Article Text


Development of tools to facilitate palliative and supportive care referral for patients with idiopathic pulmonary fibrosis
  1. Charles Sharp1,
  2. Heather Lamb1,
  3. Nikki Jordan2,
  4. Adrienne Edwards3,
  5. Rachel Gunary4,
  6. Patricia Meek4,
  7. Ann B Millar5,
  8. Clare Kendall2 and
  9. Huzaifa Adamali1
  1. 1 Bristol Interstitial Lung Disease service, North Bristol NHS Trust, Bristol, UK
  2. 2 Palliative Medicine, North Bristol NHS Trust, Bristol, UK
  3. 3 Canterbury District Health Board, Canterbury Respiratory Service, Christchurch, Canterbury, New Zealand
  4. 4 Clinical Psychology, North Bristol NHS Trust, Bristol, UK
  5. 5 Academic Respiratory Unit, University of Bristol, Bristol, UK
  1. Correspondence to Dr Huzaifa Adamali, Bristol Interstitial Lung Disease Service, North Bristol NHS Trust, Bristol BS10 5NB, UK; huzaifa.adamali{at}


Objectives Palliative care is underused in non-malignant respiratory diseases, including interstitial lung diseases (ILDs). We investigated current practices around palliative and supportive care and explored the impact of a supportive care decision aid tool.

Methods This was a single centre study in a UK ILD centre. Retrospective analysis of hospice referrals and patients with idiopathic pulmonary fibrosis (IPF) under the Bristol ILD (BILD) service were used to identify unmet palliative and supportive care needs. Using quality improvement methodology, we explored the impact of a supportive care decision aid on clinician behaviours for patients with ILD.

Results 108 patients with ILD were referred for hospice care between 2010 and 2015, representing 0.15% of all referrals, compared with a population prevalence of IPF of 0.9%. The median interval between referral and death was 124 days.

Records were reviewed for 64 deceased and 89 living patients with IPF seen on July–December 2014. The decision aid was prospectively assessed with 73 patients. The deceased patients had greater markers of severity. There were no other differences between the groups.

After introduction, the decision aid tool was completed for 49.3% of patients and resulted in significant increases in documented discussion of referral to palliative care (11.2%vs53.6%, p<0.01) and end-of-life discussions (15.7%vs91.8%, p<0.01). Tool completion led to an increase in referral for palliative care (2.7%vs16.7%, p<0.01).

Conclusion Palliative care services are underused in ILD and a supportive care decision aid can prompt consideration of palliative and supportive care needs.

  • interstitial Lung Disease
  • non-malignant respiratory disease
  • supportive care
  • chronic conditions
  • clinical decisions
View Full Text

Statistics from

The aim of palliative care is to provide symptom relief and alleviate suffering in order to maximise quality of life for patients with incurable disease and can be offered alongside other active treatments.1 This resource is underused in non-malignant respiratory diseases despite these patients often having a higher symptom burden, lower quality of life and greater social isolation compared with patients with cancer.2

Some of the difficulties faced by patients with idiopathic pulmonary fibrosis (IPF) are mirrored by patients with end stage renal failure on haemodialysis, many of whom have similarly irreversible disease with significant restrictions imposed on daily activities and on quality of life. An advance care planning intervention used in dialysis patients improved outcomes for carers, reducing anxiety and depression and improved congruence of carer and patient wishes.3 This study did not examine patient outcomes separately. Other studies have demonstrated that quality of end-of-life care is significantly better for those with cancer and dementia than for patients with end stage renal failure,4 a phenomenon also observed in chronic lung disease.2 A retrospective audit of deceased patients under the care of a regional interstitial lung disease (ILD) service highlighted that 24% of patients had unmet palliative and supportive care needs.5

This evidence and the recent publications of the European Idiopathic Pulmonary Fibrosis Patients’ Charter6 and the National Institute for Health and Care Excellence (NICE) quality standard7 suggest that care of these patients could be improved by increased use of palliative and supportive care services. In order for this to happen, clinicians would need to systematically identify the needs of their patients, have knowledge of available services and make appropriate referrals to these services.

Although there is a small amount of evidence demonstrating benefit to patients with IPF from early referral to palliative care, further work is needed as is clarification of the optimum timing of referral. In order to progress the evidence, we investigated current practices around palliative and supportive care in ILD and IPF in particular. This identified the potential for a supportive care decision aid tool to improve documentation of end-of-life discussions with patients and to ensure referral for the appropriate level of palliative and supportive care services, including community nursing, general practitioner (GP) and hospice support. This interventional tool was developed through multidisciplinary discussion and implemented, allowing further potential refinements to be identified.


