Article Text

Download PDFPDF
Parkinson’s disease and palliative care: a quality of care Delphi study
  1. Alice Rogers1,
  2. Edward William Richfield2,
  3. Sue Thomas3,
  4. Bee Wee4,5 and
  5. Sophie Anne Trotter6
  1. 1OUCAGS, Oxford University, Oxford, UK
  2. 2Medicine for Older People, North Bristol NHS Trust, Westbury on Trym, UK
  3. 3Parkinson's Academy, Sheffield, UK
  4. 4Palliative Medicine, Oxford University Hospitals NHS Trust, Oxford, UK
  5. 5Harris Manchester College, Oxford University, Oxford, UK
  6. 6University of Leeds, Leeds, UK
  1. Correspondence to Dr Alice Rogers, OUCAGS, Oxford University, Oxford OX1 2JD, UK; alice.rogers10{at}


Objectives Extending palliative care services to those with long-term neurological conditions is a current aim of UK health policy. Lack of holistic guidelines for palliative and end-of-life care, and differing models of service provision, has resulted in heterogeneity in care access and quality. There is a need for evidence-based standards of care to audit Parkinson’s services and drive improvements.

Methods A two-stage Delphi process was used to achieve consensus on statements that define quality standards in palliative care for patients with Parkinson’s disease (PD). An expert panel was selected to comprise healthcare professionals, patients and carers based in the UK; this panel evaluated the statements via a Delphi survey. Quantitative and qualitative analysis of the results informed modifications between the Delphi rounds.

Results A final set of 16 statements was produced, reflecting aspirational standards of palliative care in PD. These statements, split into four domains (‘Structures and processes of care’, ‘Preparing for the end of life’, ‘ Care in the last weeks of life’ and ‘Care in the last days of life’) underline the importance of joint working between generalist and specialist services, individualised care and early and regular advance care planning.

Conclusions The Delphi process has established a set of standards which can be integrated within and guide services, helping to improve the quality and equality of care. Further work remains to establish the effectiveness of different models of service provision, including the implementation of keyworkers and telemedicine.

  • end of life care
  • neurological conditions
  • service evaluation
  • supportive care
  • chronic conditions

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Those with end-stage Parkinson’s disease (PD) have high physical impairment and carer burden.

  • Palliative care interventions are often only considered late in the disease trajectory.


  • This study is the first UK-based study that has defined formally what is recognised as ‘gold-standard’ care for patients with PD approaching or at the end of life.

  • The aspirational standards generated via the Delphi process are holistic and advocate personalised care and integration of services.


  • The quality standards can be used to audit current PD services and drive improvements.


Parkinson’s disease (PD) is a common life-limiting neurodegenerative condition. Those with end-stage PD have evidence of increased physical impairment and carer burden,1 associated with higher rate of admissions and death in hospital.2 Nonetheless, there remains a reluctance to consider palliative care for patients given the slowly progressive nature of the disease and physician uncertainty in how to implement palliative interventions.3 Extending palliative care services to those with long-term neurological conditions is a current aim of UK and international health policy.

There are limited guidelines to advise clinicians caring for patients with palliative or end-stage PD. National Institute for Health and Care Excellence guidelines for end-stage PD are non-specific and signpost clinicians to generic guidelines on managing palliative symptoms and advance care planning (ACP). Pharmacological guidelines, such as those in the Palliative Care Formulary (PCF), focus on physical symptoms and thus do not consider the full scope of holistic management at the end of life.

Furthermore, there are different models of service provision described in the literature4 and seen in practice across the UK. This inherently leads to service variability and likely inequity of access, highlighting the need for evidence-based standards of care.

To generate care standards, a Delphi approach was adopted to develop a set of statements that would summarise aspirational ‘gold-standard’ palliative and end-of-life care for PD patients. A Delphi methodology is a heuristic process that uses a group of experts in iterative rounds to develop a consensus of ideas. Delphi studies are necessary in palliative care, whereby they exemplify the importance of expert opinion in evidence-based medicine.5


A two-stage Delphi process was adopted to achieve consensus on quality standards in palliative care for patients with PD. The methodology was informed by Conducting and REporting DElphi Studies (CREDES) guidance to ensure credibility of the Delphi process (online supplemental appendix 1). The Delphi process is summarised in figure 1.

