Planning for end of life care with older people with diabetes: hiding death in metaphors

Life expectancy is 12 years less than the general population for people with type 1 diabetes (T1D), and 8 years less for people with type 2 diabetes (T2D); especially if diabetes was diagnosed before age 40.⁶ Diabetes is linked to more than 3.7 million deaths per year, globally.⁷ In Australia, 11% of deaths and over 1 million hospital admissions were due to diabetes as a principal or associated diagnosis in 2016-2017. ⁸ In addition, most Australians who die from cardiovascular disease, cancer, stroke and renal disease have diabetes as an underlying or associated cause of death. ⁹ These data suggest diabetes is a life limiting illness (LLI), although the progression to end of life usually occurs over many years.

Life limiting illness

The term LLI describes people at high risk of dying in the subsequent 12-months.^{10, 11} Many people with an LLI are older and more likely to be admitted to hospital and/or intensive care, often through the emergency department. Many would derive more benefit from palliative care than intensive interventons.^{2, 3, 10} It could be argued that everybody with an LLI should have an Advance Care Directive (ACD).

The ‘initial’ onset of diabetes, especially T2D,and its comorbidities that reduce life expectancy, can be difficult to determine, although they often have a common underlying inflammatory process and obesity.^{6, 7}  The Gold Standards Framework (GSF) ¹¹ and other guidance recommend people document their end of life preferences while they have decisional capacity.¹² However, less than50% of people with a LLI have documented goals of care while in hospital ¹³ and many do not receive palliative care.¹⁴

Palliative care

Palliative care is an important care option at any time. It is particularly helpful for older people, including those with diabetes, who have functional changes and live with significant disease and other burdens that affect quality of life (QOL) by relieving suffering and managing distressing symptoms. ¹⁴ Palliative care emphasises the importance of good communication to support individuals to make informed decisions. ^{14, 15} ‘Good communication’ encompasses oral language, body language and written information. Usually in a conversation people  take turns to speak and listen and ask clarifying or probing questions. Listening is essential to good conversations.¹⁵

Palliative care can be used at any time and in any disease trajectory (see Table 1) and can complement usual diabetes care and continue to the end of life. ^{14, 15}Most benefit occurs when palliative care is introduced early in people with cancer, ^{16, 17}and likely in people with diabetes. Benefits of palliative care include improved comfort and function, less unnecessary treatment, fewer care burdens, increased satisfaction with care and sometimes, increased life expectancy.¹⁴ Many older people with diabetes could benefit from palliative care in addition to their usual diabetes care, and from support to document an ACD, much earlier than occurs in current practice.^{14, 55, 18}

Terminal and end of life

The last 12 months of life is generally referred to as the terminal stage. ^{10, 11.14} End of life refers to death expected in a few hours or days,^{10, 11, 14} although sudden, unexpected death occurs in ~ 25% of deaths.¹⁸ The GSF encompasses the ‘surprise question:’ ‘would I be surprised if this person died soon?’¹¹ The answer, yes or no, can be useful to guide care decisions towards those with likely benefits and the fewest risks and when to implement the individual’s ACD instructions or initiate conversations if they do not have an ACD. Thus, an ACD increases the chance the individual will have ‘a good death.’

People generally want ‘a good death’ ^{18, 19 20}The definition of ‘a good death’ is highly individual. It usually encompasses the person having some idea when death will occur to enable them and their families to plan their rest of life. Research shows people value having control over:

• their dying
• who is present at their death
• where they die – many want to die at home but  less than40% actually do.^{19 20, 21}
• their comfort e.g. being free from pain
• having time to complete unfinished business and time to say good bye to the people and pets who are important to them.¹⁹
• dying with dignity.^{14, 15}

The decision to administer or not administer interventions that do not improve QOL can be difficult for clinicians and has cost implications. People have a right to die with dignity but dignity is difficult to define and clinicians and families’ wishes often differ from the individual’s wishes, which lead to conflict and distress. ^{18, 22,   23}

Advance care planning

People admitted to hospital near their end of life are more likely to receive burdensome treatment such as resuscitation, dialysis and blood transfusions that may not be beneficial and may be futile, if they do not have an ACD. ^{18, 22 , 23} Such care is often distressing for the individual and their relatives, and for clinicians. Significantly, care that prolongs life may not accord with the individual’s values or quality of life: most people prefer QOL over longer life. ²⁴

Helping older people with diabetes plan for predictable changes in their health status and initiating timely conversations about ACD are key clinician responsibilities. These conversations include discussing the benefits of including palliative care in the individual’s current personalised diabetes care, sick day care and education plans and ACD. As indicated previously, lack of clarity about individual’s wishes can cause emotional distress for all concerned.

