A national study of practice patterns in UK renal units in the use of dialysis and conservative kidney management to treat people aged 75 years and over with chronic kidney failure

Design and setting

Qualitative study with semistructured interviews with patients recruited from nine renal units in England (Birmingham, Heartlands; Bristol, Southmead; Hull; London, King’s College; Manchester; Middlesbrough; Reading; Stevenage, Lister; Stoke-on-Trent). Renal units were purposively sampled to produce a diverse sample in terms of location in England and the scale of CKM delivery. The latter was estimated by responses provided by clinical directors to a previous survey by the UK Renal Registry. 40

Participants

Consultant nephrologists and nurses in each renal unit were asked to identify patients who were 75 years old or older and whose records indicated that they had an eGFR of less than 15 ml/minute/1.73 m² or were on dialysis. Patients were required to speak English fluently and be judged by their health-care professionals (HCPs) to be sufficiently fit, physically and mentally, to take part in an interview. The researchers aimed to sample patients purposively by stage of illness and management pathway as follows: (1) following the decision to opt for CKM (CKM pathway), (2) following the decision to have dialysis in the future, but before starting dialysis (pre-dialysis pathway) and (3) following the start of dialysis (dialysis pathway). Staff in each unit were responsible for identifying patients in each of these groups who met the inclusion criteria, and patients were invited to take part in the study either by post or in person. Demographic and other information for all patients invited to take part in the study was recorded (gender, age, ethnicity, date first seen at renal unit and date started CKM or dialysis if applicable).

Interviews

Patients were interviewed by an experienced qualitative researcher (ST-C), either face to face, in the patient’s own home or in their renal unit while they were on dialysis, or by telephone. The type of interview was determined by patient preference except for some interviews which had to be carried out by telephone because of the distance between the researcher and participant. The interviewer introduced herself as a non-clinical researcher to explain that she had both no medical training and no specific allegiance to the nephrology field. All patients gave written informed consent, either at the time of interview or before the interview if carried out by telephone. Interviews followed a semistructured interview guide which asked patients about their medical history in relation to their CKD, their contact with their renal unit, their knowledge and understanding about management options and their reasons for their management decision (see Appendix 1). A semistructured format was deemed suitable to ensure relevant questions were asked to all patients but also to allow patients the opportunity to talk about issues which were important to them. Interviews were audio-recorded and transcribed verbatim by an independent transcriptionist. Transcripts were checked by the interviewer to ensure accuracy. Recruitment and interviews continued until the interviewer was satisfied that the data indicated saturation.

Data analysis

Thematic analysis41 allowed an inductive approach to exploring the data, which lessened the likelihood that findings would be influenced by the existing literature or the researchers’ preconceptions. Transcripts were coded line by line, with codes being assigned to each meaningful segment of text. Transcripts were then compared with one another, using a constant comparison approach taken from grounded theory, to search for similarities and differences between interviews. 42 ST-C independently coded 20 interview transcripts and developed an initial set of themes. NVivo 9 (QSR International, Warrington, UK) was used to facilitate coding. Initial themes were discussed with the wider research team, and themes and subthemes were amended and renamed until a consensus was reached. This agreed framework was used to code the remaining 22 transcripts. Any new data occurring in transcripts which did not fit into the existing themes were highlighted and discussed further. New themes and subthemes were added, and existing themes amended, in the light of these new data.

Participant characteristics

Ninety patients were invited to take part in the study and 44 agreed. Eleven patients declined, five patients were unable to take part for health reasons, four patients died after being invited and 26 did not reply. Patients who did not take part were mostly CKM patients, which meant that more CKM patients had to be invited to take part (Table 1).

Forty-two patients were interviewed, 14 patients in each group (Table 2). All dialysis patients were on hospital haemodialysis (HD) except for two on continuous ambulatory peritoneal dialysis. Interviews ranged from 27 to 87 minutes with a median of 47 minutes. Interviews were carried out between May 2012 and February 2013 and were carried out in person except for five interviews carried out by telephone. There was no indication from the data that interviews carried out by telephone differed substantially from face-to-face interviews in content although interviews were slightly shorter on average. While some interviews took place with a family member in the same room, three of the CKM patients specifically wanted to be interviewed with a family member for support.

Sociodemographic characteristics did not differ substantially between the three groups. Participants’ ages ranged from 74 to 92 years with a mean of 82 years, most were men and most identified themselves as white British. Many patients were married or had a partner (n = 24, 57%) and most were living with their partner (n = 22, 52%). Other patients lived alone (n = 14, 33%), with their children (n = 3, 6%), their friends (n = 1, 2%) or in a care home (n = 2, 4%). Compared with the pre-dialysis and dialysis groups, patients opting for CKM were slightly older, were more likely to be female and were more likely to live alone or in a care home, but group numbers were small (Table 3).

At the time of interview, patients had been in contact with their renal unit for a median of 49 months (range from 5 to 131 months). Dialysis patients had been on dialysis for a median of 10.5 months (range from 1 to 120 months) and patients who had opted for CKM had done so a median of 11 months previously (range from 1 to 83 months). Although recruitment had aimed to interview CKD5 patients, some patients reported that their eGFR was between 20 ml/minute/1.73 m² and 15 ml/minute/1.73 m² in interviews, indicating that they had CKD4.

Renal unit characteristics

Results from the national survey (see Chapter 4) confirmed variation in the scale of CKM delivery in the nine units sampled (Table 4). Three of the nine renal units reported that, of their patients over 75 years old, under 10% were receiving CKM, as opposed to other units, where up to 50% of over-75-year-olds were opting for CKM. The same three units reported that the terminology used to refer to CKM was usually phrased as ‘non-dialysis’. Based on these data, units were classified into two groups: those with more established CKM pathways (units 1, 2, 5, 6, 8 and 9) and those with less established CKM pathways (units 3, 4 and 7).

Qualitative findings

Four themes emerged from the analysis of all interview transcripts (Figure 1). All themes were relevant to all patients, but differences within themes emerged between patients who had chosen different treatment options and between patients from different units.

FIGURE 1.

A thematic map of the four themes identified from the analysis of 42 interviews.

The diagram indicates how interaction with staff fed into patients’ conceptualisation of the process between understanding CKD and making (and occasionally revising) a management decision for their own CKD.

Main findings

While CKD patients consider the same factors when making a treatment decision, patients who choose different treatments hold contrasting beliefs about what dialysis can offer. These beliefs appear to be influenced by the information provided by renal staff, which can differ between units, particularly in regard to CKM. It was noticeable that few patients were aware of CKM as an option if they had not chosen it, although patients from units with a more established CKM pathway appeared to be more aware of CKM. While most acknowledged the severity of their CKD, some CKM patients did not appear to think of their CKD as serious because of its asymptomatic nature, despite information from staff indicating otherwise. There was a distinct divide between CKM patients, and dialysis and pre-dialysis patients in whether people believed they would live longer on dialysis or not and whether they expected their quality of life to get better or worse on dialysis. Information from units with less established CKM pathways appeared to focus on the number of additional years a patient was likely to live on dialysis. Patients from units with more established CKM pathways, however, described being told that living longer on dialysis was not a guarantee and in addition that choosing CKM was ‘not unusual’. Finally, few patients reported speaking to staff about the future, in terms of the consequences of either starting dialysis or opting for CKM. Patients from units with more established CKM pathways appeared to have discussed the future with renal staff and some indicated that they had begun conversations about ACP. For others, being unaware of how their disease was likely to progress added to the misperceptions some patients appeared to hold about the severity of their CKD and the need for dialysis.

Comparison with existing research

Particular results mirrored those identified in previous qualitative research. 36–39 CKM patients reported that they had chosen CKM because they felt too old for dialysis, they were worried about being a burden on family/society and they were concerned about travelling to have dialysis. While these factors can be valid reasons for making a decision there is also a sense that some patients may feel dialysis is not really a possibility or that they do not have a right to dialysis. Thinking of oneself as a burden on society, if dialysis is chosen, may lead to feelings of guilt for patients, who may benefit from extra time to talk through their decision with staff. Equally, concerns about whether transport is a possibility or not, rather than a preference to avoid transport, are different thought processes, and equity of access to treatments needs to be assured.

Reluctance to think about the future and feeling well without dialysis have been identified in previous studies37,38,43 and they linked to the current findings about perceptions of CKD severity and revising decisions. Feeling well, with few symptoms, seemed to make some patients think their CKD was less severe, which meant they did not see a reason to consider dialysis. A reluctance to think about the future and possibilities of becoming more ill had led some to choose CKM initially and then change their decision later on. This may be unavoidable for some patients who are unable to understand their CKD until they feel symptoms, although clinicians should be wary about using the term ‘CKM’ to refer to such patients. Patients who are in ESKF often want to know their prognosis but clinicians may feel uncertain or uncomfortable about when to discuss this with patients. 43 Open discussions about prognosis are likely to have a substantial influence on discussions about treatment options and feed into better shared decision-making.