Supplemental material

Figure 1

Flow chart of Delphi process.


Principles as to what ‘gold-standard’ palliative care looked like were developed during a focus group of UK-based experts in palliative care and/or PD, which was held virtually in December 2020. Thematic analysis of the focus group transcript guided the development of the initial 19 statements used in the first round of the Delphi study (online supplemental appendix 2). The 19 statements were categorised into four domains:

Supplemental material

  1. Structures and processes of care.

  2. Preparing for the end of life.

  3. Care in the last weeks of life.

  4. Care in the last days of life.

Statements were phrased to represent aspirational standards of services delivering palliative care to patients with PD. Each statement was accompanied by a clarifying rationale text to explain the context behind the statement. The statements were shared among the authors and modifications made before being sent to a pilot audience.

Delphi survey

The statements and accompanying rationales were incorporated into an online survey format using the JISC Online Survey Tool. This software enabled participants to consent to being part of the Delphi study before progression to the survey, and anonymised participant responses. The survey was designed so that participants could rate statements using a 5-point Likert scale (1=strongly disagree, 2=disagree, 3=neutral, 4=agree, 5=strongly agree). In addition to the rating scales, there was further opportunity for participants to reword the statement and to comment on their reasons for (dis)agreement.

The online survey was sent to a pilot audience comprising two healthcare professionals to ensure ease of completion and technical functioning. The pilot audience were not participants in the panel.

Selection of expert panel

We included healthcare professionals and patients/carers to ensure that the statements reflected the lived experiences of those with PD.

Healthcare professionals were recruited via purposive sampling, using members of the initial focus group as well as others with experience in palliative care, geriatrics and neurology. Healthcare professionals were selected to represent a range of professions (nurses, doctors, therapists and commissioners) and geographical locations across England, to ensure a multidisciplinary and multi-setting perspective. Twenty-eight healthcare professionals were contacted to be part of the expert panel; three healthcare professionals declined to participate due to other commitments.

An advertisement for participation in the Delphi study was disseminated by Parkinson’s UK in April 2021. Over 90 patients and carers responded via email. Thirty-five respondents were chosen on basic demographic data to ensure a mix of patients/carers and genders.

Delphi rounds

The first Delphi round took place between June and August 2021; the second round took place between September and November 2021. Each round lasted 8 weeks.

Data analysis and definition of consensus

Quantitative statistical analysis of the participants' ratings of the statements was performed using statistical software SPSS V. (15). Outcome measures included percentage of agreement (total of rated ‘agree’ and ‘strongly agree’), median rating and interquartile range (IQR) as a crude measure of consensus (with an IQR ≤1 showing a high consensus among participants).

Thresholds of agreement and definitions of consensus were based on values used in previous studies.6 Four ‘bandings’ were created:

  • Very high agreement and consensus of a statement was defined as a percentage agreement of >85%, a median rating of 5 and an IQR of 0.

  • High agreement and consensus of a statement was defined as a percentage agreement of >85%, a median rating of 4 or 5 and an IQR of 1.

  • Moderate agreement and consensus of a statement was defined as a percentage agreement of <85%, a median rating of 4 or 5 and an IQR of 1.

  • Low agreement and consensus of a statement was defined as a percentage agreement of <85%, a median rating of 4 or 5 and an IQR of >1.

Participants' comments were coded and analysed thematically.

After each Delphi round, we used the quantitative and qualitative data to modify each statement. Statements that were shown to have ‘moderate’ or ‘low’ agreement/consensus underwent major revisions after each Delphi round to reflect the comments of the participants. Such modifications included either significant rewording or were integrated with another statement. Statements that were deemed to exhibit ‘very high’ or ‘high’ agreement received minor modifications, which were limited to minor textual adjustments or further clarification in the rationale.


Expert panel

In the first round of the Delphi study, there were 43 respondents (68% response rate). This comprised 19 healthcare professionals, of which there were 10 doctors, 3 nurses, 1 pharmacist, 1 therapist and 4 assigned ‘other’, as well as 14 patients and 10 carers.

