Implementation of a complex intervention to improve care for patients whose situations are clinically uncertain in hospital settings: A multi-method study using normalisation process theory.

PURPOSE: To examine the use of Normalisation Process Theory (NPT) to establish if, and in what ways, the AMBER care bundle can be successfully normalised into acute hospital practice, and to identify necessary modifications to optimise its implementation.
METHOD: Multi-method process evaluation embedded within a mixed-method feasibility cluster randomised controlled trial in two district general hospitals in England. Data were collected using (i) focus groups with health professionals (HPs), (ii) semi-structured interviews with patients and/or carers, (iii) non-participant observations of multi-disciplinary team meetings and (iv) patient clinical note review. Thematic analysis and descriptive statistics, with interpretation guided by NPT components (coherence; cognitive participation; collective action; reflexive monitoring). Data triangulated across sources.
RESULTS: Two focus groups (26 HPs), nine non-participant observations, 12 interviews (two patients, 10 relatives), 29 clinical note reviews were conducted. While coherence was evident, with HPs recognising the value of the AMBER care bundle, cognitive participation and collective action presented challenges. Specifically: (1) HPs were unable and unwilling to operationalise the concept of 'risk of dying' intervention eligibility criteria (2) integration relied on a 'champion' to drive participation and ensure sustainability; and (3) differing skills and confidence led to variable engagement with difficult conversations with patients and families about, for example, nearness to end of life. Opportunities for reflexive monitoring were not routinely embedded within the intervention. Reflections on the use of the AMBER care bundle from HPs and patients and families, including recommended modifications became evident through this NPT-driven analysis.
CONCLUSION: To be successfully normalised, new clinical practices, such as the AMBER care bundle, must be studied within the wider context in which they operate. NPT can be used to the aid identification of practical strategies to assist in normalisation of complex interventions where the focus of care is on clinical uncertainty in acute hospital settings.

RCT Entities:   Population Interventions Outcomes

Introduction

Poor hospital care has received growing attention, particularly among the frail elderly and those approaching end of life [1]. There is increasing recognition of the challenges of caring for this growing population, many of whom face clinically uncertain outcomes in which they may improve or deteriorate further [2, 3]. These patients often have complex clinical and psychosocial needs. However, they are often inadequately addressed due to poor identification of deterioration [4] and insufficient and delayed communication from health professionals [4-6]. This has potential to negatively impact on patients and their families at a profoundly emotional level [1]. It also affects health professionals and health systems. It has been argued their intolerance of clinical uncertainty includes sub-optimal decision-making and planning, poor communication, inappropriate levels of investigation, patient safety, and use of scarce healthcare resources [6].

In response, a growing number of interventions, including the Serious Illness Conversation Guide [7, 8], the Universal Form of Treatment Options (UFTO) [9], the Recommended Summary Plan for Emergency Care and Treatment (ReSPECT) [10], and the Psychosocial Assessment and Communication Evaluation (PACE) [11], among others, have emerged. These interventions have been designed specifically to assist health professionals identify patients whose situations are clinically uncertain and more effectively navigate decision-making [10, 12, 13], communication [5, 7, 8] and care for those approaching the end of life. Broadly, these interventions aim to provide a structured approach to managing complex needs and uncertainty by developing a clearly communicated and documented care and treatment plan, that incorporates escalation or de-escalation decisions [14]. They also aim to ensure that patient preferences, in the context of clinical uncertainty, are taken into account, documented and understood across the wider clinical team [3].

However, these interventions are all inherently complex [15]. They are often situated within complicated care settings, require the successful navigation of multiple interacting components, and the involvement of health professionals, across specialties and roles, as well as patients who are acutely unwell and their families [2]. In the context of clinical uncertainty and approaching end of life, this complexity may be exacerbated, with patients, many of whom are acutely unwell, relying on family to make decisions on their behalf [16]. Further, negotiating critical decisions between health professionals, patients and their families, as well as across clinical specialities, can lead to interactional conflicts, often associated with differing vested interests and professional paradigms [16-18]. If these dynamics are not sufficiently considered and understood, such interventions may be delivered inappropriately with potentially harmful consequences for patients, or not delivered at all [19-21].

The ‘AMBER care bundle’ (where AMBER refers to Assessment; Management; Best practice; Engagement; Recovery uncertain) is a notable example of these complex interventions. It was developed in 2010 to overcome issues of inadequate and discordant decision-making and communication in the acute hospital setting [3]. It aims to improve care for patients who are deteriorating, clinically unstable with limited reversibility, and at risk of dying in the next one to two months [3]. The latter criterion was subsequently amended to be at risk of dying during a patient’s episode of hospital care, despite treatment [22]. The AMBER care bundle is designed to make clinical decision-making explicit in situations of uncertainty by encouraging health professionals to work in concert with patients, who possess sufficient mental capacity, and their families, to develop and document a clear medical plan, including consideration of anticipated outcomes, cardiopulmonary resuscitation and escalation plans, while continuing to acknowledge their situation of uncertainty [12]. The intervention encourages regular communication with the patient and family regarding treatment plans, preferences for care and any other concerns S1 Appendix [12].

A growing body of literature has attempted to shed light on the processes and outcomes associated with the intervention, however, these have demonstrated mixed findings. A comparative observational mixed-methods study of the AMBER care bundle identified increased frequency of discussions about prognosis between health professionals and patients, and higher awareness of prognosis by patients [23]. However, despite instances of communication being greater, they were often associated with a lower quality of information being communicated [10]. Interviews with health professionals in a qualitative study identified that the AMBER care bundle was often utilised as a tool to categorise patients, to change the focus of care delivery and indirectly served a symbolic purpose in influencing behaviours of individuals and teams [24]. More recently, a feasibility cluster randomised controlled trial (cRCT) of the AMBER care bundle across four UK hospital sites revealed a highly varied experience of care and communication for patients supported by the intervention and their relatives [22, 25].

Clinical and contextual equipoise therefore is still present. Concerns are amplified as this intervention has been identified by National Health Service (NHS) England as one of five key enablers to Transform End of Life Care in Acute Hospitals [26], highlighted as representing ‘good practice’ by both the Royal College of Physicians [27] and the National Institute for Health and Care Excellence [28]. Additionally, it has been widely adopted across a network of approximately 40 UK District General Hospitals [29] and a growing number of hospitals in New South Wales, Australia [30].

Choice of normalisation process theory for process evaluation

The development of this complex intervention was informed by a pragmatic clinical case note review with input from two specialties–palliative care and geriatric medicine [3]. There is now increasing evidence that the successful development and implementation of complex interventions benefit from the contribution of theoretical frameworks [31-33]. As part of the feasibility cRCT of the AMBER care bundle [22], we conducted a process evaluation to understand how the intervention was operationalised, and what modifications and refinements were needed to optimise its use in acute hospital settings. To achieve this, we made use of Normalisation Process Theory (NPT) [34], a socio-behavioural theory. NPT was chosen as it focuses on the ‘social organisation of the work (implementation), of making practices routine elements of everyday life (embedding) and of sustaining embedded practices in their social contexts (integration)’ [34] (p 538).

NPT provides a set of tools to identify and explain the social processes through which new or modified practices of thinking, enacting and organising work are operationalised in institutional settings: in this case, hospitals [21]. Moreover, it sets out a three-stage model of implementation, embedding and integration and is organised around several important questions: (i) What factors promote or inhibit the routine incorporation of the intervention in practice? (ii) What factors promote or inhibit the implementation, embedding and integration of the intervention? (iii) What factors promote or inhibit the mobilisation of structural and cognitive resources for the implementation of the intervention? [35]

The theory identifies four essential determinants of embedding or ‘normalising’ complex interventions into common practice. These are (i) ‘coherence’–the extent to which an intervention is understood as being meaningful, achievable and valuable; (ii) ‘cognitive participation’–the engagement of individuals (in this case, health professionals) necessary to deliver the intervention; (iii) ‘collective action’–the work that brings the intervention into use; and (iv) ‘reflexive monitoring’–the on-going process of adjusting the intervention to keep it in place [22]. These components are considered to be dynamic and interact within the wider context of the intervention, such as existing organisational structures and procedures [36]. Importantly, they are in keeping with the UK’s Medical Research Council (MRC) guidance on the development and evaluation of complex interventions [37] and the Methods of Researching End-of-life Care (MORECare) statement [15], which both stress the importance of theory in understanding what makes interventions effective.