This was a single centre study in a regional referral centre for ILD in the UK. During outpatient assessment, patients’ care needs were assessed, resulting in recommendations to the GP for further supportive measures and referral to community matrons (nurses who provide advanced clinical nursing care in addition to case management for those patients with high care needs) and/or specialist palliative care as appropriate. Ethical approval for this work was given by the East of England Research Ethics Committee (REC reference 15/EE/0023).

We used quality improvement theory and in particular DMAIC (Define–Measure–Analyse–Improve–Control) methodology.8 This approach seeks to define a problem, measure its impact, subsequently analysing the causes for the identified issues. The cycle continues by introducing an intervention to improve matters and test it, before a pause to ensure improvements are sustained. The process used is outlined below.

Defining and measuring the unmet care needs of patients with IPF

To assess current practice and identify unmet palliative and supportive care needs for patients with IPF and other interstitial lung diseases, a combination of approaches were taken, including an examination of variables known to indicate advanced disease state and poor prognosis. Referrals to the local hospice, St Peter’s Hospice in Bristol, were examined to identify the diagnoses and hospice services to which those with a diagnosis of ILD were referred. In addition, mortality outcomes were examined for those referred for hospice services. These services include inpatient care, Day Hospice support, home care or specialist symptom management advice. A retrospective review was performed of both those referred to the Bristol ILD (BILD) Service within a 3-month period and deceased patients with IPF. These data were collated from hospice records and demographic batch service tracing from NHS records.

For all patients with IPF referred to the ILD service, demographic details, pulmonary physiology and exercise testing results were collated, in addition to oxygen requirements and subsequent mortality as outlined above. The case notes were examined for documented resuscitation and end-of-life discussions and referral to hospice, GP or community matron services.

Developing an intervention to improve the ability to meet care needs

A supportive care decision aid tool was adapted from one in use with renal dialysis patients at our hospital. The decision aid tool was designed to highlight known poor prognostic factors in IPF, symptoms of concern and changes in functional status and was refined by an iterative process involving members of the ILD, palliative care and psychology teams. This tool was designed for completion by clinicians in an outpatient setting. Outcomes following clinic visits were recorded including any referrals made or discussions held with the patient. The aim of the tool was to prompt clinicians in the middle of a busy clinic to identify those patients with potential palliative and supportive care needs, thereby ensuring that appropriate referrals were made.

Assessing the impact of the intervention

The supportive care decision aid was trialled for all patients with IPF seen in ILD clinics over a 3-month period from 1 October 2015 to 31 December 2015. The same data were collected for this cohort of patients as for the two cohorts of retrospectively reviewed patients. The impact of this tool on clinician response to identify support needs was assessed by comparison of the same four areas of interest: (1) documented discussions about end-of-life care, (2) resuscitation, (3) documented decisions about referral for hospice and (4) community matron services or other services.

Controlling and improving the intervention

Informal feedback was gathered from clinicians using the tool, patients and members of the multidisciplinary team including palliative care specialists and clinical psychologists. This feedback was summarised at multidisciplinary meetings to inform refinement of the decision aid tool for further iteration in a quality improvement cycle.


Defining, measuring and analysing the unmet palliative and supportive care needs in IPF

Hospice referrals

Data were available from referrals to St Peter’s Hospice between 2010 and 2015. These data demonstrated that 108 patients were referred a total of 158 times with a diagnosis of ILD as defined by ICD-10 codes. Patients referred on more than one occasion were referred for different hospice services. These referrals represent 0.15% of the total of 10 299 referrals received by the hospice in this period. St Peter’s Hospice covers a population of approximately 1 000 000 and recent estimates suggest that IPF accounts for 0.9% of all deaths in the UK,9 indicating that there may be as many as six times this number with IPF and care needs at the end of life.

The mean age at the time of referral was 75.7 years and 68.5% of patients were male. 26.8% were referred more than once. Diagnoses are shown in figure 1. The hospice service for which patients were referred is shown in figure 2.

Figure 1

Diagnoses of patients referred for hospice services (% of total referrals). CPFE, combined pulmonary fibrosis and emphysema; CTD-ILD, connective tissue disease-associated interstitial lung disease; HP, hypersensitivity pneumonitis; ILD, interstitial lung disease; IPF, idiopathic pulmonary fibrosis; NSIP, non-specific interstitial pneumonia.

Figure 2

Hospice service for which patients with interstitial lung disease were referred (% of total).

Patients referred for hospice services had severe lung function impairment and reduced exercise capacity. 71.4% required oxygen with 34.2% requiring an oxygen flow rate >4 L/min to maintain exertional saturations (table 1).

