Using the Recommended Summary Plan for Emergency Care and Treatment in Primary Care: a mixed methods study

Anticipatory treatment plans

When a person develops a life-threatening condition or has a sudden deterioration in an existing illness, rapid treatment decisions may be needed. Often, the person is unable to be involved in the decision at the time, and there is limited clinical information or information about what their treatment preferences might be in this situation. Anticipatory treatment decisions and recommendations can improve this decision-making process and help to guide healthcare professionals while ensuring the person’s values and preferences are respected in the future. Anticipatory treatment plans can take several forms. A person may decide to make an advance decision to refuse treatment and specify particular treatments that should not be given to them should they be unable to speak for themselves. 1 More commonly an advance care plan (ACP) may be made between a person, their family and those caring them to document their future priorities of care more broadly, including preference for place of death. 2–4 ACP is mainly advocated in the context of end-of-life care, although an international consensus definition of ACP states that: ‘Advance care planning is a process that supports adults at any age or stage of health in understanding and sharing their personal values, life goals and preferences regarding future medical care’. 5

Emergency care treatment plans (ECTPs), sometimes called treatment escalation plans, focus specifically on treatment decisions in emergency or acute illness situations. Their aim is to make treatment recommendations that reflect the person’s preferences and values, and they are reached in discussion with the person or their family. However, the specific recommendations are made by the healthcare professional and are intended to guide future treating clinicians. 6 An important distinction between ACPs and ECTPs is that, whereas the ultimate focus of ACPs is on end-of-life care preferences, ECTPs focus on steps to be taken in the event of an acute pathophysiological deterioration in which recovery is possible, although it may be unlikely. 7 Prior to 2016, stand-alone do not attempt cardiopulmonary resuscitation (DNACPR) recommendations were the most commonly used form of emergency care treatment planning document in acute NHS hospitals (72% of hospitals). 8 However, concerns existed about their use, including lack of discussion with patients and/or their family, failure to consider cardiopulmonary resuscitation (CPR) recommendations before a patient deteriorates, and lack of transferability of forms between primary and secondary care and ambulance services. 9,10 A particular concern was that a focus on CPR leads to lack of consideration of other treatments that may or may not be appropriate. 11–13 Responding to these concerns, some hospitals and health networks developed ECTPs to provide a more holistic approach to anticipatory medical decision-making. 14–16 In 2016, the Resuscitation Council United Kingdom (RCUK) led development of a national standardised approach to support conversations about goals of care and to provide guidance to clinicians about treatments to be recommended in an urgent situation when the patient lacks capacity to decide for themselves: Recommended Summary Plan for Emergency Care and Treatment (ReSPECT). 6

The ReSPECT process

ReSPECT is both a process and a form documenting the process (see Report Supplementary Material 1 for copy of ReSPECT form, versions 2 and 3). As described on the RCUK website, the ReSPECT process creates a summary of personalised recommendations for a person’s clinical care in a future emergency in which they do not have capacity to make or express choices. The process takes into account both patient preferences and clinical judgement and records ‘agreed realistic clinical recommendations’ including a recommendation whether or not CPR should be attempted if the person’s heart or breathing stop. Implementation of ReSPECT in NHS acute trusts began in December 2016 and within 3 years it was used in 22% of trusts. While other forms of ECTPs were used in some trusts, ReSPECT accounted for 60% of the observed move away from stand-alone DNACPR forms to ECTPs during this time. 17 Since its initiation the ReSPECT form has been modified in response to feedback from clinicians and patients (in 2023 both versions 2 and 3 were in use).

Although ReSPECT implementation occurred first in acute trusts, it was always envisaged that this process would cross the primary and secondary care interface. Indeed, it was developed as a patient held form rather than a clinical record. An evaluation of ReSPECT in early-adopter trusts found that many clinicians thought that the ReSPECT conversation and documentation of recommendations fitted better into a primary-care setting. 18 Patients may have an established relationship with their general practitioner (GP), conversations can occur over an extended period, patients are often less sick and more able to engage in discussion, and conversations can be placed in a wider context of ACP. However, there are also potential difficulties in moving ReSPECT conversations to primary care; patients and families may be less ready to think about these things until a crisis emerges, GPs may be uncertain about hospital-based interventions, and both may have concerns about the effect of a conversation on the patient–doctor relationship. Another challenge to the use of ReSPECT across the primary and secondary care interface is that hospital doctors and GPs use the process in different ways and recommendations do not always translate to different settings. 19

Emergency care treatment plans and person-centred care

There is an ethical and professional obligation to balance benefits and burdens of treatment from the patient’s perspective. 20 Current NHS policy and professional guidance emphasises a model of personalised care and shared decision-making. 21 ECTPs can facilitate person-centred shared decision-making in the acute situation by summarising earlier considered assessments of the potential benefits and burdens of a range of future treatments, taking into account what is important to the patient.

People with a learning disability might particularly benefit from emergency care treatment planning as their needs and wishes are often not met in acute situations. 22 Health outcomes are often poor for people with a learning disability because health professionals do not understand their needs. 22–24 Obstacles include lack of education and training among health professionals, communication challenges and/or a perception that individuals lack capacity because of their disability. 25,26 Anticipatory care planning with people with a learning disability is minimal due to a lack of confidence and awareness among medical staff. 27–29 This further disadvantages people with a learning disability by denying them input into decisions about future treatment.

The impact of COVID-19 on emergency care treatment planning

The COVID-19 pandemic generated increased interest in the use of ECTPs in general and ReSPECT in particular. High mortality, particularly among older frail people and those with specific medical conditions, few effective treatments and restrictions on hospital visiting, all had a potential impact on decisions about whether a person would benefit from or would want hospital admission or treatment if they developed COVID-19. GPs and hospital doctors admitting acutely ill patients were encouraged to have these conversations with patients early and to document the resulting treatment recommendations. In a UK teaching hospital, completion of treatment escalation plans more than doubled during COVID-19 from 20% of inpatients to 50%, and the scope of the recommendations was broader than before COVID-19. 30 This renewed emphasis on anticipatory decision-making during COVID-19 was seen as supporting best practice in person-centred care. A paper reporting a retrospective analysis of a multicentre cohort study of patients admitted to hospital with COVID-19 also found more nuanced decision-making around life-sustaining treatments, such as CPR. Participating clinicians explained that the use of the ReSPECT process allowed more nuanced treatment recommendations to be made compared with using DNACPR forms. 31 It was also an opportunity to embed the principles and practice of ACP and emergency care treatment planning into clinical practice more generally. 32 However, there were concerns that the impetus to complete ECTPs early, and the restrictions on communicating with patients and their families in person, might lead to inappropriate use of ECTPs and result in poor patient care, leading to guidance being issued by professional and regulatory organisations. 33,34 The Care Quality Commission’s report into its review of DNACPR recommendations during the COVID-19 pandemic recommended that decisions about DNACPR should be made as part of a wider conversation about ACP, and that people and their representatives should be supported in having these conversations. The report suggested that ReSPECT was a good example of an ACP process that enabled these conversations to occur. 35 The Royal College of General Practitioners and British Medical Association recognised the specific challenges for GPs in having these conversations during a pandemic and provided guidance to its members reiterating the requirement for individualised conversations about future treatment decisions. 36,37 It is not clear to what extent practice around emergency care treatment planning has changed since the pandemic.

Evaluating the use of emergency care treatment plans in the United Kingdom

We identified six UK evaluation studies of locally developed ECTPs published between 2010 and 2017. 14,15,38–41 All reported that the ECTP improved communication and documentation of treatment escalation decisions. Three studies reported perceived improvement in appropriate decision-making around treatment escalation and resuscitation, although appropriate was not defined. ECTPs were positively evaluated by staff. Two studies demonstrated a significant reduction in patient harms associated with the use of the ECTP compared with patients who had a standalone DNACPR form documented in their notes. 40,41 Since the introduction of ReSPECT, the focus of published evaluation studies has been on this model of ECTP. An evaluation of the first ReSPECT pilot in Scotland, which included 200 ReSPECT forms completed in a range of settings (hospital, hospice and community), found that patients with a ReSPECT form were more likely to die in their preferred place of care and had a reduced chance of readmission within 3 months of hospital discharge. Patients and staff were generally positive about the process. Challenges included time pressures, staff reluctance to initiate a conversation or lack of confidence to do so. 42 A multicentre evaluation of the use of ReSPECT in early adopter sites involving six acute NHS trusts in England took place between 2017 and 2019. It found that ReSPECT conversations were mainly initiated with patients nearing the end of their life or at imminent risk of deterioration and focused on CPR decisions. However, a move towards a more holistic approach in terms of treatment recommendations and conversations with patients and their families was observed. 8 Doctors’ uncertainty about a patient’s prognosis, constraints of time and external environment, and the need to minimise patient distress, influenced both the prioritisation and content of conversations. 18 An interview study with GPs and care home staff in the West of England found a generally positive attitude to the use of ReSPECT but noted that its use was complex and there were challenges in incorporating patients’ preferences into decision-making. The authors recommended a multidisciplinary approach that engaged care staff more in the process. 43 There has been no large study of the use of ReSPECT or other ECTP in the community setting.

Work package 1

A qualitative study exploring the experience of ReSPECT in primary care and community settings drawing on interviews with GPs, patients and their families/carers and staff in care homes, and conversations with other members of staff in the participating GP practices (see Chapters 4, 5 and 6).

Work package 2

Focus groups and interviews with healthcare professionals, home care workers, members of the public and faith leaders to explore their views on the principles and practice of ReSPECT and other forms of anticipatory decision-making; two national surveys, one of the public’s attitudes to ECTPs and one of GPs’ experiences and views of anticipatory decision-making (see Chapters 3 and 4).

Work package 3

Quality assessment of ReSPECT form completion in the GP practice sites involved in WP1 (see Chapter 4).

Work package 4

Co-production workshops with adults with a learning disability to explore their understanding of and views on emergency care treatment planning and to co-create resources to support the engagement of people with a learning disability with ReSPECT; focus groups and interviews with relatives of people with a learning disability to capture their views and experiences of emergency care treatment planning. This WP was added to the original protocol following a successful application to NIHR for additional funding in response to a call for proposals to be submitted by current NIHR award holders for additional research related to social care, to be completed within their current award time frame (see Chapter 7).

Work package 5

An initial synthesis of findings across the WPs informed the content of a stakeholder meeting with participants from professional and patient organisations and implementers of ReSPECT across the UK; final synthesis of all findings (see Chapter 8).

See Report Supplementary Material 2 for study flow-chart.

Changes to recruitment of care homes

In our original protocol, we aimed to recruit two care homes associated with each GP practice site. Planned care home involvement included interviews with senior care home staff, involving care homes in recruitment of patients and their relatives, and identification of ReSPECT forms. Of 23 care homes identified by our GP practices, only 11 agreed to participate and, of those, 8 agreed to take part in the interview element only. Service delivery pressures related to COVID-19 and its consequences and frequent changes in senior staff contributed to the challenge of recruitment. To ensure that we captured the full range of care home experience, we expanded our recruitment through care home networks to include care home staff not linked to our GP sites.

Changes to recruitment approach for patient and relative interviews

Recruitment of patients and relatives of patients who lacked capacity was low. In response, we simplified our recruitment approach by removing the link between interview (WP1) and access to patient records (WP3) in the invitation letter. We also sought to advertise the interview study directly through care home networks.

