliver the [advance] care plan . . . he does not want the care plan, next of kin do not want the care plan also, they think he is too young. He is 63 years old . . . so we are in limbo with that. So these things do happen at times . . . he is not very co-operative at times . . . so we haven’t done the advance care plan we can’t do it.

SNA3S136

In NNA3, the patient and his brother were described as ‘flatly refusing’ to state their wishes. The staff suggested that the reason for this refusal was either that they did not understand or that they were choosing not to think about death.

They won’t do it . . . he’s been with us donkey’s years and he is still refusing and we still every so often say ‘would you like to?’ there’s only him and his brother and they go ‘no, no’ . . . he’s a younger man . . . he’s one of our younger gentlemen that’s in his sixties but . . . erm no they don’t want to be involved in anything like that . . . they’ve not understood or . . . they don’t want to think about it . . . we’re not here to push advance care planning but . . . we do try and do it with the person and the family as soon as possible. But like I say, some families do try and put it off.

NNA3S116

Regular review and sharing of advance care plans

Nursing staff highlighted the importance of the regular review of these plans, as it was acknowledged that patients (and relatives) may well change their minds about particular aspects of their care over time and in response to changing circumstances.

When the time comes that they are poorly . . . they still have that, how they feel and deliberating and they might change their mind . . . so we say to the family we will review the advance care plan every . . . year or if there is any significant change.

SNA1S116

Gold Standards Framework

As illustrated previously in Table 5, the GSF was reported to have been implemented in seven sites. The GSF aims to promote better quality end-of-life care in nursing homes through proactive planning, enhanced relationships with GPs and other health-care professionals, working with teams to raise staff morale and increasing the practice of ACP, all ultimately with the aim of reducing the number of hospital deaths. 59 The respondents in this study agreed with these aims. Where the GSF and the LCP had been implemented in an organisation, respondents generally believed that they had made a positive contribution to care delivery, particularly in avoiding inappropriate admission to hospital.

I think the Gold Standard and the pathway has made a massive difference that you can go home now [at the end of a shift] and know it’s not going to be left to you, that the next nurses won’t be railroaded and . . . [patients] going into hospital because it’s all organised. Everybody knows what’s going on, the family knows, the GPs know, the out of hours know and we even had a policy . . . with the hospital now . . . so if they go in for something quite acute and they sort it but they’re quite poorly . . . they can fast track them back to us.

NNA2S116

For others, the GSF (and LCP) have also helped to normalise death and dying and allow staff to feel that they have done the best for a patient, even when that patient dies.

Here there’s more of a feeling that there’s care of the dying and you did well while you were caring for the dying and then they died and you reflect on that . . . you know we had someone that came in for a day and died . . . but here it’s more like, well, that’s OK because that’s what he came in for . . . so I think it’s part of, because of the Gold Standards and because of the Liverpool Care Pathway I think it’s just making more people more aware.

SNA2S136

The GSF focuses primarily on patients deemed to be in the last 12 months of their life. The first element it seeks to address is ACP. Staff are encouraged to undertake an advance care plan with a resident ‘soon after admission’30 and there is no doubt that this requirement was successfully achieved in a number of sites; see Advance care planning for further consideration of this element.

The next requirement of the framework is to code the ‘status’ of each patient against a four-stage coding matrix and then to assess and monitor the progress of all patients against this.

We have this coding, this ABCD and A is kind of everybody, B is when people have started to go down but you know nothing drastic but you’re seeing them started to deteriorate a little bit, C is when you think there is only actually a few weeks left which means that they really are going down now and become very poorly so then you look at the whole, the whole of them and see what needs they might have that they won’t be able to do themselves and then . . . at D is like days, so we’re on days, we’re talking about you know getting the GP in and if they’re not eating and drinking, etc., which we just follow protocol, that’s when they go in the pathway [LCP] so we’ve actually got everything in ready.

NNA2S116

Each patient is then to be discussed at a monthly multidisciplinary meeting, which includes professionals external to the home, for example specialist palliative care nurses. In NNA3S116 the nurse reported that health-care assistants also attend these meetings.

But every month, everybody in the home as part of our GSF meetings, gets a full review . . . and that ties in with our care planning as well . . . the carers do come to the meetings as well . . . and it’s just a bit of an education session as well, especially when the Macmillan nurses turn up as well.

NNA3S116

Staff also reported the presence in sites of a process for monitoring, documenting and sharing any changes in the coding of the patient’s status that occurred between meetings. Finally, the GSF encourages critical reflection and discussion between those staff from within the home and those external to it (if appropriate). In NNA5S166 the nurse reported that this kind of debriefing also takes place after a patient’s death in order to highlight any lessons to be learned:

On handover we have GSF meetings so we’re discussing there what went well, what didn’t go well. Because . . . all of our residents that . . . that do pass away, we actually complete a form as to say what went well, what didn’t go well . . . and that’s open for discussion . . . so we highlight problems . . . and if we do have problems we can take it to the end-of-life team.

NNA5S166

Process for commencing the Liverpool Care Pathway to support care

Nursing homes were, by definition, situated within the wider health-care environment and were affected by wider primary care issues in their ability to deliver care at the end of life. Where the LCP was used, some issues were highlighted regarding its use in everyday practice. The main issues related to the timing and purpose of the joint assessment, the GP signature on the LCP and the prescribing of prn [pro re nata (as required)] medication for symptoms that may occur in the last hours or days of life. As with previous versions, commencement of the LCP (version 12) required the signature of the most senior medical professional caring for that patient to confirm the multidisciplinary decision that the patient was imminently dying; that is, they were believed to be in the last hours or days of their life. The use of the LCP to support care is predicated on the requirement that this assessment be undertaken by both a doctor and a nurse. The assessment should involve the consideration of any potentially reversible causes for the patient’s current presenting condition and, where appropriate, should include advice and support from relevant others. 33 After recognition that the patient is imminently dying, the GP should prescribe anticipatory medication for five key symptoms that may occur in the last hours or days of life. In version 12, it is expected that once the LCP has been started, the GP will visit the patient to undertake a further MDT review a minimum of every 3 days, or sooner if either a significant improvement or a decline in the condition of the patient is noted. Three sites (NNB6, NNA3 and NNA5) were using version 12 and four sites (SNA1, SNA2, SNA3 and NNA2) were using an earlier version.

The nurses in sites that had adopted version 12 reported that the level of involvement of the doctor was both important and reassuring.

It gives you more confidence that you’ve got a doctor actually going through it with you.

NNA3S116

This respondent, however, also recognised that gaining this level of involvement was a challenge. Here, and in NNB6, some GPs were reported to ‘struggle’ with the practicalities of a joint assessment of the patient.

Similarly, confirmation of the recognition of imminent death outside normal surgery hours involved ‘out-of hours’ GPs, who were reported to be ‘reluctant’ to make a decision to commence the LCP because they did not know the patient.

I think the only ones we’ve had . . . problems with . . . are the out-of-hours doctors . . . even though we’re a GSF home and we go to GSF meetings at the surgeries, the questions that we’re now raising is the out-of-hours team in that if someone’s condition changes overnight and we’re finding an out of hours and we want to start them on the LCP, they’re actually quite reluctant . . . because they’ve got to make that decision is that person dying? Because they don’t know enough information about that particular resident. It’s not their own GP is it? If it’s something that could be treated. Is it something that’s reversible? So they’re having to make that decision when they don’t really know the patient.

NNA5S166

Four sites that were affiliated with the GSF (SNA3, NNA2, NNA5 and NNB6) used the assessment made as part of the GSF process to recode a patient from code B (months from death) to code C (weeks from death) as a trigger for requesting the prescription of anticipatory medication, rather than on commencement of the LCP, code D, to support care.

. . . the anticipatory meds we usually do before somebody goes on the LCP, but that’s part of GSF programme. [They are prescribed] when somebody’s actually recoded as C, where they’ve got weeks to live. We actually ask for anticipatory meds then.

NNA5S166

There was, however, no evidence in any of these homes that anticipatory medications were used at any time other than when the patient was later recognised to be in the last hours or days of life and the LCP commenced (i.e. recoded on the GSF from C to D).

In three of the sites (NNA2, NNA5 and NNB6), although these medications were requested when the patient was code C, the LCP was commenced at the point of the recoding of that patient from code C to code D. For example:

. . . at C stage, is where we get the anticipated drugs in . . . that’s when we get those in the cupboard at C. And then at D is like days, so we’re on days, we’re talking about you know getting the GP in and if they’re not eating and drinking, etc. . . . that’s when they go on the pathway so we’ve actually got everything in ready.

NNA2S116

The fourth site (SNA3) which reported weekly MDT GSF review meetings and ongoing practical support from the local hospice applied the same procedure as the other three sites to trigger the prescription of anticipatory medication. However, it would appear that a LCP could be signed by a GP but not necessarily implemented immediately. When asked if the LCP could be started at the weekend (when GP cover might be compromised), the respondent explained:

We have someone going on ICP [LCP] as I explained to you we have to have a GP to sign it. Let’s say you are going to start it now, or the next day, maybe next week, maybe 2 weeks’ time. It doesn’t mean you have to start it straight away.

SNA3S136

In SNA1, the recognition of imminent death was reported to be made ‘in house’, at which point the GP could be called to prescribe the prn medication. So, in effect, the LCP could be commenced and the GP would sign the forms later:

No, no, no actually no. We explain to the family that we have started and then we call the GP and they have to sign it. We can decide. Because if in case it happened on a weekend then we don’t wait 2 days before we can start the LCP. They might be dead by then . . . if on the weekday, if the patient is fading and we call them, we are going to say ‘OK doctor can we have the advance medication because it’s Friday’, over the weekend we are going to need it and we have to start the LCP.

SNA1S166

Formal multidisciplinary team review every 3 days

The LCP version 12 promotes a formal MDT reassessment of the patient a minimum of every 3 days, or sooner if either a significant improvement or a decline is noted in the condition of the patient. Again, this was reported to be an important mechanism for the continued medical support of the dying patient.

. . . and the fact it’s got to be reviewed every 3 days by a GP which didn’t happen before . . . so GPs now have to come out to see a patient that’s been on it for 3 days.

NNA5S166

However, the views of staff and GPs were not always aligned in relation to the frequency with which GPs should visit a patient supported by the LCP.

I said ‘now they are on LCP you need to visit every 3 days’. ‘No we don’t it’s once a fortnight’. So I said ‘no, they are on LCP it’s every 3 days’. ‘No you have got it wrong and you have misunderstood it’s every fortnight we visit’. I said ‘but the point of the LCP is that it’s within the last 3 days or the last hours of life, is when we implement, you know we make an assessment to see if they meet that need’. And he said ‘no no you have got it all wrong, she is palliative care’.

NNB6S154

Decisions about future care made with or on behalf of the patient

In line with the findings in Chapter 3, ACP and the elements of care deemed appropriate in advance of a patient’s deterioration were features of these records. The most common information recorded was the patients’ preferred place of care and death (72% of cases) and their DNAR status (78% of cases). However, there were interesting differences in the recording of this information between patients who were supported by the LCP and those who were not. In all deaths where the LCP was used to support care, the care home was recorded as being the preferred place of care; where the LCP was not used, this was the case for under half (39%) of cases. Similarly, for the vast majority (97%) of patients supported by the LCP, a DNAR notification or completed form was found in the case notes, whereas this was the case for only just over half (55%) of those patients not supported by the LCP.

These advance notifications were sometimes (although relatively rarely) the result of discussion with the patient; more usually they were a result of discussions with relatives and were often represented as best interest decisions made by GPs and nursing home staff. They could be documented within specific forms (e.g. advance care plans, GSF ‘Thinking Ahead’ or ‘Allow Natural Death’ forms) and/or recorded directly in the notes, for example ‘No hospitalisation’ (SNA2R04) or ‘Not for hospitalisation’ (SNB1R01); ‘for tender loving care in the home’ (SNA1R01); ‘discussed with her sister . . . Not for Resuscitation’ (as part of admission details SNB1R05); or, simply, ‘DNAR’ (NNB6R03).

The main purpose of documenting this information seemed to be to ensure that decisions made in advance of end-of-life care were clearly articulated and accessible to all staff to prevent admission to hospital in the event of further deterioration, or inappropriate life-saving interventions being undertaken in the event of cardiac arrest.

The condition of the patient, actions taken and subsequent outcomes

In the main, regardless of whether the information was documented on the LCP or within the nursing home notes, a major focus of these written records was to record the physical condition of the patient, the care that was delivered in response to this and any subsequent outcomes.

Where the LCP was used to support care, the process required that a series of desired outcomes for patients be regularly assessed (either 4-hourly or 12-hourly) and documented as either ‘achieved’ or ‘variance’. Where outcomes were achieved, it was usual for a staff member to simply insert ‘A’ in the document. The recording of ‘variance’ should have prompted further explanation, although this was relatively rare. The outcomes to be assessed included whether or not the patient had pain, agitation, respiratory tract secretions, nausea and vomiting or dyspnoea. In addition, they included an assessment of whether or not the patient had discomfort resulting from pressure sores, dry mouth, poor micturition or bowel function as well as psychological and emotional distress.

Where the LCP was used to support care, variances were recorded for agitation, respiratory tract secretions or pain. The most common action taken in response to these variances was the use of medication, namely midazolam, glycopyrronium (or hyoscine) and morphine (or oxycodone or diamorphine), respectively. The outcome recorded following the administration of medication was generally that the patient was more comfortable and settled.

SNA3R01 – 23.30: [patient] looks agitated. Action: Injection of midazolam 2.5 mg, s/c given (@23.35). Outcome @ 24.00: Settled well.

SNA2R01 – 06.30: awake and restless. 10 mg midazolam and 10 mg morphine sulphate given. Peaceful and asleep (@ 07.30).

10.30: awake and in pain. 10 mg midazolam and 10 mg morphine sulphate given. Settled and asleep (@ 10.50).

21.00: looks in pain and calling out. 10 mg midazolam and 10 mg morphine sulphate given. Comfortable and slept now (@ 22.30).

Where the LCP was not used to support care, information about the condition of the patient and action taken was recorded as an ongoing narrative within the nursing notes. Although the focus was again on the physical, much more emphasis was placed on summarising the general condition of the patient and any resultant actions and outcomes. Phrases such as ‘All care given’ and ‘TLC’ (tender loving care) were much in evidence within these narratives. Attention was paid particularly to recording staff attempts to administer food and fluids, undertaking regular mouth care, and washing, changing and repositioning the patient.

Day 1 – night: continues to be restless and agitated, pulling everything away. Repositioned, Fluids taken, monitored via fluids chart. TLC given. Medication taken.

Day 2 – 13.05: General condition very poor. Agitated at times. Fluids taken better today using syringe. Dressing to shoulder clean and intact. Cream applied to sore legs and tubi grips. Repositioning regularly. TLC given . . .

Day 2 – night: Continues to be agitated. TLC given. Fluids taken well, please monitor output.

Day 3 – 13.40: Remains very frail. Refused drink and meals. GP visit booked re agitation. Visited by wife. Dressing to left shoulder changed. Skin condition very poor – cream applied and tubi grips to his left. TLC given.

Day 3 – night: All care needs attended to. Very frail. TLC given. Agitated and restless.

Day 4 – 13.40: Poorly, extremely frail. Very thin, writhing in pain. Taking very little fluids or food. Toileted. Repositioned. All care. BP 106/63.

Day 4 – 16.00: General condition very poor. Vomited x3 yellowish secretions – all care given. Washed and changed. Pulse very weak.

Day 4 – 19.15: At 19.15 passed away. Family informed. Doctor contacted – phoned back – is coming shortly.

SNR2R08

In non-LCP notes there were some examples of ‘extended narratives’ which tended to focus on issues or outcomes that either were contrary to those planned or involved important events or some negotiation with outside agencies.

SNB2R10, 09.05 @17.15: [name] from [hospice] visited this afternoon and set up the syringe driver for [patient]. Rang the surgery and asked to speak to [GP] for a prescription for medication to replenish the driver tomorrow. [name] from the surgery said that I could not speak to the doctor as he was running a clinic. I left a message with her for more medications and drug administration charts. I asked her to fax the chart, gave her our fax number but she said that she might not fax the document but keep it in the surgery for [home] to collect in the morning. I explained the importance and she said she will do her best to fax it through.

SNB1R04 – 15.00: At 13.00 he was sitting in an armchair, suddenly heard a big sound and found him lying on the floor. Examined, no injury or bruise noticed. Assisted him to bed, made him in a comfortable position. Informed to IMCA about the fall (left voicemail) because no next of kin. Requested for a GP visit, seen by [GP name] advised to keep him in close observation. NOT For Resuscitation or Not for Hospitalisation. Needs to refer to Palliative Care Team. Made him in a comfortable position.

Reports of communication with relatives

The recording of information relating to what was communicated to relatives and patients in the last days of life was generally limited regardless of the type of documentation used. Patient notes were more likely to capture whether or not communication had occurred rather than the content of that communication or the relatives’ response to it. There were three types of recording in the patient’s notes: (1) a minimalist record asserting that relatives had been informed or were aware of some aspect of the patient’s care; (2) a more fulsome discussion of a more complex situation connected with the patient’s care and its communication to the family; and (3) the reporting of disagreement between staff and relatives and the outcome of that disagreement.

The staff involved in the care of the patient

All of the nursing homes employed a combination of qualified registered general nurses (identified as nurses in the analysis) and health-care assistants (identified as carers in the analysis) to provide care. In the majority of cases health-care assistants, sometimes aided by a qualified nurse, undertook the physical care. Nurses were responsible for monitoring the patient and for administering medication. In all cases but one, a visit by the GP was either observed or mentioned to have taken place during the observational period.

Patient’s level of consciousness

All of the eight patients in the observations were reported by staff to be semiconscious, unconscious or comatose and none of the patients was reported to have capacity to consent to take part in the study. However, while recognising that the patient’s consciousness might fluctuate, and that some of the drugs being administered might interfere with capacity, the researchers observed interactions which suggested that there were was a higher level of consciousness in seven of the eight patients. The issue of consciousness was an interesting and important one.

From a research perspective, it raised issues of informed consent. The researchers had been informed that the patients lacked capacity to understand and discuss their participation in the study and so approval to observe their care had been negotiated with consultees. However, researchers became uncomfortable with this approach where they developed a sense that the patient was neither unconscious nor completely without capacity. In these circumstances the researchers took the opportunity to let the patient know who they were and the purpose of their presence. In all of these cases the patients were asked whether or not they were willing to be involved in the research. The research team discussed this issue of fluctuating or diminishing capacity both as a research challenge and as a research finding.

There were no formal assessments of capacity or consciousness observed. As the patient’s body became weaker as they approached death, different staff appeared to view the patient’s state of consciousness in different ways. Being ‘rouseable’, responding to loud voices and opening eyes were all seen as important indicators of consciousness and were noted as such. However, the existence of these indicators did not appear to fundamentally change how patients were viewed by staff. In the excerpt below signs of consciousness were considered important enough to be reported to the nurse in charge (NIC), although, interestingly, the subtle signs thought to be important by the carer were not recorded in any of the patient’s documentation for that time period.

Day 1;

00:00 Two carers come in to the patient’s room to check whether his pad is dry. After they have turned the patient, one carer says to the other ‘he opened his eyes when we turned him, so I will report this to [name of nurse].’

NNB4P1ON

Consciousness appeared to play a role in decisions about whether or not to undertake some aspects of care. On occasions patients were checked and left on the basis that they were sleeping, while on other occasions patients were checked and provided with care on the basis that they were unconscious. However, patients defined as unconscious were observed to respond to the care being provided with facial gestures, with noises or verbally. This variability associated with defining someone as sleeping or unconscious was reflected in the notes, as will be seen later, where within a short space of time a patient could be reported as deeply asleep, semiconscious or unconscious depending on who was recording it and the circumstances under which this was done.

It seemed that a number of things followed from the presumption that the patient was unconscious or lacked capacity. Rarely were patients accorded any sense of agency. All staff were observed to tell the patient what they were going to do and were doing in terms of the provision of care, but the patient was rarely offered choices or asked questions. This approach was in contrast with that of relatives, who, as will be seen later, despite lacking a clear understanding about the patient’s state of consciousness, were more likely to ask the patient questions and to respond to the answers given.

Defining a patient as semiconscious or unconscious appeared to change the way in which the ‘body’ was viewed. Lacking consciousness appeared to allow some observed behaviours to remain unquestioned and unexplained by both staff and relatives. In the excerpts below, the patient appears very unsettled, but neither the relatives nor the staff respond in any way.

Day 1;

15:09 Patient’s hand comes up and briefly tries to hold the cot side, and then falls back on her bed.

NNA2P2ON

Over the next few minutes these actions are observed a further three times, and then some time later:

16:01 Nurse knocks and comes into the room and looks at patient holding on to the cot side. She moves away from the bed and says ‘Do come down and get me if you need me’ and leaves the room at 16.02

16:23 Patient’s head moves left to right on the pillow briefly.

NNA2P2ON

Although it is clear that the patient’s movements here are observed, neither the staff nor the relatives question the movements or seek to understand them; they are simply witnessed by all around the bedside. While the uncharted nature of dying and death often seemed to leave relatives uncertain of how to interpret what they saw, for staff it was as though the patient’s very status of being unconscious put certain things beyond the need for comprehension.

In another observation the researcher had been informed that the patient was conscious but unable to consent due to dementia. However, on several occasions throughout the observational period this patient appeared to possess capacity.

Day 1;

15:12 Patient has just made us all roar with laughter as he’s just pretended to be dead – patient stopped breathing and stopped blinking with his eyes opened and then when he saw the expressions of alarm, he stopped doing it and grinned. Attending family say that he is always ‘doing mischievous things’.

15:45 He says he doesn’t want to sleep when his relatives ask him

Day 2;

09:12 Suddenly patient opens his eyes and looks over to me [researcher] indicating by lifting his hand to his mouth that he wants a drink. There are no sounds of a nurse or carer so I get up and give him the beaker of juice and watch as he drinks two large swigs.

Day 3;

13:00 The relatives say ‘he can hear us’ they think but is too weak to respond.

NNB6P2ON

Similarly, in SNA1, the patient was described as ‘not conscious’, ‘a bit responsive’, and ‘as though she is in a coma’.

Day 1;

15:20 Relative says ‘There’s nothing going on right now, patient is just lying there. She’s not conscious’.

16:00 Nurse comments that patient ‘is a bit responsive’ when they speak to her

16:45 Relative takes a telephone call. He describes patient’s condition saying ‘She’s not doing anything, her eyes are partially open, it’s as though she is in a coma’.

SNA1P1ON

From the observations it was difficult to assess what staff and relatives meant or understood by someone being unconscious. The relative above continues to talk to the patient and hold her hand, as do the majority of friends and relatives attending the bedside. While it was clear that the patient’s body was failing, there was uncertainty over whether or not the patient could hear or what might be going on in the patient’s mind. Certainly from the activities of the staff the patient’s identity seemed to have been merged into the deteriorating departing body.

