Establishing a peritoneal dialysis technique survival core outcome measure: A standardised outcomes in nephrology-peritoneal dialysis consensus workshop report

Abstract

Background

Technique survival, also reported with negative connotations as technique failure or transfer from peritoneal dialysis to haemodialysis, has been identified by patients, caregivers and health professionals as a critically important outcome to be reported in all trials. However, there is wide variation in how peritoneal dialysis technique survival is defined, measured and reported, leading to difficulty in comparing or consolidating results.

Methods

We conducted an online international consensus workshop to establish a core outcome measure of technique survival. Discussions were analysed thematically.

Results

Fifty-five participants including 14 patients and caregivers from 13 countries took part in facilitated breakout discussions using video-conferencing. The following themes were identified: capturing important aspects of the outcome (requiring a core event to define the outcome, distinguishing temporary from permanent events, recognising heterogeneous experiences of transfers), adopting appropriate neutral nomenclature (conveying with clarity, avoiding negative connotations), and ensuring feasibility and applicability (capturing data relevant to clinical and research settings, ease of adoption). The suggested definitions for the core outcome measure were ‘the event of a transfer to haemodialysis’, or ‘discontinuation of peritoneal dialysis’. Applying the principles described within the workshop, defining the outcome measure as a ‘transfer to haemodialysis’ was preferable.

Conclusions

It is proposed that the core outcome of technique survival is redefined as ‘transfer to haemodialysis’ and that its components are standardised using simple, neutral terminology Components considered important by stakeholders included recording the reasons for transfer from peritoneal dialysis, and focussing on permanent events whilst ensuring the outcome remains easy to implement.

Graphical abstract

This is a visual representation of the abstract.

Background

The terms technique failure or technique survival are often used to describe patients who, for a variety of reasons, have to discontinue peritoneal dialysis (PD) and most frequently transfer to haemodialysis (HD), but there is a lack of consistency in current definitions, which may or may not include temporary discontinuations, death or stopping PD to palliate. Nevertheless this is a large problem for patients and carers, illustrated both by the high rate of this outcome,^1–6 and by the widespread reporting of rates of transfer to HD by national registries. Transferring a patient from their preferred modality has adverse impacts upon healthcare expenditure, ⁷ mortality, ⁸ morbidity and quality of life, with patients frequently reporting negative experiences in the process. ⁹ Furthermore, there is significant variability between centres both nationally and internationally in rates of transfer to HD.^4,10 This is an example of unwarranted variation, strongly indicating although not definitively proving that differences in centre-level practices are directly leading to worse outcomes for patients and their caregivers, with scope for improvement.

Despite the importance of technique survival, there has been no attempt to standardise the definition of this outcome internationally. This has resulted in registries reporting the outcome according to different or unclear definitions, preventing meaningful international comparisons, although the Peritoneal Dialysis Outcomes and Practice Patterns Study has now demonstrated between-country variability in rates of transfer to HD with reasonable standardization of criteria. ² Lack of standardisation is also an issue in trials, with a systematic review of all PD trials identifying a highly heterogeneous approach in measuring and reporting this outcome, including the use of different terms (technique failure, technique survival and transfer to HD), variable inclusion/exclusion of death, and variation in the minimum duration required before transfer to HD is considered permanent. ¹¹

The Standardised Outcome in Nephrology – Peritoneal Dialysis (SONG-PD) group, representing over 900 patients, family members and health professionals from 68 countries, identified technique survival to be a critically important outcome that must be reported in all future trials. ¹² To reduce the variability in the recording and reporting of this core outcome, trials need a standardised definition and a stakeholder workshop to establish the essential features was convened. This workshop report summarises the perspectives of the patients, caregivers, and healthcare professionals on defining a core outcome measure for technique survival to be reported in all future trials involving PD patients.

SONG PD technique survival workshop

Context and scope

The SONG PD technique survival consensus workshop was held online on 15 September 2020. Patients, caregivers, healthcare professionals, researchers, policy makers and epidemiologists were invited to participate. This was held over Zoom, a modern video communication platform, to allow for global participation from the PD community.

Attendees and contributors

55 participants from 13 countries including the UK (9), Australia (17), France (3), Lebanon (1), Canada (5), South Korea (1), USA (6), Brazil (2), Singapore (3), Poland (1), Hong Kong (3), South Africa (1) and Belgium (1) attended the workshop. Thirteen participants were patients, and one was a caregiver, all with experience of PD. Other participants included paediatric and adult nephrologists, researchers and renal nurses. Healthcare professionals were invited if they had an interest in PD and/or PD expertise or leadership roles in relevant professional societies (including the International Society for Peritoneal Dialysis [ISPD]), regulatory agencies, funding organisations, patient charities (Kidney Care UK) and registries (Australia and New Zealand Dialysis and Transplant Registry [ANZDATA], UK Renal Registry). Patients and caregivers were invited by workshop conveners.

