Prevalence of Advance Care Planning Practices Among People with Chronic Diseases in Hospital and Community Settings: A Quasi-Experimental Design


 Background: Advance Care Planning (ACP) enables healthcare professionals to embrace the important process where patients think about their values in life and goals for health care, and discuss their future health care preferences with family members for a time when they are not able to make health care decisions. Despite the promotion of ACP last two decades, and well-known benefits of ACP and a written Advance Care Directive (ACD), they are still underutilised in Australia and across the world. Previous studies have provided some insights, however, an uptake of ACP and prevalence of ACDs in community setting is rarely reported.Methods: The aim of this study was to determine the uptake of ACP and prevalence of ACDs among people with chronic diseases in hospital and community settings. A retrospective medical record audit of eligible patients looking for evidence of ACP was conducted in 16 research sites (eight intervention and eight control) in hospital and community care settings. Participants included those who were admitted to one of the research sites, and who were aged 18 years and over with at least one of nine nominated chronic diseases. The primary outcome measures included the number of patients with evidence of ACP through the following practices: completion of an ACD, appointment of an Enduring Guardian (EG), or completion of a resuscitation plan. Results: The overall prevalence of ACD was 2.8% (n=28) out of 1006 audited records, and only 10 of them were legally binding. The number of EGs legally appointed was 39 (3.9%) across the sites. A total of 151 (15.4%) resuscitation plans were found across the eight hospital sites. 95% (n=144) of the resuscitation plans advised ‘Not-for-resuscitation’. Conclusions: The uptake of ACP is very low. Current medical recording system reveals the challenges in ACP lie in the process of storage, access and execution of the ACDs. Given that having an ACD or EG in place is only useful if the treating physician knows how and where to access the information, it has implications for policy, information system, and healthcare professionals’ education. Trial registration: The study was retrospectively registered with the Australian New Zealand Clinical Trials Registry (Trial ID: ACTRN12618001627246). The URL of the trial registry record http://www.anzctr.org.au/trial/MyTrial.aspx


Background
The concept of Advance Care Directive (ACDs), was rst proposed by Kutner in 1967 [1], and was adopted by the United States (US) government in 1991 [2,3]. Since then there have been growing interest and research to promote ACDs around the world. However, documentation rates of ACDs have been low to very low for three decades worldwide. Extensive research conducted across the world has led to the conclusion that the ACDs, as simply 'statements made by an individual', were neither understood nor accepted by individuals or healthcare professionals [4][5][6]. Hence, Advance Care Planning (ACP) has emerged as a response to this low uptake and practice. ACP enables health care professionals to embrace the important process where patients think about their values in life and goals for health care, and discuss their future health care preferences with family members for a time when they are not able to make health care decisions [7][8][9]. There are two main outcomes in ACP; a written ACD and an appointment of a legally binding Enduring Guardian (EG). It is ideal if ACDs are documented as a result of ACP, although individuals may decline to do an ACD. Anecdotal evidence also suggests that ACP may not result in legally binding ACDs documented by a competent adult but may result in 'other ACP documentation' such as resuscitation orders, nomination of a substitute decision-maker (SDM), and/or personally written letter to direct care.
Despite the promotion of ACP last two decades, and well-known bene ts of ACP [2][3][4], they are still underutilised in the Australian health care system [10,11] and across the world. In a Statewide-population survey [12] (n = 3055) in South Australia, more respondents reported having completed the enduring power of attorney (EPA) (22%) for nancial decisions than any of the health care-related documentsenduring power of guardianship (EPG) (13%), medical power of attorney (MPA) (11%), and anticipatory direction (12%). A recent multicentre audit study in Australia [13] reported that ACD prevalence was signi cantly higher in residential aged care facilities (RACFs) (47.7%) but still low in hospitals (15.7%) and general practices (3.2%) (p < 0.001) and varied across jurisdictions.
Previous studies that investigated the uptake of ACP and/or prevalence of ACDs have provided some insights. However, it must be noted that; 1) the terminology and documentation requirements vary locally, nationally and internationally, 2) previous attempts have relied on self-reporting, and 3) settings have been limited to institutions such as hospitals, RACFs or general practices [10,13,14]. An uptake of ACP and/or prevalence of ACDs in community setting is not widely reported. More importantly, given the various requirements, resources, forms and services available across Australia, the importance of the exact cause and effect of resources and services on the uptake of ACP is paramount for policy and service development, but is rarely reported.
In 2018, a trial of normalised ACP service was implemented for people with chronic conditions in hospital and community settings in New South Wales (NSW), Australia. This was a quasi-experimental study and the study protocol is reported elsewhere [15]. A total of 16 sites across two local health districts (LHDs) (eight in each) that covered both hospital and community settings, and both government and nongovernment agencies were selected. A medical record audit to examine the evidence of ACP was conducted before and during the 6-month intervention period at the 16 sites, to investigate the effect of the normalised ACP service delivered by the four RN ACP Facilitators.