Nowhere was this clearer than in the conversations that were had over and within the vicinity of the patient. There were many instances in which, in the presence of the patient, conversations took place involving staff and the patient’s relatives which were about the timing of the patient’s death or the disposal of the patient’s body following the patient’s death:

Day 1;

14:20 (Carer): ‘I don’t think it will be very long, she is very cold to the touch’.

14:40 Nurse to doctor: ‘Her condition has changed twice this morning. I would say she is worse.’

15:00 Relative tells me that that doctor said patient ‘was likely to pass away today, or later on’

15:32 Relative to researcher – Do you know anything about that love?

Researcher – I ask ‘What’s that?’

Relative says ‘Nurse rang yesterday and asked if we wanted her cremated or buried.’

Relative then goes on to discuss his conversation with the undertakers and a non-refundable deposit is mentioned.

16:51 Relative says ‘Doctor says it could be ages, at least she is not dying on her own.’

19:20 Nurse to researcher: She has about another 12 hours.

NA2P2ON

Day 1;

20:50 Carer – ‘She will not pass this night’

21:12 Carer – ‘She is not going to pass away tonight. Normally when patients are close to the end of life, it is 8 or 9 breaths per minute so this is okay’.

SNA1P1ON

In all of the patient notes there were references to the patient’s state of consciousness. The initial assessment recorded on the LCP documentation was always completed, but there was no accompanying evidence to suggest how decisions were made about the states of consciousness that were recorded. Moreover, and notwithstanding that a patient’s condition changed over time, the attribution of different states of consciousness sometimes appeared contradictory.

The documentation completed for SNA1P2ON/CN reported, in different records on the same day, that the patient was semiconscious, conscious, able to communicate and comatose. Similarly, the case-note record for SNA1P1ON/CN records the patient as being comatose and unconscious, and as illustrated in the earlier quote, relatives described the patient as both ‘not conscious’ and ‘as though she is in a coma’. However, a nurse was also heard to report to another carer during the same 90-minute period of the observation that the patient was ‘a bit responsive’. In another site, NNB6P2ON/CN, one set of notes reported a patient as not confused, only for another set of notes at the same time to report the patient as confused.

These reporting differences were interesting in the light of the observations. Although patients’ conditions were observed to fluctuate during the observational period, these were usually minor and on the whole patients declined over time. Clearly any intrusive assessment of consciousness would have been inappropriate for these patients at the end of life. However, it is important to consider whether or not accurately assessing an individual’s consciousness at the end of life is important. This will be discussed further in Chapter 14.

Physical decline and prognostication

Although assessing consciousness did not appear to feature in the nursing home setting, there was an emphasis on monitoring and assessing the physical decline of the patient’s body. This was evident in the monitoring that was conducted by nurses, and to a lesser extent, by health-care assistants. The following extracts reflect the emphasis, particularly by nurses, on recording the pulse rate and the number of breaths the patient took per minute.

Day 6;

20:25 Nurse feels the patient’s right wrist and looks at her watch and her [the patient’s] chest and seems to be counting.

NNB6P1ON

Day 1;

18:00 Nurse: ‘I just want to check your pulse’. She checks patient’s pulse by feeling on the wrist and watching the clock [sic] attached to her top. She comments to the carer ‘It’s 74’

20:50 A new nurse enters the room. She is carrying blood pressure equipment. ‘Hello, I’m (name of nurse). I’m going to take your blood pressure now, OK?’

21:12 Nurse pulls out her mobile phone and stands next to the patient, after a minute she says ‘She is doing thirteen breaths per minute’.

SNA1P1ON

Day 2;

09:50 Another carer comes in and feels patient’s forehead and calls her name very softly. The patient wakes up.

NNA2P1ON

Although the nurses and carers had no influence on the time of death, there was a strong emphasis on monitoring the patient’s deterioration so that the death could more precisely be expected. As can be seen from the previous quotes, staff identified time frames within which the death could be expected. In general, where this was observed to occur the staff were accurate in their predictions. The counting of a patient’s breaths became such a strong feature of nursing visits in the last hours of life that the researcher (with no clinical background at all) also began to record the patient’s breaths.

Although there was a strong emphasis on identifying how close the patient was to death, there remained a focus on care up to the point at which the patient actually died. In some homes the use of hourly charts, which ran alongside the LCP, reflected the intensity with which patients were checked and the importance that was attached to preventing pressure sores. In one case a pressure-relieving mattress had developed a fault which meant that it was not inflating/reinflating as it should have been. The staff decided to swap this mattress with one they knew to be working (SNA1P2ON). This activity involved three staff – two health-care assistants and a registered general nurse – as well as the researcher. The patient died 2 days later.

Provision of physical care

The major focus of all care for the patient was the maintenance of the body’s integrity. Containing leakages, smells and preventing further deterioration of the skin maintained the boundaries of the body. Maintaining the bounded nature of the body was conflated with keeping the patient comfortable, with the provision of care almost always being announced by carers as: ‘We are going to make you comfortable’ (e.g. NNB4P1ON). At night this verbal introduction was also accompanied by the lights being switched on fully:

Day 1;

22:05 There is a knock at the door and two carers come into the room. They switch on the overhead light which startles me [Researcher] but the patient does not react, the resident in the next bed can’t be seen.

NNB6P1ON

This emphasis on ‘keeping the patient comfortable’ and keeping the body bounded translated into three activities:

1. assessing pain, agitation and other symptoms

2. ensuring that the patient was clean, dry and turned

3. maintaining hydration and keeping the mouth moist.

Assessing pain, agitation and other symptoms

Freedom from pain at the end of life was identified by many of the patients in their advance care plan attached to their LCP documents. The assessment of pain and other symptoms took place frequently in the homes, mostly in conjunction with other activities. This assessment involved both observation and questioning, directed to any family member sitting by the bed and on occasions directed to the researcher when the family had gone home.

Given that carers and nurses made frequent visits to the patient’s bedside, ranging from hourly to 3-hourly during the day, there were many opportunities for carers and nurses to observe changes in the patient’s demeanour. In all cases but one, patients also had relatives and or visitors at the bedside for long periods of time during the last days of life. Relatives were often invited to let staff know if the patient’s condition changed or they thought the patient needed something.

Day 1;

18:35 After a period of silence, friend comments that the patient’s breathing is starting to get bubbly. Relative asks if patient has ‘had another shot?’. Friend responds ‘Not since last night.’ Friend reports that staff have told her that they ‘just need to let the staff know if another one is needed and they can give it’ Friend says ‘It’s just starting now’.

SNA1P1ON

The visible symptoms that prompted either questions of others in the patient’s room or the administration of medication by staff were associated with adjectives in the following four domains:

1. bodily discomfort – agitation, twitching, sweating, unsettled, uncomfortable

2. facial expressions – grimacing and frowning

3. noises – moaning and groaning

4. changes in breathing – rattling, bubbly, rasping, wheezy, chesty and rapid.

However, these visible expressions of the body were inconsistently interpreted and responded to. Sometimes they triggered action, while on other occasions they persisted against the backdrop of staff moving in and out of the room and relatives sitting by the bedside.

Perhaps not surprisingly, the acts of turning and washing a patient often elicited a response from the patient, usually in terms of a noise and facial expressions. Whether or not these were deemed to be signs of uncontrolled pain depended on whether or not they persisted after the patient had been turned or washed. Once they had finished providing care, staff were observed to remark that the patient either had settled or remained agitated. Whether or not medication was administered for patients who remained agitated depended on when the last dose had been given and what the NIC felt was appropriate. Carers were heard to remind each other to report the possibility of the patient being in pain to the NIC (e.g. NNB6P2ON, day 2; 10.35).

There were occasions on which patients appeared to be uncomfortable, moving their arms around, shaking their head from side to side or making movements of their legs under the bed covers. As previously mentioned above, it was not always clear why on some occasions these actions were defined as pain related, while on others they went unremarked. Certainly in a number of observed cases, the labelling of the patient as unconscious seemed to have a bearing on this. In one set of circumstances the NIC made the connection of a patient’s agitation with their having a distended stomach. The patient was catheterised and suppositories were given. As the distension was relieved the nurse reported that the patient became more settled (SNA1P2ON).

In the majority of cases where it was established, through either questioning or observation, that there was a symptom that needed relieving, the patient was given medication, with morphine and midazolam being the most common. The route of administration of these drugs was usually via a subcutaneous injection or a continuous subcutaneous infusion via a syringe driver. A small number of patients who were able to swallow were given oral medication. Over the course of the observations the route of administration was tailored to meet the patient’s requirements; tablets were replaced with liquids, subcutaneous injection or continuous subcutaneous infusion via a syringe driver.

However, there were some instances in which a clear request for help was not followed up because, in the eyes of the person who was assessing, the most obvious course of action had already been undertaken.

Day 2;

22:46 Carer and nurse come in. Nurse asks patient how he is and says gently ‘can I check your pad?’ She feels the top of the inside of it with her gloved hand. Patient groans loudly. Carer asks if he has pain. Patient says ‘Yes’ loudly and with feeling . . . Nurse says ‘I can’t give him anything yet’ to carer. She seems at a loss what to do about his pain. We all remain in silence looking at the patient who twitches and scratches his arms and chest. Eventually the carer and nurse leave the room.

NNB6P2ON

Day 1;

18:00 Patient’s mumbling is comprehensible to her son, who responds to her as she mumbles.

Relative told the carer that the patient did not seem settled. Carer said ‘She’s not settled?’ Carer then addresses patient by stating her name as a question ‘[name]?’ I think I heard patient reply ‘Go away’ but I’m not sure. The carer told the relative that ‘my colleague’ had just topped up patient’s morphine, approximately half an hour ago . . . she moved to the side of the patient’s bed and checked the place where the morphine goes into the patient’s arm. She did not appear to find anything wrong. As she was leaving she re-confirmed that the patient had been given morphine quite recently.

SNA1P2ON

All the homes in the study kept drug charts. These provided the formal means by which the administration of medication, dosage, route and timing of delivery could be tracked. In some of the homes staff recorded a narrative which provided the understanding for the administration of prn medication. The routine administration of drugs, for instance, through syringe drivers once in situ was rarely remarked on unless this had to be changed or additional pain relief was required. The variance sheets in the LCP documentation provided support for the administration of pain relief. In NNB6P2ON/CN the nursing notes record on 1 February 2013 that the patient had become very difficult with medicine and personal care, pushing staff away and becoming agitated. In the LCP variance sheet the nurse recorded ‘when personal care was given, was grimacing and became agitated’. The action was recorded as ‘painkillers given 2.5 mls diamorphine’. The outcome was recorded as ‘settled in about 20–30 minutes and went to sleep’.

Ensuring that the patient was clean, dry and turned

All the staff in the homes appeared to balance the benefits of undertaking an activity with the adverse effects of doing so, for example the short-term discomfort from being turned versus the benefit of relieving pressure on the skin. In the last days and hours of life staff were recorded visiting patients hourly to 3-hourly. There were only a small number of observations in which 4 hours had passed between visits. This was more usual at night. Ensuring that the patient was dry seemed to be a priority. If a patient was wet, pads were usually changed. However, if a patient appeared settled and not wet, staff made a decision on the spot as to whether or not they would do anything. If the patient was judged by staff to be asleep then they would most often be left sleeping. However, if the nursing care plan identified 2-hourly turns or repositioning (NNA2P2ON/CN and NNB4P1ON/CN) due to a perceived risk of skin breakdown, this pattern was usually followed. The majority of patients were being nursed on pressure-relieving mattresses and there were three patients (SNA1P2ON, SNA1P1ON and NNA2P3ON) who had pressure sores they had acquired during periods of hospitilisation, and which pre-dated their admission to the home.

In the majority of cases relatives were asked to leave the bed space while the nurses and carers attended to the patient. This sequestration of the patient’s body by nurses and carers for the purposes of care was a major feature of care at the end of life. Relatives were not asked or encouraged to play any part in this and were frequently heard to ask each other, and staff who came into the room subsequently, what care had been given and how they thought the patient was doing. The patient, however, was always given a running commentary on the care that was being provided and what was going to happen next.

Day 1;

16:02 Nurse comes in and asks the relatives to leave and checks patient’s pad by undoing it and looking down. She remarks that it is wet. She asks patient if he has any pain. Patient says ‘No’ Nurse checks his mouth with a torch and tells him he has a touch of thrush. She says ‘The girls are on their way to change you.’

NNB6P2ON

Day 1;

16:00 Nurse tells carer that patient does not appear comfortable on her back as she is making sounds. They should therefore keep her positioned on her side. Nurse tells this to the patient by leaning quite close to her face and speaking loudly ‘I can sense that you are not comfortable on your back. Don’t worry we will put you back’ . . . While carer is putting cream on patient’s heel, Nurse combs the patient’s hair with a brush. She runs the brush through the patient’s hair and then smoothes her hand over her hair.

Day 3;

13:45 Friend tells me that ‘they have been up and turned her’ I say ‘Oh right and what do they think? She says ‘I don’t know because I had to leave the room’. She does not sound particularly happy that she had to leave.

SNA1P1ON

Most patients were washed at the start of the day and additionally if they had been incontinent of urine or faeces. Although it was common for patients to experience some level of discomfort during the process of being washed and turned, this was rarely anticipated by administering pain relief beforehand. In the following example the carers recognised the need to tailor their care to the patient’s needs but were unable to respond to the patient’s pain.

Day 2;

08:23 Two carers come into the room. One of them tells the patient nicely they are going to give him a wash. She asks him if he has pain and he replies ‘I am hurting all over’. She washes his genitals very gently with a flannel and a bowl of water and then applies some sudocrem [Actavis UK Ltd, Devon, UK]. Carer A leads the washing routine and takes her time before turning him saying ‘Let’s give you time to get your breath back’. Patient is saying aargh aargh, but not too loudly. He asks for 5 minutes and says ‘I’m wet’.

08:35 Carer A says ‘We are going to leave you and come back later’. Patient says ‘Yes. Thanks’. She re-does the pad up and says to other carer ‘We will wait until he has some pain relief.’

08:40 Carer A looks at his LCP chart and says I am going to ask the Nurse about the painkillers.’

At 9:45 another carer enters the room to try and feed the patient with porridge which the patient finds too hot. The carer who has gone off to find out whether the patient can be given any analgesia ascertains that the patient was given something at 5am earlier that morning which means that according to the prescribing regime it is too early for the patient to be given any more. The patient finally gets diamorphine at 12 noon according to the notes.

NNB6P2ON

There was a sense that the work of care dominated activity at the end of life in the nursing home setting. Pain created through the provision of care, although brief, was not uncommon, and this raised questions about priorities at the end of life. That said, the adjectives used by the researchers in their observations of care were overwhelmingly positive; they included gentle, sensitive, kind, quiet, caring and soft.

The recording in the notes of the cleaning and turning of patients was largely restricted in the LCP documentation to whether or not it had happened. Where it had not taken place a rationale was supplied, but there was rarely much commentary in the notes regarding this aspect of care.

Maintaining hydration and keeping the mouth moist

All visits by the care staff involved an assessment of the need for mouth care and usually the cleaning of the mouth accompanied by the application of some sort of lubricant to the lips. Attempts at providing fluids varied, as did opinions on whether or not a patient was able to swallow. There were no recorded swallow assessments in the notes, which helps to explain why the method of providing fluids and the consistency of fluids being tried varied according to who was administering the fluids. Staff used specially designed sipping cups, teaspoons and sponges, and were observed using yoghurt, ice cream and thickened fluids such as tea, coffee, juice and Ribena as well as plain water to encourage hydration.

Views differed on whether or not a patient at any point in time could suck or swallow. In one patient’s case carers were heard to offer a patient some sips of water while being uncertain if the patient could actually swallow.

Day 1;

18:45 Two carers have come into the room to change the patient and clean the patient’s mouth. Carer 1 gives the patient mouthcare telling the patient that she is doing so. She asks ‘Would you like some sips of water?’ and I think the patient says ‘Thank you’. Carer 1 tells carer 2 she doesn’t think the patient can swallow, so she will put some water on a sponge so ‘she can suck’. She does this and comments to Carer 2 ‘See she is sucking’.

Day 2;

12:55 The nurse in charge reports that the patient should not be fed ‘otherwise she might choke’. A point that she repeats at 15.15. This particular patient was reported in the LCP progress notes to have taken a few sips of water at 12 noon on the 06.12.12 because she had said she was thirsty. By day 2 (07.12.12) the notes reflect the nurses’ direction that food be omitted, the reason given that the ‘patient appears to be semi-conscious, unable to swallow therefore oral intake omitted’.

SNA1P2ON

Mouth care was the only physical care task that was occasionally undertaken by relatives, often of their own volition.

Day 1;

18:55 Patient opens her eyes a bit. Relative gets a pink mouth sponge and dips it into the pink liquid on the bedside table and gently rubs it around the patient’s lips.

NNA2P3ON

Day 2;

14:50 After a while friend stands up and walks round to the side of the bed. She comments on how the patient’s lips ‘appear to be very dry’. Relative stands up and moves to the dresser where the mouth care sponges are . . . friend runs the sponge over the patient’s lips a couple of times.

SNA1P1ON

Mouth care, although considered to be important by both staff and relatives, was not always welcomed by the patient. In the following examples patients actively attempted to resist this activity.

Day 2;

13:28 Relative tries to clean patient’s mouth with a glycerine mouth sponge saying that she did this yesterday . . . the patient shouts at her and pushes her away. This upsets her and other relative present who both cry a bit.

NNB6P2ON

Day 1;

22:00 The older carer cleans patient’s mouth with a pink sponge and some pink cleaning fluid. The patient is very resistant, but carer is firm and gets the sponge into her mouth.

NNA2P1ON

Day 1;

19:55 The two carers I observed earlier enter the room. One says ‘knock knock’ and explains that they are going to do some mouth care before they finish their shift. They both say ‘Hello (patient’s name). Can I just clean your mouth? Can you open your mouth for me, please?’ The friend still present in the room says ‘She just closed it, didn’t she!’ and laughs. The carer says ‘She knows what we want to do!’ One of the carers spends a long time cleaning the patient’s mouth. She says ‘This is nice and clean now, see’. She comments that there is some ‘dry skin inside her mouth, on her tongue’. The carer uses two sponges and asks the other carer to get her two more because there is still some dead skin on her tongue. When they are finished, the carers say ‘Thank you (name of patient) and then says ‘See you later’. As they are leaving they ask friend if she would like the door open or closed.

SNA1P1ON

The desire to keep the patient’s mouth moist and clean was explained in terms of reducing the discomfort created by the patient breathing through their mouth. Relatives and staff alike also appeared keen wherever possible to enable the patient to eat and drink for as long as possible. Relatives were often more successful than staff in achieving this.

Day 1;

15:51 One of the relatives tries to get patient to drink the beaker of coffee; patient refuses it, grimaces but then drinks it as relative tells him that he needs to drink plenty of fluids.

NNB6P2ON

However, there was little observed discussion between staff and relatives about the strategy with regard to eating and drinking. Providing food and fluid in the absence of a swallow reflex had the potential to cause aspiration with its attendant consequences. Instances were observed where carers and relatives tried to encourage food and fluids but gave up as they could not be certain that the patient could swallow. In the following example there is a clear uncertainty about the plan for this patient with regard to food and fluids and it is the relative who, as the patient’s condition declines, restricts the administration of food and fluids.

Day 2;

22:13 Carer gently cleans patient’s mouth with oral hygiene sticks and removes a fair amount of food (?) yoghurt from her mouth – it seems to have been pocketed inside her cheeks. One carer says to the other – I don’t know how they can say that she is eating and drinking.

Day 3;

13:00 Relative is giving patient a yoghurt drink which she seems to be swallowing most of.

13:30 Relative finishes giving the patient her drink and sits back from the edge of the bed and sits in the chair. She strokes the patient’s hand. She asks patient if she would like some music on ‘or are you not bothered?’ After a while the patient replies in an audible whisper ‘No, I am not bothered.’

Day 7;

09:42 Carer tries patient with a mouthful of porridge, but she is unable to swallow it and so he has to scoop it out of her mouth, he gives her mouth a gentle clean with a swab. He tries her with a sip of tea, no joy, so he wipes her mouth clean.

09:50 Relative arrives

09:55 Carer pops in again with something, food I think but I can’t see. Relative says firmly ‘No, she can’t eat that’.

15:33 Carer comes in with a yoghurt and a beaker of tea and says to relative ‘Shall I give her some of this?’ relative says ‘No thanks. She won’t be able to take it. Just mouth care from now on’. Carer looks disappointed, nods and leaves.

17:10 Same carer knocks at the door. ‘Do you want anything for patient? Some soup or some ice cream?’ ‘No thanks, she can’t eat or drink now’. Carer stands by the screen and looks questioning. Relative adds ‘People don’t want to eat and drink at the end of life’. Carer says ‘okay’ unsurely and leaves.

NNB6P1ON

Interestingly, despite the strong line taken by this relative on eating and drinking there is no direction to care staff in any of the patient’s notes that the frequent assessment of this person’s ability or desire to take food and fluids should be replaced by mouth care or by keeping the tongue and lips moist.

The lack of a consistent line with regard to providing food and fluids was observed in a number of sites. It was not clear what staff were aiming to achieve through the provision of certain foods, for instance porridge and pureed cauliflower, or whether or not a patient’s reluctance to consume these foods had as much to do with taste as ability. Clearly, as the patient approached death the likelihood of their being able to eat reduced. Given that the staff had identified the majority of the patients observed in this study as either semiconscious or unconscious it was surprising to see the focus on feeding patients.

The notes accompanying these observations did not provide any basis for understanding the goal of feeding or providing fluids. There were contradictory references to the patient’s ability to swallow in the same time period and, as previously stated, there were no formal swallow reflex assessments recorded. The detail of foods or fluids, their type, consistency, quantity and frequency to be encouraged were also absent. On the other hand mouth care was usually identified as hourly or 2-hourly. Sips of water were always recorded in the notes. The lack of a shared understanding, between staff and between staff and relatives, about this aspect of care may be a reflection of the lack of a clearly articulated strategy in the patient’s notes.

The provision of physical care, as outlined above, continued in some shape or form up to the point of death. In one case, SNA1P2ON, the patient was washed at 21.25 having had her bowels open. The carers commented to each other that during this they believed her to have smiled. At 22.27 (about 1 hour later) the patient died. In another case, NNB41ON, the patient was turned at 02.00 and mouth care was undertaken and the patient died about 2 hours later at 04.05. One patient (NNA2P3ON) died within 5 minutes of care being provided.

Interaction with relatives

In the nursing homes the patterning of the physical care outlined above provided the framework within which the dying patient was managed. It also appeared to provide the social framework for the experience of dying by the patient’s relatives. Friends and relatives occupied an ambiguous role in the care provided in the homes. They were both at the centre and the periphery of care. Excluded from the hands-on provision of care they were often left as observers at the patient’s bedside. However, their presence was seen as central, providing a context for the patient’s identity, reconnecting the patient to their past and warding off the potential loneliness of dying.

Staff introduced themselves to relatives and always appeared polite and friendly. Staff were also observed to say goodbye to the patient and any remaining family members at the end of their shift. Relatives were provided with chairs, offered drinks and, in many homes, food (sandwiches, biscuits and cakes). Much like the patients, the relatives were well catered for in terms of their physical needs. Relatives appreciated this and were heard to remark on it to each other.

