Kidney transplantation or dialysis in older adults—an interview study on the decision-making process

Abstract Background In older patients with end-stage kidney disease (ESKD), the choice between kidney transplantation (KT) and dialysis may be more complex than in younger patients because of a higher prevalence of comorbidities and frailty. This study aims to provide greater insight into the current decision-making process by exploring the expectations, experiences and health outcome priorities of all stakeholders. Methods We performed semi-structured interviews with patients ≥65 years with ESKD (eGFR <15 ml/min/1.73m2, KT recipient or treated with dialysis), patients’ relatives and healthcare professionals (nephrologists, nurses and medical social workers). Interviews were conducted until data saturation and thematically analysed. Results We performed 36 interviews (patients n = 18, relatives n = 5, healthcare professionals n = 13). Thematic analysis revealed five themes. Older patients’ health outcome priorities were mostly related to quality of life (QOL). Individual older patients showed marked differences in the preferred level of engagement during the decision-making process (varying from ‘wants to be in the lead’ to ‘follows the nephrologist’) and in informational needs (varying from evidence-based to experience-based). On the contrary, healthcare professionals were quite unanimous on all aspects. They focused on determining eligibility for KT as start of the decision-making process, on clear and extensive information provision and on classical, medical outcomes. Conclusions The decision-making process could benefit from early identification of older patients’ values, needs and health outcome priorities, in parallel with assessment of KT eligibility and before discussing the treatment options, and the explicit use of this information in further steps of the decision-making process.


Introduction
A large proportion of the patients with end-stage kidney disease (ESKD) are older adults. In the USA, 74% of the patients treated with kidney replacement therapy (KRT) are aged 65 years or older (1). Of the two KRT modalities, kidney transplantation (KT) and dialysis, KT is generally considered superior because of a longer life expectancy and better quality of life (QOL) compared with dialysis (2)(3)(4). In 2019, 32% of the KTs in the Netherlands (±n = 300) were performed in patients aged 65 years or older (5).
However, KT might not be the best KRT modality for all older adults. Factors such as older age, comorbidities and frailty are associated with a higher risk of complications and mortality after KT (2,(6)(7)(8)(9). Consequently, the choice between KT and dialysis in older adults has to be tailored to the individual situation.
Clinical guidelines regarding the choice between KT and dialysis recommend shared-decision making (SDM) (10)(11)(12), which is 'an approach where clinicians and patients make decisions together using the best available evidence' (13). In older adults, SDM should have a goal-oriented approach instead of a disease-oriented approach (14,15). Thus, a key step of successful SDM in older adults is the early identification of patient values and goals of care (i.e. health outcome priorities) (14). Figure 1 shows the treatment options of older patients with ESKD and the people who are involved in this decisionmaking process. Little is known about the current decisionmaking process regarding the choice between KT and dialysis in older adults. For example, it is unknown whether SDM occurs, or what the health outcome priorities of the older adults facing this decision are. We therefore aimed to provide greater insight into this decision-making process by exploring the expectations, experiences, values and health outcome priorities of all persons involved in this process (i.e. older adults with ESKD, their relatives and healthcare professionals, see Figure 1). The results of this study could contribute to improved (shared) decision-making regarding the choice between KT and dialysis in older adults.

Study design
We performed a qualitative study with semi-structured interviews. The Consolidated criteria for Reporting Qualitative Health Research (COREQ) checklist (16) was used for the reporting of this study; Appendix 1 shows the completed checklist. The institutional review boards of the Radboudumc and Jeroen Bosch Hospital approved this study. This study was conducted in adherence to the Declaration of Helsinki.