Results also supported previous qualitative and quantitative work which showed how CKM patients may choose quality of life over length of life. 39,44,45 While this indicated different priorities between patients, the current results highlighted how patients held contrasting beliefs about whether dialysis would extend life or not, indicating the importance of patient expectations as well as priorities. Results also supported previous work which indicates that patients may choose CKM if they feel they have achieved everything they have wanted to in life. 39 This contentment and feeling of having had a ‘complete life’ has been suggested to lessen anxiety about death and end of life. 39

In addition to existing research, the current study identified the influence staff, and information from the renal unit, had had on patient decision-making. Patients from different units reported being provided with varying types of information, which presented CKM in a more or less positive light. It appeared that patients from units with a more established CKM pathway were more likely to know what CKM was, had been given more information about CKM and, for those who had chosen it, had discussed the consequences of their decision in more detail. While there may be recall issues with patient reports, results suggest that there are unit variations in the way in which older CKD patients are informed.

Strengths and limitations

This was the first qualitative study to explore patients’ views of choosing between dialysis and CKM across different renal units. Including nine units in the study meant that the sample contained diversity in terms of both patients’ treatment decisions as well as the service delivery experienced. We were able to include units which had more or less established CKM pathways in order to explore variation in what patients had been told.

Qualitative results cannot be generalised to other populations but data gathered can identify important issues and offer conceptual transferability. Indeed, the findings presented resonate with the existing literature. While interviewing methods run the risk of obtaining socially desirable responses, the interviewer presented herself as an impartial observer who had no link to the renal units or staff. Moreover, patients mentioned negative aspects of the care they had received, which suggests that they felt able to speak freely and in confidence. Finally, as with all interview studies, the findings do not provide a window on events as they happened but rather provide insight into how participants construct their experiences of relevant events.

Although the study had clear entry criteria, one participant under 75 years was invited to take part and three participants reported they were CKD4 rather than CKD5. Recruitment was difficult and time-consuming for renal staff and it was not always clear that participants did not match the inclusion criteria until part way through an interview. Data on patients’ eGFR over time were not collected. It is likely that patients with both advanced but stable kidney function and those with a linear, progressive, predictable decline in kidney function were recruited. Such differences may have had implications for how treatment options were discussed because of the clinician’s (un)certainty of how likely a patient was to die from their CKD rather than another condition.

The study excluded participants who did not speak English fluently, in order to ensure that participants were able to provide full responses to questions without the need for a translator. This limitation meant that the study was less likely to recruit patients from black and minority ethnic groups. Recruitment of such patients would have required additional resources which were unavailable and somewhat unfeasible for this study and may be better suited to a more focused qualitative study looking specifically at this population.

Recruitment of advanced CKM patients was particularly difficult. CKM patients most often had an eGFR of over 10 ml/minute/1.73 m² and few felt they had any symptoms from their CKD. This meant CKM patients who were symptomatic and who otherwise would have started dialysis were not well sampled; interviewing more of this type of patient might have provided additional insights into why CKM is chosen by some and would have allowed more direct comparison of CKM versus dialysis patients. However, such patients may not have been physically and mentally well enough to take part. This is likely to be a limitation of any qualitative work carried out with a chronically ill older population across several centres.

The views of dialysis, pre-dialysis and CKM patients were compared. While all patients were similar in age, patients differed greatly in their physical and mental health, within and between treatment groups. The choice between dialysis and CKM is most relevant to older adults who are frail with other comorbidities. While we tried to recruit such patients, this was not always possible and some CKD patients who were not frail or burdened with multiple comorbidity were recruited to the study. For these patients it may be unsurprising that less emphasis had been placed on CKM as an option by renal unit staff.

The number of renal units in the study captured variety in CKM delivery. However, this influenced the number of patients interviewed per unit, which had to be limited in order to obtain a manageable data set which could be explored in depth. Future qualitative work may benefit from sampling more patients from fewer units in order to follow-up initial data which suggests that patients attending different units are receiving different information about treatment options.

The categorisation of units into two groups representing units with a more or less established CKM pathway was relatively crude. Inevitably there was some overlap in policy and practice between the two groups; however, the categorisation helped us to look at general trends in the data.

Implications

Patient reports indicated that they had been given different types of information, at different time points, by their renal unit and that units with a more established CKM pathway appeared to have provided more detail about the option of CKM than others. While this may have reflected the suitability of dialysis or CKM for the individual patient, it was interesting that patients from units with a more established CKM pathway were aware of both treatments regardless of what their final decision had been. Having fewer resources dedicated to CKM may reflect a general trend in a unit to encourage dialysis for all patients. However, it may also reflect less experience in providing care for CKM patients. Staff with access to fewer resources and less experience in providing a type of care may be less likely to discuss such care with patients, which may bias discussions towards dialysis. Staff may also be less supported in these discussions if they are not provided with materials which help to explain the CKM pathway or with training in how to have such discussion with patients. Finally staff may not feel comfortable discussing CKM with patients if they believe that dialysis will always offer more benefit despite comorbidity or frailty. Research which can provide data on the comparative outcomes of CKM and dialysis patients will help to clarify whether dialysis is beneficial for certain patients or not.

It was interesting to note that some patients had opted for CKM several months before interview, in one instance 7 years previously. It was also apparent that some patients had initially opted for CKM when they were well and changed their mind once they had experienced symptoms from their CKD. Both situations suggest that the label of CKM is being used for a very broad population, arguably one that is much larger and more diverse than the label ‘CKM’ would initially suggest.

Conservative kidney management is an alternative to dialysis and therefore could be strictly defined as applying only to patients who have passed the point at which dialysis would usually have been started, though this point can be difficult to define, especially in patients who may have symptoms related to other comorbidities. While this time point will be different for individual patients and judged differently between clinicians, a consensus based on eGFR or symptoms linked to kidney failure may be possible and would reduce the number of patients with asymptomatic, stable CKD being labelled as receiving CKM. Delaying a label of CKM would allow patients more time to think about their decision. It may lead to patients experiencing more symptoms from their CKD and give them greater awareness of how much their CKD is affecting their life. Having a more standard approach to the labelling of CKM, specifying separately those patients who currently plan to have CKM in the future and those who are currently receiving CKM, would provide a clearer view of the numbers of patients on this pathway and the variation between units. Our interviews emphasise that both categories of patients retain the option of changing their plan or choice of treatment.

Both dialysis and CKM patients reported they had not discussed the future with staff. Regardless of what treatment decision is made, it is important that renal staff discuss the likely trajectory of illness with patients. Previous literature has highlighted this issue and recommended earlier, rather than later, conversations about ACP to promote optimal end-of-life care. 46,47

Finally, quantitative results from the CKMAPPS survey indicate that units use different terminology for CKM and that there is a subgroup who refer to CKM as ‘non-dialysis’. CKM is accepted by most as an alternative to dialysis for a subset of patients, and patients may benefit if the option of CKM is presented in a consistent and unbiased manner across renal units. Units that are organised to support staff in the discussion of CKM options with patients will deliver patient benefit by promoting greater shared decision-making. 48,49 Clinicians should consider the implications of having a pathway framed as a negative option. Consistency in terminology would help establish CKM as a clearly defined and appropriate pathway for some patients.

Conclusion

Our results indicated that older adults with CKD5 who have chosen different treatment options have contrasting beliefs about what dialysis will offer them. Patients’ decisions were influenced by the information provided by their renal units, which differed between units with more and less established CKM pathways. Supporting staff in discussing CKM as a valid alternative for a subset of patients across all renal units will promote informed decision-making and thereby stand to benefit patients.

Design and setting

Semistructured interviews were conducted with staff members in nine renal units across England. These units were the same as those involved in the patient interview study (see Chapter 2) and were purposively sampled to obtain a range of locations and scale of CKM practice.

Participants

Consultant nephrologists in each participating unit were approached by the principal investigators and asked to identify a minimum of five staff members who were involved in the care of CKM patients in their unit. For units that had very few CKM patients, staff who cared for patients in low-clearance clinics or for those whose eGFR was less than 20 ml/minute/1.73 m² were recruited. We stipulated that the staff members interviewed had to include a minimum of one lead nephrologist and one nurse per unit. Allied health professionals were also invited to participate in the study.

Interviews

Participants were interviewed by IO and ST-C, either face to face at their renal unit or by telephone. All participants signed a consent form either at the time of interview or before the interview if they were interviewed over the telephone. Semistructured interviews were conducted following an interview schedule (see Appendix 2). The interview schedule was initially constructed from a literature review and discussion with steering group members and was then developed iteratively as interviews were carried out. The interview schedule asked participants to discuss their views of CKM, their experiences of being involved in decision-making about CKM, how CKM was delivered in units and the role of primary and palliative care services in caring for CKM patients. Interviews were digitally audio-recorded and transcribed verbatim, and interviewers verified the accuracy of transcription by listening to the interview and reading transcripts.

Data analysis

A content analysis of the 60 interviews was carried out to inform the survey questions (see Chapter 4). This consisted of several steps and included rereading the interview transcripts, identifying frequency of words, phrases and themes, and then developing categories of words, phrases or themes with similar meaning. 52 The data from the qualitative study helped to identify the sections of the survey and facilitated the formulation of the survey questions and multiple choice options.

