View Article PDF

Advance care planning (ACP) is defined in New Zealand as the process of thinking about, talking about and planning for future healthcare and end-of-life care and identifying what is important within this care.1 The ACP process provides a person with the opportunity to document their wishes in an advance care plan (ACPlan), which may include an advance directive (AD) component to accept or refuse specific treatments that may be offered in the future. This enables the wishes of hospitalised patients to be recognised during end-of-life care if and when they lose competence. In New Zealand, clinicians are responsible for assessing whether a treatment is medically indicated. This decision takes into account the patient’s condition, projected outcome after treatment and consultation with family members.1 If treatment is medically indicated, the ACP is used to help decide whether or not to proceed.

One objective of the review described in this paper was to examine the influence of age and comorbidities on the content of ACPlans and ADs. The content of an ACPlan or AD and the expectations of care to be provided may reflect a complex multitude of factors such as the health status and prognosis of a person, expected quality of life, involvement of family members and a diversity of religious, spiritual and cultural beliefs.2–4 A person’s understanding of their clinical situation and perceptions regarding the balance between purported benefit of treatment and potential harm may also influence content.5 For patients who are seriously ill, ACP is often conducted under compromised or time constrained conditions.6 On occasions, the ACPlan may represent a compromise between a person’s expectations and the clinical resources available and can be influenced by the time the plan was prepared and the person’s relationship with the health professional assisting with preparation of the ACPlan.7

The second objective of the review was to evaluate the concordance between the wishes and directives in the plans and the care received when admitted to a public hospital. When a person writes an ACPlan they do so with the belief that their wishes and instructions will be complied with during care. Studies that have investigated the influence of ACP on end-of-life care have shown that it can reduce hospitalisation rates and increase hospice and palliative care,8 although the care received is not always consistent with a person’s preferences.9–12 This may occur in difficult clinical situations where other factors such as prognosis, perceived quality of life and the wishes of family or friends may influence compliance with the ACPlan or AD. For example, a survey of the attitudes of Australian and New Zealand anaesthetists towards resuscitation orders in the perioperative setting showed that while the majority recognised the importance of an AD, only 45% followed the directives in clinical practice.13

Methods

A project to promote ACP was introduced in Canterbury in December 2013 supported by the Canterbury Initiative, Canterbury District Health Board (DHB) (https://www.cdhb.health.nz/about-us/key-projects-and-initiatives/canterbury-initiative/). There has been a steady increase in the number of ACPlans created and shared electronically across primary and secondary healthcare that are now available immediately to medical staff in public hospitals on admission.

The current review was carried out as a student project under the conditions of an out-of-scope ethical approval granted by the Ministry of Health and Disability Ethics Committee to undertake routine evaluation of services such as ACP that are included on Community HealthPathways (http://edu.cdhb.health.nz/Hospitals-Services/Health-Professionals/Pages/Health-Pathways.aspx), a clinical guidance website for health professionals, and HealthInfo (https://www.healthinfo.org.nz/), an open-access website that provides health information for the general population.

The study involved a retrospective review of case notes of individuals who had published an ACPlan and subsequently received care in a publicly funded hospital. A published ACPlan was defined as those that had passed a two-step review process to ensure clinical interpretability and had then been loaded onto the person’s electronic health record. A total of 1,939 ACPs were published in Canterbury in the three-year period between 10 September 2014 to 31 September 2017. These ACPs were matched to hospital admissions recorded on the Canterbury DHB clinical database to identify people who had received care in a publicly funded hospital up to the end of September 2017 (n=667). One hundred and fifty people were selected randomly from this group for inclusion in the review, with this number taking into account the relatively short duration of the student project. A review of their electronic and written clinical records was carried out by the student (HBS), with all admissions involving an incompetent patient and a random subset of about 10% of other admissions reviewed independently by senior clinicians (RW and KG) to check concordance between reviewers. The following data were recorded.

1. Demographic variables: gender, age, ethnicity, residential location and deprivation index (NZDep13)14 of the suburb in which the individual lived, using the classifications of the 2013 New Zealand census.15

2. Clinical variables: medical diagnoses made before and six months after publication of the ACPlan, and medical condition assessed by the Quan age-adjusted modification of the Charlson Comorbidity Index (CCI).16 This algorithm is based on 22 conditions and predicts 10-year mortality and was chosen because of proven validity in a wide range of populations.17

3. Content of the ACPlan: all patients prepared their ACPlan using the same template18 that contained the following information and instructions. Date published, goal-of-care and specific treatment preferences in the event of future deterioration in health (see Table 1), instructions for cardiopulmonary resuscitation (CPR), preferred place of death, choices for organ donation and changes in content of the ACPlan due to progression of illness and/or change in insight of the implications of the disease.

4. Assessment of competence: competence at the time of each admission was categorised as either competent, incompetent or fluctuating competence. Competence was defined as the ability of the patient to understand the need to make a decision, process the information presented and communicate their decisions regarding treatment. This assessment was based on entries in the clinical notes of the patient that documented level of alertness and orientation and/or the level of consciousness (Glasgow Coma Scale),19 with signs of incompetence including confusion, delirium and progressive dementia.

5. Implementation of the ACPlan: the treatment received during each hospital admission for the review cohort was reviewed to determine concordance with the goal-of-care. This involved, when available, recording the investigations and treatment received and clinical outcomes of all the admissions from electronic clinical, radiological and laboratory records and written case notes. Concordance between clinical management of incompetent patients and the contents of the ACPlan, AD and CPR preference was reviewed separately, with final evaluation of the findings carried out by RW and KG.

6. Place of death: where possible, the place of death was recorded for comparison with the person’s preferred option.

Table 1: Content of the ACP including selection of goal-of-care, treatment preferences, preferred place of death and organ donation.

Data analysis

The frequencies of categorical variables in the cohort were expressed as number and percentage. Age was expressed as median and interquartile range because of the predominance of elderly people in the cohort. The chi-squared goodness-of-fit test was used to examine whether the observed frequency of demographic variables of the study cohort differed from those in the 2013 census of the New Zealand population24 and Canterbury DHB-Primary Health Organisation population.

To examine the influence of age or severity of comorbidities on the choice of goal-of-care and CPR preference stated in the ACP or AD the data were analysed using the chi square test or Fisher’s exact test. A logistical regression model was constructed to examine the relationship between changes in the choice of goal-of-care as the outcome variable and increasing age or severity of comorbidities as the explanatory variable, and between the preference to not receive CPR (outcome) and age (predictor). The level of statistical significance for all the analyses was set at p≤0.05.

Results

Demographic data

The demographic and clinical characteristics of the study cohort are shown in Table 2. One person was excluded because of a lack of adequate information in their medical records, leaving 149 people in the analysis set. The cohort was comprised of mainly elderly people (median age, 78 years), with 15% living in an aged residential care (ARC) facility. There were significantly more females than males in the cohort (57% vs 43%, p<0.01). Caucasians were over-represented (94%) compared with that in census data (81%), with a significantly reduced proportion of all other ethnicities (p=0.02). There was also a marked under-representation of the most socioeconomically deprived people and a marginally lower proportion of the least deprived individuals compared to census data (p<0.01). As shown in Figure 1, 14% of the study cohort had no comorbidities and 57% had a relatively low risk of 10-year mortality, ie, CCI ≤2. The incidence of medical conditions included in the CCI in the study cohort were cancer 40%, respiratory disorders 34%, cardiovascular disease 19%, renal disease 19%, rheumatologic disease 18%, dementia 6%, diabetes mellitus 4%, hemi- or paraplegia 3%, and liver disease 2%.

Table 2: Demographic and clinical characteristics of the study cohort.

*p value for comparison of study cohort and reference populations.
§Canterbury DHB-Primary Health Organisation
Index (0–8) predicts mortality, with scores between 1–2 and ≥5 associated with a 5-fold and 20-fold increase in the risk of dying in the next 10 years, respectively.

Figure 1: Graph showing the distribution of the Charlson Comorbidity Index in the review cohort.

Contents of the ACP

Table 3 summarises the selection of goal-of-care and other directives and preferences in the study cohort. Nearly one-half of the cohort selected Option 3 for their goal-of- care that stipulated care should be focused on symptoms only without any life-prolonging treatment. The majority (80%) did not want CPR to be carried out, 15% wanted resuscitation to be attempted if medically indicated, while 5% expressed no preference. Approximately two-thirds of the cohort had directives excluding a specific treatment. The most preferred place of death was ‘at the residential property’ (40%) compared to palliative care within an ARC or associated hospice (27%), or hospital (6%), while 27% of people had no preference.

Only three people in the cohort had updated their ACPlan, with one no longer wanting to receive CPR, one making changes following discussions with their partner, and one because of a diagnosis of cancer.

Table 3: Preferences stated in the ACP by the study cohort.

Relationship between ACP content and age and level of morbidity

Age

A significant association was observed between increased age and choice of goal-of-care. (Figure 2A). Logistic regression analysis showed that age influenced the choice of goal-of-care 2 (odds ratio [OR] 1.77; p=0.009), 3 (OR 1.93; p=0.001) and 5 (OR 2.49; p=0.04) over goal-of-care 1, but not the choice of goal-of-care 4 (OR 1.49; p=0.13) (Figure 2B). Increasing age also influenced CPR preference, such that a person was 1.7 times more likely to not want CPR than a person 10 years younger (p=0.04) (Figure 2C).

