Health and Social Care in the Community (2015) 23(5), 513–522
doi: 10.1111/hsc.12162
Advance care planning for older people in Australia presenting to the emergency department from the community or residential aged care facilities Maryann Street PhD BSc (Hons) Grad Dip Drug Eval & Pharm Sci1,2, Goetz Ottmann Megan-Jane Johnstone PhD RN FACN2, Julie Considine PhD RN FACN1,2,4 and Patricia M. Livingston BA (Hons) PhD1,2,5
2,3
BA (Hons) PhD (Sociology)
,
Eastern Health – Deakin University Nursing and Midwifery Research Centre, Box Hill, Victoria, Australia, 2School of Nursing and Midwifery, Deakin University, Burwood, Victoria, Australia, 3Uniting Care Community Options, Glen Waverley, Victoria, Australia, 4Centre for Quality and Patient Safety Research, Deakin University, Burwood, Victoria, Australia and 5 Faculty of Health, Deakin University, Burwood, Victoria, Australia 1
Accepted for publication 12 September 2014
Correspondence Maryann Street Eastern Health – Deakin University Nursing and Midwifery Research Centre School of Nursing and Midwifery, Deakin University Level 2, 5 Arnold St, Box Hill, Victoria 3128, Australia. E-mail: [email protected]
What is known about this topic
• • •
Advance care planning (ACP) provides people with an opportunity to be involved in decision-making underpinning their care and treatment options, should they be unable to do so at a later stage. Within Australia, ACP uptake in residential aged care settings has been increasing. However, the prevalence of ACP for older people in the Australian community remains unknown. When an older person is transferred to the Emergency Department, health professionals seek to consider any advance directives for treatment options.
What this paper adds
•
•
Low prevalence (13.3%) of ACP for older people attending the Emergency Department. ACP was more common for those from residential aged care with a comorbidity of cerebrovascular disease or dementia, compared to those living in the community and without cognitive impairment. Advance care planning may be associated with shorter lengths of hospital stay and a lower rate of readmission to hospital and may influence requests both for and against medical intervention.
© 2014 John Wiley & Sons Ltd
Abstract The purpose of this retrospective, cross-sectional study was to determine the prevalence of advance care planning (ACP) among older people presenting to an Emergency Department (ED) from the community or a residential aged care facility. The study sample comprised 300 older people (aged 65+ years) presenting to three Victorian EDs in 2011. A total of 150 patients transferred from residential aged care to ED were randomly selected and then matched to 150 people who lived in the community and attended the ED by age, gender, reason for ED attendance and triage category on arrival. Overall prevalence of ACP was 13.3% (n = 40/300); over one-quarter (26.6%, n = 40/150) of those presenting to the ED from residential aged care had a documented Advance Care Plan, compared to none (0%, n = 0/150) of the people from the community. There were no significant differences in the median ED length of stay, number of investigations and interventions undertaken in ED, time seen by a doctor or rate of hospital admission for those with an Advance Care Plan compared to those without. Those with a comorbidity of cerebrovascular disease or dementia and those assessed with impaired brain function were more likely to have a documented Advance Care Plan on arrival at ED. Length of hospital stay was shorter for those with an Advance Care Plan [median (IQR) = 3 days (2–6) vs. 6 days (2–10), P = 0.027] and readmission lower (0% vs. 13.7%). In conclusion, older people from the community transferred to ED were unlikely to have a documented Advance Care Plan. Those from residential aged care who were cognitively impaired more frequently had an Advance Care Plan. In the ED, decisions of care did not appear to be influenced by the presence or absence of Advance Care Plans, but length of hospital admission was shorter for those with an Advance Care Plan. Keywords: advance care planning, advance directives (medical care), domiciliary aged care, emergency departments, end-of-life care, residential aged care
513
M. Street et al.
Introduction Advance care planning (ACP) is regarded as a corner stone of end-of-life care (Piers et al. 2013). The key emphasis of ACP is to improve communication processes and assist patients and their loved ones prepare for decisions that often have to be made in uncertain circumstances when approaching the end stage of life (Perkins 2007). The process of ACP provides patients with an opportunity to be involved in decision-making underpinning their care and treatment options. This may include nominating a substitute decision-maker who can decide about options for medical treatment on their behalf should they be unable to do so. In Australia, the right to make an advance care directive, the ‘formal recording of an advance care plan’ (Australian Health Ministers Advisory Group 2011, p. 10), has statutory recognition in most Australian states and territories (Australian Health Ministers Advisory Group 2011, p. 5). Advance care directives recognise the rights of patients to refuse or request medical treatment, although the latter are not as commonly included. With ACP, there is potential to improve patient and family satisfaction with end-of-life care, and reduce the stress, anxiety and depression that can sometimes occur in surviving relatives who have had to deal with the uncertainties of how to decide what care and treatment options are best for their loved ones as their lives come to an end (Detering et al. 2010). Aspects of ACP may include one or all of the following: appointment of an Enduring Power of Attorney (Medical); documentation of person’s wishes regarding specific medical treatment; refusal of treatment certificate relating to a current medical condition; and discussion with family and significant others, including the doctor about the person’s wishes. Under the Medical Treatment Act (1988) (State Government of Victoria 2012), a competent adult may refuse treatment now, and in the future, for a specific condition, even when it is predicted that they will eventually become incompetent. However, this limitation of medical treatment order, frequently containing a ‘not-for-cardiopulmonary resuscitation’ order, only relates to the current medical condition and therefore must usually be reconfirmed on each admission (State Government of Victoria 2014). The Act also allows a competent adult to appoint an agent under an enduring power of attorney (medical treatment), who is empowered to make decisions about medical treatment on behalf of the person if they become incompetent and can complete a Refusal of Treatment Certificate on behalf of that person. 514