Subsequently, a more detailed thematic analysis41 of the qualitative data was conducted. NVivo 10 was used to facilitate coding of the data. From the 60 staff interviews, 28 interviews were sampled using maximum variation sampling to ensure variation of units and experience of being involved with renal patients. Fifteen interviews were analysed in the first instance using thematic analysis by IO. 41 ST-C independently coded 10 of the 15 interviews. These interviews were selected from four units, which were the first participants recruited to the study. This sample was selected to ensure variation in units, staff roles and their experience of being involved with renal patients. Researchers reread the interview transcripts in order to identify emerging codes and to develop an initial coding framework. A further 13 interviews were selected to ensure variation of units and experience of being involved with renal patients; they were then coded by IO, ST-C and a third researcher, CE, using the preliminary framework, which was refined and further developed into the final thematic framework. Discrepancies were discussed and resolved by consensus. Data from the remaining 32 transcripts also added to the final thematic framework and facilitated saturation of the themes, which was achieved when no new codes or themes emerged from the data.

Participant characteristics

Sixty renal staff members were interviewed between February 2012 and November 2012. Table 5 summarises characteristics of all those interviewed and of the 28 staff whose interviews were analysed using in-depth thematic analysis.

Qualitative findings

Three themes emerged from the detailed analysis of 28 interviews and the analysis of the remaining 32 interviews.

Main findings

This is the first qualitative study in the UK exploring renal staff’s views and experiences of treating patients who choose CKM. CKM was generally accepted by renal staff as a treatment option; however, they had mixed views concerning CKM, which were reflected in the varied terminology that was used to describe CKM. Most considered that CKM offered a valuable treatment option for patients, while some did not perceive that CKM offered an alternative to standard care. Most staff commented that it was difficult to assess whether patients were suitable for RRT or CKM, but all were very supportive of having open discussions and of informing patients of their treatment options, and ensuring that family members were involved in this decision-making process. While many staff considered that it was important for patients and their family to make their own decisions based on the information that they had been given, some had a more directed approach and would guide patients to a decision, particularly if they felt that patients would not benefit from dialysis. Decision-making about treatment options, including about CKM, was acknowledged as challenging for both patients and staff, as this often concerned future treatment. Explaining CKM to patients was also found difficult because staff would need to discuss the consequence of CKM (death) with patients while explaining the treatment positively. There were also some CKM patients who suddenly changed their minds and opted for dialysis, and many staff emphasised the importance of revisiting patients’ decision over time, implying that the decision-making was a process rather than a one-off event. It was perceived that having a good relationship with the patient and interpersonal continuity would facilitate good decision-making. Towards the end of their life, CKM patients were often referred back to their GP, and it was vital for renal units to work in collaboration with the primary, community and palliative care teams, often supporting them to care for CKM patients by providing them with renal-specific education.

Comparison with existing qualitative research

Only a few qualitative studies have explored renal staff’s experiences with patients on RRT. 44,50–51 None have focused on staff’s experiences of being involved in the care of patients on CKM. Despite this difference in focus, some of the previous literature can be related to our findings.

Much of the previous literature has explored the difficulties renal staff face when making the decision whether or not to initiate dialysis and/or to withdraw dialysis. 50,51 This may be because it was difficult to assess whether or not dialysis would be in the patient’s best interests and good for their quality of life. Consequently, staff members wished to have more time for careful and open discussions with patients and their family. 51 The current study also identified that staff found it difficult to assess whether patients would be more suitable for dialysis or CKM, and many staff recognised the importance of having open discussions with patients and family. Previous work also identified that renal staff tended to avoid discussing prognosis and end-of-life issues, and many believed that such discussions would be interpreted as negative and would diminish the patients’ hope. 44 Discussing end of life at an early stage was also perceived as ‘not practical’ and challenging for renal staff, who tended to engage in such conversations at a later stage or only when prompted by patients. 44,53 Our findings concur, as renal staff held similar views; for example, some staff might face difficulty in discussing the consequence of choosing CKM (i.e. end-of-life issues). It was also recognised in a previous study that having an ongoing relationship between renal staff and patients, and continuity of care was essential to good communication for end-of-life issues,54 and renal staff in this study also held the same views. The current study also demonstrated that the decision whether to opt for CKM or not was a decision patients had to make for their future treatment. Staff discussed that many patients found this difficult, and some who had opted for CKM would consequently change their minds and request dialysis when they became more symptomatic or when their family wanted them to have dialysis. This implies that decision-making about their treatment options is an ongoing process rather than a one-off activity, and is a product of an iterative process of information assessment as suggested by other literature. 55

In previous studies, not having dialysis was perceived as implying imminent death by both patients and renal staff, and this perception made it difficult to discuss issues around not initiating or withdrawing dialysis. 50,51 This perception was not shared by the participants in this current study. Instead they perceived CKM as an active treatment option that could help with the management of symptoms of kidney failure without using RRT. Other studies have suggested that renal staff were also burdened by having sole responsibility and by being the final decision-maker in life-or-death decisions for patients. 50,51 In the current study, however, renal staff emphasised the importance of patients’ informed decisions, which facilitated patient choice in the decision whether to opt for CKM or not. Open discussion was seen to enable patients to make their own choices and also alleviated the burden on the renal staff of making treatment decisions.

Strength and limitations

This is the first study to explore renal HCPs’ views and experiences with patients who opted for CKM. The strengths of this study include the participation of a wide range of renal HCPs from units from a range of locations and a range of CKM practices. However, as the primary aim of this qualitative study was to explore the key issues surrounding CKM practice patterns, we recruited only HCPs who were involved in the care of CKM patients. As a result, it is likely that only individuals with an interest in CKM were interviewed in this study, and levels of renal staff’s keenness or willingness to provide CKM in the UK in general may be much lower than shown in this study. A further limitation is that the researcher’s presence during the data collection might have elicited socially desirable responses from the study participants; however, the researcher presented as an impartial observer rather than being actively engaged with the participants during the data collection. Furthermore, collecting data through interviews does not allow us to open a window on events as they happened but can instead provide insights into how participants construct their experience of the event. These findings resonate with other literature and, although qualitative findings cannot be generalised to other populations, the data can sensitise us to important issues and can offer transferability to other populations and to other settings.

Implications for practice

The findings of this study showed that renal staff faced challenges in discussing CKM and/or end-of-life issues with patients. The provision of education, such as training regarding communication skills, CKM and/or end-of-life issues for renal staff, may be beneficial. Better decision aid may also help facilitate discussion about treatment options. Staff in our study also expressed that they sometimes found it difficult to assess whether patients were suitable for RRT or CKM, and this may make staff feel less comfortable discussing CKM with patients. Research which compares outcomes between patients who receive CKM and similar patients who receive RRT will inform shared and informed decision-making when contemplating the suitability of CKM and RRT for a patient.

Our study also demonstrated that working together with primary, community and palliative care teams was key to the provision of care for CKM patients, emphasising the importance of communication between renal units and those services. A clear record of the CKM decision in the primary care record will be of benefit in minimising inappropriate hospital admissions and/or dialysis. Educational support for primary, community and palliative care teams regarding the care for CKM and renal patients in general can facilitate this process. This may involve developing a formal strategy for knowledge translation that includes better educational tools and measurement of the effectiveness of those tools. 56 In this study some units described providing educational material to other HCPs, which renal staff reported was successful; however, the effectiveness of such education is still not clear. Research is therefore also needed to understand the effectiveness of specific education/interventions.

Currently there is no financial payment for CKM under the Payment by Results tariff scheme and this was pointed out as a barrier to seeing CKM patients in general nephrology clinics. The current Payment by Results system may need to be reviewed to support the provision of CKM. There is a need to increase opportunities for units to obtain funding dedicated to providing CKM.

There is also a need for more standardised terminology for CKM in the renal community, which will enable accurate communication for clinical and research purposes.

Conclusions

Conservative kidney management is generally accepted by renal staff, and perceived as a valuable option for patients who would not medically benefit from dialysis. Patient informed choice and open discussion at an early stage of the patient illness trajectory were considered important, even though staff reported difficulties in discussing CKM as an option with patients. Providing staff with training about communication, CKM and/or end-of-life care would facilitate the decision-making process. A more standard terminology for CKM is needed, as different terminology may lead to different conceptualisations about CKM.

Survey development and distribution

The content of the survey was developed based on existing literature and findings from the staff qualitative study and feedback from the steering group.

Based on the staff qualitative study, the following main themes were identified as areas to investigate: how CKM is provided (service components, resources available for CKM, future improvement); how CKM is discussed with patients; and roles of primary care and palliative care services. Using the staff interview guide, interview questions regarding the above areas were selected and revised to make them more suitable for the survey questions. Data from the qualitative study also helped to organise multiple-choice options under each question.

The draft survey was piloted using cognitive interviews with three nephrologists and one renal nurse specialist. Two forms (web-based and postal) of the survey were used. The survey focused mainly on the management of patients aged 75 years and over with ESKF (CKD5) who did not or had opted not to have dialysis (i.e. who were on CKM). In order to supplement the renal unit data publicly available from the UK Renal Registry, the survey also included general questions regarding the management of patients with CKD (see Appendix 3 for the survey).

Contact details of clinical directors from all 71 UK renal units were obtained from the UK Renal Registry, and both versions of the survey were sent to them in March 2013. Telephone and e-mail follow-ups were conducted after the first mailing in April and May 2013.

