PRECEDING DEATH
N Smallwood, L Ross, J Taverner, J John, A Baisch, L Irving, J Philip. A palliative approach is adopted for many patients dying in hospital with chronic obstructive pulmonary disease. COPD, 2018, In press. (Final accepted version)
A palliative approach is adopted for many patients dying in hospital with chronic obstructive pulmonary disease
Authors & Affiliations
Corresponding Author: Dr Natasha Smallwood
Dr Natasha Smallwood BMedSci MBBS SpecCertPallCare MSc MRCP FRACP
1. Department of Respiratory and Sleep Medicine, The Royal Melbourne Hospital, Royal Parade, Parkville, Victoria 3050, Australia. Tel: 03 9342 7708.
2. Department of Medicine (Royal Melbourne Hospital), University of Melbourne, Parkville, Victoria 3050, Australia. [email protected]
Dr Lauren Ross MBBS
Department of Respiratory and Sleep Medicine, The Royal Melbourne Hospital, Parkville, Victoria 3050, Australia. [email protected]
Dr John Taverner MBBS
Department of Respiratory and Sleep Medicine, The Royal Melbourne Hospital, Parkville, Victoria 3050, Australia. [email protected]
Jennifer John BDS MPH
Department of Rural Health, University of Melbourne, Northeast Health Wangaratta, 35–47 Green St, Wangaratta, Victoria 3677, Australia. [email protected]
Dr Andreas Baisch MBBS MD FRACP
1. Department of Medicine, Northeast Health Wangaratta, 35–47 Green St, Wangaratta, Victoria 3677, Australia. [email protected]
2. The Department of Rural Health, University of Melbourne.
A/Prof Louis Irving MBBS FRCGP FRCP
Department of Respiratory and Sleep Medicine, The Royal Melbourne Hospital, Parkville, Victoria 3050, Australia. [email protected]
Prof Jennifer Philip PhD FAChPM MMed MBBS
1. Chair of Palliative Medicine, University of Melbourne, St Vincent’s Hospital and Victorian Comprehensive Cancer Centre.
2. St Vincent’s Hospital, Victoria Parade, Fitzroy, Victoria 3065, Australia.
Keywords: Palliation, palliative approach, palliative care, COPD, dyspnea, end of life
Abbreviations: COPD: Chronic obstructive pulmonary disease, RMH: Royal Melbourne
Hospital, NHW: Northeast Health Wangaratta, NIV: Non-invasive ventilation, ICU: Intensive Care Unit, HDU: High Dependency Unit, & RCU: Respiratory Care Unit
Funding
This work was supported by Palliative Care Research Network, who provided a PhD scholarship for Dr Natasha Smallwood.
Abstract
Severe chronic breathlessness in advanced chronic obstructive pulmonary disease (COPD) is undertreated and few patients access specialist palliative care in the years before death. This study aimed to determine if symptom palliation or a palliative approach were delivered during the final hospital admission in which death occurred.
Retrospective medical record audits were completed at two Australian hospitals, with all patients who died from COPD over twelve years between 1/1/2004 and 31/12/2015 included.
Of 343 patients included, 217 (63%) were male with median age 79 years (IQR 71.4-85.0). Median respiratory function: FEV1 0.80L (42% predicted), FVC 2.02L (73% predicted) and
DLco 9 (42% predicted). 164 (48%) used domiciliary oxygen.
Sixty (18%) patients accessed specialist palliative care and 17 (5%) wrote an advance directive prior to the final admission.
In the final admission 252 (74%) patients had their goal of care changed to aim for comfort (palliation) and 99 (29%) were referred to specialist palliative care. 286 (83%) patients received opioids and 226 (66%) received benzodiazepines, within 1 or 2 days respectively after admission to palliate symptoms. Median starting and final opioid doses were 10mg (IQR=5-20) and 20mg (IQR=7-45) oral morphine equivalent/24 hours. Hospital site and year of admission were significantly associated with palliative care provision.
Respiratory and general physicians provided a palliative approach to the majority of COPD patients during their terminal admission, however, few patients were referred to specialist palliative care. Similarly there were missed opportunities to offer symptom palliation and a palliative approach in the years before death.