Figure 2: Graphs showing (a) The relationship between increasing age and the choice of goal-of-care and preference regarding CPR, (b) Logistic regression model showing the influence of a step-wise 10-year increase in age on the choice of goal-of-care, and (c) binomial logistic regression model showing the predicted probability for preference not to receive CPR. The small circles represent individual patients, the blue line shows the increasing probability of older people not wanting to be resuscitated (shaded area—95 % confidence intervals).

Figure 2A

Figure 2B

Figure 2C

Comorbidity index

As shown in Figure 3, individuals with a higher CCI and risk of mortality tended to choose goal-of-care 4, in which they wanted health professionals to lead decision making in conjunction with family members or friends. The distribution and median value of CCI were similar for the other four goals-of-care. No association was observed between the CCI and preference regarding CPR.

Figure 3: Graph showing the relationship between increasing severity of comorbidities and the choice of goal-of-care and preference regarding CPR.

Concordance between the ACPlan and AD content and treatment during hospital admissions

A total of 411 admissions to a public hospital since publication of the ACPlan were identified in the study cohort. Twelve percent of these admissions were to the emergency department (ED), with 88% requiring an inpatient stay. Median CCI was similar for the ED and inpatient admissions (3.3 vs. 3.2, respectively).

The majority of people (63%) were deemed competent during admission, 14% partially competent and 11% non-competent. Competence could not be assessed retrospectively in the remaining 12% of individuals. The goal-of-care was adhered to in 59 of the 60 admissions involving non-competent people. Treatment preferences stated in the AD were relevant in seven cases and were adhered to in six cases. The one case of non-concordance involved a patient with an infective exacerbation of COPD who experienced a myocardial infarction. Despite the ACPlan stating life-prolonging care or CPR for a life-threatening chest infection should not be provided, the patient was treated with intravenous fluids and antibiotics. All patients who stated they did not want CPR in the event of a cardiac arrest had their AD visible within the ACP on their electronic medical records. No patient in the study cohort received CPR.

Preferred place of death

Fifty-five people died, with the place of death identified in 44; 21 (48%) dying in their preferred place, 16 (36%) in a non-preferred place, and seven (16%) with no preference stated.

Discussion

The decision to prepare an ACPlan may be influenced by a person’s physical condition and prognosis combined with social factors such as gender, age, cultural and ethnic background, religion and education.2–4 The first aim of this review was to examine the effect of comorbidities as a clinical factor and age as a social factor on the choice of goal-of-care and decision regarding CPR. Our data indicated that age had a measurable influence on a person’s choice of goal-of-care, with older people tending to want to receive only treatments that maintained their comfort and dignity. With increasing age there was a steady decrease in the selection of goal-of-care option that accepted all available life-prolonging treatments, including CPR. In comparison, the severity of comorbidities appeared to have less influence on these decisions, with people with more severe disease tending to want healthcare professionals to make decisions regarding end-of-life care in consultation with family members or friends. This finding is consistent with other studies that reported older adults with multiple conditions were more likely to prefer active decision making with healthcare professionals.20 However, our finding of a lack of association between comorbidities and CPR preference is different from the general finding in the published literature that the severity of disease burden is a major trigger for the decision not to receive resuscitation.21

The second aim of the review was to examine compliance between ACPlans and end-of-life care. This showed that the treatment received during admission to a public hospital was generally consistent with the goals-of-care, while in six of the seven cases involving an incompetent patient, the specific treatment preferences in the AD were adhered to. It is necessary to consider the above findings from the point of view that documentation of care in case notes should not be conflated with the actual care received and that evaluation of concordance may not represent the patient’s actual wishes in the ‘real time’ context of a hospital admission. Reflecting these methodological difficulties, the findings of published studies similar to the current review are characterised by their inconsistency. For example, two recent studies in Australia reported opposite findings, with one showing about one-third of cases received treatment that contradicted preferences listed in the ACPlan,22 while the other found the majority of patients received hospital care fully concordant with their stated preferences.23 Variations in methodology and the make-up of the patients investigated may also contribute to differences between studies.24

Of the people in the cohort who died, approximately 50% died at their preferred place, a finding consistent with local audit data that showed 82% of people with an ACPlan died in a setting such as their home or aged residential care and that the proportion of people in Canterbury over 75 years who die in hospital is 30% compared with the national average of 34%. These findings are in accordance with international evidence that ACP is strongly associated with lower rates of hospital deaths.25

Our review showed ethnicity, and to a lesser extent, socioeconomic status, influenced the make-up of the cohort of people. The over representation of Caucasians and under-representation of minority ethnic groups in the cohort is consistent with previous reports in New Zealand26 and other multicultural countries.27,28 This under-representation may reflect differences in family-centred decision making and attitudes towards life-prolonging treatments and influence of religious beliefs and doctor-patient relationships.29 Our observation of a lower number of ACPlans prepared by people in lower socioeconomic groups is consistent with the findings of a large survey carried out in the US.30

The review had several limitations. The number of people investigated was relatively small and as a consequence we were able to evaluate the concordance between care received and the ACPlan in only seven cases of incompetent patients who had expressed specific treatment preferences. In addition, in approximately 20% of admissions there was insufficient information to determine competence, although in the majority of these cases the admission was for short-term, pre-arranged treatment in a medical day unit. Determining concordance of treatment with the ACP was problematic in some cases as the goal-of-care could be interpreted differently, with the true intention of care provided difficult to assess retrospectively using only clinical records. The demographics of the review cohort were also markedly different than reference populations in census data and local health authority databases, making the generalisability of our findings to the population questionable. Therefore, our results need to be considered after taking these limitations into account, but may provide useful information for future similar studies to identify populations that require further investigation.

We conclude that our review shows that age and to a lesser extent the level of comorbidities may influence the choice of goal-of-care in an ACPlan. In our small cohort of people, admission to a publicly funded hospital for both routine and end-of-life care appeared to adhere to the wishes and preferences stated in the ACP. Our findings support the usefulness of ACPs in that they provide the opportunity for patient choice and self-determination and encourage the wishes of hospitalised patients to be recognised and used to guide care.

Summary

Abstract

Aim

This retrospective review examined the influence of age and severity of comorbidities on goals-of-care in advance care plans (ACPlans) and concordance between these wishes and care received during hospital admission.

Method

The medical records of 149 people with an ACPlan admitted to a public hospital were reviewed to evaluate concordance with treatment. The associations between age and comorbidities and goals-of-care were determined using contingency tables and logistic regression analyses.

Results

The majority of the review cohort were Caucasian and elderly, with people from minority ethnic groups under-represented compared to census data. Increasing age had a measurable influence on the choice of goal-of-care, whereas comorbidity severity had less influence on this decision. In 60 of the 411 hospital admissions the patient was classified as incompetent, with the goal-of-care adhered to in 59 of these cases and treatment preferences adhered to in six of seven cases. Fifty-five people had died since writing their ACPlan, with 63% dying at their preferred place or with no preference stated.

Conclusion

Age and to a lesser extent the severity of comorbidities influence the choice of goal-of-care in an ACPlan. Our review also showed that end-of-life care appeared to adhere to the instructions in the plan.

Author Information

Henricus Berend Speelberg, Medical Student, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Brett Shand, Medical Writer, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Jane Goodwin, Advance Care Planning Facilitator, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Rachel Wiseman, Consultant Physician, Respiratory Medicine, Canterbury District Health Board, Christchurch; Laura Belluscio, Biostatistician, Hospital Services Administration, Canterbury District Health Board, Christchurch; Kate Grundy, Palliative Care Physician, Oncology Services, Canterbury District Health Board, Christchurch.

Acknowledgements

We are grateful for the help provided by the staff at the clinical records warehouse with collection of data and Ma Yi, Biostatistician, Canterbury DHB for assistance with the statistical analyses.

Correspondence

Dr Kate Grundy, Oncology Services, Canterbury District Health Board, Christchurch 8142

Correspondence Email

kate.grundy@cdhb.health.nz

Competing Interests

Nil.

1. New Zealand Ministry of Health. Advance Care Planning. A guide for the New Zealand health care workforce. Accessed on 20 November, 2019 from http://www.health.govt.nz/system/files/documents/publications/advance-care-planning-aug11.pdf

2. Ke LS, Huang X, Hu WY, O'Connor M, Lee S. Experiences and perspectives of older people regarding advance care planning: A meta-synthesis of qualitative studies. Palliat Med. 2017; 31(5):394–405. doi: 10.1177/0269216316663507

3. Jeong S, Ohr S, Pich J, Saul P, Ho A. ‘Planning ahead’ among community dwelling older people from culturally and linguistically diverse background: a cross-sectional study. J Clin Nurs 2015; 24(1–2):244–255.

4. Fakhri S, Engelberg RA, Downey L, et al. Factors Affecting Patients' Preferences for and Actual Discussions About End-of-Life Care. J Pain Symptom Manage. 2016; 52(3):386–394. doi:10.1016/j.jpainsymman.2016.03.012.