Data analysis

Data analysis was conducted using basic statistics. Cross-tabulation was conducted to explore the relationship between practice patterns and selected key factors which we hypothesised might lead to differences in such patterns. These were factors related to the general unit organisation for patients with CKD; the size of units in terms of numbers of CKM patients; and resources available for CKM (e.g. funding). Closed-ended questions related to the above factors in the survey were selected for the analysis (see Appendix 4). Units were categorised into two groups, larger and smaller, based on responses to questions on numbers of CKM patients and numbers of symptomatic CKM patients (see Appendix 5 for details). We related unit size to the number of patients aged 75 years and over on RRT derived from UK Renal Registry data on prevalent RRT in 2012.

We tested for associations using these categorical variables using χ² and Fishers exact test; given the potential for multiple testing and false positives we report only associations that were significant at p < 0.01. We report some descriptive comparisons of interest in order to measure how much time renal staff were involved in CKM, full-time equivalent (FTE) time was asked. A FTE of 1.0 indicates that a person is equivalent to a full-time worker, or two persons working half-time.

In order to analyse resources and cost to provide services for CKM patients, the following questions from the survey were used.

• Question 35. Of those, how many were on conservative care and followed up in your unit?

• (This question was under Q34. In calendar year 2012, approximately how many CKD5 patients aged 75 and over were cared for by your renal service?)

• Question 37. In 2012, how many patients aged 75 and over in your unit chose to have conservative care, became symptomatic of advanced CKD and did not have dialysis?

• Question 39. How much time do the following staff have specifically allocated for CKD 5 patients on conservative care? Please enter number of fulltime equivalent (FTE) hours for each discipline.

• Question 47. How much annual funding was dedicated to providing conservative care in the 2011/12 financial year (April 2011–March 2012)?

Services for chronic kidney disease patients

Almost all responding units (66 of 67, 99%) had an MSRT and, of those, 58 of 66 (87%) had regular MSRT meetings. Half of these teams (33 of 58) met once a week and in 14 units, teams met once a month. When asked which staff members either were involved in their MSRT or usually attended the MSRT meeting, most had a dietitian (51 of 58, 88%) and a pre-dialysis education provider (46 of 58, 79%), but only 18 (31%) and 16 (27%) of responding units had a psychologist and a counsellor respectively (Table 6).

In addition to clinics at their renal unit, the majority of units (59 of 67, 88%) also ran clinics for CKD patients in neighbouring hospitals. The number of neighbouring hospitals they served ranged from one to 12 per unit, with a median of three.

Many units (56 of 67, 84%) had a formal ‘pre-dialysis’ clinic (or equivalent) for managing patients approaching RRT. Of those, however, 15 units responded that not all consultants who had CKD patients used the pre-dialysis clinic. Reasons for this varied: three reported that it was because some consultants thought that long-term continuity of care by the same consultant was more important, while another three said it was because some consultants’ clinics were at neighbouring hospitals but the pre-dialysis clinic was in the main hospital and they wanted their patients to avoid additional travel. Two respondents indicated that both of these factors were important factors and seven gave other reasons: for example, because consultants had different clinics; there were different models across their geographical area; and MSRT members could not support all pre-dialysis patients.

The majority of units (54 of 67, 81%) had a pre-dialysis education day (a group session with other pre-dialysis patients). During the education day, many responding units reported that they covered topics such as transplantation (53 of 53 responding units covered), dietary restrictions (53 of 53), types of dialysis (52 of 53), fluid balance (47 of 53), medicines (46 of 53), side effects (45 of 53), CKD-related anaemia (45 of 53) and psychological support (42 of 53). However, some topics, such as sexual matters (20 of 53), cardiovascular risk factors (35 of 53), and renal bone disease (39 of 53) were less commonly covered. Conservative care was included in most (45 of 53).

How consultants shared responsibility of patients with each other varied between the units: 18 of 66 (27%) units shared responsibility, whereas in 19 units (29%) consultants took sole responsibility for individual patients. Twenty units (30%) indicated that they shared responsibility for most patients but took a lead role for individual patients with particular needs. Nine units (14%) had more a mixed approach: for example, they shared inpatients but not outpatients; and consultants had both dedicated patient caseloads and shared patient caseloads.

Services for patients who chose conservative kidney management

Sixty-six of 67 renal units (99%) had patients with ESKF who made an active decision not to undergo dialysis even when they were symptomatic. The unit without such patients reported that their patients were well informed of the possible limitations and side effects of RRT, and RRT was still their choice.

When asked what was the most commonly used word in their unit to refer to CKM, respondents reported a variety of terms: ‘conservative management’ (30 of 65, 46%), ‘conservative care management’ (8 of 65, 12%), ‘supportive care’ (5 of 65, 8%) and ‘conservative kidney management’ (3 of 65, 5%). Eleven out of 65 units (17%) reported they used more than one term to refer to CKM.

Fifty-one out of 66 units (77%) reported that all consultant nephrologists followed the same practice regarding CKM patients. The remaining units indicated that some consultants acted differently, their practice differing only slightly (10 of 15) or moderately (5 of 15) from one another.

Resources and conservative kidney management funding

Staff resources

About half of the units (28 of 65, 45%) reported that they had staff whose time was specifically allocated for CKD5 patients who were on a CKM pathway. The larger units were more likely to have such staff (15 of 24) than the smaller units (8 of 23) (Table 9).

TABLE 9 - Cross-tabulation of the relationship between CKM size and the availability of staff whose time is dedicated to CKM

CKM size	n	Availability of staff whose time is dedicated to CKM
		Yes (%)	No (%)
Larger	24	63	37
Smaller	23	35	65
No response	18	28	72
Total	65	43	57

Twenty-four of the 28 units provided data on how much time each staff member had specifically allocated for CKM patients (see Table 10). Sixteen units had renal nurses whose time was specifically allocated for such patients. A median FTE of these 16 units was 0.9 (IQR 0.5–1.0). Twelve units responded that consultant nephrologists had allocated time for CKM patients with a median FTE of 0.2 (IQR 0.1–0.2). Nine units had either pre-dialysis education providers or dietitians whose time was allocated for CKM patients. These units had median FTEs of 0.4 (IQR 0.18–1.0) and 0.2 (IQR 0.1–0.35) respectively. The relationship between number of CKM patients and total staff time dedicated to CKM is shown in Figure 2. Renal registrars, occupational therapists, pharmacists, counsellors and management staff were reported by up to three units for each.

TABLE 10 - The number of FTE hours dedicated to caring for patients with CKM

Staff	n	Median	Minimum	Maximum	Percentiles
					25	75
Consultant nephrologists	12	0.20	0.05	1.00	0.10	0.20
Renal nurses	16	0.90	0.30	2.60	0.50	1.00
Social workers	7	0.20	0.05	1.00	0.05	0.20
Dietitians	9	0.20	0.10	1.00	0.10	0.35
Psychologists	5	0.10	0.05	0.20	0.08	0.20
Pre-dialysis education providers	9	0.40	0.10	1.00	0.18	1.00
Anaemia nurses	6	0.15	0.10	1.00	0.10	0.44

FIGURE 2.

The relationship between total staff hours dedicated to CKM and the number of CKM patients aged ≥ 75 years per unit.

Twenty-two of 28 units (79%) with staff whose time was specifically allocated for CKM patients had staff primarily responsible for CKM, whereas 54% of the units (20 of 37) without such dedicated staff had staff primarily responsible for CKM. Half of the units with dedicated staff (14 of 28) had a written guideline for CKM, whereas fewer than a quarter of the other units (9 of 37) had a guideline. Units with dedicated staff were more likely to have dedicated CKM clinics (10 of 28) than those without (5 of 37). Many units with dedicated staff provided staff with formal training about CKM (20 of 28), whereas only one-third of other units had such training (13 of 37) (p = 0.009). Twenty-one out of 28 units with dedicated staff had a written guideline for renal end-of-life care, whereas only fewer than half of the other units (15 of 37) had such a guideline.

Staff training

Half of the units (33 of 66) provided renal staff with formal training or education regarding CKM, and 12 (18%) of 66 units were developing such training. One-third (21 of 63) did not provide their staff with formal training; common reasons for this were lack of time (11 of 21) and lack of funding (8 of 21). Five units also reported that staff did not need formal training, as CKM was an ingrained culture in the unit.

Development of conservative kidney management

The key influences on the reported development of CKM in units were nurses’ attitudes (57 of 64, 89%), nephrologists’ attitudes (51 of 64, 80%) and patients’/families’ attitudes towards CKM (45 of 65, 69%). Two units (2 of 62, 3%) indicated that the Payment by Results tariff system for dialysis negatively influenced the development of CKM. Both such units were in England.

Most units (43 of 61, 70%) indicated that frequency of late referrals did not affect the development of the CKM programme. Just over half of the units (35 of 64, 55%) reported that the availability of funding specifically for CKM had no influence, while 14 (22%) and 15 (23%) of 65 units respectively indicated that it positively and negatively influenced the development of CKM.

Discussing conservative care with patients and assessing suitability

The majority of units (56 of 65, 86%) reported that they discussed the option of CKM with all CKD5 patients aged 75 years and over. The other nine units reported that consultant nephrologists tended to decide whether or not to discuss CKM with such patients. There was no unit which only discussed the CKM option if a patient or carer asked about the alternative to dialysis.