Decision to stop data collection for congruence

We originally planned a retrospective analysis of patient records and ReSPECT forms to measure congruence between ReSPECT recommendations and clinical decision-making at the time of an acute event for people with a ReSPECT form completed in the previous 6 months. The estimated sample size for this analysis was 413 ReSPECT forms, assuming that the acute clinical event rate within 6 months of a ReSPECT form completion was 70%. Across our first 6 practice sites, we identified 233 patients with a record of a ReSPECT form completed in the previous 12 months, of whom 169 were eligible to participate. Of eligible patients, we were only able to identify and extract data from 65 ReSPECT forms. We had anticipated that GP practices may not always have a copy of a patient’s ReSPECT form as they are designed as patient held forms. However, we had expected that we would be able to identify additional ReSPECT forms through the care homes recruited to the study. As noted above, recruitment of care homes to this element of the study was extremely challenging and so this method of data collection was not possible. Additionally, analysis of the clinical records of patients with a ReSPECT form found an acute event rate of 24%. Following discussion between the study team, the study independent steering committee and the funder, it was agreed that it would not be possible to achieve the required sample size within the parameters of the current study. Therefore, data extraction from clinical records was discontinued for the remaining sites but collection of data from anonymised ReSPECT forms continued to enable a quality analysis of form completion.

We then obtained the relevant approvals to conduct a feasibility study to explore whether a different study design, using care home records only, could answer this research question. We encountered several challenges in recruitment of care homes and recruited two from a target of three homes. We found that 17/75 (22.7%) of residents had both a ReSPECT form and an acute event in a 6-month period. We estimate that we would need to work with 50–60 care homes to complete a congruence analysis (see Appendix 2 for a more detailed report). This finding and our findings on the quality of recorded recommendations in our review of ReSPECT forms in general practices lead us to conclude that at present a congruence analysis as originally conceived is not feasible.

Scope of patient and public involvement

Summary

The importance of PPI was recognised at an early stage of development of the project and was integral to its development, conduct, delivery and successful completion. Embedding the lived experience of people with a learning disability in development and outputs for WP4 was particularly important and enriched the overall study findings. The presence of PPI throughout the project also helped to ensure that the work retained its focus on the person at the heart of the ReSPECT process, and that language and communication was consistently clear and accessible.

Recruitment

We sought expressions of interest for focus group participation from members of the public with an interest in health through local Health Watch, community and patient organisations and patient groups linked to our participating general practices in each of our three study areas. 53 We asked our PPI advisory group for suggestions of relevant networks to approach and carried out desktop research to identify local patient and carer organisations in each area, including those representing minority groups. We provided a poster about the study, an invitation e-mail and brief information leaflet. Most expressions of interest came from advertising through local Health Watch organisations. Potential participants contacted the study team directly and were provided with further information and offered a range of dates for planned focus groups. If they agreed to participate, they were given details of how to join their focus group. Online participants had unique links to a Zoom™ (Zoom Video Communications, San Jose, CA, USA) meeting. One participant unable to attend a focus group was interviewed separately. Prior to data collection, participants were sent a copy of the ReSPECT form and a consent form. We obtained written consent at the start of the in-person group. Online, we went through the consent form with each participant and recorded their consent.

We advertised the study to faith leaders through community faith groups, local councils, university chaplaincy and NHS trust chaplaincy services in our three study areas and followed up expressions of interest as above.

Data collection

We developed topic guides with input from our lay advisory group. Focus groups were facilitated by researchers JH, SR or CM. Online and face-to-face focus groups and the individual interview followed the same process. After introducing the purpose and format of the discussion, and the ReSPECT process, facilitators used a topic guide to guide the discussion (see Report Supplementary Material 5). They prompted participants to consider potential benefits and disadvantages of emergency care treatment planning for patients, relatives, the health service and wider public, and consider aspects of implementation, including how and when ReSPECT conversations should be initiated, any potential challenges in using the forms, and if, and when ReSPECT forms should be reviewed.

Interviews with faith leaders followed the topic guide for the focus groups with additional prompts on how the values embedded in ReSPECT interact with key values in their faith.

Participants received a shopping voucher as a thank you for participation. We ran four focus groups; one of which was in person. Interviews were online apart from one faith leader interview. Focus groups lasted between 53 and 66 minutes; interviews with faith leaders lasted 56 and 62 minutes, respectively. All focus groups and interviews were recorded with consent, professionally transcribed and checked by the study team.

Analysis

We analysed the data as one data set using framework analysis. 54 Four study team members (FG, AS, JH, SR) initially each read a transcript from study area one, discussed the data and identified potential themes for public survey questions. JH read and coded all transcripts from area one using the topic guide as a framework, identifying themes from which we confirmed topics for the survey questions. Two researchers (CB and CM) then independently coded one focus group transcript to check for consistency and themes were discussed with JH. CM then further developed and refined the coding framework to analyse the remaining focus group and interview transcripts. Finally, AS read all transcripts and confirmed final themes with CM.

ReSPECT could help normalise discussion about emergency care

Participants across the focus groups acknowledged that death and planning for the end of one’s life were sensitive topics, and many people resist having these conversations. They agreed that ReSPECT could offer a way for individuals, families, and clinicians to begin the conversation.

It’s a really helpful tool, a kind of lever to, to have a conversation about end of life and about dying with people, and I think that’s a conversation that people find very hard to have, but it’s a really important conversation to have and I think having a document that acts as a stimulus to a conversation about end of life is really useful and really important and we don’t do enough of it, so I think it’s a good thing to have.

Public focus group 2

Participants initially discussed ReSPECT conversations as important for people approaching the end of life or with a progressive or life-limiting condition. This progressed to a discussion of relevance of ReSPECT for a much wider range of people, as emergencies could occur at any age, and that everyone inevitably would have deteriorating health. Early conversations and records of a person’s wishes could help families make difficult decisions.

I think you’re focusing more, aren’t you, on the elderly people, but you talk, it’s got the word ‘emergency’ in the title, and you could be in a car accident at 25, you know, there’s all sorts of reasons when it could be brought in.

Public focus group 2

I think life is a terminal illness. You know, … we’re all gonna get there. And I think it’s important to think about these things while you’re able to, you know, whatever your circumstances are, because it, it takes so much pressure off your family.

Public focus group 4

Good communication during the ReSPECT conversation between patient, relatives and health professional was seen as important to ensure good understanding of the purpose of the plan and the potential future scenarios and treatments. Participants were clear that the health professional needed to be skilled in having these conversations and thought that conversations would work best when the person had an established relationship with the health professional, although they recognised that this was not always possible: ‘I think it, it’s, it’s the person you have the best relationship with, not everybody has a good relationship with their doctor’ (public focus group 1).

Some participants suggested that increased public awareness of ReSPECT, for example through the media, would encourage more people to have these conversations. They suggested ideally it should become normalised ‘like the organ donor card’ (public focus group 2). The person should be at the centre of the process

A key theme across focus groups was that the ReSPECT process should focus on the individual, recording a plan that reflected their wishes and preferences about their future treatment. Some participants thought the ReSPECT plan was an excellent means for a person to communicate their wishes.

… I think they seem like a very useful tool so that you can, kind of, convey your wishes before being in that situation, if ever. At least you’ve laid out a sort of plan and your, your wishes can be respected.

Public focus group 4

However, some participants expressed concern that the ReSPECT process was insufficiently focused on the person. The form seemed medically orientated rather than person orientated. In one focus group where participants were mainly care home residents’ relatives, several compared the ReSPECT form to an ACP that they had completed for their relative on entering the home. They felt the ReSPECT plan, with its focus on emergency treatments, did not capture their relative’s reality which was more of a slow decline rather than sudden emergencies.

the one we filled in here for when my mother came in, I thought was wonderful, it was about how she wanted to be treated, you know, and it was very reassuring. I look at this and think it’s not really relevant to her, … it’s not really suitable, ’cause it’s so difficult to answer the questions. Whereas the, the form we filled in here was very warm … and personal and this is very bleak.

Public focus group 3

Participants were concerned that the person may not understand what was being discussed in a ReSPECT conversation, particularly if conversations were initiated at a late stage in a person’s illness trajectory. They commented on the potential complexity of treatment options discussed and the uncertainty of prognoses or potential future events. One participant expressed a feeling of disempowerment when considering completing a ReSPECT plan at a time of deteriorating health.

But actually, in a, a real-life scenario I think I would feel quite disempowered, that other people hold the power about things to do with my life. And I really, you know, in my most vulnerable state I don’t even know what my options are.

Public focus group 1

One faith leader suggested that it could be helpful for some people to include a faith leader in their discussions about life-sustaining treatment.

All participants across the focus groups felt strongly that the plan needed to be reviewed regularly to reflect changes in the persons health status and wishes over time.

… your views definitely change … I think it should be re-evaluated, as I say, I think to a certain degree you like to put it in the back of your mind, but sometimes I think when you’ve had more time to think about it, your views might have changed. So, I do think there should be a follow-up on it, definitely.

Public focus group 3

Yeah, that already then is a flaw in the system to me. Because what I might wish to do today may change due to experience, different circumstances. Your health might, you know, you, you, you might think your health is A today, and in two years’ time something else happens.

Public focus group 4

ReSPECT and the patient’s family

There were mixed views across the focus groups about the extent to which families should be involved in the ReSPECT process. Many thought it was important that the person’s family was involved in or aware of the ReSPECT plan to ensure that the person’s wishes were followed and take pressure off the family.

If you’ve got that understanding, there’s no fear …, you’re then at, at, at peace internally, you know, you’re at peace with what is happening to your other, to your other half, rather than not knowing what that person wants.

Faith leader 2

This will help your relatives make a difficult decision in a time when you’re not there to advise.

Faith leader 1

The ReSPECT process could also reduce conflict between family members when decisions needed to be made when the person was very unwell.

I think it helps promote clarity between the, … patient and, and their relatives about what might be wanted … It’s not necessarily going to fix all of the misapprehensions that people have, but it’s, it is a clear way of having that conversation and, and, you know, flushing out whether there are agreements or disagreement, it’s a really important thing.

Public focus group 2

[I]t, it really needs to, to be a, an open discussion which is not just linked to the immediate close person in your life. I think it has to, it has to be wider. And then the, because it’s, it will be for, for everyone it would be a very, very emotional situation to deal with.

Public interview 1

However, other participants noted that not everyone wanted their families to be involved in or know about their healthcare decisions.

And, you know, in, in some cases perhaps somebody doesn’t want this to be discussed with the whole of the family, because it is their treatment, and there are issues about confidentiality, even within the family.

Public focus group 1

Participants recognised that it was often difficult for clinicians when faced with families who wanted them to intervene for their sick relative, even when there was a form specifying that this was not what the patient wanted.

Scepticism about use of the plan in an emergency

While many participants talked about the benefits of having a ReSPECT plan, they were not confident it would be followed in the situations it was designed for.

I’m not entirely confident that having … an emergency care treatment plan will necessarily produce the result that the person was expecting, because I think there are all sorts of things that could get in the way of what somebody wants actually being implemented.

Public focus group 2

Participants identified several obstacles to effective use. They discussed practical problems in ensuring the plan was available when needed. Someone living at home may forget where the form is kept or be too sick to tell the health professional, usually a paramedic, who is attending. A potential solution would be a centrally held form or record of the plan that was accessible to clinicians at the time. Participants were generally sceptical of the ability of current NHS electronic systems to provide reliable access to these plans. Participants were not confident the recommendations on the plan would be followed. Examples were given of paramedics noting plans but then taking a different course of action because of their clinical judgement.

And both of those people had the ambulance service come, both of them showed the ambulance service the documentation, and both of them were told that they would not take that action, they were there to save life. And that upset both families very much.