Introduction
Palliative care aims to improve the quality of life of terminally ill patients and their families through the prevention and relief of suffering (1). While specialist palliative care arose in response to the needs of cancer patients, it’s goal of providing holistic care to manage distressing symptoms, psychosocial and spiritual issues, are equally relevant to patients with advanced respiratory diseases (2-6). The American Thoracic Society recommends that all health professionals who care for patients with chronic or advanced respiratory diseases should be trained in and capable of providing basic, generalist palliative care or a palliative approach to their patients (3, 7-9). Nevertheless many health professionals have received no formal palliative care or communication skills training and cite this as a significant issue when caring for patients with advanced respiratory disease (10-16). By contrast, specialist palliative care is provided to patients with complex symptoms or needs by health professionals who have completed additional training in palliative care (17).
Few patients with advanced chronic obstructive pulmonary disease (COPD) access specialist palliative care. In the United States only 1.7% of COPD patients receiving long-term oxygen therapy were referred to specialist palliative care when admitted with an exacerbation (18). Similarly large cohort studies from the United Kingdom (UK) and Australia suggest only 16.7% - 17.9% of COPD patients receive any specialist palliative care (19, 20). Difficulty estimating the prognosis due to the highly variable and unpredictable disease trajectory in COPD is an important barrier to specialist palliative care referral (13, 21). While there is usually a gradual decline over many years, any exacerbation can abruptly change function or be fatal (22-24). Therefore it is extremely challenging to determine the short-term prognosis or predict death from COPD (22, 23).
known about any palliative care these patients receive when they die. As 72% of COPD patients die in hospital in Australia (25), this study aimed to determine if symptom palliation or a palliative approach were delivered during the final hospital admission in which death occurred. Additionally we aimed to identify if end-of-life care varied according to geographical location.
Methods Study sites
The Royal Melbourne Hospital (RMH) is an Australian, metropolitan, teaching hospital with 450 acute beds. Both respiratory medicine and palliative care services (including specialist medical and nursing staff) are available. Multiple specialty outpatient clinics are available, including an integrated respiratory and palliative care clinic (the Advanced Lung Disease Service) established in 2013 for patients with non-malignant, respiratory disease (26). Inpatient care for respiratory patients is provided on medical wards or the respiratory ward, with non-invasive ventilation (NIV) available in the Respiratory Care Unit (RCU), Intensive Care Unit (ICU) or High Dependency Unit (HDU). Additionally there is an inpatient palliative care unit onsite.
Northeast Health Wangaratta (NHW) is a major, regional hospital in Australia with 140 acute beds. There is no specialist respiratory service or clinic. Outpatient care occurs through the NHW general medicine clinic or privately. A specialist palliative care nurse is available for consultation. Inpatient care is provided by general physicians on medical wards, with NIV available in the Critical Care Unit, which offers ICU and HDU level care.
Study design
Retrospective medical record audits were completed at RMH and NHW of all patients who died at each site over twelve years between 1st January 2004 and 31st December 2015, with
COPD listed in one of the first five diagnoses at death. Only those patients who died as a consequence of COPD (including respiratory infections) and not from other comorbidities were included. Patients with any co-existing malignancy were excluded as these patients may have accessed palliative care for their malignancy and not COPD. Only patients with a confirmed diagnosis of COPD (based on the history and previous spirometric testing confirming obstruction) were included.
COPD deaths were identified from each site’s mortality database using the International Classification of Diseases codes: J43-44. Medical records, administrative and clinical databases were examined to explore how clinical care was delivered, and to determine the use and duration of treatments with preventive, therapeutic and palliative intent during the final admission in which death occurred. Using definitions developed by Philip et al (27), preventive treatments include those given for comorbidities such as hypercholesterolaemia, therapeutic interventions included (but were not limited to) antibiotics, oral steroids, and non-invasive ventilation, and palliative treatments included opioids and benzodiazepines to palliate symptoms. Offering a palliative approach was defined as any of: initiating palliative treatments, discussing or documenting advance care planning, documenting that the goal of care in the terminal admission was to provide comfort care (palliation), completing a resuscitation status form or referral to specialist palliative care.