Aim and objectives

In this paper, we examine the use of NPT to determine if, and in what ways, the AMBER care bundle can be successfully embedded or ‘normalised’ into acute hospital care practice to support patients whose situations are clinically uncertain, and their families. More specifically, the paper aimed to integrate our data across sources under NPT constructs to (i) understand individual and contextual facilitators and barriers surrounding the implementation of the AMBER care bundle and (ii) identify strategies to strengthen facilitators and mitigate barriers, informing the optimisation of the intervention and its future sustainability in acute hospital clinical care.

Methods

Design

We made use of a multi-method design [38] within the wider mixed-method feasibility cRCT trial [22]. Data were collected in parallel and then analysed and integrated using NPT. The data collection approaches included (i) health professional focus groups, (ii) interviews with patients and/or relatives, and (iii) multi-disciplinary team meetings’ non-participant observations. The quantitative component involved a detailed examination of patients’ clinical notes. The study was registered in a freely accessible clinical trial registry (ISRCTN36040085).

Study setting

The feasibility cRCT took place across purposefully selected general medical wards located in clusters. For the purposes of this study, the clusters were represented by four district general hospital (DGHs) in England. Implementation of the AMBER care bundle were limited to two of the four DGHs. Selection of study wards at each site was informed by heat maps that provided contextual information at a ward level on the number of deaths during and up to 100 days after admission. Wards with the highest number of deaths per year were considered to be suitable for the study [22, 29]. In this paper, we focus on the two intervention sites, where the AMBER care bundle was implemented and delivered and who provided data for the process evaluation reported in this paper Table 1. Full details of how the sites were selected are reported elsewhere [22].

Table 1

Description of study sites.

Site	Cluster	Specialty	Number of beds	End of life care plan	CQC rating
Site 1	1 general medical ward	respiratory endocrinology	30	Individualised care plan for dying patients	Good
Site 2	2 general medical wards	care of the elderly	36	End of life care plan	Requires improvement

Implementation of the AMBER care bundle

A nurse facilitator supported the implementation of the AMBER care bundle across the two intervention sites for a period of two months. This involved: familiarisation with the ward, introducing the intervention to health professionals and training them on its use, supporting them in the practice of using the intervention, and observing how they used it in practice. Full details of the implementation of the intervention across study sites is reported in detail elsewhere [22, 29].

Sampling

Recruitment of patient and family for interviews

Patient or family participants were identified by research nurses in conjunction with health professionals for the feasibility cRCT when patients met the following criteria: were over 18 years old, deteriorating, in a clinically uncertain situation with limited reversibility, at risk of dying during their current episode of care, despite treatment. Participants also needed to be able to provide written informed consent or assent through a personal consultee (consultee declaration) prior to the interview. We made the decision to pragmatically stop recruitment when we believed we had collected an adequate amount of data to address the research questions and where we could be confident from our on-going interviews and processes associated with our framework analysis approach that new data would be considered to be redundant of data already collected [39-41]. Each participant provided written informed consent prior to the interview.

Recruitment for focus groups

Ward staff from study wards were invited via research nurses and posters to participate in the focus groups. Of those who expressed interest, we aimed to recruit a range of health professionals with different levels of experience. Written informed consent was obtained from participating health professionals.

Recruitment for non-participant multi-disciplinary observations

For non-participant observations, the researcher (EY) organised with research nurses and clinical staff an appropriate schedule of attendance. Observations included multi-disciplinary team meetings including morning handovers and board rounds across study wards. Informed consent was obtained prior to these meetings.

Data collection

Questions asked during (i) focus groups with ward staff, (ii) semi-structured interviews with patients and/or relatives; (iii) issues noted during the non-participant observations of multi-disciplinary team meetings; and, (iv) the review of patient participants’ clinical notes, were informed by our patient and public involvement (PPI) representatives, and were aligned to NPT.

(i) Qualitative interviews with patient and carers

Interviews were semi-structured and topic guides explored key constructs of NPT including patients’ and their relative’s insights into the delivery of care S2 Appendix. Interviews were conducted by a research assistant (EY) and recorded on an encrypted digital voice recorder. During transcription, all potentially identifiable information was removed or anonymised.

(ii) Focus groups with health professionals

Health professionals’ views on caring for patients whose situations are clinically uncertain and views about the intervention were explored during focus groups S3 Appendix. Focus groups were led by senior researchers (JK and CE) with experience in palliative care and qualitative research, and field notes were taken by EY and HJ. Focus groups were audio-recorded and transcribed verbatim and anonymised.

(iii) Non-participant observations of multi-disciplinary team meetings

Non-participant observations of multi-disciplinary team meetings took place at multiple time points on each of the wards. During these observations, the researcher (EY) noted who was present, the frequency and length of the meetings, and the type of conversations relating to patients identified as fulfilling the criteria to be supported by the AMBER care bundle. Notes were also made as to which professionals contributed to conversations, and the decision-making discussion and actions related to patient care. Observations were recorded as hand-written field notes throughout and immediately on leaving the meeting.

(iv) Patient participant clinical case note review

Following the implementation of the AMBER care bundle, we examined the clinical notes of patients who were in receipt of the intervention. Data were extracted by EY onto a designed-for-purpose form which captured details of admission, death or discharge, calculation of the length of stay, and documentation of the intervention components. All identifiable patient information was removed or anonymised before sharing with the rest of the research team.

Analysis

Patient participant clinical case note review

The numerical data in the case notes were analysed with SPSS [42] using descriptive statistics.

Interviews with patients and relatives, focus groups with health professionals, and non-participant observations of multi-disciplinary team meetings

Qualitative data were analysed using the Framework approach and thematic analysis [41] facilitated by NVIVO 10. Members of the research team who led on conducting the interviews, focus groups and observations (JK, EY and HJ) also led on the analysis. They familiarised themselves with the raw data and discussed their impressions of the dataset. The NPT constructs–‘coherence’, ‘cognitive participation’, ‘collective action’ and ‘reflexive monitoring’–provided a thematic framework Table 2. We took a robust approach to analysis: all interviews and focus groups were double coded in NVIVO 10 by two researchers (EY and JK, EY and HJ, or JK and HJ) independently followed by comparing results and discussion within each researcher pair to ensure uniformity of coding. We also hosted ‘data workshops’ where the researchers coded a sample of transcripts together with our patient and public involvement (PPI) members to minimise bias in interpretation and the validity of findings. Once transcripts were coded, data was exported from NVIVO 10 and were charted and mapped using Excel®. Non-confirmatory data were also explored and consideration was made about their sources to avoid making unwarranted claims about patterns in the data [43].

Table 2

NPT constructs relevant to the AMBER care bundle.