Public focus group 4

Ambiguity in the plan recommendations could result in actions that were inconsistent with the plan’s intention. For example, a recommendation about not being admitted to hospital without specifying the relevant circumstances may lead to someone being admitted/not admitted inappropriately. Participants recognised that it was impossible to specify every circumstance and provide detailed management recommendations. They recognised that recommendations would need to be read quickly in an emergency so should be quite brief, but that this created a tension with communicating detailed person specific instructions about a person’s care.

You do need something that’s fairly quick to refer to, because you can’t give a, a long kind of detailed set of instructions if it’s going to be read by paramedics and, and that sort of thing in an emergency situation.

Public focus group 2

Some participants commented that it was important to think of the recommendations as guidance but that the attending clinician would need to ultimately make the treatment decision, and people would need to trust that the clinician would act in their best interests.

I’m still convinced that the doctor makes the decision at the end of the day.

For me the form is a way to help people think about these things, express an opinion, but it’s not obligatory, as it says on the form…it’s useful for clinicians who don’t know that patient. To have something there, as guidance, and know that there has been a discussion, particularly in an emergency.

Public focus group 1

Lack of resources could result in recommendations not being fulfilled. For example, if treatment at home rather than hospital was recommended this might require urgent nursing or other support.

Reflecting on these challenges to effective use of ReSPECT plans, some participants questioned how person centred the plan was: ‘it seems to me control is taken away from you and that’s what’s so frustrating and tiring and bureaucratic’ (public focus group 3). However, even if the plan was not available or could not be followed precisely, some participants still regarded the process as helpful to ensure the person’s wishes were taken into account.

I’m a realist though and I perfectly understand that come the day when I get unplanned admission to hospital nobody is actually gonna dig through my file in the GP practice, and dig this out from seven years ago, and respect my wishes. But at least my daughters might remember something that was said.

Public focus group 1

Questionnaire development

To ensure a high-quality nationally representative population sample, we commissioned the National Centre for Social Research to include our questions in the annual British Social Attitudes (BSA) survey. 56 The BSA is the UK’s longest-running survey of public opinion. It provides a high-quality nationally representative attitudinal survey with a stratified sample based on postcode and includes face-to-face data collection using computer-assisted personal interviewing and self-completion. Question development was a collaborative process between the study team and the BSA team. Firstly, members of the study team (AS, FG, SR, JH) developed a draft set of questions, drawing on data from focus groups. Following feedback from our co-investigators and lay advisory group the questions were further refined. Next, the study team met with the BSA team and, through an iterative process, the final set of questions to be included in the BSA annual survey were agreed. We aimed to determine the willingness of members of the public to complete an ECTP for themselves or someone close to them, under what circumstances they would consider completing one and who they would prefer to complete the plan with. We also sought their views on potential benefits or harms of having an ECTP. A sample of the questions were then tested by the BSA team in cognitive interviews with 10 people, and the full set of finalised questions was piloted with a sample of 56 participants recruited from the BSA panel database and using telephone interviews. Results of the cognitive interviews and pilot were discussed with the study team and minor modifications made to question wording (see Report Supplementary Material 6 for final questions).

Data collection

Data collection was carried out in accordance with the National Centre for Social Research’s protocol for delivery of the BSA survey. Invitations to participate in the survey, with the option of online or telephone interview completion, are sent to a stratified sample of households in the UK identified from the postcode address file with up to two adults in each household able to participate. Two online access codes are provided to each household. If selected households do not wish to complete the survey online, they are able to call a freephone number and arrange to complete the survey by telephone with a specialist telephone interviewer. To maximise response up to three invitation mailings are made. Respondents are offered a conditional £10 incentive to participate. Fieldwork took place between 9 September and 30 October 2022. The target sample for administration of our question set was 1000.

Analysis

In addition to descriptive statistics for each question, we present logistic regression analyses investigating the variables associated with the three dichotomised dependent variables: (1) being in favour of anyone being able to have an ECTP (strongly and somewhat in favour vs. the other categories); (2) whether they would like an emergency care and treatment plan at present (definitely and probably would like vs. the other categories) and (3) (for those who did not already have an a emergency care and treatment plan) how comfortable (very and fairly comfortable vs. the other categories) they would feel about making such a plan for themselves. For each analysis, the independent variables were age, gender, ethnicity, educational level, having an illness lasting more than 12 months, caring for someone ill or with a disability and being close to someone with a condition that could shorten their life. These were chosen after discussion within the study team to identify ‘a priori’ the most important possible independent variables. Using the statistical software R Core Team (2023; R Foundation for Statistical Computing, Vienna, Austria) for each dependent variable, both fully adjusted model and non-adjusted models were fitted at the 5% significance level. With very few missing data, the logistic regression models for binary data fit well.

Attitudes to emergency care treatment plans

A large majority of respondents were in favour of people being able to have an ECTP if they so wished, with 908/1135 (80%) at least somewhat in favour (see Appendix 3, Table 13). Females were slightly more likely to be in favour than males [82% vs. 77%; odds ratio (OR) 1.45; 95% confidence interval (CI) 1.06 to 1.97; p = 0.02]. When compared with those with no qualifications, people with degrees were significantly (p = 0.002) more likely to be in favour (70% vs. 84%; OR 2.68, 95% CI 1.42 to 5.07). However, the absolute difference is modest and there is not a clear trend for more people with a higher educational level to be more in favour of ECTPs. Overall, ethnicity did not appear to affect people’s views, although the numbers in each group were small. However, people of Asian ethnicity might be less likely than those identifying as white British to be in favour of everyone being able to have an ECTP 67% versus 82% (OR 0.45, 95% CI 0.25 to 0.84, p = 0.012) (Table 1).

Of respondents who did not currently have an ECTP, 698/1112 (63%) felt they would be at least fairly comfortable making an ECTP, with only 49/1112 (4%) saying they would be very uncomfortable having an ECTP discussion (see Appendix 3, Table 14). Compared with those with no qualifications, people with degrees were more likely to feel comfortable having an ECTP (70% vs. 49%; OR 2.48, 95% CI 1.40 to 4.39; p = 0.002; Table 2). Comfort with completing an ECTP was not significantly associated with age, gender, or experience of caring for someone with a long-term health condition. Of those respondents who said they would like to have an ECTP for themselves, half (316/618, 51%) would want their GP to complete this with them while one-quarter (161/618, 26%) would want a doctor or nurse who did not know them but was trained in making ECTPs to do this (see Appendix 3, Table 12).

Who would like an emergency care treatment plan for themselves and when would they want to have it in place?

Half of respondents without an ECTP (620/1112; 56%) would want one at present (see Appendix 3, Table 15). Overall, fewer people in older age groups would want an ECTP at present; however, the difference was only statistically significant for one comparison; 63% in those aged 18–24 years compared with 47% for those aged 65–69 years (p = 0.045, OR 0.5, 95% CI 0.26 to 0.98; Table 3). People with a chronic physical or mental health condition that reduces their activity were more likely to want a plan at present compared with those in good health (64% vs. 52%; OR 1.78, 95% CI 1.30 to 2.45; p < 0.001; see Table 3). Knowing or caring for someone with a long-term physical or mental health condition did not affect whether they would want an ECTP for themselves (see Table 3). In answer to a more specific question of when you would want an ECTP in place for yourself, 152/1112 (14%) would want one ‘now’ with just 36/1112 (3%) saying that they would never want one. Developing a long-term condition or becoming disabled would lead 467/1112 (42%) and 481/1112 (43%) respondents respectively to want an ECTP. Even more 534/1112 (57%) would want a plan if they developed a life-threatening condition (see Appendix 3, Table 16). Of the 441 people who said they would want an ECTP when they were older, the peak decades identified were their 60s and 70s with 270/441 (61%) wanting an ECTP in place by the time they were aged 70 years (see Appendix 3, Table 17). When asked if they would like to be involved in completing an ECTP for a close family member if they were not able to do this for themselves, 930/1135 (82%) said that they definitely or probably would (see Appendix 3, Table 14).

Perceived potential benefits and harms of emergency care treatment plans

To explore how responders felt about potential benefits and harms of ECTPs, they were asked whether they agreed to a series of statements in relation to ECTPs. Predominately, 938/1135 (83%) respondents agreed or strongly agreed that ECTPs would help avoid their family needing to make difficult decisions on their behalf, and that a plan would ensure doctors and nurses knew their wishes (Table 4). Nevertheless, a small majority 628/1135 (55%) agreed that there is a serious risk that a plan could be out of date and so not reflect their current views or health condition, and a substantial minority 330/1135 (29%) agreed that in having a plan they might not get a treatment that would save their life (see Table 4).

For each of our regression models, the results of our univariable analyses were not materially different from our multivariable analyses (see Appendix 3, Tables 18–20).

Strengths and limitations

To our knowledge, we have completed the first community survey of the UK public’s views and attitudes to the completion of ECTPs. The use of the annual BSA survey to collect these data means that we can be confident that our sample is nationally representative and that the data quality is good. However, this approach will exclude many for whom ECTPs are most relevant, that is frail older people and those with cognitive impairment. The slight under-representation for people from minority ethnic groups is typical of push surveys of this nature but is not large enough to affect our overall conclusions. The rigorous process for developing our questions, including preparatory qualitative work, and their piloting by the BSA survey team minimises question ambiguity. Nevertheless, it is still possible that some respondents misunderstood the purpose of each question. Some caution is needed when interpreting statistical significance because of the large number of analyses done.

Despite contacting a wide range of community and faith groups, several of whom supported advertising the study, we received very few expressions of interest for focus group participation. Feedback suggested that many groups had insufficient resources to support recruitment. GP patient groups had often stopped during the COVID-19 pandemic. We were unable to purposively select participants for demographic variability because of lack of interest in two study areas. The small number of faith leader participants meant that it was not possible to achieve our aim of exploring the impact of different faith values on how people might view or engage with ReSPECT. All participants had a particular interest in, or experience of ReSPECT or ACP so may not have reflected the views of the general public. However, because of their interest and experience, our participants were able to provide in depth reflection on the topic.

Questionnaire development

Drawing on initial analysis of our interviews with GPs, and reflecting the questions in our public attitudes survey, members of the study team (AS, FG, JH, CB) developed a draft set of questions. We aimed to measure experience of using ECTPs compared with standalone DNACPR forms, views on the use of ECTPs in primary care, and how likely they were to complete a plan for their patients. During question development, we identified that knowing which factors might predict how comfortable GPs were at having an emergency care treatment planning consultation with a patient, or family member, as a key questions of interest, and likely to reflect how likely a GP was to complete a plan with their patients. We originally planned to use 11-point numerical rating scales to assess GPs’ attitudes to aspects of emergency care treatment planning. Input from our qualitative work and the work we did refining the questions led us to conclude that five-point Likert scales were more likely to be satisfactorily completed and the findings would be easier to interpret; that is, ‘very comfortable’ to ‘very uncomfortable’ for our main outcomes of interest.

Following feedback from our lay advisory group and GP co-applicants, a revised set of questions was further refined through cognitive, think-aloud interviews with a small number of GPs not involved in the study. The final set of questions were piloted within the format of the online survey (see Report Supplementary Material 6).

Data collection

The market research company, medeConnect, draw their sample from GPs who are registered with Doctors.net.uk and who have been active on this online platform. Opportunity to take part in the monthly survey is posted on the home page of all GPs who have been active on the site within the previous 90 days and an e-mail is sent inviting participation to all GPs who have given consent to be approached to take part in research. Each GP can only participate once. Doctors.net.uk undertakes weekly automated checks of its membership base against the publicly available licensed status of practising doctors in the UK as published and maintained by the General Medical Council. Before a survey starts, the questions are tested by the internal research team, the sponsor (in this case the research team based at Warwick) and a small sample of the target group (Miles N, personal communication, medeConnect, Abingdon, 16 February 2023).

