The role of family confidants and caregivers in the care of older cancer patients: Extending the concept of “shared decision‐making”

Abstract Background and aims Family caregivers play an important role in assisting their family members with cancer, but their influence on the treatment decision‐making process has not yet been adequately investigated. This exploratory study approached this topic via reconstructive methodology, focusing on assessing patient‐caregiver relationships. Methods We conducted semi‐structured interviews with 37 mostly elderly cancer patients (median age: 74 years) about the context of their diagnosis, treatment decision, and family support. Additionally, we interviewed 34 caregivers of cancer patients. Of these, 25 were related to patients interviewed. We analyzed the interviews via a multi‐step coding method informed by Grounded Theory methodology toward characterizing patient‐caregiver relationships, the treatment decision‐making process, and the caregivers' role therein. Results In the majority of cases (86%), patients were being supported by caregivers. We categorized patient‐caregiver relationships in regards to the caregivers' involvement in the therapy decision‐making process. We found patient‐caregiver interaction patterns that indicate the potential of caregivers to decidedly influence the therapy decision‐making process. Yet, only in 38% of cases, a caregiver attended relevant patient‐physician‐consultations. Conclusion Depending on the nature of the patient‐caregiver relationship, the traditional concept of shared decision‐making, which assumes a dyadic relationship, needs to be extended toward a more dynamic concept in which caregivers should be involved more frequently. This could enable physicians to better understand a patient's reasons for or against a therapy proposal and ensure that the patient's wishes are communicated and considered. On the other hand, strong caregiver‐involvement bears risks of over‐stepping elderly patients' wishes, thus violating patient autonomy.

caregiver-involvement bears risks of over-stepping elderly patients' wishes, thus violating patient autonomy.

K E Y W O R D S
family, older cancer patients, qualitative research methods, quality of life, survivorship

| INTRODUCTION AND BACKGROUND
The care of cancer patients of advanced age requires special caution due to age-associated limitations of organ functions, co-morbidities, and co-medication. 1,2 New pharmacological developments, however, represent increased therapeutic options also for older patients. 3,4 These developments, combined with the demographic shift to more elderly patients, underline the need for sensitive and reasonable patient/doctor conversations that include a clear understanding of the treatment goals by both parties involved.
Several studies have shown that the adherence of patients to their doctors' treatment recommendations is surprisingly low. This is especially true of elderly cancer patients and oral cancer therapy, with adherence rates down to 46%. 5 Good physician and patient communication was identified as one of the major contributing factors for adherence. 6,7 One of the reasons for the complexity of patient/doctor talks especially with older patients is the multilayer situation of rational decision-making regarding the proposed therapy in a frightening situation in which a patient is facing a potentially deadly illness. 8,9 Family confidants and caregivers (CGs) play an important role in supporting their ill family members in many respects. Some of their major duties are managing the home and housekeeping and support in domestic chores including preparing and assisting with meals, dealing with mobility issues, arranging for transportation to medical care facilities, making medical appointments, managing prescription medication and also offering emotional support. However, the extent of CGs' influence on patients' decision-making regarding therapy recommendations given by the oncologist is not clear. Therefore, we conducted a study in which we analyzed interviews with older cancer patients (median: 74 years) and their CGs, focusing on the patterns of relationships between them.