Communication with relatives varied with the members of staff involved. Staff entered the bed space to undertake physical care or offer the relatives refreshments, but were rarely seen to spend time with the patient and or relatives solely for the purposes of communication. ‘Come and get me if you need me’ (NNA2P2ON) was a familiar refrain from staff as they left the bed space.

As previously highlighted, the sequestration of care of the body by staff was reinforced by the exclusion of the relatives from the bed space while this was taking place. Although there are a number of reasons why relatives may be asked to leave, the relatives were never given explanations. Once care had been given, staff made little attempt to report to relatives what had been done or to tell them anything about how they found the patient’s body. In one case where a patient was turned and then turned back because it had become clear that she was struggling to breathe on her back, the lack of communication between the nurse, the carer and the relative led the relative to ask the researcher what had taken place.

Day 1;

16:00 ‘What did they do? Turn her?’ and I say ‘Yes, they turned her, and washed her mouth, and brushed her hair’ relative looks at the patient and touches her shoulder as he sits down.

SNA1P1ON

Interactions with relatives were rarely recorded in any of the LCP documentation. All homes identified that discussions had taken place with relatives about the patient’s deteriorating condition and that an understanding had been reached that the patient was imminently dying. Where the spiritual needs of the patient involved the administration of last rites or the visit of a local vicar, this was observed to be reported to relatives as well as recorded in the nursing notes.

The role of the family

In contrast with the activity of nurses and carers around the bedside, relatives generally sat around the patient’s bed. Being present in this way took many forms and varied over time. There was only one observation in which a relative (a daughter), who was a nurse, took part in providing some aspects of physical care. This particular daughter acted as a co-ordinator as well as a provider of her mother’s care while she was present. She was observed to initiate and provide care, and in relation to food and fluid administration she decided the point at which these activities should cease. The care she provided appeared to draw on her nursing knowledge and skills as well as her knowledge of her mother.

Day 2;

13:05 Relative puts on the patient’s glasses which are in the bedside drawers. She cleans them and puts them on the patient and says ‘Is that better?’ Patient says clearly and audibly ‘No’ Relative says ‘Let’s keep them on and see if you get used to them’.

NNB6P1ON

Although other relatives were less involved in the provision of physical care, helping patients to drink and moistening the mouth and lips were the activities most frequently undertaken. Monitoring the patient’s condition from a lay perspective was actively encouraged by staff, who would frequently ask the relatives how they found the patient that day.

Day 2;

14:50 Friend comments that the patient’s breathing is starting to sound bubbly again. She moves her head to the patient’s chest and listens.

SNA1P1ON

Some relatives appeared to feel empowered and almost at home in the patient’s room. These relatives also appeared more assertive and familiar with the nursing home staff and routines. Other relatives appeared less certain and more passive. While one set of relatives felt able to adjust the cot sides on the patient’s bed to get physically closer to the patient, another set of relatives worked awkwardly around the physical barrier presented.

Day 1;

18:40 The cot side is down where the relative is sitting which means she can lean over and get closer without being obstructed by the cot sides.

Day 7;

09:51 Relative puts the cot sides down . . . She pulls her armchair up close to patient and holds her hand talking to her, telling her information about people they know.

NNB6P1ON

Day 1;

21:20 Both sisters talk over the patient and each holds her hands. I am struck again by the thought that the cot sides act as a barrier to them getting close to the patient.

NNA2P3ON

In another case, relatives were unable to carry out mouth care for their relative because they were uncertain about how to resolve the lack of oral fluid.

Day 1;

14:55 The family are wondering about his dry mouth. The relative takes the oral hygiene pack and looks at it. There are mouth sponges but no liquid so they don’t know what to do. One of the relatives says ‘I don’t know who to ask’ The other relative says ‘We will ask the nurse when they come.’

NNB6P2ON

The watchful waiting that was observed to characterise the presence of all relatives was interspersed with physical contact and communication. There were no gender differences observed; sons and husbands touched, hugged and kissed the patient in the bed as often as did daughters and wives.

Day 1;

18:15 (Male) relative has been standing at the bedside the whole time, looking at and talking to the patient. Relative held her hand for a while; he commented that this was the warmest she’d been for a while. He repeated ‘Don’t worry, I’m here. I’m here still.’

Day 2;

14:05 Relative says ‘Hello mum, only me, I’m back, you can’t get rid of me that easily’. Relative tells patient about his evening dinner plans and then says ‘Shame you can’t be there!’ Patient shrugs her shoulders and lets out a breath as she relaxes them.

14:20 Her breathing begins to get more irregular and relative says ‘Hello, hello mum.’ Relative continues to stand at the bedside and patient’s breathing becomes regular again. Every now and then relative rubs the top of patient’s right arm with his left hand.

SNA1P2ON

Watching the patient in the bed was punctuated by everyday activities such as reading newspapers and magazines, texting and chatting with other visitors. Occasionally, one of the relatives at the bedside dropped off to sleep. In one instance the breaths of the sleeping relative overlapped with the audible breaths of the patient.

Day 1;

16:45 Relative sits back down and continues to read the magazine. He taps along to the music on the radio. Relative puts down the magazine and looks at his mobile phone . . . Relative texts on his phone, every now and then glancing at the patient.

SNA1P1ON

There were few conversations around the bedside that reconnected the body of the patient with the life that had been lived. Occasionally, visitors appeared. The stories that were told in these circumstances located the visitor’s position in the patient’s life, projecting into the room a character and identity no longer recognisable as the patient in the bed. On these occasions the body in the bed was brought to life.

Day 3;

11:40 The dinner lady walks into the room and says ‘I’ve bought a visitor’. An elderly lady walks in. Conversation continues between the friend who is at the bedside and the new visitor. The new visitor says she’s one of those people who is a good friend of the patient and who could not cope with her deteriorating condition.

They share stories about the patient and the new visitor says ‘I’ve been a terrible friend over the past few years’.

12:00 The new visitor says to patient ‘I must love you and leave you. Thank you for everything you have done over the years.’ She strokes the patient’s hair. She stands up and kisses the patient. When she turns to leave I see she is crying.

SNA1P1ON

Day 2;

16:17 A woman wearing a winter coat enters the room. She tells me that she is the patient’s neighbour. After a moment neighbour stands up and walks to patient’s bedside. She says ‘name of patient’. Patient opens her eyes and appears to look at her. She says ‘It’s (neighbour’s name) . . . she pulls up the comfy chair so it is angled away from me and towards the patient’s face. After several minutes she turns to me and asks me if she has ‘been breathing like this all the time you’ve been here?’ A number of other questions follow about whether the lighting is left on and whether the patient is more comfortable today than yesterday.

16:40 The neighbour continues to sit in the chair watching the patient . . . She told me that she wouldn’t be staying long, she just wanted to sit with patient for a bit. She said ‘There’s no point talking to (name of patient) because she can’t hear, although that’s one of things we don’t know, how much they take in . . . She says ‘But I got a smile . . . so.’

SNA1P2ON

Although it was rare for patients to receive visits from people other than relatives or religious leaders, it was the case that some of the patients had friendships with other residents in the home. However, staff were rarely observed facilitating visits from other residents, with one exception that was notable because the visit was facilitated only once the patient had died.

Day 8;

12:58 Following patient’s death the nurse asks the relative if the patient’s best friend in the home can come in and say her goodbyes. Relative nods.

13:03 I follow Sister to patient’s best friend room. Sister sits best friend down in a chair and explains gently but clearly that patient has just ‘passed away.’ Best friend is really upset and cries and says ‘Can I see her? She’s my best friend’. We go to patient’s room and best friend says her goodbyes tearfully to patient’s body.

SNA1P1ON

The failure to support social relationships where they existed between the dying patient and other residents in the home reflects the common concern of the nursing home staff with the patient’s body rather than with ideas of the patient’s personhood.

The notes were largely silent on the role of the family in care at the end of life. Occasionally the notes referred to the presence of a relative or relatives at the bedside, but more often than not the relatives were absent from the records. They were identified in the notes in terms of next of kin details and whether or not they wanted to be contacted during the night should the patient die, and reappeared in the notes following the patient’s death.

Managing the environment

Although the physical care of the patient received considerable attention from staff throughout the observations, much less attention was paid to the management of the physical environment. In particular, temperature control and noise control were inconsistently managed.

Although carers and relatives were heard to pass comment on the temperature of the patient’s room, most frequently that it was too hot, this rarely resulted in any action. Only once was a window observed to be opened, and once a fan was switched on. However, staff did respond when the environment appeared to have a more direct bearing on the patient. On one occasion, during an observation, the researcher noticed the sun streaming in through a window directly onto the patient. Subsequently, while the researcher was out of the room, action was taken by one of the carers to remedy this.

Day 2;

14:55 When I walk in carer is pouring some water from the jug on the tray stand into a glass. He puts the glass down on the tray. I see the curtain is drawn and he says ‘I’ve just pulled the curtain because the sun was shining’ and he gestures on to his face.

SNA1P2ON

Adjusting the environment to suit the patient at the end of life is difficult because it is not easy to discern what the patient may like. Relatives and staff reported that at the very least they thought that patients were able to hear. As a result, attempts were made to talk to the patient. The use of background music, the radio and the television were in evident in all homes. However, this was not always regulated or tailored to the individual in the bed.

Day 3;

24:11 Carer comes in with a CD player and starts a CD of an Irish male singer, but as he’s singing about dying she switches it off, a bit embarrassed and says ‘I’m not sure it is appropriate’ and leaves.

Day 4; Irish male singer CD playing in the background.

Day 7;

10:55 Relative gets up from chair and puts on music – She shows me the holder it is Handel’s water music. No obvious reaction from patient. Relative sits and strokes patients [sic] hand and forearm.

14.40 Pan pipes are playing in the background.

NNB6P2ON

Day 1;

15:00 Friend reports that the reception is quite poor and that she had had to tune the radio into a number of stations before she found one that would stay on without fading . . . Relative says he will keep radio on. The friend said something like; she liked to dance. The relative now remains on his own sitting with the radio on his lap, changing the station regularly until he finds a song, listens to it and then when it ends switches to another station. After doing this twice he settles on one station.

18:35 When friend arrived, straight after the relative left, friend switched the radio off. Friend reported that it was competing with the TV that was on in another room.

Day 2;

14:50 The radio is on, tuned into a radio station playing what I would describe as light hits from the past. The song currently playing is ‘Ground control to Major Tom’.

SNA1P1ON

Staff often reported, as they switched a radio on, that they liked to have music on in the background to foster a sense for patients that they were not alone. However, as the researchers sat observing, they recognised the incongruence of what was frequently being broadcast; for example, Karen Carpenter singing ‘Top of the World’ might have been difficult for patients to relate to. The unremitting nature of the radio textured the physical space like ever-changing wallpaper, at times soft and inviting and at other times loud and intrusive.

As identified at the beginning of this section, only a few patients were being nursed in a single-occupancy room. Inevitably, the sharing of a room with a person who was dying required some management and this usually took the form of taking the other resident out of the room during the day into some communal space elsewhere. There were times during which the resident sharing the room had visitors who congregated at their bedside.

Day 1; Patient in next bed – relatives of other patient gathered around bed.

16:02 Patient A in next bed clears her throat behind the curtain.

18:05 Patient A is now not in the room and her curtain is drawn back.

19:05 Patient A goes back to bed.

NNA2P2ON

Day 1;

19:10 The other family are talking to patient in next bed in very loud voices drowning out any conversation between patient’s family.

NNA2P3ON

The management of noise elsewhere within the home was hugely variable. The noise of the television emanating from other residents’ rooms, the sounds of patients being escorted from one location to another, and the everyday domestic noises of cleaning and pots and pans, cups and saucers all penetrated the rooms in which people were dying. It is not possible to say what impact, if any, these sounds had on patients; patients rarely appeared to react to them directly. There was only one instance in which the researcher wondered if a patient was affected by noises emanating from outside the room.

Day 2;

13:00 Patient moves left arm out from the blanket. She moves it around for a very short period and then rests it back down. She has begun to make short grunting noises irregularly, or perhaps they are better described as short moans. She seems to react to sound and there are a variety of sounds coming from outside the room. The door is propped open and I can hear a couple of people talking, things being moved about etc. At 13.10 the call bell begins to sound in the Nursing Home. Almost immediately the patient groans and begins to move both of her arms although she does not lift them very high off the bed. This stops after a few moments although the call bell continues to sound and I wonder if the two events . . . are related?

SNA1P2ON

It was interesting that one of the cleaning staff, in acknowledgement of the patient’s imminent death, used a dustpan and brush rather than the vacuum cleaner to clean the patient’s room.

Day 7; The cleaner is using a brush and dustpan on the whole of the floor. She says ‘It’s not fair to use a hoover.’ The relative sits on the edge of the bed so that the cleaner can do under the armchair.

NNB6P2ON

Although the nursing gaze was largely focused on the body, the caring staff passed through the space inhabited by the body, possibly unable to feel the heat and hear the noise that had become the patient’s world.

Impact of the observations on patient care

As observations continued it became clear, from remarks made following a patient’s death by a few relatives, that the presence of the researcher had been reassuring. Relatives suggested that the care their relative had received had been of a higher standard than they might otherwise have received had the researcher not been present.

Day 4;

01:40 As relative is leaving, he says ‘I think she received better care because you were here’. I say ‘Really?’ He says Well, I don’t want to go into it now, but yes.’

SNA1P1ON

Relatives’ perception of patient’s deteriorating health prior to admission to the nursing home

The primary purpose of these interviews was to understand relatives’ perspectives of care in the final hours and days of the patient’s life in the nursing home. For the relatives in this study, however, their perspective on care at the end of life was heavily informed by the much bigger story of the patient’s deteriorating health, which began some time prior to admission to the nursing home. These narratives were important to the relatives and considerable effort was expended in the interviews ensuring that the correct chronology was recounted. It became clear, given the universality of the presence of these narratives, that they provided an important context for portraying the circumstances in which the patient was admitted to a nursing home. A central feature of these narratives was the physical and emotional labour expended by relatives to support the patient to live independently. The physical decline of the patient was mapped by doctors’ appointments, emergency admissions and clinical investigations (NNB6P1F1, SNA1P1F1, NNA2P2F1 and NNA2P3F1). Admission to a nursing home heralded a critical point in an individual’s trajectory of decline and failed attempts to keep the patient living independently. However, once the patient had been admitted to the nursing home, the narrative continued to chart repeated infections, failure of wounds to heal effectively, development of pressure sores, renal failure and the progression of dementia. So, in some sense, the admission to a nursing home was simply part of the physical decline associated with old age. What happened after admission was a continuation of the decline that had been in evidence prior to admission. In another sense, admission to the nursing home became the point at which relatives were forced to acknowledge that the strategies that had been put in place in the community were no longer working and that more radical solutions were required to protect the patient.

We sort of looked after her for . . . well as long as you can do in a house . . . we didn’t want her in a home . . . we ended up having to get carers in . . . me still going and the wife . . . four carers a day . . . getting near the end of her being at home . . . she was becoming a danger to herself then.

NNA2P2F1

The physical and emotional demands of caring for these elderly and increasingly infirm patients were present in all of the accounts. The bereaved relatives interviewed in this study were not young (the majority were over 55 years of age) and often held other caring responsibilities. Nevertheless, the admission of a patient to a nursing home was undertaken reluctantly and, in some cases, was associated with a huge sense of guilt and desperation.

I couldn’t cope any more looking after her . . . I couldn’t manage me mum and [daughter’s] baby as well and . . . me sister who probably would have helped, her husband died . . . She [mother] said ‘you must really hate me for putting me in here’ . . . I mean I didn’t want it to happen but I just couldn’t manage any more.

NNA2P3F1

At the point of a patient’s admission to the nursing home, few relatives felt that there were any options other than admission. Medical professionals were reported to be highly persuasive, largely, it would seem, as a result of the way in which they framed the risks that the patient posed both to themselves and to others. ‘A catastrophe waiting to happen’ was how one relative reported a doctor’s view on his mother (NNA2P2F1).

Three patients had lived in the nursing homes between almost 5 months and 2 years prior to their death (see Table 6). Four patients were ‘fast-tracked’ under the continuing care scheme, three from the hospital setting to the nursing home and one from home to the nursing home (NNA2P2F1, NNA2P3F1, SNA1P2F1 and SNA1P1F1). These patients had lived in the home for only a very short period of time before they died.

In the relatives’ accounts of the patient’s death, the age and state of health of the patient were of central importance. All but two of the patients observed in this study (whose relative was interviewed) were over the age of 90 years when they died. The reality of death was a feature of discussions between the patient and their relatives as their lives became underscored by the burden of physical decrepitude. Relatives reported discussions instigated by patients about the manner in which they would like to die and the arrangements for the disposal of their body.

But she told me in the last sort of 3 or 4 months, whatever happens, you must tell people that I’ve had enough now. I’ve had everything in my life, I’ve had a good life, I’ve had extra years as well and I’m not to be resuscitated.

SNA1P2F1

The relatives’ accounts of the progressive deterioration and decline of the patient, while sad and distressing, were tinged with the expectations of a life coming to an end. In short, the death of these patients was expected, sometimes welcomed, often planned for and in the main accepted to be timely. It is this lens through which the care in the nursing homes was viewed and should be understood.

Without exception the nursing homes in the study were described as welcoming and friendly. Staff greeted the relatives, were reported to be courteous and sensitive, offered the relatives unlimited access to the patient and supported visits with refreshments.

. . . the nurse said anything you want . . . they were bringing tea and biscuits all the time and stuff like that they were very good . . . I don’t think nothing [sic] was too much trouble for them.

NNA2P2F1

. . . it was quite a relaxed place, I think we felt no pressure when anything like, it felt like we could just do what we wanted to. I mean we could come in and we could go, we could come and go when we wanted to and nobody was really bothered about that and uh in that sense you’re not being sort of shooed out because of visiting hours or anything like that, you just felt that it was very laid back in that sense yeah, so that was quite nice.

SNA1P1F1

The relatives’ perspectives on care in the final hours and days of the patient’s life in the nursing home fall into four themes:

• a comfortable and peaceful death

• interactions between staff, patients and visitors

• being ‘there’ for the patient

• the LCP.

A comfortable and peaceful death

Contrary to the view expounded by poet Dylan Thomas, that death in old age should be resisted and fought, the relatives interviewed in this study hoped for a comfortable and peaceful death in which the patient slipped away quietly. The absence of pain and suffering were central components of the deaths they wanted for the patient.

There didn’t seem to be any need for pain relief or anything because she didn’t seem to be in any pain. She was just asleep . . . Basically the important thing was that she died peacefully and comfortably.

SNA1P1F1

Making her comfortable so she wasn’t too agitated which they seemed to do quite well I have to admit. They knew when she hurt and things like that, it was keeping her comfortable . . . I just wanted to say ‘I want you to do anything you can to make mum peaceful and happy and not have to suffer, she’s suffered enough now’.

SNA1P2F1

She wasn’t in any pain at all. So I think, I think they were basically on top of everything. It wasn’t pain. I don’t know what will have caused that agitation. I don’t know what will have caused it. It was err . . . restlessness if you like.

NNA2P2F1

It is important to understand that the deaths in this study were viewed in the context of old age, increasing frailty and hands-on experience of attempts at staving off progressive dependence. Relatives recognised the inevitability of death and focused on the quality of care provided to the patient up to the point of death. The quality of care was measured in terms of the minimisation of physical symptoms (pain, agitation and distress, breathlessness) and the maximisation of comfort care (turning, washing, dressing wounds, providing food and fluids and administering mouth care). These were features of all the relatives’ accounts of care at the end of life for the patient. In the main, relatives felt that this was achieved.

There wasn’t much more they could have done for her, I don’t think . . . it was peaceful, it was in nice surroundings and everybody was nice to her and you couldn’t really ask for more . . . She was given her dignity, they did look after her . . . because some of the awful things I’ve seen previously, because they’re so rushed and stressed they just chuck you in a room and that’s it they forget about you almost, you’re a bit of a nuisance but I never felt mum was like that. I felt they were caring for her, they were actually coming up and checking her and making sure she was all right, trying to talk to her ‘is there anything you want’ . . . which I think was nice.

SNA1P2F1

There was nothing much else they could do, they did everything. They bent over backwards for anything I wanted and all I had to do was phone up and they did the rest.

NNA2P1F1

Everything was just superb.

NNA2P2F1

However, there was one exception to this. One of the relatives, who had spent some time at his mother’s bedside, did not feel that his mother had received enough attention. He reported that from conversations with other family members they felt that the care had been better when the researcher was observing than when she was absent. When asked what in particular had been different, the son reported the regularity and frequency of ‘checks’.

Quite frankly apart from the fact that somebody goes round every now and then, she could have died and nobody been there possibly . . . I mean that might have been fine if she’d died peacefully but she might have been in distress and nobody would have known possibly.

SNA1P1F1

Getting the balance right may be easier to achieve where a dialogue exists between the relatives and the staff. The previous relative’s view is contrasted with that of another relative who felt that, on balance, the staff (in a different site) did too much checking and turning.

I blamed them for wanting to, sort of, go in (to her room), because it’s a very difficult situation. You’ve got somebody that’s really comfortable and you don’t want to be disturbing them too much but the other side to that is you also don’t want to end up with pressure sores if somebody’s going to survive a lot longer.

NNB6P1F1

Interactions between staff, patients and visitors

Relatives viewed staff as intermittent visitors to the patient’s bed space. They interrupted the long vigils adopted by the majority of relatives. Staff were reported to ‘pop’ in, checking the body, offering the relatives refreshments and encouraging relatives to seek them out if they needed anything.

When I was there they kept popping in and out sort of ‘you OK?’ ‘just let us know if you need anything’. They were pretty good like that . . . they just said ‘if you need anything just give us a . . . you know where we are’ and then sometimes they’d pop in anyway just to check her and check her blood or check everything you know check her pulse or something . . . she had pressure relieving mattress and things like that.

SNA1P2F1

From what I could see she was never left on her own. Because do you know I went everyday and (other relative) went at different times. So they didn’t know when I would turn up. But every time I did go . . . there was somebody there.

NNA2P2F1

The relatives all reported being told that if they wanted anything they should find a member of staff or use the room’s buzzer if one was available. One relative reported that they had to go and look for a nurse after they used the buzzer and no one answered.

It was just me and [other relative] at the time . . . what they do they give us a buzzer to buzz ‘em, to sort of see if my mum does go worse, and we buzzed it, but no-one came so [other relative] went looking for the nurse.

NNA2P1F1

Relatives reported that once they had found a nurse, either to seek clarification of an observed behaviour that the patient was exhibiting or to ask the nurse for something, the staff responded quickly. Relatives reported to staff that the patient was in pain and needed medication or needed to be propped up.

I’d have to ask, ‘can you prop her up a bit?’, she seems to sort of, sometimes she slides down the chair and . . . So I’ve had to ask, the nurses did it . . . picked her up and pulled her back again and, you know, but basically they was brilliant, basically you’ve just got to try and catch them ‘cos they’re busy and it was a bit short staffed and they was all looking after everyone else.