Workshop overview

Prior to the workshop, background materials were sent out to all participants. First, we provided a brief overview of the SONG-PD initiative with preliminary results of a systematic review on the outcome measures for technique survival. Following this, participants were allocated into break-out rooms for 35 min to discuss the core outcome measure. Each group consisted of eight to 10 participants, which included professionals and patients/caregivers. A facilitator in each group followed a standard run sheet including questions to bring to the group.

Each group was asked to discuss the pro and cons of the terms ‘technique survival/failure’ and ‘transfer to haemodialysis’. They were then asked to focus on what is the best/most meaningful definition for the core outcome measure, what aspects are important to consider for the definition, and what duration of haemodialysis would be regarded as a ‘transfer or technique failure’. Following the break-out sessions, a representative from each group was allocated to feedback to the large consensus workshop group. This involved a brief summary of the discussion from every group. The session was concluded with a final wrap-up from the chair (ML).

All breakout discussions were recorded and transcribed verbatim. Transcripts were entered into HyperRESEARCH (ResearchWare Inc, Version 3.0) to facilitate coding and analysis of the data. Investigators EE and ML read the de-identified transcripts line-by-line and inductively identified and coded concepts into themes reflecting participants’ perspectives and suggestions regarding the development and implementation of a core outcome measure for technique survival. Preliminary themes were discussed and revised with YC, AJ and KM who independently read the transcripts. The agreed findings were then sent to all workshop attendees for review and feedback to ensure that it reflected the full range of perspectives and additional comments integrated into the final report.

Summary of the workshop discussion

We identified three main themes that encompassed the discussions across the breakout groups: (1) capturing important aspects of the outcome; (2) adopting appropriate, neutral nomenclature; and, 3) ensuring feasibility and applicability.

Capturing important aspects of the outcome

Recognising heterogeneous experiences of technique survival. Patients could choose to transfer to HD, for example, patients who felt ‘burnout’ with home dialysis and wanted relief by switching to in-centre HD. ‘… my take on in-center hemo was, “Oh good, three days a week I get a break. I get to watch cable that I don’t have. I can have a nap if I need it” (Patient)’. However, some patients had no choice but to transfer to HD, for example, because of infection. This variability in patient experience was recognised by healthcare professionals. ‘I guess the question is whether hemodialysis transfer will be viewed as a positive outcome or a negative outcome. Because it is a mixture of those things (Healthcare professional)’. Thus, participants recommended to record the reason for the event defining technique survival, to help with interpretation of the measure.

Requiring a core event to define the outcome. Participants agreed that technique survival should be defined clearly by an event. Some suggested that technique survival should be defined as stopping PD to be able to identify all problems leading to discontinuation of PD, irrespective of whether patients then transferred to HD. It was uncertain whether kidney recovery, withdrawal of dialysis, temporary stoppages, death either due to complications of PD or unrelated deaths, or kidney transplantation should be included and how it should be reported. A ‘global’ outcome measure of stopping PD, with specific events recorded was suggested.

Some asserted that transferring from PD to HD was the key event that should define the core outcome measure. Other events (e.g. kidney recovery, death) were assumed to be recorded as separate outcomes. Transfer to HD was regarded as a clear and simple definition that would remove causes that might cause confusion, such as stopping PD as part of end-of-life care, that was easily interpretable by healthcare professionals and patients and could be easily captured in trials and registries.

Distinguishing temporary from permanent events. The reasons for transferring to HD or stopping PD were considered by participants in relation to the permanency of the event. An intended permanent change (e.g. patient choice relating to burnout) was viewed as having different implications compared with an unintentional temporary change (e.g. acute severe peritonitis), but it was noted that the intention could be difficult to collect accurately. The duration of the transfer to HD (i.e. time off PD) was regarded to be the optimum method to determine whether an event was ‘permanent’. However, it was recognised that the duration of qualifying permanence would vary between patients, centres and countries requiring an individualised approach.