It is difficult and stressful for clinicians and family to make care decisions when they do not know the individual’s values and care preferences. These situations often lead to conflict and/or moral (decisional) distress and uncertainty for families and clinicians. ²⁵Moral distress and uncertainty is significantly reduced when individuals have an ACD. Key dilemmas that cause uncertainty and moral distress for clinicians as people near their end of life include:

• giving people ‘bad news,’ and ‘talking about death’
• discussing ACDs
• managing blood glucose and HbA1c g. deciding when and how to change the focus from  ‘tight blood glucose management’ to a focus on comfort and QOL without making  the individual/family feel clinicians are ‘giving up on them’ and reducing hope
• dealing with family conflict over treatment decisions
• withdrawing glucose lowering medicines and other treatment.^{18, 26}

A clearly documented ACD can significantly reduce uncertainty about many of these dilemmas and appropriate education can help clinicians learn how to discuss ‘bad news.’ Thus, it is important to encourage older people with diabetes to document their values, the things that give meaning and purpose to their life, the care they want to receive, and the care they do not want, in an ACD. ACDs can be seen as an outcome of clear open communication among the individual, their families and clinicians. It does not need to happen all at once, particularly if it is commenced early. ACDs enable medical treatment decision makers, family and clinicians to make the decisions in the individual’s best interests, when needed.^{12, 18, 22, 26}

However, ACDs are often first documented when the individual has a Medical Emergency Team (MET) call to assess sudden deterioration in hospital.²⁷ Significantly, 20% – 30% of people who die in hospital receive a MET call and 20% – 35% of MET calls result in death.²⁸ A clearly documented ACD can help the MET teams and other carers decide whether to actively treat the episode of deterioration, whether palliative care could be implemented to reduce suffering, or whether the person is dying and end of life care is required.

Older people with diabetes are likely to have multiple comorbidities and many die in hospital. They often have many episodes of deterioration and recovery, especially in the terminal stage of life, which is typical of the chronic disease trajectory. ^{2, 3} Deterioration can be triggered by inter-current illnesses or be due to preventable and/or treatable hypoglycaemia, ketoacidosis or hyperosmolar states. Episodes of deterioration and hospital admissions increase yearly and often lead to admission to an aged care facility.²⁸ They are a poor prognostic sign.

Diabetes clinicians and general practitioners have many opportunities to identify when older people with diabetes could benefit from palliative care and initiate conversations about ACP. Every clinical encounter represents an opportunity to initiate ACD discussions.²⁹ Significantly, the focus of care needs to change from tight blood glucose management to prevent complications to managing complications to promote comfort and QOL: that is, palliative care.

Many opportunities to begin ACD conversations are missed because many clinicians are reluctant to discuss prognosis and dying.³⁰ despite people with diabetes’ willingness to discuss these issues. ^{31, 32} Consequently, many people with diabetes who could benefit from palliative care are not identified and/or not referred to palliative care specialists or to experts in ACP.³³ Significantly, the National Safety and Quality Health Service Standards (NSQHS) Standards (2017)³⁴ encompass the need to identify people’s values and end of life care needs by engaging them in shared decision making, as key care standards.

Our research suggests many people with diabetes want to discuss palliative and end of life care and ACP but are reluctant to initiate conversations because they do not want to upset their clinicians and because clinicians do not initiate such conversations.^{15, 35} Other research shows doctors wait for the individual or family to raise the issue,  the individual and family waits for the doctor to do so, and consequently nobody does.¹⁷

Aim

The paper reports the outcome of focus groups conducted as the external stakeholder consultation phase of a larger study that developed and evaluated tailored information designed to facilitate ACP.^{15, 35} The information contained the words death, dying and dead and was developed with an advisory group of older people with diabetes and family members. The draft information was then discussed with consultative groups of people with diabetes and families who were not part of the advisory group, and clinicians.