Towards the end of the 1990s, reports on ACP in Australia gave rise for concern. An Australian population-based study published by Nair et al. (2000) reported that out of a population of 2764 older people in residential aged care facilities in the Hunter Valley (72% response rate), only 0.2% of residents had a formal Advance Care Directive, 1.1% had a not-for-resuscitation order documented in their medical records, 5.6% had a formal guardian and 2.8% had an enduring guardian (Nair et al. 2000). Research in South Australian facilities suggested that ACP documentation in residential aged care was haphazard and unsystematic (Brown et al. 2005, Shanley et al. 2009) and that ACP discussions often occurred very late during a person’s illness trajectory (Shanley et al. 2009). These and similar reports prompted a number of initiatives aimed at increasing ACP uptake in Australian residential aged care settings (Blackford et al. 2007, Clayton et al. 2007, Jeong et al. 2007, 2010, Lyon 2007, Fernandes 2008). It has been suggested that to increase uptake and adherence to Advance Care Directives in Australia, focus on the whole health system and inclusion of the family is required (Blackford et al. 2007, Seal 2010). Even so, debate remains surrounding lack of training and education for staff members, levels of compliance with ACP provisions, inadequate consultation with families and negative attitudes towards ACP by older people and their family members (Blackford et al. 2007, Fernandes 2008, Shanley et al. 2009, Jeong et al. 2011). The difficulties of implementing ACP in residential care in regional and rural settings have also been highlighted (Mitchell et al. 2011). Johnstone and Kanitsaki (2009) have emphasised the importance of cross-cultural considerations in end-of-life care decisions and planning, and why certain ethnic minority groups may be wary of ACP. Considerations in developing a culturally appropriate ACP approach for Aboriginal communities have similarly been highlighted (Sinclair et al. 2014). Meanwhile, specific ACP needs of gay, lesbian, bisexual and transgender elders have been identified (Cartwright et al. 2012, Rawlings 2012). Two issues emerge from the literature on ACP in residential and community aged care published to date. First, Nair’s foundational study conducted over 14 years ago in 2000 has remained the only large population-based study focusing on ACP in residential aged care in Australia; and second, while ACP in residential aged care settings has received growing attention over the last decade, we were unable to identify any peer-reviewed publication exploring the uptake of ACPs within a community aged care context. To address these issues, we undertook a study to determine the prevalence of ACP among older © 2014 John Wiley & Sons Ltd
ACP among older people presenting to an ED
Australians presenting to ED from the community or residential aged care facilities.
Methods Study design This study was a retrospective cross-sectional matched cohort study of emergency presentations comprising two cohorts of people who presented to any of three EDs at Eastern Health, Victoria, Australia. A mixed methods approach (Creswell 2014) using both quantitative and qualitative analysis was used to examine the uptake of ACP by older people and explore the deeper context of ACP adherence. The quantitative and qualitative data were collected concurrently with priority given to analysis of the quantitative data. Setting and sample Eastern Health provides more than 800,000 episodes of patient care each year across 25 sites and has the largest geographical catchment area of any metropolitan health service across the state of Victoria (Eastern Health 2013). The study sample was 300 patients aged 65 years or older, and who presented to any of three EDs at Eastern Health between 1 July and 31 December 2011. During this period, there were 2051 presentations to the ED from residential aged care facilities and 150 were randomly selected using the random number function in Microsoft Excel 2007 (Microsoft Pty Ltd., Seattle, WA, USA). A total of 150 older people attending the ED from the community (from a pool of 14,072) were matched to the patients from residential aged care by age, gender, presenting problem and clinical urgency on arrival (triage category). Those who presented to the ED with a documented ACP were compared to those who did not have an ACP. Data collection Data were collected from the medical record and health service information systems and included: 1 patient characteristics: age, gender, country of
birth, living arrangements, diagnostic category, comorbidities, limitation of medical treatment orders, details of any ACP; 2 transfer data: day and time of ED arrival, mode of arrival, Australasian Triage Scale (ATS) category, reason for presentation, whether accompanied by another adult to the ED, Glasgow Coma Score; 3 outcome data: investigations and interventions conducted in ED (radiology, pathology, urinalysis, cardiac monitoring, oxygen, IV fluids and © 2014 John Wiley & Sons Ltd
medications), ED discharge destination (return to place of residence, hospital admission), ED length of stay, hospital length of stay, in-hospital serious adverse events (unplanned ICU admission, Medical Emergency Team call, cardiac arrest), ED re-presentation within 30 days of index visit, readmission to hospital within 48 hours or 30 days of index visit and in-hospital mortality. Documentation of ACP included in the medical records varied in both style and content. However, they commonly included some prescribed directives requiring a Yes/No response and some free text items for the person to detail preferences. The content of Advance Care Directives was transcribed from the medical record to the data collection form and included in the data set for analysis. Comorbidity status, an indication of higher burden of disease, was calculated using the Charlson index (Charlson et al. 1987). Comorbidities were grouped based on ICD-10-AM codes (National Centre for Classification in Health (NCCH) 1998, Quan et al. 2005), and a score calculated using weightings which varied from 1 (e.g. acute myocardial infarction, peripheral vascular disease, cerebral vascular disease, chronic pulmonary disease) to 6 (e.g. metastatic solid tumour, HIV/AIDS) (Frost et al. 2009). The aim of this study was to determine the prevalence of ACP among older Australians presenting to EDs from the community or residential aged care facilities. The primary outcomes were length of stay in ED, investigations and treatment given in ED, and discharge destination from ED. Secondary outcomes were length of hospital admission and re-presentation to ED or readmission to hospital. Rigour Rigour was ensured first by matching the randomly selected cohort of people from residential aged care facilities to a cohort of people from the community by age, gender, reason for presentation and clinical urgency on arrival at the ED. In this way, several potential confounders were addressed. Second, the data collected from the medical record were relevant to the research question of ACP uptake and adherence. Data included in the data set for analysis were validated and any inconsistencies addressed by reviewing the original data. Appropriate statistical comparisons were applied. Statistical analysis The quantitative study data were summarised using descriptive statistics [means, standard deviations 515
