Downloaded from http://spcare.bmj.com/ on June 28, 2015 - Published by group.bmj.com

Research

Introducing the Palliative Performance Scale to clinicians: the Grampian experience Gordon Linklater,1 Sally Lawton,1 Shona Fielding,2 Lisa Macaulay,1 David Carroll,1 Dong Pang3 1Department

of Palliative Medicine, NHS Grampian, Aberdeen, UK 2Division of Applied Health Sciences, University of Aberdeen, Aberdeen, UK 3Institute of Health Research, University of Bedfordshire, Luton, UK Correspondence to Gordon Linklater, Palliative Medicine, NHS Grampian, Roxburghe House, Ashgrove Road, Aberdeen AB25 2ZH, UK; [email protected] Received 27 September 2011 Accepted 9 February 2012 Published Online First 31 March 2012

ABSTRACT Objectives The Palliative Performance Scale (PPS) was introduced across NHS Grampian. Our aim was to determine how practical and useful the PPS was for clinicians looking after palliative patients in a variety of settings. Methods A prospective audit approach was used in primary, secondary and nursing home care settings who. Demographic and assessment data were gathered for 3 months; feedback was gathered at the end of the data collection phase. Patient follow-up status was determined at 12 months. Results Fifteen clinical sites participated and feedback was obtained from all clinical areas (n=30). Most respondents found the PPS easy to use and that it helped recognise disease progression in cancer patients, but not in patients with dementia/frailty. Assessment data were gathered on 666 patients. Sixty per cent had a malignant diagnosis and 62.5% of the sample died within 12 months. Lower PPS scores at initial assessment indicated poorer prognosis. Median survival figures differed from previously published data. Falling PPS scores increased the risk of death compared with patients whose PPS scores remained static or improved. Conclusion Clinicians found the PPS to be a quick, useful way of assessing and reviewing functional changes in palliative patients. However, it may not identify the subtle changes in individuals with advanced dementia. The survival figures confirm that caution is needed in generalising survival data across different settings and populations. Further work is needed to examine changing functional status in patients with nonmalignant diseases or dementia/frailty.

INTRODUCTION Palliative care is part of the daily work for almost all healthcare professionals. Patients are likely to spend 90% of their fi nal year of life in a primary care setting,1 and at least 12% of hospital inpatients are in their last 3 months of life. 2 Functional decline is a feature of many palliative conditions, but the rate of decline may vary according to the particular diagnosis3 4 and, therefore, it can be difficult to recognise when patients are entering the palliative or terminal phase of their illness. These uncertainties about prognosis can impede appropriate advance care planning. 5 6 It has been recognised that there is a need for more work on prognostic indicators that can inform end-of-life care planning.7 A number of functional assessment tools that link changes in functional status with survival have been developed BMJ Supportive & Palliative Care 2012;2:121–126. doi:10.1136/bmjspcare-2011-000133

and validated. The Scottish Government published ‘Living and Dying Well: a national action plan for palliative and end of life care’ in 2008. Action Point 2 states that all palliative patients should be assessed and reviewed in all care settings using recognised tools currently available.8 One such tool is the Palliative Performance Scale (PPS). The PPS was developed from the Karnofsky Performance Scale in Canada in 1996 as a way of providing a framework to assess the ‘progressive decline in palliative patients’.9 The PPS is a measure of performance, where an assessment is made of a person’s ability to ambulate, undertake self-care, eat and drink independently, level of consciousness and presence of disease, based on which a score ranging from 0 to 100 is allocated. The assessment takes less than 1 min to carry out and the score correlates with survival time.10 The lower the score, the shorter the survival time. There are reports of the use of the PPS in palliative care settings—palliative care units, hospices or hospital-based palliative consultation services—with the majority of patients having a cancer diagnosis.11–19 A meta-analysis exploring the results of studies undertaken in Canada and the USA reported that although the PPS is a strong predictor of survival from initial assessment following referral to specialist palliative care services, caution is needed in applying these fi ndings to other care settings.11 To our knowledge, there are no published data looking at the use of the PPS in hospital and community settings in the UK. The NHS Grampian Palliative Strategy Group was allocated the task of rolling out the Action Plan across the Grampian region. However, we felt it would be beneficial to understand more about the PPS and its use in practice, prior to full rollout across the region. This paper describes the introduction of the PPS across a range of primary and secondary care clinical settings in the northeast of Scotland and feedback from the clinicians using it. This is a pragmatic study describing the ‘real-life’ utility of the PPS. It is not a controlled clinical trial. The aim was to investigate how practical and useful the PPS is for clinicians looking after palliative patients in primary and secondary care settings across northeast Scotland. The specific objectives were: ▶ To gather feedback from clinicians about the ease of use of the PPS in the assessment of palliative patients. ▶ To gather feedback from clinicians about the helpfulness of the PPS in the identification of palliative nursing needs. 121

