Identifying features associated with higher-quality hospital care and shorter length of admission for people with dementia: a mixed-methods study

Care environment

Lack of privacy, including the impact of noise, was regarded by service users and their families as a significant issue that compromised patients’ basic care on occasions, as well as their sense of identity and self-esteem. 51 Organised activities were regarded as contributing to the creation of a homely and comfortable environment conducive to quality care. 51,57 Basic safety and cleanliness, as well as other inexpensive but effective touches, were identified as important in making the patient feel more comfortable, which, in turn, were believed to ameliorate fear and anxiety. 57

Cultures of care

The frenetic nature of activity on the ward, combined with a sense of professional abnegation of responsibility, were viewed as barriers to assessing the mental health needs of the patients. 52 Staff felt as if there was insufficient time to care for patients, resulting in a focus on task management.

Patients were often excluded from decision-making, which, aside from leading to inaccurate judgement calls in relation to patients’ needs, led to a fractured caring relationship with professionals. 54

Attitudes

The resilience of staff to manage their workloads was significantly undermined by the lack of organisational training, access to pastoral and professional support, and time for emotional processing. 56 When person-centred training had been offered to staff, positive changes in approach to work were observed, in terms of a professional’s approach both to their work and to the patients. 59

Challenges for people with dementia as an acute patient

Lack of consultation with patients often made them feel frustrated, which had an impact on their ability to function in a non-aggressive way on the ward. 54 Patients often coped better when they were able to access family support, instead of relying entirely on staff, particularly in circumstances when the family felt welcomed and were provided with feedback as to how to manage the care needs and welfare of the patient. 57 Patients often felt insecure without family support and missed them. 39

Challenges for carers

Carers’ dissatisfaction was found to be mostly associated with discharge planning, clinical management and poor communication. When patients were treated on specialist medical and mental health units, carers were less dissatisfied. 58 Carers who were allowed to become more involved with their loved one while they were a patient helped bridge the difficulties in communication about welfare needs between patient and staff. This was even more beneficial when visiting hours were flexible and staff were not rotating frequently. Greater consistency in staffing was viewed as improving rapport between patients and staff. 57 When carers were not involved appropriately, this led to significant strains, especially around discharge planning. Care diaries, family meetings and routine engagement by staff (including consultants overseeing a patient’s care)39 with family members were all considered helpful and inexpensive methods to achieve a joined-up experience of care for patients and their family. 52

Challenges for staff

Nursing staff on surgical wards reported that providing care for patients with dementia was demanding and that they felt that they did not have the skills needed to effectively manage challenging behaviour. 53 The importance of assessing capacity and knowing how to manage care in consultation with the patient was highlighted in the research undertaken by Ibrahim et al. 54 The ability of staff to identify and manage delirium was considered a barrier to the assessment of capacity. 49,50 Staff reported wanting to have better information about what community-based services they can involve or refer people on to. 50 In their analysis of qualitative comments from front-line staff working with people with dementia, Banks et al. 49 reported that some staff had positive experiences of using patient-held documents. These documents summarise a person’s history, cultural and family background, and their preferences.

Service models

Specialist nurse involvement improves patient experience and outcomes, in terms of both cost and delivery of specialist care. Specialist nurses can also assist in training other staff and thus enhance the general levels of skills and care afforded to patients with dementia. 55 The dementia champions training programme was also assessed to lead to positive changes in confidence, aptitude and work satisfaction in staff dealing with patients who have dementia. 56

Unit champions were considered to promote the welfare of patients, particularly if they had a strong interest in gerontology, and were prepared to act in a mediation-type role between staff, patients, family support and administrative leadership. 59

Specialist units were found to provide more consistent and enduring attention to a patient’s mental health needs, despite a need to concentrate more time to basic physical health care,51 which was enhanced even more so if the patient was cared for on a specialist mental health ward. 52

The results of the scoping review of the literature provided additional evidence to support the roles that the ward environment, staffing levels and staff training, and supervision and support for staff and carer involvement play in the quality of inpatient care that people with dementia receive. We also found, largely qualitative, evidence supporting the deployment of specialist nurses and patient-held records. The results of the scoping review led to our decision to include the deployment of specialist staff and the availability of personal information as items in our secondary analysis of quantitative data from the NAD. We also enquired about these elements of care in the subsequent qualitative case studies.

Study setting and sample

The setting for the study was acute hospitals in England and Wales. We analysed data from all acute hospitals that took part in the third round of the NAD. Overall, 203 hospitals were asked to take part in the audit and 200 (98.5%) participated. Each hospital that took part in the audit was asked to submit anonymised data on a consecutive sample of at least 50 people (and a maximum of 100 people) admitted to the hospital from 1 April 2016. To be eligible to take part in the audit, patients had to have been given a diagnosis of dementia and to have been in hospital for ≥ 72 hours. Case notes were identified by hospitals using a list of International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, codes for dementia. When patients had more than one admission during the data collection period, data from the first admission only were used.

Data collection methods

Each hospital that took part in the audit was asked to collect and submit five different sets of data:

1. a hospital-level organisational checklist

2. a retrospective case note audit of at least 50 sets of patient records

3. an examination of the availability of personal information on 10 people with dementia

4. a survey of carers of people with dementia

5. a staff survey.

Data from the first four sets of data were included in this analysis. Data from the staff survey were not included.

A senior member of staff at each hospital was asked to complete the organisational checklist. The checklist consisted of 42 questions on topics including collection and reporting of data to the executive board; staff training; carer engagement; environmental review; the collection of personal information about patients with dementia; and food provision. The audit team asked for the member of staff who completed the organisational checklist to be aware of how senior managers and front-line staff aimed to support people with dementia, and that they consulted with colleagues in the hospital to ensure that the information they provided was accurate. Data were entered by staff in each hospital via a secure online survey portal.

Hospital staff were asked to enter data extracted from clinical records for between 50 and 100 patients with dementia. This approach meant that larger hospitals with greater numbers of admissions could submit more data if they chose to. The case note audit covered 36 items, including the person’s admission, their assessment, use of personal information, care planning and delivery, and the discharge process. Data were submitted by audit leads from each hospital, with input from colleagues from audit departments, junior doctors and dementia champions. Data from patient records were entered by staff in each hospital via a secure online survey portal.

Staff at each participating hospital were also asked to examine the extent to which personal information on people with dementia was available to front-line clinical staff. To establish this, the audit lead at each hospital was asked to select three acute adult wards with the highest admissions of patients with dementia. They were then asked to identify 10 patients with dementia on these wards and to check whether a personal information document (e.g. the Alzheimer’s Society’s This is Me document63) was present at the bedside or in the patients’ clinical notes.

A median of 10 patients were checked per site. Of the patients checked at each site, a mean average of 49% had a personal information document present. This ranged from 0% to 100% of patients checked.