Normalisation Process Theory (NPT) Constructs	NPT framework questions relevant to AMBER care bundle
NPT construct 1- ‘Coherence’The work people engage individually and collectively when they are faced with the problem of operationalising a set of practices	Is the AMBER care bundle easy to describe? Is it distinct from other ward-based interventions? (i.e., meaning and sense-making by participants) Does the AMBER care bundle have a clear purpose for all relevant participants i.e. ward staff? Do ward staff have a shared sense of its purpose? What benefits will the AMBER care bundle bring, and to whom? It is AMBER care bundle expected to improve the performance and the clinical outcomes of patients and their families. Are these benefits likely to be valued by potential participants? Does the AMBER care bundle fit with the overall goals and activity of the organisation?
NPT construct 2-‘Cognitive participation’‘Buy-in’ or relational work people do to build and sustain a community of practice around a complex intervention.	Do ward staff consider the AMBER care bundle to be a good idea? Will they see the point of the AMBER care bundle easily? Will ward staff be prepared to invest time, energy and work in it?
NPT construct 3-‘Collective action’The operational work that people do to enact a set of practices around a complex intervention.	How will the AMBER care bundle affect the work of ward staff? Will it promote or impede their work Will ward staff require extensive training before they can use it?
NPT construct 4-‘Reflexive monitoring’The monitoring work that people do to understand and appraise the ways that a new set of practices affect them and others around them.	How are ward staff likely to perceive the AMBER care bundle once it has been in use for a while? Will the AMBER care bundle to be perceived as advantageous for patients or ward staff? Will it be clear to them what the effects of the AMBER care bundle intervention have been? Can users/staff contribute feedback about the AMBER care bundle once it is in use? Can the AMBER care bundle intervention be adapted/improved based on experience?

Data across all sources were then discussed further by the researchers and triangulated using the NPT constructs to understand the operation of the AMBER care bundle on each ward. At this stage, researchers identified areas of confirmation and contradictions across sources which were used to greater researchers understanding of the operation of the AMBER care bundle across health professionals, and the contributing facilitators and barriers involved in the sustainable use of the intervention.

Research governance and ethical approval

Ethical approval was obtained from the National Research Ethics Committee—Camden and King’s Cross (20.12.2016, REC Reference: 16/LO/2010) and Health Research Authority (25.01.2017). Local research governance approvals were obtained from participating hospitals.

Results

We conducted two focus groups (26 health professionals), nine non-participant observations, 12 interviews (two patients, 10 relatives) and 29 patient participant clinical note reviews. Demographics of those involved in focus groups, interviews, non-participant observations and clinical note reviews are provided in S1–S4 Tables respectively. Using multi-methods, the implementation process across sites based on the four NPT constructs of coherence, cognitive participation, collective action and reflexive monitoring were examined. Within each construct, we present the barriers and facilitators to implementation and discuss strategies for optimising implementation of this complex intervention within the acute hospital setting.

Coherence–making sense of, and finding meaning in, the AMBER care bundle

Coherence represented the process through which ward staff shared a common and valid interpretation of the purpose and value of the AMBER care bundle. Overall, ward staff were observed as having a good practical understanding of the intervention and its constituent components. For example, during the non-participant observation of a morning handover at Site 1, a range of health professionals were noted as being confident and clearly explaining the intervention to a new consultant on the ward.

There was broad agreement from health professionals that the intervention represented a positive shift in the emphasis of care for a patient group who were previously overlooked in clinical practice. Ward staff recognised that the intervention prompted them to recognise and prioritise patients whose situations were clinically uncertain and further, engage in important discussions with them regarding preferences for care, escalation decisions and medical treatment. Many noted the value of this for ensuring that patient and family preferences were discussed, captured and communicated with those involved in the patients’ care in a timely manner.

The AMBER care bundle was also perceived to be valuable in supporting some staff to provide care for this patient group. At Site 2, the ward manager highlighted the value of the intervention for junior staff. First, in increasing their understanding of clinical uncertainty, deterioration, and end of life and secondly, their increased confidence to engage with this patient population by using the intervention as a platform to broach such topics.

… But now with the ‘AMBER’ I think they can talk, and they will feel more confident to talk with relatives.

(Site2-014, female, Ward Manager)

Whilst most health professionals suggested that the AMBER care bundle represented a fundamental change in care, a small number did not believe it differed noticeably from their existing practice. These individuals did, however, note that the intervention acted as a means to formalise their current practice, a view typified by the following junior doctor:

Speaking for myself, if I have someone who I’m worried and has the potential of deteriorating, I would always within my best capacity to try and inform the family about their situation. I don’t think it changed our practice. The only thing we’re doing is just formally documenting it. We were doing everything we could, uhm yeah, even prior to AMBER bundle.

(Site2-005, Female, Senior House Office)

Despite health professionals holding a coherent view of the value of the AMBER care bundle, a lack of clarity surrounding the intervention’s eligibility criteria resulted in a varied understanding of what patients were most appropriate for the AMBER care bundle. In practice, the clinical team were frequently observed making judgements on patients’ suitability to be supported by the intervention, based on the patients’ level of co-morbidity, frailty, disease progression, the likelihood of responding to treatment and medication, or their ‘ceiling of care escalation’, that represented proxies for ‘clinical uncertainty’. They rarely referred to a patient’s ‘risk of dying’ during the episode of care [the admission] to inform decisions. This was further exemplified by the responses received by health professionals during focus groups who when asked to describe the patients who were suitable to be supported by the intervention used descriptors such as ‘those who are aged and frail’, or ‘those with an unpredictable recovery’.

One junior doctor on Site 2, explained that they had focused on identifying patients in clinically uncertain situations rather than those at risk of dying, due to the latter requiring them to prognosticate, which they did not feel skilled or confident to do. This posed fundamental challenges and often made this aspect of the eligibility criteria impossible to operationalise coherently. Further, this reluctance to acknowledge a patient’s ‘risk of dying’ was noted by patients and their family, who were often more aware of their own, or their loved one’s likelihood of death.

I remember having a conversation with the doctor and saying, “Do you really actually think he’s going to be discharged out of here? Because he looks like he’s a dying man to me.” The doctor just said to me “You have to be optimistic”, and I just said, “Optimistic or realistic?” you know?

(Site1-017, Carer)

Cognitive participation—Commitment and engagement with the AMBER care bundle

Cognitive participation represented the extent to which key stakeholders (health professionals) were adequately motivated to incorporate this complex intervention into their practice and how well it fitted in with existing approaches. All staff were expected to engage in the active identification of patients whose situations were clinically uncertain, discuss plans for treatment and care with patients and their family, and document these in patients’ clinical notes.

Across sites, staff participating in the focus groups were unanimous that the role of the nurse facilitator was critical in successfully engaging with and operationalising the intervention. She encouraged them to appraise which patients might be suitable for the AMBER care bundle, as well as, importantly, prepare and reflect on conversations with patients and their families which were often difficult and emotionally demanding. Observations of multi-disciplinary team meetings supported these views. At numerous points, the nurse facilitator was observed encouraging health professionals to complete AMBER care bundle components, particularly in identifying instances of clinical uncertainty, ensuring important discussions took place with patients and families and reminding staff of the requirement to document these circumstances.

He’s ‘AMBER’, so while writing the discharge letter, we should remember to note the things discussed and escalation decisions.

(FIELD NOTE: Site2-011, Female, Nurse Facilitator: observed during Handover meeting)

Importantly, because of their pressured workloads, ward staff did not believe they had additional capacity to ‘champion’ the intervention in the ward within their roles. During focus groups, they questioned their ability to engage in the delivery of the intervention without the continued dedicated support from the nurse facilitator. Specifically, they highlighted the challenges and time required to train new staff about the intervention, a situation amplified by increasing levels of staff turnover.

Staff were fearful of potential negative consequences which may come from the use of the intervention without a dedicated ‘champion’ to facilitate engagement. At the very least, there was concern that they would overlook patients who might benefit from the intervention and, more seriously, a wariness of engaging in potentially emotionally laborious conversations with patients and or their relatives.

I think the nurse facilitator was excellent actually in helping us implement and ‘do the AMBER’. Since she’s gone, I think it dropped off a little bit.

(Site1-022, Female, Ward Manager)

Collective action—Work required to make the AMBER care bundle function

Collective action represented the notion that ward staff performed actions or tasks based on principles outlined by the intervention. The data from across the different components of the study shed light on the resources and procedures associated with its integration into routine practice. Since the general principle of delivering patient-centred care was shared across staff, for some, the intervention was not perceived to represent a radical departure from their current way of delivering care for some individuals. This consequently meant that the work associated with the collective delivery of the intervention was generally well accepted into daily clinical practice.