NNA2P2F1

In all cases except one the provision of care by staff was undertaken in the relative’s absence. Relatives were reported to be asked to leave the room while care was being delivered. Rarely did relatives know exactly what care was being delivered and rarely did they ask the nurses and carers who were providing it. There seemed to be a tacit acceptance that the body and its physical needs and care was the remit of nurses and carers in the nursing home.

We went out of the room and they did something around her tummy area I think because I saw them pull the sheets back and do something around her tummy area and . . . then when we came back in my mum was peaceful again.

NNA2P2F1

They must have changed the dressing and things like that . . . yeah generally while that kind of stuff was going on we were out of the room . . . and that’s fair enough.

SNA1P1F1

Although relatives accepted that the provision of care was the work of nurses and carers, they reported receiving little information about the care that had been provided in general.

Really there was nobody else there to ask . . . [you] found out by yourself.

NNA2P3F1

Indeed, in their observations the researchers recorded frequently being asked by the relatives about the care that had been provided during the period that the relatives had been out of the room.

While the provision of care was dominated by the nurses and carers, relatives reported their own role as ‘being there’ for the patient and not letting the patient die alone. In one exceptional case a member of staff was reported by relatives to have sat with them, providing comfort and monitoring the patient’s slow decline into death.

[Other relative] went to get the nurse, the nurse came up she . . . stayed until my mum died . . . I don’t like saying it but it took an hour for my mum to die . . . and even the nurse . . . she was crying a bit as well . . . she was just stood at the bed comforting me and [other relative] and just every so often go and taking the pulse . . . she’s not in any pain, don’t worry about the pain, she said it [the breathing] will just go shallower and then all of a sudden it will just stop.

NNA2P1F1

As previously mentioned, one relative, a daughter who was a trained nurse, reported orchestrating the management of the physical and emotional aspects of her mother’s care as well as the hands-on delivery of it. Indeed, her account reflects what the researcher observed. This relative combined the role of daughter and nurse in her narrative.

. . .and it was also good that I was able to do a lot of the care for her and they were helping me almost and I think because again, of my background, I didn’t have to be taught to do it, I was just doing it and it came second nature to me almost . . . at least I felt I could do things and check on her and other things that needed to be done.

NNB6P1F1

This daughter portrayed her role as ‘active’, based on her knowledge, skills and training. For her, using her skill set made her feel useful in an environment in which she could not alter the outcome of certain death. She reported that ‘it took her mind off things, it normalised things’. She contrasted this with other relatives who, lacking this skill set, sat, watched and waited.

I’m sure people that haven’t got the background would feel helpless because you’re almost sitting, waiting and watching . . . Passive relative . . . that’s why I say I just think for anybody else in that situation it would have been maybe a lot harder for them.

NNB6P1F1

Indeed, sitting, watching and waiting was what the majority of other relatives reported doing. But contrary to the daughter’s view that this rendered them helpless, relatives took comfort from the emotional support they believed their presence offered and from their ability to monitor the care that was provided. In general, relatives accepted the environment in which they found themselves and rarely questioned or saw the need to try to change it. They did not expect to be involved in the hands-on care of the patient, but needed to be present for the patient.

‘Being there’ for the patient

Being there for the patient seemed to represent a complex set of thoughts and feelings which related both to the relative and to the patient.

I wanted to be present, I was glad I was, I was glad I was there anyway, ‘cos I missed out on my dad dying, ‘cos I couldn’t get there in time, at least I could hold my mum’s arm, her hand, which I did all the way until she died. Just sort of saying I’m there for her.

NNA2P2F1

The presence of the relative lent a status to the patient as, for example, someone’s mother, father or husband. Bringing the identity to the body was important and in the interviews relatives talked at great length about the person behind the patient. Reconstructing personhood in this way was important because in the last days of life the relatives were often unsure about the patient’s state of consciousness. Most relatives struggled to define or understand the patient’s state of consciousness. One relative (SNA1P1F1) variously described his mother as having been ‘just asleep’, ‘[in] a sleep that went on for a number of days’ and ‘in a black coma’. He also remarked that ‘it didn’t really feel like she was alive in the care home’ and ‘they said she was in a coma [but] I’m curious whether she actually was conscious at all . . . I don’t really know’. Uncertainty over the conscious state of the patient was echoed throughout all the relatives’ accounts. Interestingly, it did not feature as a concern; it existed as a puzzle to which they did not have an answer. This confused and sometimes contradictory sense of a patient’s consciousness was always accompanied by a belief that ‘hearing is the last sense to go’. Relatives reported talking to the patient, playing them their favourite music and reading to them. All of these activities were designed to provide familiarity and comfort to the patient, to recreate a connection with past interests and to reassure.

[T]hrough her dementia she seemed to be more and more into music . . . she would be tapping away . . . we kind of thought . . . if she could hear anything, she’d quite like to hear music . . . if anything was going to make her do something it might be music . . . but . . . there was no feeling of any connection really in the home.

SNA1P1F1

I suppose I was managing to get something through to her. She realised I was there . . . and on Thursday when I came back . . . well, I think she knew I was there, she did sort of say ‘oh hello’. Sort of startled me and said my name but on the Friday when I went through, I thought she’s not mum at all . . . [but] because they say the hearing’s the last thing to go . . . I was talking to her normally, trying to talk to her . . . but she wasn’t really with me.

SNA1P2F1

The provision of food and fluids for patients

The relatives held very mixed views on the appropriateness of providing food and fluids to patients and the timing of this. There was a general agreement, however, that at the very end of life, patients often neither wanted nor were physically able to eat or drink, and to force them to do so was ‘cruel’. For many relatives the acceptance of the patient’s death provided the starting point for understanding this view.

I was asked if I wanted to send her into hospital for subcutaneous fluids by one of the staff nurses. So I sort of said ‘well no, we’ve got an advanced plan in place and I really don’t think that it’s appropriate on two counts: (a) because she’s got C. diff [Clostridium difficile] – I don’t think the hospital’s going to be very impressed at sending somebody in with C. diff and (b) I don’t honestly think it’s going to improve mum’s outcome or ultimately what’s going to happen. I think this is the inevitable and all it would do is to prolong the inevitable for her.

NNB6P1

Another relative reported that as a family (SNA1P1F1) they had already ‘made a conscious decision not to keep feeding, not to keep her alive unnecessarily’.

Three patients were reported by relatives to have been consuming nothing or very little for some considerable period of time. For example, two relatives (SNA1P1F1 and NNA2P1F1) reported that their mothers had not taken any food or drink for the 3 days prior to their deaths, and a wife reported that her husband (NNB6P2F2) had refused food and drink in the last day or so of his life. The relatives in these cases were keen to emphasise that it was not a case of staff ‘saying you’re not going to eat any more’ (SNA1P2F1); it was that the patients had come to a point at which eating and drinking was too much effort or not physically possible.

I think it came to a point where she couldn’t physically do it [feed] herself and she wasn’t accepting stuff and they were giving her the, you know, the supplement drinks and these sorts of things.

SNA1P2F1

Other relatives, however, reported observing staff attempting to get patients to eat and drink up to the point of death. There was one occasion when the participant (who was a qualified nurse) questioned the utility and safety of continuing to try a patient with food and fluids in the face of resistance and a perceived inability to tolerate it, as there seemed to be little or no agreed strategy for feeding.

Bless ‘em they still kept bringing food in and one day they brought something in . . . I wasn’t too sure even what it was . . . was it cauliflower cheese? . . . some of the stuff to be fair they were just winging it . . . there’s no way that’s going down and again I think that was a lack of experience . . . she was struggling to swallow like 6 months ago but she’s certainly not going to swallow that now.

NNB6P1F1

The issue of food and fluids at the end of life is often an emotive one, but in this study relatives adopted a pragmatic approach, principally because they had all decided that the patient’s life was coming to a natural end and they had no desire to prolong the patient’s suffering.

The Liverpool Care Pathway

The inevitability of death and the concomitant desire to promote comfort up to the point of and including the death provides a useful context in which the LCP can be examined. The care of all of the patients whose dying phase was observed in this study was supported, to some extent and in a variety of ways, by the LCP. The relatives’ views on the LCP illustrated a broad spectrum of knowledge, attitudes and beliefs about the LCP in general and, more specifically, its use in the support of their dying relative. Many of the homes in the study used the GSF, although rarely did relatives identify it as such. Where it was used relatives referred to elements of the documentation which are unmistakeably GSF. Most relatives had at least a vague recollection of ACP, including spiritual and religious beliefs and requirements at the end of life, preferred place of death, cardiopulmonary resuscitation and the most important issues for care at the end of life. These conversations also included preferences regarding music and information on key family members, next of kin and important friends.

In a similar vein the majority of relatives had heard of the LCP but could not really describe it. Three relatives were unclear about whether or not the LCP had been used to support the care of their relative.

She mentioned the Liverpool Care Pathway, but to be honest I don’t think we got it.

NNA2P3F1

[T]hey weren’t actually using it as such, were they, on my mum at all? I don’t think they were actually using it on my mum at all.

SNA1P2F1

Knowledge about the pathway was attributed to the media, to discussions with friends and neighbours and to health-care professionals.

I had heard about it in the news and things like that and read about it and it’s been in the paper as well.

SNA1P2F1

Her auntie, she had this pathway thing. And she, well they were very pleased with it.

NNB6P2F1

Relatives also described a range of thoughts and ideas associated with the pathway. These ranged from general care or TLC, to the administration of drugs and stopping someone eating, through to assisted death.

Just formalising what we’ve all known as TLC for so many years . . . when you do feel that somebody’s coming to the end of their life and it’s inevitable . . . that you just make them as comfortable as you can do without causing them any more distress I think.

NNB6P1F1

What’s different about Liverpool Care Pathway to just stopping . . . somebody eating and it would just seem common sense. I can’t really understand what’s odd about it to be honest . . . we wouldn’t have wanted to force feed her . . . we wouldn’t have wanted to have prolonged her life . . . I don’t think she’d have wanted to prolong her life . . . I mean it’s just common sense . . . it seemed normal to me.

SNA1P1F1

I think it’s good myself, other people might not . . . I have heard some people call it assisted death you get all different terms don’t you, but I think it’s good . . . it’s something I’d want if it was for me and I was in the same state as my mother . . . I’m a great believer in things like that anyway . . . whether you want to call it assisted death or pathway or whatever . . . I think if a person is terminal, my mother at that stage was terminal without question or doubt and I know what her feelings would be she didn’t want to die so let her go in her sleep and if that means help her to go to sleep I don’t see a problem with that.

NNA2P2F1

Interestingly, two relatives emphasised that it was just a pathway and not an irreversible route to death. However, one of the relatives, whose mother had had previous experience of the LCP in another home, seemed to suggest that his mother had survived the LCP as though it were a test designed to sort the survivors from those who were going to die.

But they said . . . they can go on it, but . . . they can easily take them off it.

NNA2P3F1

Because when my mum was on the pathway in that [previous care home]. Totally different. She was on like a morphine drip; some medication and morphine drip as well but it was tubes going into her and they explained . . . they said eventually what will happen is basically it will be the morphine that puts her to sleep . . . [but] she kept jumping up and starting hallucinating I thought she’s not going anywhere here making her better if anything . . . I felt this isn’t working . . . anyway she survived that . . . and the doctor took her off it, obviously because she survived it.

NNA2P2F1

These accounts serve to illustrate the complex landscape of knowledge, attitudes and commonly held beliefs that exist and within which end-of-life care resides.

Nursing home philosophy

All of the nursing staff interviewed had a fundamental belief in the benefits of end-of-life care in the nursing home. They reported that their knowledge of the patient and the fact that it was the patient’s home was an important starting point. They reported being passionate about ACP and keeping patients comfortable in the nursing home environment. The nurses talked about making sure that end-of-life care was pain free and comfortable.

Advanced [sic] care plans are not done just as a paper exercise, but in that we fulfil our promise, when we do that advance care plan because if the decision is to keep them here and for them to die here, then we are saying that we will ensure that they have a peaceful death.

NNB6P1S24

Understanding early on what a patient’s wishes would be were they to become unwell was recognised to be important for staff in achieving the best possible care for that patient.

The majority of the people are on an advanced [sic] care plan; not all of them, because it depends on the circumstances of the resident and the families on the advanced [sic] care plan . . . Occasionally we’ll have an advanced [sic] care plan that says I want my relative to be nursed in hospital; it’s rare but it does happen and people change their minds.

NNA2P1S214

Indeed, the religious and spiritual needs of the patients and their particular preferences for music and foods were all recorded, and observed in most cases to be delivered.

In NNB6 it was recognised that for people with dementia finding out what the patient might want could involve talking to a wider group of people.

Most of ours have dementia, so we need to do what we call a ‘best interest advanced [sic] care plan’ and I always like to include relatives or if there is no relatives, a friend who knows the resident before they came in and would know something about the character and personality to say, what do you think she would like, just to make what they are calling a ‘good death’.

NNB6P1S24

Although the staff were keen on advance care plans, in the case of NNA2P3 one of the relatives raised a note of caution. She was reported to have thought these conversations offensive when taking place too soon after a nursing home admission.

She said it’s something I don’t want to talk about now. Yes, you know, force her mum before her death to discuss how people would treat her . . . they don’t like it.

NNA2P3S214

Clearly such conversations need to be conducted sensitively and it is possible that the enthusiasm of the nurses in the sites might blind them to the individual’s fears about death and dying.

The staff were committed to providing a high level of care and believed that nursing homes were the best environment for elderly and frail people to die.

Training of the nursing home staff

The staff providing care to the patients in this study reported having received end-of-life care training. In some of the nursing homes the GSF was used in conjunction with the LCP.

All qualified actually had a day course on the Liverpool Care Pathway at the Hospice and the senior carers also did the same course the beginning of this year, end of last year. In conjunction with that – because of us doing the Gold Standard Framework – we would have our study days and then every 2 weeks, one of the GSF trainers – would come to the home and have an afternoon of training, so that all members of staff knew about the Gold Standard Framework and end-of-life care because it is important, everybody is included, cleaners, the cooks because they didn’t appreciate what their role was, in caring.

NNB6P1S24

In one site (NNA2) where three members of staff were interviewed, all three nurses reported receiving training. The training was repeated because of the sometimes long gaps between patients dying and the need to keep up to date.

All had training from Sister X several times, because what happens now, maybe one patient becomes poorly and after that there is long gap, but it’s happened about two or three times. They changed the set of papers as well so. So we had about 2–3 sessions from Sister X.

NNA2P3S214

Perhaps not surprisingly, there were issues regarding the way in which night nurses got their training. Four staff were working on nights at the time of the study; not all reported receiving training.

. . . it’s difficult to do on nights, I cannot do the training most of the time because if you’re on night you don’t get the skills and not the training so I won’t get to do it.

NNB4P1S24

The patient’s deterioration

All of the nurses interviewed talked about the patient’s deterioration and this formed an important context for the diagnosis of dying and the steps that needed to be taken by the home to provide end-of-life care. Four of the patients had been admitted with serious health issues and with the expectation that their health would decline. However, the short length of residency in the home set against this decline was reported by staff to make the deterioration difficult to judge. Where patients had been resident for 6 months or more staff reported being better able to chart their deterioration.

I knew that slowly, slowly she was deteriorating. We knew when she’d really stopped eating and drinking.

NNA2P3S214

She was mobile when I came with a frame for short distances and she was quite a feisty lady but suffered a lot of bouts of paranoia where she would think we were poisoning her with any medication or any drinks or anything that we gave her, or any food and on those days she would have nothing. Not a drop of anything because she just mistrusted us. I would say that for a lot of this year you could see she was beginning to deteriorate.

NNB6P1S24

Alongside the general reduction of mobility and the loss of continence and desire to eat and drink, level of consciousness was perceived as another key indicator of dying.

I think her consciousness was starting to go then and she wasn’t fully aware any more then really and that’s when I decided, in my mind, she was probably truly dying at that point then.

NNA2P1S214

I washed and changed her at one point and she never opened her eyes once. So there was varying levels of responsiveness from her.

NNB6P1S24

As observed, consciousness was linked to responsiveness rather than to a detailed neurological assessment. The rule of thumb that seemed to be applied to the definition of a patient’s irreversible decline was their state of consciousness, the nature of their breathing and the colour of their skin.

I started like checking her nails and checking the soles of the feet because these are usually the indications if they’re to die in the next few hours if it’s going to happen.

NNA2P1S214

In reporting the patient’s physical deterioration, nurses made a clear distinction between being old and frail and dying.

There is just a lot of this . . . it’s an attitude – they are elderly, they are frail and it’s immediately they are end of life and she isn’t.

NNB6P1S24

Recognition of dying

The distinction between the recognition of dying and the commencement of the LCP is an important one, as the LCP should be used only when this recognition has occurred. The staff realised that recognising dying was not easy, and in reality a number of things confused the situation.

This is where it’s sometimes difficult with some of our residents. It’s deciding when is the right time.

NNA2P1S214

While the nursing care provided in nursing homes was adjusted to the patient’s bodily decline, the recognition of dying and the commencement of the LCP required nursing homes to involve the GP. Considerations around the involvement of the GP appeared to distort the use of the LCP. The majority of staff reported difficulties getting GPs to visit.

The assessment should be with a doctor but I can’t get the GPs to come out and go through the initial assessment with me and what happened with patient – and it’s happened with all the others – is that I have done the initial assessment on my own, and then I think, yes, I will put her on the Liverpool Care Pathway but then I will get a GP within 24 hours, and when they come I say, please read my assessment and what my comments are and sign if you agree; now they don’t ever read, they just sign it, so I could have written anything to be honest.

NNB6P1S24

As a result, many of the nursing home staff reported bringing forward the GP’s involvement, particularly if the patient’s decline coincided with the weekend.

Before the weekend usually Friday, when the patient is poorly we’ll call the doctor, because if it falls on Saturday we can’t do much. So because Saturday the pharmacy is closed and sometimes, GP won’t be there.

NNA2P3S214

The nurse interviewed in NNB6 reported that she was part of a nursing team involving the local hospice that had been active locally trying to raise the understanding of GPs in relation to end-of-life care in nursing homes. In her view the reluctance of GPs to be involved in the LCP stemmed from the requirement that they reassess the patient at regular short intervals. She reported on several occasions throughout her interview that she thought GPs found the reassessment ‘too much to do’ (NNB6P1S24).

There is no GP will come every 3 days, as they should do when somebody is on the LCP, because they should be doing the reassessment with me but there’s nobody and I am on my own – or whatever nurse is on duty.

NNB6P1S24

Behind the scenes nurses reported talking to people with specialist expertise and taking help and advice where they felt uncertain. In some cases this was from other staff in the nursing home. In other cases it was from hospice experts about drug dosages and maintaining a patient free of pain.

I said that she still complains about pain and she looks uncomfortable when attending her and they said ‘OK you can increase the morphine to 7.5 mg on the syringe driver and keep her with a stat dose of 1.5, 1.25 when needed’.

SNA1P2S23

Commencement of the Liverpool Care Pathway

There seemed to be a variation across sites and patients as to when the LCP was signed and started. Unpicking this variation was complicated because staff often talked about the tools they used, the GSF and LCP in particular, as though they were one and the same thing. In discussing the patient’s deterioration one nurse was able to create a distinction between what she called a ‘palliative period’ and the LCP.

If we feel that we are in what I call a palliative period then we will get the doctor in and say, can we have the four core drugs? We think we are getting within the last few weeks and we will have all that set up and established and from that point on, we will then request a palliative care review every 2 weeks and the doctors are quite happy to come out and do that; it’s just once they’re on the LCP.

NNB6P1S24

However, there was little consistency about the point at which the LCP was started. Views ranged from an informal sense that the patient was dying to the application of an algorithm.

I always go through the algorithm very carefully and the algorithm is a good leader in helping you to make that decision. I found that helpful because I am doing them on my own; it is a bit more support for me and I do find having done the algorithm, that it does tend to be right.

NNB6P1S24

They have a criteria of going on the Pathway, it usually means being nursed in bed; which we would have been doing. Sips and fluids still given but it’s at our discretion; medications are stopped, or acting medications are stopped, although occasionally sometimes antibiotics continue for a few days if it’ll help with their comfort. But most medications are stopped and we no longer actively try and feed them. Food is offered if we think that their swallowing is up to it but we no longer actively try and feed somebody. It gives us permission to say yes this person may be dying.

NNA2P1S214

In the same site in which the nurse reported using the most precise formula for its commencement, another nurse stated ‘Usually you commence, then do the initial assessment, then the doctor comes later’ (NNB6P2S214).

Additional reported difficulties were associated with getting out-of-hours doctors involved who, while attending, were frequently reported not knowing the patient well enough to know whether or not the decline that was observed was as a consequence of the patient being at the end of their life. Again, it is important to remember that it was beyond the scope of this study to collect the views of GPs. As a consequence, nurses reported that while medication changes might be agreed and signed off the LCP documentation was not completed or signed at this stage.

The Care Pathway really commenced around 7–7:30 o’clock. I called the doctors, doctor came, but he has to sign all the forms because he is out of hours and don’t know the patient. So he saw the booklet, so wrote ‘stop giving orally, stop all the oral tablets’ then he signed that.

NNA2P3S214

In this case the LCP documentation was eventually signed off by the patient’s own GP when he was called in as the patient’s condition had deteriorated.

Towards the morning on that shift around 6 o’clock she became a bit breathless and a bit cold in her extremities. Looked very pale, so we called the doctor because the chest was a bit rattlely.

NNA2P3S214

Care in the last hours or days of life

Care at this time in a nursing home was a nursing task rather than a medical one. Relatives were only rarely involved in the provision of care and in the interviews, with the exception of one family member a daughter who was also a qualified nurse, it was not felt appropriate that the relatives should undertake nursing tasks. During the observations a number of aspects of care were done frequently. These included mouth care, feeding and encouraging drinking, pressure area care, checking for secretions and pain assessment.

In the interviews, nurses were asked about the relative importance of these tasks and how judgement was applied to the frequency with which they were undertaken. Nurses juggled two different ideas in their answers. The first was the notion that certain things should be assessed, if not carried out, at regular intervals. The second was that discretion and the application of expertise should be allowed to define the frequency with which assessments were made and care was conducted. In relation to the patients whose care was observed, a combination of these two approaches was adopted. However, it seemed that in the fusion of these two ideas some oddities emerged, one of which was explored in the interviews: the relatively common practice of putting the room lights on fully during the checks of a patient at night.

I think the lights should be on because staff are observing as well, you can’t observe breathing and changes or anything in the dark but then always what happened in that room I think the regulator is not perfectly working so the light is quite bright. We can’t reduce it, it was a bright light but as soon as you put off also it was coming on, like then they thought there was a ghost or something because as soon as they put it off it was coming up light.

NNA2P3S224

The frequency with which a task was undertaken and how it was undertaken often reflected the views of the staff rather than the views of the patient or their relatives.

This can also be seen in the pressure area care undertaken for the patients.

Pressure area care

All of the patients observed in the study received regular visits from care assistants to turn the patient. Carers were observed to use their judgement about whether or not to turn a patient but in the interviews there remained an emphasis on 2-hourly turns, reducing to 4-hourly at night.