M. Street et al.
(SD), medians and interquartile ranges (IQR)]. Between-group comparisons were made using chisquare tests for categorical data and independent t-tests or Mann–Whitney U-test for continuous data. The level of significance was set as P ≤ 0.05. Data were analysed using SPSS version 22 (IBM SPSS Statistics, Inc., Chicago, IL, USA). Qualitative data analysis was conducted using a five-stage approach: (i) familiarisation; (ii) identifying a thematic framework; (iii) indexing; (iv) charting; (v) mapping and interpretation (Ritchie & Lewis 2003).
Analysis of the whole cohort showed the median ED length of stay was 6.5 hours (IQR = 4.8–8.2) and median length of stay for those admitted to hospital (n = 213) was 5 days (IQR = 2–9). One in four (26.4%; n = 79) had a visit to the ED within 30 days (before or after) of the index ED visit included in this study. Only 85 people (28.3%) were discharged from ED; 213 (71.0%) were admitted to hospital and two people died in ED. The inpatient mortality rate was 13/213 (6.1%). Advance care planning
Research ethics This study was approved by the Human Research Ethics Committees of Deakin University and Eastern Health and it conforms to the provisions of the Declaration of Helsinki.
Results Demographic and clinical characteristics of the 300 people included in this cohort study are shown in Table 1. The mean age was 86.2 years (SD = 6.5) and 67.7% were female. Of the 286 people whose country of birth was recorded, the majority were born in Australia or New Zealand (66%, n = 188), with 16% (n = 46) from the United Kingdom, 15% (n = 43) born in Europe and a small proportion born in Asia (2.4%, n = 7). The most common reasons for presentation were respiratory (n = 57, 19%), neurological (n = 52, 17%), a fall (n = 42, 14%), trauma/wound (n = 42, 14%) or gastrointestinal problem (n = 28, 9%). The majority of patients (n = 140, 47%) were triaged as category 3 using the ATS, so could safely wait 30 minutes before being seen. A further 39% of patients (n = 117) were triaged to ATS category 4, requiring care within 60 minutes and only one person was assessed as safe to wait for 2 hours for emergency care (ATS = 5). Differences were evident in the mode of arrival. All those transferred from residential aged care arrived by ambulance; while 68.6% (n = 103) of those from the community arrived by ambulance, the remainder (31%, n = 47) arrived by a private vehicle. While all those from residential aged care lived with other adults, only half of those from the community were living with other adults (56%, n = 84) and 44% from the community (n = 66) were living alone. Supported living arrangements for the community participants were living with aged care assistance in their own home (13.3%, n = 20), living with a carer (11.3%, n = 17) and in assisted living arrangements (3.3%, n = 5), leaving 72.1% (n = 108) living without any formalised aged care assistance. 516
The overall prevalence of ACP was 13.3% (n = 40/ 300); over one-quarter (26.6%, n = 40/150) of those presenting to the ED from residential aged care had an Advance Care Directive with them on ED arrival, compared to none (0%, n = 0/150) of the people from the community. In the medical record, it was recorded that one other person from residential aged care had an Advance Care Directive, but did not bring it with them and one person from the community had an enduring power of attorney (medical) for the next of kin. Those having Advance Care Directives were less likely to be accompanied to the ED by another adult. Limitation of medical treatment orders was associated with ACP, such that 23/40 (57.5%) people with an Advance Care Directive also had a limitation of medical treatment order, while only 37/260 (14.2%) people without an Advance Care Directive also had a limitation of medical treatment order in place (Table 1). Higher Charlson comorbidity score was associated with ACP. Those with a comorbidity of cancer (malignancy) were no more likely to have ACP (n = 3/25; 12%) compared to those without this comorbidity (n = 37/275; 13.4%) and none of the seven people with metastatic cancer had an Advance Care Directive. However, those with a comorbidity of cerebrovascular disease or dementia were more likely to have an Advance Care Directive (Table 1). The median ED length of stay for those with ACP was 6.3 hours (IQR = 4.3–7.9) which was not significantly different from those without ACP of 6.7 hours (IQR = 4.8–8.5; P = 0.424). There was no association between day (weekday or weekend) or time of arrival (in-hours vs. out-of-hours) at the ED and having ACP. Also, the time from arrival at the ED to being seen by a doctor was no different for those with ACP (mean = 65.8 minutes, SD = 98.4) compared to those without ACP, 57.1 minutes, SD = 63.4; t(285) = 0.736, P = 0.398. There was no significant difference in the clinical urgency on arrival in ED or the number of investigations or interventions performed for those © 2014 John Wiley & Sons Ltd