Downloaded from http://spcare.bmj.com/ on June 28, 2015 - Published by group.bmj.com

Research ▶

To assess the usefulness of PPS as a predictor of mortality.

METHOD Approach A prospective audit approach was used. Clinicians in primary, secondary and nursing home settings who care for palliative patients were eligible to participate. The sample included three general practices (primary care), a specialist palliative care unit, hospital specialist palliative care team and Community Macmillan Nurse referrals, an acute oncology ward, two longstay dementia wards, three medical wards for the care of older people and three care homes. The settings were chosen to reflect the range of palliative care work undertaken, both in primary and secondary care. Patients were selected for inclusion if the clinician looking after them believed they were ‘palliative’. We deliberately did not enforce more specific inclusion or exclusion criteria as we wanted to assess how the PPS would fit with ‘normal’ clinical practice. Each area was visited by the Senior Lecturer in Palliative Care and a short tutorial was given to explain the project and how to use the PPS. Clinicians were asked to record an initial PPS score for their palliative patients, then either regularly or whenever they felt there was a change in the patient’s condition for 3 months. Each clinical area identified a key contact (nurse or doctor) who forwarded assessment data as well as limited demographic data, which was collated by one of the investigating team (LM).

Clinician feedback After data collection was completed, feedback was sought from the key contact in each pilot site. The key contact was also invited to seek feedback from other members of their clinical team who had used the PPS. The location and designation of the respondent was recorded. Three questions were included in the feedback form with ‘yes’, ‘no’, ‘unsure’ responses and space for additional comments. The questions were: ▶ 1. Has the PPS been easy to use? ▶ 2. Do you think the PPS has been useful in helping to recognise disease progression in your palliative patients? ▶ 3. Is the PPS helpful in identifying palliative nursing needs? The feedback data were collated according to the comments made about each of the questions and grouped into recurring themes.

Statistical analysis Audit data collected included PPS assessment scores, patient location, diagnostic group,4 gender and date of death or status 12 months after the initial PPS assessment. Each participant had at least one PPS assessment with some as many as 13. The patients were grouped into three diagnostic categories (malignant, non-malignant and dementia/frailty) as these groupings are widely used to distinguish different disease trajectories, notably within the Gold Standard Framework. 20 SPSS (V.18) was used to analyse the data. To investigate whether initial PPS predicted mortality, Cox regression for time from the fi rst PPS assessment to death (censored at 1-year follow-up) was carried out. Initially unadjusted models were fitted followed by an adjusted model (for gender and location). For the purposes of modelling, the PPS was recalculated as a score out of 10, with 10 representing death, and included in the model as a continuous variable to allow the interpretation of 122

Table 1

Clinician feedback on the Palliative Performance Scale (PPS)

Question

Yes

No

Unsure

No answer

Total

PPS easy to use PPS helps to recognise disease progression PPS helps to recognise palliative nursing needs