Nevertheless, formally operationalising the daily clinical activities associated with the intervention provided a welcomed opportunity for other ward staff to be involved in discussions about the decision-making process associated with the delivery of care for this patient group. Instances of teamwork and shared decision making were observed during the handover meetings. At Site 2, for instance, a range of health professionals contributed to discussions about patients’ suitability for the AMBER care bundle. In one focus group, a consultant also highlighted that whilst it was doctors who were typically perceived as being pivotal in patients’ care, the views of other professions increasingly began to contribute to patient-centred decisions and their medical plans. In this respect, the presence of the intervention enabled and empowered those who had previously not been called on for their views, to actively share them. A consultant on Site 1 emphasised the shift from clinical decisions being made independently to decisions made as an MDT, with health care assistants who often have more insight into the day-to-day care of the patient, and allied health professions now being involved in decision making.

Perspectives from some patients and relatives substantiated health professionals’ views, highlighting teamwork across professionals and specialties during their care, or the care of their family member:

100% work together well.

(Site2-009, Male, Patient)

The OTs especially have been brilliant there … and yeah they seem to, they seem to all know what’s not and what’s going on.

(Site2-016, Female, Carer)

The case note review of patient participant notes provided further insight into ward staffs’ involvement in patient-centred decisions and medical plans Fig 1. In Site 1, 90% of patient participants’ notes detailed discussion and agreement of patient medical plans with ward staff, and in Site 2, this was noted in 67%, demonstrating documented involvement of other health professionals.

Fig 1

Documentation of AMBER care bundle components in patient participants’ clinical notes.

Further analysis of the patients’ clinical notes provided additional insight into the collective action taken to complete the required documentation associated with the delivery of the intervention. Overall documentation of medical plans and escalation plans for patients across sites were high (medical plans 90% and 78% and escalation plans 80% and 78%, for Site 1 and 2 respectively).

This high level of compliance may be due to health professionals perceiving the documentation required as part of the AMBER care bundle as valuable, simple, and easy to complete.

Despite this, there were, however, collective action-related concerns with some of the intervention procedures, particularly those that relied on timely discussions between health professionals and patients and/or their relatives. The combined findings from the focus groups, interviews with patients and relatives, and review of clinical notes identified that the completion of initial conversations and follow-ups with patients or their family were often delayed. In focus groups, health professionals reflected that this was often because many of the patients suitable for the intervention were elderly, confused and lacked mental capacity. Consequently, engaging in conversations with these patients about their situation, and making treatment and care plans, necessitated the presence of family.

Relatives, however, were not always available to have important discussions in a timely manner. For instance, some were unable to visit or were only able to visit on weekends or evenings, when health professionals most familiar with the patient’s circumstances were not on the ward. Observations of multi-disciplinary team meetings further evidenced the challenges ward staff reported experiencing in accessing some family members. For instance, when talking about a patient on Site 1, a consultant at the multi-disciplinary team meeting was overheard suggesting a conversation was required with the patient’s wife to discuss the possibility of a referral for a dementia assessment. However, an occupational therapist mentioned that the patient’s wife was herself unwell and currently receiving chemotherapy, so any discussion soon would be unlikely.

A key component and goal of the AMBER care bundle is the requirement for staff to document patients’ clinical situations and medical plans to ensure all staff, including weekend and out of hours staff, are familiar with a patients’ situation. While patient participant clinical notes highlighted that this documentation was occurring in most cases, the detail in these notes often varied from patient to patient, translating to instances of inconsistency in care.

This is exemplified by a daughter of a patient on Site 1 who, when asked if she felt like she was getting consistent information from nurses and doctors, highlighted the importance of all staff being aware of a patient’s clinical situation to ensure they are receiving consistent care.

I think probably some of the nurses, I don’t want to criticise them but I think probably some of the nurses who were maybe, bank nurses and things, who didn’t know her [the patient] so well didn’t quite realise about the delirium and thought it was the dementia.

(Site 1–002, Carer)

Similar experiences were shared by other relatives. Another daughter of a patient on Site 1, highlighted instances where her father had been in receipt of unnecessary tests from some health professionals who did not seem aware of the de-escalation for his clinical situation.

During engagements with patients and relatives, health professionals were encouraged, as part of the intervention, to discuss future plans, including preferred place of care and preferred place of death. Despite this, the clinical notes of patients demonstrated that these preferences were often inadequately recorded, particularly place of death. This was documented in only 30% and 44% of patient clinical notes for Site 1 and Site 2, respectively. Compliance with these components of the intervention appeared to be lower than other care plan elements, such as completion of escalation plans and do not attempt resuscitation (DNAR) discussions. This corresponded with the views of patients and relatives who suggested that, while other areas of care were being discussed (such as treatment), preferred place of care and preferred place of death were often absent from these conversations.

No, we didn’t have any of those conversations in the hospital, and to be honest I haven’t had any of those conversations since he’s gone to the nursing home, which we need to have really to be honest.

(Site 1–017, Carer)

One consultant explained that discussions surrounding the preferred place of care or preferred place of death were difficult for health professionals, due to concerns they might not fulfil patients’ or relatives’ expectations if preferences could not be realised due to over-riding clinical reasons. One consultant explained that for some patients, while the preferred place of death is home in reality, it is unlikely that these patients would be able to be supported at home, due to the services they need and care they would require. There was therefore reluctance to engage in these conversations with patients in fear that they may be disappointed if preferred place of care or death could not be achieved.

Findings also highlighted that communication with patients and families was often variable due to workload pressures. During conversations, staff felt rushed and unable to provide adequate time to have these important discussions with patients and their relatives. This proved to be frustrating and health professionals believed that, whilst the intentions of the intervention were honourable, the fidelity of its delivery was often compromised. This was typified in the following comment from a registrar:

Sometimes our communication is brief, just because of the time pressures we are in nowadays. Even when I have the conversation, I sometimes feel that if I spent 10 minutes more or 15 minutes more with a patient, I’d probably explain a lot better.

(Site2-002, Male, Registrar)

Communication and daily review

Engaging in difficult conversations was also heavily dependent on health professionals’ skills and confidence. This varied considerably. While some consultants reported possessing well-honed skills and expertise in this area, other staff, particularly nurses, believed their skills were inadequate. One ward manager reflected that this was due to nurses having inadequate training in advanced communication, whereas doctors have more focus on this during their training.

Due to this perceived disparity in skills and confidence of ward staff, there were repercussions associated with the delivery of core collective activity in planning care and treatment in concert with patients and relatives. Since some ward staff were reticent to engage in these challenging conversations, this task was delegated by default to consultants, often leading to delays in these discussions occurring, due to consultants already pressured workloads. This finding was supported by the interviews with relatives who were frustrated by the lack of information provided by nurses, health care assistants and other front-line staff, particularly as they were told they would need to wait for the consultant or doctor to provide an update.

The nurses never really actually talked about his, erm, terminal diagnosis …, it was only the doctors, I’m sure the nurses were aware… his erm, condition would change, but the nurses didn’t really talk about that.

(Site 1–001, Carer)

Most of them were healthcare assistants but they were able to say functionally able to say how he was, but if you asked them anything medically, they wouldn’t answer your questions.

(Site 1–001, Carer)

The emotive nature of discussions relating to clinical uncertainty and the impact of these on some health professionals, particularly junior doctors’, also caused issues.

I think as a med student, we didn’t have to broach that with the family, so we just meet the patient and put them on ‘AMBER care’. So, I’m meeting you for the first time and they may cry, and you know I found it quite hard, emotionally.

(Site 2–023, M, Foundation Doctor Y 1)

While the nurse facilitator provided some support in this respect, there was a strong desire to provide greater access to clinical supervision and protected time for staff to debrief after these disconcerting discussions.