Setting and participants
As we aimed to gain broad insight into the decisionmaking process, the study population consisted of all persons involved in the decision-making process in Dutch clinical practice: patients, patients' relatives and healthcare professionals (nephrologists, nurses and medical social workers). Because we wanted to collect data about expectations (about KRT) and experiences (with the decisionmaking process), we included both patients who were not yet on KRT (i.e. patients currently facing the decision between KT and dialysis) and patients already treated with KRT (i.e. patients who made the decision between KT and dialysis in the past). Additional inclusion criteria for patients were ESKD (i.e. estimated glomerular filtration rate (eGFR) < 15 ml/min/1.73m 2 , dialysis treatment or KT recipient), age 65 years or older (at start of KRT) and eligibility for both KT and dialysis (at start of KRT, judged by the treating nephrologist).
Participants were recruited from two Dutch hospitals: the Radboudumc (transplantation centre with dialysis ward) and the Jeroen Bosch Hospital (hospital with dialysis ward). The standard pre-dialysis and pre-transplantation educational programmes of the participating centres are described in Appendix 2. Potential participants were approached face to face, by telephone or via email. None of the approached subjects refused or withdrew participation. To assure broad diversity in the opinions and ideas to be collected, patients and healthcare professionals were recruited using purposive sampling. We aimed for variation in patient characteristics (e.g. age, gender, current KRT modality, treatment preference) and professional characteristics (e.g. age, gender, profession, clinical experience). Relatives were recruited via participating patients. Again, we aimed for variation in personal characteristics (e.g. age, gender, relation to patient). All participants provided written informed consent.

Data collection
The first version of the interview guide was developed by the research team (TS, AK, MP) based on an extensive literature search and clinical experience. During data collection and analysis, the interview guide was constantly adjusted based on new insights. Appendix 3 contains the final version of the interview guide.
Interviews were performed from July to December 2020 by two interviewers. Appendix 1 (domain 1) shows the personal characteristics of the interviewers. During the interviews, only the participant and interviewers were present. Interviews were performed face to face (in the hospital or at the participants' home), or, due to the coronavirus pandemic, via video call or telephone. Interviews were audio recorded and transcribed verbatim afterwards. Interviews were conducted until interviews did not reveal new concepts and themes (i.e. data saturation). The level of data saturation was assessed during frequent research meetings (TS, LHu, MP, AK) using the 'information power' model. This model enables systematic reflection on five items with an impact on the information power (i.e. study aim, sample specificity, use of established theory, quality of dialogue and analysis Figure 1. Treatment options of an older patient with ESKD, scope of this study and people that are involved in the decision-making process. This study focuses on the decision between KT and dialysis. Blue = people that are involved in the decision-making process. Red = main question of the decision making process. Green = treatment options. * Nephrologists, nurses, medical social workers * * Beyond the scope of this study strategy) (i.e. the more information the sample holds, the lower amount of participants is needed) (17).

Data analysis
Interviews were analysed using thematic analysis (18), supported by ATLAS.ti (version 8.4.20). In the first phase (open coding), all phenomena in the text were labelled by two researchers (TS, LHu) separately. Next, codes were compared, and similar codes were grouped together into axial codes (TS, LHu). New transcripts were analysed using a codebook with these axial codes by two researchers (TS, LHu) separately. During this phase of axial coding, existing axial codes were rephrased and new axial codes were added if necessary. The results were discussed in frequent research meetings (TS, LHu, MP, AK). Finally, all axial codes were categorised and themes were defined.

Results
We performed 36 interviews: 18 with patients, five with patients' relatives and 13 with healthcare professionals. Participants were recruited from the Radboudumc (n = 26) and the Jeroen Bosch Hospital (n = 10). Interviews were performed face to face (n = 25), via video call (patients n = 2, relatives n = 2, healthcare professionals n = 2), or by telephone (patients n = 2, healthcare professionals n = 3). Mean interview duration was 42 minutes (standard deviation 11 minutes). Table 1 summarises the population characteris- Quantitative data are presented as median (IQR) and categorical data as number of patients (percentage). Abbreviations: DD = deceased donor; DM = diabetes mellitus; HD = haemodialysis; IQR = interquartile range; LD = living donor; PD = peritoneal dialysis; TIN = tubulo-interstitial nephritis. * Educational level was categorised into low (no education, primary education or pre-vocational education), moderate (vocational education or selective secondary education) and high (university). # For patients' relatives: this displays the patients' (time on) current KRT modality (none of the relatives had kidney disease themselves). $ One patient pre-emptive.
tics, and shows that purposive sampling successfully resulted in variation in most predefined characteristics such as age and KRT modality. However, the number of participants with a high educational level among patients and relatives was limited. All participating patients were retired. Thematic analysis identified five themes. Table 2 shows these themes and the codes on which they are based. The five themes are described below and are illustrated by participant quotations (Table 3). Patients' relatives agreed with patients in all cases and are therefore not described separately.