5. Boot M, Wilson C. Clinical nurse specialists perspectives on advance care planning conversations: a qualitative study. Int J Palliat Nurs. 2014 Jan; 20(1):9–14.

6. Lund S, Richardson A, May C. Barriers to advance care planning at the end of life: an explanatory systematic review of implementation studies. PLoS One. 2015 Feb.

7. Sharp T, Malyon A, Barclay S. GPs' perceptions of advance care planning with frail and older people: a qualitative study. Br J Gen Pract. 2018; 68(666):e44-e53. doi: 10.3399/bjgp17X694145.

8. Brinkman-Stoppelenburg A, Rietjens JA, van der Heide A. The effects of advance care planning on end-of-life care: a systematic review. Palliat Med. 2014; 28(8):1000–1025. doi: 10.1177/0269216314526272

9. Hardin SB, Yusufaly YA. Difficult end-of-life treatment decisions: do other factors trump advance directives? [published correction appears in Arch Intern Med. 2004 Oct 25; 164(19):2124]. Arch Intern Med. 2004; 164(14):1531–1533. doi:10.1001/archinte.164.14.1531:

10. Hopping-Winn J, Mullin J, March L, et al. The progression of end-of-life wishes and concordance with end-of-life care. J Palliat Med. 2018; 21:541–5. doi: 10.1089/jpm.2017.0317. doi: 10.1089/jpm.2017.0317.

11. Khandelwal N, Curtis JR, Freedman VA, et al. How often Is end-of-life care in the United States inconsistent with patients' goals of care? J Palliat Med. 2017; 20:1400–4. doi: 10.1089/jpm.2017.0065.

12. MacKenzie MA, Smith-Howell E, Bomba PA, Meghani SH. Respecting choices and related models of advance care planning: A systematic review of published evidence. Am J Hosp Palliat Care. 2018; 35:897–907. doi: 10.1177/1049909117745789.

13. Keon-Cohen Z, Myles PS, Story DA. A survey of Australian and New Zealand anaesthetists' attitudes towards resuscitation orders in the perioperative setting. Anaesth Inten Care J. 2017; 45:396–402.

14. New Zealand Ministry of Health. NZDep2013 Index of Deprivation. Accessed on 20 November, 2019 from http://www.health.govt.nz/publication/nzdep2013-index-deprivation

15. Stats NZ. The New Zealand Census of Population and Dwellings, 2013. Accessed on 20 November, 2019 from: http://www.stats.govt.nz/ 2154:188–94. doi: 10.1097/MLR.0000000000000471.

16. Quan H, Li B, Couris, CM, et al. Updating and Validating the Charlson Comorbidity Index and Score for Risk Adjustment in Hospital Discharge Abstracts Using Data From 6 Countries. Am J Epidemiol. 2011; 173:676–82.

17. Austin SR, Wong Y, Uzzo RG, et al. Why summary comorbidity measures such as the Charlson Comorbidity Index and Elixhauser score work. Med Care 2015; 53:e65–e72. doi: 10.1097/MLR.0b013e318297429c

18. Canterbury Initiative. “My advance care plan”. Accessed on 20 November, 2019 from http://www.healthinfo.org.nz/patientinfo/MyAdvanceCarePlan.pdf

19. Teasdale G, Maas A, Lecky F, et al. The Glasgow Coma Scale at 40 years: standing the test of time. Lancet Neurol. 2014; 13:8.

20. Chi WC, Wolff J, Greer R, Dy S. Multimorbidity and Decision-Making Preferences Among Older Adults. Ann Fam Med. 2017; 15:546–551. doi:10.1370/afm.2106

21. Mockford C, Fritz Z, George R, et al. Do not attempt cardiopulmonary resuscitation (DNACPR) orders: a systematic review of the barriers and facilitators of decision-making and implementation. Resuscitation. 2015; 88:99–113. doi:10.1016/j.resuscitation.2014.11.016

22. Johnson SB, Butow PN, Kerridge I, Bell ML, Tattersall MHN. How Well Do Current Measures Assess the Impact of Advance Care Planning on Concordance Between Patient Preferences for End-of-Life Care and the Care Received: A Methodological Review. J Pain Symptom Manage. 2018 Feb; 55(2):480–495. doi: 10.1016/j.jpainsymman.2017.09.008.

23. Panozzo L, Ward B, Harvey P, Fletcher J. Content and implementation of advance care plans: A retrospective cohort study. Aust J Gen Pract. 2019 May; 48(5):323–325. doi: 10.31128/AJGP-12-18-4783.

24. Oo NM, Scott IA, Maggacis R, Rajakaruna N. Assessing concordance between patient preferences in advance care plans and in-hospital care. Aust Health Rev. 2019 Aug; 43(4):425–431. doi: 10.1071/AH18011.

25. Brinkman-Stoppelenburg A, Rietjens JAC, van der Heide A. The effects of advance care planning on end-of-life care: A systematic review, Palliat. Med. 2014; 28:1000–1025.

26. Cawley LM. Racial, cultural, and ethnic factors influencing end-of-life care. J Palliat Med. 2005; 8:S58–S69.

27. Portanova J, Ailshire J, Perez C, Rahman A, Enguidanos S. Ethnic Differences in Advance Directive Completion and Care Preferences: What Has Changed in a Decade?. J Am Geriatr Soc. 2017; 65:1352–1357. doi: 10.1111/jgs.14800.

28. McLaughlin L, Braun K. Asian and Pacific Islander cultural values; considerations for healthcare decision making. Health & Social work. 1998; 23:116–126.

29. Oetzel J, Simpson M, Berryman K, Iti T, Reddy R. Managing communication tensions and challenges during the end-of-life journey. Perspectives of Maori kaumatua and their whanau. Palliat Med. 2015; 29:764–766. doi: 10.1177/0269216315583619.

30. Dobalian A. Advance care planning documents in nursing facilities: results from a nationally representative survey. Arch Gerontol Geriatr. 2006; 43:193–212. doi:10.1016/j.archger.2005.10.007.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Advance care planning (ACP) is defined in New Zealand as the process of thinking about, talking about and planning for future healthcare and end-of-life care and identifying what is important within this care.1 The ACP process provides a person with the opportunity to document their wishes in an advance care plan (ACPlan), which may include an advance directive (AD) component to accept or refuse specific treatments that may be offered in the future. This enables the wishes of hospitalised patients to be recognised during end-of-life care if and when they lose competence. In New Zealand, clinicians are responsible for assessing whether a treatment is medically indicated. This decision takes into account the patient’s condition, projected outcome after treatment and consultation with family members.1 If treatment is medically indicated, the ACP is used to help decide whether or not to proceed.

One objective of the review described in this paper was to examine the influence of age and comorbidities on the content of ACPlans and ADs. The content of an ACPlan or AD and the expectations of care to be provided may reflect a complex multitude of factors such as the health status and prognosis of a person, expected quality of life, involvement of family members and a diversity of religious, spiritual and cultural beliefs.2–4 A person’s understanding of their clinical situation and perceptions regarding the balance between purported benefit of treatment and potential harm may also influence content.5 For patients who are seriously ill, ACP is often conducted under compromised or time constrained conditions.6 On occasions, the ACPlan may represent a compromise between a person’s expectations and the clinical resources available and can be influenced by the time the plan was prepared and the person’s relationship with the health professional assisting with preparation of the ACPlan.7

The second objective of the review was to evaluate the concordance between the wishes and directives in the plans and the care received when admitted to a public hospital. When a person writes an ACPlan they do so with the belief that their wishes and instructions will be complied with during care. Studies that have investigated the influence of ACP on end-of-life care have shown that it can reduce hospitalisation rates and increase hospice and palliative care,8 although the care received is not always consistent with a person’s preferences.9–12 This may occur in difficult clinical situations where other factors such as prognosis, perceived quality of life and the wishes of family or friends may influence compliance with the ACPlan or AD. For example, a survey of the attitudes of Australian and New Zealand anaesthetists towards resuscitation orders in the perioperative setting showed that while the majority recognised the importance of an AD, only 45% followed the directives in clinical practice.13

Methods

A project to promote ACP was introduced in Canterbury in December 2013 supported by the Canterbury Initiative, Canterbury District Health Board (DHB) (https://www.cdhb.health.nz/about-us/key-projects-and-initiatives/canterbury-initiative/). There has been a steady increase in the number of ACPlans created and shared electronically across primary and secondary healthcare that are now available immediately to medical staff in public hospitals on admission.

The current review was carried out as a student project under the conditions of an out-of-scope ethical approval granted by the Ministry of Health and Disability Ethics Committee to undertake routine evaluation of services such as ACP that are included on Community HealthPathways (http://edu.cdhb.health.nz/Hospitals-Services/Health-Professionals/Pages/Health-Pathways.aspx), a clinical guidance website for health professionals, and HealthInfo (https://www.healthinfo.org.nz/), an open-access website that provides health information for the general population.