Table 1

Pulmonary physiology and exercise testing results at the time of referral for hospice services

Referral for hospice services was late in the course of disease for the majority of patients with a median interval between referral and death of 124 days. Those referred for Day Hospice support had greater survival after referral (table 2). This was also seen for those patients with connective tissue disease-associated ILD (median survival 214 days compared with overall median survival 124 days, p<0.01). 5.6% of patients died before they were assessed by the hospice, with a median interval between referral and death of 11 days.

Table 2

Interval between referral for hospice services and death

Patients reviewed in ILD clinics

The review of patients seen in ILD clinics identified 89 patients with IPF seen between July and December 2014 (‘pretool’) and 64 patients with IPF who died between 2011 and 2015 (‘deceased’). Seventy-three patients were seen in outpatient clinics from October to December 2015 (‘post-tool’) during the initial trial of the supportive care decision aid. Demographic and pulmonary physiology variables are shown in tables 3 and 4. There were no differences between the two cohorts of patients identified in the time periods identified. However, deceased patients had greater physiological impairment and were more likely to be on long-term oxygen therapy than either of the outpatient cohorts.

Table 3

Demographic and physiological parameters for the three cohorts of patients: pretool introduction, deceased patients and after-tool introduction

Table 4

Characteristics of the three cohorts: pretool introduction, deceased patients and after-tool introduction

Developing an intervention to improve the ability to meet care needs

A supportive care tool was developed as described in the methods and was assessed during an initial trial phase, gathering feedback from clinicians and patients about its use. During this trial phase, the supportive care tool was completed for 49.3% of patients with IPF reviewed in ILD clinics. Significant increases were observed in documentation of resuscitation decisions and end-of-life discussions, in addition to decisions related to GP, community matron and specialist palliative care referral (figure 3). Discussions and consideration of referral to palliative and supportive care services did not always result in referral, and there was no evidence of an increase in referral for community matron support or for specialist palliative care. This remained highest (32.9%) in those patients with IPF who subsequently died.

Figure 3

Impact of the supportive care decision aid tool on documented resuscitation decisions, documentation of consideration for community matron or specialist palliative care referral and completed referral and documented end-of-life discussion in the pretool (pre), deceased and tool assessment (post) cohorts. *p<0.001. DNACPR, Do Not Attempt Resuscitation.

When exploring the impact of the decision aid, comparison was made between those patients for whom the decision aid was completed and those where this was not done. There were no significant differences between these two groups in demographics or markers of disease severity. The decision aid resulted in much greater documentation of all identified areas.

There were significantly more referrals to specialist palliative care and community matron care where the decision aid was completed (figure 4). Documentation of end-of-life discussions was also significantly increased in those for whom the decision aid was not used (83.8% compared with 15.7% before its introduction, p<0.0001 and 39.1% in those who died, p<0.0001). This may represent raised awareness of these issues among clinicians while the trial was taking place.

Figure 4

Impact of supportive care decision aid completion on documented resuscitation decisions, documentation of consideration for community matron or specialist palliative care referral and completed referral and documented end-of-life discussion in the pretool (pre), deceased and tool assessment (post) cohorts. *p<0.01. DNACPR, Do Not Attempt Resuscitation.

Controlling and improving the intervention

The results of the initial trial phase informed development and refinement of the supportive care decision aid. Multidisciplinary discussion was held on three occasions, with patient views represented by an ILD specialist nurse and additional expertise from a consultant and specialist nurse in palliative care, two clinical psychologists, two ILD specialist respiratory consultants who had used the tool in clinical practice. This iterative process led to revision of the tool on each occasion, followed by further discussion.

As has been reported in the literature, clinicians can sometimes find it difficult to initiate conversations with patients about poor prognosis and end-of-life issues when there is no invitation from patients to discuss these issues10 and this was reflected in the multidisciplinary discussions. Our patient feedback included expressions of surprise at these topics being raised and an expressed lack of awareness of the severity of their disease. This led to the modification of the decision aid and the addition of a patient completed, holistic assessment tool. This form is currently in use in ILD outpatient clinics and is due to be evaluated in the future as part of the continuous cycle of quality improvement.


Patients with ILD are referred late in the course of disease for hospice services, and decisions related to supportive, palliative and end-of-life care are poorly documented, including in those in the period immediately before their death.

A simple decision aid tool resulted in both improved recognition of patients with IPF with potential need for palliative and supportive care and documentation of decisions and discussions related to this care and end-of-life management. There was also an increase in the frequency of documentation of end-of-life discussions even without completion of this decision aid. This may reflect heightened awareness of the need to discuss these issues following the development of the decision aid within the department. Use of the decision aid also resulted in an increase in the rates of referral to community matron or specialist palliative care.