The carer questionnaire was developed specifically for the NAD by staff at the Patient Experience Research Centre at Imperial College London. The questionnaire consisted of 10 items, out of which eight were ranked high in terms of relevance to the care of people with dementia, by a panel of carers. The panel thought that all carers and family members who might visit people with dementia would answer these questions. The questionnaire also incorporated the Friends and Family Test question for validation and comparison,64 and a final question, added at the suggestion of the audit reference group, on support provided by the hospital to the carer. Paper versions of the carer questionnaire were distributed by staff to both paid and family carers visiting patients during June to September 2016. Questionnaires were returned directly to the audit team at the RCPsych. Carers could also complete an online version of the questionnaire, which was publicised via social media and on posters displayed around participating hospitals. The questionnaire did not include personal identifiable information and could not be linked to individual patients. All tools used in the audit were piloted by 10 hospitals in 2015 and changes made prior to the full audit in 2016.

Data for the second national survey of liaison psychiatry services for working-age adults and for older adults in England were collected between 14 January 2015 and 30 April 2015.

The team conducting this survey contacted liaison psychiatry teams at all acute trusts in England that included an emergency department. Data were returned on 179 hospitals in England. 62 The questionnaire included 27 items, covering the composition of teams and the hours that the liaison service covered.

Outcome measures

We set out to examine factors influencing three aspects of the quality of acute care that people with dementia receive: (1) length of admission, (2) carer-rated quality of care and (3) quality of patient assessment. The date of a patient’s admission and discharge from hospital was included in the case note audit. We used this to calculate the length of admission for each patient.

We used scores on carer-rated quality of care for each hospital, which had been calculated for the third round of the NAD. This score was derived from responses carers made to a single item, ‘overall, how would you rate the care received by the person you look after during the hospital stay?’, which could be rated excellent, very good, good, fair or poor. The maximum rating for a hospital was 100 (awarded if all carers rated the hospital as excellent) and the minimum rating for a hospital was 0 (if all carers rated the hospital poor). Details of how carer responses were used to calculate the aggregate hospital score can be found at URL: www.rcpsych.ac.uk/improving-care/ccqi/national-clinical-audits/national-audit-of-dementia/nad-reports-and-resources (accessed 18 March 2020).

We used data from the case note audit records to assess the quality of assessment that each patient received. 44 Each patient was given a score of between 0 and 7, according to whether or not seven components of a high-quality assessment had been conducted (mobility, nutritional status, pressure ulcer risk assessment, continence needs, presence of any pain, a structured assessment of functioning and assessment of delirium). These seven items were selected because they are included in the guidelines produced by the British Geriatrics Society. 65 Failure to assess these aspects of health care have also been shown to be associated with the likelihood of early readmission to hospital of people with dementia. 33,66

Predictor variables

We used a three-stage process to generate a list of predictor variables for work package 1. We selected these variables from those assessed in the dementia audit and survey of psychiatric liaison services. In the first stage, we used data from our scoping review of the literature and from our previous research examining factors associated with the quality of assessment of people with dementia who are admitted to acute hospital beds,44 to draw a long list of items that we judged could influence one or more of our three study outcomes. We presented this long list to the Project Management Group and Stakeholder Reference Group for their views (see Patient and carer involvement). This generated a short list of variables paired with each outcome. Having examined the distribution and frequency of these measures, we noted that in some instances there was very little variation and excluded any predictor variables that were present or absent in > 90% of hospitals. In all instances, variables were excluded from the analysis because they were nearly always present. A list of these variables is provided in Appendix 1. The final list of variables included in the main analysis is presented in Tables 2–4.

Carer-rated quality of care was measured at the hospital level; therefore, all variables needed to be aggregated at hospital level. We were therefore unable to include primary diagnosis and type of ward in this analysis.

Data management and analysis

A number of the predictor variables included large numbers of categories with small numbers of patients in them. To simplify the analysis, we combined some categories prior to data analysis. When possible, we used categories that were developed by the audit team. Case notes included > 100 descriptions of primary diagnosis. These were grouped together in 11 categories, such that myocardial infarction was combined with other vascular conditions, and kidney and urological conditions were grouped together. Ethnicity was not well recorded in the case notes. The majority of the sample were recorded as white or white British (82.1%) and 12.4% of the sample were recorded as ‘other’. Less than 2% of records indicated other ethnicities (black/black British 1.2%, Asian/Asian British 1.9%, Chinese 0.1% and mixed 0.1%). We therefore combined these groups with ‘other’ to create a dichotomous variable white and white British, or black, Asian and minority ethnic (BAME).

We used univariate tests to conduct preliminary analysis of the relationship between dependent variables (length of admission, carer-rated quality of care and quality of assessment of patient needs) and patient-level (age, gender, presenting complaint), ward-level (type of ward) and hospital-level (access to liaison mental health services, deployment of specialist dementia nurses, etc.) predictor variables. Further exploratory analyses were carried out to look at which combinations of these exploratory variables could best explain the variations in dependent variables.

Given the nested structure of the data (patients within hospitals), the final analysis was carried out using hierarchical models. 67,68 Traditional regression methods were not deemed appropriate because they assume independent observations. For instance, measurements taken from patients in the same hospital can no longer be assumed to be independent (i.e. they are correlated). Hierarchical models take this into account to draw valid statistical inferences. 69 Using hierarchical modelling made it possible to compare hospitals in terms of patient outcomes. This was achieved by testing cross-level interactions, which combine the effects of explanatory variables at the patient and hospital level. Interactions between explanatory variables within each level were also tested.

Using this methodology, we combined the information across hospitals with patient-level information to identify predictors that may have an impact on length of admission and quality of assessment of needs. Unlike length of admission and quality of assessment, which were measured at the patient level, carer-rated quality of care was measured at the hospital level. Therefore, for this analysis, variables that were not at the hospital level were aggregated. Multiple linear regression was used for this analysis.

Patient-level scores on quality of assessment were not continuous; they were ordered categorical variables, ranging from 0 (no items completed) to 7 (all items completed). We therefore used ordered logistic regression applied to hierarchical models to analyse these data. Most patients (n = 9260, 91.6%) had a score of ≥ 5 on this measure. We therefore collapsed the seven categories to create a variable with four categories: 0 (zero to four items completed), 1 (five items completed), 2 (six items completed) and 3 (seven items completed).

All data were analysed using statistical packages Stata^® (version 13; StataCorp LP, College Station, TX, USA) and SPSS 23 (IBM Corporation, Armonk, NY, USA).

The outcome measure, length of admission, was skewed so transformation of data on length of admission was required. For the outcome measure, we excluded all those who died during their admission from all analyses, including the sensitivity analyses.