The study involved a retrospective review of case notes of individuals who had published an ACPlan and subsequently received care in a publicly funded hospital. A published ACPlan was defined as those that had passed a two-step review process to ensure clinical interpretability and had then been loaded onto the person’s electronic health record. A total of 1,939 ACPs were published in Canterbury in the three-year period between 10 September 2014 to 31 September 2017. These ACPs were matched to hospital admissions recorded on the Canterbury DHB clinical database to identify people who had received care in a publicly funded hospital up to the end of September 2017 (n=667). One hundred and fifty people were selected randomly from this group for inclusion in the review, with this number taking into account the relatively short duration of the student project. A review of their electronic and written clinical records was carried out by the student (HBS), with all admissions involving an incompetent patient and a random subset of about 10% of other admissions reviewed independently by senior clinicians (RW and KG) to check concordance between reviewers. The following data were recorded.

1. Demographic variables: gender, age, ethnicity, residential location and deprivation index (NZDep13)14 of the suburb in which the individual lived, using the classifications of the 2013 New Zealand census.15

2. Clinical variables: medical diagnoses made before and six months after publication of the ACPlan, and medical condition assessed by the Quan age-adjusted modification of the Charlson Comorbidity Index (CCI).16 This algorithm is based on 22 conditions and predicts 10-year mortality and was chosen because of proven validity in a wide range of populations.17

3. Content of the ACPlan: all patients prepared their ACPlan using the same template18 that contained the following information and instructions. Date published, goal-of-care and specific treatment preferences in the event of future deterioration in health (see Table 1), instructions for cardiopulmonary resuscitation (CPR), preferred place of death, choices for organ donation and changes in content of the ACPlan due to progression of illness and/or change in insight of the implications of the disease.

4. Assessment of competence: competence at the time of each admission was categorised as either competent, incompetent or fluctuating competence. Competence was defined as the ability of the patient to understand the need to make a decision, process the information presented and communicate their decisions regarding treatment. This assessment was based on entries in the clinical notes of the patient that documented level of alertness and orientation and/or the level of consciousness (Glasgow Coma Scale),19 with signs of incompetence including confusion, delirium and progressive dementia.

5. Implementation of the ACPlan: the treatment received during each hospital admission for the review cohort was reviewed to determine concordance with the goal-of-care. This involved, when available, recording the investigations and treatment received and clinical outcomes of all the admissions from electronic clinical, radiological and laboratory records and written case notes. Concordance between clinical management of incompetent patients and the contents of the ACPlan, AD and CPR preference was reviewed separately, with final evaluation of the findings carried out by RW and KG.

6. Place of death: where possible, the place of death was recorded for comparison with the person’s preferred option.

Table 1: Content of the ACP including selection of goal-of-care, treatment preferences, preferred place of death and organ donation.

Data analysis

The frequencies of categorical variables in the cohort were expressed as number and percentage. Age was expressed as median and interquartile range because of the predominance of elderly people in the cohort. The chi-squared goodness-of-fit test was used to examine whether the observed frequency of demographic variables of the study cohort differed from those in the 2013 census of the New Zealand population24 and Canterbury DHB-Primary Health Organisation population.

To examine the influence of age or severity of comorbidities on the choice of goal-of-care and CPR preference stated in the ACP or AD the data were analysed using the chi square test or Fisher’s exact test. A logistical regression model was constructed to examine the relationship between changes in the choice of goal-of-care as the outcome variable and increasing age or severity of comorbidities as the explanatory variable, and between the preference to not receive CPR (outcome) and age (predictor). The level of statistical significance for all the analyses was set at p≤0.05.

Results

Demographic data

The demographic and clinical characteristics of the study cohort are shown in Table 2. One person was excluded because of a lack of adequate information in their medical records, leaving 149 people in the analysis set. The cohort was comprised of mainly elderly people (median age, 78 years), with 15% living in an aged residential care (ARC) facility. There were significantly more females than males in the cohort (57% vs 43%, p<0.01). Caucasians were over-represented (94%) compared with that in census data (81%), with a significantly reduced proportion of all other ethnicities (p=0.02). There was also a marked under-representation of the most socioeconomically deprived people and a marginally lower proportion of the least deprived individuals compared to census data (p<0.01). As shown in Figure 1, 14% of the study cohort had no comorbidities and 57% had a relatively low risk of 10-year mortality, ie, CCI ≤2. The incidence of medical conditions included in the CCI in the study cohort were cancer 40%, respiratory disorders 34%, cardiovascular disease 19%, renal disease 19%, rheumatologic disease 18%, dementia 6%, diabetes mellitus 4%, hemi- or paraplegia 3%, and liver disease 2%.

Table 2: Demographic and clinical characteristics of the study cohort.

*p value for comparison of study cohort and reference populations.
§Canterbury DHB-Primary Health Organisation
Index (0–8) predicts mortality, with scores between 1–2 and ≥5 associated with a 5-fold and 20-fold increase in the risk of dying in the next 10 years, respectively.

Figure 1: Graph showing the distribution of the Charlson Comorbidity Index in the review cohort.

Contents of the ACP

Table 3 summarises the selection of goal-of-care and other directives and preferences in the study cohort. Nearly one-half of the cohort selected Option 3 for their goal-of- care that stipulated care should be focused on symptoms only without any life-prolonging treatment. The majority (80%) did not want CPR to be carried out, 15% wanted resuscitation to be attempted if medically indicated, while 5% expressed no preference. Approximately two-thirds of the cohort had directives excluding a specific treatment. The most preferred place of death was ‘at the residential property’ (40%) compared to palliative care within an ARC or associated hospice (27%), or hospital (6%), while 27% of people had no preference.

Only three people in the cohort had updated their ACPlan, with one no longer wanting to receive CPR, one making changes following discussions with their partner, and one because of a diagnosis of cancer.

Table 3: Preferences stated in the ACP by the study cohort.

Relationship between ACP content and age and level of morbidity

Age

A significant association was observed between increased age and choice of goal-of-care. (Figure 2A). Logistic regression analysis showed that age influenced the choice of goal-of-care 2 (odds ratio [OR] 1.77; p=0.009), 3 (OR 1.93; p=0.001) and 5 (OR 2.49; p=0.04) over goal-of-care 1, but not the choice of goal-of-care 4 (OR 1.49; p=0.13) (Figure 2B). Increasing age also influenced CPR preference, such that a person was 1.7 times more likely to not want CPR than a person 10 years younger (p=0.04) (Figure 2C).

Figure 2: Graphs showing (a) The relationship between increasing age and the choice of goal-of-care and preference regarding CPR, (b) Logistic regression model showing the influence of a step-wise 10-year increase in age on the choice of goal-of-care, and (c) binomial logistic regression model showing the predicted probability for preference not to receive CPR. The small circles represent individual patients, the blue line shows the increasing probability of older people not wanting to be resuscitated (shaded area—95 % confidence intervals).

Figure 2A

Figure 2B

Figure 2C

Comorbidity index

As shown in Figure 3, individuals with a higher CCI and risk of mortality tended to choose goal-of-care 4, in which they wanted health professionals to lead decision making in conjunction with family members or friends. The distribution and median value of CCI were similar for the other four goals-of-care. No association was observed between the CCI and preference regarding CPR.

Figure 3: Graph showing the relationship between increasing severity of comorbidities and the choice of goal-of-care and preference regarding CPR.

Concordance between the ACPlan and AD content and treatment during hospital admissions

A total of 411 admissions to a public hospital since publication of the ACPlan were identified in the study cohort. Twelve percent of these admissions were to the emergency department (ED), with 88% requiring an inpatient stay. Median CCI was similar for the ED and inpatient admissions (3.3 vs. 3.2, respectively).

The majority of people (63%) were deemed competent during admission, 14% partially competent and 11% non-competent. Competence could not be assessed retrospectively in the remaining 12% of individuals. The goal-of-care was adhered to in 59 of the 60 admissions involving non-competent people. Treatment preferences stated in the AD were relevant in seven cases and were adhered to in six cases. The one case of non-concordance involved a patient with an infective exacerbation of COPD who experienced a myocardial infarction. Despite the ACPlan stating life-prolonging care or CPR for a life-threatening chest infection should not be provided, the patient was treated with intravenous fluids and antibiotics. All patients who stated they did not want CPR in the event of a cardiac arrest had their AD visible within the ACP on their electronic medical records. No patient in the study cohort received CPR.

Preferred place of death

Fifty-five people died, with the place of death identified in 44; 21 (48%) dying in their preferred place, 16 (36%) in a non-preferred place, and seven (16%) with no preference stated.