| METHODS
We conducted semi-structured, 10,11 guideline-based interviews 12 with 37 mostly elderly (55-89 years, median age: 74 years) recently diagnosed cancer patients and 34 CGs, most of whom were related to the patients interviewed. Cancer patients were recruited in a German cancer center and two rehabilitation clinics and asked for contacts to persons who had supported them during the course of their illness. As additional patients' recruitment would have posed a burden for the clinics, we opted for an alternative recruitment path to address the CGs of elderly cancer patients directly. Interview guidelines were developed in an interdisciplinary process with the collaboration of oncologists, social scientists, and philosophers based on the SPSS-method, according to Helfferich. 12,13 After a pre-test with four patients and two CGs, the interview guidelines were deemed suitable and understandable. Interviews were held in various settings between November 2017 and May 2018 (see Figure 1 for the participant-recruitment process and interview settings, Table 1 for the relevant patient and CG characteristics).
The interviewers prompted the participants to give free and indepth answers, as our approach was exploratory. The notion that CGs are an important group in the context of elderly cancer patients' illness did not only emerge during the analysis of patient interviews. As we did not pursue a preconceived research question and did not a priori focus on a particular dimension, we did not opt for an iterative data collection process (e.g., theoretical or targeted sampling. 14 The interviews were transcribed verbatim, pseudonymized and anonymized prior to publication. Data analysis was performed using a multi-step coding method, informed by "Grounded Theory" methodology. 15 Open, axial, and selective coding processes were applied subsequently. 16 The emerging categories and subcategories concentrated on the patient's autonomy and self-determination, related limitations and burdens, and subsequently, the patient-CG relationship. This study was approved by the ethics committee of the University of Luebeck (file number: 17-288) and informed consent was sought from all participants.

| RESULTS
Of 37 patients interviewed, 32 (86%) stated to have been supported by a CG during the course of the illness. We analyzed information on 41 patient-CG relationships (32 patients with CGs and 9 CGs interviewed separately). We distinguished patient-CG interaction patterns into "passive" (30 cases, 73%) and "active" (11 cases, 27%) regarding the respective CG's involvement in the therapy decision-making process (see below). Patients with actively involved CGs were aged 73 to 89 (median: 80) compared to a range of 55 to 89 (median: 73) for patients with passive CG involvement. Passive CGs were present at relevant doctor-patient conversations in 31% of cases. In cases of active CG-involvement, the participation rate was 45%. These CGs were also concerned with the patient's general mental well-being by creating a sense of normality:

| Passive caregivers
…what was a priority for us was to provide encouragement. Not to make him understand that life was over now and: 'We don't care about you any-more', but always to be there for him'.
[…] we deliberately did not want to talk about these decision-making processes or exclusively about this topic. So, we tried, simply to keep everyday life going, to put it simply: 'When we come (for a visit), we don't just talk about this disease'.
Because we saw that that was important …(Neighbor and friend, AM107) It should be noted that "passivity" only refers to the CG's involvement in the patient's decision-making process and is not a value judgement and should not be understood as indifference toward the patient's illness.

| Active caregivers
Active CGs were concerned with shaping and steering the therapy decision-making process. They considered their strong involvement in this process as vital for adequately meeting the patient's needs. Taking into account the multitude of motives for such involvement, we expanded on the broad category of "active" CGs into two different categories: (a) patient and CG as a unit with the CGs identifying themselves with the patients' illness, and (b) reversal of the care relationship.