It’s 2-hourly usually once they go on the Pathway, which is roughly about the same otherwise. Some of the night staff were saying that on their shift . . . if they’re actually on the machine, they go down to 4 hours, they’re not disturbing them too much.

NNA2P1S214

You know the one who is really ill or bed bound we need to do it to prevent the pressure sore . . . most of the residents here we check 2-hourly. We check 2-hourly the residents who move themselves but the one you know, they cannot move they need to be turned 2-hourly to prevent the pressure sore.

NNB4P1S24

In relation to the observed death in NNB4 the nurse reported that she had instructed the staff to reduce the frequency of the turns as it was clear that the patient was dying.

4 o’clock I told them, don’t touch him because I know that’s the last, when I go I told them don’t touch him again because it looks now. So . . . because as long as he’s comfortable not to worry to turn every 4 hours . . . not to turn every 2 hours.

NNB4P1S24

Nevertheless, there was one site in which the nurse seemed to suggest that the prevention of a further deterioration in the patient’s pressure sore demanded frequent turns, despite their painful consequences for the patient.

Always it’s very painful for the patient but we have to do it to prevent pressure sores as we can’t prevent the pressure sore totally. It’s impossible and it’s very painful for patients in the last days . . . and we use a lot of cream to protect the skin but it’s not working, it’s not working and I see the expression on their face when we turn the patient side by side, that expression looks painful.

SNA1P1S23

Food and fluids

The provision of food and fluids was another area in which the goals of care were observed to be inconsistent. In one site care staff were observed to make frequent attempts to get a patient to eat and drink. It appeared that there was no coherent or consistent approach, with different members of staff adopting different strategies. The interview with the member of staff was important because it revealed the various attempts made by the nurse to get the staff to desist from feeding the patient.

The biggest issue is feeding and drinks, and they can’t seem to grasp that the patient wasn’t able to swallow, which she hasn’t been all this week. But they would still go in and because they might get response from her they will try and feed her or give her drinks. So, I have to say, ‘No, you come to me and I will say yes or no whether to give her anything’ and I find that the best way.

NNB6P1S24

The nurse recognised that the actions of the care staff did not match the patient’s ability.

I was doing her mouth care this morning and I was going quite a bit way down her tongue to try and clean it and there was no gag reflex whatsoever, yet one of the girls came to me and said, ‘Patient opened her eyes, shall I give her her breakfast?’ And that was this morning . . . and I just think it’s some guilt.

NNB6P1S24

Despite the nurse’s exhortations, the observations recorded the frequency with which the care staff entered the patient’s room to provide food and fluids. On the patient’s last day of life, her daughter, who was a trained nurse, repeatedly told care staff that her mother was not able to eat or drink.

The nurse’s explanation for the staff’s persistence was interesting.

It does distress them. It is their nature, caring, and they’re thinking she has gone all these days without food and they don’t really grasp that she can’t swallow and it might be a cultural thing because we have quite a few foreign carers and nurses, and I just think they feel they are neglecting if they are not feeding; they don’t see the mouth care and everything like that as part of what the patient needs at that time, if you will.

NNB6P1S24

The nurse went on to report that, for a small group of nursing home staff, there remained a belief that not providing food and fluids right up to the point of death was ‘quickening the death’.

In another site (NNB4P1S24) the nurse reported that, as a member of the night staff, the lack of clear reports in the notes had left her uncertain about whether or not the patient we were observing in the study was able to eat or drink.

Checking for secretions

In all of the sites, nurses reported the importance of managing secretions, as these were uncomfortable for the patient and distressing for the relatives to observe. This was not straightforward, as nurses weighed up the use of the suction machine versus the use of drugs.

If you see them dying so you know, all that suction, it’s not fair to be suctioned regularly. It’s not fair to be suctioned all the time.

NNB4P1S24

In NNA2P1 the nurse reported the difficulty of disentangling the discomfort of secretion from pain and from the agitation associated with dementia.

The main problems I had was excess secretions then, as she moved to the next stage we were unsure whether it was a pain or whether it was an agitation because she’s got dementia it was hard to work out exactly what the signs were so we went down the sedation line first of all with her.

NNA2P1S214

The management of pain

All nurses talked about a pain-free and peaceful death as the goal of care, and the ability to administer pain relief was seen as an important element in achieving this. Staff reported using a range of drugs and routes of administration, and in their accounts staff relied on both the patient’s self-report and the assessment of the patient by the care staff.

On that last day as well to try to keep her comfortable I was the one making decisions, there was definitely more secretions, so she got extra glycopyrronium and there was definitely some sign of pain. I’d decided it was pain because my colleagues, the girls, when they had turned her they said she seemed to be a bit uncomfortable on turning.

NNA2P1S214

She was kept comfortable for quite a while. Then she went on and that’s when we put the syringe driver on . . . because no two people are the same, we have minimum doses we start on and then we . . . obviously decide what this is doing for the patient and if not, we increase it and if then if it’s something we can’t increase, we phone the GP and if it’s something not within our scope . . . but we didn’t need to involve the GP because we managed OK.

NNA2P1S214

The nurse in NNB6P2 talked in detail about the way in which the dosages of pain relief were escalated for the patient in the study. Initially managed on codeine and paracetamol, the patient was believed not to be in any pain. However, it became clear to the care staff that the patient had begun to experience pain and so they began to administer the diamorphine written up in conjunction with the LCP.

He never says he’s got pain . . . but you could see from the face especially facial expressions. That he was in pain . . . They wanted to turn him and he said to them ‘I am in pain’ so they just leave him.

NNB6P2S214

The staff left the patient because the NIC reported that she was unable to increase his diamorphine at that time and it was deemed to be poor nursing to continue while the patient was suffering.

There was a belief that the ability of the nursing home to take care of the patient’s pain had removed some of the incentive to have patients at the end of their life admitted to hospital.

So the patients don’t need to go hospital in a rush, their wishes will be followed, also that they want to stay in their own rooms with their families and everything. Because in the hospital also at an elderly age they don’t want any interference with the resuscitation and everything. It’s like a natural death and with the medication after the doctor has prescribed, after the family agree and the patient has stopped eating and drinking when generally they deteriorate.

NNA2P3S214

However, the administration of drugs often left nurses feeling uncomfortable, particularly if the patient died soon after their administration.

Dying alone

None of the staff thought that their patients should die alone. Two nurses were present at the bedside when the patient died and for one of these nurses this was the first time she had actually seen a patient die. The nurse in NNA2P1 had been summoned by one of the patient’s sons, who had noticed that his mother’s breathing had changed. When the nurse arrived she recognised that the patient was dying and communicated this to the sons.

As soon as I walked in the room I thought she’s dying because I could just tell because of the sort of change because she was blue. She was blue and at that point she wasn’t breathing either, and I thought . . . she was taking her last breath, this was the impression I got because she wasn’t breathing, she went like . . . and her chest was still . . . I said to the boys ‘I think she’s going . . . So the boys both got by her side and then she took a breath.

NNA2P1S214

Having alerted the sons that their mother would die very soon, the nurse stayed in the room and the patient died an hour and a quarter later.

Dying in intensive care units

In ICUs, death often followed a decision to withdraw or withhold life-sustaining interventions. For this reason, judgements about the point at which interventions were no longer effective received considerable attention in the interviews with doctors and nurses in ICUs.

Doctors viewed the ICU as a specialist resource most appropriately reserved for patients for whom there was an expectation of recovery and survival. Admission to ICU was rationed on the basis of a prospect of recovery; doctors reported that they were effectively labelling those not suitable for ICU admission as ‘dying’.

I would say, particularly when we see referrals . . . from other specialities from medicine or surgery or something like that, and say ‘will you come and see this patient, will you take them to intensive care’ and we go up there and say ‘no that’s not appropriate because this patient is dying’ . . . OK, and dying is a sort of natural physiological process. And sometimes to try and intervene in that isn’t the right thing to do.

SIA2S161

End of life care is really quite difficult in intensive care because generally speaking one has admitted a patient into the intensive care unit because you think they’ve got a chance of survival. And you believe that with the appropriate support, to whatever level, that that patient will then survive . . . If we think someone’s not going to survive, we really shouldn’t admit them.

SIA1S158

The doctors described dying as ‘recognisable’ and ‘obvious’ and generally based their assessment on factors such as the patient’s history, current condition and prognosis, including the patient’s lack of progress, deterioration or reduced prospect of recovery. Nurses were not responsible for the admissions to ICU but, like doctors, reported being able to judge an ICU patient’s failure to respond to treatment. Although nurses reported that the formal acknowledgement of dying was based on repeated medical assessment, they also reported that, as nurses, they were in a prime position to recognise a patient’s deterioration and treatment limitations.

. . . because the nurses are at the bedside 24/7, we are the ones that see the subtle changes in the patient so we are going to pick something up a lot quicker than a doctor will. So I mean, from that point of view they [doctors] do ask for our input during the ward round but at the end of the day the final decision is a medical decision.

SIB2S113

Decision-making about withdrawal

There was agreement between doctors and nurses that decisions regarding withdrawal of treatment were primarily medical but usually drew on the views of the wider team. Although what constituted the ‘team’ varied between sites, and indeed between patients within sites, minimally the consultants reported seeking nurse input. Depending on the unit, such decisions might involve two or three consultants, junior or middle-grade doctors, the senior nurse and the bedside nurse, as well as the patient’s surgical or medical team if applicable.

Doctors were agreed on the importance of collective decision-making, for a number of reasons. There was a view that seeking a range of perspectives from colleagues resulted in more balanced decision-making, ‘balanced’ because the views that contributed to the decision-making were often reflective of different ways of seeing the patient. Medical staff outside the ICU tended to be seen as bringing a more optimistic perspective than their ICU counterparts. Although these views were invited, ICU doctors tended to locate them in ‘unrealistic expectations’ of what could be achieved in ICU. ICU nurses’ views were identified as the ‘most realistic’ in relation to patient outcomes because nurses were at the patient’s bedside every day, observing and recording the patient’s condition. ICU doctors reported having less face-to-face contact with the patient than ICU nurses and, as a result, being more likely to be ‘overly optimistic’ and more willing to keep treating. Where doubts were expressed or opinions were divided, active treatment was more likely to be pursued. However, doctors and nurses agreed that disagreement between staff was rare.

Nurses emphasised a decision to withdraw was often arrived at only when treatment options had been exhausted and that MDT meetings, both formal and informal, ensured that these decisions were thoroughly explored.

Often there’s two consultants nowadays on a ward round, because it’s a like a handover period or whatever, plus senior registrars, nursing staff including senior nursing staff. Physios are nearly always there, dieticians and pharmacists; MDTs every day. So I suspect that’s why that kind of conversation starts to take place because of the volume of expertise . . . and a microbiologist is there.

SIA3S163

While doctors and nurses agreed that decisions to withdraw were usually unanimous, the nature of withdrawal and the speed with which it occurred was reported as hugely variable and often influenced by the senior doctor in charge of the patient’s care. Nurses recognised the individual variability between doctors regarding withdrawal decisions.

People are individuals, consultants are individuals, and they are the ones ultimately making these decisions. Even as a group their name is going to be there so they have think carefully. And I’m never for one minute saying it’s an easy . . . it isn’t easy, but some consultants or intensivists can be much more decisive and others kind of play at the edges a bit. I’m not sure if that’s a kind of defeatist – like ‘I’ll just give them 24 hours, we’ll just give them 24 hours longer’, so it can be . . . I think variability can be unsettling. But then you’ve got to perhaps understand where they’re coming from and talk to them about it. So people think that’s when they’re slow in making that decision when it’s screamingly obvious but then you say to someone ‘well it’s not you making the decision’. You know, you have to talk to them about it.

SIA3S163

I think that sometimes certain consultants go too far and I have actually said to one consultant ‘look you are flogging a dead horse. What exactly do you hope to gain by this? You know this patient is not going to survive. What we are doing is cruel’ [. . .] They just, you know, there comes a point when you know the outcome is not going to be good no matter what you do but some doctors will persist in doing everything, even though you know that outcomes are never going to be positive, and it’s cruel.

SIB2S113

There was a consensus that withdrawal was a doctor-led clinical decision, and that the MDT, family members and relatives, although part of discussions, were not ultimately responsible for the decision.

Ascertaining the views of the patient and their relatives

Involving patients and family members in discussions about withdrawal was considered to be of paramount importance. Within the ICU setting, it was unlikely that patients could be questioned directly, as they were generally either unconscious or heavily sedated and unable to take part in discussions. Doctors reported starting from the premise that where the medical evidence indicated that treatment was not effective there was a default option of withdrawal.

We kind of have a default opinion really where most patients wouldn’t want to be entirely dependent on full time nursing care and be severely mentally impaired and we would generally withdraw treatment, unless we were hearing something very different about what that particular patient would have wanted, but if they had expressed I want that at all costs no matter what the disability, we would change our position in their best interests.

NIA6S111

This default option was reviewed in the light of evidence about what the patient’s wishes might have been in these circumstances. While doctors talked about the patient’s best interests, the determination of best interests was a complex fusion of general and proxy ideas. It did not conform to the definition or test laid out in the Mental Capacity Act (2007). 65 Best interests was constituted in terms of a generalised negative default option: what patients in general would not want as opposed to the considerations each individual might have liked applied to their situation. In determining the patient’s best interests, doctors were reliant on people who had known the patient well and who had some idea about the patient’s views. These people were usually deemed to be the patient’s relatives. Doctors described strategies that they had used with relatives in identifying what a patient’s wishes might have been.

I usually say ‘the decision is made in their best interests but I didn’t know [the patient] and could [you] speak to me, and offer me an opinion, I just wonder what [the patient] would say if he/she was watching us now and knew what we knew now, whether he would offer an opinion to carry on with whatever’. So I try and make it specific that I’m using them as a conduit for the patient’s opinion, whether they’ve either specifically expressed it or what they would think he/she would say in the circumstances and so that is so I am very explicit with it, that ‘it’s our decision but you can help us reach that decision’.

NIA6S111

I try and make it very clear that I’m asking their views on what the patient would want, not because I want them to make a decision about what we want to do long term, but because my decision, our decision on the unit of how we manage that patient must be based on what we think that patient would want us to do. An informed decision of what they’d want us to do.

SIA3S141

Nurses recognised that, in consulting relatives in this way, there was a risk that relatives might begin to feel that they were getting involved in making a decision.

We listen to their [relatives] . . . but we do make it clear that it isn’t their decision at the end of the day. That it is ultimately a medical decision because it’s the doctors that have the knowledge . . . it’s knowing how to kind of say to them ‘I’m sorry but it isn’t your decision’. I think if you can get that over to your relatives it kind of puts the ball back in our court then, therefore we are in control of the decision-making and some of them are grateful for that, it takes the pressure off them.

NIA1S113

Only one nurse reported experience of doctors who routinely involved relatives in decisions about withdrawal.

Pre-withdrawal communication

Medical and nursing staff reported that, prior to the ‘withdrawal conversation’, they worked as a team to establish what information the patient’s relatives understood about the situation and to ensure that the relatives knew that the patient’s condition was deteriorating. Consultants reported that having conversations ‘early on’ with relatives was particularly important in cases when the patient was not improving despite receiving treatment. Much like the nurses, doctors believed that preparatory conversations provided an important context to lessen the shock should the delivery of bad news be required. Erring on the side of pessimism and being proved wrong was seen to be better than giving relatives ‘false hope’. Early conversations were believed to facilitate relatives’ understanding, providing a basis for wider discussions with other family members as well as preparation for discussions about withdrawal. Only one of the doctors interviewed adopted an optimistic approach right up until the time that deterioration was seen as irreversible. He argued that a 20% mortality rate equates to an 80% survival rate (SIA1S158). Most of the doctors, however, felt that it was important to be ‘realistic’ about a patient’s poor prognosis.

I’ll always give them a heads-up early on that things aren’t great.

SIA3S141

As discussed above, the point at which withdrawal was decided was often preceded by a period of deterioration on the part of the patient, during which staff would prepare the relatives for the ‘bad news’ of withdrawal via teamwork and a series of ‘realistic’ conversations. Nurses placed considerable emphasis on communication with relatives that was continuous, honest and consistent.

If they are hearing from each successive person that the condition is deteriorating that the patient is not getting better, eventually it sinks in [. . .] if they are getting people telling them the same thing, on successive days and they are seeing the monitor and the numbers are going down and the support is going up and machines are being added, they know. I mean they might not want to know but they know.

SIB2S113

In the context of ongoing discussions about a patient’s deterioration, doctors and nurses recognised that some relatives might have already come around to the idea of withdrawing treatment before it was actually raised as a possibility. Several of the doctors reported that on occasions relatives ‘jumped at’ the decision to withdraw because they believed that it was in the patient’s interests and that the situation was irreversible.

Sometimes it’s the relatives who want you to stop and it’s us that’s saying ‘well we’re not quite at the hopeless stage yet, it’s very likely that I will be talking to you tomorrow and telling you that we are now, but he might still get better, so we’re still trying’.

NIB3S131

On the other hand, some doctors identified instances in which relatives did not accept that the point of withdrawal had been reached. In line with the strategies reported by the nurses, doctors tended to draw in a second opinion from a colleague in order to confirm both diagnosis and prognosis.

Our job then is kind of to present the information as clearly as possible and from different angles really so they don’t feel they’re just getting one opinion, they’re getting a clear consensus from several doctors ideally of what’s occurring.

NIA6S111

Following the decision to withdraw treatment from a patient, it was the senior doctor’s responsibility to communicate this to the patient’s relatives, and this generally occurred within the context of a formal meeting. Nurses reported that it was important that they attend this meeting in order to provide support and advice following the meeting. If the bedside nurse was inexperienced or the relatives were ‘difficult’, a more senior nurse might replace the bedside nurse.

Mostly the nurse will go, the bedside nurse, will go into the meeting with the consultant. To see the family and also if they expect any problems from the family then the nurse in charge – you know, if it’s a junior bedside nurse – nurse in charge will also go in.

SIA1S143

Depending on the nurse at the bedside, or if it’s a family that we’ve had issues with, then a senior member of the team would often go in with the consultant to be there, rather than maybe a brand new Band 5.

SIB1S143

Nurses were keen to emphasise that the senior doctors were in charge of discussing withdrawal with the relatives and all had different approaches. Doctors reported that these conversations should be clear and free of the use of jargon.

And you say, ‘look, they are dying. We have the ability . . .’, and one of the phrases I commonly use is ‘we have the ability to prolong people’s death as well as their lives’.

NIA1S141

So my technique would be that I would say that the patient, if they were dying, that we need to move towards making him comfortable rather than trying to make quality of life rather than quantity of life. It doesn’t mean we are going to withdraw care it just means we are stopping for example antibiotics, fluids, and explain why.

NIB4S112

Then we would get the family in and then we would explain to the family what the situation is and say that we would not – ‘it is not in the patients best interest to carry on’ [. . .] The majority of the cases they agree because they can see what is going on. And if they don’t agree it’s our job to convince them really to be honest with you. It’s not a decision that they have to make. The GMC [General Medical Council] is very clear as to that you don’t have to offer treatment which you think is inappropriate.

SIA2S161

Although doctors reported the importance of being honest, clear and careful not to provide ‘false hope’ in their communications with patients, nurses reported observing communication which fell short of what doctors themselves recognised as important. Nurses reported that their role in the aftermath of the consultant’s discussion with the relatives was to make sure the relatives had understood what they were being told, answer any questions that might arise and provide comfort.

We have got a particular doctor who really talks way over their [relatives’] heads. And you can see it on their face that they are so confused by what is going on, that they come out of it and they go . . . I have actually had a relative say to me ‘so what exactly did he just say to me?’ And that’s when I say to them, ‘OK look, this is what was just said. We have done everything we humanly can; there is nothing more we can do. And they are like ‘is that what he just said?’ and I am ‘yeah, that’s what he just said’.

SIB2S113

The families always tend to ask questions from the nurse afterwards rather than the doctor. That often happens.

NIA2S113

Then the nurse can back that up again after the meeting. We go over it again. Because sometimes when they’re in the meeting, and they’re upset, they don’t always take everything in.

NIB4S133

Although nurses reported the importance of clear direct communication with relatives, they themselves shied away from using the term ‘dying’ when talking to relatives.

Well, we think of a different word that we can still convey this message you know, ‘because the treatment is not making any better’, and ‘we are going to stop the treatment and make him comfortable’. These kinds of words will give: ‘at the end of the day he is going to die’. And many times, we may probably also tell them ‘we don’t know how long he is going to last’.

NIA6S153

Relatives’ responses

Nurses, perhaps rather simplistically, identified relatives as either accepting of the patient’s condition or in denial of it. For example, some nurses interpreted relatives’ questions about the length of time it would take for the patient to die as an acceptance of the patient’s irreversible condition and inevitable death:

Sometimes they approach us and say things like: how long’s this going to go on for? Well, quite often it is when they’re not close relatives. You know, when they’re nieces and nephews and they say: ‘how long’s this going to go on for’? Erm, so they’re, quite often they are quite accepting.

NIB3S153

We are talking to the relatives every day, every minute, every hour of the day, over the phone, straight to their face. We’re continuously telling them how their relatives are doing. Yeah? So we have no problem. We can understand. They might tell us first: ‘we don’t [want] them to be treated any more’.

SIA2S113

On the other hand, staff also had to negotiate situations in which relatives were defined as being in denial. Either the relatives were not ready for, not understanding or not in agreement that the patient was deteriorating and/or with the decision to withdraw active treatment. In the nurse interviews, the most common approach to managing relatives who did not agree with the withdrawal of active treatment was to ‘give them time’. This allowed nurses time to talk to the relatives, and the relatives time to process the information.

We will just discuss it with them, find out what their worries and concerns would be and see if we could just address it and just get them to understand the process that we are going through and why we’ve reached that decision.

NIB5S153

They just don’t grasp it, even though you know I have explained to them many times. And that could be difficult really. All you can do is just sit down and then just basically, just try to explain that they’ve come to the point that, you know, they will be able to, you know, understand it, try to understand it better.

SIB3S153

Very occasionally over the years, you’ve had a family that’s said ‘no’. And if they say no [then] there’s no point proceeding with it really, we might as well just back off a bit. Because they will go home and think about it and they will come back ready, either the evening or the next day, they’ll come back and we’ll have the discussion again and they’ll say yes.

NIB3S153

Although the goal of discussions between the medical and nursing team and the relatives was to come to a consensus on the withdrawal of active treatment, there were occasions on which relatives disagreed with the decision. In these very rare cases the nurses reported that the consultant might simply state that the decision was a non-negotiable medical one or that a ‘ceiling of treatment’ had been reached, beyond which there was nothing left.