ACP among older people presenting to an ED
Table 1 Demographic and clinical characteristics of those arriving at the Emergency Department (ED) with a documented Advanced Care Plan (ACP) compared to those who did not have an ACP on arrival at ED Advanced care plan
Lived in Community RACF Accompanied by another adult (n = 279) No Yes LOMT orders None NFR Charlson comorbidity score Low (0–1) Medium (2–3) High (>4) Comorbidities Malignancy Cerebrovascular disease Dementia Mode of arrival (n = 284) Ambulance Private vehicle
Total sample
Yes
No
N = 300
N = 40
N = 260
N
%
N
%
N
%
150 150
50.0 50.0
0 40
0 100
150 110
57.7 42.3
v2(1) = 46.154, P < 0.001 v2(1) = 13.077, P < 0.001 174 105
62.4 37.6
33 4
89.2 10.8
141 101
58.3 41.7
240 60
80.0 20.0
18 22
45.0 55.0
222 38
84.2 15.8
v2(1) = 35.337, P < 0.001 v2(2) = 16.038, P < 0.001 146 105 49
48.7 35.0 16.3
8 20 12
20.0 50.0 30.0
138 85 37
53.1 32.7 14.2
25 72 75
8.3 24.0 25.0
3 15 22
7.5 37.5 55.0
22 57 53
8.5 21.9 20.4
237 47
83.4 16.6
38 0
100 0
199 47
80.9 19.1
v2(1) v2(1) v2(1) v2(1)
= = = =
0.042, P = 0.838 4.612, P = 0.032 22.154, P < 0.001 8.700, P = 0.003
LOMT, limitation of medical treatment; NFR, not for resuscitation; RACF, Residential Aged Care Facility.
having ACP compared to those without (Table 2). However, proportionally more people with ACP had a Glasgow Coma Scale score less than 15 compared to those without ACP (Table 2). The rate of hospital admission was high and not significantly different for those with (79.5%, n = 31/ 39) or without ACP (70.3%, n = 182/259; Table 2). Median length of hospital stay for those with ACP was half (3 days) of that for those without ACP (6 days; Table 2). During hospital admission, a Medical Emergency Team call was made for 3/31 (9.7%) people with ACP, compared to 4/182 (2.2%) for those without ACP, v2(1) = 4.662, P = 0.031. Four people were admitted to ICU; however, no one with ACP was admitted to ICU. Although not achieving significance, the rate of re-presentation to the ED within 30 days before or after the index visit was lower for those with ACP (17.9%, n = 7/39 vs. 27.9%, n = 72/258; Table 2). The rate of readmission to hospital within 30 days of the index admission for those with ACP was significantly lower than for those without ACP (0%, n = 0/31, compared to 13.7%, n = 25/182; Table 2). The in-hospital mortality rate was 9.7% (n = 3/31) for those with ACP compared to 4.4% (n = 8/182) for © 2014 John Wiley & Sons Ltd
those without ACP; however, this did not reach statistical significance, v2(1) = 1.509, P = 0.219. Directives specified within the advance care plans Qualitative analysis of the 40 Advance Care Directives in this study revealed four themes: request for medical intervention, directives against medical intervention, consultation with family members and pre-/ post-mortem preferences. While 25 of the 40 Advance Care Directives included instructions for medical intervention, 16 of these also included directives against medical treatment. The data revealed that people’s knowledge and understanding of ACP appeared limited regarding the meaning, policies and practice of ACP. Request for medical intervention Twenty-five of the forty Advance Care Directives requested medical intervention in the event of clinical deterioration for example: For resuscitation (RAC003) Hospitalisation if necessary (RAC004)
517
M. Street et al.
Transport by ambulance (RAC013) Transfer to acute care hospital (RAC023) Go to hospital if deterioration in health (RAC032) Yes to investigations and diagnosis, Yes to surgical and medical procedures, Yes to pain and symptom relief (RAC048) All attempts to be made for resuscitation (RAC056) [I] wish to have antibiotic therapy and pain-relieving medications (RAC123)
Other requests centred on end-of-life care such as pain management and symptom control. Some people wanted to return to the residential aged care facility and receive palliative care.