Healthcare professionals and patients agreed that different time frames would provide data on different patient populations. Shorter timeframes, such as 30 days, would likely include more patients with temporary problems, for example post peritonitis. Longer time frames (e.g. 180 days) would generally identify problems that required permanent cessation of PD. Two patients/caregivers commented that even a very short period on HD was relevant and meaningful, and some healthcare professionals felt that identifying short periods on HD/off PD would help to identify problems as part of a governance process. However, others viewed the permanent transfers as more relevant, thereby justifying a longer period to define a transfer to HD. ‘The definition doesn’t really interpret the groups that we really want to focus on like the group where they permanently fail to go back on PD’. (Caregiver)

This included a suggestion by a healthcare professional that in some situations a short period of PD, such as for a hernia repair, should not be viewed as a problem, but as part of the therapy. There was a concern that shorter durations would reflect differences in the speed of catheter re-insertion for temporary transfers, and also inflate the reported rate, potentially discouraging prospective PD patients due to concerns about permanent transfers to HD. The main concern about longer durations on HD/off PD is related to problems with trials (see below). Specific durations discussed were 30, 90 and 180 days, and where a preference was expressed it was for 30 or 90 days

Adopting appropriate, neutral nomenclature

Conveying with clarity. With regards to potentially adopting the definition of technique survival as ‘transfer to HD’, the alternative terms to ‘transfer’ were ‘switch’ or ‘change’. For stopping PD, suggestions for nomenclature were more variable, including ‘cessation of PD’, ‘termination of PD’ and ‘discontinuation of PD’.

The use of the word ‘technique’ (for technique failure) was considered to be ambiguous as it was not clear if death, transplantation, recovery of kidney function, and dialysis withdrawal were captured in this term. Patients and health care professionals felt it was important to report the death, which could be viewed as a failure of the treatment, and the discussion included one proposal to specify death as related to complications of PD or unrelated causes, although how this could be applied in practice was not discussed.

Use of the term ‘technique survival’ was considered but this was thought to imply the lack of an event, introducing a lack of clarity given that defining an outcome measure as an event would be clearer, and easier to record and analyse. Similarly, the word ‘failure’ was potentially inaccurate, in that this did not capture situations where there was not a failure, but a planned integrated care strategy or patient choice.

Avoiding negative connotations. The use of the word ‘failure’ was felt to have negative connotations, with patients expressing the perception that this may be attributing blame to them, and that it was their fault PD had ‘failed’. As one patient highlighted, ‘I know this is for trial purposes and studies, but inevitably that creeps into clinical language as well’. Some patients felt the term ‘failure’ was ‘… putting the blame on us for it failing, whereas it could just be the body doing it … (patient)’. Similarly, when considering stopping PD, ‘termination’ was viewed as having negative connotations that should be avoided. Instead of the term ‘failure’, others wanted more clarity in the definition to describe the event of interest, ‘I … prefer the transfer to hemo[dialysis]. When someone chooses to be transferred to hemo[dialysis], that's not a failure. That's a choice (patient)’.

Ensuring feasibility and applicability

Capturing data relevant to clinical and research settings. As a core outcome for studies, participants suggested to limit the duration of follow-up for feasibility. For example, to define an event requiring 90 days of follow-up could exceed the duration and resources of some trials, ‘Let's say for a one year follow up trial, definition that includes 90 days means that you extend the trial for 30% of the time, and I think that's complicated from a trial logistic perspective (Health Professional)’. This need to minimise the duration of follow-up required was recognised as conflicting with the need to emphasise permanent events ‘30 days in terms of length of time for either technique survival or transfer to hemodialysis should be meaningful because from the patient's point of view, 30 days there's a fair chance that it would be a permanent transfer … … seems to be a bit of a compromise in terms of feasibility as well, not having to wait for example 180 days doing a trial (Health Professional)’.

Ease of adoption. While the core outcome measure of ‘technique survival’ was primarily intended for clinical trials, participants felt that reporting of technique survival by registries was of similar importance. Thus, the definition of technique survival had to be applicable to and feasible to implement in registries. Recording of the reason for transfer to HD was potentially important in the interpretation of international differences in this outcome, particularly when comparing countries with restricted availability of HD. The burden of data collection was cited as the biggest issue when considering adoption by registries, as there could be a need to capture whether a transfer was intended to be temporary or permanent at the point of transfer, reasons for a transfer to HD and all episodes where PD was temporarily stopped.

Discussion

Participants in the consensus workshop for defining a core outcome for ‘technique survival’ in PD were in agreement that it should be defined as a single, clear event, with a majority view that the outcome measure should be defined as the event: ‘transfer to HD’. However, there was a minority view that the definition should be the event of ‘stopping PD’, whether or not the patient had HD subsequently. There was uniform support for changing the nomenclature describing the core outcome measure, avoiding the terms ‘technique’, ‘failure’ and ‘survival’ and instead using simple language reflecting the event defining the outcome measure without any positive or negative connotations. Other principles that should determine aspects of the core outcome were also agreed upon, particularly the competing priorities to be balanced when considering the reason for the event, and how long the change must have been sustained for (Table 1).