In the second round of the Delphi study, there were 36 respondents (57% response rate): 17 healthcare professionals comprising 10 doctors, 3 nurses and 4 ‘other’, 11 patients and 8 carers.


In the first round, 19 statements were reviewed. Three were considered to meet ‘very high’ agreement/consensus, eleven were categorised as ‘high’, three ‘moderate’ and two ‘low’.

Of the five ‘moderate’ or ‘low’ statements, three statements underwent major revisions in the wording of the statement and rationale, and two were combined with another statement in the same domain. Two statements that met ‘high’ agreement/consensus (statement 6 and 7) were combined given that their associated qualitative comments were very similar, and it felt more fitting for the statements to be blended.

In the second round, 16 statements were evaluated by the expert panel. Seven statements increased in their consensus/agreement banding, six statements remained in the same banding and three statements decreased in banding. Of these latter three, one statement (statement 5) dropped from a ‘high’ agreement/consensus to ‘moderate’; it was decided by the authors to keep the statement given that the wording remained almost identical between rounds. Statements 10 and 11 dropped from ‘very high’ to ‘high’ bandings due to a small decrease in consensus.

Table 1 summarises the agreement and consensus of statements between the rounds. Online supplemental appendix 3 details the evolution of statements between each round.

Supplemental material

Table 1

Agreement and consensus of statements

Final statements

The final 16 statements (figure 2) reflect aspirational 'gold standards' of palliative care for patients with PD.

Figure 2

Finalised statements and rationales

Domain 1: structures and processes of care

Statements in this domain aimed to highlight the need for integrated care that was multidisciplinary, multi-agency and multi-setting. There was overall agreement of the need for better education of the whole multidisciplinary team (MDT) and improved access to specialist support for generalists caring for this cohort of patients.

The impacts of inadequate communication between providers can be significant if the person with Parkinson’s no longer has the ability to communicate their needs effectively or does not have someone else to help them do so. Round 2, Participant 32

The notion of a keyworker to coordinate care was seen to be beneficial, however, many of the respondents commented that this should be expanded to a ‘keyworker team’ to ensure that continuity can be maintained if staff members were not always at work due to sickness or leave.

I agree a key worker would co-ordinate care better and support patient and the family. Most of the time problems happen out of hours when the key worker is not around or when they are on leave. In addition to a key worker we need everyone to use a single system….that not only traverses primary and secondary care but also health and social care. Round 1, Participant 43

There was widespread appreciation of the vital need for access and sharing of health records. Ideally, records should be in an electronic format, but in many cases having readily available paper copies (eg, of DNACPR or ReSPECT forms) is more appropriate. Equally, access should not just be limited to healthcare teams, but also available to the local authority, social workers and hospice services. A number of respondents commented on personal distress caused by issues of knowledge transfer between teams.

Through my mother’s advancing journey with PD—what made things harder was having to explain about her advanced decision with every professional we met in every healthcare setting… Round 1, Participant 34

Statement four regarding red flags received the most modifications in this domain. Ideally patients should be known to specialist services (PD or palliative) working in an integrated fashion. If not there should be ‘red flags’ or triggers that prompt referral from primary teams (eg, primary care, neurology) to specialist teams. Red flags should reflect social as well as clinical milestones and should be patient specific.

Domain 2: preparing for the end of life

This domain focused predominantly on ACP discussions. There was an acknowledgement of the tensions between ensuring that patients and their family were supported but not forced to have such discussions.

Some palliative care team members can be a bit gung-ho with ACP conversations and put people off, studies suggest that some people simply don't want to engage in ACP and upset can be caused by trying to force it. Round 1, Participant 19

ACP should be available as early in the disease trajectory, but participants voiced concern over whose role it was to initiate the discussion. Many felt that the whole MDT should be engaged with ACP, with the caveat that this could create confusion as to whose responsibility it was.

One of the main issues with ACP is that it is everyone’s responsibility—but because everyone ‘can’ do it, often people assume that others will, and with time and resources being tight, it often gets left as a job for other professionals Round 2, Participant 17

The need for regular review of ACP wishes was highlighted, but there was a lack of consensus as to how prescriptive this should be (eg, yearly/monthly); instead, it was felt that as ‘changes can happen suddenly’ there should be ‘reasonable intervals’ to facilitate patients coming to terms with their own situation.