Various factors influenced staff when they contemplated the suitability of conservative care for a patient (Figure 3). All 65 respondents reported that patients’ preference for CKM influenced staff strongly or very strongly when considering the suitability of conservative care. This factor was followed by the extent and severity of comorbidities (61 of 65, 94%), frailty (58 of 65, 89%), functional status (52 of 65, 80%), patient’s current quality of life (50 of 65, 77%) and cognitive status (48 of 65, 74%). In addition, 12 of 64 (19%) and 11 of 65 (17%) units respectively reported that carer and consultant preferences for CKM strongly or very strongly influenced them. The distance from the dialysis unit to a patient’s home had the least influence on decisions about the suitability of conservative care (2 of 64, 3%). Uraemic symptoms and rate of decline in kidney function did not have a major influence.

FIGURE 3.

Factors likely to influence staff when contemplating the suitability of conservative care for a patient (n = 65). a, ‘Would I be surprised if this patient died in the next year?’ b, n = 64 because of missing responses.

About half of the units (37 of 65, 57%) indicated that the option of CKM was most commonly first raised with a patient when they were referred to the pre-dialysis clinic. Fifteen of 65 units (23%) reported that CKM was first raised when a patient’s eGFR reached a certain level: 10 units when eGFR reached 20 ml/minute/1.73 m², two units at eGFR level 19 ml/minute/1.73 m² and three units at eGFR level 15 ml/minute/1.73 m². In six units (9%) the option of CKM was usually first raised with patients at a specific time before the anticipated start of dialysis (from 3 to 12 months with a median of 9 months).

Working with primary care and general practitioners

No unit reported referring all CKM patients aged 75 years and over back to GPs to be cared for under primary care only. Some units referred patients back to GPs but GPs shared care of patients with the renal unit (12 of 65, 19%), whereas in some units patients were primarily cared for by the renal unit with little GP involvement (7 of 65, 11%). More than half of the units (40 of 65, 62%) indicated that GPs’ involvement was varied in line with the three options above. Thirty-one units indicated that this decision varied by patient/carer preference and nine units indicated that the decision varied by nephrologist preference.

When asked about the role of GPs in the management of CKD5 patients receiving CKM, a majority (59 of 65, 91%) of units reported that GPs liaised with the renal unit for specialist support. Forty-four of 65 units (68%) also reported that GPs and primary care staff were involved by providing/organising palliative care support at the end of life. Half of the units (34 of 65, 52%) reported that GPs also checked patient medication and arranged blood tests for patients but liaised with the renal unit for their interpretation. A minority of units indicated that GPs assessed patients regularly/routinely (not on demand) in the GP surgery (9 of 65, 14%) or via home visits (13 of 65, 20%).

A majority (57 of 65, 88%) of units provided GPs and/or their practice team with information or advice regarding the treatment of CKD5 patients receiving conservative care. Information given to GPs and/or their practice team was mainly written advice/guidelines (55 of 57, 97%) and verbal advice (46 of 57, 81%). One-third (18 of 57, 32%) of units held educational meetings to provide information to GPs. Among the units that did not provide any information/advice to GPs, four of eight units said it was for lack of time and/or funding. One unit reported lack of time and funding as well as the opinions of consultants as a precluding factor.

End-of-life care

Just over half (36 of 65, 55%) of units had a written guideline for renal end-of-life care, and 11 units (18%) were preparing such a document. Thirty-six of 65 units (55%) used a register to identify CKM patients approaching the end of life. The following factors were indicated as influencing very strongly or strongly their decision whether or not to add a patient to the register: frailty (31 of 35, 89%), frequent hospitalisation (30 of 35, 86%), symptoms (28 of 35, 80%), comorbidities (27 of 35, 77%), quality of life (27 of 35, 77%), and the ‘surprise’ question of survival of less than 12 months (26 of 35, 74%). Of those units without a register (29 units), three indicated that they were in the process of developing a register; 11 units used their medical notes or database instead; and other units identified CKM patients approaching end of life based on individual assessment, clinical judgement and discussion with other health professionals, GPs or family.

Fifty-one (79%) of 65 units used ACP in end-of-life care; those who were involved in ACP were mainly consultant nephrologists and nurses (13 units). Eight units indicated that they also had palliative care specialists involved in ACP, and six units had social workers and counsellors as well.

In terms of staff training in palliative/end-of-life care for renal patients, a majority of responding units (57 of 64, 89%) indicated that their staff had had such training; however, 39 of 64 (61%) said only a small number of the staff had received training. Only two units (3%) reported that all staff had had such training, and seven (11%) reported the majority of the staff had.

All 65 responding units appeared to liaise with some services to provide care for patients receiving CKM and approaching the end of life. Specialist palliative care services within the hospital were used most commonly (59 of 65, 91%), followed by a primary care team (58 of 65, 89%). Almost all units (62 of 65, 95%) used more than one service; for instance, 38 of 65 units (58%) liaised with specialist palliative care services within the hospital, those from a local hospice, those in the community and the primary care team. Patients received these services mainly at home (59 of 65, 91%), at a hospice (55 of 65, 85%) or within the hospital as an inpatient (53 of 65, 82%). About half of the units provided specialist palliative care services within the hospital to outpatients (38 of 65, 59%) and at the GP practice (33 of 65, 51%).

The major services that specialist palliative care services provided for renal patients on CKM were symptom management at the end of life (61 of 62, 94%), followed by patient support at home out of hours (54 of 62, 83%) and admission to the hospice, when required (52 of 62, 80%).

More than half of the units (42 of 65, 65%) provided palliative care specialists with training or advice regarding the management of renal patients. Verbal advice was most commonly given (35 of 42, 83%), followed by written advice or guidelines (24 of 42, 57%) and educational meetings (22 of 42, 52%).

Evaluation and future provision of conservative care

Fewer than half (25 of 65, 39%) of responding units reported that the quality of CKM provided in their unit was regularly evaluated. Place of death was most commonly used as information when evaluating CKM (22 of 25 units), followed by symptoms (19 of 25), survival (16 of 25) and quality of life (16 of 25).

When asked about factors that could help improve the provision of CKM in their unit, 52 of 64 units (81%) strongly or very strongly agreed that both providing renal staff members with more CKM education/training and increasing communication/involvement with GPs could help this. These were followed by providing GPs with more education/training regarding CKM (51 of 64, 80%); increasing communication/involvement with community teams (50 of 64, 78%); increasing communication/involvement with palliative care teams (49 of 64, 77%); improving computer systems by integrating primary care data with renal data (48 of 64, 75%); and more funding to develop CKM within the unit (48 of 64, 75%). Forty-two of 63 (67%) units also strongly or very strongly agreed that having funding models specifically designed to reimburse the cost of delivering CKM could help improve the provision of CKM. Fifty-one units (80%) also supported better evidence of the comparative outcomes between patients who receive CKM and those who receive dialysis (Figure 4).

FIGURE 4.

Factors that could help improve the provision of conservative care in renal units. a, n = 63. b, n = 62 because of missing responses.

In terms of actual reported plans to improve CKM services, over half of the responding units (36 of 63, 57%) planned to provide renal staff members with more CKM education/training, and to provide better end-of-life care by implementing ACP (33 of 63, 52%). Half of the units (32 of 63, 51%) planned to increase communication/involvement with GPs, 25 units with palliative care teams (40%) and 24 units (38%) with community teams. Other plans were to improve computer systems by integrating primary care data with renal data (21 of 63, 33%), and to establish a system for evaluating the provision of CKM (21 of 63, 33%). Only a quarter of units (16 of 63) planned to increase the number of staff dedicated to CKM (including nine units with dedicated staff and seven without such staff) and to implement a written up CKM policy (only 6 of 25 units without CKM policy planned to have a written policy). Fewer than 20% of units were planning to have dedicated CKM clinics (12 of 63, 19%) and to obtain funding to develop CKM (10 of 63, 16%) (Figure 5).

FIGURE 5.

Planned changes in renal unit regarding the provision of conservative care (n = 63).

Future research

Participants were presented with theoretical descriptions for a randomised clinical trial (RCT) and an observational study, which could compare conservative care with dialysis (see Appendix 3).

When asked if their unit would consider entering a patient aged 75 years or over with CKD5 into an RCT comparing conservative care with dialysis, more than half of respondents (42 of 65, 65%) indicated that they would consider it for some patients. Of those, 18 units said that their unit would definitely be willing to participate in such a trial, and 20 units said they might consider doing so. Only one unit reported they would be unwilling to participate in such a trial.

When asked if they would consider entering CKD5 patients aged 75 years and over into a prospective multicentre observational study to compare conservative care and dialysis, a majority of units (60 of 65, 92%) said they would consider it for some patients; of those, 28 units indicated that they would definitely be willing to participate in such a study, and another 28 units might like to do so. No unit said it would not be willing to participate.

There was no significant relationship between the factors such as size of CKM patient population, availability of staff primarily responsible for CKM, or availability of staff whose time was specifically allocated for CKM patients and whether or not units would consider entering their patients into an RCT or an observational study (Table 12).

Main findings

This national survey had a very high response rate and, using a detailed structured questionnaire, it provided a robust assessment of the state of CKM in the UK for older patients with CKD5. Several important findings emerged.