The daily review component of the intervention also experienced low compliance; this intervention component was documented in only 40% and 67% of clinical notes at Site 1 and Site 2, respectively. This component of the intervention required ward staff to take time each day to, first, review whether a patient’s situation of clinical uncertainty remained and, second, to follow-up on any patient or family concerns and preferences identified during the initial discussion. Two issues were identified by health professionals to explain variance in the daily review. Firstly, some health professionals were of the view that appraising patients’ clinically uncertain recovery status daily was unnecessary. This was exemplified by a consultant at Site 1 who preferred for his team to feel that there was always an element of uncertainty and to review patients less often, unless there is a drastic change in their situation of clinical situation. Secondly, others believed that frequent discussions with patients and families about their situation of clinical uncertainty had the potential to cause additional and undue distress.

As part of the daily review, staff were also required to place a yellow sticker on patients’ clinical notes each day that they were on the AMBER care bundle, alongside any relevant notes. In practice, the ‘A’ stickers were not perceived as adding value, were burdensome and therefore were quickly abandoned by staff.

However, in contrast to health professionals’ concerns surrounding the daily review, relatives reported valuing the frequent updates from health professionals. Accepting the extensive workload and time pressures of health professionals, some suggested that they did not expect lengthy discussions, but brief updates would have been welcomed each time when they visited the ward.

We wanted more communication. We were there every day, so there was no reason why they did not stop and spoken to us.

(Site1-003,Carer)

Reflexive monitoring—Opportunities to appraise the AMBER care bundle

Reflexive monitoring refers to the appraisal process which health professionals, formally and informally, undertook to assess and understand the ways that the AMBER care bundle affected themselves, patients and/or families and others around them. Opportunities for formal reflection and monitoring were not integrated into the intervention. However, we did identify, through observations and focus groups, attempts by health professionals to locally modify or reconstruct the AMBER care bundle and its delivery, to successfully enable implementation within their acute hospital setting. This included adapting the frequency of reviewing patients’ situation of clinical uncertainty and removal of the requirement to place a yellow ‘A’ sticker in patients’ notes Table 3.

Table 3

Modifications to the AMBER care bundle component.

Suggested Modification	Rationale for Modification
Remove prognostication from eligibility criteria	Health professionals highlighted the difficulty of predicting whether patients were going to die during their current hospital admission. Consequently, many were reluctant to make decisions on patients’ suitability for the AMBER care bundle based on their risk of death and instead focused on identifying situations of clinical uncertainty to inform their decisions. Additionally, health professionals suggested that simplification of the eligibility criteria to concentrate solely on ‘clinical uncertainty’ rather than ‘deterioration’ and ‘risk of dying’ would not only ensure that a wider group of patients would be identified and benefit from the AMBER care bundle, but it would mean that staff would not be required to use the ambiguity of prognostication as a decision-making tool.
Removal of daily review stickers	Health professionals saw little value in the requirement of placing a yellow ‘A’ sticker delineating ‘AMBER’ on patients’ clinical notes to prompt staff to think about their situation. In practice, this task associated with the intervention was rarely completed. Health professionals therefore recommended that the sticker should be disposed with.
Daily review of the patient’s situation of clinical uncertainty	Health professionals suggested that reviewing patients’ clinical uncertainty within the clinical team was not required daily since patients’ situations did not tend to change between recovery and deterioration that often. Further, some health professionals perceived that the requirement to revisit conversations on a daily basis were distressing for patients and family members. Staff therefore recommended that it would be more valuable and efficient to review patients’ clinical situations only where there was evidence of a more profound change in their situation.
Daily re-engagement with patients and/or family	Paradoxically, patients and particularly relatives suggested that staff should provide a brief practical update to the patient and family each day regarding their general overall care. Aware of workload pressure of staff, patients and relatives suggested that these did not need to be lengthy discussions. Instead, they should be brief, covering any notable events that occurred throughout the previous day and provide a general update regarding their care. It was suggested that these brief updates could be undertaken by nurses and other ward staff who were present more often on the wards.

Optimising implementation, integration and sustainability of the AMBER care bundle

Fig 2 presents a model of the facilitators and barriers of the AMBER care bundle alongside modifications that must be considered to enable the normalisation of the intervention within the acute hospital setting. In response to these barriers, improvement strategies are identified that are likely to contribute to improving aspects of cognitive participation and collective action to sustain practices associated with the delivery of this complex intervention [14, 44].

Fig 2

Facilitators, barriers and strategies for normalisation of the AMBER care bundle in acute hospital settings.

Discussion

Introducing new models of working in health care settings such as hospitals is often extremely challenging [45]. This multi-method process evaluation study, residing within a wider cluster cRCT, has demonstrated a varied response to implementing the AMBER care bundle into routine practice to serve patients whose situations are clinically uncertain and their families. Whilst it is evident from our findings that the intervention was perceived as enhancing some aspects of care for this previously under-served patient group, we also identified barriers that block its successful integration and normalisation as it was intended and consequently its sustainability in routine practice. Below, we critically reflect on the findings to understand the intervention through an NPT lens, describing which aspects of the intervention work, which did not and importantly, why. We also outline what is vital to improve the intervention so it can be normalised and sustained in ‘real-world’ acute hospital ward settings and deliver favourable patient and family outcomes.

Aspects that aided normalisation of the AMBER care bundle

The intervention was used by health professionals to prompt them to actively consider the care and management of a patient group which were not previously prioritised in their clinical practice. The intervention itself acted as a platform to broach the issue of clinical uncertainty within the acute hospital setting, helping to shift the focus of care provided. This symbolic use of the AMBER care bundle is in line with previous qualitative explorations which identified health professionals use of interventions to prompt changing the focus of care [24, 46]. For some health professionals, the AMBER care bundle was perceived to differ little from their current ways of working. Considering the growing number of end of life care-related support aids or tools, for example ReSPECT [10], a lack of differentiation may be present and problematic, with some staff resisting interventions they believe may replicate existing practices and reflect an unnecessary administrative burden [14]. Yet most health professionals perceived little administrative burden associated with the AMBER care bundle, with the required documentation considered simple, clear and easy to complete. Further, while health professionals recognised the value of the AMBER care bundle for improving their own processes (simple and clear documentation, the involvement of all staff in decision making), they were also clear on the value and benefit of the intervention for patients and their families, including empowering patients to share their preferences earlier, and involving patients and their family in critical decisions. This is a welcomed shift away from an overemphasis on processes to a greater focus on the ultimate goals of intervention delivery: improved care for patients and their families [46].

Critical to initial and sustained engagement with the intervention was the provision of a ‘champion’ (nurse facilitator) who ensured that health professionals were prompted to consider all components of the intervention and were supported throughout delivery. The importance of a ‘champion’ has been previously emphasised for the successful implementation of integrated end of life pathways within organisations [10, 47]. For the AMBER care bundle, the ‘champion’ promoted active engagement, trained new staff members and provided emotional support to health professionals delivering the intervention.

Aspects that were disadvantageous to the normalisation of the AMBER care bundle

Staff shared views on their concerns associated with identifying patients suitable for the AMBER care bundle which currently require staff to judge a patient’s ‘risk of dying during their episode of care’ alongside their clinical uncertainty and deterioration. In practice, health professionals made decisions about who was suitable for the intervention based more broadly on their capacity to benefit from it, rather than specifically their ‘risk of dying’. This highlights health professionals unwillingness to make judgements based solely on prognostication and is in keeping with current evidence which notes the difficulty some health professionals have recognising impending death [48]. Further, research suggests these judgements are highly subjective and are frequently inaccurate, which raises important issues for the operationalisation of the intervention [48]. Solutions to this issue were present during the development of ReSPECT [10]. During the development phase of the intervention designers incorporated cognitive interviewing to understand exactly how terms were interpreted across different professional groups and levels of seniority in the field, prior to its wider-scale implementation [10]. The results of this work contribute to ensuring intervention fidelity and also offer insights into where a more nuanced, flexible interpretation of the intervention eligibility criteria may be required. In its current form, we provide evidence that the eligibility criteria of the AMBER care bundle warrants serious consideration since it limits opportunities to potentially escalate care for those patients who might benefit from critical or high dependency care settings [49], as was originally intended by this intervention [3]. This has potential implications for patient experience and outcomes.