Theory development: a ‘twin-track’ approach

To optimise the contribution of work package 2 to the study, we designed and implemented a ‘twin-track’ approach to theory development. 71

Selection of case study sites

Case study sites were selected from hospitals participating in the NAD. We thought that there was likely to be a complex relationship between service organisation and delivery, experience of care and length of stay, so hospitals were carefully chosen for detailed case study analysis based on their capacity to (1) provide insights into this relationship, and (2) provide a better understanding of what works for whom and under what circumstances.

We had planned to use data from work package 1 to help select the first four case study sites. However, time constraints required us to select the first site before the results of the third round of the NAD had become available. Our selection of the first hospital was based on its performance in previous rounds of the audit and other intelligence, which led us to predict (correctly) that it would score comparatively highly in relation to carer-rated quality of care. Findings from interim analysis of the qualitative data from this first site clearly demonstrated that, despite lack of governance infrastructure, this hospital scored highly on carer-rated quality of care. After discussions with the project working group it was decided that governance score should be chosen as the other main criterion for selecting sites. This would provide the opportunity to explore how and why governance infrastructure might or might not result in desirable patient outcomes and whether or not there are some other organisational and/or contextual factors that lead to high-quality care.

When the audit results subsequently became available, we reviewed those results alongside other data sources, such as Care Quality Commission reports, and undertook telephone interviews with key stakeholders, before selecting three further hospitals. A paired sampling approach was used to help us understand similarities and differences between sites (Table 5).

Members of the Project Management Group and Stakeholder Reference Group informed us about hospitals that had put measures in place to reduce length of stay. We were also interested in collecting data from hospitals that were geographically different from our four main study sites, which were all in London or the South East/West. Selecting two additional study sites in the south-west and north of England allowed us to test our theories in hospitals with different characteristics to the first four study sites.

Our sampling approach resulted in the inclusion of hospitals with differing levels of performance. We define good performance or ‘outcome’ in realist evaluation terms, as comparatively positive carer-rated experiences of care plus shorter admissions. It also resulted in the inclusion of hospitals operating in differing contextual conditions for delivering good-quality acute care for people with dementia, including key components in the organisation and delivery of care identified through the third round of the NAD, such as adequate staffing levels and access to liaison mental health services.

Data sources

At each hospital we conducted in-depth interviews and assembled documentary evidence at three key organisational ‘levels’ relating to the quality of care: (1) clinical governance, (2) middle management and (3) the staff–patient interface.

Data management and analysis

Sensitivity analysis

We conducted two sensitivity analyses in which we restricted the analysis to subsamples of (1) patients who were admitted with hip fractures or trauma and (2) hospitals that scored above the median carer satisfaction ratings. The results of these analyses are presented in Tables 14 and 15.

Among the restricted sample of patients admitted to hospital with hip injuries, fewer associations were found with average length of stay. Patients whose records included documented evidence of discussion of discharge with carers had longer periods of time in hospital, and patients who had documented evidence of discharge planning being initiated within 24 hours of their admissions spent less time in hospital. Shorter lengths of stay were seen among women from BAME communities, but not among men from these communities. We also found evidence that when the hospital deployed dementia nurse specialists, female patients had shorter average periods of time in hospital.

In the subsample of data from 74 hospitals with high carer-rated satisfaction with care, associations between length of stay and primary diagnosis, and length of stay and ward type, were still seen. Among female patients, there was evidence of shorter admissions to hospital among those for whom there was no documented evidence of discussion with carers about the patient’s discharge. When liaison teams included a specialist old age psychiatrist, shorter lengths of stay were seen when discharge planning had been initiated within the first 24 hours and in hospitals in which patients did not have a designated social worker.

Length of stay

The associations found between outcome measure ‘length of stay’ and predictor variables are described as follows. As presented in Chapter 5, in the primary multivariate analyses, these associations were:

• shorter average length of stay and documented evidence that discharge planning is initiated within 24 hours of admission

• shorter stays and executive boards reviewing delayed discharge for people with dementia (the effect is borderline)

• longer stays and carer involvement in discharge planning.

Each of these associations were explored in our interviews. Here, we present the findings and use direct quotes to put these findings into context. To ensure anonymity and identity protection, the participant identifiers accompanying direct quotes will be restricted to the participating site, whether they are carers or staff and participant number (e.g. 3ST012 indicates that the participant is a member of staff from site number three and recorded as participant number 12 in this site).

Carer-rated quality of care

As detailed in work package 1 results (see Chapter 5), we found that hospitals in which there is a strategy for carer involvement in place have lower carer-rated quality of care scores. We found a borderline significant association between hospitals in which the care assessment contained a section dedicated to collecting information from the carer and having lower carer-rated quality-of-care scores. In hospitals with higher proportions of patients and carers from BAME communities, we found that levels of carer-rated quality of care was higher when there was greater availability of personal information about patients in documents, such as This is Me. 63

In our qualitative interviews we sought to explore the reasons why carer satisfaction may be lower in hospitals when there appeared to be more evidence of organisational efforts to involve carers. Factors that might explain this association include raising expectations and not delivering what was promised, whereas early planning, clear communication and consistency were things staff could do to help manage carers’ expectations.

Interviews with front-line staff and managers suggested that carer involvement could also be a source of contention and may lead to certain challenges. In circumstances in which staff obtain information from carers, but do not address their concerns, the collection of such information could adversely impact the carer-rated quality of care, as care will be perceived as not having been delivered up to expectations:

If someone says oh she likes going for walks outside, or she doesn’t like to eat this or that. We can’t take people outside realistically, there is not enough staff to give care on the wards, let alone to take someone out and walk with someone outside. And the food that’s available is limited by what’s on the menu. And now with health and safety there is sometimes governing what people can bring in, etc. So sometimes you can ask for a lot of information, but you can’t necessarily do what you would like to do with it.

4ST006

Interviews with carers also highlighted that collecting information that is not fully followed through could potentially lead to carers’ dissatisfaction with care provision:

Mum tended to eat with that little spoon, don’t like the big one going in, and I said to them can you feed her with the little spoon because she doesn’t like that big one going in. I even got about 20 out of McDonald’s and leave them up there, but when I come up, they still were feeding her with the big silver one.

1CA001

Staff believed that discharge arrangement is an area that is most likely to be the source of family carer dissatisfaction. Although staff make efforts to involve carers in discharge planning, they can only advise other parties involved (e.g. social services) and are not necessarily in control of what happens, which might be perceived as not delivering what was promised:

I think it makes us a little bit tentative, to say that the person can definitely go home with four times a day care, or they can definitely go to a care home, because although we are going to give information to the social worker, we are not the decision-maker in that situation. I think that makes communication slightly more difficult.

2ST007

Some family carers might be seen by staff to be obstructive to the discharge process, as they prefer their family members to stay longer in the hospital for various reasons. Disagreements about time of discharge will lead to carer dissatisfaction.