The increase in documentation of decisions related to palliative and supportive care show that this simple tool was an effective means to screen poor prognostic markers of disease to highlight those patients with potential care needs. However, multidisciplinary feedback has suggested refinement to combine the screening process with a patient-led holistic needs assessment to identify the issues of concern and thus guide appropriate palliative and supportive care referrals.

The introduction of this decision aid was relatively straightforward within a small department and was introduced at a time of readiness for improvement in palliative and supportive care provision and in the context of other initiatives to achieve this. There was therefore increased awareness of these issues and that may explain some of the improvements seen.

These data do have limitations. The hospice data are difficult to verify, being based on diagnoses recorded on referral to hospice services, often by non-specialist clinicians. While it was possible to verify a large number of the diagnoses, it is likely that due to the nature of the records kept, patients will have been missed, therefore underestimating use of hospice resources. These data are also based on a single, large hospice in an urban setting. As such, generalisability of the findings may be limited.

The range of slow or rapid progression and unpredictable precipitous decline in IPF11 often leads to an underappreciation by patients and carers, in addition to some clinicians, of the severity of their disease.12 13 This finding is supported by an analysis of Swedish registry data surrounding end-of-life care in ILD and lung cancer.14 This demonstrated that in patients with ILD, death was more likely to be unexpected and less likely to occur in a palliative care setting. Additionally, it demonstrated that the initiation of long-term oxygen was a clear poor prognostic marker for ILD, with death occurring 8 months later in that group. End-of-life discussions were less frequent despite a greater burden of dyspnoea, which was unrelieved.

Considering the high levels of stress, depression and anxiety related to the often remorseless progression of disease, patients with IPF could benefit from early referral for palliative care.15 16 The best time point for such referrals remains unclear and data are lacking surrounding the provision of palliative care in IPF, despite this being indicated a quality standard by the NICE.7

In metastatic lung cancer, early referral to palliative care resulted in improved quality of life and improved survival by 3 months, despite less aggressive treatment.17 In IPF, one single centre study demonstrated that 57% of those dying with IPF did so in hospital, with only 13.7% referred to palliative care, of which the majority were within a month of their death.18

A separate study involving qualitative, semistructured interviews with patients, their carers and health professionals highlighted a high frequency of uncontrolled symptoms of dyspnoea, cough and insomnia, in addition to frustration and anger at the manner in which their disease impacted on their quality of life.15 In addition, health professionals expressed varying knowledge and confidence in managing these symptoms, along with underestimation of the psychosocial needs of their patients.

There has been only one randomised controlled trial examining the impact of a palliative care intervention in advanced fibrotic lung disease; the early implementation of a community case conference was compared with delayed implementation in a group of 53 patients.19 Benefit was seen on the Palliative Care Outcome Scale, a measure of symptoms and concerns.

There is limited evidence in the literature of tools to facilitate palliative and supportive care in ILD. Boland et al adapted a Needs Assessment Tool originally designed for use in cancer.20 Using an iterative process, they refined this to reflect the experience of patients, their caregivers and clinicians and an increased emphasis was given to respiratory symptoms as a result of this process. This was also reflected in the refined patient-completed tool in this study.

A universal finding from literature of palliative care in fibrotic lung disease and non-malignant lung disease in general is that palliative and supportive care are underused, with resulting poor symptom management in this group,1 that patients would appreciate improved communication and co-ordination of care12 13 and that palliative care measures are poorly integrated into routine care.21


The work reported here underlines these findings of underuse of palliative and supportive care services in ILD. We have demonstrated that a simple supportive care decision aid can result in significant increases in consideration of palliative and supportive care and end-of-life issues, with associated increases in the rates of documented referral for these services. It is not possible to conclude that referrals themselves have increased as a result of the decision aid as this methodology would not capture previous referral behaviours of clinicians with less comprehensive documentation of these decisions. The impact and outcomes of a refined approach with a supportive care decision aid alongside a patient completed holistic assessment tool will be evaluated as part of a continuing improvement project within the BILD service. This aims to empower patients and their carers to optimise quality of life using all available resources.


View Abstract


  • Contributors CS, ABM, CK and HA conceived and designed the study. CS, HL, AE and HA collected the data. CS performed the data analysis. All authors participated in tool development and refinement and in writing and approving the manuscript.

  • Competing interests None declared.

  • Ethics approval East of England Research Ethics Committee (RECreference 15/EE/0023).

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

  • Correction notice This paper has been amended since it was published Online First. Owing to a scripting error, some of the publisher names in the references were replaced with ’BMJ Publishing Group'. This only affected the full text version, not the PDF. We have since corrected these errors and the correct publishers have been inserted into the references.

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.