Withdrawal of active treatment in practice

Moving patients from the category of ‘active treatment’ to ‘withdrawal of treatment’ implied a number of changes in care. However, the degree to which active treatment was actually withdrawn varied between sites and between patients. In very few cases did consultants report the implementation of complete withdrawal of active treatment. In practice, withdrawal ranged from non-escalation, to limitation, to consultants imposing a ‘ceiling’ of treatment for the deteriorating patient. In general, variation in the nature of withdrawal appeared to be based on the individualised nature of care within ICU. For patients on a high level of support – for example, maximum inotropic and ventilator support – it was reported to be unnecessary to withdraw all treatment; simply reducing the inotropes on a patient might bring about the patient’s death before anything else was withdrawn. For patients on lower levels of support, active treatment might be withdrawn in a progressive or step-wise manner. Withdrawal of active therapies included limiting or stopping inotropes, filters, antibiotics, intravenous (i.v.) fluids, nasogastric (NG) or other feed and/or ventilation. If the patient was ventilated, their ventilatory support might be reduced, changed or stopped altogether, with or without subsequent extubation.

I think for us on ITU [intensive therapy unit] it’s easy because you say well, you can do things to bring the end to its end, you know, not continuing infusion or making adjustments to ventilation. You know what you do would in terms of withdrawal of care.

NIA2S161

Decisions about what to discontinue were not straightforward; the ‘grey’ area of (dis)continuing food and fluids was particularly mentioned by doctors. The decision to withdraw food and fluids from patients who did not die quickly was viewed as ‘tricky’.

I don’t feel that maintaining fluids is going to delay death necessarily. There are similar situations we find in ITU and there are patients who do have some sort of consciousness and awareness that sensation of thirst isn’t a very pleasant symptom. I’ve got less of a problem with withdrawing feed because . . . and again, this is in an ITU environment, with sedated patients I’m not sure that . . . I don’t know, can you perceive hunger through sedation? I’m not sure. I suppose just from a personal point of view I’ve got more of a problem with withdrawing fluids than I have with withdrawing feed.

SIB1S121

The withdrawal of active treatment did not mean removing everything; drugs to ensure the patient’s comfort could be commenced, altered or continued if these were already in place. For the doctors, changing from active treatment to the withdrawal of treatment signified a fundamental shift in the care of the patient, namely from medical to nursing care. Typically, good post-withdrawal care was described by consultants as being about patient comfort, dignity, sympathy and peace. The use of such concepts reflected the consultant’s view that withdrawal was also about a shift from intervention to ‘softer’ nursing care. For example, several of the doctors stated that, following the decision to withdraw treatment, patients left the sphere of medical treatment and became ‘nurse level’, with little consultant input unless there were subsequent problems.

But it basically becomes a talk with the nurse looking after the patient ‘Is everything OK? Is the LCP running okay?’ and that will be the ward round for that patient. There wouldn’t be anything active, it’s more ‘is there a problem? Do you need the data changed or anything?’ So it all becomes . . . so it no longer becomes medical it’s all nurse led and then only gets fed up to a doctor if there is a problem with it. So that patient sort of drifts into the background for us and it takes a specific effort almost to remember and at least check in. I remember to check in but I haven’t . . . not everyone checks in on that patient.

NIA6S111

Yes I feel I would be tending to step back [. . .] once that conversation has been had with the family, I would talk with the nursing staff, particularly talking to them or sort of senior experienced sisters who’ve got a lot of experience with this sort of situation, and we would agree a plan, a course of action. Then, if there were any specific complications, talk through them at that particular time . . . and I would expect really then not to have very much input other than say there were problems during the course of that.

SIB1S121

It was clear from the doctor interviews that the ‘withdrawal of treatment’ was a complex directive which took its precise meaning and shape from the individual context in which it was being used.

The nurses supported the idea that once withdrawal of treatment was planned it was their responsibility to manage that withdrawal. The ‘process of withdrawal’ depended on the severity of the patient’s illness and the current level of drug support, and the consultant’s views on the needs of the patient and family. In general, depending on the patient’s condition, the ‘steps’ of withdrawal might include:

1. stopping inotropes (e.g. noradrenaline)

2. reducing oxygen (e.g. from 100% to 21% room air)

3. extubation (if possible)

4. stopping all other active treatment (e.g. NG feed, i.v. fluids)

5. introduction, or continued provision, of i.v. drugs to control pain and ensure comfort.

The order and degree of withdrawal of active treatment was a main focus of discussion for the nurses because they were the staff responsible for providing patient care once the decision to withdraw active treatment was made. In particular, the nurses differentiated between ‘quick deaths’ and ‘slow deaths’, and described how controlling different elements of the withdrawal facilitated a degree of control over the patient’s death and the relatives’ experience of that death.

Doctor and nurse views on the Liverpool Care Pathway

Six sites, three in the north-west and three in London and the south-east, were recruited to the study as LCP-using sites. However, as we have already reported, sites recruited as LCP-using sites did not always use the LCP to support care, while some LCP documentation was evident in some of those sites recruited as non-LCP-using sites. Regardless of whether or not their sites had been recruited as LCP-using sites, the majority of the doctors and nurses interviewed knew about the LCP and were supportive of it both as a document and as a pathway of care. The views of the doctors and nurses about the LCP demonstrated that there was no single or clear-cut understanding of the LCP as either a document or as a plan of care, except perhaps that it was ‘too long’.

On the one hand there were examples of doctors who were signed up to the principles on which the LCP was based.

I firmly believe what we’re doing is a significant advance on managing dying patients and in their best interests and making the dying process a comfortable and dignified process rather than uncomfortable and undignified.

SIA3S141

On the other hand there were doctors who reported that the LCP was ‘just a piece of paper’.

What makes the difference at the end of life is people and not paper and the people aspect of it, they give that personal touch.

NIA1S141

The relationship between withdrawal and the Liverpool Care Pathway

Doctors who had some knowledge or experience of using the LCP reported a conceptual separation between commencement of ‘withdrawal of treatment’ and commencement of the LCP. It was common for doctors to state that one activity took precedence over the other. One doctor described a stepwise process involving the decision to withdraw, and implementation of the LCP prior to the commencement of withdrawal or treatment limitation.

The consultant will have seen the patient and made the decision, will then fill the LCP in at that point having spoken to the family, obviously first.

NIA2S161

The LCP pathway doesn’t teach you to withdraw treatment. LCP pathway tells you and prompts you once you have made that decision.

NIB5S161

For some doctors it seemed that use of the LCP prompted withdrawal of treatment, while for others withdrawal of treatment preceded commencement of the LCP, if indeed the LCP was commenced at all. It was suggested that if the use of the LCP was not separate from and conditional on a decision to withdraw, there was a danger that, almost by default, the patient would be deemed to be dying.

If somebody is sort of being sort of lined up for the LCP 2 days in advance of the decision to withdraw treatment in people’s minds, they may be viewing it as an inevitable end product. It is inevitable that this patient is going to be on the Liverpool Care Pathway; they are already in their minds a dying patient and may stop treating them or having the same degree of vigilance that they would for another patient and I worry about that a little bit.

NIA6S111

For one doctor, the use of the LCP was further qualified by time. This doctor reported that the LCP was not usually commenced unless the patient was still alive half an hour following withdrawal of treatment. This was because it was ‘administratively cumbersome’ to undertake half an hour of paperwork for a patient who would die in 5–10 minutes.

Not a lot of our dying patients go on it. Well, putting them on the Pathway administratively is cumbersome. Signing the Pathway and signing all the drugs and so on takes about half an hour and . . . this is purely a practical consideration but there are a group of patients that would probably die within 5 or 10 minutes of stopping life-sustaining treatment and they will require no medication or anything from the Pathway, so very often they don’t formally get instituted on the Pathway, those patients that appear that they’re going to die within 5 or 10 minutes, because they’re in such a degree of life sustaining treatment, that they are overwhelming likely to die almost instantaneously. Though occasionally they do get formally put on it, but they don’t actually receive any interventions from it. They don’t get their lines removed and so on because they just die within 5 or 10 minutes really, and so pragmatically, what quite often happens is that life sustaining treatment has stopped and then about half an hour or an hour they get put on the LCP . . . then if they’re still alive at half an hour they’re then generally get put on the LCP, simply because of the administrative burden of filling it all in.

NIA6S111

Nurses were more likely than doctors to have been involved in dealing with the completion of the LCP paperwork. For those nurses who had used it, the LCP was regarded as a useful tool for encouraging staff to think about the processes of withdrawal.

I find it useful because it kind of makes you think about, think more about, what levels of care are needed. And it’s kind of indication to think about organ donation, you think about all the comfort of the patient, you think about the relatives, you think about the spiritual care and sometimes you could forget, you know, when you are busy we could, but that paperwork is actually good reminder on all the other areas.

SIA1S143

Another nurse emphasised the way in which the LCP allowed individual level care within a standard of care. Some doctors in this study also reported the LCP as a way of standardising care. It was common for the ICU staff to report the use of the LCP as a checklist and as a reminder to explore a number of domains with the individual patient in mind.

I like it. I think it’s a good tool for people, it has consistency, it’s almost like a checklist to make sure that you do things that’s in the best interests of the patient and the relatives. But maybe that comes from getting experience.

SIA3S163

It allows any variances to be documented so it’s quite nice in the fact that a lot of it is tick box. So you can tick to say ‘yes, I have done that’, ‘yes, I have spoken to relatives’, ‘yes, I have done this, I have considered that’, and then, if I’ve said there’s a variance, then you document at the side why it is different from the actual LCP.

NIA1S113

Some nurses, however, thought that there was a trade-off between tick boxes, which facilitated speedy use of the documentation, and accountability.

Some of it’s good because it’s tick boxes, but it’s got the disadvantage of tick boxes that some of the tick boxes are not signed. If you’re just ticking it, you don’t know who’s ticked it. Really it should be signed because then you know who’s signed it off and who’s done what. Whereas if you’re just ticking it and then sign the bottom and then you come on the next shift, and some of it’s not ticked, the next nurse wants to make sure it’s filled in. But it’s half been done through a shift. There’s no accountability for the ticks except for the line at the bottom of the page.

NIB4S133

This nurse also took issue with the ‘communication’ sections of the document, stating that the information recorded in these sections varied because staff understanding of what constituted ‘communication’ in itself varied. There were broader concerns expressed about the relevance of some of the domains to a setting in which many people died very quickly and as a result of withdrawal of treatment.

In my experience, the patients die so quickly once you stop treatment, then it doesn’t really matter: you’re never going to get through all those aspects of care.

NIB3S153

If the patient is likely to not maintain their own airway and you take the ventilator away, they’re just going to die. And so all the protocols involved in the Liverpool Care Pathway, there would not be time to implement any of those. If the patient is likely to maintain their own airway, then they’re going to be able to be sort of cared for, for a little bit longer. But the idea would be that you wouldn’t then be doing intensive monitoring, and that they may have what was considered to be treatment drugs and the necessary medications discontinued, and just need to have medications for palliation. And that’s where that framework seems to be appropriate, to me.

NIA2S113

We try to use it [LCP] but obviously . . . I think if they die very, very quickly, so that we haven’t really had a chance to fill it all in and you’d be doing it retrospectively [then no point] really no, because some of them, as soon as you withdraw they can die within minutes. Depends on how much support you see we’re taking from them.

NIA1S113

The amount of time required to complete the LCP documentation was discussed in all the doctor and nurse interviews. The documentation was perceived to be too long and an ‘unnecessary distraction’ for which consultants did not have time. There were some doctors who considered that the benefits of using the LCP outweighed the length of time it took to complete the documentation, but these doctors were unusual.

We have to fill out a whole bunch of paperwork when we do any form of intervention in intensive care. It’s a shame if there’s a reluctance. I mean, one must try and advocate practice that limits paperwork wherever possible, but if this is paperwork that occurs around discussions of withdrawal of therapy that will end up in someone dying, and the incrementation [sic] of drugs and therapies that could, in the wrong hands, contribute to the speed of their dying, then it’s appropriate that it’s done, there’s time taken and it’s carefully documented.

SIA3S141

As discussed previously, there was a general view that patients who were expected to die quickly were not considered to receive any benefit from the use of the LCP to support care. In the context of ICU care, the LCP was not considered by most doctors to be a useful tool, being too long and cumbersome. Perhaps ironically, one doctor reported that when the LCP was first introduced into the unit there was ‘initial resistance from the medics’ until they realised that it was mainly to be completed by the nursing staff, and then it got better’ (NIA1S141).

Few of the consultants felt it important, or indeed practical, when talking to relatives to identify the LCP by name as the plan of care to be used to support withdrawal of treatment. In fact, it was common for doctors to report that they referred to the LCP as a ‘protocol of palliative care’ or an ‘end-of-life care plan’. The clinicians reported avoiding use of the term ‘Liverpool Care Pathway’ because of the negative associations with the name reported in the press.

Up and until the press got onto the Liverpool Care Pathway, very few relatives, in my experience, really wanted to go into the details of what is meant by ‘palliative care pathway’. They were happy, they had trust in us to say ‘OK you’ve been looking after my relative to the best of your abilities, we’ve failed to get them, or we’ve not been successful in going through this illness, we don’t have the technology to do it, so we trust you to palliate them as necessary’, but I think that trust has been lost. So the way we’re going to move forward because we’ve identified that it is a problem, we’ve had families proving it’s a problem, we think the problem is with the name. We think the pathway itself is appropriate, but what we must highlight every single time is the communication.

SIA3S41

In ICU the management of pain, secretions and hydration is a central clinical concern for all patients. This is therefore an area in which doctors considered themselves to have developed considerable expertise. Although the majority of clinicians reported that, in general, the LCP as a tool was too cumbersome and lengthy, for some it came into its own in relation to the drug algorithms, which were reported to provide an important structure for drug provision.

We don’t use it from a philosophical point of view, we use it as a purely practical tool to establish what is acceptable and what’s not acceptable. It feels like it gives you, in terms of using drugs, purely to treat, stressing symptoms rather than giving drugs to shorten life. It makes you feel that you are doing the right ethical thing, treating symptoms rather than giving drugs to shorten life. It makes me feel more confident in a way – dare I say, medically, legally – more protected that I am operating within an established framework and, retrospectively, somebody might . . . would not be able to say ‘Why did you give that diamorphine? Was it to shorten their life?’ I said ‘No, it was to treat their symptoms of breathlessness’, so it makes me feel confident that I am acting within medical norms, and to be honest, that’s why I’m using it really.

NIA6S111

In addition, one senior consultant reported that the LCP framework communicates the clinician’s intentions regarding drug administration, which may enable a nurse to react in a timely and responsive manner to the needs of the patient:

It’s been sanctioned as a best practice pathway guideline because we’re not restricted to it in any way; we can deviate from it at any point if we feel the need to. But it means there are clear parameters as to what the nursing staff can do in terms of providing comfort care, which is what they want and what they need in order to make sure it happened. The problem otherwise, without set guidance, is you can have a patient who’s distressed but because a drug isn’t written up or the nurse isn’t sure whether she can give it, the patient doesn’t get the drug and therefore they continue in a state of discomfort.

SIA3S141

Narratives drawing on clinical data to justify a transition from cure to palliation

The narratives that underpin the transition from cure to palliation provide the backdrop for communication to relatives regarding care of the patient at the end of life. The discursive components shown in Table 12 draw on a number of linguistic repertoires.

Some or all of the Table 12 categories of information formed the narrative argument that justified the transition from cure to palliation in the ICU notes. While there was variation between sites, there was also variation within sites, and this may relate to individual doctor/nurse styles of record keeping and to the fact that a range of documentation was evident (LCP and non-LCP) regardless of the category in which these sites were recruited to the study. For this reason, distinctions between LCP and non-LCP documentation are not drawn. There also appeared to be an overarching logic behind the ways in which the categories were employed in the descriptions of the conversations held with the patient’s relatives. Usually the record of these conversations started with A, the patient’s overall state of health, and/or followed by B, the patient’s disease progression. These two categories of information were used separately or together to orientate family members towards the ‘bad news’: C, physiological markers of the patient’s disease, and/or D, the futility of active treatment. These last two categories, C and D, were used, again either separately or together, to back up or reinforce A and/or B, and to introduce E, the patient’s prognosis. All of the notes contained category E information, perhaps not surprisingly as the patient’s prognosis was a fundamental pre-requisite for F, the withdrawal or withholding of active treatment. Of the last two categories of information conveyed to the relatives, G, the notion of palliative care, was most often discussed in terms of freedom from pain and suffering as well as the promotion of peace and comfort. The types of information subsumed under category H related to the need to place the patient’s medical and legal considerations at the heart of decision-making.

Extracts from NIB3R09 case notes demonstrated the logic outlined above in the medical note record of the conversation with relatives:

Medical note: ‘30/10/2012 09:25 . . . speak to relatives – I think we’ve come to the end.

[consultant]

30/10/2012 11:40 Spoke to mum + others. Staff nurse [name] present.

Told that [patient] is very unwell and getting worse all night and all organs are failing. I told them she is on the maximum O² and the O² is not getting through. Pressure on the ventilator is high any more will cause more harm such as a pneumothorax. Her inotropes are back again and gone into very high amount and we cannot give her more since her leg (Lt) is going more purple. So we have to leave it like this.

She has got infection which is getting worse despite changing the antibiotic and I have added another one today.

I also told them her kidneys are not working and overall she is much worse + in these circumstances it is better not to resuscitate if she has a cardiac arrest. This is not in her best interest. I also informed them that I have changed some treatment and going to restart the filter. If she is not responding to the treatment, we should call it a day and keep her comfortable.

They seem to understand this + asked how long she will last. I said ‘I do not know’. We will give her a few hours if she deteriorates, we will discuss again with the view to keep her comfortable.

DNAR form completed. Signed [consultant].

31/10/2012 11:00 42 F bifida/severe CAP + sepsis

(Day 10 in ITU)

Condition worsened yesterday + continued to decline overnight.

DNAR completed yesterday + ceiling of care initiated . . .

31/10/2012 – 11:25 Discussion with family (mum) consultant + SN [staff nurse].

Explained [patient] has deteriorated further + hasn’t responded to maximum treatment overnight. I discussed withdrawal of inotropes and family agreed that we will not renew the adrenaline infusion when it runs out in a few hours- Family will contact other relatives.

These discursive categories were used strategically to prepare the ground for the difficult news that a patient was dying and that medical intervention had reached the limit of its powers. It was clear from the data that doctors’ discussions with a patient’s relatives about the patient’s imminent death reflected a shift from cure to care. This was not just a linguistic device but a very real change in the nature of the patient’s care. The requirement for relatives to make a cognitive shift away from the survival and recovery of the patient to symptom control and death has been discussed elsewhere. 26,67 The speed with which these conversations progressed varied in line with the patient’s condition. This is demonstrated in the cases of SIB3R08 and SIA2R04, below:

D/W partner/son [name]/children. They themselves can see that [patient] is dying. And are in full agreement that we have come to [the end]. Curative Rx [radiotherapy] is not possible. New focus of care will change to palliation.

SIB3R08

I explained to the wife + 2 daughters that the LCP is a comprehensive pathway whereby a medical team identify that a patient is dying and their care switches from active management to palliation. The LCP is designed to ensure the patient is kept calm, pain free and comfortable during the dying process, and that unnecessary/potentially distressing treatment is avoided. The reason for starting the LCP is that [the patient] has a systematic bacterial infection (sepsis) which has affected all his organ systems. The infection + its systemic effects have failed to respond to 72 hours or more of aggressive ITU care and his condition has deteriorated to the point where we do not feel he can ever get better . . .

SIA2R04

Despite the careful construction of a narrative which was designed to lead to the conclusion that the patient was dying, notes often also revealed a glimmer of caution, a small degree of uncertainty which could hold out to a patient’s relative the slim hope of recovery. In some of the notes the use of terms such as realistic, significant recovery, meaningful recovery and significant possibility left some linguistic room for recovery against all odds:

No realistic chance of improving and his deterioration is worsening.

SIA2R09

Dr explained likelihood of significant recovery is poor.

NIA2R10

Without prospect for a meaningful recovery.

SIA3R04

There is a significant possibility he might not survive.

NIA6R04

Will more than likely not recover from this and is most likely at the end of her life.

SIA3R06

While this lack of absolute certainty in the language used to communicate information about prognosis may reflect a wider problem with evidence and its interpretation,68 it is generally considered to be very difficult to predict accurately that a patient is dying. 22,26,45,69 It is clear that despite medical evidence to the contrary there are cases which defy medical expectations; these are often framed as the ‘miracle’.

In this study, the most commonly identified physiological markers used to support the decision that active treatment was no longer of benefit included multiple organ failure (involving at least three organs) and irreversible brain injury/damage. Ravenscroft and Bell’s survey in ICUs also reported multiple organ failure as one of the most common reasons for withdrawal of therapy. 63 In the cases of SIA1R01 and NIB4R03, the role of organ failure supported the notion of a poor prognosis:

Overall, very poor prognosis given 5 organ failures. Family aware of this; DNAR form completed today.

SIA1R01

Discussed end-of-life decisions: – if signs of multi organ failure – esp renal failure this will be indicative of terminal prognosis -will likely withdraw at any time.

NIB4R03

In this study, examination of the notes suggested a multipronged approach to defining the limits of treatment in relation to each individual patient. First, it was reported that staff had done their best, with the limits of active treatment reached or there being no possibility of doing more. Expressions such as on maximum treatment, our best efforts, full treatment and most aggressive therapies were commonly contextualised by disease progression or lack of response or improvement from patient. This is exemplified in the following extracts:

She is going to die regardless of what we do because she has deteriorated despite extremely aggressive treatment on ICU, but also on ward X.

NIB5R07

Advised he is receiving maximum support possible + any further deterioration + we may not be able to support him.

NIA6R04

I explained that his condition has deteriorated further today and we are reaching the limits of our supportive therapies.

SIA3R04

Despite maximal multi-organ support his condition was not improving.

SIA2R10

Told that any attempts to improve [patient’s] health did not help – he is deteriorating rapidly . . . explained that despite all the treatment given there has been no improvement in her condition.

NIB5R06

In many cases, the notes revealed how the concept of futility was actively tested by allowing a trial period during which treatment was continued or given to see if patients improved, responded or deteriorated. The Intensive Care Society’s guidance suggests that ‘when outcome is uncertain, it is worth considering a trial of intensive care treatment on the understanding that it will be withdrawn if ineffective’. 70 However, the nature of active treatment and the length of time allowed to assess efficacy were frequently discussed in the notes, as shown in the following extracts.

[Information Given to Relatives Form] 29/7/12 09:30 Present at interview Son, Daughter + Staff Nurse.

I explained the surgical procedure + the serious nature of a hole in the [-].

Her organs are failing + she is not responding to our treatment. We will continue treatment but I think it is likely that in the next few hours we will reach the limit of our treatments + if she is still continuing to deteriorate continuing Rx [radiotherapy] at this point would not be in her best interests.

Family understand. No further questions. [Signed by the consultant]

29/07 12:05 Present at interview Daughter, Staff nurse, Consultant

I explained that BP is continuing to drop.

Continuing Rx is futile.

I discussed withdrawal of Rx.

[Medical notes] Daughter upset but understands. [Signed by the consultant]

NIA6R09

14/6/12 12:14 Spoke to the relatives at 10:15 Notes written later because of other patient unwell. Told the daughter that [patient] is unwell with multi organ failure + ischemic bowel. Surgeons to have a look at it to see whether they can do anything. If they feel all the bowel is dead it will be open + close + we will withdraw treatment and keep her comfortable. In case of cardiac arrest we will not do CPR [cardiopulmonary resuscitation].