Theme 1: KT and dialysis in older patients: attitudes and expectations
The attitude and expectations differed between patients preferring KT, patients preferring dialysis and healthcare professionals.
Patients preferring KT were convinced that KT is, without doubt, superior to dialysis. These patients choose KT because they expected to retain their 'normal' life after KT. After KT, they expected to be able to do what they want (e.g. hobbies, holidays, social contacts), to become more independent, to have more energy and/or to live longer. Some patients referred to KT as 'a miracle' or 'a cure' and could not mention any disadvantages of KT. Other patients acknowledged that KT has risks and that they would remain a patient for the rest of their life. However, all patients were convinced that they would not experience any (severe) complications or side effects. Their attitude towards dialysis was mainly negative. They expected (or experienced) that dialysis would have a negative impact on their daily life, QOL and/or life expectancy.
Patients preferring dialysis did not consider KT superior to dialysis. They felt that individual circumstances determine whether KT or dialysis is the best KRT modality for an Kidney transplantation or dialysis in older adults Abbreviations: HCP = healthcare professional.    'Of course, there are stories and information from doctors. But really learning about the experience itself. . . patients are the ones who are actually undergoing it. That will happen to me as well. That ís happening to me as well.'

Healthcare professionals Nephrologist (male)
'Most important is that we properly inform and educate patients. older patient. Their expectations regarding KT were more negatively charged than patients preferring KT. For example, some dreaded the surgery or the use of immunosuppressive drugs. All of them mentioned multiple disadvantages of KT, such as the risk of infections or surgical complications, which they considered a significant risk for themselves. Regarding dialysis, these patients expected (or experienced) few dialysisrelated restrictions or symptoms. In fact, they mentioned advantages of dialysis such as relieve symptoms since start of dialysis and the opportunity for social contacts they gained at the dialysis ward. Patients already on dialysis considered it as a normal part of their daily life. Healthcare professionals considered KT the best KRT modality, also for older patients. However, healthcare professionals emphasised that KT has significant risks, and therefore not every patient is eligible for KT. Given these risks and the scarcity of donor kidneys, healthcare professionals did not consider KT as something to 'just try' for all of their patients. Similar to patients preferring KT, healthcare professionals predominantly described dialysis in a negative way. However, they were more nuanced and mentioned several advantages of dialysis, such as the possibility for frequent monitoring of the patient, which could increase medication adherence, and potential social contacts for patients. Moreover, all of them also knew dialysis patients who did not have many (medical) problems.

Theme 2: KT or dialysis in older patients-Who makes the decision?
All participants reported that, in the end, the nephrologist decides whether a patient is eligible for KT or not. For this, patients trusted on the expertise of their nephrologist and most of them would accept it if he/she said that KT was not possible. For some nephrologists, this felt as a heavy responsibility because of the impact of their decision on their patients' life. They frequently consulted other healthcare professionals, such as fellow nephrologists.
In addition, older patients preferred different levels of patient-engagement, which varied from 'wanting to decide myself' to 'not wanting to decide at all'. Patients in the first group felt that they are the only one able to decide which treatment is best to them personally, whereas patients in the latter group felt that they lacked relevant (medical) knowledge, or they did not want to be responsible themselves. Healthcare professionals mentioned to be aware of these differences in preferred level of patient-engagement. Irrespective of patient preference, the majority of patients wanted to know the opinion of their nephrologist because they felt that this made the choice easier.
Nephrologists indicated that they feel somewhat uncomfortable, when an eligible patient decides to forego KT. Some nephrologists indicated that they try to convince such older patients to choose KT instead. However, all of them reported that the patient has the final say during this stage.