Carers also commented on the importance of clear communication during their relative’s hospital stay. Carers spoke of the many occasions where they had felt that they were left in the dark by staff about the progress of their relative’s illness, and commented on how they would have appreciated a more structured and frequent communication:

If you see where I’m coming from . . . an update for dementia patients would be brilliant, because you can’t go to your relative and say, what did the doctor say to you today? Because they can’t know . . . they wouldn’t be able to tell you.

3CA001

To minimise subsequent conflicts and manage carers’ expectations, hospitals have put some measures in place. Staff believed that envisaging potential problems early on means that meetings are arranged with the relatives to discuss the discharge plan. Staff were adamant that in these meetings they would make it clear to relatives that although they will continue to address the family’s concerns, hospitals will progress with the discharge, as the plan is based on the offer from the commissioners:

. . . unfortunately, a lot of it is the expectations of carers, and it’s the expectations that we can’t always deliver. Because obviously the system can’t deliver what sometimes, what they want . . . we will offer them all the support to manage that message [discharge plan], but ultimately that is the message that we have to give.

5ST002

Staff believed that, in some cases, patients do not require the level of care their relatives might ask for. Although these disagreements might lead to some very difficult conversations, engaging carers early on and keeping them informed could minimise the contentions.

Lack of consistency was something else that was considered to contribute to carer dissatisfaction. In some hospitals, there are designated members of staff whose role is to liaise with carers and to ensure that all staff members are fully aware of the hospital policy and discharge strategy. Barriers include when hospitals have staff shortages and need to employ agency staff who might not be fully aware of the strategies and discharge plans for any individual patient.

This feedback contrasted with examples that staff gave of when use of documents, such as This is Me,63 which contain details about a person’s background and important events, people and places from their life, had supported the delivery of better care:

And things like how people eat and drink as well, having that written down, because if you are vegetarian and you keep getting delivered meat and you haven’t got the ability to communicate that, then staff just goes, she’s off her food, she’s not eating and drinking. Well actually it’s because you keep giving her the wrong food, or somebody delivers like a really milky cup of tea with no sugar and someone likes it strong with three sugars, we assume it’s because she’s off her food and drink, well actually you just made her a really bad cup of tea. It’s nothing to do with anything else. So, it does help having that written down.

1ST004

Indeed, staff offered us many tangible examples on how ensuring that staff have access to personal information can lead to better-quality care for patients with dementia. For example, in situations in which a patient is distressed and might mention a family member, if staff are aware who and what the patient is talking about, they can comfort the patient and reassure them. Staff believed that having this information not only reassures patients and calms them, but also lowers the level of anxiety in staff who otherwise would not know how to react and what to do to calm the patient:

We are using This is Me. Some people I’ve spoken to have said, God, this is amazing because my dad gets car sick and, there’s never a space to tell anybody that. Because he gets transport home, because there might be an issue, and I can actually write it down. But then again, if you’ve done it all and then people are not taking any notice of it that must be really annoying.

2ST001

Staff training

One of the recurrent themes from the interviews with staff was the importance of dementia-specific training for staff who provide care for patients with dementia. Staff members unanimously believed that dementia training plays an important role in ensuring good-quality care. Staff education in dementia care has the potential to make a significant impact on the quality of dementia care provision, as it enables staff to realise that they can be flexible in their approach to deliver care in a more individually appropriate manner.

We now present some CMO configurations that explain how, why and under what circumstances dementia-specific training can lead to higher-quality care and shorter inpatient stays.

Carer involvement

Carer involvement comprises various strategies that hospitals might adopt to actively involve family carers in different stages of patient stay in acute settings, from admission through to discharge planning. Overall, staff believed that having family carers on board will lead to providing better quality of care for patients with dementia. Presence of family carers allows patients to settle, as they are better able to make patients feel at home by following the routines the patients are used to.

Staff also spoke about difficulties when there is no family carer or relative present who can advocate for the patient and be consulted on various issues. Lack of family carers also means that hospitals need to organise more social care for patients on discharge.

In the subsequent series of CMO configurations, we elaborate on how, and why, carer involvement strategies might lead to positive outcomes.

Culture and attitudes

An organisational culture that encourages and supports staff to be flexible in delivering patient-centred care fosters a certain degree of autonomy. This provides staff with a degree of freedom whereby they can use their discretion and show flexibility in applying the ‘rules’, which leads to patients receiving more individualised and dignified care.

In better-performing hospitals, senior clinical staff are seen and present on the wards, and encourage more junior staff to ask questions when dealing with complex scenarios. Senior staff champion a culture of ‘it is OK not to know everything’ about dementia and encourage front-line staff to ask for help or second medical opinions. Furthermore, in situations when things go wrong, there is a culture in which the responsibility is shared, rather than adopting a ‘punitive approach’. This means that mistakes are learned from and used to improve care provision:

I think it’s a really fine line . . . what I’ve tried to do is foster a way of working that when things go wrong we ask ourselves the difficult questions, but when actually, if I found that I was in that position I would’ve made exactly the same decision I will say it, regardless of what other people think we should be saying. So, actually, I remember we had a pressure ulcer on one of our older people’s wards and I went through and said I can see you’ve done everything, actually, is there anything else you think they could’ve done that could’ve protected this patient? No. Well, then I think this is unavoidable. And we need to recognise that. You carry on doing what you’re doing. Don’t be disheartened. I can’t see what else you would’ve done. Because, there can be a culture of fear where people think, like, just because something goes wrong that somebody’s made a mistake. Sometimes things go wrong. As long as we’re honest about that and we’re challenging, and we don’t accept the unacceptable, then we can create a culture that says OK, we will have a difficult conversation, but at the end of that difficult conversation, if we’ve done the right thing, then I will stand behind you, 100%.

1ST015

Staff believed that people choose nursing because they genuinely care about others. However, constant heavy workloads might lead people to temporarily lose touch with the human side of their jobs and become very ‘task orientated’ to achieve the targets:

You see different people every day. You come across one or two who have genuine empathy, I suppose is the best word, others who . . . I don’t know if it’s they’re rather more concerned with the hours they have to work. It is difficult, I do sympathise greatly with a lot of their situation, but nevertheless, this is a very special place for people, and they know that emotions are generally high, and people are very much more concerned than they might be in a different situation.

2CA001

In hospitals with high ratings on quality of care, we found evidence that there is a culture that leaders are frequently seen on the wards and can identify when staff find it overwhelming to provide ‘compassionate care’. Staff had different views on whether or not compassion is innate, and if and to what extent staff can be taught how to practise compassion or dignity. However, most agreed that there are simple measures that can be incorporated into everyday practice that do not require any extra resources. For example, language used on a ward has a huge impact on how staff relate to patients. Using patients’ first names in staff meetings and ward rounds instead of referring to them by their diagnosis or their characteristic behaviour (e.g. ‘Mr Wanderer’) makes staff more aware and in touch with the human side of their jobs; and is a reminder for staff that there is a person behind the disease who needs care. Furthermore, it helped if senior staff champion a culture that encourages more junior staff to ask questions when dealing with complex scenarios:

Ideally, anybody with dementia, regardless of what’s wrong with them, should go to a dementia ward, but that isn’t what happens. Because then the team that works on that ward, even if they don’t know your particular individual, have a far better understanding of dementia. But sometimes it was just basic compassion that was missing, because if somebody is making no sense when they talk to you, and they look ancient when you look at them, then even if you don’t know for sure they’ve got dementia, it’s quite a, you know, a logical conclusion to come to.