Analysis

For a binary outcome on our key question of interest (very comfortable or comfortable completing a plan vs. all other responses) a sample size of 1000 would, if 50% were ‘comfortable’, provide precision of 6.2%, or if 80% were comfortable a precision of 5%. This is the size of the medeConnect monthly survey and hence our original sample size. However, this includes UK GPs based outside England who were not covered by our research ethics approval. Thus, a smaller number of responses was expected. We present descriptive statistics for each question. For the outcome of how comfortable GPs are in having emergency care and treatment plan discussions, with patients or with someone close to the patient, using the statistical software R Core Team (2023), we first did univariable logistic regression analyses at the 5% significance level with gender, GP role, NHS region, type of area (major conurbation, large town/city, medium town/city, small town/city, hamlet) years since completion of GP training, and use of ReSPECT form versus DNACPR/Other as explanatory variables. We then constructed a fully adjusted logistic regression model. We repeated this using a backward elimination approach to select the most statistically significant variables.

Recruitment

General practices and their staff

Three CRNs recruited practices across three diverse CCG areas in England (Figure 1). During site initiation visits, we obtained consent from the designated principal investigator at each site for their practice to be involved in the study. We arranged a suitable time for the researcher to spend a day at their practice to learn about their ReSPECT procedures from clinical and administrative staff. Each principal investigator was asked to identify two to three GPs and/or specialist nurses who were involved in conducting the ReSPECT process in their practice. Invitations to participate were e-mailed and interviews arranged.

FIGURE 1.

Recruitment flowchart for GP practices.

Data collection

Participating practices identified patients with ReSPECT form completion documented in the previous 12 months and provided anonymised demographic data for these patients to the study team.

During visits to GP practices the researcher aimed to find out about the ReSPECT process from initiation through to form storage. Anonymised field-notes of behaviours, actions and conversations were taken and typed up and expanded after each visit.

All interview and focus group participants were provided with a participant information sheet and blank consent form to read in advance. Focus group participants received a shopping voucher as a thank you for participation. Interviewees were given the choice of remote [telephone, Zoom™ or Microsoft Teams® (Microsoft Corporation, Redmond, WA, USA)] or face-to-face interview. Focus groups were held remotely (using Zoom). For face-to-face data collection, we obtained written consent. For remote data collection, the consent form was read out and consent recorded by the researcher for each item. During practice visits, we obtained and documented verbal consent at the start of each conversation. All interviews were audio and/or video-recorded, apart from one GP interview, one group interview with care home staff and one focus group with paramedics. Consent to audio-record was not given for these interviews; therefore, detailed fieldnotes were made instead.

Our topic guide for the semistructured interviews and focus groups (see Report Supplementary Material 5) was informed by our research questions and developed with input from the co-applicant team and lay advisory group. It covered experiences of initiating, conducting and organising the ReSPECT process in the context of their work, completing forms with patients/residents, managing any challenges or tensions arising from implementing ReSPECT, and how ReSPECT form recommendations were used.

Five researchers (CJB, JH, KE, CM and JW) carried out semistructured interviews with GPs, care home staff and other health and social care professionals. Focus groups were facilitated by one or two researchers (CM, JH, SR and KE). Researchers recorded and shared reflective notes about the data collection process to refine the topic guide to consistently address any new issues or topics emerging during the interviews.

Audio-recorded data were transcribed by a professional transcribing company. We checked and anonymised the interview transcripts and allocated each transcript a study identification number.

Analysis

For GP practices and care homes, we initially grouped the data by practice/care home then read and reread the data before writing a summary of how ReSPECT was approached and managed within each practice or care home. Then for each interview/focus group, we used an a priori coding framework developed using the stages of the ReSPECT process and the main constructs of NPT: coherence building (conceptualisation), cognitive participation (initiation and engagement), collective action (use and documentation) and reflexive monitoring (evaluation). 44,45 The study team (FG, CJH, CJB, JH, CM, AS) made several refinements to the coding framework to increase its relevance to the data set. We used the agreed framework to code in NVivo™ software (NVivo v1.6.1, QSR International, Warrington, UK), a purposive sample of 50% of the interviews with GP and care home staff (sampled across all practices and care homes and types of staff) and all focus groups/interviews with other health and social care professionals. For emergency department staff, we limited our analysis to data about their engagement with community completed ReSPECT plans. During the coding process, we wrote memos of the key messages and themes emerging from the data. CM performed intercoder reliability on four interviews early in the process. The team then read the remaining data, looking for new codes or additional nuances in the data. 64 The ethnographic fieldnotes were used to supplement the clinician interview data. We analysed the coded data in three groups according to the participant’s relationship to the ReSPECT process: GPs and their staff who take responsibility for completing ReSPECT forms in the community, staff of care homes as residents commonly have ReSPECT forms and the staff use them in emergencies, and other staff who engage with community-based ReSPECT processes and forms. Within each data set, we compared data coded for each stage of the ReSPECT process and report our analysis following the ReSPECT process. Finally, we compared findings across the data sets and summarised the results.

In presenting our results, we mostly use the term ‘patient’ as this is the term used by many of our participants. Care home staff use the term ‘resident’, so we use that term in reporting data from these participants. Home care staff use the term ‘client’. As the data from these participants are reported together with health professionals who use the term ‘patient’, we use the word patient in this section to avoid confusion. Unless we specify a clinical role, we use the term ‘clinician’ for registered health professionals.

The views and experiences of general practitioners and their staff of the ReSPECT process

The views and experiences of care home staff of the ReSPECT process

The views and experiences of the ReSPECT process among other health and care professionals who engage with community-based ReSPECT process and resulting completed forms

Evaluation framework for completed ReSPECT forms

We used a modified version of a tool developed during our previous study of ReSPECT in secondary care. 66 This was based on the ethical framework of accountability for reasonableness, which focuses on fair decision-making processes rather than on specific ethical principles. 67 The framework stipulates that decision-making should be transparent, based on reasons agreed as justifiable by the relevant stakeholders and open to review. 67 The process should be consistently applied to justify the ethical principle of equity. The evaluation tool is structured to answer the following questions:

1. Consistency: what is the variation in content in terms of the detail requested on the ReSPECT form and associated guidelines?68

2. Transparency: how complete is the information about the patient’s preferences, diagnoses, reasons for treatment recommendations and who participated in the ReSPECT discussion?

3. Ethical justifiability: do the reasons for treatment recommendations indicate that clinicians considered the patient’s preferences and clinical conditions, as well as the benefits and burdens of specific treatments? How complete is the documentation of patient capacity and who provided input to completion of the ReSPECT form?

The tool provides a structure for evaluating each ReSPECT form and providing counts of how many ReSPECT forms in the survey meet quality criteria.

Team members CJH, KE, CAH, AS and FG modified the existing tool and associated guidance to take account of changes to the ReSPECT form since our previous study (introduction of version 3). We each independently used the tool to evaluate three completed ReSPECT forms, using a total of eight forms, so several team members evaluated each form. We met to discuss adaptations needed and repeated the process again to ensure each version of our tool provided consistent evaluation of each domain (see Report Supplementary Material 7). The relevant version of the tool was then used to assess each completed ReSPECT form (version 2 or version 3).

Sampling and recruitment

The 13 GP practices identified adults registered with the practice recorded as having a ReSPECT form completed within the previous 12 months. Following eligibility screening, patients or their next of kin if it was known to the GP practice that they lacked capacity, were sent information about the study, that the study involved accessing their medical records and what to do if they did not wish to participate.

Data collection

A study researcher or member of practice staff accessed the records of all eligible patients who did not opt out of the study and identified whether they contained a ReSPECT form. Identified ReSPECT forms were scanned (hard copies) or electronically copied. Data collection took place from February 2021 to January 2023.

Data extraction and evaluation of each ReSPECT form

CJH extracted data and evaluated content of all forms. KE, CAH, AS and FG extracted data and evaluated 25% of the total. To ensure consistency, the team met to compare findings, discuss and resolve discrepancies after evaluating half of the forms and again after evaluating the remaining forms.

Ethics considerations

Data were collected with approval from the CAG (reference 21/CAG/0089). Personal identifiers were redacted prior to transfer to the study team using NHS encrypted e-mail or an encrypted NHS platform. Data were stored in password-protected files on a secure University of Warwick server.

Consistency of recorded decision-making processes

Of the 141 forms, 64 (45%) had a mark on the scale for patient preference [Balance of priorities (version 2)/What matters to me (version 3)]. Over half (81/142, 57%) of the forms recorded free-text personal preferences (54%, 37/68). For people with capacity, this increased to 69% (47/68).

Fifty-three forms (37%) had a clinician signature in one of the CPR recommendation boxes (42/55, 76%) for handwritten forms and (11/86, 13%) for electronic forms). Of the electronic forms, 54/86 (63%) had the CPR recommendation typed in place of a signature.

Most forms recorded free-text recommendations for treatments other than CPR (87%, 122/141) but the focus of the clinical recommendations (extending life/prioritising comfort) was less well recorded (52/141, 35%).

Most forms recorded that the patient or (if the patient lacked capacity) someone close to them was involved in discussing the ReSPECT form content (133/141, 94%).

Most forms (108/141, 77%) recorded the clinician’s role or registration number but less than half (64/141, 45%) included a signature. The senior clinician signature was only present on 14/141 (10%) forms. This may be because of confusion about where the clinician completing the form should sign if they are also the senior responsible clinician (usually the GP).

Confirmation of validity of the form was completed in 11 forms (8%).

Transparency

Nearly all forms (129/1411 92%) recorded the person’s health conditions. Some provided a detailed list of diagnoses; while others used terms, such as ‘frailty’ and ‘decreasing health’ without clarification, or used abbreviations (e.g. ‘AF, CKD, IHD, OA’). Only 27/141 (19.0%) forms documented reasons underpinning the clinical recommendations.

Where free-text personal preferences were recorded, text was often brief and usually referred to a preference for being kept comfortable or remaining at home. More detailed explanations were noted in version 3 forms where the clinician is prompted to record what the person most values and what they most fear/wish to avoid for example:

Values; ‘Being kept comfortable, familiar surroundings, husband and daughter present’. Fears; ‘Hospital admission, uncertainty’.

Most forms recorded free-text recommendations for treatments other than CPR (122/141, 87%). Most (94/122, 76%) provided detailed recommendations on what treatment should be provided in a range of situations, for example: ‘Not for resuscitation. Not for admission under any circumstances, apart for suspected fracture or head injury that requires A&E assessment’, while 28 (23%) gave general instructions, such as ‘Not for hospital admission’. The focus of the clinical recommendations (extending life/prioritising comfort) was less well recorded (52/141, 35%).

Ethical justifiability of recorded decision-making processes

Of the 26 forms where reasons for clinical recommendations were provided, 22 (85%) included a reference to the patient’s wishes. Of these, 6 (27%) did not complete any part of the personal preference section.

Only five forms included a reference in the recommendations to the patient’s condition and ability to benefit from future interventions. We found only one form which explicitly referred to the patient’s preferences and considered the benefits and burdens of treatment.

In 131/141 forms (93%), the person’s capacity was recorded. Version 3 requires clinicians to say in what way the person lacks capacity. Of the 23 relevant forms analysed, 17 (73%) provide a reason which includes in some cases reference to capacity ‘unable to process or retain information’ or a more general reference to the person’s cognitive status ‘dementia’ or ‘cognitive impairment’.