A one off conversation isn't enough as patients may not be able to think about these issues initially and brush the idea aside. Round 1, Participant 3

Domain 3: care in the last weeks of life

Participants advocated for patient-centred management in the last weeks of life, with a focus on maintaining comfort and not inadvertently prolonging life. Three participants commented on physician assisted suicide.

Surely it is more important to focus on keeping someone comfortable and as peaceful as possible. Investigations and procedures should only be done if absolutely essential to their whole well-being. Round 1, Participant 34

Medications, including but not limited to dopaminergic medication, should be regularly reviewed with the support of specialists to ensure that the benefits are weighed against potential side effects.

My experience was that the taking of the prescribed dose of Sinemet potentially had more of an adverse effect on cognitive behaviour than it did on enabling movement and such there could have been an argument that a lower dose may have in effect allowed a better quality of life in later days. Round 2, Participant 32

The use of guidelines, such as the PCF, was deemed to be less important compared with assessing each individual situation and prescribing accordingly. It should be noted that the PCF is behind a paywall, so not easily available for non-palliative care specialists.

The PCF is a fantastic resource, hugely helpful, but not available to all - only really used by palliative care specialists at present Round 1, Participant 19

Domain 4: care in the last days of life

Feedback from statements demonstrated that uncertainty in patient trajectory and prognostication should be sensitively discussed with patients and their carers. The dying phase should be promptly identified by healthcare professionals, but carers wanted ‘training’ in the recognition of what dying looked like.

I remember that when my dad died I was not aware of the changes in breathing that were representative of someone being in their last period of life and as such his death came as quite a surprise which shocked me greatly. It would have been helpful if someone had forewarned me and made me more aware of the fact that there were specific aspects that I could maybe look out for and be prepared. Round 1, Participant 35

Comments from carers demonstrated that there was a sense of fear around the prescription of anticipatory medications; the need for sensitive communication to facilitate better preparation was highlighted. A ‘good death’ was broadly defined as honouring patient wishes and not overly prescriptive in how or where it occurred. Participants felt that while it was important to signpost the availability of bereavement support, this should be guided by need.

Bereavement support is available but that doesn’t mean people know about it. They may not feel they need anything initially and not know they can access it after a long time. It is still a bit taboo for many. Round 1, Participant 34


Models of care

There was overall agreement from the Delphi process of the need for integration between services, with strengthening and education of the whole MDT. How to transform this from a theoretical aspiration into a functional reality is the challenge for emerging services and likely to depend on local circumstances. This makes a compelling case for care standards to guide development.

Questions arise as to who should take the primary responsibility for providing palliative care for patients with PD. As discussed in the 2016 European Association for Palliative Care (EAPC) Consensus,7 palliative care and neurology physicians can lack the training to be competent in each other’s field, creating a knowledge gap where patients with end-stage PD fall through. A neuropalliative physician with skills in both may be the solution,4 advocated by some Delphi participants.

The temporal integration of specialist services is also an uncertainty, with possibilities including specialist palliative support being offered based on patient triggers.8 Qualitative comments from the Delphi stressed the importance of relationship building between generalist and specialist services to ensure that patients receive timely care when they deteriorate. The use of a coordinating keyworker was championed by participants of the Delphi. The logistics of how the keyworker role should be assigned remain uncertain and would be dependent on local resources and service structures.

The setting of care is also important. A recent randomised control trial showed beneficial outcomes to quality of life and caregiver burden for patients with PD seeing specialist palliative care in an outpatient clinic setting.9 Nonetheless, this leaves a whole cohort of homebound patients ‘missing’ to secondary services. The importance of ensuring that integrated services can reach into the community was echoed by a recent study involving an interdisciplinary team visiting end-stage PD patients at home.10 Telemedicine was used in the COVID-19 pandemic and is another resource that can be used for homebound patients, but many consider it to not be a sustainable supportive long-term option.11 Exploration of acceptability of telemedicine for palliative care in PD would be relevant and should be addressed in future work.