I told her that the decision is a medical one and in all probability she will die.

She accepted the fact.

Pt [patient] returned from theatre. Myself + the medical SPR [specialist registrar] spoke to the whole family + told them that we cannot do any surgery since the entire bowel is black. So we are going to withdraw treatment + keep her comfortable. They agreed. For TLC.

NIB3R06

Reports of what relatives were told and relatives’ responses to that information

Interactions between staff and patients were much in evidence in the ICU patients’ notes. There was lengthy consideration of the families’ responses to what had been communicated and in particular, where there were divergent opinions, how these differences were resolved.

A consistent line was articulated with regard to the relatives’ role in decision-making. Families were not expected to make medical decisions, as this was defined as the doctor’s responsibility. This was evident in the notes:

Aunt informed me that there were more relatives travelling from [name of city] & they were not able to make any decisions until then. I reassured them that there was no decisions for them to make – that once the family were here we would put [patient] onto a ‘t-piece’/macrophon + [patient] would be kept comfortable.

SIA1R06

I have mentioned that we (ITU team) along with medical team have made a decision of DNAR in view of maximum therapy and 4/s organ failure. The outcome is going to be very poor and no further questions from husband.

Have explained that we may make the decision to withdraw active management in next 24 hours . . . [Consultant Anaesthetist].

NIB5R07

[Patient’s] friend spoke to the consultant regarding [patient’s] chances in theatre. [Friend] was very upset following consultation stating that he felt that if there was a chance that the operation could work he should have it irrespective of the likely outcome. I apologised to [friend] and told him that he shouldn’t feel like he has to make the decisions re his friend’s health, that the decisions are for the Drs. NIC also spoke to [friend] and said the same. ITU consultant also spoke with [friend]. I wasn’t present at the time.

SIA1R07

However, although families were not expected to take medical decisions, the voice or opinion of the relative was not absent from the notes. There were three ways in which families’ views were represented in the notes: (1) family agreed to care plan made by staff; (2) family provided proxy views on patient’s preferences for care; and (3) family suggested care plan which was accepted by staff.

The plan of care

As Billings points out, in intensive care facilities, deeply anxious patients and families are thrown into a hospital setting that is often strange and frightening, and must work with unfamiliar health-care professionals. 72 In order for families to cope with and accept a terminal prognosis, they need effective and empathetic communication. 73,74 Seymour’s study45 and the study by Cosgrove and Bari74 found that both nursing and medical staff in ICUs saw the delivery of information to families as a major aspect of their role. Nursing and medical staff portrayed their role in giving information as a means of enabling families to adjust to the likelihood of death and facilitating the primary grieving process.

While focusing on communicating the patient’s condition to relatives, the notes also portrayed a narrative around the care that was being provided. Keeping families ‘updated’ and ‘informed’ was invariably recorded in the notes. The word ‘reassure’ or ‘reassurance‘ was used frequently in the notes across all sites and there appeared to be a clear and consistent line articulated in the notes as to what the relatives were told. Once the plan of care had been communicated and was being put into practice, the notes highlighted the relatives’ response, usually in terms of their acceptance of it. There was frequent use of terms such as:

• they (the relatives) are [(well) aware of]

• they have clear insight of

• they (appear to) understand

• they are (very) understanding

• they have a good understanding of

• they (appear to, seem to) accept

• they (seem) accepting of

• they are happy with

• they seem at peace with.

Just as clinicians appeared keen to get the relatives’ support for their treatment decisions, so clinicians also appeared keen to reflect conversations they had had with relatives about a proposed plan of care:

Have explained that we may make the decision to withdraw active management in next 24 hours. Husband was initially very upset to which I have reassured that the ITU management is always to maintain organ support to allow the recovery and in view of the both: had very poor nutritional and very poor natural immunity due to alcoholic liver disease and multiple nutritional deficiencies and ICU outcome is likely to very poor. Husband had no further questions, left the interview room to contact his daughter [Consultant Anaesthetist].

NIB5R07

The use of the Liverpool Care Pathway in Liverpool Care Pathway-using sites

For non-LCP-using sites, there was no evidence to suggest the presence or influence of an end-of-life care pathway. However, for LCP-using sites, the use of the LCP varied enormously between sites and on occasion between patients in the same ICU. The data reported here relate to the 11 observed deaths in five sites which were recruited because they reported using the LCP. Despite the finding in the retrospective case note analysis (see Chapter 9) that some LCP-using sites did not always use the LCP to support care, and that a minority of non-LCP-using sites had supported deaths using the LCP, all 11 observed deaths in LCP-using sites were supported by the LCP.

This section begins by examining the use of the LCP, and in particular the relationship between decisions about care and the use of the LCP. All of the sites used the LCP in different ways.

The major differences in the use of the LCP were:

1. timing of the use of the LCP

2. perceived purpose of use of the LCP

3. the use of documentation

4. communication of LCP use to relatives.

Withdrawal

As has been highlighted, in addition to the initiation of the LCP, the concept of withdrawal was central to the care of these patients in ICU. The following section focuses on the issue of withdrawal in all observed cases regardless of whether or not the LCP was used to support care. It focuses on four specific elements in the process of withdrawal:

1. the decision to withdraw

2. the management of withdrawal

3. the meaning of withdrawal

4. the observed process of withdrawal.

Deterioration of the patient

All of the bereaved relatives were keen to reconstruct the patient’s history of deterioration. Two types of deterioration were described. The first was a slow decline in the patient’s health; the second was a fast and sudden episode resulting in a speedy admission to ICU. The nature of the patient’s deterioration created different expectations and held very different implications for the relatives.

Withdrawal of active treatment

In the observations the withdrawal of treatment was a critical point. Withdrawal represented, for staff and relatives, the failure of treatment, the irreversible decline of the body and the imminence of death. The dependency of the patient on medical technology for life meant that its removal was highly symbolic in the eyes of the relatives.

. . . but in their eyes your mum’s no longer doing anything by herself . . . It’s all the machinery doing it, you know.

SIA1P2F1

However, the fine detail around withdrawal was absent from the relatives’ accounts of the patient’s care in ICU. This was curious. The act of withdrawal was presented almost as less significant than what had gone before and what came afterwards. Clearly the relative’s accounts regarding the withdrawal of active treatment were filtered through the patient’s death. This is important to understand because the accounts provided were highly selective when compared with the observations. Some aspects of care that were observed to be discussed or argued over in the period up to the patient’s death, and in relation to withdrawal in particular, were absent from the relatives’ accounts. Further, without exception, the relatives interviewed were unable to remember some aspects of the care provided and there were some very clear examples of what was remembered being inaccurate.

At first it seemed that the time between death and the point at which the relatives were interviewed had allowed elements of the accounts to become simplified. However, it also became apparent that in reconstructing the death with hindsight, the aspects of withdrawal that may have seemed important at the time had been replaced by other elements.

The narratives of withdrawal were structured around three key stages: the decision to withdraw, the process of withdrawal and the outcome – waiting for death.

Use of the Liverpool Care Pathway

The majority of bereaved relatives (n = 11) had been present in sites which used the LCP. The standard of care described by the relatives was very high and it was difficult for researchers to elicit from relatives ways in which the care could have been improved. However, very few of the bereaved relatives had heard of the LCP. The most common response was that relatives could simply not remember whether or not they had been told anything.

I think two of the doctors went into more lengthy discussion but I really can’t tell you now because I just can’t remember.

SIA1P1F1

I’d never heard of it. It was sort of said that he’s going to be on this Liverpool Path Care, but, now whether the lads knew I don’t know.

NIA6P2F1

I just couldn’t take it in. They were trying to . . . I just didn’t take it in.

NIA1P1F1

When I picked up the belongings, I took a card and some chocolates to the ITU staff because I was so kind of grateful for how they’d looked after him. Um yeah, yeah, it was I think . . . it was called the Liverpool Care Pathways.

SIA1P3F1

Two sets of relatives reported hearing something about the LCP. In the case of NIA6P3F1, the patient’s mother and the sister reported that they thought it had been mentioned but it had not been explained to them at all. The two relatives discussed it between themselves in the following excerpt:

But they didn’t actually explain it to us . . . the only thing they kept saying was that she was being starved to death . . .

Somebody, and I forgot who it were now, he wasn’t in the hospital, it was somebody that said to me. They said the Liverpool Care thing, if you’re old and nobody goes, they give you an injection and you die. That’s what they were told.

NIA6P3F1

Another relative reported that the LCP had been mentioned but she could remember very little about it.

He also mentioned ‘Have you heard of the Liverpool Pathway?’ He was talking and it was, it was all over my head a bit and I’m thinking ‘Is this to do with euthanasia? Oh my God what is this? . . . I thought it was a new way of dying. You know that it was similar to letting somebody go.

SIA3P1F1

This relative went on to explain:

And of course I was crying as well because I was upset and I was trying to be brave. Because I didn’t want to go back and upset everybody else. So I was trying to stop. And he was, he did talk about it but I don’t know what he was saying.

SIA3P1F1

Although all the relatives in the LCP sites reported a high standard of care and attention, they knew little about the LCP and its relationship to the care provided.

Deterioration of the patient: the clinical context

When discussing the care of the patient who had died on ICU, all of the nurses framed their accounts in terms of the patient’s deterioration. This seemed to be important as all the patients had been admitted to ICU on the basis that there was some hope of recovery.

While you’re still treating them there’s still that sort of glimmer of hope. And when you’re actually withdrawing treatment, you sort of feel like . . . Not that you’ve failed but . . . I don’t know. Like I say just a bit helpless really.

NIA6P1S214

The interviews were undertaken after the patient’s death and this may have influenced the structure of these accounts which described the interplay between the patient’s clinical decline, the clinical interventions instituted to address the patient’s decline and the outcome of these interventions.

His heart was restarted in Casualty and he came up to intensive care at 7 o’clock in the evening for monitoring and to see what we could do about this slow heart rate. We needed to give him an infusion of something that’s quite caustic, so it’s only done here, not in general wards . . . And I was taking handover . . . his heart stopped again. And we had to restart it another two times in that episode. By 11 o’clock that night he was then intubated, ventilated, sedated. So, we could get a good look at what was going on with his heart and kidneys. He was on as much noradrenalin as he could be on. Which was maintaining his blood pressure and his blood pressure was still falling. And ventilation-wise he wasn’t really able to get enough oxygen into his blood in spite of his ventilation so it became very apparent that his, his lungs were failing, his heart had failed, his kidneys had failed, he had too many organs failed for us to be able to make him better. But we carried on and on and on, still for full care though if his heart stopped we had said we wouldn’t resuscitate.

SIA1P1S21

She was very severely sick on admission and got intubated because she wasn't able to breathe by herself, she was on a ventilator. And then she became very, very unwell over the last few days where she had maximum treatment of care, ventilation wise and also on medication to support her blood pressure, kidney function, yesterday she had a CT [computerised tomography] scan. She ended up having a multi-organ failure, all her organs failed and . . . she ended up in heart failure, she had a left right sided heart failure and so the heart wasn’t pumping enough and the treatment we were giving was just supporting her blood pressure and the cardio output but she wouldn’t be able to live to survive.

SIB1P2S21

These accounts existed across all sites and formed the basis for discussing detailed aspects of the observed care. In these discussions there were similarities between nurses working in LCP ICU sites and nurses in the non-LCP-using sites, but there were small but subtle differences between the LCP-using and non-LCP-using sites. The differences lay in the role of the nurse following withdrawal. The nature of this distinction will be discussed in the context of the LCP-using and non-LCP-using ICUs.

Liverpool Care Pathway-using sites

Non-Liverpool Care Pathway-using sites

Six staff were interviewed who were working on ICUs where the LCP had not been used to support a patient in the last days and hours of life. In addition, a student nurse, present during the observations and being mentored by a senior sister, was also interviewed. All of the staff interviewed, with the exception of the student nurse, were aware of the LCP and some had even used the LCP in the past. The more senior members of nursing staff were more likely to have reported using the LCP before.

As previously mentioned, all of the nurses described the clinical context of the patient’s decline and this seemed to be as important for understanding the care in the non-LCP-using sites as in the LCP-using sites.

Goals of care: Liverpool Care Pathway-using and non-Liverpool Care Pathway-using sites

The goals of nursing care in the LCP-using and non-LCP-using sites were identical: keeping a patient comfortable and pain free and ensuring their dignity. In all but one of the non-LCP cases, the religious and spiritual needs of the patient were reported to have been considered by the nurse.

I make sure that the patient is comfortable and is pain free and I am sure there is no distress, so once that is sorted and I see that then I, then I can focus on the family and make sure they are happy how their loved one died.

SIB1P2S21

From active treatment to withdrawing treatment, the main thing is to just ensure that she’s comfortable really, at that point . . . She had the priest came in.

SIB1P3S24

Achieving privacy for dying patients and their families on ICU was recognised to be a challenge, which all nurses reported managing in the same way, by using the curtains around the bedside.

Men don’t like to be emotional in public. So I drew the curtains just to give him some privacy and also some time alone with his father . . . And then they can let their emotions out and not feel that everybody else is watching them.

NIA1P1S24

She was in an open bay, so I tried to pull the curtains round to get that bit of privacy, even though it wasn’t probably ideal as such, but I feel that it was probably good enough though.

SIB1P3S24

We don’t like anybody to think that anybody dies, we pull all the curtains round all the beds and, everybody kind of looks the other way and pretends that nothing’s happening. Then magically as the curtains are drawn back, there is an empty bed space there . . . it’s a bit like a pantomime really.

NIB3P1S213

Patient observations

For the purposes of the economic analysis the patient observations were used to understand the quality of care provision and the resources used for each patient. Only in this manner could sufficient detail be generated of the broad range of public and private resources utilised by patients in their final hours or days of life. It is well recognised that effective communication between staff and patient/relatives is fundamental to the provision of optimal end-of-life care. In addition a range of subjective elements, for example ‘did the nurse smile and interact positively with the patient and the relatives?’, are likely to be as important in determining ‘quality’ as objective elements such as drug regimens. Equally, accurate allocation of the most important resource (staff) between individual patients requires a detailed knowledge of interactions between staff and patients. All of these aspects featured in the patient observations.

All staff costs were taken from a combination of NHS salary data and allocated in relation to the level of care provided, as measured by the interaction between each staff member and each patient in relation to hospital stay, and Unit Costs of Health and Social Care [Personal Social Services Research Unit (PSSRU)]78 costs.

• For nursing time in ICU we assigned two different costs: one for contact with the patient, which is £1.62 per minute, and the other for nursing duties for the patient or family without direct contact at £0.67 per minute. These costs are £97 and £40 per hour, respectively. Registered general nurse costs were based on the PSSRU costing document; see Appendix 29 for full costing details.

• Nursing home nursing staff costs were accounted for at the same level as the ICU nursing staff (£1.62 per minute) for the purpose of the study, in order to keep both sets of costs equal; therefore, if there was a true difference in care it would be due to the amount of care rather than cost differences. In addition, this method was adopted because nursing home providers around the country have varied pay structures and rates, making comparisons difficult. Similarly, costs relating to auxiliary nurses/health-care assistants and carers were calculated at the same rate throughout both sites, assigning a wage of £9.77 per hour, as per NHS Agenda for Change79 band 2 salary structure.

• Costs assigned to porters within the ICU observations were based on band 2 point 1 in the NHS Business Services Authority, at a cost of £9.77 per hour. 80

• The school-aged assistance in the south nursing home was paid help; therefore, for the purpose of this study costs have been applied at the National Minimum Wage for 16- to 17-year-olds, which in 2013 was £3.68 per hour. 81

• In one of the sites the nursing home manager was involved in giving care. For the purpose of the study, the cost applied is at PSSRU rate for a nurse team manager with patient contact at £2.37 per minute. However, we acknowledge that this would vary depending on the nursing home and location.

• One of the nursing home sites had a sister in the unit. This has been costed at PSSRU rates for a nurse team leader, with patient contact at £2.03 per minute.

• An important part of end-of-life care is care of the family; during the observations it was noted at times that their family members were brought drinks (tea and coffee). It was not always known which member of staff carried out these hospitality duties, although in some cases it was known, for instance, that volunteers were responsible for this role. For the purposes of consistency it was assumed that individuals who provided tea and coffee in the ICU were hospital volunteers and hence costs have not been assigned.

• For a number of the resource use items, it has been assumed that they can be separated easily; for example, mouth swabs come in packs of 25, with three swabs in each pack (75 swabs in a box). Therefore, for the purpose of the study it has been assumed that one item can be individually costed. This process has also been carried out for items such as gloves and incontinence pads that come in bundled packs.

The primary analysis was derived from the observational notes captured by the researchers during the last few hours or days of life. Once all observations had been completed, an eligible sample was selected (i.e. had complete data for all of the sources of data being assessed) from each LCP-using/non-LCP-using ICU/nursing home. The information relating to resource use was entered into a Microsoft Excel^® table (Microsoft Corporation, Redmond, WA, USA). From this, and using the times noted down by the researchers, a time period was assigned to each action that required resource use.

The final aim was to micro-cost all of the actions observed, using the data sources available. The costs chosen to allocate to the resource use are current NHS figures and recently published costs, although some assumptions still had to be made throughout, and these are outlined later in this chapter (see Limitations and assumptions).

Drug charts and interventions

A detailed analysis was undertaken of drug charts to identify comparative use of medications. In addition, all other diagnostic or therapeutic interventions were identified for each patient. Drug prices were taken from the British National Formulary. 82 Using the drugs charts gathered from the hospitals and nursing homes, any medication that the patients were administered was recorded and costed to the nearest amounts assigned.

When costing drug usage we have costed the whole value of the nearest dosage, therefore not accounting for waste or vial sharing, as it is assumed that the remainder of drug would have been disposed of. All of the drug costs used were sourced from the British National Formulary online 2013 list. For a detailed list of resources, see Appendix 29.

Bereaved relative economic tool

The aim was to review existing measures and identify a suitable quality of dying and death tool for use in this study. If no suitable tool was identified, the next objective was to develop a tool suitable for use in the study.

Quality of Dying and Death

The Quality of Dying and Death (QODD) questionnaire is a 31-item questionnaire and was developed in order to assess the quality of the dying process. It is completed by family members or clinicians after the patient has died. 83 The QODD questionnaire comprises a number of quantitative questions. The responses from the questions are totalled to a value between 0 and 100, with a higher score denoting a better quality of dying and death. Although there were some elements of the tool that were relevant to this study, such as how the family members’ needs were assessed and their views were taken into consideration in the care of the patient, in this study we were concerned only with care in the last days and hours of life.

Views of Informal Carers – Evaluation of Services

The Views of Informal Carers – Evaluation of Services questionnaire contains 158 questions in eight sections. 84 Section F asks questions specifically about the last 3 days of life and the care given over that time period. The majority of the questions have the options to respond ‘yes’, ‘no’ or ‘don’t know’ and there are spaces to add comments to some questions. Degrees of preference are not measured, which limits the use of this questionnaire for an economic assessment of the quality of care at the end of life.

Evaluating Care and Health Outcomes – for the Dying

Evaluating Care and Health Outcomes – for the Dying has 94 questions divided into nine sections. 85 The time frame of focus is the last days of life and the period immediately after the death of the patient. As the tool was developed based on the LCP domains of care, there was some concern that the use of this tool in the study would be biased towards the LCP-using sites.

Elements of a good death

Before developing a tool it was important to consider the conceptual and practical elements of a good death that are important to patients, their families and carers at the end of their life. Two key studies have been undertaken to this end.

Steinhauser conducted focus groups and interviews with patients, relatives and individuals providing care to patients at the end of life to establish a list of the components contributing to a good death. 86 Six key factors were identified, which included the management of symptoms and pain; clear communication between health professionals, patients and their families which enabled collective decision-making; being prepared for death; enabling the patient to participate in any spiritual rituals or being able to settle conflicts and say goodbye (termed ‘completeness’); the ability of patients to play a role for the benefit of others (this may be in the form of gifts or spending time with other people); and treating the patient as an individual no matter their level of illness or consciousness.

Low and Payne87 developed a questionnaire to obtain the perceptions on death of health-care professionals working in palliative care. They were concerned specifically with the factors that were perceived to contribute to a good and a bad death. The authors produced a factor analysis of the survey results and found a patient dying without distress to be of fundamental importance to a ‘good death’. This factor encompassed aspects such as a lack of pain and anxiety, as well as family members accepting the prognosis and being at peace. The factors that contributed to the perception of a bad death included elements such as the patient dying without people whom they are close to being around them. In general, the health-care professionals perceived a good death to address any physical or psychological distress of the patient while preparing both them and their family members for what was to come. Aspects of spiritual and cultural rituals were deemed an important contribution to the holistic care of the dying patient.

Development of a de novo outcome measurement tool

Given that most patients at the very end of their life will not be able to complete an outcome measurement tool, it was decided that the views of the patient’s family would be sought. For this reason, the tool’s design needed to take into account the sensitivities of patients and close family members.

The tool in its piloted format contained seven factors which family members were asked to rate on a scale from 1 to 10. This scale was chosen to provide the ability to vary the answers to different elements of the questionnaire by a number of points; in addition, a 1–10 scale is commonly used, which provides familiarity to the assessor and prevents introduction of unnecessary complexity. The poles of the scale represented extremely bad care (1) and care as good as it could be (10).

In line with the layout of the QODD questionnaire, there was an additional box within each question which enabled the responder to discount the domain as an issue, that is, that the patient did not experience the problems identified in the domain.

The relative weight of each of the factors that play a role in enabling a ‘good death’ is thought to differ. 88 In order to apply variable weights in this quality of dying and death outcome tool, the precise magnitude of each of the weights, and their contribution to the estimation of the quality of dying and death as a whole, first needed to be developed. Members of the research team and a convenience sample of relatives were asked to allocate weights to each of the elements in an initial attempt to quantify the preference weights. In this way, it was hoped that the preferences and experiences of the bereaved relatives would contribute to the development of a more sophisticated weighting of the contributing factors.

The information generated from these exercises does not enable precise quantitative weights to be determined. It does, however, allow a form of sensitivity analysis to be conducted, which overcomes the issue of all the factors being weighted with equal importance.

The bereaved relative completed the revised tool at the end of the interview. Respondents were also asked to talk through their response to each item as they made it, and their answers were recorded to provide additional feedback on how the relatives understood and completed the tool.

The final component was then to use the tool to evaluate the quality of dying and death of the patients as perceived by the relatives/carers. A simplifying assumption of linearity was made by which each increment (1–2, 2–3, etc.) was held to have equal value. In this manner the individual values assigned for each patient could be summed and averaged.

Once a detailed cost table for each of the patients sampled had been developed (see Appendix 30 for an example of the detailed cost table for case study 4), a summary of the resource use was derived to show the individual amounts of each resource consumed during the period observed.

It is important to note that the costs accounted for in each of these case studies are not directly comparable, as the length of observed periods differed between cases. Resource use can be calculated only for the time periods that were actually observed, and, similarly, the costs accounted for are not the whole costs of care in the final hours or days of life. The purpose was to focus on the time periods observed to assess any extra care undertaken, including staff care, and time that may have been driven by use of the LCP to support care.