Among the people who lacked capacity, the involvement of someone close to them (relative or friend) was recorded on 39/59 (66%) forms and this person’s name and role was recorded on 25/59 (42%). Whether the person who lacked capacity had a legal proxy was recorded on 45/5 forms (76%).

Strengths and limitations

To our knowledge, this is the first survey of GPs’ use of ECTP. We obtained a high-quality data set with no missing data. Although we did not achieve our original target of 1000 respondents, this is still much larger than most GP surveys. 69,70 We have sufficient data to produce a precise estimate for all our outcomes of interest and a robust statistical model exploring factors influencing how comfortable GPs are having emergency care and treatment planning discussions. Our respondents were representative in terms of region, age, practice size, years since qualification and region. Nevertheless, using a market research company working through a commercially funded, free to use website for data collection may have introduced bias into the sample selection.

Our qualitative data provide us with evidence of how a wide range of clinicians and carers conceptualise ReSPECT and an in-depth account of how the ReSPECT process is operationalised in general practice and care homes. Recruiting clinicians and carers working in other community/care services and emergency departments was challenging; recruiting more participants may have provided further and different accounts of ReSPECT. We used NPT as a framework for our analysis but to gain traction on the bulk of the data which related to the ‘cognitive participation’ and ‘collective action’, we used the ReSPECT process itself as an analysis framework.

A strength of our evaluation of completed ReSPECT forms is that it evaluates the actual content of ReSPECT forms. However, ReSPECT forms are designed to be held by the patient, we were limited to copies of ReSPECT forms which GP practices had chosen to store in the patient’s clinical notes. Some practices stored electronic copies that did not include signatures, so we do not know what signatures were on the copy held by the patient.

Participant recruitment

Patients who had a ReSPECT plan completed in the previous 12 months were identified by the 13 GP practices who participated in the study (see Figure 1). Eligible patients and the next of kin of patients without capacity were sent information letters. Interested patients and relatives sent expressions of interest, which we followed up with an introductory phone call and a detailed information sheet and consent form (by e-mail and/or post); patient participants were told they could invite a relative to accompany them during the interview. Consent was taken before each interview, either written (in-person interviews) or verbal (telephone interviews). For verbal consent, the researcher went through the consent form with the participant, documenting their consent for each item. A copy of the completed consent form signed by the researcher was sent to each participant following the interview.

Data collection

We developed semistructured interview topic guides (see Report Supplementary Material 5). The questions were designed with reference to the NPT framework and focused on understandings of the ReSPECT process and form (coherence), narratives of the process from first consideration to completion and updating (collective action), views of the process (cognitive participation), experiences of the ReSPECT conversation and expectations as to how ReSPECT might be used (reflexive monitoring).

Participants were interviewed either in person or by telephone, depending on their preference. The interviews were conducted by one of three researchers (JH, CJH, KE) and were audio-recorded and professionally transcribed.

During the consent process, participants were asked if they could be contacted for a follow-up interview. If participants agreed, we checked with their GP surgery, at least 6 months after the first interview, if they could be approached. We then contacted eligible participants to ask if they would agree to an interview. Follow-up interviews were semistructured and were conducted by JH, KE and CJH by telephone. These interviews explored whether there had been changes in the patient’s health, whether the ReSPECT form had been used or reviewed, and whether the participant’s views of ReSPECT had changed.

Analysis

We initially intended to use framework analysis. The smaller than expected sample size meant inductive thematic analysis was more appropriate. KE coded the first interview transcripts using a semantic and latent approach to capture both direct and implied meanings. 73 Providing intercoder reliability, JH and CJH analysed between them eight interviews. After abstracting from initial to higher order codes, KE drafted candidate themes and met with CJH and JH to discuss these. Following their discussion, the themes were refined and critically reviewed and approved by the research team.

KE read through the follow-up interview transcripts, to ascertain whether participants reported any changes and whether ReSPECT had been used, either of which prompted further analysis.

Descriptive findings

Thematic findings

Strengths and limitations

A strength of this study is that it includes the voices of patients and their relatives who have had a ReSPECT plan completed in the recent past mitigating problems with recall. The patient voice adds richness to the overall qualitative analysis of how ReSPECT is used and experienced in primary care. However, the small number of participants recruited means that their views and experiences may not be a reflection of the wider patient population with a ReSPECT plan in place.

Participant recruitment and data collection

Interviews were conducted with GPs and nurses, care home managers and staff, and patients and relatives (refer to Chapters 2 and 4 for participant recruitment and data collection procedures).

Analysis

Interview data were analysed using a modified framework approach and the empirical ethics model of grounded moral analysis. 54,74 Six co-authors (AS, CB, CJH, FG, JH, KE) coded 12 interview transcripts (4 GP, 3 care home and 5 patient/relative interviews), with 11 transcripts intercoded. This initial coding identified semantic (explicit) and latent (implicit) references to ethics. 73 Using these initial codes and the group discussions that followed, AS developed a draft coding framework, including inductive and deductive codes. Topic guides for future interviews were refined to explore the ethical issues and concepts identified.

Interviews were summarised in individual pen portraits by KE (GPs and nurses, care home managers and staff, and patients and relatives) and CJH (GPs). 75 We created 37 pen portraits, capturing all patient and relative interviews (n = 13) and half of the GP/nurse (n = 13) and care home (n = 11) interviews. GP/nurse and care home interviews were purposively sampled to include data from all sites and different types of staff. These sample sizes allowed us to reach code saturation. 64 KE coded each pen portrait using the draft coding framework. AS intercoded 16 pen portraits (7 GP, 5 care home and 4 patient and relative interviews); comparisons showed a high level of consistency. During coding, KE and AS added new codes to the draft framework. To ensure meaning saturation, AS reviewed the interviews not summarised in pen portraits (13 GP and nurse and 11 care home interviews) to check for new codes (none identified) and extract illustrative quotes.

Using the pen portraits, KE outlined the ReSPECT process timeline as described by the participants, from preconversation to the enactment of recommendations, identifying ethical dilemmas at each stage and abstracting candidate themes from the codes to capture the salient ethical values for each dilemma. These themes were further refined and finalised in discussion with AS.

When is the right time to have the ReSPECT conversation? Respecting autonomy and protecting from harm

When deciding when to initiate the conversation, GPs and nurses weighed the risk of physical or moral harm, wherein a patient may be subjected to invasive and distressing interventions without their wishes being considered, alongside the risk of emotional harm caused by having the conversation before the patient and/or their relatives were ready. These competing risks posed a dilemma:

Sometimes if you go in too early with the [conversation], and not approach it in the right way … they can sometimes step back … but we don’t want them so unwell that they’re not able to be involved as part of that decision-making and really don’t want them to be going into hospital and just having them made.

GP 18, area 2

In working through this dilemma, unless the patient was imminently approaching death, GPs and nurses tended to favour preventing emotional harm. This was linked to GPs’ and nurses’ perceived need to maintain trust and rapport. GP and nurse participants spoke of postponing conversations if they felt the patient was not emotionally ready or had more pressing needs, and of introducing ReSPECT gently and gradually before discussing the ReSPECT form in detail. One nurse explained this approach led to better patient experiences and higher quality outcomes:

[I]t just feels like it’s a bolt out of the blue, you know, all of a sudden somebody’s talking to you about end of life, it might not be the right time for them and actually if you can invest a bit of time, you get the better quality ReSPECT Form or care plan than if you, you try and force something on to somebody.

Specialist nurse 5, area 3

Trusting and having good rapport with the healthcare professional who facilitated the ReSPECT conversation was important to patients and relatives. As explained by relative 9, although ReSPECT ‘was totally new’ to them,

Because I know the doctors well, I think the process was good. I, I trusted the doctor and we, we know, I know that she knows Mum well, I know her quite well, so we were able to do it on the phone. I think if, if it had been a doctor I didn’t know very well then it would have been much more difficult.

Relative 9

While deferring, waiting with, or gently introducing ReSPECT conversations was a preferred approach among the GP and nurse participants, care home staff favoured early completion to prevent physical/moral harm to residents in an emergency. Not having a ReSPECT form during an emergency could substantially harm patients. A harrowing example was narrated by a carer who, in her previous assignment as a home care worker, witnessed CPR being performed on the frail, elderly woman for whom she had cared – an event which both harmed the patient and traumatised the participant: ‘She was really, really fragile. Really fragile. And because she didn’t have the ReSPECT form, they had to resuscitate her. And I can hear the “crack” sounds’ (care co-ordinator/senior carer 14).

As several participants explained, not having a clinical background or decision-making authority, but still being held responsible by residents’ relatives, carers depended on the ReSPECT form for critical guidance:

Whenever we have a new admission … straight away [Name] will put them on a GP round next day … I can’t imagine how difficult it would be to look after someone if this was not in place … being a care home there’s only so much care and treatment you can provide.

Care home manager/deputy manager 11

Thus, folded into carers’ concern with preventing harm to residents was a concern with protecting care home staff, whose timing may not match that of the resident or their GP/nurse.

The conversation: balancing patients’ wishes, families’ wishes and what is clinically possible

ReSPECT is not a direct record of a person’s wishes, but rather of a shared decision-making conversation. However, when asked what ReSPECT meant to them, most patient participants said it was about their wishes; as patient 12 explained, ReSPECT ensured her wishes were recorded ‘in black and white’. Patient participants also described recording their wishes to reduce their family’s future caregiving and decision-making responsibilities (see Chapter 5).

General practitioners, while emphasising that ReSPECT focused on the patient, viewed family involvement in the ReSPECT process as crucial. When patients lacked capacity, families were essential in representing their wishes. But even when patients had capacity, GPs involved families in ReSPECT conversations to ensure that patients’ wishes, reflected in the recommendations, would be followed. For example, one GP explained that a patient’s ReSPECT form.

needed to be redone so that the family were on board with it really and so that they understood the process and so that they were happy to follow through with the ReSPECT form and not call an ambulance if she deteriorated.

GP 3, area 2

Despite the weight given to relatives’ views, patient participants were clear that their wishes should prevail over those of their family. For example, one patient described negotiating with his children, who disagreed with his treatment preferences: ‘I did explain to them that I didn’t want to be brought back if I was going to be like a cabbage … and they sort of thought about it and then they agreed’ (patient 11).

In most cases described by the GP and nurse participants, patients’ and relatives’ wishes matched what was clinically possible; for example, terminally ill patients asked not to be admitted to hospital or receive CPR. However, when patients’ expressed wishes opposed clinical judgement, this presented an ethical dilemma. If patients requested treatment likely to be unavailable, not beneficial or even harmful, GPs and nurses had to consider what counted as an ethically justifiable constraint on patient autonomy. Faced with this dilemma, some GPs and nurses explained to patients that their wishes may have limited impact on future decisions:

[S]ome people are still very much, like, ‘No, I want all treatment to be offered’, and I say, ‘Well, it’s kind of up to the clinician at the time and I can make your views clear in the form, but I also have a responsibility to make the clinical judgment of what your other comorbidities are and what your baseline is in order for that decision to be made’.

GP 9, area 3

Other GPs and nurses preferred to defer or avoid recording specific patient wishes, often suggesting they revisit the conversation. One GP describing this approach explained they tried to respect patient autonomy while being honest about what was possible:

If a person were telling me that they wanted to be resuscitated and I knew it wasn’t physically possible or it [would be] detrimental, I might, although I might not make the DNAR decision, I might make it clear that it’s something that we might have to come back to or it might be a decision that’s made medically if they were to be unwell.