Data were collected regarding pre-admission disease severity and previous inpatient and outpatient hospital care, including previous access to specialist palliative care. Whilst all patients had spirometry results available, only results within two years of death were
included. Ethics approval was granted for this study by the Melbourne Health Research Office (Approval number: QA2014171) and by the Northeast Health Wangaratta Human Research Ethics Committee (Project number: 169).
Statistical analysis
Data are reported descriptively using counts and frequencies, with median values and interquartile ranges reported for variables with significant distribution skew. Data from each site were compared using Student’s t test for continuous variables and the Chi Square test for categorical variables. The twelve years included in the study period were divided into three 4-year blocks: 2004-2007, 2008-2011 and 2012-2015. Logistic regression was then used to examine associations between palliative care provision and year of admission. Statistical analyses were performed using IBM SPSS Statistics version 24. Statistical significance was indicated by p<0.05.
Results
Patient characteristics
Six hundred and ninety-one COPD deaths were identified, of which 343 met eligibility criteria with 221 patients from RMH and 122 from NHW (Figure 1). Almost two thirds (217, 63.3%) of patients were male, with median age 79.4 years (IQR 71.4-85.0) (Table 1). At NHW there were significantly more never smokers (p<0.0001) and more patients with coexisting respiratory diseases (p<0.0001) or co-existing cardiac disease (p=0.005) compared with RMH. At RMH significantly more patients had documented cor pulmonale (p<0.0001). There were no other significant differences in demographic or disease characteristics between the two groups.
While the diagnosis of COPD was confirmed for all included patients, only 166 (48.4%) had completed pulmonary function tests within the 2 years before death. The modified Medical Research Council (mMRC) breathlessness score was documented for only 9 (2.6%) patients (all from RMH) and was inferred from outpatient notes for remaining RMH patients, with 153 (69.2%) having a score of 3 or 4, suggesting severe breathlessness.
Previous respiratory management
Approximately half (195, 56.6%) the cohort had attended an outpatient appointment in the two years prior to death, with significantly more patients receiving outpatient care at NHW (p<0.0001) (Table 2). Sixty (17.5%) patients had accessed specialist palliative care (from the Advanced Lung Disease Service, community palliative care and/or inpatient specialist palliative care team) prior to the terminal admission and 17 (5.0%) had written an advance directive (AD). There was no significant difference between the two sites in access to specialist palliative care or completion of advance care planning (ACP), however, the median time between writing an AD or discussing ACP and death was significantly longer at NHW (462 days, IQR= 233-614) compared with RMH (152 days, IQR-56-303; p=0.042). Twenty-one (6.1%) patients were prescribed opioids for chronic breathlessness as outpatients prior to death. One hundred and six (30.9%) patients had previous admissions in which a resuscitation status decision was made, with this occurring more frequently at RMH than NHW (p=0.001).
Terminal admission care
One third of patients (111, 32.4%) had pneumonia on admission, with similar proportions at each site (RMH 32.6% and NHW 32.0%), and the majority (254, 74.1%) were admitted under general medicine (Table 3). Of the 160 patients not admitted under respiratory medicine at RMH, 61 (27.6%) were referred to respiratory medicine within 1 (IQR=0-4) day from
admission. In keeping with the greater availability of HDU/ICU beds at RMH, significantly more patients were admitted to these beds at RMH compared with NHW (p=0.04).
Medications prescribed during the terminal admission to all patients and by each site are shown in Figures 2 and 3. All treatments were initiated on admission, except for statins and antiplatelet medications, which had been commenced previously. The median number of antibiotics prescribed was 2 (IQR=2-3) with no difference by site, however, antibiotics were significantly more likely to be ceased before death at RMH compared with NHW (p<0.001). Nebulised bronchodilators (p=0.005), oxygen therapy (p=0.011) and low molecular weight heparin (p=0.007) were significantly more likely to be ceased before death at NHW. Non- invasive ventilation (NIV) was significantly more likely to be both initiated (p=0.014) and then ceased before death (p<0.0001) at RMH compared with NHW.