Good communication is essential in all health encounters, especially when discussing emotive topics like death. Thus, the focus groups aimed to determine:

• The words/language older people with diabetes, families and clinicians use or prefer to use to discuss dying, end of life and ACP.
• Compare people with diabetes/families’ preferred words/language with clinicians’ language.
• Describe clinician’s experience of ACP.

Method

The method used to collect the data and guidelines for managing diabetes when people need palliative and end of life are described in other publications.^{15, 35} The current paper reports the findings concerning the words and language used to hide death in metaphors that emerged in the focus groups undertaken with purposive samples of:

a) older people with diabetes and family members from community settings using flyers placed in relevant settings

b) diabetes, geriatric and palliative care clinicians recruited using advertisements distributed to potential participants through electronic emailed newsletter, and news bulletins of relevant clinician professional associations e.g. Australian Diabetes Educators Association, Palliative Care Australia and the Royal Australian College of General Practitioners.

The group discussion was semi structured and used probing and clarifying questions when indicated, consistent with the Focus Group Toolbox guidelines.³⁶ The discussion was audio recorded, with permission, and transcribed verbatim. An independent expert in focus groups facilitated the discussion to limit researcher bias. Field notes were compiled by a scribe during the focus groups and used to summarise the main points of the conversation to participants towards the end of the focus group to reach consensus and make sure key issues were captured. The key focus of the data analysis for the current paper was to identify the words and language participants used to discuss death.

Data analysis

The transcripts and field note data were combined for analysis. Data analysis was performed independently by two researchers using Spencer and Richie’s Framework method.³⁷ The transcripts were read several times to enable the researchers to become immersed in and familiar with the data, and identify codes and then themes and subthemes. The researchers independently undertook the analysis and then met to discuss their findings, reach consensus and ensure the codes and themes were saturated.

Barwon Health and Deakin University Human Research Ethic Committees approved the study.

Findings

Sixty-two people participated: 21 older people (3 groups with 8, 7 and 6 participants), 4 family members in one group), 30 diabetes clinicians and 7 palliative care clinicians (mixed groups 10,10, 11, 6 participants). All clinicians were female: 6 males participated in the older people group and one male in the family group. Most diabetes clinicians had little experience with and no education to undertake ACP, whereas palliative care clinicians were more familiar with discussing changing life expectancy and ACP. Most older people with diabetes and family had witnessed the death of family and friends: ‘we’re in that age group so we’ve had some experience of death.’

There were differences between the language clinicians and older people with diabetes/families preferred. Most diabetes clinicians indicated they preferred soft words like slowing down and gone. In comparison most older people, family members and palliative care clinicians indicated they were ‘comfortable’ using words such as death, dead and dying, when relevant. Most diabetes clinicians indicated they were uncomfortable discussing death and using ‘those words.’ Thus, hiding death in metaphors emerged as the main theme and had two subthemes: the need to use clear language and the need to find common ground.

Language of death – hiding death in metaphors

Most diabetes clinicians preferred ‘soft’ words and were uncomfortable using ‘death words’ because ‘they are harsh,’ ‘confrontational’ and ‘likely to upset patients.’ They indicated their training did not equip them to have ACP conversations or discuss prognosis, death and dying: that is giving ‘bad news’. Thus, many avoided having ACD conversations and when they did, they ‘hid death in metaphors such as ‘gone,’ ‘past’ and ‘passed away.’

One geriatrician suggested:

Clinicians avoid saying ‘death’ and ‘dying’ as a way of psychologically distancing themselves [from discussing death] so they soften the blow by using metaphors to ‘protect’ the individual and/or their family, and most likely themselves, from the reality.

Clinicians said they found it more difficult to use ‘dying’ and ‘died/dead’ when people were receiving curative treatment. They indicated that the ‘language of death’ was inappropriate when people were not dying, and that ‘it would be detrimental to or diminish hope for them and the family’ ‘to speak about death’. Some clinicians found it easier to use ‘straight talk’ rather than death metaphors when the person was receiving end of life care and clearly dying.