3CA004

Carers who we interviewed also pointed out the potential difficulties of communication with a patient with dementia and emphasised the significance of compassionate care when dealing with patients with dementia:

So, you shouldn’t be taking what a person with dementia says on face value, . . . And I know they have to explain what they’re doing, but it would be, right, sir, we’re going to reinsert your catheter for you, right? Firstly, he doesn’t know what reinsert means. Secondly, he doesn’t know what a catheter is. Thirdly, he doesn’t know where you’re going to put it. So, he’d sort of nod at you, and say, yes, so they think, right, he knows what we’re talking about – in they’d go, and he’d scream the place down . . . I had one doctor, registrar, describing him as aggressive – I said, no, my dad’s not aggressive, never has been aggressive, he’s the most placid person you would ever want to meet. I said, what he was, was frightened. I said, if you were doing that to a two-year-old, and they started screaming, would you call them aggressive? You wouldn’t, you’d just say, poor little frightened child. And that’s exactly what he was at that point. And I think it’s just a lack of understanding, that . . . I can understand that they might not know exactly how to handle each individual, but we were there, saying he’s got advanced dementia, that wasn’t good enough really . . .

3CA004

Resources (staffing, recruitment and retention)

Adequate staffing level was mentioned by many staff as one of the most important factors in enabling staff to provide high-quality care. Some were explicit about the challenges that they face in delivering high-quality care when there are staff shortages:

If I’m honest, we have a recruitment problem. So, to really understand what the care needs are we need to spend quite a lot of time with the individual patients and we probably haven’t got the time to do that because the turnover of patients is quite fast . . . I mean, part of it is taking time to discuss if there are relatives and friends coming in. We’ll eventually use patient passports and that obviously will help, but I’m not sure how frequently they’re completing the individual patient passports, if I’m honest.

2ST010

It emerged that in hospitals with higher scores on carer-rated quality of care, specific attention is paid to the recruitment process, in which challenges and details about working with patients with dementia is explicitly discussed in the interviews with candidates for nursing jobs. Staff believed that these strategies would increase the likelihood of recruiting people who are willing and ready to work with patients with dementia. Awareness of potential difficulties means that they are more likely to stay and carry on working on the wards. This was further supported when staff spoke about problems with temporary workers:

One of the barriers for us is our staffing levels . . . I never want to make that an excuse, but I can’t deny it. We’ve got up to about 30% vacancy of nurses on our wards. So, we’re using an awful lot of agency and temporary staff. So, are they consistent with your processes and your pathways? No! If you’ve only got one trained substantive member and you’ve got at best two other trained, you’re really struggling on a ward of 26. So, for me, there’s something around temporary staffing use and making sure that they’re aware of what our expectations of them are . . .

3ST013

It also emerged that in better-performing hospitals there is a culture that focuses on individual staff members’ strengths and this would be considered when allocating more sensitive tasks, such as one-to-one patient observation, so that tasks are fitted to personalities. In such a culture, senior staff recognise and thank front-line staff for their hard work during challenging times. Not surprisingly, hospitals in which staff feel valued and supported have a good record of staff retention, which leads to patients receiving more consistent and personalised care:

I mean a lot of our nurses here we’re lucky they have stayed here for quite a long time now and, you know, we are quite good at retaining our staff. I think there’s something about valuing your staff and investing in them as well and making sure that they are supported, and they know they can turn to you when things are difficult. And that’s kind of coming from leadership I think really.

1ST007

Discharge planning

The factor that had the greatest bearing on length of stay was whether or not discharge planning was initiated within 24 hours of admission. Discharge planning involved setting a planned discharge date, identifying tasks that needed be taken to achieve the planned discharge date and assigning staff who would be responsible for working to complete these tasks. Although it has long been recognised that early discharge planning is helpful,79,80 our analysis of data from the audit allowed us to quantify this, and the qualitative data we analysed suggested factors that support and impede effective discharge planning for people with dementia. A coefficient of –0.24 is equivalent to an average length of stay that is 27% lower among people whose discharge planning was initiated within 24 hours of their admission to hospital. The association in the subsample of people admitted with hip fracture and associated problems was greater, equivalent to a 46% reduction in length of stay. With an average length of stay of 12 days, this amounts to about 3 days per patient across the group as a whole and 5.5 days per patient admitted with hip fracture. With > 100,000 people with dementia admitted to hospital per year, the impact of increasing the use of early discharge planning for people with dementia is considerable. In some instances, we found that staff believed that early discharge planning would not be needed because the patient’s needs were not sufficiently complex. In other instances, staff felt that the patient’s presentation was too complex to make early discharge planning a useful exercise. In wards in which early discharge planning appeared to be most effective, named staff were allocated the task of keeping track of the discharge planning process and there were established processes for raising concerns with more senior staff when plans were off track, so that additional resources could be provided. The results of this study suggest that more could be done to ensure that discharge planning for inpatients with dementia is initiated within 24 hours of the person’s admission to hospital.

Reviewing delayed discharges among people with dementia

Although most trust boards routinely review delayed discharges, data from the NAD suggest that only one-third of trusts specifically review delayed discharges among people with dementia. 8 The association we found between length of stay and trust board reviews was not statistically significant. However, a strong trend emerged with people with dementia spending, on average, 1 day less in those hospitals in which trust boards regularly reviewed delayed discharges of people with dementia. In our sensitivity analysis of patients admitted with hip fractures and related conditions, the association between trust boards reviewing delayed discharges and shorter length of stay attained statistical significance for older patients. Qualitative data from senior managers and front-line staff suggested that the response that trust boards make to this process is important, with staff reporting greatest benefit from trusts that use data to identify gaps in the provision of community services and have discussions with local commissioners to try to ensure that these are filled. With plans in England for strengthening co-operation between health and social care services via the development of sustainability and transformation plans, and the development of accountable care organisations, the value of reviewing delayed discharges among people with dementia is likely to increase. 81 In the meantime, an increase in the proportion of trusts that specifically review delayed discharge among people with dementia could help shorten the amount of time that these patients spend in hospital, and having someone in this meeting who understood what happened on the ward was seen as important.