Theme 3: Assessment of KT eligibility: medical and personal values
Appendix 4 shows an overview of all variables that healthcare professionals and patients reported to use for assessing eligibility for KT.
Healthcare professionals reported that eligibility for KT is mainly determined by medical history and comorbidity. In addition, assessment of the level of independence, therapy compliance, vitality/physical condition and cognitive status were essential factors during the eligibility assessment. Ethical factors, mainly related to organ allocation and a responsibility to the donor, were also considered and played a more prominent role than in younger patients. Nephrologists felt that a patient's life expectancy should be long enough to justify the use of a scarce donor kidney. Nephrologists strived for objectivity in the eligibility assessment. However, all of them reported that, in reality, the eligibility assessment is not purely objective because of, for example, the influence of nephrologists' personal values and previous experiences.
In contrast to healthcare professionals, medical factors played a less prominent role for the majority of the patients. For example, some patients did not consider their medical history (e.g. diabetes mellitus) to be relevant for the eligibility assessment. Instead, patients frequently used personal values to assess if KT would be possible for them. Examples of such personal values are their attitude towards KT and dialysis and whether they felt they deserved the kidney. Previous (medical) experiences also influenced the way patients assessed the risks of a treatment. For example, a patient that experienced severe complications after a previous surgery was very concerned about the KT surgery.

Theme 4: Health outcome priorities
Appendix 5 displays all health outcomes that were considered by participants while making the choice between KT and dialysis. Although most health outcomes were mentioned by all participants, only healthcare professionals (not patients) mentioned 'kidney function' and 'patient satisfaction'. Concerning the latter it is important to note that patients mentioned several other factors that will eventually result in (dis)satisfaction with the treatment. In addition, only patients (not healthcare professionals) mentioned 'ability to perform caregiver duties', 'burden for relatives', 'feeling normal', 'feelings related to the presence of a donor organ' and 'freedom (being able to do what you want)'.
Healthcare professionals were quite unanimous in the way they valued health outcomes. They considered personal outcomes such as 'QOL' equally, or even more, important than medical outcomes such as 'life expectancy' and 'surgical risks'. The input of medical social workers, who frequently discussed personal goals of care with patients, was highly valued. However, most healthcare professionals, including medical social workers, struggled with the fact that these personal goals of care frequently do not have clear definitions and cannot be measured objectively. As a consequence, in the actual decision-making phase, healthcare professionals mainly focused on medical outcomes.
The health outcome priorities of most older patients included QOL-related factors instead of general medical outcomes, such as life expectancy. Frequently mentioned health outcome priorities were 'feeling normal', 'maintaining independence' and 'impact on daily life'. However, the goals of care differed greatly between individual patients. For instance, 'life expectancy' was considered the most important health outcome by some patients, whereas others did not consider 'life expectancy' relevant at all.

Theme 5: Information provision and older patients' needs
Healthcare professionals mentioned that they inform their patients in great detail about medical-technical aspects and risks of KT and dialysis. In addition to such factual, medical information, medical social workers also inform patients about more personal aspects, such as the impact of KRT on their (daily) life. They considered this information essential for patients to make a well-considered choice. Information provision was, if necessary, adjusted to the patients' (expected) level of education and cognitive status. Most healthcare professionals wanted to provide objective information only. However, some of them acknowledged that their personal values, preferences and opinion influence the way they counsel patients.
This way of information provision seems suitable for some of the patients: our results show that some older patients indeed want to know all medical details. These patients also use other sources such as websites of patient federations to gain information. However, other patients have different informational needs. We found that some older patients were not interested in medical details at all. These patients felt that knowledge about all potential risks and disadvantages of a treatment would only make them (even more) worrisome. They considered the experiences of other patients far more valuable than (only) factual, medical information.