If hospitalised, to go back to the Nursing Home as soon as possible for palliative care (RAC121) Pain relief only (RAC148) Palliative care in the event of gradual decline, to be kept comfortable (RAC222)
However, often these instructions were in the same Advance Care Directive besides requests for medical intervention and at times leaving the decision to the clinician or a family member as to what was best for the person at the time. For example: Would like cardiopulmonary resuscitation while medically appropriate and in best interests (RAC259) Wishes to have what is necessary without life support. Cardiopulmonary resuscitation only if medically beneficial (RAC204)
[I want] to be kept comfortable and pain free (RAC028) Table 2 Patient assessment and outcomes
Advanced care plan Total sample
Yes
No
N = 300
N = 40
N = 260
N
%
N
7 33
Length of admission (days) Admitted patients
N = 31
Median (IQR) 5 (2–9) Readmission to hospital within 30 days of index visit Yes 25 11.7 No 188 88.3
N
% v2(2) = 4.941, P = 0.085*
Triage category ATS = 1 3 1.0 ATS = 2 39 13.0 ATS = 3 140 46.7 ATS = 4 117 39.0 ATS = 5 1 0.3 Glasgow Coma Scale (GCS) (n = 289) 13–15 270 93.4 9–12 16 5.5
doi: 10.1111/hsc.12162
Advance care planning for older people in Australia presenting to the emergency department from the community or residential aged care facilities Maryann Street PhD BSc (Hons) Grad Dip Drug Eval & Pharm Sci1,2, Goetz Ottmann Megan-Jane Johnstone PhD RN FACN2, Julie Considine PhD RN FACN1,2,4 and Patricia M. Livingston BA (Hons) PhD1,2,5
2,3
BA (Hons) PhD (Sociology)
,
Eastern Health – Deakin University Nursing and Midwifery Research Centre, Box Hill, Victoria, Australia, 2School of Nursing and Midwifery, Deakin University, Burwood, Victoria, Australia, 3Uniting Care Community Options, Glen Waverley, Victoria, Australia, 4Centre for Quality and Patient Safety Research, Deakin University, Burwood, Victoria, Australia and 5 Faculty of Health, Deakin University, Burwood, Victoria, Australia 1
Accepted for publication 12 September 2014
Correspondence Maryann Street Eastern Health – Deakin University Nursing and Midwifery Research Centre School of Nursing and Midwifery, Deakin University Level 2, 5 Arnold St, Box Hill, Victoria 3128, Australia. E-mail: [email protected]
What is known about this topic
• • •
Advance care planning (ACP) provides people with an opportunity to be involved in decision-making underpinning their care and treatment options, should they be unable to do so at a later stage. Within Australia, ACP uptake in residential aged care settings has been increasing. However, the prevalence of ACP for older people in the Australian community remains unknown. When an older person is transferred to the Emergency Department, health professionals seek to consider any advance directives for treatment options.
What this paper adds
•
•
Low prevalence (13.3%) of ACP for older people attending the Emergency Department. ACP was more common for those from residential aged care with a comorbidity of cerebrovascular disease or dementia, compared to those living in the community and without cognitive impairment. Advance care planning may be associated with shorter lengths of hospital stay and a lower rate of readmission to hospital and may influence requests both for and against medical intervention.
© 2014 John Wiley & Sons Ltd
Abstract The purpose of this retrospective, cross-sectional study was to determine the prevalence of advance care planning (ACP) among older people presenting to an Emergency Department (ED) from the community or a residential aged care facility. The study sample comprised 300 older people (aged 65+ years) presenting to three Victorian EDs in 2011. A total of 150 patients transferred from residential aged care to ED were randomly selected and then matched to 150 people who lived in the community and attended the ED by age, gender, reason for ED attendance and triage category on arrival. Overall prevalence of ACP was 13.3% (n = 40/300); over one-quarter (26.6%, n = 40/150) of those presenting to the ED from residential aged care had a documented Advance Care Plan, compared to none (0%, n = 0/150) of the people from the community. There were no significant differences in the median ED length of stay, number of investigations and interventions undertaken in ED, time seen by a doctor or rate of hospital admission for those with an Advance Care Plan compared to those without. Those with a comorbidity of cerebrovascular disease or dementia and those assessed with impaired brain function were more likely to have a documented Advance Care Plan on arrival at ED. Length of hospital stay was shorter for those with an Advance Care Plan [median (IQR) = 3 days (2–6) vs. 6 days (2–10), P = 0.027] and readmission lower (0% vs. 13.7%). In conclusion, older people from the community transferred to ED were unlikely to have a documented Advance Care Plan. Those from residential aged care who were cognitively impaired more frequently had an Advance Care Plan. In the ED, decisions of care did not appear to be influenced by the presence or absence of Advance Care Plans, but length of hospital admission was shorter for those with an Advance Care Plan. Keywords: advance care planning, advance directives (medical care), domiciliary aged care, emergency departments, end-of-life care, residential aged care
513
M. Street et al.
Introduction Advance care planning (ACP) is regarded as a corner stone of end-of-life care (Piers et al. 2013). The key emphasis of ACP is to improve communication processes and assist patients and their loved ones prepare for decisions that often have to be made in uncertain circumstances when approaching the end stage of life (Perkins 2007). The process of ACP provides patients with an opportunity to be involved in decision-making underpinning their care and treatment options. This may include nominating a substitute decision-maker who can decide about options for medical treatment on their behalf should they be unable to do so. In Australia, the right to make an advance care directive, the ‘formal recording of an advance care plan’ (Australian Health Ministers Advisory Group 2011, p. 10), has statutory recognition in most Australian states and territories (Australian Health Ministers Advisory Group 2011, p. 5). Advance care directives recognise the rights of patients to refuse or request medical treatment, although the latter are not as commonly included. With ACP, there is potential to improve patient and family satisfaction with end-of-life care, and reduce the stress, anxiety and depression that can sometimes occur in surviving relatives who have had to deal with the uncertainties of how to decide what care and treatment options are best for their loved ones as their lives come to an end (Detering et al. 2010). Aspects of ACP may include one or all of the following: appointment of an Enduring Power of Attorney (Medical); documentation of person’s wishes regarding specific medical treatment; refusal of treatment certificate relating to a current medical condition; and discussion with family and significant others, including the doctor about the person’s wishes. Under the Medical Treatment Act (1988) (State Government of Victoria 2012), a competent adult may refuse treatment now, and in the future, for a specific condition, even when it is predicted that they will eventually become incompetent. However, this limitation of medical treatment order, frequently containing a ‘not-for-cardiopulmonary resuscitation’ order, only relates to the current medical condition and therefore must usually be reconfirmed on each admission (State Government of Victoria 2014). The Act also allows a competent adult to appoint an agent under an enduring power of attorney (medical treatment), who is empowered to make decisions about medical treatment on behalf of the person if they become incompetent and can complete a Refusal of Treatment Certificate on behalf of that person. 514