Having a documented goal of care aiming for comfort (palliation) was associated with ceasing antibiotics (p<0.0001), oral steroids (p<0.0001) and NIV (p<0.0001). Forty-four (12.8%) patients received antibiotics and oral steroids, and 88 (25.7%) received nebulised bronchodilators despite having a documented goal of care aiming for palliation.
In the 275 patients in whom oxygen therapy was continued until death, for the majority (193, 70.2%) the documented intention was for breathlessness palliation. Of the patients who continued using NIV until death, significantly more used this for symptom palliation (not to treat hypercapnoeic respiratory failure) at RMH (16, 11.3%) than at NHW (10, 4.5%; p=0.034).
For the whole cohort the median time from admission to death was 5 (IQR=2-10) days, with no significant difference by site (Figure 4). Within the whole cohort for patients in whom
antibiotics were ceased before death, the median time to stop antibiotic therapy was on day 3 (IQR 1-7) of the admission. However, antibiotics (5 days, IQR=1-10; p=0.008) and oral steroids (2.5 days, IQR=1-9.8; p=0.001) were continued for significantly longer at NHW compared with RMH (antibiotics: 3 days with IQR=1-6, and oral steroids: 2 days with IQR=0- 4). There was no difference between sites for the duration of other treatments.
Palliative approach
Documentation regarding resuscitation status (68.8%) and that the goal of care had been changed to aim for comfort (73.5%), was high. Referral to specialist palliative care (p<0.0001), documentation of goal of care (p<0.0001), documentation of resuscitation status (p<0.0001), and place of death (p<0.0001) differed significantly by site (Table 3).
Opioids were prescribed to 294 (85.7%) patients with morphine (286, 83.4%) being the most commonly prescribed opioid, however, the opioids prescribed differed significantly by site (p=0.02) (Table 4). The median initial and final doses of opioid consumed were 10mg (IQR=5- 20) and 20mg (IQR=7-45) oral morphine equivalents/day respectively, with no difference by site. The time from admission to initiation of an opioid (started on day 1, IQR=0-5 days) or benzodiazepine (started on day 2, IQR=0-7 days) or did not differ significantly between sites (Figure 4). The prescription of an opioid was associated with having a documented goal of care as palliation (p<0.0001), and prescription of a benzodiazepine was associated with having a documented goal of care as palliation (p<0.0001) and referral to specialist palliative care (p<0.0001). Hyoscine (71, 20.7%) and Glycopyrronium (19, 5.5%) were also prescribed to reduce respiratory secretions.
Patients admitted during the last four years of the 12 year study period (2012-2015) were significantly more likely to receive opioids (OR=5.1, 95%CI=1.8-14.0, p=0.002), receive
benzodiazepines (OR=8.9, 95%CI=4.1-19.4, p<0.0001), be referred to specialist palliative care (OR=5.1, 95%CI=2.7-9.4, p<0.0001), and to have their goal of care documented as palliation (OR=8.2, 95%CI=3.6-18.6, p<0.0001), when compared with patients admitted over the first 4 years of the study (2004-2007). Patients admitted in the middle four years of the study period (2008-2011), were more likely to be referred to specialist palliative care (OR=2.0, 95%CI=1.1-3.7, p=0.031) when compared with patients admitted between 2004-2007, however, other elements of palliative care were not significantly different.
Discussion
In this large Australian study, respiratory and general physicians provided a palliative approach to the majority of COPD patients during their terminal admission. However, few patients were referred to specialist palliative care and therapeutic interventions were continued for some until death. While many aspects of end-of-life care varied significantly by hospital site, the overall adoption of a palliative approach for inpatients significantly improved over time. Nonetheless there were also significant missed opportunities to offer symptom palliation and palliative care in the years before death.
Terminal admission care
In addition to receiving therapeutic treatments, most patients received opioids and benzodiazepines, which were started promptly and with doses adjusted throughout the admission. Furthermore some treatments (such as oxygen therapy and NIV) that are usually given with therapeutic intent were purposefully continued until death with the documented aim being to provide symptom palliation. Therefore the importance of palliating symptoms such as breathlessness, pain and psychological distress was understood. By comparison similar, but smaller, recent studies of COPD deaths in Taiwan (n=91) and Australia (n=34)
identified that only 11%-29% received opioids before death (28, 29), with the duration of use being very short (28).