Discussion

The decision to prepare an ACPlan may be influenced by a person’s physical condition and prognosis combined with social factors such as gender, age, cultural and ethnic background, religion and education.2–4 The first aim of this review was to examine the effect of comorbidities as a clinical factor and age as a social factor on the choice of goal-of-care and decision regarding CPR. Our data indicated that age had a measurable influence on a person’s choice of goal-of-care, with older people tending to want to receive only treatments that maintained their comfort and dignity. With increasing age there was a steady decrease in the selection of goal-of-care option that accepted all available life-prolonging treatments, including CPR. In comparison, the severity of comorbidities appeared to have less influence on these decisions, with people with more severe disease tending to want healthcare professionals to make decisions regarding end-of-life care in consultation with family members or friends. This finding is consistent with other studies that reported older adults with multiple conditions were more likely to prefer active decision making with healthcare professionals.20 However, our finding of a lack of association between comorbidities and CPR preference is different from the general finding in the published literature that the severity of disease burden is a major trigger for the decision not to receive resuscitation.21

The second aim of the review was to examine compliance between ACPlans and end-of-life care. This showed that the treatment received during admission to a public hospital was generally consistent with the goals-of-care, while in six of the seven cases involving an incompetent patient, the specific treatment preferences in the AD were adhered to. It is necessary to consider the above findings from the point of view that documentation of care in case notes should not be conflated with the actual care received and that evaluation of concordance may not represent the patient’s actual wishes in the ‘real time’ context of a hospital admission. Reflecting these methodological difficulties, the findings of published studies similar to the current review are characterised by their inconsistency. For example, two recent studies in Australia reported opposite findings, with one showing about one-third of cases received treatment that contradicted preferences listed in the ACPlan,22 while the other found the majority of patients received hospital care fully concordant with their stated preferences.23 Variations in methodology and the make-up of the patients investigated may also contribute to differences between studies.24

Of the people in the cohort who died, approximately 50% died at their preferred place, a finding consistent with local audit data that showed 82% of people with an ACPlan died in a setting such as their home or aged residential care and that the proportion of people in Canterbury over 75 years who die in hospital is 30% compared with the national average of 34%. These findings are in accordance with international evidence that ACP is strongly associated with lower rates of hospital deaths.25

Our review showed ethnicity, and to a lesser extent, socioeconomic status, influenced the make-up of the cohort of people. The over representation of Caucasians and under-representation of minority ethnic groups in the cohort is consistent with previous reports in New Zealand26 and other multicultural countries.27,28 This under-representation may reflect differences in family-centred decision making and attitudes towards life-prolonging treatments and influence of religious beliefs and doctor-patient relationships.29 Our observation of a lower number of ACPlans prepared by people in lower socioeconomic groups is consistent with the findings of a large survey carried out in the US.30

The review had several limitations. The number of people investigated was relatively small and as a consequence we were able to evaluate the concordance between care received and the ACPlan in only seven cases of incompetent patients who had expressed specific treatment preferences. In addition, in approximately 20% of admissions there was insufficient information to determine competence, although in the majority of these cases the admission was for short-term, pre-arranged treatment in a medical day unit. Determining concordance of treatment with the ACP was problematic in some cases as the goal-of-care could be interpreted differently, with the true intention of care provided difficult to assess retrospectively using only clinical records. The demographics of the review cohort were also markedly different than reference populations in census data and local health authority databases, making the generalisability of our findings to the population questionable. Therefore, our results need to be considered after taking these limitations into account, but may provide useful information for future similar studies to identify populations that require further investigation.

We conclude that our review shows that age and to a lesser extent the level of comorbidities may influence the choice of goal-of-care in an ACPlan. In our small cohort of people, admission to a publicly funded hospital for both routine and end-of-life care appeared to adhere to the wishes and preferences stated in the ACP. Our findings support the usefulness of ACPs in that they provide the opportunity for patient choice and self-determination and encourage the wishes of hospitalised patients to be recognised and used to guide care.

Summary

Abstract

Aim

This retrospective review examined the influence of age and severity of comorbidities on goals-of-care in advance care plans (ACPlans) and concordance between these wishes and care received during hospital admission.

Method

The medical records of 149 people with an ACPlan admitted to a public hospital were reviewed to evaluate concordance with treatment. The associations between age and comorbidities and goals-of-care were determined using contingency tables and logistic regression analyses.

Results

The majority of the review cohort were Caucasian and elderly, with people from minority ethnic groups under-represented compared to census data. Increasing age had a measurable influence on the choice of goal-of-care, whereas comorbidity severity had less influence on this decision. In 60 of the 411 hospital admissions the patient was classified as incompetent, with the goal-of-care adhered to in 59 of these cases and treatment preferences adhered to in six of seven cases. Fifty-five people had died since writing their ACPlan, with 63% dying at their preferred place or with no preference stated.

Conclusion

Age and to a lesser extent the severity of comorbidities influence the choice of goal-of-care in an ACPlan. Our review also showed that end-of-life care appeared to adhere to the instructions in the plan.

Author Information

Henricus Berend Speelberg, Medical Student, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Brett Shand, Medical Writer, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Jane Goodwin, Advance Care Planning Facilitator, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Rachel Wiseman, Consultant Physician, Respiratory Medicine, Canterbury District Health Board, Christchurch; Laura Belluscio, Biostatistician, Hospital Services Administration, Canterbury District Health Board, Christchurch; Kate Grundy, Palliative Care Physician, Oncology Services, Canterbury District Health Board, Christchurch.

Acknowledgements

We are grateful for the help provided by the staff at the clinical records warehouse with collection of data and Ma Yi, Biostatistician, Canterbury DHB for assistance with the statistical analyses.

Correspondence

Dr Kate Grundy, Oncology Services, Canterbury District Health Board, Christchurch 8142

Correspondence Email

kate.grundy@cdhb.health.nz

Competing Interests

Nil.

1. New Zealand Ministry of Health. Advance Care Planning. A guide for the New Zealand health care workforce. Accessed on 20 November, 2019 from http://www.health.govt.nz/system/files/documents/publications/advance-care-planning-aug11.pdf

2. Ke LS, Huang X, Hu WY, O'Connor M, Lee S. Experiences and perspectives of older people regarding advance care planning: A meta-synthesis of qualitative studies. Palliat Med. 2017; 31(5):394–405. doi: 10.1177/0269216316663507

3. Jeong S, Ohr S, Pich J, Saul P, Ho A. ‘Planning ahead’ among community dwelling older people from culturally and linguistically diverse background: a cross-sectional study. J Clin Nurs 2015; 24(1–2):244–255.

4. Fakhri S, Engelberg RA, Downey L, et al. Factors Affecting Patients' Preferences for and Actual Discussions About End-of-Life Care. J Pain Symptom Manage. 2016; 52(3):386–394. doi:10.1016/j.jpainsymman.2016.03.012.

5. Boot M, Wilson C. Clinical nurse specialists perspectives on advance care planning conversations: a qualitative study. Int J Palliat Nurs. 2014 Jan; 20(1):9–14.

6. Lund S, Richardson A, May C. Barriers to advance care planning at the end of life: an explanatory systematic review of implementation studies. PLoS One. 2015 Feb.

7. Sharp T, Malyon A, Barclay S. GPs' perceptions of advance care planning with frail and older people: a qualitative study. Br J Gen Pract. 2018; 68(666):e44-e53. doi: 10.3399/bjgp17X694145.

8. Brinkman-Stoppelenburg A, Rietjens JA, van der Heide A. The effects of advance care planning on end-of-life care: a systematic review. Palliat Med. 2014; 28(8):1000–1025. doi: 10.1177/0269216314526272

9. Hardin SB, Yusufaly YA. Difficult end-of-life treatment decisions: do other factors trump advance directives? [published correction appears in Arch Intern Med. 2004 Oct 25; 164(19):2124]. Arch Intern Med. 2004; 164(14):1531–1533. doi:10.1001/archinte.164.14.1531:

10. Hopping-Winn J, Mullin J, March L, et al. The progression of end-of-life wishes and concordance with end-of-life care. J Palliat Med. 2018; 21:541–5. doi: 10.1089/jpm.2017.0317. doi: 10.1089/jpm.2017.0317.

11. Khandelwal N, Curtis JR, Freedman VA, et al. How often Is end-of-life care in the United States inconsistent with patients' goals of care? J Palliat Med. 2017; 20:1400–4. doi: 10.1089/jpm.2017.0065.

12. MacKenzie MA, Smith-Howell E, Bomba PA, Meghani SH. Respecting choices and related models of advance care planning: A systematic review of published evidence. Am J Hosp Palliat Care. 2018; 35:897–907. doi: 10.1177/1049909117745789.

13. Keon-Cohen Z, Myles PS, Story DA. A survey of Australian and New Zealand anaesthetists' attitudes towards resuscitation orders in the perioperative setting. Anaesth Inten Care J. 2017; 45:396–402.

14. New Zealand Ministry of Health. NZDep2013 Index of Deprivation. Accessed on 20 November, 2019 from http://www.health.govt.nz/publication/nzdep2013-index-deprivation

15. Stats NZ. The New Zealand Census of Population and Dwellings, 2013. Accessed on 20 November, 2019 from: http://www.stats.govt.nz/ 2154:188–94. doi: 10.1097/MLR.0000000000000471.

16. Quan H, Li B, Couris, CM, et al. Updating and Validating the Charlson Comorbidity Index and Score for Risk Adjustment in Hospital Discharge Abstracts Using Data From 6 Countries. Am J Epidemiol. 2011; 173:676–82.