Towards the end of the 1990s, reports on ACP in Australia gave rise for concern. An Australian population-based study published by Nair et al. (2000) reported that out of a population of 2764 older people in residential aged care facilities in the Hunter Valley (72% response rate), only 0.2% of residents had a formal Advance Care Directive, 1.1% had a not-for-resuscitation order documented in their medical records, 5.6% had a formal guardian and 2.8% had an enduring guardian (Nair et al. 2000). Research in South Australian facilities suggested that ACP documentation in residential aged care was haphazard and unsystematic (Brown et al. 2005, Shanley et al. 2009) and that ACP discussions often occurred very late during a person’s illness trajectory (Shanley et al. 2009). These and similar reports prompted a number of initiatives aimed at increasing ACP uptake in Australian residential aged care settings (Blackford et al. 2007, Clayton et al. 2007, Jeong et al. 2007, 2010, Lyon 2007, Fernandes 2008). It has been suggested that to increase uptake and adherence to Advance Care Directives in Australia, focus on the whole health system and inclusion of the family is required (Blackford et al. 2007, Seal 2010). Even so, debate remains surrounding lack of training and education for staff members, levels of compliance with ACP provisions, inadequate consultation with families and negative attitudes towards ACP by older people and their family members (Blackford et al. 2007, Fernandes 2008, Shanley et al. 2009, Jeong et al. 2011). The difficulties of implementing ACP in residential care in regional and rural settings have also been highlighted (Mitchell et al. 2011). Johnstone and Kanitsaki (2009) have emphasised the importance of cross-cultural considerations in end-of-life care decisions and planning, and why certain ethnic minority groups may be wary of ACP. Considerations in developing a culturally appropriate ACP approach for Aboriginal communities have similarly been highlighted (Sinclair et al. 2014). Meanwhile, specific ACP needs of gay, lesbian, bisexual and transgender elders have been identified (Cartwright et al. 2012, Rawlings 2012). Two issues emerge from the literature on ACP in residential and community aged care published to date. First, Nair’s foundational study conducted over 14 years ago in 2000 has remained the only large population-based study focusing on ACP in residential aged care in Australia; and second, while ACP in residential aged care settings has received growing attention over the last decade, we were unable to identify any peer-reviewed publication exploring the uptake of ACPs within a community aged care context. To address these issues, we undertook a study to determine the prevalence of ACP among older © 2014 John Wiley & Sons Ltd
ACP among older people presenting to an ED
Australians presenting to ED from the community or residential aged care facilities.
Methods Study design This study was a retrospective cross-sectional matched cohort study of emergency presentations comprising two cohorts of people who presented to any of three EDs at Eastern Health, Victoria, Australia. A mixed methods approach (Creswell 2014) using both quantitative and qualitative analysis was used to examine the uptake of ACP by older people and explore the deeper context of ACP adherence. The quantitative and qualitative data were collected concurrently with priority given to analysis of the quantitative data. Setting and sample Eastern Health provides more than 800,000 episodes of patient care each year across 25 sites and has the largest geographical catchment area of any metropolitan health service across the state of Victoria (Eastern Health 2013). The study sample was 300 patients aged 65 years or older, and who presented to any of three EDs at Eastern Health between 1 July and 31 December 2011. During this period, there were 2051 presentations to the ED from residential aged care facilities and 150 were randomly selected using the random number function in Microsoft Excel 2007 (Microsoft Pty Ltd., Seattle, WA, USA). A total of 150 older people attending the ED from the community (from a pool of 14,072) were matched to the patients from residential aged care by age, gender, presenting problem and clinical urgency on arrival (triage category). Those who presented to the ED with a documented ACP were compared to those who did not have an ACP. Data collection Data were collected from the medical record and health service information systems and included: 1 patient characteristics: age, gender, country of
birth, living arrangements, diagnostic category, comorbidities, limitation of medical treatment orders, details of any ACP; 2 transfer data: day and time of ED arrival, mode of arrival, Australasian Triage Scale (ATS) category, reason for presentation, whether accompanied by another adult to the ED, Glasgow Coma Score; 3 outcome data: investigations and interventions conducted in ED (radiology, pathology, urinalysis, cardiac monitoring, oxygen, IV fluids and © 2014 John Wiley & Sons Ltd
medications), ED discharge destination (return to place of residence, hospital admission), ED length of stay, hospital length of stay, in-hospital serious adverse events (unplanned ICU admission, Medical Emergency Team call, cardiac arrest), ED re-presentation within 30 days of index visit, readmission to hospital within 48 hours or 30 days of index visit and in-hospital mortality. Documentation of ACP included in the medical records varied in both style and content. However, they commonly included some prescribed directives requiring a Yes/No response and some free text items for the person to detail preferences. The content of Advance Care Directives was transcribed from the medical record to the data collection form and included in the data set for analysis. Comorbidity status, an indication of higher burden of disease, was calculated using the Charlson index (Charlson et al. 1987). Comorbidities were grouped based on ICD-10-AM codes (National Centre for Classification in Health (NCCH) 1998, Quan et al. 2005), and a score calculated using weightings which varied from 1 (e.g. acute myocardial infarction, peripheral vascular disease, cerebral vascular disease, chronic pulmonary disease) to 6 (e.g. metastatic solid tumour, HIV/AIDS) (Frost et al. 2009). The aim of this study was to determine the prevalence of ACP among older Australians presenting to EDs from the community or residential aged care facilities. The primary outcomes were length of stay in ED, investigations and treatment given in ED, and discharge destination from ED. Secondary outcomes were length of hospital admission and re-presentation to ED or readmission to hospital. Rigour Rigour was ensured first by matching the randomly selected cohort of people from residential aged care facilities to a cohort of people from the community by age, gender, reason for presentation and clinical urgency on arrival at the ED. In this way, several potential confounders were addressed. Second, the data collected from the medical record were relevant to the research question of ACP uptake and adherence. Data included in the data set for analysis were validated and any inconsistencies addressed by reviewing the original data. Appropriate statistical comparisons were applied. Statistical analysis The quantitative study data were summarised using descriptive statistics [means, standard deviations 515