Specialist nurse 5, area 3

Another GP described recording a ‘for CPR’ recommendation to establish trust and open possibilities for a future conversation and reversal of decision:

But again, sometimes it’s helpful to do the form, put the [for] resuscitation, get their trust, get them to know that you’re not trying to force that on them, and then as, once you have got that trust and as you have further conversations and you can slip that in, they will then listen to those things.

GP 19, area 2

Some participants, however, described deferring to the patient even if their wishes opposed clinical recommendations, citing patient autonomy. For example, one GP did not record a ‘not for CPR’ recommendation for a patient who had terminal cancer because the patient had capacity: ‘It felt like it was a bad decision, but he was, had the capacity to make that bad decision at the moment’ (GP 11, area 2).

In another example, a GP described deferring to the wishes of the family of a patient lacking capacity, despite their own clinical judgement:

[I]t seems crazy to do lots of tests and investigations and charging around for this sort of very elderly patient who’s really who is really like quite frail and demented, who wouldn’t really probably want this and there wouldn’t be any benefit anyway. But the, the relative wants that and so therefore it just has to be put down. I mean I guess the patient, the, the patient or the relative can, you know, they have their own autonomy.

GP 12, area 2

In an emergency: how should recommendations be interpreted? The challenge of acting in the patient’s best interests

In an emergency, if a person lacks capacity to contribute to a decision, healthcare professionals have an ethical and legal obligation to make a decision in the person’s best interests; ReSPECT recommendations inform this best-interests decision. Our data show that ReSPECT recommendations can help determine what is best for the patient and ensure a decision is person centred. However, uncertainties inherent in anticipatory recommendations, ambiguities in how recommendations are articulated and conflicting views over interpretation can lead to patient harm.

Among the patient and relative participants, there was near consensus that ReSPECT form recommendations would be followed in an emergency. For example, one patient said that while his preference for comfort care may not be followed immediately, it will be followed eventually: ‘there might be a delay agreeing what I want … which would then change as the message leads the acute medical team knowing this person has a ReSPECT form in place’ (patient 3). In practice, however, care home managers and GP and nurse participants said that recommendations recorded on ReSPECT forms were not always followed. GP and nurse participants explained that ReSPECT recommendations left space for clinical decision-making in the patient’s best interests on a case-by-case basis. They conceptualised ReSPECT recommendations as guidance and acknowledged that, as the recommendations could not encompass every clinical scenario, there needed to be some flexibility in interpreting these recommendations, either implicitly or explicitly: ‘I’ve put that care plan into place with the proviso of that’s, if in the event of a fall or a head injury if the clinician at the time can assess if he does need to be admitted’ (GP 8, area 3).

General practitioner and nurse participants said that, at times, interpretive flexibility was stretched too far, with recommendations not being followed and thereby causing harm to patients, through either not respecting their wishes or providing inappropriate treatment. One nurse described feeling disempowered after a patient died in hospital, having been admitted contrary to carefully worded, detailed recommendations: ‘She didn’t actually receive anything in hospital that we couldn’t have done at home … So it’s left me sort of like thinking, “Well, what could I have put on there that makes any difference?”’ (specialist nurse 3, area 1).

Many care home staff participants perceived ReSPECT recommendations as prescriptive, feeling that deviations could potentially harm patients. This could lead to conflict with the healthcare professionals who had decision-making authority in emergency situations. As one care home manager described: ‘We do tend to speak up and try and fight for the residents and say, “No, this is what it states, this is what should be happening”’ (care home manager/deputy manager 10).

When ambulance staff did not follow ReSPECT recommendations, some carers felt this compromised person-centred care:

[O]ften we find that if the ReSPECT form says ‘not for admission to hospital’, that the healthcare professional will make their own decision based on their assessment at that time. And then the, that person could end up going to hospital … And I find that confusing and a bit demoralising at times.

Care home manager/deputy manager 5

Care home staff also felt that healthcare professionals’ insistence on following ReSPECT recommendations despite their apparent inappropriateness could cause patient harm:

And we have had a bit of resistance from paramedics saying no, their form says not for hospital admission. But if that person needs to go into hospital to be hydrated because they’re dehydrated, or to have intravenous antibiotics, that’s where we get a little bit of resistance.

Care home manager/deputy manager 2

[T]he amount of times we’ve had paramedics come in and refuse to take someone in for sepsis because their [form] says, ‘Not for treatment’, is ridiculous … We, we waste hours and hours of management time trying to get another ambulance out … they don’t understand, because documentation is not clear enough, and the guidelines are not clear enough.

Care home manager/deputy manager 9

When conflict arose in emergency situations, carers sometimes called on family members to influence the decision. One staff member described how they would ‘ring the families then, and then the families are, like, “Yes, but they can go for this”, so then they’re going’ (care home manager/deputy manager 11) – an example that captures the family’s central role in enacting ReSPECT recommendations.

At its extreme, underinterpretation (or misinterpretation) of the ReSPECT form could lead to patient harm through equating the form with DNACPR. This worried several GP and nurse participants:

My biggest worry is that when somebody says, ‘Have they got a ReSPECT Form?’ they’re meaning, ‘Have they got a DNAR decision?’ … And to be fair, I tend to not fill in a ReSPECT Form if I’m not doing a DNAR decision as well for that reason.

Specialist nurse 5, area 3

Another type of conflict was described by a GP whose palliative patients asked to die in hospital: ‘they are scared of dying and their family are scared of dying and they don’t want to see that’ (GP 19, area 2). Despite recording a ‘for admission’ recommendation, hospital clinicians refused to comply, raising questions about how the person’s best interests were understood: ‘I definitely had experience with patients who have wanted to die in hospital, and you ring and try and admit them and they’re like, “What? They, they’re palliative. What, what are we going to do for them?” ’ (GP 19, area 2).

Aim

To understand how health and social care professionals can optimally engage people with a learning disability in the ReSPECT process and to co-produce relevant support materials.

Research questions

1. What are the views of people with a learning disability about ECTPs in general and the ReSPECT process in particular?

2. What are the views of carers of people with a learning disability about emergency care treatment plans in general and the ReSPECT process in particular? What might be the challenges for people with a learning disability and those working with them to access and engage with emergency care treatment planning?

3. What kind of information, processes and support around emergency care treatment planning and ReSPECT would be useful for people with a learning disability and those caring for or supporting them?

Sampling and recruitment

Workshop development

The reference group met before the workshops began to familiarise themselves with the research and to plan the first workshop, then met in between workshops to receive feedback from the previous workshop(s) and plan the next (see Report Supplementary Material 9 for flowchart of meeting activity).

The researcher (JL), who was embedded in both the reference group and workshop process, linked issues raised from the workshops and drew on findings from other areas of the wider study to sustain critical reflection with reference group members. Worksheets and materials co-produced by reference group members and the CHANGE team were also included in the iterative development of the workshops content (see Report Supplementary Material 10). With consent, meetings were audio-recorded to supplement fieldnotes made by the researcher.

Co-production of workshop outputs

Researchers (JL and AR) reviewed all workshop summaries to identify key messages and suggestions around information, process, and support for engaging with the ReSPECT process. The researcher (JL) then presented these to the reference group, with a simplified summary of the report and recommendations from the stakeholder meeting. Two reference group members, one workshop participant who had experience of work on ACP, together with JL and the CHANGE team, used this information to establish the content and format of the recommended outputs. These were drafted by JL, reviewed by AR and AS, and then submitted to CHANGE to convert into easy-read format. The full reference group reviewed the draft versions and recommended minor modifications. Additional comments were sought from a further two workshop group members with a specific interest in supporting end of life decision-making for people with a learning disability.

Reflecting on the process: collaborative autoethnography

Following completion of the workshops, a separate meeting was held in which reference group members evaluated their experiences of working on the project. This was facilitated by JL using a series of questions to aid reflection on the reference group process from the personal perspectives of the people with a learning disability who took part.

The researcher (JL) began by asking the question; ‘What do we think and feel about what we did?’ with the following prompts:

• What meant the most to you?

• Why did this mean the most to you?

• Has listening to other people’s stories affected you?

• In what way?

The reflective process drew on the model of collaborative autoethnography. Collaborative autoethnography extends the idea of autoethnography by involving multiple participants in the process, encouraging the collaboration of multiple voices and perspectives to create a multifaceted account that reflects the complexity and diversity of the shared experiences. 84 The researcher (JL) who led the reference group participated in the collaborative autoethnography.

Workshops

Workshop summaries were analysed using thematic analysis. 73 JL developed summaries based on the workshop recordings and her reflections and notes from each workshop. JL and AR read a sample of summaries and identified codes: (1) related to information, process, and support for engaging with the ReSPECT process (which fed into development of resources) and (2) reflecting the meaning of emergency care treatment planning for people with a learning disability. JL coded all summaries and drafted candidate themes which were then discussed and refined with AR.

Collaborative autoethnography

With consent the reflective discussion was recorded. The recording was transcribed by JL and coded by AR. JL reviewed the coding and any disagreements were discussed and resolved. AR then re-read the coded transcript and identified key themes following Braun and Clark. 85

Workshops

Analysis of the workshop summaries identified three themes: autonomy, trust and process.

Development of resources to support health and social care professionals to engage with people with a learning disability about the ReSPECT process

The development of materials was a staged process to evaluate the explanation of what ReSPECT was, the setting, the preferred format of any resources, key information to be included and suitable terminology to use when explaining ReSPECT and health emergencies. In the workshops, multiple methods were used to support the participants to evidence their thoughts and feelings about the ReSPECT process and to highlight what they felt should be the key elements of any discussion of ReSPECT.

We found that the six-part story method was a popular way to think about the ReSPECT conversation and enabled workshop members to articulate their needs and concerns before, during and after a ReSPECT conversation. 86 This format was then used in the final materials. The concerns and needs of people with a learning disability expressed in the workshops fed into the development of a set of questions or prompts for professionals to discuss with a person with a learning disability to ensure all reasonable adjustments needs are considered and all concerns would be addressed during the process of the ReSPECT conversation.

Reference group reflections on the co-production process (autoethnography)

Analysis of the recording of the reference group reflections identified four themes.

Recruitment

We sought expressions of interest for focus group participation from carers of people with a learning disability through CHANGE’s national networks of learning disability and learning disability carers’ support organisations, and through online carers support forums and carers groups identified through social media platforms. We provided an invitation e-mail describing the study for circulation through these groups. Potential participants contacted the study team directly and were provided with further information and offered a range of dates for planned focus groups. If they were unable to join a planned focus group, they were offered an individual interview. Participants were given a unique link to a Zoom meeting for the relevant focus group or interview. Prior to the focus group or interview, participants were sent a copy of the ReSPECT form (see Report Supplementary Material 1) and consent form. Initially, consent was taken at the start of each focus group with each participant in turn. However, this process proved challenging and time consuming online so for focus group 3 onwards a separate online/telephone discussion was arranged between the researcher and participant to obtain consent. Consent for individual interviews was taken prior to the interview. Demographic questionnaires were either returned by focus group participants by e-mail to the study team, completed on the telephone or were completed during individual interviews. Participants received a shopping voucher as a thank you for participation.

Data collection

Focus groups were facilitated by up to three researchers (JL, AR, JH). All individual interviews were carried out by JL. At the beginning of the focus group/interview the researcher briefly described the purpose and format for the discussion and the ReSPECT process. A topic guide was developed by AR and JH and refined by a family carer and chief executive officer of a charity supporting people with learning disabilities: Special Needs Outreach Project (SNOOP). 88 The topic guide was used to structure and facilitate the discussion (see Report Supplementary Material 5). The topic guide prompted participants to consider potential benefits and disadvantages of emergency care treatment planning for people with learning disability and their carers; how an emergency care treatment planning process could be implemented with people with learning disability, the role of carers in the process; and potential benefits for the wider health service.