All but one unit had a pathway alternative to dialysis. There was no agreed terminology to describe this pathway, although ‘conservative management’ was most frequently used. The lack of agreed terminology means that there is no standard definition for CKM. Although data about numbers of CKM patients were obtained in this survey, the lack of definition for CKM hampered evaluation of such data. CKM is used to characterise a future treatment option as well as non-dialysis care for ESKF (i.e. a disease state equivalent to being on dialysis); the number of patients in the latter group on CKM was relatively small (median 8, IQR 4.5–22) in the units that responded to this question.

In order to assess suitability for CKM for a patient, similar criteria were used across the units, foremost among them being patients’ preference for CKM. The CKM decision-making process was also similar across units: most reported that they undertook informed decision-making with CKD5 patients aged 75 years and over by presenting treatment options including CKM to all such patients, and decision aids were widely used when discussing the options. Not only patients but also family/carers were actively involved in decision-making about CKM. After the initial CKM decision-making, their decision was always reviewed at clinic visits. Family members were often involved when a decision was reviewed. All units had patients who had changed their mind after deciding not to have dialysis; however, from our survey we could not quantify numbers of such patients.

Conservative kidney management practice patterns varied across units; some showed considerable investment of staff time (both nursing and medical) and consequent processes with evidence of dedicated clinics, a written CKM guideline and staff training initiatives. There was also some evidence that having staff responsible for CKM was associated with some of these practice patterns, such as having a dedicated CKM clinics and a CKM guideline. The number of patients was related to the number of patients aged 75 years and over on RRT. The components of CKM varied, for example provision of EPO and iron therapy, and symptom assessment and management were uniform, whereas formal psychological support and use of home visits was more varied. Only a minority of units had funding dedicated to providing CKM. Whether they had funding or not was associated with numbers of CKM patients; however, our survey could not identify whether they had more CKM patients as a result of receiving funding or units with the larger number of CKM patients had more opportunities to apply for and/or to receive such funding. Only a quarter of units had clinics exclusively for CKM patients, and this was closely related to availability of staff responsible for CKM. Again, we could not identify whether they had CKM clinics because they had responsible staff or vice versa.

All responding units worked collaboratively with primary care and palliative care teams, indicating their vital role in the care for patients receiving CKM especially at the end of life. For instance, many units provided GPs and their practice team with information or advice regarding the treatment of CKD5 patients receiving CKM, and all responding units appeared to liaise with some palliative care services in order to provide care for patients receiving CKM and approaching the end of life. Some units involved patients and families in the decision regarding sharing care for CKM with the primary care team.

In relation to palliative care, almost all units appeared to have staff who had training in palliative/end-of-life care for renal patients. There was, however, variation in the proportion of such staff: from two units with all staff trained to 39 with very small proportions of such staff. This may be related to a lack of resources, but may also be because higher priority is given to involving palliative care specialists in providing care to ESKF patients rather than developing the skills of renal staff.

To improve the provision of CKM in future, increasing communication/involvement with GPs, community teams and palliative care teams were thought to be very important by many units. For better collaborative working, information sharing was also thought to be vital, which could be facilitated by the integration of primary care and renal computer-based data. Many units reported that providing more education and training regarding CKM was important not only for renal staff but also for primary care and palliative care teams. However, it was also clear that lack of funding and time was a big issue for many units. Although many units thought that more funding could help develop CKM in their units, only a minority were actually planning to apply for funding. This may indicate that either it is difficult to obtain such funding, or that funding to cover CKM may not be widely available. Moreover, currently there is no financial payment for CKM under the Payment by Results tariff scheme. There is a need for a better understanding of the resources and costs of CKM to improve the commissioning of CKM. In terms of planned changes regarding the provision of CKM, increasing the number of staff dedicated to CKM and setting up dedicated CKM clinics were not prioritised as much as other items. This may be also because of a lack of funding; however, many units delivered CKM without receiving funding. It would be useful to understand how units without funding currently manage this provision.

Comparison with existing research

This is the second national survey to explore the practice patterns of renal units regarding care for older adults who choose not to have dialysis. Our survey showed similar variation in practice patterns to those found by the previous national surveys of the provision of CKD and of palliative care provision to renal patients in UK renal units. 31,57

The previous study regarding CKD practice patterns in the UK demonstrated that the most common perceived staff shortages in the care were of counsellors and psychologists: only 33% and 25% of units respectively had such staff in their MSRT. 57 Figures are similar in the current study. This suggests that formal emotional and psychological support is still considered to be a low priority in many units.

The proportion of units with a pre-dialysis clinic has increased from 50 of 70 (71%) to 56 of 67 (84%). 57 Similarly, regarding geographical accessibility to pre-ESKF care, the proportion of units that run clinics for CKD patients in neighbouring hospitals has increased from 49 of 70 (70%) to 59 of 67 (88%). The previous research showed that, even in units where outreach clinics were already set up, some patients were not receiving the same MSRT input as those being followed up in the main unit57 because these MSRT personnel were not funded to work outside the remit of the employing hospital. Our survey did not investigate whether CKD patients in the neighbouring hospitals could access to MSRT services or not; therefore, the quality of care in neighbouring hospitals remains unclear.

Many units that had a pre-dialysis education day covered CKM and psychological support as topics. From this survey, however, it is not possible to find to what extent they were discussed during such programmes. Compared with the previous study,57 psychological support was more commonly reported.

Many renal units first raised the option of CKM with patients either when they were referred to the pre-dialysis clinics or when their eGFR level was about 20 ml/minute/1.73 m². Our survey was not set to explore if the timing of referral to pre-dialysis was governed by eGFR or timing before the expected start of dialysis. According to an Australian study conducted by Morton et al. ,59 84% of patients received information about treatment options when their eGFR was < 15 ml/minute/1.73 m². They suggested that education on treatment was probably best commenced in stage 4 (eGFR < 30 ml/minute/1.73 m²) rather than CKD5. Although our survey did not collect actual patients’ eGFR values at the time of the first discussion on treatment, our results may indicate that many UK renal units provide patients with such information earlier, as suggested by Morton et al. 59 Only a minority of units used timing rather than eGFR level in order to decide when to raise the CKM option with a patient. Many CKD patients have a non-linear eGFR trajectory or a prolonged period of non-progression in contrast to the traditional notion of steady eGFR progression over time. 60 This suggests that careful assessment of kidney disease progression is needed to decide when to discuss CKM with patients rather than solely using an eGFR level.

In the past, UK renal units did not have ready access to palliative care services, although most were willing to provide ongoing care in terms of outpatient follow-up. 31 Our survey, however, showed that all units liaised with specialist palliative care service for CKM patients approaching the end of life. Many liaised with such services not only within the hospital but also from a local hospice. Access to palliative care used to be restricted to patients with malignant disease, and specialist palliative care services were involved only in the minority of ESKF patients’ care. 31 However, data from the more recent study undertaken by Hobson et al. 61 demonstrated that specialist palliative care services accepted more referrals of ESKF patients than before. In our survey, precise figures related to referrals were not collected; however, this survey may support Hobson’s data, suggesting that palliative care services are now more widely available to patients with ESKF. Similarly, the current study showed that a larger number of renal units had more standardised care for renal end-of-life care than before: currently over a half of responding units had a written renal end-of-life care protocol compared with only 20% at the time of the past study. 31

Strengths and limitations

Although we achieved an excellent response rate, some limitations need to be addressed. First, the survey responses were self-reported; therefore, we could not check their validity and they were predominantly completed by clinical directors. Some of these data were necessarily estimates, such as the number of CKM patients aged 75 years and over, and the number who were symptomatic. Such estimates should be regarded with caution. Our division of units into larger and smaller categories was arbitrary and for exploratory purposes. This survey focused on CKM patients aged 75 years and over. There are patients on CKM aged under 75 years;62 therefore, the results do not fully address all CKM patients. While we examined if some selected factors were associated with the use of certain practices regarding CKM, we were unable to determine causal relationships. Furthermore, most questions in the survey were multiple-choice questions where respondents were asked to select the best possible answer (or answers) out of the choices from a list. This may have limited their responses, although a selection of ‘other’ was provided in case the respondent could not find any appropriate items in the list. Finally, these findings represent a current snapshot of CKM practices in the UK, which is one of the fastest-changing areas among the renal community.

Implications for research and practice

There is a need for a standard terminology and more precise definition for CKM, since a non-uniform terminology and definition may lead to both conceptual errors and misinterpretation of data. Our study highlighted a vital role of primary care and palliative care teams, which emphasises the importance of communication between renal units and those services. Improving computer systems by integrating renal data with primary care and out-of-hours databases is essential in order to ensure good continuity of care across different care providers. Lack of resources was seen as a big problem. More opportunities for funding dedicated to providing CKM will be needed, and the current Payment by Results tariff scheme and renal commissioning may need to be revised in order to support CKM. Resources available for delivering CKM varied between the units; however, the value of resources such as dedicated staff and dedicated CKM clinics in terms of quality of service is still not clear. There is a need to quantify the value of such services in order to facilitate development of CKM services.

Having identified the individual components comprising CKM in renal units around the UK, there is an urgent need to establish the most cost-effective combination of these components in a CKM service and indeed the optimal way to organise a CKM service across primary and secondary care.

At this stage, an observational approach is likely to be the most appropriate next step to achieve this. With practice patterns now defined and measurable, provision/receipt of these individual components of CKM could be captured at a unit level/individual patient level, and novel statistics such as propensity score matching, instrumental variable analysis or marginal structural modelling used to deal with the confounding due to treatment by indication bias. The UK Renal Registry, which has recently had its Section 251 approval extended to pre-dialysis CKD, could provide a very cost-effective infrastructure for obtaining some of these data.