17. Austin SR, Wong Y, Uzzo RG, et al. Why summary comorbidity measures such as the Charlson Comorbidity Index and Elixhauser score work. Med Care 2015; 53:e65–e72. doi: 10.1097/MLR.0b013e318297429c

18. Canterbury Initiative. “My advance care plan”. Accessed on 20 November, 2019 from http://www.healthinfo.org.nz/patientinfo/MyAdvanceCarePlan.pdf

19. Teasdale G, Maas A, Lecky F, et al. The Glasgow Coma Scale at 40 years: standing the test of time. Lancet Neurol. 2014; 13:8.

20. Chi WC, Wolff J, Greer R, Dy S. Multimorbidity and Decision-Making Preferences Among Older Adults. Ann Fam Med. 2017; 15:546–551. doi:10.1370/afm.2106

21. Mockford C, Fritz Z, George R, et al. Do not attempt cardiopulmonary resuscitation (DNACPR) orders: a systematic review of the barriers and facilitators of decision-making and implementation. Resuscitation. 2015; 88:99–113. doi:10.1016/j.resuscitation.2014.11.016

22. Johnson SB, Butow PN, Kerridge I, Bell ML, Tattersall MHN. How Well Do Current Measures Assess the Impact of Advance Care Planning on Concordance Between Patient Preferences for End-of-Life Care and the Care Received: A Methodological Review. J Pain Symptom Manage. 2018 Feb; 55(2):480–495. doi: 10.1016/j.jpainsymman.2017.09.008.

23. Panozzo L, Ward B, Harvey P, Fletcher J. Content and implementation of advance care plans: A retrospective cohort study. Aust J Gen Pract. 2019 May; 48(5):323–325. doi: 10.31128/AJGP-12-18-4783.

24. Oo NM, Scott IA, Maggacis R, Rajakaruna N. Assessing concordance between patient preferences in advance care plans and in-hospital care. Aust Health Rev. 2019 Aug; 43(4):425–431. doi: 10.1071/AH18011.

25. Brinkman-Stoppelenburg A, Rietjens JAC, van der Heide A. The effects of advance care planning on end-of-life care: A systematic review, Palliat. Med. 2014; 28:1000–1025.

26. Cawley LM. Racial, cultural, and ethnic factors influencing end-of-life care. J Palliat Med. 2005; 8:S58–S69.

27. Portanova J, Ailshire J, Perez C, Rahman A, Enguidanos S. Ethnic Differences in Advance Directive Completion and Care Preferences: What Has Changed in a Decade?. J Am Geriatr Soc. 2017; 65:1352–1357. doi: 10.1111/jgs.14800.

28. McLaughlin L, Braun K. Asian and Pacific Islander cultural values; considerations for healthcare decision making. Health & Social work. 1998; 23:116–126.

29. Oetzel J, Simpson M, Berryman K, Iti T, Reddy R. Managing communication tensions and challenges during the end-of-life journey. Perspectives of Maori kaumatua and their whanau. Palliat Med. 2015; 29:764–766. doi: 10.1177/0269216315583619.

30. Dobalian A. Advance care planning documents in nursing facilities: results from a nationally representative survey. Arch Gerontol Geriatr. 2006; 43:193–212. doi:10.1016/j.archger.2005.10.007.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Advance care planning (ACP) is defined in New Zealand as the process of thinking about, talking about and planning for future healthcare and end-of-life care and identifying what is important within this care.1 The ACP process provides a person with the opportunity to document their wishes in an advance care plan (ACPlan), which may include an advance directive (AD) component to accept or refuse specific treatments that may be offered in the future. This enables the wishes of hospitalised patients to be recognised during end-of-life care if and when they lose competence. In New Zealand, clinicians are responsible for assessing whether a treatment is medically indicated. This decision takes into account the patient’s condition, projected outcome after treatment and consultation with family members.1 If treatment is medically indicated, the ACP is used to help decide whether or not to proceed.

One objective of the review described in this paper was to examine the influence of age and comorbidities on the content of ACPlans and ADs. The content of an ACPlan or AD and the expectations of care to be provided may reflect a complex multitude of factors such as the health status and prognosis of a person, expected quality of life, involvement of family members and a diversity of religious, spiritual and cultural beliefs.2–4 A person’s understanding of their clinical situation and perceptions regarding the balance between purported benefit of treatment and potential harm may also influence content.5 For patients who are seriously ill, ACP is often conducted under compromised or time constrained conditions.6 On occasions, the ACPlan may represent a compromise between a person’s expectations and the clinical resources available and can be influenced by the time the plan was prepared and the person’s relationship with the health professional assisting with preparation of the ACPlan.7

The second objective of the review was to evaluate the concordance between the wishes and directives in the plans and the care received when admitted to a public hospital. When a person writes an ACPlan they do so with the belief that their wishes and instructions will be complied with during care. Studies that have investigated the influence of ACP on end-of-life care have shown that it can reduce hospitalisation rates and increase hospice and palliative care,8 although the care received is not always consistent with a person’s preferences.9–12 This may occur in difficult clinical situations where other factors such as prognosis, perceived quality of life and the wishes of family or friends may influence compliance with the ACPlan or AD. For example, a survey of the attitudes of Australian and New Zealand anaesthetists towards resuscitation orders in the perioperative setting showed that while the majority recognised the importance of an AD, only 45% followed the directives in clinical practice.13

Methods

A project to promote ACP was introduced in Canterbury in December 2013 supported by the Canterbury Initiative, Canterbury District Health Board (DHB) (https://www.cdhb.health.nz/about-us/key-projects-and-initiatives/canterbury-initiative/). There has been a steady increase in the number of ACPlans created and shared electronically across primary and secondary healthcare that are now available immediately to medical staff in public hospitals on admission.

The current review was carried out as a student project under the conditions of an out-of-scope ethical approval granted by the Ministry of Health and Disability Ethics Committee to undertake routine evaluation of services such as ACP that are included on Community HealthPathways (http://edu.cdhb.health.nz/Hospitals-Services/Health-Professionals/Pages/Health-Pathways.aspx), a clinical guidance website for health professionals, and HealthInfo (https://www.healthinfo.org.nz/), an open-access website that provides health information for the general population.

The study involved a retrospective review of case notes of individuals who had published an ACPlan and subsequently received care in a publicly funded hospital. A published ACPlan was defined as those that had passed a two-step review process to ensure clinical interpretability and had then been loaded onto the person’s electronic health record. A total of 1,939 ACPs were published in Canterbury in the three-year period between 10 September 2014 to 31 September 2017. These ACPs were matched to hospital admissions recorded on the Canterbury DHB clinical database to identify people who had received care in a publicly funded hospital up to the end of September 2017 (n=667). One hundred and fifty people were selected randomly from this group for inclusion in the review, with this number taking into account the relatively short duration of the student project. A review of their electronic and written clinical records was carried out by the student (HBS), with all admissions involving an incompetent patient and a random subset of about 10% of other admissions reviewed independently by senior clinicians (RW and KG) to check concordance between reviewers. The following data were recorded.

1. Demographic variables: gender, age, ethnicity, residential location and deprivation index (NZDep13)14 of the suburb in which the individual lived, using the classifications of the 2013 New Zealand census.15

2. Clinical variables: medical diagnoses made before and six months after publication of the ACPlan, and medical condition assessed by the Quan age-adjusted modification of the Charlson Comorbidity Index (CCI).16 This algorithm is based on 22 conditions and predicts 10-year mortality and was chosen because of proven validity in a wide range of populations.17

3. Content of the ACPlan: all patients prepared their ACPlan using the same template18 that contained the following information and instructions. Date published, goal-of-care and specific treatment preferences in the event of future deterioration in health (see Table 1), instructions for cardiopulmonary resuscitation (CPR), preferred place of death, choices for organ donation and changes in content of the ACPlan due to progression of illness and/or change in insight of the implications of the disease.

4. Assessment of competence: competence at the time of each admission was categorised as either competent, incompetent or fluctuating competence. Competence was defined as the ability of the patient to understand the need to make a decision, process the information presented and communicate their decisions regarding treatment. This assessment was based on entries in the clinical notes of the patient that documented level of alertness and orientation and/or the level of consciousness (Glasgow Coma Scale),19 with signs of incompetence including confusion, delirium and progressive dementia.

5. Implementation of the ACPlan: the treatment received during each hospital admission for the review cohort was reviewed to determine concordance with the goal-of-care. This involved, when available, recording the investigations and treatment received and clinical outcomes of all the admissions from electronic clinical, radiological and laboratory records and written case notes. Concordance between clinical management of incompetent patients and the contents of the ACPlan, AD and CPR preference was reviewed separately, with final evaluation of the findings carried out by RW and KG.

6. Place of death: where possible, the place of death was recorded for comparison with the person’s preferred option.

Table 1: Content of the ACP including selection of goal-of-care, treatment preferences, preferred place of death and organ donation.

Data analysis

The frequencies of categorical variables in the cohort were expressed as number and percentage. Age was expressed as median and interquartile range because of the predominance of elderly people in the cohort. The chi-squared goodness-of-fit test was used to examine whether the observed frequency of demographic variables of the study cohort differed from those in the 2013 census of the New Zealand population24 and Canterbury DHB-Primary Health Organisation population.