Methods
The aim of this paper is to report the outcomes of the audit of the prevalence of ACP before the intervention was implemented at the 16 sites. For this study, the following chronic diseases included Cancer, Chronic Kidney Disease, Chronic Obstructive Pulmonary Disease, Congestive Heart Failure, Coronary Artery Disease, Dementia, Diabetes, Frailty and Hypertension.
Sites from two LHDs were invited to participate after obtaining the Ethics approval (Approval number: 17/12/13/4.16). Participating sites were eight wards in acute hospitals, which were pair-matched based on admission rates, patient characteristics, average length of stays, number of deaths per month/year, and number of referrals from/to hospital and community. The eight community settings included four public and four non-public home and community care service providers.
Medical records of those patients/clients; 1) who were aged 18 and over, 2) who had at least one of the chronic conditions, and 3) who were admitted to a selected site during April -May 2018 were requested from relevant medical records departments. It should be noted that the term admission includes service requests/home visits for patients receiving care from hospital or community health nurses. Records were audited for the evidence of ACP. For the purpose of this study the evidence of ACP included the following.
1. ACD: a legally binding document made by a legally capable person about the person's speci c wishes and preferences for future care. This includes treatments they would accept or refuse if they had a life-threatening illness or injury, their values in life and goals of care [16]. For an ACD to have su cient authority to act on, the four standards should be satis ed including speci city, currency, competence and witnessing [17,18].
2. An Enduring Guardian: an individual(s) who is legally appointed by the person and who can legally make decisions on behalf of the person about the person's medical and dental care, if the person loses capacity to make decision [16].
In NSW, 'a resuscitation plan' guides medical and healthcare professionals in using or withholding resuscitation measures and other aspects of treatment relevant at end of life [19]. Resuscitation plans are written by medical o cers in hospitals and do not consistently include discussion with the patient. For this reason, it does not meet the de nition of ACP in NSW but it is a common practice in hospital settings.
Hence, presence of 'resuscitation plans' were captured but not included in the overall prevalence of ACP.
Audits were conducted by two trained research assistants using a pilot-tested and standardised approach. The method used to search for evidence of ACP varied on the type of medical records kept to paper or electronic. All identi ed ACDs were reviewed for validity according to NSW jurisdiction. The detailed audit process is provided in Additional le 1.
All analyses were performed using Stata Version 15.0 (StataCorp LP, College Station, TX, USA). Due to the small number of the evidence of ACP, descriptive statistics were used to summarise the data.

Demographic characteristics of patients
In total 1006 patients' medical records were audited with 477 records in LHD-1 and 529 records in LHD-2. Demographic characteristics are presented in Table 1. In total, 48% of patients were male and 52% were female. The mean age of patients in LHD-1 was 72.75 years and 76.14 years in LHD-2. Prevalence of ACP A summary of the audit results is shown in Table 2. There was a low number of ACDs found across all sites. The overall prevalence of ACD was 2.8% (n = 28) out of 1006 audited records with a slightly higher prevalence of 4.7% (n = 25) in LHD-2, compared to 0.6% (n = 3) in LHD-1. Out of 28 ACDs found, only one (0.2%) in LHD-1 and nine (1.7%) of ACDs in LHD-2 were legally valid. Two (0.4%) and 16 (3.0%) ACDs in LHD-1 and LHD-2 respectively were legally invalid. Those invalid ACDs were due to being signed by 'Power of Attorney' who does not have a legal authority to make medical and lifestyle decisions in NSW. Two other patients who were transferred from a residential aged care facility had an ACD which was signed by 'Power of Attorney for Health Care' which is not a legalised term nor has a legal authority to make medical and lifestyle decisions in NSW. There was one case of staff referring to 'resuscitation plan' as ACDs. In another case, there was a written note by a social worker that "patient has come to terms with end of life/ ACP was discussed and documented", but ACD was not located in medical record. The prevalence of ACDs in hospital setting was 0.2% (n = 1) in LHD-1 and 1.3% (n = 7) in LHD-2. The prevalence of ACDs in community setting in LHD-2 was 0.4% (n = 2) and no ACD was found in community setting in LHD-1.
The relationship of the EG to the patient is presented in Fig. 1. Of those (n = 39) who appointed legally valid EG, 72% (n = 28) elected their children as their EG. It should be noted that this included stepchildren and children-in-law, although this was a small fraction of the cases.

Low prevalence of ACDs and EGs
Given the lack of available evidence on the practice of ACP in community setting, this study adds new insights about the prevalence of ACP in both government and non-government community settings. Like previous studies [13,18], low number of ACDs and EGs were found in patient records. Only 10 (1%) of the patients had a legally binding ACD and 39 (3.9%) of the patients had an EG. It is beyond the scope of this study and extensive research [2,3,8] already suggests that the reasons for low awareness of ACD and appointment of EG include; 'Don't know how to do it', 'Di cult to understand the form' and 'Di culty to understand what it is'.

