Increasing advance care planning in the secondary care setting: A quality improvement project
ABSTRACT
Background Advance care plans (ACP) provide patients the opportunity to communicate their goals and wishes for future care.
Local problem A retrospective case note review of 50 inpatient deaths in 2017 confirmed a doctor had discussed expected death in 90%, however only 2% had an ACP.
Methods Patients appropriate for ACP were identified on a single geriatrics ward. Interventions were implemented with monthly data collection. Patients with an ACP were followed prospectively. The initiatives were subsequently applied across six geriatrics wards.
Interventions Interventions included improved identification of patients appropriate for ACP, doctor education and improved communication to general practitioners and healthcare providers.
Results Before initiation of interventions on the pilot ward, ACP was completed for 38% of appropriate patients; this increased to a mean of 78.6% over 4 months post-interventions. During the pilot, 44 patients had an ACP. Of those discharged, 75% avoided readmission over the following 6 months. After applying the interventions across all geriatric wards, ACPs increased to a mean of 81.2% and was maintained 12 months later at 72%.
Conclusions The initiatives formed a structure to promote the use of ACP on the wards. Care plans focused on individualising care and effective communication resulted in reduction of readmissions.
Introduction
The nature of the problem
The Gold Standards Framework defines advance care planning (ACP) as the ‘conversation between people, their families and carers and those looking after them about their future wishes and priorities for care’.1 In 2018, The Royal College of Physicians (RCP) published key recommendations that included to ‘initiate shared decision-making including ACP in line with patient preferences’.2
ACP is an ongoing process through which a patient can express their preferences for future health and end-of-life care. In an era when medical interventions can prolong life even beyond what a person may wish, an ACP can help navigate through these difficult medical decisions so that the patient, rather than their condition, is treated.3,4 If a patient lacks capacity, a care plan can be generated with a designated proxy decision-maker acting in their best interests.
A retrospective review of deaths in the hospital was undertaken to determine the frequency with which an ACP was discussed. These data confirmed that significant improvement was required to ensure more patients were provided the opportunity to receive an ACP and consequently a better end-of-life experience.
The evidence for advance care planning
The disease trajectory in chronic conditions is unpredictable and even more challenging in frail older people.5 Most patients do not want to die in hospital and do not enjoy the hospital experience.6 However, when patients are approaching the end of life, they have most contact with hospital and social services.7,8 Multiple transfers between care settings do little to improve quality of life or symptom control and may threaten continuity and safety.9 ACP, in conjunction with specialist palliative care, reduces hospitalisation, multiple readmissions and helps patients to die in their preferred location.10
The RCP's Talking about dying report discusses when to initiate ACP discussions, suggesting that ‘conversations about the future should be initiated at any point’.2 Patients state they want their doctor to have these discussions sooner rather than later.11,12 Others, however, may not feel ready or able to have this conversation and the timing of the discussion is likely to influence its acceptability and effect.13 When a patient is acutely unwell, making treatment decisions can be stressful and rushed. By planning for these eventualities in advance, it allows patients and their caregivers to feel prepared and more in control, improving their experience.14 Patients and families do not mind having these ACP discussions with someone other than their ‘regular doctor’ as long as the doctor is willing to open the conversation with them.15 Patients describe an enhanced doctor–patient relationship after discussing end-of-life plans.16 It is suggested that caution is required because ACPs can oversimplify the decision-making process if not followed up appropriately. Decisions will likely change over time due to social, cultural, financial and other circumstantial factors.17
At the end of life, there is often a disparity between the patient's expressed preference and what a physician or family member thinks they would choose.18 Contention can arise within families and with healthcare providers (HCPs) when the patient's preferences are unknown.19 Conducting an ACP discussion removes the uncertainty and fear which can result in conflict. In addition, the bereaved relatives of patients with ACPs feel less anxiety and guilt after the death.20
Barriers to advance care planning
Relatives and carers may adopt the ‘not yet’ approach and the need to ‘stay positive’.4 Culturally, discussion and acceptance of death can be difficult.21 Patients and relatives may lack the confidence to initiate conversations about death and prognosis.