| Patient and caregiver as a unit: Caregivers identifying themselves with the patient's illness
Especially in cases of married couples, we observed CGs who identified themselves strongly with the patients' illness and their situation.
In these couples' accounts, the aspect of experiencing the therapy process together was emphasized and perspectives were mutually shared. Accordingly, the cancer was perceived as a challenge that F I G U R E 1 Illustration of the interview recruitment process and interview settings could only be overcome through joint efforts, and therapy decisions were always portrayed as a consensus, weighing the consequences for each individual and the relationship.
And it was clear to us that we go for the (medical) discussion and then we also went together directly, we went to the discussion, (thinking) it may be that he (the doctor) now will tell us this is now a carcinoma. And that turned out to be the case. [ The extent to which a CG influences the therapy decision in these cases is difficult to determine since the eventual therapy decision is the result of a complex interpersonal and individual process. In the case cited above, the CG stated that it was the patient's decision. At the same time, the CG mentions that the burden that would have to be carried by her was also a relevant factor in the decision. The relevant steps and decisions were framed as being taken together. It can therefore be assumed that in such a patient-CG relationship, the values, views and assessments of the CGs play a significant role in the choice of cancer therapy and the way its consequences are handled.
CGs and patients thus influenced each other's perception of cancer, seeing it as a test that also conferred meaning on their relationship:  The interviews analyzed made it clear that, in particular, the coordination of medical appointments and discussions could represent a great challenge for older patients. In several cases examined, this was a central topic which also determined the role of the CGs: these patients wished to have their CG present at medical consultations, and the CGs, for their part, were convinced that their presence at medical discussions was necessary to ensure, among other things, that no information was lost, neither the relevant information about the patient to be communicated to the doctor nor that given by the doctor to the patient.
And then to the gynecologist and he discussed the therapy options with us and there I was able to discuss the questions we had considered (earlier) and so on very well. And I think that maybe that helped her once again, because I think one is simply overwhelmed in the situation, or one has so many thoughts in one's head that sometimes one can't ask questions clearly, although one would simply like to know so many things, isn't it? (Daughter, AM105) The transformation of the CGs' role was a key observation in several interviews. This transformation has led to a shift of the CG's role in the patient-physician-CG triad and can be described as a role reversal: The interviews analyzed here were mostly mother-daughter relationships. The assumption of responsibility by the CGs thus meant a reversal of the care relationship. While the "mother" bears responsibility and provides care, the "child" is freed from responsibility for itself and its actions. These CGs reported that they became aware of this role reversal when facing the necessity to make vital and immediate cancer therapy decisions. Even though they had already assumed greater responsibility for their parents or parents-in-law, the illness accelerated this process.
…but now that it has -but it has totally changed. The patient had decided not to undergo further treatment. The daughter felt that this decision did not fit her life-history and identity.
The CG saw her function as contextualizing and counteracting a decision that she perceived to be irrational. In her view, she protected her mother from the consequences of a decision made under distress. She infringed the patient's autonomy to find a treatment decision consistent with her mother's values ("not giving up"). This illustrates that the concept of "autonomy" in elderly cancer patients' therapy decision making is not absolute.

| DISCUSSION
The relationship between patients and their CGs can be complex and multilayered, as has been shown, for instance, by studies on patients' feelings of being a burden in the context of palliative care decision-making. 17 The analysis of the interviews in this study shows that the forms of support offered by CGs to their ill family members can differ considerably. We found interactions that can be When interpreting the interviews with CGs, it must be considered that not all CGs were willing to be interviewed. Forty-six percent of CGs declined to be interviewed. The reasons given were heterogeneous, most often, no reason was given, and some CGs indicated that the situation was too stressful for them. This high rejection rate may have led to a selection of engaged CGs (bias) in our study. It is also interesting to note that although the CGs regularly criticized the health care system or organization details, which they considered needed improvement, they rarely questioned the specialist's competence as far as therapy proposals were concerned. This corresponds to the experience of the authors in oncological clinical practice. The CGs rather questioned whether a therapy concept that was generally GIESELER ET AL.
correct was also the optimal (meaningful) one in the particular case of their family member. 8 Although most patients (86%) indicated that they were supported by a CG, CG participation in patient-specialist consultations was surprisingly low (38%). There is little information on whether this percentage is representative, but it corresponds to the authors' experience. We suggest that CGs should be more often integrated into the decision-making encounters between patients and their oncologists.
The patient is the sole addressee of all medical clarifications during medical consultations. This principle is not only ethically correct but also legally enshrined. It should be preserved especially in the CG-patient relationship we considered "care relationship reversal," where CGs take upon themselves the responsibility for their patients where the CG appears to be having total control over the patient's therapy wishes, the physician must be alert to the possibility of the patient being rendered a silent and possibly nonconsenting partner to his therapy plan. In such cases, it is necessary to actively and directly include the patient in discussions and to find an opportunity to talk to the patient alone to find out what their real wishes are.
The gold standard in patient-doctor communication is the concept of "shared decision-making" (SDM), which has meanwhile been revised and modernized several times. [20][21][22] As early as 1997, Cathy Charles published an article with the subtitle "it takes at least two to tango." At least two suggests that more than two may be meaningful.
However, it is still not usual to involve more than the doctor and the patient in making therapy decisions. In cancer therapy, it would be particularly important not only to achieve just adherence. The patient follows the doctor's instructions, but concordance is based on the patient's conviction that the doctor's proposal is appropriate for her or him personally. 23,24 In this context, CGs can play a constructive and meaningful role by mediating between a patient and oncologist, conveying information in both directions.

TRANSPARENCY STATEMENT
Frank Gieseler is the guarantor. He declares that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.