Involving carers of people with dementia

Interviews with staff and carers of people with dementia highlighted the important role that the presence of carers on wards can play in helping patients feel more settled, while also supporting nutrition, hydration and adherence to treatment. In this way, carers can help reduce the incidence of challenging behaviour and may be able to minimise the likelihood of complications arising, which can increase the length of the patient’s admission. 82 In discussions with members of the Stakeholder Reference Group, carers highlighted practical barriers that reduce the presence of carers on the wards, including restrictions on visiting times, parking fees and the availability of alternative forms of transport, especially for elderly carers.

Carers of people with dementia also play an important role in helping to ensure that hospital staff are aware of the patient’s needs. Communication with carers needs to take place at an early stage of the admission. Our finding that patients whose records included evidence of discussion with carers about discharge spent longer in hospital supports the views of both carers and staff, who told us that, if communication with carers does not start until discharge is being arranged, then new concerns, which could have been addressed earlier in the patient’s admission, may be raised.

The collection of information from carers seems unlikely, by itself, to improve the quality of care that people with dementia receive in hospital. Indeed, analysis of data from the audit suggested that hospitals had lower levels of carer-rated quality of care when a greater proportion of case notes had information recorded following discussions with carers. Data collected from carers and feedback from our Stakeholder Reference Group highlighted difficulties that can arise when carers are asked about patient needs or preferences but then find that these are not met. When discussion with carers starts soon after the admission to hospital, there is greater opportunity for these to be acted on. There is also the potential for needs or concerns to be raised, which staff in hospitals are unable to meet. For instance, acute hospitals may not be able to reverse reductions in health and functioning that are a result of health problems precipitating the admission to hospital. Early discussions with carers may help to ensure that carers have realistic expectations about what can and cannot be achieved during the hospital admission. 19 Some hospital staff may also have existing beliefs about what carers want and have concerns that this could slow down the discharge process. However, interviews with carers and staff suggest that many carers are aware of the reasons why prolonged lengths of stay of people with dementia should be avoided, and that other carers will better appreciate the need for a timely discharge when these issues are discussed.

Our finding of lower levels of carer-rated quality of care at hospitals that reported having a carer strategy in place was unexpected. In response to this unexpected finding, members of the audit team conducted a simple univariate analysis of the relationship between carer-rated quality of care and whether or not hospitals reported having a carer strategy over all four rounds of the audit. This confirmed that in each of the first three rounds (2010/11, 2012/13 and 2015/16), hospitals with a carer strategy had lower levels of carer-rated quality of care. However, in the fourth round of the audit, for which data were collected in April–May 2018, the relationship changed, with hospitals with a carer strategy in place having higher levels of carer-rated quality of care. We do not know the reasons for this change. It is possible that there was a delay between the development of a carer strategy and implementing the strategy in practice. Therefore, any positive impact on carer-rated quality of care would have taken time to appear. An alternative explanation is that hospitals that were the first to develop carer strategies did so in response to negative feedback from carers.

Availability of personal information

We found some evidence from our analysis of audit data that the degree to which wards ensured that staff had access to personal documents, such as This is Me,63 was associated with improved carer-rated experience of care. This attained statistical significance in hospitals that had a greater proportion of patients and carers from BAME backgrounds. Interviews with front-line staff generated multiple examples of when access to personal information had enabled them to comfort an anxious or distressed patient. If efforts to reduced distress and agitation are successful, they have the potential to reduce the amount of time that people with dementia stay in hospital. 83 Staff also highlighted the role that access to personal information plays in helping colleagues feel more confident that there are things they can do or say that may help calm a patient if they become agitated.

These documents were developed in response to evidence highlighting the value that people with dementia place on a personal approach to their care. 84,85 There is evidence from the NAD that hospitals have successfully taken steps to improve staff access to personal information;8 however, there is still marked variation in practice across hospitals and between wards in the same hospital. Our data emphasise the importance of making these documents a central component of the resources that staff use to help to improve the quality of care that patients with dementia receive.

Specialist staff and teams

One of the key areas of enquiry in this study was the role that specialist staff might play in improving the quality of care that people with dementia receive. In the third round of the NAD, > 80% of hospitals reported that they employed a dementia champion at a directorate level and just under one-third reported that, in keeping with Royal College of Nursing guidance, they employed at least one full-time dementia specialist nurse for every 300 admissions of people with dementia per year. 86 Data from the second annual survey of liaison psychiatry in England62 showed that just over half (53%) of mental health liaison teams were providing a 24-hour service and that 49% had a specialist in old age psychiatry as part of their team. Accounts of discussions with specialist staff did not feature widely in the qualitative interviews we conducted with managers and front-line clinicians, and were absent from interviews with carers of people with dementia. However, indications of the impact of specialist staff were found in our secondary analysis of audit data. In the main multivariate analysis of factors associated with length of stay, patients generally had shorter admissions in hospitals in which the mental health liaison team included a specialist old age psychiatrist. This association reached statistical significance for patients for whom there was no record of discussions about discharge with a medical consultant. In the subanalysis of factors associated with length of stay in the hospitals that had higher ratings for carer satisfaction, we found evidence of lower average lengths of stay in hospitals that had teams with an old age psychiatrist when there was no record of discussions about discharge with a social worker. In the subanalysis of factors associated with length of stay in the hospitals among people treated for hip fractures and related complaints, we found evidence of lower average lengths of stay among female patients in hospitals that employed higher numbers of dementia specialist nurses. We also found some evidence that quality of assessments of patients with dementia was higher in hospitals that deployed at least one dementia specialist nurse per every 300 admissions of people with dementia per year. Although these data provide some evidence to support the role that specialist staff play in delivering high-quality inpatient care to people with dementia,28 previous studies have reported that most people with dementia do not have direct input from specialist staff. 87 The high prevalence of dementia among people in general hospitals highlights the need for all staff to develop expertise in the assessment and treatment of people with dementia. 88 In a recent systematic review of effective dementia training programmes, authors highlight (among other factors) the importance of supporting staff in applying taught materials in clinical practice and providing dementia-specific guidelines. 89

Hospital culture

A central finding from the qualitative interviews we conducted was how the culture of hospitals can have an impact on the quality of care that people with dementia receive. Changing the culture of any organisation is a considerable task that can take time to achieve. 90,91 It is also influenced by resources and may prove especially challenging at times of financial restraint. We have highlighted some of the factors that appear to have a bearing on hospital culture in Chapter 6, Organisational and contextual factors, including attitudes, staffing, recruitment and retention. Although staffing levels require funding, other components, such as adopting a flexible approach and encouraging delivery of compassionate and person-centred care, may be possible with existing resources. The role of ward managers and other senior staff appears to be particularly important in creating an environment in which solely ‘task-oriented’ practice is discouraged and front-line staff are encouraged to use a flexible approach to delivering patient-centred care. Identifying individual strengths in staff, allocating tasks that are compatible with their abilities and acknowledging staff for their hard work are among factors that lead to higher rates of staff retention. Previous research has highlighted that hospitals in which staff are given the permission to adapt clinical care to patient needs and in which person-centred care is encouraged by senior managers, can provide better dementia care. 92