M. Street et al.
(SD), medians and interquartile ranges (IQR)]. Between-group comparisons were made using chisquare tests for categorical data and independent t-tests or Mann–Whitney U-test for continuous data. The level of significance was set as P ≤ 0.05. Data were analysed using SPSS version 22 (IBM SPSS Statistics, Inc., Chicago, IL, USA). Qualitative data analysis was conducted using a five-stage approach: (i) familiarisation; (ii) identifying a thematic framework; (iii) indexing; (iv) charting; (v) mapping and interpretation (Ritchie & Lewis 2003).
Analysis of the whole cohort showed the median ED length of stay was 6.5 hours (IQR = 4.8–8.2) and median length of stay for those admitted to hospital (n = 213) was 5 days (IQR = 2–9). One in four (26.4%; n = 79) had a visit to the ED within 30 days (before or after) of the index ED visit included in this study. Only 85 people (28.3%) were discharged from ED; 213 (71.0%) were admitted to hospital and two people died in ED. The inpatient mortality rate was 13/213 (6.1%). Advance care planning
Research ethics This study was approved by the Human Research Ethics Committees of Deakin University and Eastern Health and it conforms to the provisions of the Declaration of Helsinki.
Results Demographic and clinical characteristics of the 300 people included in this cohort study are shown in Table 1. The mean age was 86.2 years (SD = 6.5) and 67.7% were female. Of the 286 people whose country of birth was recorded, the majority were born in Australia or New Zealand (66%, n = 188), with 16% (n = 46) from the United Kingdom, 15% (n = 43) born in Europe and a small proportion born in Asia (2.4%, n = 7). The most common reasons for presentation were respiratory (n = 57, 19%), neurological (n = 52, 17%), a fall (n = 42, 14%), trauma/wound (n = 42, 14%) or gastrointestinal problem (n = 28, 9%). The majority of patients (n = 140, 47%) were triaged as category 3 using the ATS, so could safely wait 30 minutes before being seen. A further 39% of patients (n = 117) were triaged to ATS category 4, requiring care within 60 minutes and only one person was assessed as safe to wait for 2 hours for emergency care (ATS = 5). Differences were evident in the mode of arrival. All those transferred from residential aged care arrived by ambulance; while 68.6% (n = 103) of those from the community arrived by ambulance, the remainder (31%, n = 47) arrived by a private vehicle. While all those from residential aged care lived with other adults, only half of those from the community were living with other adults (56%, n = 84) and 44% from the community (n = 66) were living alone. Supported living arrangements for the community participants were living with aged care assistance in their own home (13.3%, n = 20), living with a carer (11.3%, n = 17) and in assisted living arrangements (3.3%, n = 5), leaving 72.1% (n = 108) living without any formalised aged care assistance. 516
The overall prevalence of ACP was 13.3% (n = 40/ 300); over one-quarter (26.6%, n = 40/150) of those presenting to the ED from residential aged care had an Advance Care Directive with them on ED arrival, compared to none (0%, n = 0/150) of the people from the community. In the medical record, it was recorded that one other person from residential aged care had an Advance Care Directive, but did not bring it with them and one person from the community had an enduring power of attorney (medical) for the next of kin. Those having Advance Care Directives were less likely to be accompanied to the ED by another adult. Limitation of medical treatment orders was associated with ACP, such that 23/40 (57.5%) people with an Advance Care Directive also had a limitation of medical treatment order, while only 37/260 (14.2%) people without an Advance Care Directive also had a limitation of medical treatment order in place (Table 1). Higher Charlson comorbidity score was associated with ACP. Those with a comorbidity of cancer (malignancy) were no more likely to have ACP (n = 3/25; 12%) compared to those without this comorbidity (n = 37/275; 13.4%) and none of the seven people with metastatic cancer had an Advance Care Directive. However, those with a comorbidity of cerebrovascular disease or dementia were more likely to have an Advance Care Directive (Table 1). The median ED length of stay for those with ACP was 6.3 hours (IQR = 4.3–7.9) which was not significantly different from those without ACP of 6.7 hours (IQR = 4.8–8.5; P = 0.424). There was no association between day (weekday or weekend) or time of arrival (in-hours vs. out-of-hours) at the ED and having ACP. Also, the time from arrival at the ED to being seen by a doctor was no different for those with ACP (mean = 65.8 minutes, SD = 98.4) compared to those without ACP, 57.1 minutes, SD = 63.4; t(285) = 0.736, P = 0.398. There was no significant difference in the clinical urgency on arrival in ED or the number of investigations or interventions performed for those © 2014 John Wiley & Sons Ltd
ACP among older people presenting to an ED