HCPs may struggle with discussing the end of life. Addressing their own acceptance of disease progression can be difficult as they may view it as a therapeutic failure.22 It can be difficult to know when to initiate conversations and HCPs may fear doing it too early.15 Many chronically ill patients are already thinking about their end-of-life decisions and would be willing to discuss them.23 Few doctors have had any formal training in having these discussions and many believe that they lack the appropriate skills, experience or knowledge to comfortably perform ACP.24,25 Time constraints for hospital physicians have also been identified as an important barrier.26
The project aims and rationale behind the interventions
Having identified a need to improve ACP for older people towards the end of life, key priorities for interventions were established. Considerable emphasis has been placed on leadership for the sustainability of successful interventional projects with ACP leads providing face-to-face training, mentoring, updates, networking and development of processes.26 The role of the ACP lead in this study was to coordinate the initiative to ensure a sustained improvement.
Aims
Improve recognition of patients who are appropriate for ACP
The time constraints that HCPs face in the hospital often result in ACP discussions being considered less important than daily routines with more measurable outcomes.27 An organisational commitment to ACP promotes implementation.26 Routinely highlighting patients requiring an ACP at the daily multidisciplinary meeting could help to prioritise and legitimise the task. Physicians acting as ACP champions have successfully promoted ACP discussions among colleagues.26 The ‘ACP champions’ role was designated to the ward registrars.
Improve doctor education in ACP and end of life
Education and communication programmes instil confidence in doctors to engage with early ACP conversations.28,29 To address this, a series of educational events were designed to increase doctors’ confidence and ability to have effective ACP discussions with patients.
Improve communication of ACP discussions between primary and secondary care
Effective communication of information is essential to implement patients’ choices in end-of-life decisions.30 The NHS Five year forward view recognises the collaboration and communication between HCPs.31 To ensure this, emphasis was placed on detailed written documentation and, on occasion, telephone calls to primary care providers.
Methods
For the quality improvement project (QIP), a plan, do, study, act (PDSA) cycle was used. To identify the nature of the problem and ‘plan’ the QIP, a pre-project audit was performed.
Pre-project audit
A retrospective case-note review of all hospital deaths over a 4-week period was undertaken. To improve ACP, three key strategies were highlighted for implementation. Firstly, identification of patients appropriate for ACP discussion; secondly, doctor education; and finally, improved communication with primary care.
Identifying appropriate patients for initiation of ACP discussion
Triggers to consider which patients were appropriate for initiation of an ACP discussion included trust guidelines, a high admission clinical frailty score, consideration of the ‘surprise question’, a pre-existing do not attempt cardiopulmonary resuscitation (DNA-CPR) form and patient choice.1,32 For patients who lacked capacity, care planning was discussed with the next of kin. A new ACP column on the patient whiteboard was created to prompt initiation of discussion of appropriate patients with further prompting from the designated ‘ACP champion’.
Doctor education
Three lectures and multiple informal tutorials on ACP discussions were given to doctors within the geriatrics department. Posters detailing when and how to do ACP were created. To increase doctor confidence, junior team members shadowed seniors performing ACPs.
Improved communication with primary care
ACP stickers were printed for the medical notes highlighting documentation. Junior doctor education focused on improving communication with primary care on discharge summaries and verbally. Doctors reviewed the general practitioner (GP) summary care records and asked patients or their proxy decision-maker whether or not a prior ACP discussion had taken place. Prior to a patient's discharge, doctors were encouraged in their ACP training to telephone GPs communicating information from the ACP discussion.
Data collection
Following the retrospective review, baseline data were collected on the pilot ward. The current patient notes and any corresponding online information were reviewed. Patients considered appropriate for an ACP and those who had an ACP already in place were identified.
Patients who had an ACP during the intervention period were prospectively followed up over 6 months. Hospital records were reviewed to determine if the patient had been readmitted or died and the location of death.
Interventions introduced across remaining five general geriatric wards
Once the pilot was established and the strategies had improved the number of patients with an ACP, the project expanded to include five other geriatric wards. Through studying the outcomes of the pilot as part of the PDSA cycle, it was noted that the team needed regular prompting at the board round to consider each patient for ACP. ACP champions were therefore recruited to the remaining five wards to ensure the interventions outlined occurred on each ward. Teaching sessions were conducted for all doctors in the geriatrics department.