29

0

1

0

30

23

4

3

0

30

16

5

7

2

30

the HR to be with respect to a 10% decrease in the PPS score (a worsening). Separate models were then fitted for each of the three diagnostic groups. To compare mortality at 12 months between those whose PPS scores had remained unchanged/showed improvement and those whose PPS scores worsened from the initial to the second PPS assessment, χ2 tests were used. Separate logistic models for each diagnosis for 12-month mortality from the fi rst PPS assessment were fitted with change in PPS as the explanatory variable (no change/improvement vs worsening). To account for the differences in timing between assessments, these models were adjusted for time between the fi rst and second assessments (≤14 days, 15–28 days, >28 days) for the malignant group and (≤28 days, >28 days) for the non-malignant and dementia/frail groups (due to smaller numbers).

Ethical approval Confi rmation was obtained from the Regional Ethics Committee that full ethical approval was not required as this was a clinical audit. The audit was registered with NHS Grampian’s clinical effectiveness department. All data received from the clinical settings were stored on an encrypted, password protected computer database in a locked office and were anonymised.

RESULTS Fifteen clinical sites participated in this audit. We fi rst present the feedback from clinicians on using the PPS, followed by analysis of the PPS to predict 12-month mortality.

Feedback from clinicians Completed forms were received from all (15/15) key contacts at the clinical sites. Forms were also received from other clinicians who had used the PPS at the clinical sites. In total, 30 forms were returned. Nineteen forms were returned from secondary care locations, 10 were from primary care and one person did not identify his/her location. The forms were completed by 23 nurses, 3 doctors, 3 nursing home managers and 1 person who did not identify his/ her role. Responses to the questions about how easy the assessment tool was to use, whether it helped in the recognition of disease progression and the identification of nursing needs are shown in table 1. Most respondents found the PPS easy to use and that it helped to recognise disease progression in cancer patients, but they were less likely to report that it helped to recognise palliative nursing needs.

Ease of use of the PPS assessment tool A number of respondents commented that the PPS was user friendly, easy to explain and understand as well as being quick to use. There were a number of comments about the difficulty in placing a patient on a specific percentage band BMJ Supportive & Palliative Care 2012;2:121–126. doi:10.1136/bmjspcare-2011-000133

Downloaded from http://spcare.bmj.com/ on June 28, 2015 - Published by group.bmj.com

Research

Table 2

Patient characteristics, N (%)

Proportion dead at 12 months Gender Male Female Location Specialist palliative care unit Primary Secondary Nursing home

Table 3

Malignant (N=397)

Non-malignant (N=163)

Dementia/frailty (N=106)

Total (N=666)

296 (74.6%) 186 (46.9%) 211 (53.1%) 83 (20.9%) 195 (49.1%) 116 (29.2%) 3 (0.8%)

81 (49.7%) 60 (36.8%) 103 (63.2%) 4 (2.5%) 10 (6.1%) 126 (77.3%) 23 (14.1%)

39 (36.8%) 35 (33.0%) 71 (67.0%) 0 (0) 9 (8.5%) 60 (56.6%) 37 (34.9%)

416 (62.5%) 281 (42.2%) 385 (57.8%) 87 (13.1%) 214 (32.1%) 303 (45.4%) 63 (9.4%)

Cox regression for time from first PPS assessment to death Unadjusted HR

Diagnosis

Gender Location

Initial PPS

Malignant Non-malignant Dementia/frailty Male Female Specialist palliative care unit Primary Secondary Nursing home 10% decrease

1.00 0.53 0.32 1.00 0.78 1.00 0.30 0.41 0.13 1.37

Adjusted* 95% CI

p Value

(0.41 to 0.67) (0.23 to 0.45)

Introducing the Palliative Performance Scale to clinicians: the Grampian experience.

The Palliative Performance Scale (PPS) was introduced across NHS Grampian. Our aim was to determine how practical and useful the PPS was for clinician...
179KB Sizes 0 Downloads 3 Views