The present study has also highlighted the importance of measuring patient quality of life and satisfaction with treatment when comparing the outcomes achieved with CKM and dialysis, and addressing carer and family burden. Given the age, comorbidity and cognitive function of the patients for whom there is likely to be equipoise for this decision, such matters will be particularly important to consider when planning and designing future observational or interventional studies.

While an observational study may provide high-quality comparative effectiveness evidence for CKM versus dialysis, the gold standard is likely to remain a RCT. Although 65% of centres indicated willingness to participate in such a study, some serious doubts persist among clinicians about whether such a study is really possible or not. A previous attempt to undertake a RCT comparing the outcomes of HD and peritoneal dialysis was unsuccessful because of the limited number of patients who were uncertain over their therapy choice and willing to participate. 63 A feasibility study exploring the willingness of clinicians and patients to be randomised to CKM versus dialysis, with embedded qualitative research to identify recruitment, would likely be a necessary first step. 64 Data from the observational work described above would also be very useful for refining both the hypothesis and the intervention for such a study.

Conclusion

Conservative kidney management is widely practised in UK renal units. Provision and organisation of CKM varies in many ways, such as availability of CKM funding, dedicated clinics, staff whose time is especially allocated to CKM patients, written CKM policy and availability of staff training. Our survey also highlighted the vital role of primary and palliative care teams for the provision of CKM. Palliative care services for renal patients approaching the end of life are more widely available than previously reported. In order to improve the provision of CKM, better understanding is needed of the comparative outcomes and costs between CKM and RRT.

Design and setting

This was a mixed-methods study involving quantitative data collection from UK renal units and semistructured interviews with GPs. The same nine English renal units were included as in the staff and patient studies. GPs were identified from general practices which were in the catchment area of four of the nine renal units (King’s College, London; Southmead, Bristol; Heartlands, Birmingham; and Lister, Stevenage).

Data linkage

Nine renal units were contacted to identify an information technology (IT) professional who could obtain patient data. Two data sets were required, renal unit data and clinical biochemistry data from their local laboratory. Renal unit data included any patients who were known to the renal unit at the time of the laboratory data extract. Laboratory data included patients, aged 75 years or older, with two eGFR results < 15 ml/minute/1.73 m² for the first time on record between January 2010 and June 2011. These individuals were therefore new cases of CKD5. Specific data were required in each data set as detailed in Appendix 7. IT professionals sent secure versions of the data to the UK Renal Registry. Approval was obtained from the National Information Governance Board. Laboratory data provided the number of patients with new CKD5 in the 18-month period (as determined by eGFR blood results). Renal unit data provided information on all patients known to a renal unit; in some units, such patients included those who had received phone advice only. Laboratory data were matched with data from renal units to identify patients with new CKD5 who were and were not known to a renal unit.

Participants and interviews

The original approach for identifying GPs to take part in the qualitative section of the study involved inviting the GPs of patients identified through the quantitative data collection. Patients of interest were those who were identified as having new CKD5, but who were not known to local renal units. As only a small number of patients were identified from the quantitative data, an alternative approach was adopted in order to recruit GPs. This approach aimed to invite any GP working in the catchment area of four of the nine renal units of interest. Additional ethical approval was granted for this.

Primary care trusts (PCTs) in the catchment area of the four renal units of interest were identified. General practices within each PCT were identified through the QOF database. One GP in each practice was invited to take part in the study by post. If a GP did not respond or declined to take part, a second GP from the same practice was invited. GPs returned a reply slip and a signed consent form by post to indicate they were happy to take part in the study. GPs were then contacted by telephone or e-mail to arrange a time for interview. Interviews were carried out by telephone by ST-C, an experienced qualitative researcher. Interviews followed a semistructured interview guide which asked about GPs’ previous experience of managing patients with CKD, referring CKD patients to secondary care and providing palliative care to end-of-life CKD patients (see Appendix 8). GPs were reimbursed for their time to take part in interviews at £40 per interview. Interviews were audio-recorded and transcribed verbatim by an independent transcriptionist. Interview transcripts were checked by ST-C.

Qualitative data analysis

Data analysis was carried out using an inductive thematic approach in order to allow results to be driven by the data. 41 This approach minimised the influence of the researchers’ preconceptions or existing knowledge on analysis. Transcripts were analysed once they had been checked and reread by ST-C. Line-by-line coding was used to produce basic codes, and new codes were added as more transcripts were analysed. NVivo 9 was used to organise data and facilitate coding. ST-C independently coded 10 interviews and developed an initial set of themes. Themes were developed by comparing codes and grouping them into similar categories to produce themes and subthemes. Themes and supporting quotes were discussed with MS and revised to produce a consensus framework. This framework was used to code the remaining transcripts. Any new codes were incorporated into the existing framework, which was amended if necessary. Final themes were developed by ST-C and MS and discussed with the wider research team until a consensus was achieved.

Data linkage

Data were linked for only three of the nine pairs of renal units and laboratories (Table 13). In all three, renal units were aware of the majority of CKD5 patients; however, there were patients in all areas who were not known to the local renal unit.

In the remaining units, four contacts reported that it was not possible to obtain the data required. This was most often because of the computer systems used by the various units and laboratories and the ability to search for the required data. For two units, it was not possible to identify an appropriate IT contact whom we could liaise with, and in one, the CKD5 patients were not broken down by age.

Participant characteristics

A total of 353 GPs were invited to the study across the four areas. Twenty-five GPs responded to say that they were interested in taking part in the study. Other than two responses from GPs who replied to say they did not want to take part, no response was received from any of the other GPs. The overall response rate was 7%, although this varied between areas, with a response rate of 4% in London and 13% in Bristol. It was not possible to obtain characteristics of non-respondents.

Nineteen GPs were interviewed, with three to six GPs being interviewed in each of the four regions. Six GPs who had initially indicated they would like to take part were unavailable to be interviewed. GPs were mostly male (63%), had a mean age of 46 years, a mean of 16 years in practice and worked in practices with five doctors on average. Three GPs worked in inner-city practices, five in rural areas and the remaining 11 in suburban areas. Nine of the GPs said they had no special interest in kidney disease or other associated conditions; the remaining 10 GPs indicated that they were interested in diabetes, palliative care and/or care of the elderly, so there was some selection effect.

Qualitative findings

Five themes were identified from the analysis of transcripts. All themes were relevant to all interviews but GPs occasionally reported different views and behaviours within themes depending on the context in which they worked and their experience.

Main findings

Quantitative results indicated that the majority of patients with CKD5 were known to local renal units, indicating that GPs were referring most patients, although it was difficult to obtain and evaluate such data, and our data set was limited to three renal units. The qualitative analysis identified several factors which influenced GPs’ management and referral of patients with CKD4 and 5 to help explain these numbers.

Previous experience of treating patients with CKD was a good indicator of how familiar GPs were with guideline recommendations and when to refer. Some GPs had relatively little experience of managing patients with CKD4 or 5 and of CKM per se and found it harder to comment about their experiences with these patients. GPs with older patient populations felt more comfortable managing patients with CKD, who were often older adults with comorbidities. While some GPs felt that patients with CKD3 might not know of their diagnosis, all felt that patients diagnosed with stages 4 and 5 would be aware of their CKD and that patients were generally happy to be referred. Most GPs reported that they had good communication with their local renal units and particularly sought advice on when to refer patients and how to manage patients if they were not referred, or after they had been sent back from nephrology. There was variation in when individual GPs said they would refer patients with CKD4. They were influenced by the eGFR, decline in eGFR, the general well-being of the patient, the patient’s age and comorbidities, and occasionally the distance to the renal unit. In general, GPs emphasised that it was preferable for patients to stay under primary care where possible and to refer patients based on their individual characteristics. Some GPs identified older adults with comorbidities as complicated cases where referral decisions were more difficult. Finally, GPs identified patients receiving palliative care or with advanced dementia as less likely to be referred to the renal unit and were happy to care for such patients. Most interviewees had little experience of managing palliative care for patients with CKM those who had were comfortable with this as long as the patient had made an informed decision in secondary care and that this had been communicated to primary care.

Comparison with existing research

Qualitative research in this area is limited, with only a few studies having been carried out in the UK and the USA;66,71,72,74,75 however, there were similarities between the results of the current study and such previous research. GPs reported that they felt some anxiety about telling patients with early-stage CKD of their diagnosis71,72 and that GPs felt patients and the public had little understanding of kidney disease. 66,72 In addition, most GPs accepted that they did not know a great deal about the management of CKD4 and 5 because they had relatively few patients with these and they welcomed any guidance or additional advice that they could get to inform their practice. 75

Other studies have found that GPs can be sceptical about using the measure of eGFR to make a diagnosis of CKD. 66,76 It was interesting to note that GPs did not seem to have this concern in the current study. Although a few GPs described using eGFR as a factor when deciding whether or not to refer, it was most often the decline in eGFR rather than the single reading that was referred to. This may have meant that GPs were more confident using a decline in eGFR as evidence of impaired renal function or that GPs had had longer to get used to referring to eGFR results since the introduction of NICE guidelines.