ACP discussions

From the Delphi rounds, there was consensus that ACP discussions should be done sensitively, early and at regular intervals.

The need for ACP discussions is paramount, given that for patients with PD there are complex decisions to be made, such as those regarding interventions like gastrostomy feeding or ventilatory support as well as withdrawal of care.8 Cognitive decline is inevitable, with many Parkinson’s patients eventually developing dementia12 which adds to the necessity for earlier discussions.

In a recent systematic review, ACP was shown to be inconsistent, lacking in standardisation of what should be covered and often not adhered to.13 Barriers to ACP, such as denial and disease progression heterogeneity, meant that discussions often occurred in a crisis, such as an emergency admission to hospital. Patients feel that the onus was on them to initiate these conversations13 and clinicians have been shown to disagree as to when to start ACP discussions.14 MDT involvement in ACP is reported to provide greater holistic plans for patients, but can lead to a lack of ownership of the process, as was echoed in the Delphi.

Interventions at end of life

Holistic care in the last weeks and days of life is the cornerstone of palliative medicine. Psychological and spiritual care is equally as important as the management of physical symptoms; indeed, poorer spiritual health is associated with increased palliative symptoms.15

Recognition of both the palliative stage and the terminal phase remain uncertain for clinicians. In the focus group held prior to the Delphi study, clinicians reported difficulty in diagnosing dying in this cohort of patients. It was surprising to see how many carers from the Delphi panel wanted to be trained to recognise the dying phase. Management of terminal symptoms is challenging and many of the treatments for complex symptoms such as intractable nausea are informed by case studies; the need for specialist opinion and support for generalists is paramount.

Statements in domain three and four triggered some comments surrounding physician assisted suicide. Given this remains an evolving debate and not the focus of our study, we have not included any reference to this in the finalised statements and/or rationales.


Once published, it is hoped that these statements will be taken up by practitioners, service providers and commissioners to measure services against standards and drive improvements.

The development of PD palliative care services is likely to be different depending on location, health setting and population. Establishing core principles which can be embodied within and guide the development of these different service models is an important step towards achieving equality of provision for people affected by PD. The challenge for all stakeholders is how to achieve these ideals, in a post pandemic environment, which has highlighted particular issues in PD16 and increased demands on specialist palliative services. Future work is needed to establish the clinical and cost-effectiveness of these developments.


Not all healthcare professional groups were represented in the study, with no commissioners responding to the Delphi survey, and only one therapist and one pharmacist being involved in round 1. In order to anonymise participant responses, we could not guarantee that the same participants who responded to the survey in round 1 were the same in round 2, thus impeding consistency of opinions. The study protocol that we developed did not have a definitive plan as to how to modify statements based on their ‘banding’ from participant ratings. Instead, we adopted a common-sense approach, as advocated by Black et al,17 whereby modifications were based on banding as well as qualitative comments from participants.

We have not performed sub-analysis of the contributions from HCPs and patients/carers. While this would be interesting, it may detract from the aim of the project which was to produce a comprehensive set of statements with equal value to all stakeholders.


The Delphi process has established a set of standards which can be integrated within and guide heterogeneous services, helping to improve the quality and equality of care. The standards have underlined the importance of joint working between generalist and specialist services, individualising care and early and regular ACP. Further work remains to establish the effectiveness of different models of service provision, including the implementation of keyworkers and telemedicine.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Ethics approval

This Delphi study was reviewed and received ethics clearance through the University of Oxford Central University Research Ethics Committee (CUREC), reference R75435/RE001. Participants gave informed consent to participate in the study before taking part.



  • Twitter @alicerogers95

  • Contributors AR, with guidance from EWR, ST and BW, conceptualised and developed the project. Data collection from the focus group was performed by AR. Thematic analysis of the focus group and development of the initial statements was performed by AR and SAT. The Delphi survey was written by AR. Data analysis of the Delphi survey results was performed principally by AR, with contribution from EWR, ST and BW. The primary report was authored by AR with editing and review by EWR, ST and BW.

  • Funding Author AR completed this study as part of her Academic Foundation Programme.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.