Most older people with diabetes and the four family members indicated they preferred clinicians to use clear, unambiguous language to discuss death to avoid misunderstanding. They suggested soft words and metaphors were ambiguous and caused confusion and miscommunication. One family member said she would have appreciated clear language about her parent’s health status when she knew they were dying:

When they’re dying you don’t want the language to be too sensitive. I’m not a doctor or a nurse but both my parents died soon after each other. I wanted the doctors to use clear language to hear ‘this is the situation. They’re dying’. I needed the truth and some certainty’

Clinicians’ reluctance to use clear language and have open conversations about dying resulted in uncertainty, miscommunication and lost opportunities for them to discuss and educate people about, possibly, their most significant life event. Further, older people with diabetes and family members indicated being able to talk about death with clinicians was important as they got older: one said it was an ‘important part of my care,’ which elicited nods of agreement from the group. One man said he wanted clinicians ‘to be present and listen to my concerns’ in clinical encounters. He said:

I feel it is important for them [clinicians] to ask whether there is anything you [the individual] wants to talk about like dying then listen to the answer and respond to my questions or tell me when you don’t know.

People with diabetes and families made some suggestions that could help clinicians meet their communication needs and discuss changing life expectancy and ACP. Their suggestions are paraphrased in Table 1.

Need for clear language

The need to use ‘clear language’ also emerged. One palliative care nurse said:

Not using clear language is a big problem. I remember a time being in a family meeting and the consultant told the family their loved one had ‘gone.’ When he left, the family said ‘where has he [the relative/patient] gone and how long will he be?’

This anecdote led clinicians to  name other metaphors for death, which included: ‘no longer with us,’ ’gone to God,’ ’pushing up daisies,’ ‘with the angels’ and ’fell off the perch.’ This discussion caused laughter and reduced the stress associated with the ‘death conversation.’ The group indicated most of these metaphors were inappropriate in health care, unless they reflected the individual’s words.

One palliative care nurse suggested death metaphors are frequently used in clinical documents:

I saw ‘the patient passed away’ written in patient’s notes last week up on the medical ward. I think doctors should say the patient is dying or died – it’s clearer but we get conditioned to use them, metaphors.

A general practitioner agreed:

It’s in the media too. They say we lost so-and-so or somebody passed away. That just perpetuates the language.

A diabetes educator said:

I try to find a middle ground to talk about their care so we understand each other.

Most clinicians, people with diabetes and family members in all the focus groups indicated it is essential to understand what each other is saying and to find ‘the middle ground’ to have respectful, sensitive conversations.

Finding the middle ground

Most clinician participants indicated it is important to reflect the words individuals used. Clinicians said they tried to use or reflect the language the person with diabetes and family used to be respectful and avoid words that could be distressing. A diabetes educator said:

You need to find some middle ground and use the words the person uses. You have to be respectful. ‘Passed on’ is not disrespectful, if the person uses it: it is respectful at a time that’s very difficult.

A palliative care physician agreed, but suggested:

Using the same language as the patient and family is only useful if everybody in the conversation has a clear understanding of what the language means, otherwise it still leads to miscommunication.

This concept also emerged in the people with diabetes and family focus groups. A woman with diabetes said:

I have cancer and diabetes. My oncology doctors and nurses tell me what to expect in clear language and I prefer that but I can see that it could be OK to use the same words as the person as long as you are both [person with diabetes, family and clinician] clear about what you’re talking about and people’s culture and religion accept it.

The person’s comment alluded to the fact that the care setting and clinician’s discipline could influence the language used.

Influence of the care setting

The suggestion that care settings and the clinicians’ discipline influenced the language used emerged in all the focus groups. The older people with diabetes and family members indicated they had encountered death ‘in many guises’. ’Palliative care clinicians were more likely to have experience caring for people who were actively dying and ACP and were generally more comfortable talking about these issues compared to diabetes educators. A palliative care nurse said:

In acute care people think they will be saved and recover and go home. In palliative care, people are very much more comfortable talking about and planning for death but that does not mean it is not stressful.

An endocrinologist said:

We know about diabetes and the complications that cause death but we generally don’t talk about it [death] with patients. Some do but they are the minority. I think we could do with some training in it.

Discussion

The potential impact of language on the health and wellbeing of people emerged in several papers over the past five years. ^{38, 39, 40} None of these publications focus on the seemingly common practice of hiding death in metaphors, which was at odds with several clinician’s and almost all people with diabetes’ comments that the language needed to be clear to everybody in a conversation.