Only one previous study mentioned GPs’ approaches to referring patients. 75 Its results were similar to the current study, showing that GPs varied in when they made a referral and on which factors a referral was based. The study also indicated that GPs working in more rural locations tried to manage patients themselves as much as possible to avoid unnecessary travel for patients. 75

Patients who have opted for CKM are not currently identified and recorded by the UK Renal Registry or any other national registry. While our data suggest that most are known to nephrology services, it is likely that some are kept in primary care and are never referred. A report by Kidney Health Australia identified the number and proportion of patients with ESKF who did not start RRT by age, based on death certificates and registry data. 77 The results identified a substantial proportion of patients who did not receive RRT, with a higher proportion being seen in older age groups. 77 More data that links prevalent CKD cases to dialysis is needed in the UK, and linking primary care databases to the UK Renal Registry may help with the identification of patients with CKD and an assessment of the proportion of patients treated without dialysis.

Strengths and limitations

This was the first study to identify the proportion of CKD patients known to renal services through the collection of laboratory data. It was difficult to obtain the data needed to identify the proportion of patients with CKD5 not known to renal services, and only three of nine units could provide the full information. Moreover, we accessed the clinical biochemistry laboratory of only the hospital in which the renal unit was based; this might have led to an overestimation of the referral proportion referred. This also meant that GPs could not be sampled based on individual patient cases and that some GPs had to talk hypothetically in some instances because they had not treated a patient with CKD5 recently. Despite this, GPs were able to answer all questions posed to them and there was a similarity between those who had specific patients in mind when answering questions and those speaking hypothetically, indicating that the approach taken by GPs was consistent between practices and areas.

General practitioner participation was voluntary; this probably led to sampling GPs who had a particular interest in CKD or a related specialty (half the volunteers). Data did indicate that there were differences in GPs’ knowledge of guidelines and confidence in managing patients between groups who had a special interest and those who did not. This may mean findings have less relevance to all the GPs in the UK.

There was a particularly low response rate from GPs invited to take part in the qualitative study. Study materials emphasised that GPs did not have to have specialist knowledge in the management of CKD. However, GPs may have been reluctant to participate because it was something they did not commonly see in practice. Previous research has identified low response rates in general practice research and the difficulties of recruiting GPs to research studies involving surveys. 78 Postal invitations, Freepost^TM envelopes for replies and reimbursement for time spent taking part in the study have been shown to help increase participation by GPs and were all used in the current study design. 79

Unlike previous studies, GPs were recruited from practices in specific areas of the UK, rather than from practices which were taking part in a larger research study. 66,71 Invitations sent to individual GPs, rather than practice managers, avoided recruiting only from those practices that were research-focused and probably added to the relevance of findings to GPs in general.

Additional insight into primary care practice might have been obtained by including practice nurses or district nurses in the sample. Practice nurses might have provided insights into the management of CKD patients and discussions with CKD patients when diagnosed, and district nurses might have been able to offer insight into palliative care for patients with CKM, although these are likely to form a small minority of their workload. It was not possible to obtain data on nurses working in all practices and this would have limited recruitment. In addition, nurses were less likely to have views on the decision to refer patients to secondary care, as this was something most commonly discussed between GP and patient.

Implications for practice

General practitioners distinguished between younger and older adults with CKD, of whom many had comorbidities. Some mentioned that there was a conflict between treatment for CKD and treatment for another chronic condition and others felt unsure about referring older adults because of the many other conditions they had. It appeared to be difficult for GPs to judge which condition was most detrimental to a patient and, given contrasting advice from specialists, there may have been a preference to keep the patient under primary care. While this approach may avoid anxiety and multiple hospital appointments for patients, there may be missed opportunities to educate patients about their CKD and prevent further decline. Guidelines which focus on single conditions are particularly problematic in CKD, where so many patients have comorbidities. Further advice on managing common comorbidities would be helpful for GPs. 80

General practitioners appeared to be comfortable managing patients with CKD and were confident that they could obtain advice from renal specialists when necessary, for example about medication. Some GPs suggested that current communication with secondary care could be improved by the provision of a dedicated telephone helpline to answer renal queries. GPs reported that this already existed for other conditions and that it would make it easier for them to obtain fast responses to their queries. GPs felt patients needed to be seen by secondary care when they needed to talk about future options for treatment.

Some GPs reported that they had already received specific input from their local renal units through local guidelines or educational meetings. These had been positively received and appeared to have offered clarity on issues about referral. There is potential for other renal teams to emulate such approaches and to initiate better communication strategies with primary care professionals, which will probably improve care for renal patients in local areas.

General practitioners used decline in eGFR as a key factor in the decision if and when to refer to a nephrologist. Highlighting trends in eGFR in pathology reports by clinical chemistry laboratories would aid this decision process. 81

Conclusion

General practitioners make decisions about referrals to nephrology services on an individual patient basis. A substantial majority of CKD5 patients are known to renal units. GPs have access to guidelines on management of CKD and, although they sometimes found it difficult to refer all patients suggested in national guidelines, they generally felt supported by additional information from local nephrology services. GPs may benefit from advice delivered from their local renal unit and clinical chemistry laboratory to clarify referral criteria and additional guidance on the management of older adults with multiple comorbidities. Most had little direct experience of managing palliative care for patients with CKM.

Clinical/service implications

• Conservative kidney management is widely recognised and delivered across the UK, but through differing models of care and sizes of CKM programmes.

• An agreed terminology and definition of CKM are needed to enable future evaluation. The designation of a patient as having CKM must recognise the two key points on the CKM pathway: (1) stating a preference or intention to opt for conservative care rather than have dialysis in the future, made at an unspecified level of kidney function and an unspecified time before dialysis is indicated – ‘the early conservative care pathway’ – and (2) a decision to opt for CKM made at a level of kidney function or despite symptoms that would otherwise justify starting dialysis – the ‘alternative to dialysis pathway’. The systematic routine recording of the CKM pathway within renal data systems would provide a foundation for evaluation of CKM patterns and outcomes.

• Communication and information given to patients with ESKF who are older and or multimorbid needs to (1) support patients’ understanding and perceptions of their kidney disease and their other comorbidities; (2) routinely include details of the CKM pathway, settings of care and codependencies with other pathways of care that individuals are receiving; (3) include realistic discussions (according to preferences) of what is likely to happen in the future; (4) recognise that shared decision-making in this setting is a process rather than an event and that decisions made need to be periodically reviewed; (5) recognise the importance of identifying and where possible ameliorating cognitive impairment and ensuring that plans are in the best interests of individuals.

• In order to address the challenging nature of decision-making, and communication about end-of-life issues, renal staff need education and training in (1) advanced communication skills, adapted to include the specific issues around dialysis decision-making, and (2) how to discuss and address palliative and supportive care needs, including end-of-life care.

• Better communication and information sharing is needed with primary care teams and training in the renal-specific elements of CKM care.

• End-stage kidney failure is a unique kind of organ system failure in that a replacement therapy is potentially available for the majority. As a consequence, planning for care occurs earlier than in other conditions, such as heart failure. However, a better understanding of patient and their carers’ wishes, needs and values in advancing CKD and of the optimum service for CKM is likely to have wide interest across the health-care system. This is, first, because multimorbidity, frailty and functional decline in those with long-term conditions is a major system challenge, and, second, because the appropriate balance between advanced technology and palliative care is also relevant for other conditions.

Research implications

Research is required to measure the benefits and costs of CKM and dialysis and to inform decision-making by staff, patients and their families. There is some support in principle for a randomised trial, which would be the most robust method in terms of internal validity given the problem of selection bias in observational designs. A feasibility study exploring the willingness of clinicians and patients to be randomised to CKM versus dialysis is likely to be needed before embarking on a definitive interventional study. Data from such observational research would be very useful for refining both the hypothesis and the intervention for such an interventional study.

• An alternative would be a prospective observational study, and there was strong support for that approach. The fact that there is variation in the scale of conservative care between units suggests that similar patients are being treated by these two approaches across the UK, so it should be possible to undertake such a study.

• Nested within the CKM versus dialysis question is the need to establish the most cost-effective combination of various components in a CKM service and the optimal way to organise a CKM service across primary and secondary care. An observational approach could capture the provision/receipt of individual components of CKM at a unit level/individual patient level.

• In observational studies, propensity score matching, instrumental variable analysis or marginal structural modelling could be used to deal with the confounding caused by treatment indication bias.

• Given the age, comorbidity and cognitive function of patients for whom there is likely to be equipoise with regard to CKM versus dialysis, patient and carer quality of life and satisfaction with treatment will need to be carefully measured in studies comparing outcomes.

• Given the underascertainment of non-English speakers, a focused qualitative study is needed looking specifically at this population.

For renal clinicians and senior nurses to answer: willingness to randomise

There is a lack of high quality evidence for patients and clinicians to consider when deciding whether to have dialysis or CKM.

1. Would you consider it appropriate to enter any patients aged > 75 years with progressive CKD5 into a RCT of CKM versus dialysis?

   1. Prompts:

      1. More detail on types of patient that might be considered?

      2. Reasons against/for a trial?

      3. Would your unit be willing to participate in such a trial?

If no, willingness to enter into a prospective observational non-randomised study comparing outcomes of patients with CKM with similar patients starting RRT.