Discussion
To our knowledge, this is the first qualitative study that explored the decision-making process regarding the choice Kidney transplantation or dialysis in older adults between KT and dialysis in older adults. Our study shows that individual older patients report marked differences in prioritisation of health outcomes, the preferred level of engagement during the decision-making process and informational needs, while healthcare professionals were quite unanimous in their views. Older patients' health outcome priorities were mostly related to QOL. Healthcare professionals also considered QOL-related outcomes important but found it difficult that they were neither objective nor measurable. Consequently, their focus tended to move towards more classical medical outcomes.
Previous research in older adults showed that, in general, older adults have different health outcome priorities than younger adults. For example, life expectancy is considered more important by younger adults (aged 21-60 years) than by older adults (aged >60 years) (19). Moreover, in community-dwelling frail older adults, goals of care are as often related to issues of well-being, such as living situation, social and family relationships, and mobility, as they are to medical problems (20). Similarly, many older adults with chronic kidney disease (CKD) prioritise QOL-related health outcomes, such as level of independence, above medical health outcomes, such as life expectancy (19,21,22). This highlights the importance of future research in the subgroup of older patients. SDM in older adults differs from 'regular' SDM in two aspects: it is goal-oriented instead of disease-oriented (14,15) and has different steps (14). After a short preparatory step, SDM in older adults should start with the identification of patient values and goals of care, and when discussing treatment possibilities, these factors should be taken into account. In other populations, such as patients with atrial fibrillation and older patients with cancer, explicitly discussing outcome priorities indeed provided new insights to healthcare professionals, and helped patients to make a treatment decision (23,24).
Our study shows that, in the current decision-making process, patients, relatives and healthcare professionals collaborate and share their views. For instance, the majority of the patients and relatives reported that they wanted to know the treatment preference of their nephrologist. Moreover, healthcare professionals were motivated to inform their patients about all treatment options and also advised them during the actual decision-making phase. This collaboration is in line with the 'collaborative deliberation' model that describes interpersonal aspects that affect decisions and with the model of successful SDM in older adults (14,25).
However, SDM in the decision-making process regarding the choice of KRT in older adults might be impeded by the discrepancy we observed between patients' health outcome priorities (mainly QOL-related outcomes) and the variables that healthcare professionals reported to take into account during the decision-making process (mainly medical outcomes) (14,26,27). Although medical social workers explicitly discussed values and goals of care with their patients, these factors were not prioritised during the actual decisionmaking phase. The main reason for this finding, according to healthcare professionals, was that QOL-related outcomes are difficult to objectify.
Previous research already showed this tendency towards a disease-oriented approach by healthcare professionals (14). Moreover, clinical guidelines are frequently diseasespecific and biomedically focused as well. Indeed, guidelines about decision-making recommendations in patients with ESKD mainly provide information about general medical outcomes, such as mortality and graft survival (2)(3)(4). Previous research also showed poor concordance between the health outcome priorities of patients with ESKD and healthcare professionals' perceptions on these priorities (21,28,29). Thus, the decision-making process could benefit from early identification of older patients' values and goals of care (in parallel with assessment of KT eligibility and before discussing the treatment options), and the explicit use of this information in further steps of the decision-making process.
This study has four limitations with potential impact on data saturation. First, we included only one patient with a high educational level. Educational level is associated with differences in patient involvement and needs during SDM (30). Socio-economic status, which is closely related to educational level, also influences doctor-patient communication (31). For example, patients from higher social classes receive a less directive and more participatory consulting style than patients from lower social classes (31). However, since the interviewed patients already presented a wide variety in personal values, prioritisation of health outcomes and preferred level of engagement during the decision-making process, we decided not to include additional patients with a high educational level. Second, we have only included participants from two hospitals and, in theory, inter-subject variation might be relatively small among participants from the same hospital. However, participants were recruited from one academic and one non-academic hospital, and they were successfully purposively sampled (e.g. on age, gender, KRT modality). A third limitation of this study is that patients already on dialysis and KT recipients might experience difficulties in recalling experiences regarding the decision-making process. However, we considered it valuable to include a wide variety of patients, including those that had completed the decision-making process. Lastly, data collection was not uniform: interviews were conducted face to face (n = 25), and, because of the COVID-19 pandemic, also via video call (n = 6) and by telephone (n = 5). However, previous research showed that face to face interviews are only marginally superior to video calling (32). Similarly, for interviews by telephone, there is no evidence of significant data loss, data distortion or reduced quality of the findings (33). This justifies the use of data collection via video calling and/or telephone, especially in case of limited time, budget, and/or access to people (32,34).
This study meets all important quality criteria for qualitative research. First, we ensured that all possible perspectives were captured in our data by applying purposive sampling and by including all stakeholders involved in the decisionmaking process. We approached the issue from multiple perspectives (data triangulation) by including different types of patients (i.e. patients not yet on KRT, dialysis patients and KT recipients), while most other studies focused on the perspective of a single patient category (21,(35)(36)(37)(38). Second, data collection and data analysis were parallel processes, which enabled exploration and verification of preliminary hypotheses during further data collection. Lastly, the influence of researchers' personal characteristics and presumptions was minimised using two coders who independently analysed the data, and by frequently discussing the results during research meetings. The researchers involved in data analyses also have different personal characteristics, occupations and (clinical) experience, which contributes to investigator triangulation.
Supplementary Material: Supplementary material can be found at Age and Ageing online.