Clinicians regarded reflecting ‘the patients’ language’ as respectful and indicated it enhanced communication. It was also referred to as ‘mirroring language,’ which is discussed elsewhere.³⁵ Interestingly, clinicians did not appear to consider that people with diabetes and family might actually be reflecting/mirroring clinician language in order to communicate with them.³⁸

Willingness to talk about death also depended on the clinician’s discipline. For example, discussing ACD and dying is part of palliative care, which could be one reason palliative care clinicians are comfortable discussing these issues. While most diabetes clinicians know people with diabetes die; the ‘death language’ is not part of their discipline language, which could help explain why they preferred death metaphors. Significantly, people with terminal illnesses are usually aware of their prognosis and are better able to set realistic goals and treatment preferences, especially if their family and clinicians are willing to discuss these issues with them.

Our research suggests structured education programs improve clinicians’ ability to have conversations about prognosis and document goals of care⁴¹ and provide dying care.²¹ Table 2 lists some useful resources as well as books written by people who are dying and clinicians managing LLIs. These stories can help clinicians understand the words and language ‘patients’ use and ‘respectfully reflect their language’ in consultations. The books could also help clinicians reflect on their readiness to have such conversations!

Three important documents were released in in 2016 and 2018, which increased the focus on ACP:

1. The Medical Treatment Planning and Decisions Act 2016 that has implications for Advance Care Planning and Advance Care Directives.⁴²
2. The second edition of Australian Commission on Safety and Quality in Health Care National Safety and Quality Health Care Standards (NSQHS) (2017).³⁴
3. The Victorian Assisted Dying Legislation date 2018.⁴³

It is not clear what impact, if any, these documents had on the participants’ thinking or practice. The documents do discuss the importance of clear communication and ACP but do not indicate whether clinicians should hide death in metaphors or reflect a persons’ language during discussions. The NSQHS standards ³⁴ largely apply to hospital care but they are applicable in all care settings. They include shared decision-making, which can only occur if people use clear, unambiguous language and listen to each other: that is good communication.

Death is a fact of life for everybody. Palliative care is everybody’s’ responsibility.¹⁴ ACP should be part of holistic personalised diabetes care and initiated in a timely manner, particularly for older people with diabetes and comorbidities such as dementia, cardiovascular and renal disease and frailty. These people are likely to benefit from palliative care rather than tight metabolic management, which can result in death due to hypoglycaemia.⁴⁴ Hypoglycaemia and hyperglycaemia can cause unpleasant symptoms, some of which can be improved using palliative care.¹⁴

Comprehensive predictors of mortality, including functional status and QOL, especially the individual’s QOL issues, and diabetes and other comorbidity-related indicators need to be considered when deciding care.^{14, 15, 35} Prognostication is challenging, despite the availability of tools to help predict life expectancy (see Table 3). Many of these tools rely on subjective assessment and often use ‘tick boxes;’ some require value judgements. Most do not mention diabetes specifically. These issues should be considered when using prognostic and other assessment tools.

Significantly, risk assessment, prognostic indicators and care must be discussed with the individual, and sometimes their family, to support the individual’s autonomy and be personalised as part of shared decision making. Helping people document their values can guide care when the individual lacks decisional capacity and reduce the likelihood of futile treatment and moral distress for all concerned. It is a privilege for clinicians and gift for individuals with diabetes and families.

Strengths and limitations

A strength of the study is that the data were saturated and the researchers and advisory group members did not participate in the focus groups. A limitation is that only four family members participated; thus the views of  family members who participated and the participants may differ from other clinicians and people with diabetes. Another limitation is that all the clinician participants were female. Possible reasons were there are more females than males in the sampling populations and only/or only females elected to participate. Thus, male clinician voices were not heard and could be different for female voices. In addition, few people from culturally and linguistically diverse backgrounds participated, although one participant mentioned the need to consider cultural and religious customs and beliefs. Likewise, no Aboriginal or Torres Strait Islander people participated. Thus, the findings cannot be applied? to these groups and may not be transferable outside the study.

Areas for further study

1. Repeat the research in other settings and participant groups including Aboriginal and Torres Strait Islander People.
2. Determine whether structured ACP training could improve diabetes clinicians’ capacity to engage in discussions about diabetes and end of life care.