Baseline data were collected for each of the five geriatric wards. Once the interventions were implemented, the percentage of patients who were appropriate for ACP and who had one initiated was recorded weekly over a 5-week period. ACP champions received weekly reminders to ensure the boards were updated.
12-month follow-up
The project continued on all six wards over the subsequent year with the same interventions. When the junior doctors rotated, training was implemented and new ACP champions were recruited. At 12 months, data were recorded weekly for a 6-week period.
There was no need for formal ethics review and no conflict of interest.
Results
Pre-project audit
There were 80 deaths in the hospital in September 2017. It was only possible to retrieve and review 50 of these case-notes.
The mean age of the patients was 86 years with a mean Rockwood clinical frailty score of 6.2. The mean duration from admission to death was 16.9 days.
Ninety-eight per cent of the patients had a DNA-CPR form and 90% of families had discussed the anticipated death with the doctors. One patient had a documented ACP. Applying the triggers discussed, ACP discussion would have been appropriate in 98% of the patients.
Thirty-four per cent of this cohort had a recent hospital admission (within a mean of 3 months from discharge to readmission).
Identifying appropriate patients
At baseline, 13 (59%) patients on the pilot ward were considered appropriate to have an ACP discussion. Of these, ACP was initiated in five (38%).
Doctor education
Following introduction of the ACP column and prompting from the ACP champions, there was an initial increase in ACPs completed to 60%. Following targeted junior doctor education on the ward and with lectures, this increased to 83.3% (Fig 1).
Improved communication with primary care
The pilot study data showed that 93% of ACP discussions were communicated to primary care in the discharge summary.
Data collection
During the pilot study, a mean of 78.6% of patients had an ACP discussion initiated appropriately. The percentage of appropriate ACPs completed dropped to 20% when the ACP champion was absent but increased to 91% within 2 weeks of their return (Fig 1).
Over the 4-month period of the pilot, a total of 44 patients had an ACP; 20 (45%) of these patients died during admission. Of the 24 (55%) who were discharged, hospital records were monitored to see if they were readmitted or died. Eighteen (75%) patients were not readmitted in accordance with their wishes. Of the patients discharged and not readmitted, 15 (83%) died in their chosen location in the community. None of the patients chose hospital as their preferred place of death. Six patients were readmitted and their reasons are charted in Fig 2.
Interventions introduced across the remaining five geriatrics wards
At baseline before introduction of the interventions, 19.2% of patients considered appropriate had an ACP initiated. After 5 weeks of the project, the mean rate of ACPs initiated when appropriate was 81.2% (Fig 3).
Follow-up at 12 months
At 12 months, data were collected for each ward over 6 weeks. The mean rate of ACPs initiated across all six wards during this time period was 72.3% (range 55.3% to 90.6%; Fig 4).
Discussion
The project demonstrates that through simple interventions, education, an ACP section on the board round and use of ACP champions, there was a marked increase in the number of ACPs initiated in appropriate patients. The ACP champion was key to the project's success and sustainability. When the ACP champion was absent, the number of ACPs reduced significantly but was restored when they returned. In the second month of the project, there was a reduction of ACPs completed when the junior doctors rotated (and before the new juniors received training).
There was improved communication with primary care, with 93% of ACP discussions included on the discharge summaries; 75% of patients discharged were not readmitted (in the time of the study) and, of these, 83% died in their preferred location.
The project has increased initiation of ACPs across the whole department in appropriate patients with a sustained change in practice.
Interpretation
In the pre-project audit, 98% of patients who died would have been appropriate for an ACP discussion; 34% of these patients had a recent hospital admission with a mean time of 3 months suggesting that there may have been an opportunity on the previous admission for an ACP discussion.
There was a higher proportion of ACPs completed at baseline on the pilot ward (38%) than in the retrospective review (2%). The retrospective data reviewed all hospital deaths, and the pilot data was collected from a designated geriatrics ward where staff are more likely to be familiar with ACP.
The three main interventions were:
identifying appropriate patients using the ACP column on the multidisciplinary board round
doctor education and ACP champions
improved communication with primary care.
The ACP column served as a daily reminder to prioritise these discussions.
Education was key to giving confidence to doctors to initiate conversations. The skills learned are transferable to other specialties.