Inadequate skills and confidence in communication also led to issues in the delivery of the intervention. Variation in staff skills meant that the delivery of difficult and daily follow up conversations with patients and their family were frequently allocated to specific health professionals to complete (those with greater experience and confidence in communicating), creating workload pressures and delays in delivery. Further, the emotive context surrounding difficult conversations relating to clinical uncertainty has the potential to impact on health professionals’ well-being, particularly among some junior doctors [50]. This is pertinent in light of the ongoing global COVID-19 pandemic [51] where a growing number of relatively inexperienced junior doctors are being called upon to engage with profoundly unwell patients whose situations may be clinically uncertain, in addition to their very distressed families. Caring for patients in this situation is especially challenging, requiring additional skills, excellent communication, and importantly, systems of support for those delivering care [16, 52]. The absence of such systems may contribute to health professionals experiencing what has been referred to as ‘moral injury’ [53], representing deviations or transgressions from health professionals’ moral beliefs and expectations that are witnessed, perpetrated, or allowed by that individual. When not acknowledged and supported, this can lead to stress, depression, anxiety and post-traumatic distress.

Contextual issues, particularly workload pressures and time constraints were also cited as a significant barrier to the delivery of intervention components. Similar to previous studies exploring the implementation of complex interventions into clinical practice [47, 54, 55] frequent staff-turnover and competing priorities led to challenges and meant that health professionals felt unable to ‘champion’ the intervention themselves, raising concerns about the sustainability of the intervention.

Notably, many of the patients who the health professionals believed suitable to be supported by the AMBER care bundle often lacked mental capacity. This was associated with additional time needed to contact the family to be involved in decision-making, creating further workload pressures, and often resulting in delays in important discussions. Having a clear understanding of the population the intervention intends to serve is critical to minimise harmful outcomes and is emphasised in the Medical Research Council and MORECare guidance [15, 37, 56].

Supporting normalisation of the AMBER care bundle in acute hospital settings

To support the normalisation of the AMBER care bundle in acute hospital wards, important considerations regarding the skills, knowledge and attitudes of those delivering the intervention, and an understanding of the context in which it exists, is required. First, staff need to have a coherent view of who is suitable for the AMBER care bundle. It is therefore imperative that the AMBER care bundle eligibility criteria focus solely on a patient’s ‘clinical uncertainty’ and their clinical needs. Criteria that avoid the need for prognostication may ensure patients who would benefit from being supported by the intervention can be quickly, and more efficiently, identified by health professionals [16].

Second, staff need to be equipped with adequate communication skills to engage clearly and compassionately with patients and their family and are allocated protected time to de-brief following difficult encounters. Advanced communication training and integrated opportunities for emotional support therefore need to accompany the AMBER care bundle to ensure staff are skilled to deliver the intervention effectively. This is imperative for nursing and ward staff who are often more present on wards and therefore have potentially more opportunities to communicate with patients and their families. Additionally, previous research highlights the need to ensure this training is continually re-visited to further develop and strengthen staff knowledge and skills [50, 57, 58].

Last, while patients’ clinical uncertainty may not change on a day-to-day basis, and the clinical team should adapt the frequency of reviewing this to their local needs, it is imperative that brief daily updates for patients and their family are still prioritised. These brief, but frequent updates should aim to provide patients and their family with an update on any medical and/or care changes and provide an opportunity for patients and family to discuss any changes in preferences.

Implications for clinical practice

The AMBER care bundle carries the potential to address a crucial gap in clinical practice and further refinements, as suggested above, would help ensure it is adequately normalised into practice. Patients and their families, who experience uncertainty in light of their conditions, appreciated communication from health professionals and found the opportunity to discuss issues that matter to them valuable. One advantage of the AMBER care bundle is the fact that it relies on the shared human experience of uncertainty. For health professionals, this uncertainty is about the prognosis and likelihood of imminent death or the possible outcomes of the active treatment. Clinical uncertainty can exist for many conditions and has been argued as being an inevitability [59]. While emotive discussions, linked to expressing clinical uncertainty, have generally been seen as the job of designated health professionals, the recent COVID-19 global pandemic [60], has brought with it instances where more health professionals are required to do this work. Specifically, the COVID-19 pandemic has witnessed health professionals dealing with an extraordinary number of cases in acute hospital settings where clinical uncertainty is omnipresent across multiple levels including within the health-system, among professionals, and the patients and their families, for which the course of disease deterioration and potential for recovery is still relatively unknown [51, 61, 62]. As such, with considered modification, the AMBER care bundle has potential to offer health professionals with an approach to better serve patients affected by the COVID-19 and their families.

Strengths and limitations

Our approach used multiple data sources to inform our findings. This represented a practical and feasible way to explore NPT constructs associated with the implementation and operationalisation of the AMBER care bundle in acute hospital care settings, with qualitative components permitting us to explore salient contexts and mechanisms in more detail. Furthermore, the use of four data sources, and data being analysed iteratively by an inter-professional team, including input from our PPI members, should increase the dependability of the study findings and interpretations.

The use of the four NPT constructs as an analytic framework enabled us to provide an understanding of how the AMBER care bundle did, and in many instances could not become normalised within an acute hospital setting. Associated with this, we also identified barriers and facilitators to the future successful integration and implementation of the AMBER care bundle.

This study is also subject to limitations. It represents one component of a wider feasibility cRCT of the AMBER care bundle and was restricted to just two study sites.

Second, we are also mindful of the absence of nursing representation in one of the focus groups, who may have provided additional insights into the use of the intervention in practice from this professional group. Although, several nurses expressed interest, and confirmed their availability beforehand, on the day of the focus group no nurses were able to attend, due to urgent clinical commitments. This highlights the issues faced while researching in a ‘real-world’ context and could be overcome in future studies by considering additional flexibility and resources in the study design to accommodate the unpredictable nature of clinical work. For example, holding two smaller health professional focus groups at each site, or following up individually with health professionals who were unable to join focus groups at the allocated time.

Third, we were not able to conduct direct observations of care due to ‘real-world’ resource and logistical constraints associated with the feasibility study. Direct observations of clinical practice would be possible if research ethical approval for the study included a provision for consent at the level of the unit/cluster level [63].

Last, whilst the model has been informed by NPT constructs, developed as a result of detailed discussions within the research team, the modifications suggested as a result of this study to the AMBER care bundle have yet to be tested in the field.

Future exploration of the AMBER care bundle across other care settings and professional groups will be valuable in providing further understanding of the normalisation of this intervention in practice.

Conclusions

Our findings support growing evidence that to be successfully implemented, scalable and to be of value, new clinical practices such as the AMBER care bundle must consider the social, organisational and environmental context in which they are required to operate [45]. Whilst individual health professional change is necessary, the local context in which the intervention is intended to operate must also be supportive when implementing new, often highly complex, clinical interventions. Omitting this has potentially direct implications for patient and family experience and outcomes, patient safety and staff well-being, including issues raised by the Independent Review of the Liverpool Care Pathway [64] that specifically stressed the importance of understanding interventions focused on clinical uncertainty and communication when caring for the dying.

The importance of in-depth examination of implementation processes should proceed with feasibility studies of complex interventions to identify and incorporate modifications required so that the intervention operates as intended. Our findings highlight both facilitators and barriers and offer practical strategies for normalising multiple inter-related components of a complex intervention where the focus of care is on clinical uncertainty and end of life care in acute hospital settings. This has particular resonance during a time when the global COVID-19 pandemic is challenging patient care, shared decision-making and planning, and exercising health professionals in an unprecedented manner [62].

It is central for the normalisation and successful sustainability of such interventions that the health professionals who deliver the intervention feel empowered, and supported, in contributing reflexively to making recommendations about the workability of intervention. As the AMBER care bundle is already being operated in some DGHs, key stakeholders involved in implementation development must be receptive to these findings and scrutiny. The costs of not doing so are now regrettably well known in palliative and end of life care.

The AMBER care bundle tool.

(DOCX)

Click here for additional data file.