Reflections on patient and carer involvement

This study benefited from the contribution of people with dementia and carers of people with dementia. As active members of both the study’s Project Management Group and the Stakeholder Reference Group, people with lived experience helped determine the content and focus of the study, and the interpretation of study findings. By working with the Steering Group for the NAD, we were able to obtain helpful feedback from a range of people with mild dementia and from carers of people with dementia throughout the study. One of the advantages of working with this existing group was that they were already familiar with the data we were analysing, and they were keen to develop learning from the audit process to which they had already contributed. As a well-established group that has a culture of actively valuing contributions from people with lived experience, patient and carer members of the group had the confidence to comment on a question plan for the study and how the data would be interpreted. For instance, the initial response of some service providers to our finding of an inverse association between the collection of information from carers and lower satisfaction with carer, was that hospitals that promoted this may have done so in response to negative feedback from carers. However, carer members of the group raised the possibility that collecting information from carers could increase expectations and lead to dissatisfaction of the carer if the suggestions were not acted on.

Dementia is a progressive condition and some of the patients involved in this study at the start were not able to contribute throughout the study period. However, because the Steering Group for the NAD had strong links with a range of voluntary sector organisations, we were able to ensure that people with lived experience contributed to each stage of the study, even if people attending each meeting varied.

During the course of this study, a number of clinical and academic staff that were involved had personal experiences of supporting a relative or friend with dementia during an inpatient admission. These experiences will also have influenced the contributions that people made to the study and support the notion that dementia is, increasingly, ‘everybody’s business’. 95

Reducing the need for inpatient admission

Although there are a variety of steps that staff at acute hospitals can take to improve the quality of care and decrease the length of stay of people with dementia, their implementation will not overcome the inherent challenges associated with caring for people with memory problems in an unfamiliar environment, surrounded by people who do not know them well. During the course of this research, we became aware of initiatives that acute hospitals are developing to try to prevent admission to hospital of people with dementia. These include steps taken in emergency departments to divert people from an admission, by strengthening community-based support, services and technologies that aim to treat acute conditions in a person’s home. Future research could identify and describe the CMOs of policies and practices that support the delivery of acute care for people with dementia in the community, and avoid an admission to hospital.

Ethnic minority differences in acute care for people with dementia

Our findings that both the length of stay and the quality of assessment of people with dementia from ethnic minority backgrounds differed from those of other patients was unexpected. The qualitative data we collected were not sufficient to fully explain reasons for these differences. Demographic changes in Britain mean that the numbers of people from BAME communities with dementia will rise in the coming years and steps need to be taken to ensure that all patients with dementia receive high-quality inpatient care. Further research using both qualitative and quantitative methods could examine associations between ethnicity and the quality of inpatient care that people with dementia receive.

Carer involvement

Not all people admitted to hospital have an informal carer who is able to support the care of a person with dementia during their admission to hospital, but a substantial minority do. There is increasing recognition of the support that carers provide to people with dementia during an admission to hospital, but there has been little research into how the role of carers on wards can be strengthened and supported. The age, occupation, health and proximity of carers have an impact on their ability to provide care for their loved ones. It is important that future research considers how carers who are able and willing to contribute to the acute care of people with dementia can be better supported. This could include training and support for carers to co-produce aspects of inpatient care. Our finding of a negative correlation between having a carer strategy and carer satisfaction was not fully explicated. We believe that this would benefit from further research, with a focus on the relationship between intention, execution and organisational culture.

In summary, we recommend future research examines the following areas:

• Identifying the CMOs of policies to deliver acute care for people with dementia in the community.

• Exploring associations between ethnicity and the quality of inpatient care that people with dementia receive.

• Exploring the relationship between hospitals having a carer strategy and levels of carer satisfaction, with a focus on relationships between intention, execution and organisational culture.

• Examining how carers who are able and willing to contribute to acute care of people with dementia can be supported.

• Investigating methods for enhancing carer involvement by co-producing optimal care packages and co-delivering dementia training for staff.

Furthermore, we recommend that future rounds of NAD consider collecting data on the median length of stay as one of the key outcome measures, due to the marked variation across hospitals. We also recommend collecting information on carer-rated quality of care and the presence of a personal information document, and to explore their associations with length of stay.

Dementia-friendly environments

Hypothesised impact of dementia-friendly environments on all three outcomes (Table 23):

By reducing agitation due to unsuitable furnishings, noise, etc., a dementia friendly environment will improve patient experience of care (as rated by carers) and other outcomes.

(Note, the measure of this in the NAD is whether an ‘environmental review’ has been carried out, but this will only make a difference to the dementia friendliness of the environment if the results of the review had been acted on. If considered important, this could be checked in the case study visits.)

Dementia-friendly food provision

Hypothesised impact of dementia-friendly food provision on length of stay (Table 24):

By encouraging better nutrition and hydration, dementia-friendly food provision (particularly finger food and 24-hour food availability) will help patients recover faster, resulting in shorter stays in hospital.

Hypothesised impact on carer-rated experience:

By encouraging better nutrition and hydration, dementia-friendly food provision will improve patient experience of care (as rated by carers). (Note, this is an aspect of care which is of great concern to carers, so when it is done well, or not done well, it is likely to be noticed.)

Dementia training for staff

Hypothesised impact of dementia training for staff on carer-rated experience (Table 25):

Through staff communicating to patients and observing carers that they have good awareness of the needs of people with dementia, dementia awareness training will improve carer-rated patient experience.

Staff with designated dementia roles

Hypothesised impact of staff with designated dementia roles on carer-rated experience (Table 26):

By leading and implementing dementia care in a hospital, the dementia lead supported by dementia champions and dementia working groups will improve all four outcomes.

Dementia-specific care

Hypothesised impact of dementia-specific care on length of stay (Table 27):

By prompting the development and implementation of an action plan to improve carer engagement including in the discharge planning process, self-assessment via the triangle of care tool will reduce length of stay for patients with dementia.

Hypothesised impact on carer-rated experience (1):

By prompting the development and implementation of an action plan to improve carer engagement including in the discharge planning process, self-assessment via the triangle of care tool will improve carer-rated experience.

Hypothesised impact on carer-rated experience (2):

By giving staff visual reminders of which patients have dementia, dementia-specific care in the form of alerts like This is Me will make them aware that the patient might need different or additional forms of help, which carers will notice.

Better involvement of carers

Hypothesised impact of better carer involvement on carer-rated experience (Table 28):

By having an ID [identification] scheme that allows carers to visit at any time, better involvement of carers via the Carer’s Passport Scheme will improve carer-rated experience.