Table 1.

Summary table of thematic analysis with illustrative quotes.

Capturing important aspects of the outcome
Requiring a core event to define the outcome	‘… I would prefer to use the term with the word hemo in it, rather than death, personally…’ P ‘So my understanding is of course death is very important but it doesn't usually tell you much about the way that unit is treating patients’. HP ‘Maybe the way to have a more global term would be to say PD stop or PD cessation and to distinguish inside different causes because as it has been said, it can be a good issue if it's for a transplantation or a kidney function recovery, but also for a bad issue and for transfer to hemo’. HP ‘I think we should make the definition in such a way that actually we ask for both death related and death unrelated technique failure as part of the definition, part of this reporting process’. HP ‘There are two fundamental questions. How long did you last? A day, a week, a month, a year, five years? Then why did you stop?’ HP ‘It seems like we would favor a broader definition and apply some exceptions. I think, from what I heard, exceptions would be death or planned conservative care, that are not related to kidney complications as exceptions’. HP HP ‘What is most important to you? Leaving PD at all or leaving PD permanently? Where would you put the weight as being more important to you?’ P ‘After experiencing both, I would say permanently. The technique failure or transfer to HD means I cannot use my PD at all. I've lost it’.
Distinguishing temporary from permanent events	‘I think for me, transfer to HD is usually what I write as well. But I always say temporary, likely long-term, so I give it some quantification, in terms of timeframe instead of specific times to generalise it’. HP ‘even at three or six months, you may not be able to say it's permanent. Once you get to 180 days, you can then say it's definitely permanent because so few people at that point will ever get back to on PD’. HP ‘And I think six months is probably a good time, when you say long-term. Because I have patients who went into HD after peritonitis, and they come back after three to four months, and the reinsertion and restart it back. By six months they usually if they wanted to come back, they would have been back’. HP ‘I also agree with the 30 day definition if we want to have a standardised definition it doesn't mean that we cannot collect different definitions and do our study. But 30 days is meaningful clinically’. HP ‘I personally think 90 days is probably right timeframe for two reasons: one is most people that rest following peritonitis would return to dialysis within three months’. HP “The reason 30 days is (too short) that as soon as you hit 31 days, if you can't get a surgeon to do the revision or you hit 32, that's not really a technique failure. On the other hand, 180 days, if that's somebody that you're resting the peritoneum, you can rest the peritoneum for a year. I don't think that's really a success. I think somewhere in the middle, like 90 days…” HP ‘From my experience, definitely it would be 90 days. That would be a perfect period to say, well somebody is not coming back’. HP ‘To a point that I couldn't do PD anymore. Yes, we have to consider I wouldn't want to do hemo for one day, I'm not sure if I'd want to do it for 180 days, so it would have to be like either 30 day brackets. And my fear about HD is what I've heard over the years. You're less likely to be able to wee and the swelling of the legs et cetera. There's a bit of a fear there. So again, we have to consider the patient's view on whether or not it should be. I'd settle for the 30 days if I had to do hemo. I'd probably settle for 30 days provided I can go back to my PD, but if it was permanent then of course it would be different’. P
Heterogeneous patient experiences of transfers	‘so it's a real concern for us, safety concern’ (in reference to a high rate of transfers to HD). P ‘… whether it's HD transfer or technique survival, it's a mixture of positive reasons and negative reasons. Sometimes it's quite appropriate to have a timely transfer to HD … And other times it's a catastrophic, unexpected, unwanted transfer because they get a peritonitis or some other major mechanical issue….’ HP ‘Yeah, look my priority is to stay on PD for as long as I possibly can. But I'm very aware that there might come a time when I can't stay on PD, so I guess I'm prepared for that. Yeah hopefully I don't have to do that too soon’. P ‘I'd settle for the 30 days if I had to do HD. I'd probably settle for 30 days provided I can go back to my PD, but if it was permanent then of course it would be different’. P ‘I had peritonitis once but I was never ever asked if I wanted to do HD and it was the furthest thing from my mind because the treatment (PD) was so good’. P
Adopting appropriate neutral nomenclature
Conveying with clarity	‘…I think I can probably better understand transfer to HD over technique survival. It just seems to me to make a little bit more sense’. P ‘I was actually pondering on the word technique, because it's not something, I don't think patients use’. P ‘So whether it was infection, whether it was volume overload, whether it was family issues, I think that's more important to quantify why you're transferring instead of just saying transfer and not explaining it for that’. HP ‘I think it's much clearer than using wording like technique survival or failure. I think the words transfer to HD are clearer, less ambiguous, easier for a patient to understand. I still have an issue with the use of the word technique because it implies certain things and some of the reasons you'd transfer to HD are not about technique at all’. P
Avoiding negative connotations	‘By calling (the emergency hotline) means I want to get better and I want the problem solved. I think I would prefer the word survival more than failure’. P ‘So, that makes sense to me to call that survival. I always have a problem with the word failure because it's so … Whether we could just use transfer or discontinuation or something like that’. P ‘it tends to suggest that it's something we've done wrong. That we're at fault for it’ (referring to ‘technique failure’) P ‘I prefer the term transfer to HD, rather than a technique survival or failure. I don't particularly like being told I've failed’. P ‘I think as a carer, if the patient aware what is going on, and you elect your situation, I think for my father, if he choose at the end of the day, he choose to be “I want to have conservative care, because I don't want to be so vigorously go for HD, so sick every time.” I think this is not, I would say, don't see it as a failure, because it's an active choice’. C
Ensuring feasibility and applicability
Capturing data relevant to clinical and research settings	‘We don't capture, for example, the decision whether you would want to have conservative care inset, for example, as a very important decision’. HP ‘I think that a short-term time is … too short is still that time to be able to go back to PD, because that's been my experience anyway. It's been surgery-related. So, yeah. I think it needs that time to go, “Yep, we can rather move back,” or, “No, look, it's not working again. You need to stay on HD, and either then, that's going to be where you are now.” So I think it needs to be three months, at least, I would say’. P ‘Definition, the transport to HD is a very practical and simple, however this cannot reflect the specific temporary transport [sic] to HD’. HP
Ease of adoption	‘In the setting of a clinical trial, every time you extend the need for additional followup to define an endpoint, you bring difficulties. Let's say for a one year follow up trial, definition that includes 90 days means that you extend the trial for 30% of the time’. HP ‘Of course I do believe it should be reported as a continuous variable, but the problem is this is going to be captured in trials where they haven't really thought about this in that sense’. HP