Bereaved relative economic tool results

Data from a total of 22 complete observations and the bereaved relative economic tool formed the focus of this analysis. The challenges to recruitment have been detailed elsewhere in this report, but it is important to highlight that the limited resultant sample has also had a negative effect on the economic analysis. Table 17 shows a breakdown of each of the questionnaires filled in by location and site type. It is important to highlight that the observations in the north non-LCP-using care homes were from the same site, and, although at recruitment this site did not use the LCP to support care, by the time the observations were undertaken the LCP was being used.

In order to analyse the results by institution type and whether or not it was a LCP user, a series of tables were created. This allowed for assessment of overall trends to establish what (if any) were the most important factors to the patient’s family.

The first set of results shows the north nursing home LCP-using sites. As identified in Table 18, the pain management scores are quite varied between patients even though all three observations were undertaken at the same site. The relative of patient NNA2P3 scored pain management at 3 because they believed that the patient had trapped wind and the nurses did not attend to this, although they thought that the pain management through the use of drugs was controlled well. It is interesting to see that the relative of patient NNA2P2 scored 10 across all areas except for breathlessness, which they believed was not an issue. They also believed that the spiritual issues were most important and were pleased with how this was managed. The relatives of patients NNA2P1 and NNA2P2 allocated a score of 10 to personal care and dignity/respect as well as the staff keeping them informed, demonstrating that they felt that both they and the patients were treated well, which was particularly important to the relative of patient NNA2P1. The lowest-rated score from the three observations was for the pain management of patient 3. However, the narrative data collected during the relatives’ allocation of this score reflects uncertainty about the time period to which the pain management question was relevant. In this particular case the patient had been treated on a ward as well as in ICU. Pain management was also the most important factor to this relative, which may account for such a low score. Of the 17 numbered results, nine were scored at 10, two at 9, three at 8, two at 6 and one at 3.

Looking at these results in line with the LCP-using nursing home in the south of England, shown in Table 19, we can see that in these two observations there were no scores of 10, although the results were still at the higher end of the scale. All of the results given (except for the three ‘not an issue’ responses and one declined to answer) were between 7 and 9. It was interesting to see that the relative of patient 2 thought that the management of the patient’s dignity and respect had been very good, although the item was only allocated a score of 8.

Ideally, the results for the LCP-using nursing homes should be contrasted with the results for those not using the LCP. However, as we have seen, this home began to use the LCP during the course of the study. For the purposes of analysis this site had to be referred to as a LCP-using site in order to avoid introducing bias. Additionally, there are no data at all for the south non-using LCP site.

Table 20 shows the results from the north non-using LCP site (evaluated as a LCP site). There were some interesting reasons given for the scores that were allocated. In the case of the management of pain, breathlessness and agitation, the relative of patient 2 chose not to score 10 because she was not rating for herself and did not really ‘know for sure the way her relative was feeling’. No two of the aspects were scored the same by the relatives, although the scores were always relatively close by 1 or 2 points. The relative for patient 2 believed that all of the care aspects were equally as important, while the relative of patient 1 believed that dignity and personal care were the most important.

The results from the family tool for the ICU LCP-using sites are shown below in Table 21. There were six observations of this kind from three sites. As can be seen, the lowest score on any of the care characteristics was 7, which occurred twice: once for keeping the family informed at site NIA6P2 and the other for the management of agitation at site NIA1P1. Observations NIA6P1, NIA6P3 and NIA2P1 all received scores of 10 where rated. The relative of patient NIA6P1 reported being certain that pain management had not been an issue because she had consulted the patient about pain at the time. This relative also reported that the quality of personal care was important to her. The results for patient NIA6P3 were rated as 10 across all care aspects with no ‘not an issue’ given; the results here for this patient were interesting as both the mother and the sister were present during the interview to answer the family tool. As the mother was the primary and intended respondent, her results are shown in Table 21. The sister mostly agreed with her mother’s scores; however, she would have answered 7.5 for ‘staff keeping the family informed’ because she felt that she had not known what the LCP was or how it worked. The relative of patient NIA2P1 believed that breathlessness and agitation was not an important aspect, although all of the other aspects were scored a 10. The relative of patient NIA1P1 believed that all of the aspects being scored were of equal importance, and scored the care 7, 8, 9 and 10, with personal care the best and agitation the worst. Likewise, the relative of patient NIA2P2 scored 8, 9 and 10 as the results for care, with a 10 score for pain management, agitation, dignity and keeping the family informed as the highest results, and for this patient keeping the patient pain free and comfortable were the most important aspects.

The data from the south LCP-using ICU are shown below in Table 22. More observations were undertaken in the south sites than in the north. Results here are more varied; for example, the lowest score was 0 for spiritual care for SIA3P1, in which the family member would have liked some spiritual care for the patient, but had not been asked about this. The relative scored breathlessness, agitation and dignity as 10; however, for breathlessness the relative contemplated answering 8, as they did not know exactly how to respond because the patient was on a ventilator and it was hard to assess breathlessness. The next overall lowest scores were from the relative of observation SIA1P1, who graded pain as 5 and breathlessness as 6. The relative, however, commented that as they did not know the level of pain (if any) that their relative was in they did not want to score it any higher. The remaining aspects – personal care, dignity/respect, spiritual and staff keeping the family informed – were rated 10. This relative believed that keeping the patient out of pain and clean and comfortable were the most important aspects. Observation SIA1P2 also scored 10 for all aspects except for agitation, which was not an issue; the relative believed that all of the aspects were equally important, but so was patient self-respect. For observation SIA1P3 the relative said it was difficult to tell if the patient was in pain, but that they were satisfied when the nurses told them their relative was not in any pain. Overall, they responded with scores of 9 and 10.

Only one observation was undertaken in the north non-LCP-using ICU (Table 23). The results were varied and no scores of 10 were awarded, although the relative commented that they were unsure how best to score pain management as their relative had been heavily sedated. The patient did not have breathlessness issues and they were not religious. The relative scored personal care as 6, pain, agitation and dignity as 9 and keeping them informed as 8.

Three observations were undertaken in the south non-LCP-using ICU sites (Table 24). The lowest score assigned was from SIB2P1 for agitation, to reflect that fact that the relative felt that patient had been uncomfortable. On the other hand, the relative believed that dignity, respect and staff keeping the family informed were all good enough to be scored 10, while the management of pain and breathlessness scored 9 and personal care scored 8. The relatives of SIB1P1 rated 8 for all of the aspects of care overall, with ‘not important’ for spiritual aspects. However, they explained that if they had been asked to rate only the last 6 hours of the patient’s life, the scores would have been much lower: breathlessness, pain and agitation would have been scored as 3, and personal care, dignity and being kept informed as 1. For site SIB1P2, the relatives scored all of the management aspects as 10, except for spiritual, which they did not feel comfortable answering, and staff keeping the family members informed, which they gave an 8, as they believed there were some inconsistencies in what the family had been told by the staff. Overall, the family believed that pain management was most important to the patient, but communication was most important to the family.

Case study 1: north Liverpool Care Pathway-using nursing home

The cost table for the LCP-using nursing home identifies an overall observation period of just under 9 hours. This time period covers 1 day and the observer’s breaks have been removed. The results from this site highlight the different levels/types of staff that were employed and the structure of the nursing home itself. Here, most of the care was undertaken by the carers; there were also sisters employed in this nursing home who had more contact with the patient than the nurses over the period of the observation. The death occurred in the nursing home and a doctor was called in relatively early, but at the point of death the observer did not see a doctor. The total additional cost of care over the 8 hours 55 minutes was £157.03. However, had the nurses or sister provided as much care for the patient as the carers had, this would have been much higher, owing to the wage rate. If we look at the cost table in relation to the family questionnaire results, the scores were all 10, indicating that the family was happy with the care and attention given to the patient (Table 25).

Case study 2: south Liverpool Care Pathway-using nursing home

The observation for the south nursing home LCP-using site took place over 2 days, with much shorter observational periods. In total this amounted to around the same number of hours as the north LCP-using nursing home. The cost of care here was much lower than in the north site, although this could have been due to the two short periods as opposed to one full observational period. During the first day, the observer had just missed the administration of morphine to the patient, which cannot be included as, although we knew about it, it was outside the observational period. Similarly, on day 2 of the observations, midazolam, morphine sulphate and haloperidol had been administered 30 minutes before the observer arrived, and so could not be included in the observations. The staffing of this nursing home differs from that of the north as there are more registered general nurses working with auxiliary nurses/carers. There were also no doctors or sisters seen during this time, although the nursing home manager herself took care of patients.

The total cost of care for this case was £70.42, although as the patient died overnight the cost of post-death care could not be included as it was not observed. The cost of care for day 1 was greater than for day 2, although the observation time for day 2 was longer; this was due to the level of staff time spent with the patient. In this site a school-aged helper was employed to bring drinks to the family and patients.

Assessing the cost of care against the family tool we can see that the results are primarily 8, except for agitation of the patient, which scored 7, although it is known that the patient was prescribed medication to ease this. It is also noted in the costing table that the nurses performed mouth care and so maintained personal care (Table 26).

Case study 3: north non-Liverpool Care Pathway-using nursing home

The north non-LCP-using nursing home was actually a LCP-using site by the time the observation was commenced. The observation was undertaken over 4 days and the combined observation time was 17 hours and 20 minutes, after the break periods were removed, and the total additional cost was £137.67. The greatest costs were incurred on the second observation day as a result of more nurse time and drugs being administered. Once again, the lower than expected cost for the time period could have been because of the short observations over a few days and the time of visits undertaken. A point to note in this site was that it was observed on day 2 that prednisolone and cyclizine were administered, but that they could not be found on the drugs charts and were not therefore included in the costing. In addition, on the first day the observer was told that the patient had received diamorphine hydrochloride, but we also could not include this in the cost, as it was not observed. Nurses and carers predominantly undertook the care in this site, but no sisters were employed and no doctors were observed to provide care. The nursing home environment meant that there were no porters, but undertakers were observed to attend to the patient following death.

The costs in relation to the family questionnaire table show that the family rated personal care as the highest and we can see that the nurses/carers attended to mouth care and making sure that the patient was clean and dry (Table 27).

Case study 4: north Liverpool Care Pathway-using intensive care unit

The north ICU LCP-using site was observed over 1 day. This was a relatively short observation, although, perhaps not surprisingly, the costs were greater than those for the nursing home. The difference between the two settings makes comparisons futile. The principal costs were nursing staff and doctors, rather than the lower-band nurses or carers employed in nursing homes. Two different cost levels were attributed to the ICU nurses depending on whether or not they had direct contact with the patient.

The costs depicted in Table 28 shows that the largest component of the total £221.93 cost came from the time that the nursing staff spent with the patient directly, at £152.28, and also the doctor’s time spent with the patient at £48.60. The time spent with the patient by staff appears to coincide with the family member questionnaire results given by the relative, as all of the answers were 10, except for pain, which was not an issue.

Case study 5: south Liverpool Care Pathway-using intensive care unit

The south LCP-using ICU site observation was undertaken over 1 day (as was the case in the north site); this observation was 4 hours and 20 minutes longer, at a total time (after breaks) of 9 hours and 30 minutes. The total additional cost for this site was £306.64. Similar to the north site, the biggest cost driver came from the nursing time spent with the patient, which accounted for £288.36 of the total cost. In this case there was no observed time spent from the nurses without direct contact, or from a doctor. There were porter costs, which were for two porters taking 10 minutes each. In this case, in addition to the drugs prescribed, the patient was given cream for dry skin, mouth swabs, mouth wash and eye drops, which is reflected in the results for personal care and dignity. The responses for pain management and breathlessness were scored lower, but the relative justified this because they did not know what the patient’s level of pain was and so they gave a score of 5. The family member also responded with scores of 10 for spiritual and being kept informed by staff (Table 29).

Case study 6: north non-Liverpool Care Pathway-using intensive care unit

The non-LCP-using ICU site for the north in Table 30 is the shortest observation of all the case studies (3 hours 15 minutes), but despite this the costs are quite substantial. The majority of these costs were derived from two nurses spending 23 minutes each to clean and care for the patient for post-death care. There was only 6 minutes less nursing time spent with the patient that was not direct. In this short period of observed care, the patient was given morphine for pain control. The results from the family tool show that the family thought pain management was high, but were unsure as how best to answer. The relative did not rate personal care during the patient’s end-of-life stage as high, scoring only 6, although we do not know why this is. The remaining results were varied, and owing to the short observation period we cannot be certain of the results in line with the costing. It is important to note that in the absence of attending relatives the nurse remained present at the patient’s bedside until the patient died.

Case study 7: south non-Liverpool Care Pathway-using intensive care unit

The final non-LCP-using ICU case study, in contrast to the north non-LCP-using site, was relatively long. Conducted over 1 day, the observation period is 10 hours and 40 minutes after breaks were removed. This also had the highest costs of £415.87, attributable to the length of the observational period and the involvement of nursing care. The most interesting point to note, however, that is both the doctor and the nurse spent 47 minutes each with family members to explain the patient’s care. Surprisingly, it is questionable if this is reflected in the family’s scoring of being kept informed by staff, as they rated this aspect of care as only 8, the same as the results given to the other aspects of care that were rated. The relatives were not pleased with care in the last few hours during the period close to the patient’s death (Table 31).

End-of-life care in the nursing homes

The original research design had attempted to match LCP-using and non-LCP-using nursing homes on the basis of a range of criteria in order to assess the impact of the LCP on care. Although, in the main, we were successful in recruiting the sites we had selected on the basis of the predetermined characteristics, as the study progressed it became clear that the distinction between LCP-using and non-LCP-using sites in this setting was blurred. Homes that were reported not to be using the LCP at the time of recruitment were, by the time of data collection, using some or all of its elements. Moreover, as the research progressed only a sample of homes identified deaths for us to observe. Thus, the sample became even more skewed.

The recent drive to reduce the number of hospital deaths6 has reinforced the need to focus attention on improving end-of-life care delivery in the care home setting92 which is currently the second most likely place for people over the age of 65 years to die in the UK. 3 The majority of homes in this study were involved in end-of-life care initiatives (other than the LCP) including the GSF, regional ACP initiatives and the North West End of Life Care Six Steps to Success programme for care homes. 43 Therefore, the environment in which the observations took place was, on paper at least, equipped to undertake end-of-life care using established frameworks either singularly or together.

The patients whose deaths were observed in the nursing homes in this study were extremely elderly. All patients but one were in their eighties or nineties, and there was an expectation by staff and relatives that, in the context of these patients’ declining health, death was a certainty.

End-of-life care in intensive care units

In order to understand the complexity of care at the end of life in the ICU it is important to draw on all the data relating to the observations of care. This is a complex world, in which people are admitted on the basis of some chance of recovery, care is provided on a one-to-one basis and the environment is textured by the latest lifesaving technology and machinery. The patient can be stripped of their personhood, through either the use of sedatives or the severity of their condition, and their body is transformed by the lines and tubes that attach it to the ICU life-support equipment. Relatives entering this world are often highly stressed and anxious108–110 and their thoughts and understandings can be disordered. 111 Even where the admission has resulted from a long-standing deterioration in health, and death is recognised as a possibility, many relatives retain hopes of recovery.

In this study, the patients whose deaths were observed were being cared for by staff in five LCP-using sites and three non-LCP-using sites and, where possible, the following discussion will take account of similarities and differences in the care delivered.

Researchers27,28 have described a ‘good death’ in ICU as one where a consensual decision (between staff and between staff and patients and relatives) has been achieved regarding the futility of ongoing treatment; where death is expected and prepared for; where death is neither prolonged nor rushed; where symptoms are managed effectively; where death is natural (free from technology) and conforms to the wishes of the relatives and any previously expressed wishes of patients; and where death is in the presence of family and is characterised by clear and non-conflicting communication. In the main the deaths observed in this study met the criteria outlined above for a ‘good death’, regardless of whether or not the dying was supported by the LCP. Indeed, in line with a study conducted by Fridh et al. ,112 the relatives of 15 patients who had interventions removed and died on the ICU in our study reported that the patient had been treated as a person with their dignity and care maintained right up to and including their death. In two cases in which the patient was transferred out of the ICU once interventions had been withdrawn, the relatives were extremely unhappy about the care that was subsequently provided.

Care of people dying in the generalist settings of nursing homes and intensive care units

On the surface, it is hard to imagine two more diverse health-care settings in which death could routinely be expected to occur. At one end of the spectrum, the nursing homes recruited to this study attempted to provide a ‘home from home’ for their residents, and those who died in these settings were among the ‘oldest old’. Here, death was an expected outcome following a lengthy period of increasing frailty and decline. Much of the ‘hands-on’ care in this setting was delivered by non-qualified staff. Staffing levels were often challenging and medical input was basic and relatively infrequent. At the other end was intensive care, where the focus was on cure and rehabilitation, and the latest life-saving equipment and interventions were available and utilised. These sites were characterised by high levels of medical and nursing expertise and input and the patients dying in this setting were likely to be considerably younger than those dying in nursing homes. While some deaths occurred after a long-standing disease, others were the result of trauma or injury.

There were a small number of issues common to both settings. Some of these issues were often solved in slightly different ways.

Economic evaluation

The economic evaluation was based on the observations of care, analysis of the drug charts and completion of the economic tool by bereaved relatives at the end of the interview conducted by the researchers. This tool, in conjunction with the observational and resource data, provided a strategy for calculating comparative costs and benefits experienced by patients supported in LCP-using and non-LCP-using units in the last hours of life.

The results obtained should be interpreted as being indicative rather than definitive as despite attempts to ‘match’ service providers the comparison inevitably contained a wide range of confounding factors. Many ‘non-LCP-using’ units were found to utilise aspects of the LCP process, and ICUs using the LCP to support care varied in their application of it.

In both settings, relatives were able to discriminate between the different aspects of care provided; so, for instance, different scores were allocated to pain and agitation by a relative. Relatives found completing the tool to be a useful exercise and welcomed the opportunity to justify outlying scores, whether they were high or low. The seven case study analyses demonstrated that it was possible to attribute costs to end-of-life care, but in the context of this study the costings could not be used to compare the costs of care provided using the LCP versus standard care.

The methods

This study was designed to use a number of data sources to examine care at the end of life in the two different settings. Data were collected on care at the end of life in general as well as in relation to the individual patient. In each site, interviews with a range of staff were undertaken (identified in this report as point 1 interviews) about end-of-life care in general, and the case notes from 10 recent patient deaths were analysed to gain an understanding of end-of-life care in these sites. At the patient level, observations of 25 patients in the period up to death were undertaken and combined with the views of staff (referred to as point 2 interviews in this report) and relatives to provide a more nuanced understanding care at the end of life. This also included analysis of the written case notes relating to the patient’s care.

Three distinct methods were therefore used at different times and in different ways to bring together an understanding of the landscape of end-of-life care in these two settings. None of the individual methods used was unproblematic and all had their inherent limitations, as discussed below.

Nursing homes

• Documentation of key elements of communication: the retrospective case note analysis of deaths revealed that very little, if indeed anything, had been recorded in the patient’s notes regarding the communication that staff had had with the patient’s relatives. Maintaining good records of these conversations is important to ensure that all staff are aware of what relatives have been told and to assist in the delivery of information in a consistent and accurate way.

• Consistency and availability of policies to guide care: when staff were asked to produce policies relevant to the care they provided they found it difficult, more so in nursing homes than in ICUs. Policies that were produced varied enormously both in content and in date. It was hard to assess how staff viewed the policies in relation to the care they provided. Clearly, in the provision of end-of-life care there needs to be an up-to-date and coherent approach, which should be identifiable in appropriate policy documents.

• Lack of distinction between appropriate levels of care and routinised care: although patients in the nursing home setting received frequent visits from staff, these visits were mostly for the purposes of monitoring the patient or to provide physical care. The imperative to relieve pressure areas or provide food very close to death did not appear to be critically assessed by staff. Very few discussions were observed where staff decided not to do something on the basis that it was not appropriate. Given the uncertainty over the timing of death it is possible that staff undertook activities because they simply did not know how long the patient was going to live. Achieving a balance between providing routine care and appropriate care is difficult but the judgements underpinning this distinction may be influenced by training.

• The focus on the delivery of physical care: while physical care received a lot of attention by nursing home staff, less attention was paid to the patient’s physical environment. The patient’s experiences of noise, light and temperature were frequently not considered, and often seemed to reflect the needs of the staff. This was particularly the case in relation to the playing of background music.

• Lack of attention to the assessment of consciousness and swallowing: in the nursing home setting there were occasions on which a more formal set of assessment procedures would have improved care. These related to the assessment of the swallowing reflex and the patient’s state of consciousness.

• The pacing of ACP discussions: the trend towards establishing a patient’s advance wishes regarding their care at the end of life is well established. However, in this study the speed at which staff appeared to want to establish these wishes was in some cases reported to be unseemly. While recognising the importance of establishing the patient’s wishes, it should also be recognised that the pace of these conversations should be individually determined.

• Reported lack of consistency of GP cover: the reported involvement of the GP in end-of-life care in this study is of concern. A number of consequences flowed from the uncertainty reported by nursing staff of whether or not and when a GP might attend the patient. New meaning was given to the notion of anticipatory prescribing, as nursing staff sought to ensure that a range of drugs designed to make a patient more comfortable at the end of life were available for use. In many cases these were reported to have been prescribed well in advance of need and were often not used. This is clearly a cost which with improved medical cover might be avoided.

• Systematic delivery of end-of-life care training: finally, it was apparent in all the interviews in the nursing home that although training for end-of-life care was recognised as important, it was rarely delivered in a systematic, rigorous way for all staff across each of the organisations involved in the study. Unless all staff receive high-quality ongoing training, none of the above will change.

Intensive care units

• Lack of practical information for relatives and carers: the one-to-one care in ICU was perceived by relatives to be of an extremely high quality. Nurses carried out the majority of care once a decision to withdraw treatment had been made. Although the communication between staff and the patient’s relatives about the patient was consistent, on a few occasions relatives reported lacking vital information about purchasing food and hot drinks. Relatives also talked about their car parking concerns and keeping the parking meter topped up. A general reluctance to leave the bedside compounded these issues. More often than not this information was provided, but, where it was not, the relatives identified the stress it produced, over and above that they were already experiencing.

• Distraction for relatives of the monitor at the bedside after the withdrawal of life-saving interventions: The presence of the monitor at the patient’s bedside was for some relatives a distraction. Relatives found it compelling watching although it took their focus off the patient. In many sites, but not all, the monitor at the bedside was either turned off or rotated out of the relatives’ sight.

• The impact of staff–relative divergent views on care and treatment: while there were few cases of staff–relative divergent views over a patient’s care and treatment, when these did occur the consequences were far reaching. The divergences in this study occurred as a result of the staff’s inability to understand the relatives’ perspective and the relatives’ inability to understand the goals of care being provided. Despite acknowledging this difference, no attempt was made by staff to resolve it. There were also many cases where family members disagreed with each other. This was not always reflected in discussions between the clinical team or in the notes where the tendency was to refer to the family as a single entity.

Terms of Reference