Focus groups lasted between 38 and 76 minutes and interviews between 45 and 62 minutes. All focus group and interviews were recorded and transcribed.

Problems with online connections and participant engagement meant that no reliable data was obtained from one of the focus groups (four participants) so this was excluded from analysis.

Analysis

Jacqui Lovell and Amy Russell coded focus group transcripts and the joint interview transcript using the topic guide areas as a framework and then met to compare and refine codes. Following agreement of codes, JL developed a coding framework and coded all remaining interview transcripts against this framework. Reading across the coded data, themes were identified which were discussed and refined by JL, AR and AS.

Process: (the process should acknowledge the complexity of care and caring for someone with a learning disability)

In general, carers were supportive of ReSPECT planning as a way of providing clear guidance to clinicians in an emergency when the carer may not be there to support the person with a learning disability. They also saw it as a means of helping healthcare professionals in an emergency when they don’t know the patient.

So yes, in that regard erm it could say all the things that you, you want it to say and you could put your numbers down, your involvement down, you could put a second number down if you weren’t available and all that.

Parent-carer interview 4

And also with [person with a learning disability’s name] you can’t learn anything straightaway because he might start gabbling and you think, ‘Oh, he’s just ... is he ...’ you don’t expect him to be, have the, the disabilities that he does have, you know, so ... yeah, it’s got to be a help. It’s got to be a help, having some more information.

Parent-carer interview 6

Carers’ views on the process of ReSPECT planning reflected their role in understanding and advocating for the person with a learning disability and the complexity of care needs to be considered. Carers were clear that they should lead the ReSPECT process including deciding which clinicians should be involved.

Our loved ones are known to a multidisciplinary team and I think we knowing our young people as we do should be the ones to decide who we elect to be that clinician who knows them best rather than one-size-fits-all GP.

Parent-carer interview 4

They emphasised that the ReSPECT conversations needed to be collaborative, possibly involving several family members, and that they required sufficient time to do this properly citing the need for a number of ReSPECT meetings before the completed form is signed. One participant suggested that ReSPECT conversations should follow the format of a best-interest meeting for a person who lacked capacity.

It would actually be it would have to be, it’s got to be over a certain amount of time. It’s not something that can be just, just done in one appointment … and I want to rip it apart so that I know I’m understanding absolutely everything that’s on there so that there’s no regrets.

Parent-carer focus group 2

Different people have different views on what is a meaningful life, and they make decisions based on that … and therefore the best interest meeting, who attended, what was discussed and what were the conclusions is essential in order to make sure that there is no cause for concern.

Sibling-carer focus group 1

All carers felt that it was important to do this process early and review regularly.

Lack of trust in the system

While carers were generally supportive of the ReSPECT process, they expressed doubt about how the process would work in practice. Their past experiences of direct discrimination toward themselves as carers who were not trusted to know and understand the needs of the person in their care, and a general disregard of the quality of life that people with a learning disability may have, influenced carers’ lack of trust in the system. A recurring concern for carers was whether ReSPECT recommendations would be acted on in an emergency. They lacked confidence in the efficiency with which ReSPECT plans would be available when needed.

but sometimes you go to A&E and they or you go to, you know, and they can’t even access a doctor’s letter, can’t you bring your records up. So, God only knows if they’reable to, to, to find this ReSPECT form.

Parent-carer interview 3

Would they ask? Would they ask whether there’s one in existence or not? Would they ask that question?

Parent-carer interview 2

Even if the plan was available, carers expressed fear that recommendations may not be followed unless they were present to reinforce them: ‘I’m, I’m not so sure it would be followed. I would like to feel it would be followed but what, what proof have I got that it’s ever going to be followed’ (parent-carer interview 5). Indeed, one person expressed surprise that the ReSPECT plan was not legally binding.

I think it’s appalling that it’s not a legally binding contract. It needs to be because that security needs to be given in a really difficult time to them families … that you know, yes, we will do what you’ve asked for.

Parent-carer, focus group 2

However, some carers felt that if they had good a relationship with their doctor, they would be confident in trusting that the plan would be followed.

I, I think if my doctor was proving to me that he was being a very good GP and this was going for a couple of years, I’d have no reason to mistrust him. I would then assume that if he’s involved with this plan that he was going to stick by his word and do what he needed to do.

Parent-carer interview 4

Carers as advocates

Carers saw value in the ReSPECT process, but primarily as a secondary enforcement of their own advocacy role and to counter clinician stigma around learning disability. They felt it could be used to indicate their support network and care needs, so the person was not hospitalized without an advocate who knew them: ‘I do have an inherent fear of [son’s name] just being admitted to A&E, somebody forgetting to tell me and then doing whatever they want to him’ (parent-carer interview 4).

For some carers, their role as advocate negated the need for a ReSPECT plan.

So I don’t see the need for one for my family … in the sense that if, if there is an emergency then you know I’m going to be there, I’m going to be talking to them … you know you will have to deal with the emergency incident on its merits at the time and that piece of paper that ReSPECT is almost historical.

Parent-carer interview 3

This carer’s strong sense of advocacy for the person in their care also influenced how they perceived completing a ReSPECT plan for themselves.

I think I would not want anything that is going to make my family’s life more difficult. And for me, [son’s name] is the priority. Therefore, I don’t want anything for me that’s going to make her [daughter] feel, ‘oh Lord, now I’ve got to look after mum too’, so I’m not entirely sure I would want something that’s enabled me to survive but disabled me entirely.

Parent-carer interview 3

Strengths and limitations

This WP has embedded the voices of people with a learning disability and their carers in an evaluation of the use ReSPECT in primary and community care, ensuring that their perspectives and experiences inform future developments. Collaboration with CHANGE strengthened the project, providing expertise in working with people with learning disability and production of easy-read resources and information.

The reference group initially struggled to retain members due to their own personal circumstances and had to enter a second round of recruitment. This meant the group had less time to work together and bond which may have impacted on their effectiveness and their experience of participation. Recruitment to carer focus groups was challenging and technical issues meant that online engagement was intermittent for some participants.

The consistent messages from across our results

The public survey indicates overwhelming support in Great Britain for people being able to have an ECTP. 55 Our qualitative results indicate similar support from all participant groups. 89 There was consistency in the results related to key issues from across our participant groups. All agreed that ReSPECT provides a valued opportunity for patients to express their wishes to the clinicians making treatment decisions, with parent/carers of people with a learning disability seeing it as a means of reinforcing their voice as an advocate for the person they care for. The qualitative data indicate ReSPECT is commonly understood as primarily about resuscitation, including among patients and their relatives. Its role for providing other clinical guidance particularly for patients nearing end of life is also recognised. Across all data, there was agreement about when to initiate an ECTP such as ReSPECT: when there was worsening health especially if the person had a life threatening or life limiting condition. Family members are willing to be involved in completing ECTPs such as ReSPECT and their involvement is considered important to prepare them for potential medical emergencies, recognising that not everyone wants to involve their family in healthcare decisions. For people with a learning disability the involvement of a family member or carer was seen as essential to advocate for or to support the person in the planning process. The range of health and social care professionals who could lead ReSPECT conversations could be broader than at present, if skilled in communication and trained in the ReSPECT process.

There was concern among the public and all those working in health and social care in the community that ECTPs such as ReSPECT could be out of date. On the use of ECTPs such as ReSPECT in an emergency, the consistent message from the qualitative results was that it was problematic. The public and GP survey both asked about the possibility that having an ECTP might result in someone not receiving lifesaving treatment and substantial minorities agreed. Our qualitative results from GPs and their staff and other participant groups working in community health and social care indicate that completed ReSPECT plans are not always perceived as an important influence on treatment decisions in a medical emergency in the community. This perception has potential to undermine the ReSPECT process in the community.

The role of the ReSPECT form in a medical emergency

The ReSPECT form is designed to assist with decision-making across the whole range of potential medical emergencies that can occur in the community. We first consider the role of ReSPECT in the community when a patient appears to have collapsed.

Our home-based care workers and care home staff provide insight into the experience of staff who are not clinically trained when a person they care for appears to be unresponsive. They know the first decision is whether to start CPR and, here, the recommendation recorded in the ReSPECT form is useful. However, they also need help from a clinician as they do not know the cause of the unresponsiveness and whether it can be treated. A call to the emergency services may elicit the request to start CPR, which creates for them a dilemma if the ReSPECT recommendation is not for CPR. If the patient has become severely ill, care staff need help from a clinician to work out what the problem is before the clinical recommendations on the ReSPECT form become useful. Care home staff experiences of a lack of understanding of their dilemma from GPs and paramedics caused them distress. There was no mention on any of the ReSPECT plans we examined of information about which service care staff, or family, should call in a medical emergency.

Paramedics and emergency department doctors attending patients who have apparently collapsed or become severely ill make their assessment and treat following established clinical pathways taking account of patient safety. ReSPECT plan content rarely provided sufficient justification for deviating from these pathways. Although there was recognition among clinicians and care staff that it is the attending clinician’s responsibility to make the clinical decisions, the perception that the attending clinicians ignored ReSPECT recommendations was a source of frustration for care home staff. However, those who completed ReSPECT plans struggled with what to write in the plan that might help the decision-making of the clinicians attending their patient in an emergency. For the attending clinicians knowing patient preferences was valued.

For the ReSPECT process to continue with support from the public, patients, families, clinicians and care staff, we need to find out if it is possible to ensure the content of a completed ReSPECT plan is useful in medical emergencies in the community, and if so, how.

Control of the sharing of ReSPECT forms

The ReSPECT form was designed to be patient held. Our survey results suggest widespread understanding of the role of the ReSPECT plan in ensuring clinicians know the patient wishes. For this to happen, the patient must share the ReSPECT plan with the relevant clinician. There are patients and families who understand this, perceive the plan as having considerable power to influence treatment decisions and actively control access to their patient-held ReSPECT plan, making it accessible or inaccessible to other people. Patients and families active in managing their ReSPECT plan are likely to ensure that it is up to date. Our results indicate that some patients do ask for the plan to be updated. Some patients and families appear not to be active in controlling the sharing of the ReSPECT plans. There is some evidence that they forget about them. Clinicians and care home staff expressed concern that repeated reference to emergency care treatment planning through regular review could cause distress for patients and families who prefer not to think about their plan.

Based on their experience of lost, out of date or inaccessible forms clinicians and care staff felt an electronic form available to all relevant health and care staff would solve the problem of form accessibility and validity. However, this would affect patient control of the form. Our qualitative results include concern that people can change their mind as their circumstances, including their health status, changes. Both surveys identified the risk of the form being out of date.

The tension between patients/families maintaining control of their completed ReSPECT plans and the need for clinicians and care staff to access the plan when needed is a challenge for the current ReSPECT process.

Patient preferences recorded on ReSPECT form

There was agreement on the value of patient preferences recorded in the ReSPECT plan and the importance of clinicians knowing the patient’s wishes when making treatment decisions. People with a learning disability emphasised not only patient wishes but also clinicians knowing what was important to and about the person to minimise clinician bias when making emergency treatment decisions. However, among completed ReSPECT forms we evaluated, we found completion rates of the patient preference scale were under half, and completion of free text preferences only just over half. When text was recorded it usually referred to being kept comfortable and place of death. Our qualitative results suggest the preference scale was not always useful. The public focus groups drew out the tension between the need for text to be brief to be read in an emergency, and the need to communicate details about preferences. Care home staff and GPs talked about ACPs or other care plans holding more detailed information about patient preferences.