Topic guide for semi-structured interviews with patient and carers.

(DOCX)

Click here for additional data file.

Topic guide for focus groups with health professionals.

(DOC)

Click here for additional data file.

Demographics of health professionals involved in focus groups at each site.

(DOCX)

Click here for additional data file.

Demographics of patients and carers involved in qualitative interviews at each site.

(DOCX)

Click here for additional data file.

Demographics of health professionals involved in non-participant observations at each site.

(DOCX)

Click here for additional data file.

Demographics of patient participants involved in clinical case note review.

(DOCX)

Click here for additional data file.

31 Jul 2020

PONE-D-20-11538

Implementation of a complex intervention to improve care for patients whose situations are clinically uncertain in hospital settings: a mixed method study using normalisation process theory

PLOS ONE

Dear Dr. Johnson,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by August 20th. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Alessandra Solari, M.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please ensure that you refer to Figure 2 in your text as, if accepted, production will need this reference to link the reader to the figure.

3. Please include a copy of Table 4 which you refer to in your text on page 20.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: N/A

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: There are no statistical concerns as there is very limited to none quantitative analyses. However, the authors may want to clarify the quantitative phase of the mixed methods approach. The quantitative data collection and analyses is unclear and study seems more of a qualitative study.

Reviewer #2: Thank you for asking me to review this very interesting and well-written paper. I really enjoyed reading it.

For the editor’s benefit I should make clear my conflicts of interest at the outset. My first conflict of interest is that I have led the programme of empirical research and theoretical development that led to Normalisation Process Theory, which is employed as the conceptual framework in this manuscript. My second conflict of interest is that I have also jointly led a programme of research in the closely related area of treatment escalation plans that has used a similar method (notes review doi: 10.1136/bmjopen-2018-022021; and qualitative process evaluation https://www.sciencedirect.com/science/article/pii/S0277953620303622) formed around a similar theoretical framework.

My comments on this paper are as follows,

1. Introduction

1.1. This is a very interesting and well written paper which deals with a highly complex problem: forming and shaping decision-making processes as people with complex diseases that sometimes include problems of decisional capacity progress towards end of life.

1.2. The AMBER Care bundle is a widely used to support professionals, patients and families in this context. I think it would be helpful to the reader to do more work at the front end of the paper that describes both the intervention, and the ensemble of practices that using it puts in train. The introduction assumes a fairly well-informed reader. I think you can also frame this as a representative of a class of interventions (decision-making tools) that have a complex and contested history in end of life care – AMBER (along with UFTO and ReSPECT) make sense only if we know a little of the of the interactional context missing here.

1.3. I wonder if it might also be helpful to consider very briefly in the introduction some of the interactional problems that follow from negotiating decisions about end of life care. There is a large literature on this, and I think it would be helpful to think about the context in which AMBER and is operationalised. These a both organisationally and interactionally fragmented, and they are definitely not cognitively or affectively neutral. Again, this is context is important if the reader is to make sense of the assertion that clinical and contextual equipoise is present.

2. Methods

2.1. This is a really interesting study design, that could do with just a little elaboration. My concern here is that it is stated that this is a mixed methods study in which qualitative and quantitative data were collected in parallel and then a form of integrative analysis was undertaken. It is not clear to the read what the quantitative data is. Was it extracted from the notes review. Or, is the mixed methods design the feasibility trial plus the qualitative data. I’m not sure which it is. This is a simple matter to solve.

2.2. Qualitative data was collected from two sites. Although it’s stated that details of site selection are available elsewhere, it would be useful to have at least a precis of this here. This is important because the qualitative work reported here is drawn from two sites only. Are these the only sites at which qualitative data were collected? Or are some data (e.g. notes review) drawn from all sites? Was there a comparative element to site selection (e.g. good and bad CQC ratings)?

2.3. My understanding is that ‘saturation’ is a criterion applied to interpretive data collection in grounded theory studies, especially those using the constant comparative method. But you also say that you used framework analysis. I think we just need a couple of sentences to say how using these to techniques together worked.

3. Results

3.1. In the introduction to results you again say that a mixed methods study was performed. Again, I can see that several qualitative techniques were used, but conventionally ‘mixed methods’ is used to signify qualitative plus quantitative data and analyses so this is something that needs to be made clear.

3.2. I like the results section very much. I think the really important result that you draw attention to is that a new category of patient—the ‘clinically uncertain’ patient—is defined and brought into service, as a proxy for the patient who is likely to die. We found something similar in our study of the ReSPECT Treatment Escalation Plan, which is that when our key informants talked about clinically uncertain patients their examples were of people they believed were likely to die. Although I think that prognostic uncertainty is important here, and alternative interpretation is that these tools are intended to be used by different professional groups and hierarchies and that the category ‘clinically uncertain’ permits a high level of flexibility about their use. So, implementing the decision-tool depends to some extent on implementing a new category of patients. (This speaks to the NPT sub-construct of differentiation.)

4. Discussion

4.1. I wonder if Table 5 should be in the discussion rather than the results section?

4.2. I really like figure 2 and I really like table 3: I did wonder if they overlapped a bit , and whether it would work better to just have figure 2?

4.3. In the strengths and limitations section it is stated (i) that restricting the study to two sites, and (ii) the absence of nursing representation in a focus group limit the transferability of the findings. I’m not sure that they do. This study explores a common problem and its management. Janice Morse introduced the concept of ‘theoretical generalisability’ in qualitative research, and well theorised studies of common problems often have a very high level of transferability. After all, we’re still using fundamental concepts from Glaser and Strauss 60 years after their empirical research on awareness contexts and the dying patients was conducted in single wards at three general and one VA hospital in San Francisco.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

13 Aug 2020

Response to Review Questions:

1. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

In response to reviewer 2, we note that we have included sufficient extracts both in tables and within the manuscript to replicate and support our study findings, as such we consider our manuscript to be compliant with the PLOS data policy.

Response to Review Comments:

Reviewer 1:

1. There are no statistical concerns as there is very limited to none quantitative analyses. However, the authors may want to clarify the quantitative phase of the mixed methods approach. The quantitative data collection and analyses is unclear and study seems more of a qualitative study.

Thank you for offering your time to provide a detailed review of our paper and for raising this important issue. Our process evaluation took place within a mixed-methods feasibility cluster randomised controlled trial, and used multi-methods including the following qualitative components (i) focus groups; (ii) semi-structured interviews; and (iii) non-participant observations of multi-disciplinary team meetings. The quantitative component involved a detailed examination of patient’s clinical case notes, to provide an understanding of compliance with each component of the AMBER care bundle. On reflection, we believe using the term ‘multi-methods’ is more appropriate for our study and made this amendment to the title, abstract, and methods section (pg. 7) of the revised manuscript and have subsequently referred to this throughout the manuscript.

Reviewer 2:

1. Thank you for asking me to review this very interesting and well-written paper. I really enjoyed reading it.

For the editor’s benefit I should make clear my conflicts of interest at the outset. My first conflict of interest is that I have led the programme of empirical research and theoretical development that led to Normalisation Process Theory, which is employed as the conceptual framework in this manuscript. My second conflict of interest is that I have also jointly led a programme of research in the closely related area of treatment escalation plans that has used a similar method (notes review doi: 10.1136/bmjopen-2018-022021; and qualitative process evaluation https://www.sciencedirect.com/science/article/pii/S0277953620303622) formed around a similar theoretical framework.

Thank you for your comprehensive and helpful review of our paper, and for sharing some of your work which we have found most helpful to reflect on and have now referred to this where appropriate when revising our manuscript (please refer to pages 3- 4 in the introduction, and pages 27-28 in the discussion of the manuscript).

2. This is a very interesting and well written paper which deals with a highly complex problem: forming and shaping decision-making processes as people with complex diseases that sometimes include problems of decisional capacity progress towards end of life.

Thank you very much for your favourable view of our paper.