To examine the influence of age or severity of comorbidities on the choice of goal-of-care and CPR preference stated in the ACP or AD the data were analysed using the chi square test or Fisher’s exact test. A logistical regression model was constructed to examine the relationship between changes in the choice of goal-of-care as the outcome variable and increasing age or severity of comorbidities as the explanatory variable, and between the preference to not receive CPR (outcome) and age (predictor). The level of statistical significance for all the analyses was set at p≤0.05.

Results

Demographic data

The demographic and clinical characteristics of the study cohort are shown in Table 2. One person was excluded because of a lack of adequate information in their medical records, leaving 149 people in the analysis set. The cohort was comprised of mainly elderly people (median age, 78 years), with 15% living in an aged residential care (ARC) facility. There were significantly more females than males in the cohort (57% vs 43%, p<0.01). Caucasians were over-represented (94%) compared with that in census data (81%), with a significantly reduced proportion of all other ethnicities (p=0.02). There was also a marked under-representation of the most socioeconomically deprived people and a marginally lower proportion of the least deprived individuals compared to census data (p<0.01). As shown in Figure 1, 14% of the study cohort had no comorbidities and 57% had a relatively low risk of 10-year mortality, ie, CCI ≤2. The incidence of medical conditions included in the CCI in the study cohort were cancer 40%, respiratory disorders 34%, cardiovascular disease 19%, renal disease 19%, rheumatologic disease 18%, dementia 6%, diabetes mellitus 4%, hemi- or paraplegia 3%, and liver disease 2%.

Table 2: Demographic and clinical characteristics of the study cohort.

*p value for comparison of study cohort and reference populations.
§Canterbury DHB-Primary Health Organisation
Index (0–8) predicts mortality, with scores between 1–2 and ≥5 associated with a 5-fold and 20-fold increase in the risk of dying in the next 10 years, respectively.

Figure 1: Graph showing the distribution of the Charlson Comorbidity Index in the review cohort.

Contents of the ACP

Table 3 summarises the selection of goal-of-care and other directives and preferences in the study cohort. Nearly one-half of the cohort selected Option 3 for their goal-of- care that stipulated care should be focused on symptoms only without any life-prolonging treatment. The majority (80%) did not want CPR to be carried out, 15% wanted resuscitation to be attempted if medically indicated, while 5% expressed no preference. Approximately two-thirds of the cohort had directives excluding a specific treatment. The most preferred place of death was ‘at the residential property’ (40%) compared to palliative care within an ARC or associated hospice (27%), or hospital (6%), while 27% of people had no preference.

Only three people in the cohort had updated their ACPlan, with one no longer wanting to receive CPR, one making changes following discussions with their partner, and one because of a diagnosis of cancer.

Table 3: Preferences stated in the ACP by the study cohort.

Relationship between ACP content and age and level of morbidity

Age

A significant association was observed between increased age and choice of goal-of-care. (Figure 2A). Logistic regression analysis showed that age influenced the choice of goal-of-care 2 (odds ratio [OR] 1.77; p=0.009), 3 (OR 1.93; p=0.001) and 5 (OR 2.49; p=0.04) over goal-of-care 1, but not the choice of goal-of-care 4 (OR 1.49; p=0.13) (Figure 2B). Increasing age also influenced CPR preference, such that a person was 1.7 times more likely to not want CPR than a person 10 years younger (p=0.04) (Figure 2C).

Figure 2: Graphs showing (a) The relationship between increasing age and the choice of goal-of-care and preference regarding CPR, (b) Logistic regression model showing the influence of a step-wise 10-year increase in age on the choice of goal-of-care, and (c) binomial logistic regression model showing the predicted probability for preference not to receive CPR. The small circles represent individual patients, the blue line shows the increasing probability of older people not wanting to be resuscitated (shaded area—95 % confidence intervals).

Figure 2A

Figure 2B

Figure 2C

Comorbidity index

As shown in Figure 3, individuals with a higher CCI and risk of mortality tended to choose goal-of-care 4, in which they wanted health professionals to lead decision making in conjunction with family members or friends. The distribution and median value of CCI were similar for the other four goals-of-care. No association was observed between the CCI and preference regarding CPR.

Figure 3: Graph showing the relationship between increasing severity of comorbidities and the choice of goal-of-care and preference regarding CPR.

Concordance between the ACPlan and AD content and treatment during hospital admissions

A total of 411 admissions to a public hospital since publication of the ACPlan were identified in the study cohort. Twelve percent of these admissions were to the emergency department (ED), with 88% requiring an inpatient stay. Median CCI was similar for the ED and inpatient admissions (3.3 vs. 3.2, respectively).

The majority of people (63%) were deemed competent during admission, 14% partially competent and 11% non-competent. Competence could not be assessed retrospectively in the remaining 12% of individuals. The goal-of-care was adhered to in 59 of the 60 admissions involving non-competent people. Treatment preferences stated in the AD were relevant in seven cases and were adhered to in six cases. The one case of non-concordance involved a patient with an infective exacerbation of COPD who experienced a myocardial infarction. Despite the ACPlan stating life-prolonging care or CPR for a life-threatening chest infection should not be provided, the patient was treated with intravenous fluids and antibiotics. All patients who stated they did not want CPR in the event of a cardiac arrest had their AD visible within the ACP on their electronic medical records. No patient in the study cohort received CPR.

Preferred place of death

Fifty-five people died, with the place of death identified in 44; 21 (48%) dying in their preferred place, 16 (36%) in a non-preferred place, and seven (16%) with no preference stated.

Discussion

The decision to prepare an ACPlan may be influenced by a person’s physical condition and prognosis combined with social factors such as gender, age, cultural and ethnic background, religion and education.2–4 The first aim of this review was to examine the effect of comorbidities as a clinical factor and age as a social factor on the choice of goal-of-care and decision regarding CPR. Our data indicated that age had a measurable influence on a person’s choice of goal-of-care, with older people tending to want to receive only treatments that maintained their comfort and dignity. With increasing age there was a steady decrease in the selection of goal-of-care option that accepted all available life-prolonging treatments, including CPR. In comparison, the severity of comorbidities appeared to have less influence on these decisions, with people with more severe disease tending to want healthcare professionals to make decisions regarding end-of-life care in consultation with family members or friends. This finding is consistent with other studies that reported older adults with multiple conditions were more likely to prefer active decision making with healthcare professionals.20 However, our finding of a lack of association between comorbidities and CPR preference is different from the general finding in the published literature that the severity of disease burden is a major trigger for the decision not to receive resuscitation.21

The second aim of the review was to examine compliance between ACPlans and end-of-life care. This showed that the treatment received during admission to a public hospital was generally consistent with the goals-of-care, while in six of the seven cases involving an incompetent patient, the specific treatment preferences in the AD were adhered to. It is necessary to consider the above findings from the point of view that documentation of care in case notes should not be conflated with the actual care received and that evaluation of concordance may not represent the patient’s actual wishes in the ‘real time’ context of a hospital admission. Reflecting these methodological difficulties, the findings of published studies similar to the current review are characterised by their inconsistency. For example, two recent studies in Australia reported opposite findings, with one showing about one-third of cases received treatment that contradicted preferences listed in the ACPlan,22 while the other found the majority of patients received hospital care fully concordant with their stated preferences.23 Variations in methodology and the make-up of the patients investigated may also contribute to differences between studies.24

Of the people in the cohort who died, approximately 50% died at their preferred place, a finding consistent with local audit data that showed 82% of people with an ACPlan died in a setting such as their home or aged residential care and that the proportion of people in Canterbury over 75 years who die in hospital is 30% compared with the national average of 34%. These findings are in accordance with international evidence that ACP is strongly associated with lower rates of hospital deaths.25

Our review showed ethnicity, and to a lesser extent, socioeconomic status, influenced the make-up of the cohort of people. The over representation of Caucasians and under-representation of minority ethnic groups in the cohort is consistent with previous reports in New Zealand26 and other multicultural countries.27,28 This under-representation may reflect differences in family-centred decision making and attitudes towards life-prolonging treatments and influence of religious beliefs and doctor-patient relationships.29 Our observation of a lower number of ACPlans prepared by people in lower socioeconomic groups is consistent with the findings of a large survey carried out in the US.30

The review had several limitations. The number of people investigated was relatively small and as a consequence we were able to evaluate the concordance between care received and the ACPlan in only seven cases of incompetent patients who had expressed specific treatment preferences. In addition, in approximately 20% of admissions there was insufficient information to determine competence, although in the majority of these cases the admission was for short-term, pre-arranged treatment in a medical day unit. Determining concordance of treatment with the ACP was problematic in some cases as the goal-of-care could be interpreted differently, with the true intention of care provided difficult to assess retrospectively using only clinical records. The demographics of the review cohort were also markedly different than reference populations in census data and local health authority databases, making the generalisability of our findings to the population questionable. Therefore, our results need to be considered after taking these limitations into account, but may provide useful information for future similar studies to identify populations that require further investigation.

We conclude that our review shows that age and to a lesser extent the level of comorbidities may influence the choice of goal-of-care in an ACPlan. In our small cohort of people, admission to a publicly funded hospital for both routine and end-of-life care appeared to adhere to the wishes and preferences stated in the ACP. Our findings support the usefulness of ACPs in that they provide the opportunity for patient choice and self-determination and encourage the wishes of hospitalised patients to be recognised and used to guide care.