Together, these results suggest that, apart from patients on a relatively clear trajectory towards death when options are limited, expressing preferences about medical treatment is not straightforward, which may in part be because the options are not clear. This is a challenge for the ReSPECT process in the community. However, version 3 of the ReSPECT form, which prompts clinicians to record what the patient most values and what they most fear/wish to avoid may have gone some way to meeting this challenge as clinicians had recorded more detail on these forms than on version 2. However, our sample included few version 3 forms and our other data did not distinguish versions of the form.

Clinical recommendations and reasoning recorded on the ReSPECT form

Most of the ReSPECT forms we evaluated contained some detail in the clinical recommendations text box and recorded whether the patient was recommended for CPR. However, clinicians who wrote clinical recommendations talked about their struggle with this, in part because of the wide range of possible medical emergencies that could arise in the community.

The prospect of writing clinical recommendation text influenced the timing of some ReSPECT conversations, as over time as a patient’s health deteriorated, treatment options reduced, and patients and families came to terms with the patient’s condition. However, not having a ReSPECT plan risks patients having difficult treatment experiences including CPR.

Very few ReSPECT forms we reviewed included explicit reasons for the treatment recommendations made with reference to the patient’s condition and ability to benefit from future interventions. Only one included patient preference in their reasoning about benefits and burdens of treatment. Our qualitative data suggest that the ReSPECT process is challenging, and clinicians need training to do it well. In the public survey, one-quarter of respondents wanted a clinician who did not know them but was trained in making ECTPs to complete the plan with them. This minority view suggests a dilemma about whether a clinician who knows the patient well, or one that is highly trained in ReSPECT, is best placed to undertake the ReSPECT process.

In the community, the completion of medical recommendations in ReSPECT plans is challenging because of the level of medical uncertainty around future clinical events and treatment options. This uncertainty reduces as someone’s disease trajectory progresses or they near the end of their life, but waiting to complete a ReSPECT plan until this medical uncertainty is reduced risks increased treatment burdens for patients. A further challenge is to bring together medical uncertainty and patient preferences for weighing of benefits and burdens to make clinical recommendations.

Patient trust and ReSPECT

Across our data there was agreement that the person holding a ReSPECT conversation with a patient needed good communication skills and to establish a trusting relationship. Although half of the members of the public surveyed indicated they would want their GP to complete an ECTP with them, other results suggest a wide range of types of professionals who could hold ReSPECT conversations. What was important was a trusting relationship between the professional and the patient.

Qualitative results from our patients/relatives’ interview data indicate they trust that the recommendations on the ReSPECT form would be followed which other results suggests is not necessarily the case. Parent-carers of people with learning disabilities were less trusting that preferences would be followed based on their past experiences of clinical decision-making in an emergency. The public focus group results remind us that patients have little choice but to trust the clinician in a medical emergency and the health service to provide the most appropriate care.

The ReSPECT process requires trust between patients, their families and clinicians but also trust between different healthcare professionals. Engendering patient trust in the ReSPECT process depends on patients and public trusting the health system more generally.

Initiating the conversation

Stakeholders thought it was important that ReSPECT planning should not only be considered in the context of end of life or palliative care. They suggested other potential triggers for initiating a ReSPECT conversation, for example, when a person has specific risk factors for severe acute illness, or during regular reviews of someone with a chronic disease. They distinguished between conversations to prepare someone for discussion of a plan and the planning conversation itself, suggesting that preparation conversations can be initiated by a wide range of people, with appropriate training and support.

Having and recording the conversation

Stakeholders identified several challenges that could occur in the process of the ReSPECT conversation, including lack of time to have what is a difficult and complex conversation. A strong message from the meeting was that ReSPECT conversations needed to be open and honest, and that the healthcare professional should take time to understand what was important to the person and what their preferences were. For good conversations to occur, communication challenges needed to be considered, such as language barriers, and alternative formats for the form. There was a consensus that a digital version of the form was the preferred option to ensure accessibility.

Articulating and interpreting recommendations

Stakeholders agreed it was important that recommendations, including the reasons for them, were understandable to the person and their family and to the health and social care professionals who interpret them in the emergency situation. As it is not possible to cover all potential scenarios, emphasis was placed on the need to carefully document the person’s values and preferences and for healthcare professionals to take account of these when interpreting recommendations. Stakeholders noted the difference between hospital completed recommendations, which were specific to hospital-based scenarios, and primary care-based recommendations, highlighting the need for review as a person moves between healthcare settings.

Training and raising awareness of ReSPECT

Stakeholders agreed that there is a need for training around the whole ReSPECT process from preparatory discussions through to use of the form in an emergency, with training content and format being tailored to the needs of the different groups of health and social care professionals who may be involved in the ReSPECT process. They highlighted the importance of training in interpreting ReSPECT recommendations in an emergency. Examples of training content and format were provided. There was consensus in the meeting of the need to raise awareness of ECTPs including ReSPECT within the general population.

Recruitment

We worked with the ENRICH team in West Midlands to recruit three care homes not involved in the main study. Initial information about the feasibility study was provided through ENRICH to research active care homes in the area. Care homes expressing an interest were put in touch with the study team by ENRICH. The study team provided a detailed information sheet and arranged an online meeting between a senior member of the study team and the manager of the care home to discuss the study and answer questions. We received approval from the HRA Confidentiality Advisory Group for an extension of our existing permission under section 261 of the Health and Social Care Act 2012110 to access records without consent. For care homes who agreed to participate, a check was made by the care home to identify if any residents had registered under the national data opt out to ensure their records were not accessed during the study. The study team provided participating care homes with information letters about the study for all residents, and for the designated next of kin of those residents who lacked capacity. Letters were either given to residents or posted to their designated next of kin. The information described the purpose of the study, that the researcher would have access to their care record, and the ways in which the person could let the study team know that they do not want their care home records to be used in the study. No data collection occurred until 4 weeks had elapsed to allow a resident or their next of kin on their behalf to opt out of the study. Opt-outs were recorded in the study screening log kept by the care home and no identifiable data was accessible to the study team.

Data collection

Four weeks after delivery of the information, letters two senior study researchers (FG and AS) with a clinical background in general practice each visited one of the care homes and identified the records of all residents who had not opted out of the study. For each resident, the researcher checked the care home record and noted in a field spreadsheet the following:

• Presence of a completed ReSPECT form.

• Any record of acute medical event or significant deterioration in clinical condition within 6 months of ReSPECT form completion (or within the last 6 months if no ReSPECT form or ReSPECT completed more than 12 months previously and number of events if ˃ 1).

• If the record of decisions for each event sufficiently detailed to enable an analysis of congruence with a ReSPECT recommendation; that is, does the decision specify any treatment given or reasons for decision not to escalate treatment.

We did not review recommendations in ReSPECT forms as part of the feasibility study as a review of ReSPECT forms had been undertaken as part of the main study.

No personal identifying information or specific information about treatment decisions was recorded.

The researcher also identified with the care home manager the process for archiving and retrieval of care home records of deceased residents.

Following the data collection visits, the two researchers compared their findings and discussed the level of detail provided in the clinical records to agree if this was sufficient to perform a congruence analysis against a completed ReSPECT recommendation.

Recruitment challenges

Despite circulation of information from the ENRICH team, including personal approaches to research active homes, we received very few expressions of interest. Six sites expressed interest, of which two chose not to proceed after receiving further information and one withdrew because of manager turnover. To comply with the CAG requirements, all three recruited sites needed to obtain confirmation of compliance with NHS Digital Data Security and Protection (DSPT) Toolkit. 111 Navigating this process with NHS Digital proved challenging and one home withdrew at this stage. The remaining two participating homes were then required to check eligible residents against the national data opt-out records. This process involved several stages and was highly technical. The two recruited homes were part of a larger organisation with a dedicated data protection officer and information technology support, but even with this specialist support the process was time consuming and complicated. Overall, the time from initial recruitment to confirmation of data opt out compliance was 3 months. No opt-outs were recorded on the national data opt-out system for residents in the participating homes and no opt outs were received in response to the information letters.

Systems for archiving and retrieval of records of residents who have died

The research team confirmed with each care home manager that the home had a policy and system for archiving the records of residents who have died, and a system for retrieving the record if required for future review. Length of archive period varied but was at least 6 years. Therefore, it would be possible to access these records to compare clinical decisions around end of life with ReSPECT recommendations for residents who had died for a congruence analysis study.

Recruitment of care homes

Recruitment of care homes for this study, and the main study was challenging. In both studies, we found staff turnover to be a major obstacle to establishing a relationship between the home, ENRICH and the study team. Care homes are under immense workload pressure at present and any additional work may be too much for them to consider, particularly for small independent homes. Studies that involve accessing records without consent under a Section 251 of the NHS Act 2006 approval from HRA CAG add an additional onerous layer of work for care homes to obtain confirmation of DSPT compliance by NHS Digital and to navigate the complex technical system for identifying membership of the national data opt-out for eligible residents. For a future study, it may be advisable to recruit care homes within large national care organisations where there is likely to be central data protection and information technology support. Study teams should allow a long set-up period to ensure all homes can demonstrate compliance with the relevant data protection and data opt out requirements. However, exclusion of smaller independently run homes may bias study findings. Sufficient resources for dedicated and specific research support for care homes may lead to increased recruitment and retention of study sites.

Sample size requirements for a large congruence analysis study

In our feasibility study, homes 17/75 (22.7%) residents had a completed ReSPECT form and an identifiable acute event in the preceding 6 months. In our original sample size calculation for congruence, assuming congruence of around 75%, we needed data on 289 people with an acute event. On a simple calculation, this would mean that we needed to work with homes to recruit 1275 residents. However, the two study homes may not be representative of homes overall. To allow for this uncertainty in our calculations, we have used the lower limit of the 60% CI for events (i.e. 18.6%). This means that we would need to examine the records of at least 1554 residents to ensure that we identified 289 acute events. The mean capacity (number of beds) of all care homes involved in the ReSPECT study (main and feasibility studies) was 49.7, meaning we would need to work with 32 homes. Allowing for some uncertainty about the actual congruence rate level we anticipated that we needed to work with around 40–50 homes to complete this study.

Clarity of ReSPECT recommendations to enable congruence to be assessed (findings from main study)

To enable an accurate assessment of congruence between clinical decision-making and ReSPECT recommendations, there needs to be a clear record of the clinical decision-making in the patient record, and also clear and specific recommendations on the ReSPECT form. Our findings from review of completed ReSPECT forms and qualitative data from interviews with clinicians and care home staff in the main study suggest that recommendations are often not sufficiently specific to inform action in an emergency. The challenge of interpreting recommendations in the context of an emergency or acute clinical episode may also be encountered in evaluating the congruence of decisions with recommendations. Studies of congruence between anticipatory treatment recommendations in the POLST and clinical decisions have been carried out. 112 However, POLST recommendation forms are more structured and do not use free text recommendations which may facilitate the assessment of congruence. A review of methods to evaluate the impact of ACP on concordance between patient preferences for end-of-life care and the care they received identified several challenges. These included a lack of clarity over what is being measured, the use of medical documentation to assess patient preference, and failure to adequately deal with the complexity of the constructs being measured. 113 Based on our main study findings, we anticipate similar challenges in any congruence analysis to assess the impact of ReSPECT plans on patient treatment and care.

In summary, our feasibility study has demonstrated that the substantial logistical challenges in recruiting care homes to a study of this nature, and the large sample size that would be needed, mitigate against embarking on a large-scale study. Additionally, the conceptual challenges identified by Johnson et al. 113 in similar studies of ACP interventions suggest further research is needed to inform the development of an appropriate assessment tool.