Table 1 Demographic and clinical characteristics of those arriving at the Emergency Department (ED) with a documented Advanced Care Plan (ACP) compared to those who did not have an ACP on arrival at ED Advanced care plan
Lived in Community RACF Accompanied by another adult (n = 279) No Yes LOMT orders None NFR Charlson comorbidity score Low (0–1) Medium (2–3) High (>4) Comorbidities Malignancy Cerebrovascular disease Dementia Mode of arrival (n = 284) Ambulance Private vehicle
Total sample
Yes
No
N = 300
N = 40
N = 260
N
%
N
%
N
%
150 150
50.0 50.0
0 40
0 100
150 110
57.7 42.3
v2(1) = 46.154, P < 0.001 v2(1) = 13.077, P < 0.001 174 105
62.4 37.6
33 4
89.2 10.8
141 101
58.3 41.7
240 60
80.0 20.0
18 22
45.0 55.0
222 38
84.2 15.8
v2(1) = 35.337, P < 0.001 v2(2) = 16.038, P < 0.001 146 105 49
48.7 35.0 16.3
8 20 12
20.0 50.0 30.0
138 85 37
53.1 32.7 14.2
25 72 75
8.3 24.0 25.0
3 15 22
7.5 37.5 55.0
22 57 53
8.5 21.9 20.4
237 47
83.4 16.6
38 0
100 0
199 47
80.9 19.1
v2(1) v2(1) v2(1) v2(1)
= = = =
0.042, P = 0.838 4.612, P = 0.032 22.154, P < 0.001 8.700, P = 0.003
LOMT, limitation of medical treatment; NFR, not for resuscitation; RACF, Residential Aged Care Facility.
having ACP compared to those without (Table 2). However, proportionally more people with ACP had a Glasgow Coma Scale score less than 15 compared to those without ACP (Table 2). The rate of hospital admission was high and not significantly different for those with (79.5%, n = 31/ 39) or without ACP (70.3%, n = 182/259; Table 2). Median length of hospital stay for those with ACP was half (3 days) of that for those without ACP (6 days; Table 2). During hospital admission, a Medical Emergency Team call was made for 3/31 (9.7%) people with ACP, compared to 4/182 (2.2%) for those without ACP, v2(1) = 4.662, P = 0.031. Four people were admitted to ICU; however, no one with ACP was admitted to ICU. Although not achieving significance, the rate of re-presentation to the ED within 30 days before or after the index visit was lower for those with ACP (17.9%, n = 7/39 vs. 27.9%, n = 72/258; Table 2). The rate of readmission to hospital within 30 days of the index admission for those with ACP was significantly lower than for those without ACP (0%, n = 0/31, compared to 13.7%, n = 25/182; Table 2). The in-hospital mortality rate was 9.7% (n = 3/31) for those with ACP compared to 4.4% (n = 8/182) for © 2014 John Wiley & Sons Ltd
those without ACP; however, this did not reach statistical significance, v2(1) = 1.509, P = 0.219. Directives specified within the advance care plans Qualitative analysis of the 40 Advance Care Directives in this study revealed four themes: request for medical intervention, directives against medical intervention, consultation with family members and pre-/ post-mortem preferences. While 25 of the 40 Advance Care Directives included instructions for medical intervention, 16 of these also included directives against medical treatment. The data revealed that people’s knowledge and understanding of ACP appeared limited regarding the meaning, policies and practice of ACP. Request for medical intervention Twenty-five of the forty Advance Care Directives requested medical intervention in the event of clinical deterioration for example: For resuscitation (RAC003) Hospitalisation if necessary (RAC004)
517
M. Street et al.
Transport by ambulance (RAC013) Transfer to acute care hospital (RAC023) Go to hospital if deterioration in health (RAC032) Yes to investigations and diagnosis, Yes to surgical and medical procedures, Yes to pain and symptom relief (RAC048) All attempts to be made for resuscitation (RAC056) [I] wish to have antibiotic therapy and pain-relieving medications (RAC123)
Other requests centred on end-of-life care such as pain management and symptom control. Some people wanted to return to the residential aged care facility and receive palliative care.
If hospitalised, to go back to the Nursing Home as soon as possible for palliative care (RAC121) Pain relief only (RAC148) Palliative care in the event of gradual decline, to be kept comfortable (RAC222)
However, often these instructions were in the same Advance Care Directive besides requests for medical intervention and at times leaving the decision to the clinician or a family member as to what was best for the person at the time. For example: Would like cardiopulmonary resuscitation while medically appropriate and in best interests (RAC259) Wishes to have what is necessary without life support. Cardiopulmonary resuscitation only if medically beneficial (RAC204)
[I want] to be kept comfortable and pain free (RAC028) Table 2 Patient assessment and outcomes
Advanced care plan Total sample
Yes
No
N = 300
N = 40
N = 260
N
%
N
7 33
Length of admission (days) Admitted patients
N = 31
Median (IQR) 5 (2–9) Readmission to hospital within 30 days of index visit Yes 25 11.7 No 188 88.3
N
% v2(2) = 4.941, P = 0.085*
Triage category ATS = 1 3 1.0 ATS = 2 39 13.0 ATS = 3 140 46.7 ATS = 4 117 39.0 ATS = 5 1 0.3 Glasgow Coma Scale (GCS) (n = 289) 13–15 270 93.4 9–12 16 5.5