During the admission recognition that death may occur or was imminent was high as most patients had a documented resuscitation status and their documented goal of care was changed to aim for comfort. Yet despite recognising that patients were actively dying, 12- 25% received potentially burdensome therapeutic treatments and a few continued to receive long-term preventive treatments until death. Additionally, approximately one third of patients died in ICU, HDU, or RCU and a few died in the emergency department, yet such settings are deemed not ideal (25). These results represent a significant improvement in end-of-life care compared with earlier studies from Taiwan and the United States, in which 67-70% of COPD patients died in ICU receiving multiple interventions including mechanical ventilation (29, 30). Nevertheless there remains a reluctance to cease therapeutic treatments and move solely to measures aimed at providing comfort. This hesitance may relate to rapid clinical deterioration or clinicians’ and families’ concerns about “giving up” on patients or taking away hope. Importantly, these concerns cannot be dismissed, as sometimes COPD patients do survive even when death is expected. Therefore given the challenge of accurately prognosticating in COPD, providing both non-burdensome therapeutic treatments together with a palliative approach is reasonable for some patients.
Small studies from the UK, Australia and Taiwan have demonstrated that 0-18% of COPD inpatients are referred to specialist palliative care teams during the final admission in which death occurs (28, 29, 31). By contrast, in our study 29% were referred to specialist palliative care, with this occurring more frequently at RMH, where only the specialist palliative care team is permitted to manage palliative medications delivered subcutaneously. Importantly, while referral to specialist palliative care was low, a palliative approach (including
prescribing palliative medications), was provided to most patients in this study by general or respiratory physicians. Indeed many guidelines recommend such an approach, with referral to specialist palliative care required for those patients with more complex or challenging symptoms or issues (2, 3, 6, 9). Therefore focussing only on specialist palliative care in research studies risks underestimating the quality and quantity of palliative care provided by other physicians to COPD patients. Nevertheless improved access to specialist palliative care services is required to support general physicians in providing a palliative approach, to facilitate communication regarding what palliative care is, and to enable patients with challenging symptoms or issues to access optimal care.
While individual clinicians’ practices account for some of the geographical variation in end- of-life care, many practices (including outpatient opioid prescription for breathlessness, admission to high level care, use of NIV, referral to specialist palliative care and place of death) are explained by differing hospital policies and limited access in regional Australia to high level care beds, multidisciplinary specialist, respiratory and palliative care services, and inpatient palliative care units. These gaps in service provision highlight some of the many challenges regional physicians face in geographically vast countries like Australia.
Missed opportunities
The patients in this study were older, with multiple comorbidities, severe lung disease, and significant functional limitation, with over half requiring previous admissions for a respiratory illness. Additionally severe, chronic breathlessness was documented for two thirds of RMH patients, albeit no breathlessness score was recorded. Each of these characteristics is a well-recognised clinical predictor of poor prognosis (32). Yet in the two years before death few patients accessed symptom palliation, ACP opportunities, or specialist palliative care, even though more than half the patients were attending outpatient
appointments. Therefore both the prognostic predictors and earlier opportunities to offer a palliative approach or specialist palliative care referral were missed. Notably this occurred equally at both sites, even though specialist palliative care is readily accessible at RMH. Similarly recent studies suggest that clinicians neither recognise the symptom of severe, chronic breathlessness nor the need for treatment (33), and that there is a mismatch between self-reported practice managing chronic breathlessness and actual care delivered (34). Therefore to avoid prognostic paralysis and under-treatment of distressing symptoms, referral to palliative care should be considered on the basis of individuals’ needs not clinicians’ estimates of prognosis (3, 21, 35, 36).
While improved access to specialist palliative care is needed in regional Australia (and in many countries), importantly clinician education regarding symptom palliation, communication skills and providing a palliative approach, as well as time in outpatient or