Resuscitation Plans
There was a much greater number of resuscitation plans recorded in patient records compared to both EG and ACD. One concern with resuscitation plans found in the audit was that 32 of the total 151 were not signed by a medical o cer. This is troubling as it calls into question the legal validity of the document should a doctor have to decide about whether to attempt resuscitation. Another question one must ask is why doctors initiate and complete more resuscitation plans than ACDs. Discussion and completion of a resuscitation plan should prompt discussion for completion of an ACD. The fact that resuscitation plan can be written by and for the person whilst ACDs can be only written by a legally competent person [17], may explain one aspect. The exact reasons for more prevalent use of resuscitation plan than ACD require more investigation.

Inconsistency and inaccessibility of location and storage of ACDs and EGs
The process of auditing revealed that the current medical record system used to record patient information are either ill-equipped or underutilised to record the presence of an ACD. For LHD-2, most of the records were searched using 'PowerChart' for hospital patients (See Supplementary le 1). This system was the only one to speci cally have an 'Advanced Care Planning' tab in an obvious and accessible location. This made it easier to nd ACDs in the auditing procedure. It would also make it easier for treating healthcare professionals to access when necessary. Yet ve out of nine ACDs found in LHD-2 were all located by manually searching through patients' paper medical records. Treating health care professionals in the Emergency Department are unlikely to have access to the paper record in subsequent admissions. On the other hand, LHD-1 used medical record access programs such as Digital Medical Record (DMR) and Community Health Information Management Enterprise (CHIME), where the recording of ACDs was far more inconsistent. This made evidence of ACP di cult to nd, which means a greater chance of reporting a false negative for their patients.
The auditing also revealed inconsistent or inappropriate recordings by healthcare professionals. LHD-1 Control Hospital site 3 and 4 were more closely examined as it was the only hospital sites to show no evidence of ACDs. On admission, nurses routinely ll out an 'Adult inpatient admission and risk assessment' form electronically. This form includes a check box asking if patient does have a pre-existing adult resuscitation plan/ACD/ACP. In 17 of 148 patients (11.4%) this was ticked yes but there was no evidence of the ACP/ACD in any other part of the patient's record. Twenty-three (15.5%) of the 'Adult inpatient admission and risk assessment' forms were either missing or blank. Inconsistent recording was also evident in EG, as there were numerous mentions of EGs in patient notes with no formal documentation found. This was evident in patient records at all sites, but in particular, in community sites.
In summary, storage of ACP related documents is inconsistent and does not facilitate timely access to or execution of these documents. These challenges are noted in other studies [15,18]. One study by Cheang et al. [20] revealed that out of 100 patients, they could not nd any record of ACDs in their medical les, despite 12% of patients reporting to have an ACD in place in their interview. This has a serious implication for clinical practice, policy and system to ensure an easily accessible storage to be available for all involved. With the launch of the national digital health record system 'My Health Record' in Australia, it opens a potential solution for all, although some individuals may opt out. It also has implications for treating healthcare professionals' education regarding accurate and comprehensive collection of ACP information, followed by consistent storage in a designated, accessible location.
The study is limited as it was relied on medical records only although the rigour and validity were ensured as in the detailed audit process (Additional le 1). ACP/ACD documentation may be kept by the patient, general practitioner, specialist or EG which could not be found in the audit. The ndings of the audit may be limited to the 16 study sites.

Conclusion
Despite the efforts made to promote ACP, its prevalence remains very low in NSW. In addition to this, the current medical record systems appear ill-equipped to correctly record ACP practices and ACD documents. Having an ACD or EG in place is only useful if the treating physician knows where and how to access the information. To ensure the most bene t for all, ACP needs to be promoted to both patients and medical health practitioners including nurses. This will lead to a normalisation of ACP practices and have it consistently recorded in patient's records, where it can be of bene t.

Declarations
Ethics approval: The study has been approved by the Hunter New England Human Research Ethics Committee, approval no. 17/12/13/4.16 Consent for Publication: Not applicable.
Availability of data and materials: All data generated or analysed during this study are included in this published article.
Competing interests: None declared.
Funding: The study was funded by the NSW Health Translational Research Grant Scheme (Round II #116). The funding body was not directly involved with the conception and development of the project. The funding body reviewed and approved the design of the study and the method of data collection and analysis. The funding body has no role in the design of the study, collection, analysis or interpretation of data, and in writing the manuscript.
Author Contributions: SJ was involved in the conception, design, implementation of the intervention, data acquisition, analysis and interpretation, and drafting the manuscript. SO and PC were involved in implementation of the intervention, data acquisition, analysis and interpretation, and drafting the manuscript. TB was involved in the conception, design, analysis and interpretation, and drafting the manuscript. RD was involved in data acquisition, analysis and interpretation, and drafting of the manuscript. All authors approved the nal version of the manuscript.