3. The AMBER Care bundle is a widely used to support professionals, patients and families in this context. I think it would be helpful to the reader to do more work at the front end of the paper that describes both the intervention, and the ensemble of practices that using it puts in train. The introduction assumes a fairly well-informed reader. I think you can also frame this as a representative of a class of interventions (decision-making tools) that have a complex and contested history in end of life care – AMBER (along with UFTO and ReSPECT) make sense only if we know a little of the of the interactional context missing here.

Thank you for your comment, on pages 3-4 of the revised manuscript we have now provided an additional paragraph in our introduction highlighting the challenges of decision making and communication for those in uncertain situations, and towards the end of life. We have referenced a number of examples of the increasing number of tools responding to these challenges and have highlighted how the AMBER care bundle fits among them.

4. I wonder if it might also be helpful to consider very briefly in the introduction some of the interactional problems that follow from negotiating decisions about end of life care. There is a large literature on this, and I think it would be helpful to think about the context in which AMBER and is operationalised. These a both organisationally and interactionally fragmented, and they are definitely not cognitively or affectively neutral. Again, this is context is important if the reader is to make sense of the assertion that clinical and contextual equipoise is present.

Thank you for this suggestion, we agree this would be a helpful addition to our introduction. We have now included additional contextual information at the start of our introduction located on page 3-4 of the revised manuscript, referencing both the Campling et al., 2018 and Higginson et al, 2016 papers among others, who highlight some of the difficulties of negotiating decisions at end of life between and across health professionals, patients and their families.

5. This is a really interesting study design, that could do with just a little elaboration. My concern here is that it is stated that this is a mixed methods study in which qualitative and quantitative data were collected in parallel and then a form of integrative analysis was undertaken. It is not clear to the read what the quantitative data is. Was it extracted from the notes review. Or, is the mixed methods design the feasibility trial plus the qualitative data. I’m not sure which it is. This is a simple matter to solve.

Thank you for raising this issue. On page 7 of the revised manuscript, we have now clarified in our methods that this process evaluation sat within a mixed methods feasibility cluster randomised controlled trial. Specifically, data collection for the process evaluation involved multi-methods, including qualitative components (the qualitative interviews, focus groups, non-participant observations of multi-disciplinary team meetings and a quantitative component (the examination of the patients’ clinical notes).

6. Qualitative data was collected from two sites. Although it’s stated that details of site selection are available elsewhere, it would be useful to have at least a precis of this here. This is important because the qualitative work reported here is drawn from two sites only. Are these the only sites at which qualitative data were collected? Or are some data (e.g. notes review) drawn from all sites? Was there a comparative element to site selection (e.g. good and bad CQC ratings)?

Thank you for your suggestion, we have now added further detail on the ‘heat maps’ used to inform study ward selection on page 7 in the methods section of revised manuscript.

7. My understanding is that ‘saturation’ is a criterion applied to interpretive data collection in grounded theory studies, especially those using the constant comparative method. But you also say that you used framework analysis. I think we just need a couple of sentences to say how using these to techniques together worked.

Thank you for raising this issue. On pages 8-9 of the revised manuscript we now state that we made the decision to pragmatically stop recruitment when we believed we had collected an adequate amount of data to address the research questions and when we could be confident from our on-going interviews and processes associated with our framework analysis approach that new data would be considered to be redundant of data already collected.

8. In the introduction to results you again say that a mixed methods study was performed. Again, I can see that several qualitative techniques were used, but conventionally ‘mixed methods’ is used to signify qualitative plus quantitative data and analyses so this is something that needs to be made clear.

Thank you for raising this, on reflection we agree that this should be referred to as multi-methods and we have adapted this throughout the manuscript.

9. I like the results section very much. I think the really important result that you draw attention to is that a new category of patient—the ‘clinically uncertain’ patient—is defined and brought into service, as a proxy for the patient who is likely to die. We found something similar in our study of the ReSPECT Treatment Escalation Plan, which is that when our key informants talked about clinically uncertain patients their examples were of people they believed were likely to die. Although I think that prognostic uncertainty is important here, an alternative interpretation is that these tools are intended to be used by different professional groups and hierarchies and that the category ‘clinically uncertain’ permits a high level of flexibility about their use. So, implementing the decision-tool depends to some extent on implementing a new category of patients. (This speaks to the NPT sub-construct of differentiation.)

Thank you for raising this important point. We acknowledge that the concept of clinical uncertainty is nuanced and potentially open to wide interpretation. On page 28 of the revised manuscript we refer to the Hawkes et al. 2020 study – ‘Development of the Recommended Summary Plan for Emergency Care and Treatment (ReSPECT)’ highlighting the importance of intervention designers to conduct cognitive interviews to understand how key terms and concepts are understood so that modifications or refinements can be made to ensure intervention components can be operationalized with fidelity, or to understand where greater flexibility is needed.

10. I wonder if Table 5 should be in the discussion rather than the results section?

Thank you for making this suggestion. Whilst we understand why it may be reasonable to present this table (previously table 5, now table 3) in the discussion section of the paper, we are minded that the contents of the table reflect suggested modifications resulting from the reflexive monitoring NPT construct. Moreover, in this table we provide contextual information from the perspectives of the health professionals, patients and their relatives as to why these changes were suggested which present key results. We would therefore be very grateful if the Editor and reviewer concede to retaining this table in the results section.

11. I really like figure 2 and I really like table 3: I did wonder if they overlapped a bit, and whether it would work better to just have figure 2?

Thank you for your comment. We agree that some of the content of the figure and the table are overlapping. The table presents the facilitators, barriers and implications in relation to the constructs of NPT, whereas the figure presents these against the components of the AMBER care bundle. We agree that the figure presents the more valuable contribution to the paper, and by and large contents from table 3 can be ascertained within the main text of the results section of the manuscript. Consequently, we have now removed Table 3.

12. In the strengths and limitations section it is stated (i) that restricting the study to two sites, and (ii) the absence of nursing representation in a focus group limit the transferability of the findings. I’m not sure that they do. This study explores a common problem and its management. Janice Morse introduced the concept of ‘theoretical generalisability’ in qualitative research, and well theorised studies of common problems often have a very high level of transferability. After all, we’re still using fundamental concepts from Glaser and Strauss 60 years after their empirical research on awareness contexts and the dying patients was conducted in single wards at three general and one VA hospital in San Francisco.

Thank you for your comment and reminding us of this formative study. We agree that the number of sites involved in the study, and absence of nursing representation from one of the focus groups is not a critical limitation of our study. Rather we believe it would have been useful to have more sites and nursing representation to inform potential modifications and improvements to the intervention. On page 32 of the revised manuscript we have changed the text in this section to highlight this. Moreover, we note that future research exploring additional contexts and professional groups may provide additional and potentially useful insights.

References highlighted in response:

Campling N, Cummings A, Myall M, Lund S, May CR, Pearce NW, et al. Escalation-related decision making in acute deterioration: a retrospective case note review. BMJ Open. 2018;8(8):e022021.

Higginson IJ, Rumble C, Shipman C, Koffman J, Sleeman KE, Morgan M, et al. The value of uncertainty in critical illness? An ethnographic study of patterns and conflicts in care and decision-making trajectories. BMC Anesthesiol. 2016;16(1):11.

Hawkes CA, Fritz Z, Deas G, Ahmedzai SH, Richardson A, Pitcher D, et al. Development of the Recommended Summary Plan for eEmergency Care and Treatment (ReSPECT). Resuscitation. 2020;148:98-107.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

2 Sep 2020

Implementation of a complex intervention to improve care for patients whose situations are clinically uncertain in hospital settings: a multi-method study using normalisation process theory

PONE-D-20-11538R1

Dear Dr. Johnson,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Alessandra Solari, M.D.

Academic Editor

PLOS ONE

7 Sep 2020

PONE-D-20-11538R1

Implementation of a complex intervention to improve care for patients whose situations are clinically uncertain in hospital settings: a multi-method study using normalisation process theory

Dear Dr. Johnson:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Alessandra Solari

Academic Editor

PLOS ONE