Preferences of elderly cancer patients in their advance directives

1. Introduction 

The relationship between patients, proxies and physician during end-of-life decision-making is complex and changes have appeared in Europe since approximately 10 years only; there is a shift away from paternalism, yielding to a growing emphasis upon patient's autonomy [1], [2]. Some patients want to be active in end-of-life decisions and would not leave them to doctors [3], [4]. Furthermore, regardless of the proximity to death, patients often express concerns surrounding death, like fear of pain, fear of indignity, fear of abandonment and fear of the unknown [5]. Open and direct discussions can ease many of these fears.

Advance directives (ADs) are important documents to allow competent patients to record the nature and the kind of medical procedures they want to be implemented in the case they would become incompetent. The role of ADs at the end-of-life has been questioned in studies that mainly focused on resuscitation measures and costs. ADs have at least three notable benefits: they increase patient's empowerment, they remind caregivers to value patients as partners and they encourage planning for death [6], [7], [8]. We recently demonstrated that ADs are vehicles for in-depth and ongoing discussions among health care professionals, patients, and families [9]. We were able to show increased satisfaction with end-of-life decisions of patients with an advanced oncological disease. Not only did patients have medical concerns, but also religious beliefs values and lifestyle concerns. The aim of the present study was to better identify preferences and values expressed by elderly cancer patients in their ADs and to evaluate the practical utility of ADs when patients become incompetent.

2. Methods 

Retrospective chart review of all elderly patients with an advanced oncological disease hospitalized at our institution. Consecutive patients with advanced cancer were included if they completed their ADs during hospitalization from 2004 to 2007.

Collected data included demographic information, religion, education level, main medical diagnoses and Charlson's co morbidity score [10]. Mini Mental State Examination (MMSE) was used to assess the cognitive status at admission [11].

From 2004, each hospitalized patient was given short information about ADs at admission. Patients wishing to go further in the process and complete their ADs received a written support and more extensive oral information. Patients were not asked to complete a pre-established existing check-list, but were asked to write down their preferences and values on a sheet. They were encouraged to designate one or more proxy (-ies). Health professionals offered to support the patients during the process. ADs were considered completed upon signature.

Main motivations of patients to complete their ADs were recorded. Item's content of ADs was analyzed and defined in broad categories and then classified accordingly. The practical utility of ADs was measured for patients who died under our care.

3. Results 

During the study period, approximately 1600 patients with an advanced oncological disease were hospitalised and 149 patients (9%) showed interest to complete ADs. Ninety-nine patients initiated the process, but were unable to complete the process for the following reasons: too difficult to anticipate (n: 38), decision turned down after discussion with the caregivers (n: 33), onset of delirium or worsening clinical condition (n: 28). Thus, a total of 50 ADs were available for analysis. Characteristics of the patients are described in Table 1. Thirty-four patients died during hospitalisation after a mean length of stay (LOS) of 51.2±37.8 days, 14 went home (LOS=34.7±41.2) and 2 were discharged to a nursing home (LOS=53.2±35.2).

Table 1. Characteristics of patients.

		n: 50
	Age (mean±SD) (n)	71.9±9.6
	65–69 years	13
	70–79 years	16
	80–89 years	18
	≥90 years	3
	Gender (n)
	Male	19
	Female	31
	Marital status (n)
	Married	23
	Widowed	18
	Divorced	6
	Single	3
	Level of education (n)
	Primary school	16
	Secondary school	22
	University	12
	Nationality (n)
	Swiss	38
	Other	12
	Religion (n)
	Catholic	19
	Protestant	18
	Non-religious	12
	Other	1
	Main cancer disease (n)
	Gastrointestinal tract	16
	Respiratory system	10
	Genitourinary tract	9
	Breast	6
	Haematological	4
	Other	5
	Mean reason for admission (n)
	Uncontrolled symptoms	31
	Infection	5
	Rehabilitation after surgical intervention	5
	Acute complication as ileus, delirium	9
	Charlson comorbidities score (mean±SD)	6.9±1.2
	MMSE (mean±SD)	28.0±1.8

Five patients had already completed some ADs prior to hospitalization. Two patients had completed a pre-established questionnaire and three patients had written a short sentence to express their fear of over treatment. All of them agreed to fully complete ADs.

ADs were completed at a mean of 17±19 days after admission. Thirty patients decided to complete their ADs, either because they had some previous knowledge or information on the process or after reading the written information provided on ADs. The other twenty patients were encouraged to write down ADs by the caregivers.

Main motivations of patients to complete their ADs were: to enhance autonomy (n: 31), to enhance communication with caregivers (n: 20), fear of over treatment (n: 17), not to be a burden (n: 14), to enhance communication with their surrogates (n: 7) and to be sure their preferences will be respected (n: 6). On average, 3 formal discussions (mean duration: 34.7±19.6min) with caregivers were necessary to complete the process. ADs contained two categories of mixed items that are described in Table 2. Main medical concerns of the patients were resuscitation and introduction of artificial nutrition. One patient expressed the wish to have everything attempted to prolong his life. Seven patients expressed their wish to have everything done to keep or improve their cognitive function even if symptoms like pain would be experienced. Four patients expressed the wish not to die alone. Two patients nominated proxies they wanted to stay along during the dying process. One patient asked to listen to music and to benefit from massage.

Table 2. Main items cited or not in the ADs.

	Main items	Wanted	Do not wanted	Not cited
	To be resuscitated	1	49	0
	To be transferred in another hospital in case of hip fracture for example	5	5	40
	The use of high doses of analgesics in case of intractable pain even with lower consciousness	31	7	12
	Introduction of artificial nutrition	3	30	17
	Introduction of artificial hydration	31	4	5
	Tracheostomy or colostomy	0	2	48
	Introduction of antibiotics	29	7	14
	Administration of blood transfusion	12	13	25
	Complementary exams	0	3	47
	Introduction of a palliative sedation in case of refractory symptom	27	0	23
	Hospital place wish if new hospitalisation	15	0	35
	Particular wishes for dying process	5	0	45
	Religious support	4	1	45
	Autopsy	9	22	19
	Funeral wishes	17	0	33

Main funeral concern was about cremation. One patient organised his funeral ceremony and another choose the clothes he would wear after his death.

All patients designated a surrogate: children (n: 27), husband–wife (n: 22), both (n: 3), sister–brother (n: 2), mother (n: 1), friends (n: 3), general practitioner (n: 1), psychologist (n: 1). Three patients could not designate their spouse because of cognitive impairment. One child refused to be designated as a proxy because the task would be too emotional for him. Only seven patients designated two surrogates.

Five patients among those who went home were re-admitted in another hospital where they died (Table 3).

Table 3. practical utility of ADs when patients become incompetent.

At our institution, mean number of days during which patients could not communicate because of delirium or impaired level of consciousness were 4.2±2.4. Six patients could communicate until their death. ADs were fully applied to all 33 patients except one who died at our hospital and who were unconscious during 1 or more days. No attempt at resuscitation was undertaken. No patients who declined artificial nutrition, hydration or blood transfusion was given any of them. All patients who refused to be transferred to an acute care hospital in case of complication stayed at our institution. Antibiotics were not administered to two patients who formally refused them despite the development of clinical signs and symptoms of infection. Two patients who wished to keep their cognitive function as long as possible experienced high intensity of pain and refused adequate analgesic treatments during several days to eventually died peacefully. Two patients received palliative sedation for refractory dyspnoea. Surrogates were regularly informed about the clinical condition of the patient. In one single situation, a surrogate refused permission for autopsy of her mother who had previously accepted the procedure and this was respected. No patients asked for assisted suicide or euthanasia.

4. Discussion 

A small minority of elderly patients with an advanced oncological disease completed ADs despite encouragement through an established 10-year-old specific legislation. The same minority was willing to appoint a proxy and to ask caregivers about the existence of ADs. Completion rate is thus much lower than previously reported in studies conducted in the United-States where active promotion was done and which varied from 15 to 23% [12], [13], [14], [15], [16]. In the Netherlands, one study based on interviews with surrogates of 342 deceased persons who were included in the Longitudinal Aging Study Amsterdam concluded that 9% of patients had a written declaration about medical end-of-life decisions and 27% had designated a health care representative [17]. These results were similar in another study conducted in Finland [18].

One possible explanation of such a low rate in our study may be that patients were not asked to complete an existing check-list containing specific treatments to be received or forgotten, but were requested to write down their values and preferences on a blank sheet. Such process may request additional attention and delicate formulation of personal wishes. In addition, this process requires time with a mean of three appointments that were necessary to build a dialogue and to negotiate between patients, family and health care provider. Finally, some patients must have discussed their preferences orally with caregivers and their choices were then reflected in the patient's chart.

Given the high prevalence of cognitive impairment (dementia and delirium) in this population, it seems important to promote discussion about goals of care and the completion of ADs earlier in the course of the disease and whenever possible before hospital admission. This would save preliminary information done by health professionals who were previously not involved in the patient care.

Our format of ADs promotes the expression of personals preferences and dialogue matching the high level of information needs at all stages of the disease process [19]. In our study, only one patient had unrealistic expectation as cardiac resuscitation applied to cancer patients may restore cardiac function but it does not prevent death. In one chart review of 406 cancer patients undergoing attempt at resuscitation, only 7 patients (2%) survived until hospital discharge. This issue together with such published results has to be discussed in depth with the patient. Moreover, it should not be offered as a reasonable therapeutic intervention in patients with a progressive malignancy since it could propagate misperception of long term survival after resuscitation [20].

Some patients refused interventions such as artificial nutrition or antibiotics which health professionals would have otherwise introduced in patients without completed ADs. Health professionals have thus to learn and accept to give up potential useful treatments to fully respect patients’ autonomy [2]. Open discussion between patients and caregivers about the care process, assumptions and preconceived ideas during ADs completion is therefore crucial. Caregivers thus need education and training to discuss these issues with patients.

Recent recommendations to introduce palliative sedation for refractory symptom have requested to either actively involve the patient, providing this is possible and appropriate or the designated surrogate, to elicit the patient's values, beliefs, and personal goals. In the case of patient's inability to participate, one should refer to previous discussions or documentation eluding at patient's values, wishes, or directives [21].

As often as possible, an advanced treatment plan about sedation, partial or complete, transient or terminal, reversible or irreversible, continuation of hydration and nutrition should be discussed [22]. In our study, palliative sedation was discussed with more than half of the patients but it was used only twice.

Content of ADs not only involves life-sustaining technology, but also psychosocial items and religious beliefs and values. Discussions that focus solely on resuscitation fail to recognize important psychosocial concerns [23]. The perspective to leave funeral arrangements to the family is a good example of that because patients want to prepare their death and secure the minimum burden for family members in the event of death.

Three patients were members of a “right to die” association but died without asking for assisted suicide suggesting that the process of ADS completion may decrease the fear to suffer at the end of life and may avoid to ask for an assisted suicide procedure. We would thus encourage patients who have asked for assisted suicide or euthanasia to complete their ADs.

The designation of a surrogate is an important task, but the person is unlikely to be able to consider what should be done in each situation. Nevertheless we must not forget that discussion may be emotional for surrogate when faced with the task of advocating for the patient. They often think they should take the decision for the patient. By involving family members in these discussions, relationships within the family can be strengthened, and can reduce the isolation experienced by the dying person [24], [25], [26].

Our study has several limitations. The item's content was influenced by the fact that patients actively participate in the treatment decision process thus not reflecting the wishes of other patients. A larger prospective study could be conducted to compare outcomes (satisfaction, death in hospital with or without resuscitation attempt, death at home or in nursing home and length of stay) of patients with oral transmissions and those who wrote ADs. In conclusion, only a minority of elderly cancer patients in our cultural environment are prepared to put their personal wishes in writing. Completion of ADs takes time and energy. Help and assistance from a qualified team is efficient. The whole process may ease many fears as suggested by a decrease in anxiety level. Elderly cancer patients should be informed about ADs and encouraged to complete them early in the course of their disease. ADs should not be considered as another questionnaire, but as a process to improve open and honest communication between patients, proxies and health professionals.

Conflict of interest 

No conflict of interest.

Reviewer 

Dr. Miriam B. Rodin, The University of Chicago, Section of Geriatrics, 5841 S Maryland Ave W-700, Chicago, IL 60637, United States.