Aims and objectives

We are conducting qualitative case studies of dementia care in general hospitals. Hospitals have been sampled to include pairs of hospitals that organise and deliver care in a similar way, but which achieve different outcomes. The overarching aim of this study is to identify aspects of organisation and delivery of general hospital services associated with three key outcomes measured by the National Audit of Dementia (NAD):

1. patient experience for people with dementia (as rated by their carers)

2. the length of stay that people with dementia have in hospital

3. comprehensive assessment of people with dementia.

At each case study site, we are aiming to undertake strategic interviews with those who have an oversight of dementia care in their hospital. These interviews will help shape the content and nature of subsequent interviews at the case study site and are an opportunity to understand any changes since NAD data collection (2016). The specific aims of these are to:

• identify things participants think their hospital is doing (or not doing), which in their view make a difference to the quality of dementia care in the hospital

• understand the rationale and theory underpinning these ‘programmes’ [i.e. how they think it makes a difference to outcomes listed above (carer-rated experience of care; length of stay; comprehensive assessment)]

• explore explanations underpinning predictors of outcome (from work package 1).

Aim

To explain the research, answer any questions and ensure informed consent.

• Introduce self.

• Purpose: brief explanation of DemCARE and purpose of this interview.

  • We want to understand what makes a difference to:

    • patient experience for people who have dementia (as rated by their carers)

    • the length of stay people with dementia have in general hospitals

    • comprehensive assessment of people who have dementia.

  • We would like to explore:

    • the hospital approach to dementia care

    • the rationale of this approach to dementia care

    • any key changes in dementia care in acute hospital since the last year (when NAD data were collected)

    • who else we should speak to from their hospital.

• Explain voluntary nature.

• Interested in views and experiences: no right or wrong answers.

• Confidentiality and anonymity: everything is confidential, might use quotations in written reports, not possible to identify individuals.

• Recording: would like to audio-record with permission, data stored in accordance with the Data Protection Act.

• Timings: reiterate start and finish times.

• Any questions?

Start recording: record verbal consent.

Aim

To establish rapport, help participant feel at ease and gain important background information.

Professional background

• Briefly explore professional background (what do they do, how long for, length of time working in current hospital/setting).

• Explore any dementia-specific experience (training, role, champion, old age specialism).

Time dependent.

Hospital context

• Briefly explore hospital make-up and demographic (what specialist wards or treatment centres for elderly people and those with dementia).

Exploring recent developments

• Explore how similar or different dementia care is in the hospital now compared with a year ago.

• Map what (if anything) has changed and why.

• In your experience, what sorts of things for a patient with dementia are likely to make a difference to the outcomes of their stay in hospital? (Start openly then probe: financial situation, social support, health condition, other. )

• What characteristics in the way staff work with patients with dementia help to create better outcomes (e.g. quicker discharge)? What makes a difference?

• Can you think of any external factors that can support or impede the delivery of effective care? (Start openly then probe: onwards health care, local social services, other.)

• Other contextual factors? (Start openly then probe: beds, staff experience, dementia training, organisational culture, other.)

Aim

To explore what dementia care looks like in the acute hospital and identify priority areas.

Identifying relevant programme and subprogramme theories

Aim

To explore explanations underpinning predictors of outcome (from work package 1).

Overall approach to optimising dementia care

• Shared understanding of terms. [Briefly reiterate (from introduction) that our focus is on comprehensive assessment, length of stay, carer-rated quality of care (briefly explain that our definition of carer-rated quality of care is informed by NAD and show the handout).]

• Explore any other outcomes they think should feature in high-quality dementia care.

• Most important ‘thing’ (intervention/policy/programme).

  • Explore what they think is the most important ‘thing’ the acute hospital does to optimise care for people with dementia, in relation to:

    • comprehensive assessment

    • length of stay

    • carer-rated quality of care.

• For each ‘thing’ identified, probe in depth on:

  • Why they think this makes a difference?

  • How they think this makes a difference?

• Explore any links between predictors and outcomes (Table 30), go through each in turn, probe on:

  • how far this seems unusual or surprising

  • extent to which they think it is true of their own hospital

  • any explanations (i.e. why is there this link).

Aim

Give a chance to add anything else and wind down.

Check whether there is anything else to add, that should be kept in mind for remaining interviews.

Stop recording.

• Wind down.

• Reassure about confidentiality.

• Check whether or not there are any questions (when recorder is off).

• Check if they want to be informed of findings of the study (note e-mail address if they do).

Aim

To introduce the research and set the context for the discussion.

• Introduce self and Imperial College London/RCPsych.

• Introduce the study: who is it for, what is it about?

• Talk through the key points:

  • purpose of the interview

  • length of the interview

  • voluntary nature of participation and right to withdraw

  • reasons for recording the interview.

• Confidentiality, and how findings will be reported.

• Check for any questions they have.

• Start recording, then ask:

  • what relationship the interviewee has to the person they look after/care for (e.g. spouse or partner)

  • whether or not they regard themselves as one of the person’s main carers

  • how they would like the person they care for to be referred to (e.g. ‘your husband/wife’, ‘the person you care for’).

Aim

To explore carers’ views and experiences of the quality of care, things they think may have influenced care quality and how those things have made a difference.

• Overall, how would you rate the quality of care received by [the person you look after] during their stay in hospital? (Start openly then probe, eliciting examples when possible.)

  • Staff:

    • well informed and understood person’s needs

    • treated person with respect

    • kept you informed (e.g. about progress)

    • involved you/person as much as you wanted in care decisions

    • asked you about person’s needs for care planning

    • other.

• In your experience, what things may have influenced the quality of care [the person you look after] has received? (Start openly then probe for contextual factors and dementia care programmes/interventions, eliciting examples where possible. )

• How do you think those things have made a difference?

Aim

To explore carers’ views on length of stay, things they think may have influenced this and how those things have made a difference.

• What are your views on how long [the person you care for] is expected to stay in hospital? (Start openly then probe, eliciting examples when possible: too long/too short, other.)

• In your experience, what things may have influenced their length of stay here? (Start openly then probe for contextual factors and dementia care programmes/interventions, eliciting examples when possible. )

• How do you think those things have made a difference?

Aim

To explore carers’ views on discharge planning, things they think may have influenced this and how those things have made a difference. (Establish if they are aware of any discharge planning; if not, skip to next section. )

• Overall, how would you rate the discharge planning process? (Start openly then probe, eliciting examples when possible: involved in decisions, other. )

• In your experience, what things may have influenced discharge planning? (Start openly then probe for contextual factors and dementia care programmes/interventions, eliciting examples when possible. )

• How do you think those things made a difference?

Aim

To explore carers’ views on other outcomes that they perceive to be important, things they think may have influenced those outcomes and how those things have made a difference. (Ask these questions about each outcome. )

• Overall, how would you rate [outcome X, Y, Z]? (Start openly then probe, eliciting examples when possible. )

• In your experience, what things may have influenced [this outcome]? (Start openly then probe for contextual factors and dementia care programmes and interventions, eliciting examples when possible. )

• How do you think those things made a difference?