The support of senior management is key to the sustainability of improvement projects.26 The ACP project was strongly supported by the departmental lead and senior doctors within the geriatrics department. The role of the ACP champions was essential, highlighted by the drop in ACPs when they were absent. They prompted daily ACP discussion of patients until the practice was embedded. To ensure sustainability, ACP champions were also tasked with handing over the role to their replacement colleague when they rotated.
We hope that the QIP has laid the foundations for an established change in practice using the multidisciplinary board rounds and a skilled workforce who are educated in initiating, documenting and communicating ACP discussions.
In the pre-project audit, patients spent on average 16.9 days in hospital before dying. In the follow up study, 18 out of 24 patients who were discharged avoided readmission. If these patients had been readmitted and died in hospital, it is assumed they would have also spent approximately 16.9 days each in hospital, equating to 304 bed days. The Department of Health and Social Care estimates that one patient bed day costs £400.32 If these 18 patients had been readmitted and died in hospital, this would have cost an estimated £121,000. This figure represents the potential saving for only one ward over 4 months.
End-of-life care costs are significant in the community. However, these have been shown overall to be cheaper than hospital.10
In the pilot study, 44 patients had an ACP but 20 died before discharge indicating that some ACP discussions needed to be initiated earlier.
Limitations
In the pre-project audit, it was not possible to retrieve 30 case-notes from the 80 deaths recorded. However, the notes reviewed were assumed to be representative of the total cohort.
When deciding if a patient was appropriate for an ACP, the teams discussed this between themselves based on the advised triggers. This was not formally standardised. The data collection relied on the teams accurately recording if the discussions took place. Variability in what was viewed as ‘initiating an ACP discussion’ may have limited the project.
Discharge summaries were reviewed to establish if ACP discussions were documented. The ward notes were not reviewed to confirm whether a telephone conversation between the ward team and the GP had occurred. Consequently, it is possible that discrepancies exist between the ACP content discussed on the ward and what was communicated to the GPs.
Data on the outcomes of patients for whom ACP was appropriate and not yet initiated were not collected. Thus, the true impact of the QIP on hospital readmissions is unknown.
Conclusions
This quality improvement project demonstrates a simple and successful strategy for improving ACP in the hospital setting. The interventions have shown a sustained change in practice although it is essential that education must be ongoing.
The project could be applied to many hospital departments or community settings.
If these strategies are employed across primary and secondary care, more patients would be able to express their preferences regarding goals of treatment and so receive individualised care.
Suggested next steps include establishing mandatory training in ACP communication skills for all HCPs and training for online systems for communicating information, such as ‘Coordinate My Care’.
Through local campaigns, patients should be encouraged to ask HCPs to have these discussions and feel empowered to take control of the remainder of their lives.
In summary, the simple interventions used in this project increased ACP within the secondary care setting, embedded it within geriatric ward practice and ensured more patients were provided with the opportunity to have their preferences at the end of life achieved.
Supplementary material
Additional supplementary material may be found in the online version of this article at www.rcpjournals.org/clinmedicine:
S1 – Timeline.
Acknowledgements
The authors would like to thank Dr Lucy Owen, Dr Sonal Sharma, Dr Joanna Pleming, Dr Tamar Ish-Horowicz, Dr Areej Benafif, Dr Annie Pervaiz, Dr Suresh Sivarajah and Dr Naomi Newman.
- © Royal College of Physicians 2020. All rights reserved.
References
- ↵
- The Gold Standards Framework
- ↵
- Royal College of Physicians
- ↵
- Seymour J
- ↵
- Schickendanz AD
- ↵
- Amblàs-Novellas J
- ↵
- Dying Matters
- ↵
- ↵
- ↵
- ↵
- Kernick LA
- ↵
- Davidson SN
- ↵
- ↵
- ↵
- ↵
- Simon J
- ↵
- Davison SN
- ↵
- Yates P
- ↵
- ↵
- ↵
- Detering KM
- ↵
- ↵
- Prod'homme C
- ↵
- Mason B
- ↵
- Brighton LJ
- ↵
- ↵
- Dixon J
- ↵
- Llewellyn H
- ↵
- ↵
- Wuensch A
- ↵
- Mudge AM
- ↵
- ↵
- Department of Health and Social Care
Article Tools
Citation Manager Formats
Jump to section
Related Articles
Cited By...
- No citing articles found.