Summary

Abstract

Aim

This retrospective review examined the influence of age and severity of comorbidities on goals-of-care in advance care plans (ACPlans) and concordance between these wishes and care received during hospital admission.

Method

The medical records of 149 people with an ACPlan admitted to a public hospital were reviewed to evaluate concordance with treatment. The associations between age and comorbidities and goals-of-care were determined using contingency tables and logistic regression analyses.

Results

The majority of the review cohort were Caucasian and elderly, with people from minority ethnic groups under-represented compared to census data. Increasing age had a measurable influence on the choice of goal-of-care, whereas comorbidity severity had less influence on this decision. In 60 of the 411 hospital admissions the patient was classified as incompetent, with the goal-of-care adhered to in 59 of these cases and treatment preferences adhered to in six of seven cases. Fifty-five people had died since writing their ACPlan, with 63% dying at their preferred place or with no preference stated.

Conclusion

Age and to a lesser extent the severity of comorbidities influence the choice of goal-of-care in an ACPlan. Our review also showed that end-of-life care appeared to adhere to the instructions in the plan.

Author Information

Henricus Berend Speelberg, Medical Student, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Brett Shand, Medical Writer, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Jane Goodwin, Advance Care Planning Facilitator, The Canterbury Initiative, Canterbury District Health Board, Christchurch; Rachel Wiseman, Consultant Physician, Respiratory Medicine, Canterbury District Health Board, Christchurch; Laura Belluscio, Biostatistician, Hospital Services Administration, Canterbury District Health Board, Christchurch; Kate Grundy, Palliative Care Physician, Oncology Services, Canterbury District Health Board, Christchurch.

Acknowledgements

We are grateful for the help provided by the staff at the clinical records warehouse with collection of data and Ma Yi, Biostatistician, Canterbury DHB for assistance with the statistical analyses.

Correspondence

Dr Kate Grundy, Oncology Services, Canterbury District Health Board, Christchurch 8142

Correspondence Email

kate.grundy@cdhb.health.nz

Competing Interests

Nil.

1. New Zealand Ministry of Health. Advance Care Planning. A guide for the New Zealand health care workforce. Accessed on 20 November, 2019 from http://www.health.govt.nz/system/files/documents/publications/advance-care-planning-aug11.pdf

2. Ke LS, Huang X, Hu WY, O'Connor M, Lee S. Experiences and perspectives of older people regarding advance care planning: A meta-synthesis of qualitative studies. Palliat Med. 2017; 31(5):394–405. doi: 10.1177/0269216316663507

3. Jeong S, Ohr S, Pich J, Saul P, Ho A. ‘Planning ahead’ among community dwelling older people from culturally and linguistically diverse background: a cross-sectional study. J Clin Nurs 2015; 24(1–2):244–255.

4. Fakhri S, Engelberg RA, Downey L, et al. Factors Affecting Patients' Preferences for and Actual Discussions About End-of-Life Care. J Pain Symptom Manage. 2016; 52(3):386–394. doi:10.1016/j.jpainsymman.2016.03.012.

5. Boot M, Wilson C. Clinical nurse specialists perspectives on advance care planning conversations: a qualitative study. Int J Palliat Nurs. 2014 Jan; 20(1):9–14.

6. Lund S, Richardson A, May C. Barriers to advance care planning at the end of life: an explanatory systematic review of implementation studies. PLoS One. 2015 Feb.

7. Sharp T, Malyon A, Barclay S. GPs' perceptions of advance care planning with frail and older people: a qualitative study. Br J Gen Pract. 2018; 68(666):e44-e53. doi: 10.3399/bjgp17X694145.

8. Brinkman-Stoppelenburg A, Rietjens JA, van der Heide A. The effects of advance care planning on end-of-life care: a systematic review. Palliat Med. 2014; 28(8):1000–1025. doi: 10.1177/0269216314526272

9. Hardin SB, Yusufaly YA. Difficult end-of-life treatment decisions: do other factors trump advance directives? [published correction appears in Arch Intern Med. 2004 Oct 25; 164(19):2124]. Arch Intern Med. 2004; 164(14):1531–1533. doi:10.1001/archinte.164.14.1531:

10. Hopping-Winn J, Mullin J, March L, et al. The progression of end-of-life wishes and concordance with end-of-life care. J Palliat Med. 2018; 21:541–5. doi: 10.1089/jpm.2017.0317. doi: 10.1089/jpm.2017.0317.

11. Khandelwal N, Curtis JR, Freedman VA, et al. How often Is end-of-life care in the United States inconsistent with patients' goals of care? J Palliat Med. 2017; 20:1400–4. doi: 10.1089/jpm.2017.0065.

12. MacKenzie MA, Smith-Howell E, Bomba PA, Meghani SH. Respecting choices and related models of advance care planning: A systematic review of published evidence. Am J Hosp Palliat Care. 2018; 35:897–907. doi: 10.1177/1049909117745789.

13. Keon-Cohen Z, Myles PS, Story DA. A survey of Australian and New Zealand anaesthetists' attitudes towards resuscitation orders in the perioperative setting. Anaesth Inten Care J. 2017; 45:396–402.

14. New Zealand Ministry of Health. NZDep2013 Index of Deprivation. Accessed on 20 November, 2019 from http://www.health.govt.nz/publication/nzdep2013-index-deprivation

15. Stats NZ. The New Zealand Census of Population and Dwellings, 2013. Accessed on 20 November, 2019 from: http://www.stats.govt.nz/ 2154:188–94. doi: 10.1097/MLR.0000000000000471.

16. Quan H, Li B, Couris, CM, et al. Updating and Validating the Charlson Comorbidity Index and Score for Risk Adjustment in Hospital Discharge Abstracts Using Data From 6 Countries. Am J Epidemiol. 2011; 173:676–82.

17. Austin SR, Wong Y, Uzzo RG, et al. Why summary comorbidity measures such as the Charlson Comorbidity Index and Elixhauser score work. Med Care 2015; 53:e65–e72. doi: 10.1097/MLR.0b013e318297429c

18. Canterbury Initiative. “My advance care plan”. Accessed on 20 November, 2019 from http://www.healthinfo.org.nz/patientinfo/MyAdvanceCarePlan.pdf

19. Teasdale G, Maas A, Lecky F, et al. The Glasgow Coma Scale at 40 years: standing the test of time. Lancet Neurol. 2014; 13:8.

20. Chi WC, Wolff J, Greer R, Dy S. Multimorbidity and Decision-Making Preferences Among Older Adults. Ann Fam Med. 2017; 15:546–551. doi:10.1370/afm.2106

21. Mockford C, Fritz Z, George R, et al. Do not attempt cardiopulmonary resuscitation (DNACPR) orders: a systematic review of the barriers and facilitators of decision-making and implementation. Resuscitation. 2015; 88:99–113. doi:10.1016/j.resuscitation.2014.11.016

22. Johnson SB, Butow PN, Kerridge I, Bell ML, Tattersall MHN. How Well Do Current Measures Assess the Impact of Advance Care Planning on Concordance Between Patient Preferences for End-of-Life Care and the Care Received: A Methodological Review. J Pain Symptom Manage. 2018 Feb; 55(2):480–495. doi: 10.1016/j.jpainsymman.2017.09.008.

23. Panozzo L, Ward B, Harvey P, Fletcher J. Content and implementation of advance care plans: A retrospective cohort study. Aust J Gen Pract. 2019 May; 48(5):323–325. doi: 10.31128/AJGP-12-18-4783.

24. Oo NM, Scott IA, Maggacis R, Rajakaruna N. Assessing concordance between patient preferences in advance care plans and in-hospital care. Aust Health Rev. 2019 Aug; 43(4):425–431. doi: 10.1071/AH18011.

25. Brinkman-Stoppelenburg A, Rietjens JAC, van der Heide A. The effects of advance care planning on end-of-life care: A systematic review, Palliat. Med. 2014; 28:1000–1025.

26. Cawley LM. Racial, cultural, and ethnic factors influencing end-of-life care. J Palliat Med. 2005; 8:S58–S69.

27. Portanova J, Ailshire J, Perez C, Rahman A, Enguidanos S. Ethnic Differences in Advance Directive Completion and Care Preferences: What Has Changed in a Decade?. J Am Geriatr Soc. 2017; 65:1352–1357. doi: 10.1111/jgs.14800.

28. McLaughlin L, Braun K. Asian and Pacific Islander cultural values; considerations for healthcare decision making. Health & Social work. 1998; 23:116–126.

29. Oetzel J, Simpson M, Berryman K, Iti T, Reddy R. Managing communication tensions and challenges during the end-of-life journey. Perspectives of Maori kaumatua and their whanau. Palliat Med. 2015; 29:764–766. doi: 10.1177/0269216315583619.

30. Dobalian A. Advance care planning documents in nursing facilities: results from a nationally representative survey. Arch Gerontol Geriatr. 2006; 43:193–212. doi:10.1016/j.archger.2005.10.007.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE