Who Should Discuss Advance Directives With Patients With Cancer and When? A Systematic Literature Review CURRENT STATUS: UNDER REVIEW

Among patients with cancer, discussion about the desired care is important, because progression of the disease, particularly signs of impaired consciousness, can lead to unwanted care in situations where no advance directives (ADs) have been provided. The aim of this study was to identify the most appropriate individual and time for addressing the subject of ADs with patients with cancer. A systematic literature review was carried out between September 2014 and August 2015, using the following databases or sets of databases: PubMed, Web of Science, Cochrane Library, EM Premium, and the French database Public Health Data Bank. The methodological aspects of the articles were selected and evaluated with the help of PRISMA international guidelines.

Only 5-25% of the population prepare ADs. This segment of the population includes both individuals in good health and those suffering from an acute, chronic, or deadly disease.
These figures have changed little [9][10][11][12][13], despite mostly favorable attitudes toward AD arrangements [15] and information and promotion action undertaken in drafting them [16]. Likewise, we find difficulties in dialog regarding ADs, as well as an indifference among the participants on the subject of drafting them. Of the 62 patients affected by cancer who were invited to be part of a study on ADs, 29 refused to participate [17]. The majority stated a lack of interest, others mentioned fatigue or pain symptoms, and others explained that they did not want to discuss ADs. Of all those who agreed to participate in the investigation, 57% did not wish to prepare their ADs.
The barrier in communicating with the patient about ADs most cited by doctors is the fear of hurting the patient by bringing up this subject [15]. Compiling ADs requires a certain amount of caution because this questioning reminds the patient of a difficult road ahead, and it may make him or her aware too abruptly of the seriousness of his or her illness [16]. It is only to be expected that the doctor would fear creating or increasing the patient's anxiety. Answering the question about the appropriate individual and time for addressing the subject of ADs thus becomes essential.
Given the limited information in French guidelines, we have chosen to develop an inventory of the implementation of ADs among patients with cancer.
The main objective of our work was to evaluate who was the most appropriate individual to bring up the subject of ADs and when, according to patients with cancer.

Methods
We carried out a systematic literature review in accordance with PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) international guidelines, from September 2014 to August 2015. Bibliographical monitoring was carried out up to September 2015.

Search Strategy
We searched the following databases or sets of databases: PubMed, Cochrane Library, Web of Science, EM Premium and the French database Public Health Data Bank. We also searched scholarly organization sites. We have also consulted the gray literature and the bibliographical references of the articles included.
The search equations should have made it possible to select articles that would help to answer the following question: "With whom and when do patients with cancer want to address the subject of advance directives?" Each search equation was performed using English terms from the MeSH (Medical Subject Headings) database, along with (if possible) French versions of the terms, also from MeSH.
The search question was itemized under 2 themes: advance directives and cancer. For advance directives: we used the English terms "advance directives" and "living wills," and we excluded the term "advance care planning." For patients with cancer, we used the terms "neoplasms" and "cancer patients." Subsequently, we performed another search, adding a third term: "communication." MeSH definition of the descriptors: Advance directives: "Declarations by patients, made in advance of a situation in which they may be incompetent to decide about their own care, stating their treatment preferences or authorizing a third party to make decisions for them." (Bioethics Thesaurus) Living wills: "Written, witnessed declarations in which persons request that if they become disabled beyond reasonable expectation of retrieval, they be allowed to die rather than be kept alive by extraordinary means." (Bioethics Thesaurus) Neoplasm: "New abnormal growth of tissue. Malignant neoplasms show a greater degree of anaplasia and have the properties of invasion and metastasis, compared to benign neoplasms." Communication: "The exchange or transmission of ideas, attitudes, or beliefs between individuals or groups." Each MeSH term was compiled by the Boolean operator [AND].

Selection Criteria: Inclusion and Exclusion Criteria
The articles included had to be written in English or in French. The study location was not limited to any particular area. The population studied consisted of patients over 18 years of age, suffering from a cancer that is solid or of hematological origin, with no differences in the stage or evolution of the disease, and who were hospitalized or seeking inpatient or outpatient consultation. The intervention needed to involve the act of addressing the subject of ADs with the following main judgment criteria: the preferred professional, the most appropriate time to address the subject of ADs, or both. The article publication date should have been between January 1, 2000, and August 31, 2015.
Studies in which the population was inappropriate, the objective was unsuitable (i.e., did not involve the evaluation of the content of the ADs, the factors influencing their drafting, the method applied to improve their drafting, etc.), or the intervention did use a strict definition of advanced directives (i.e., did not discuss advanced care planning, end-of-life discussion, do-not-resuscitate orders, ethics consultation, etc.) were excluded. Studies for which the full text could not be retrieved were also excluded. Editorials, letters, and case analyses were not selected.

Evaluation of Study Quality and Data Extraction
Selection of the articles was carried out in several steps: reading the title, then the summary, and then the full text.
To guarantee the quality of the articles included in the literature review, we evaluated them by using the following guidelines: STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) for observational studies [18] and SQUIRE (Standards for QUality Improvement Reporting Excellence) for interventional studies [19].
The analysis guide for the literature published by the National Authority for Health has also made it possible to define the items essential for the inclusion of articles and the extraction of data: precise definition of the objectives, description of the relevant context, locations and dates, definition of the eligibility criteria, definition of the sources and population selection methods, description of the data and the data collection method, description of the statistical analysis and results that can be verified using raw data. The articles were selected by a single researcher, but they were read and analyzed by 2 researchers. Table 1 presents the search equations used and the number of references found in the various databases. The selection process using the databases made it possible to identify 1,170 articles, and 25 articles were identified from reading the bibliographical references in the included articles. Of the total 1,195 included articles, 26 articles were read in full, and 7 were ultimately selected ( Fig. 1).  With regard to methodology, 5 studies followed a quantitative method, 4 used specifically designed standardized questionnaires, and 1 study used a previously established questionnaire. Finally, 2 studies followed a qualitative method using semi-structured interviews; for the first one, a guide designed specifically for the study was followed, and for the other one, a guide derived from a previous study with added questions was followed.

With Whom Should Patients With Cancer Discuss ADs?
Six out of the seven articles included in this literature review provide answers to this question.
(1) Doctors The 2 types of specialists most often mentioned as those who should address ADs with patients with cancer were the general practitioner and the oncologist. In 2 studies, the general practitioner was the most mentioned. did not give a reason, and 5% (4/75) provided another reason.
(2) Patients' relatives This group of individuals was examined in only 2 studies. In the study by Hubert et al. [20], they were mentioned as individuals who were eligible to provide assistance in the drafting of ADs by 40.1% (first place) of patients who had issued ADs, whereas 93% (first place) mentioned them as knowledgeable individuals when it comes to drafting ADs. In the study by Sahm et al. [23], when patients were asked to answer whether their relatives should initiate the discussion of the drafting of the ADs, 44.2% of patients answered "YES" and 55.7% answered "NO" (n = 95). It should be noted that, in the study by Dow et al. [21], out of the 54 patients who did not want to discuss ADs with their oncologist, 22% believe that it was more of a family or personal decision.

(3) Legal professionals
Legal professionals were considered in 3 studies.
In the study by Pfistinger et al. [17], of the 285 patients without ADs, 8% (19/285) wanted to obtain legal advice. In a study by Van Oorschot et al. [24], for the item "the directives must be certified by a notary," the 139 patients with an AD predominantly disagreed or strongly disagreed; the 314 patients without ADs but who were likely to obtain them in the future predominantly disagreed, and the 53 patients without an AD and who were unlikely to obtain them in the future predominantly disagreed or agreed. Finally, individuals in the notary/lawyer category were mentioned among those who should provide assistance in the drafting of ADs (25.5%) in the study by Hubert et al. [20].

(4) Other type of individuals
In 2 studies, 2 other types of individuals were considered: the clergyperson and the psychologist. In the study by Hubert et al. [20], for 3% of patients without an AD, a pastor is mentioned as a source of information. In the study by Sahm et al. [23], in response to the question "who would initiate the discussion about drafting ADs?" 11.5% of patients answered "YES" to the suggestion of "the clergy or psychologists," and 88.4% answered "NO" (n = 95). Finally, in the study by Hubert et al. [20], with regard to the information source, the media is mentioned by 22% of patients with ADs and by 19.6% of patients without ADs.

When Is the Best Time to Discuss ADs With Patients With Cancer?
Four out of the seven articles included in this literature review provide answers to this question.
(1) During a state of "good health" In 3 out of 4 studies, patients suggested that the most opportune time to address the subject of ADs was prior to the disease, "when one is in good health." Vinant et al. [22] reported that out of 20 patients who evaluated the time for information about advance directives, 11 patients answered "in good health." This finding is in agreement with those of the study by Pfistinger et al. [17], in which, out of 285 patients without ADs, 47% (135/285) believed that this discussion should take place independently of the neoplastic disease. In a South Korean study by Keam et al. [25], in a state of good health was the third most preferred time; of the 1,242 patients, 17.6% (217/1,242) believed that the drafting of ADs should take place while the patient was in good health. These findings were in line with those of the study by Sahm et al. [23], in which, in response to the question of "who should initiate the discussion about drafting ADs?", for "only affected individuals," 81.6% (n = 98) of respondents answered "NO" and 18.3% answered "YES." (2) During the diagnosis In the study by Pfistinger et al. [17], during the diagnosis was the third most preferred time; 12% (33/285) of patients without ADs chose "after the cancer diagnosis news." In the study by Keam et al. [25], during the diagnosis was the second most preferred time, chosen by 29.5% (364/1,242) of oncology patients. In the case of the study by Vinant et al. [22], 9 patients out of 20 suggested that the time of information about ADs take place "at the outset of the treatment for a serious illness." (3) During palliative status Keam et al. [25] reported that 37.2% (459/1,242) of patients, which was the majority, in their study believed that this discussion should take place during terminal status, and 15.7% (193/1,242) believed it should take place when death was imminent. During "incurable disease" status was the least preferred time in the study by Pfistinger et al. [17]; only 9% (25/285) of patients in the study chose this option. Likewise, "when the treatments for the disease are no longer enough" and "upon reaching the end of life" were not chosen by patients in the study by Vinant et al. [22]. Of note, in the study by Pfistinger et al. [17], "when the patient requests it" was chosen by 33% of patients without an AD (93/285).

Discussion
Our literature review demonstrated that a doctor, particularly a general practitioner or an oncologist, is the individual with which patients prefer to discuss the subject of ADs. The most appropriate times were when patients were in good health (before the appearance of a serious pathology) and at the time of the disease's diagnosis.
We limited our search to publications from the year 2000 onward in order to study the most recent data. We focused on the most recent data because the field is constantly evolving, along with the ethical, medical, and legislative issues involved. The first observation that we can make is that the subject of drafting ADs and, more broadly, endof-life medical conditions, is universal. However, legislative progress regarding the end of life is not the same on all continents.
The main limitation of our study is due to a selection bias because, despite the articles being read and analyzed by two researchers, there was only a single researcher in place to select and evaluate the articles. However, the inclusion of articles from various countries and various cultures makes it possible to gain broader external validity.

With Which Professional Do Patients Want to Discuss ADs?
Our study finds that the doctors who would most likely be chosen by patients are oncologists and general practitioners. Palliative care doctors, nurses, and legal professionals were not cited or evaluated.
The American study by Dow et al. [21] showed that patients would allow this discussion to occur during the admission process. The question could then be addressed when opening an outpatient medical file or upon hospitalization, for example. This form of standardization would make it possible to prevent associating ADs with an imminent endof-life scenario. This hypothesis is supported by the French study by Vinant et al. [22], where patients wanted the doctor addressing the subject of ADs to not be involved in the medical care.
Moreover, the importance of the general practitioner should be pointed out; the general practitioner was the most preferred professional in 2 studies and the second most preferred professional in a third study. This can be explained by the fact that, as a primary care physician, the general practitioner knows the patient the best, from a medical and personal point of view, and for the longest amount of time. He or she is the "patient care coordinator" [26]. He or she is often the family doctor who knows the patient's socio-family context and could initiate the discussion on ADs with the patient and his or her family during a home or medical office consultation.
O'Sullivan et al. [27] found that patients who consider the subject of ADs extremely important prefer that their family doctor be the individual to initiate the discussion.
When Is the Right Time to Address the Subject of ADs?
For patients, various times were identified: apart from a serious pathology, when the patient is in "good health," (from 3 studies).
This raises the question of the validity of ADs and the variation in choices over time. In 2007, Haas et al. [8] confirmed that the view of patients can vary widely upon approaching end of life; it is not fixed, and evolves, as Aubry had already suggested in 2006 [28]. Priorities change as death approaches. Are the wishes of an individual in good health the same as those of a sick individual? Also, how does one anticipate the medical consequences and possible treatments for a disease that one does not have?
In 2 studies, the "time of diagnosis" was identified as being the most appropriate time.
However, this time, in particular in the context of cancer, already requires several explanations: the diagnosis itself, the therapeutic strategy, and access to support care [29]. A Swiss study from 2014 [30] reported similar types of answers from the point of view of the general practitioner. This study was conducted with 23 general practitioners of various nationalities (Italian, German, French, and Swiss) and its main objective was the evaluation of the "right" time to discuss ADs. Three trends in the answers emerged: "when the individual is in good health," "when he/she is still in good health but has a serious or chronic illness," and "systematically upon the first consultation, during a life event or when entering a retirement home." This also raises the question of the patient's competence and medical knowledge [31] (i.e. whether and when to decide what is best for him or her). How would an individual who does not know the possible evolution of his or her illness or the technical care required be in a position to make decisions about his or her medical care? However, it is also the ability to make clear judgements and to be able to give consent, which, for example, raises the question of the value to be placed upon a request to cease medical care that is made by a depressed patient. How do you distinguish, in the case of a request to cease medical care, how much is due to a lucid and legitimate desire, how much is due to depression that can be treated with therapy, and how much is due to fatigue associated with the treatment? In Report N°58 of the National Consultative Ethics Committee [32], the following is stated: "the act of consenting implies a dual capacity: that of being able to understand and that of being able to freely determine." This also raises the problem of patient information and therefore the time to be granted to them for drafting ADs. A study conducted in Geneva in 2008 [33], evaluated the time necessary for the drafting of advance directives for patients assisted by 2 nurses. The average length of time required was 3.2 hours.
In France, the Claeys-Leonetti law was passed in February 2016. It specifies the procedures for drafting ADs and the appointment of the support person and confirms the role of the primary care physician with regard to the information and the reporting thereof. The National Authority for Health published documents for health professionals and patients in order to spread awareness of ADs and to direct their drafting [13,34]. In

Conclusions
We have carried out a systematic literature review in order to determine the professional with which patients with cancer prefer to discuss ADs and when they would prefer to have this discussion. We have highlighted, at the end of this work, that the most appropriate times were when patients were in good health (before the appearance of a serious pathology) and at the time of diagnosis. The preferred professional was a doctor, in particular a general practitioner or an oncologist. We believe it would be worthwhile to suggest a consultation dedicated to drafting ADs, whether outpatient or inpatient. Our results resonate with the 2016 French law concerning the rights of patients receiving endof-life care, but they also indicate that health professionals need to dedicate time to this issue.

List Of Abreviations
ADs: Advanced Directives

Consent for publication
Not applicable

Not applicable
LG : writing manuscript, critical review, commentary and revision VD : critical review, commentary and revision MERB : critical review, supervision.
All authors read and approved the final manuscript.