P = patient; HP = health professional; C = caregiver; HD = haemodialysis.

The distinction between transfer to HD and stopping PD as the outcome measure was the main point of discussion raised in the workshop, suggesting that this was viewed as the key issue. The event of transferring to HD is unambiguous and clearly related to technique survival, but whether transplantation, dialysis withdrawal, renal recovery and temporary PD stoppages without any form of dialysis should also be included in an outcome measure of technique survival is less clear. Furthermore, the implications of including temporary PD stoppages without any dialysis in the outcome measure are not clear because, although there is considerable published research on transfers to HD, there is little reported on rates of, and relative importance of different causes of, temporary PD stoppages. PD stoppages are an important problem, often reflecting problems with the technique, but many reasons for PD stoppages do not reflect a problem with the technique, making this a suboptimal measure of technique survival. Using the principles identified within the workshop, a comparison of using transfer to HD or stopping PD is shown in Table 2, supporting the majority view within the workshop that the core outcome measure should be defined as transfer to HD. PD stoppages will need to be captured better to holistically address technique issues in PD, but the complexities and details of this were beyond the scope of this workshop.

Table 2.

Comparison of options for core event defining the outcome according to principles identified during the workshop.

	Transfer to HD	Stopping PD
Capturing important aspects of the outcome	Publications and registry reports available illustrating frequency of outcome	Little research to understand the implications of using this event instead of transfer to HD
	Known to be associated with increased mortality, healthcare expenditure and adverse patient experiences and clear reporting of the reasons that cause this event	Includes all possible reasons for stopping PD
Adopting appropriate neutral nomenclature	Majority in favour of transfer to HD	No consensus on most suitable nomenclature
Capturing data relevant to clinical and research settings	Most registries and trials already collect transfer to HD events so issues mostly around standardisation, definitions and possible need for extra contextual data such as reasons for transfer	Not currently captured in most registries and trials so extra data burden
		Introduces ambiguity in how to distinguish from other events, such as deaths which are a core outcome measure in them self
		If the definition is meaningfully different from transfer to HD, thereby increasing the event rate, this will risk introducing bias against PD by reporting higher event rates, when equivalent stoppages are not reported for haemodialysis

HD = haemodialysis; PD = peritoneal dialysis.

Competing arguments were identified for the duration of the event recorded (i.e. time on HD/off PD). Shorter periods would identify more clinical problems thereby potentially improving governance, ¹³ and they would be easier to implement in clinical trials. Longer periods were felt to be focussing on the more significant permanent events. The duration of time on HD used to define the event of transfer to HD also affects the relative contribution of different reasons.^2,14

Capturing more detail around reasons for the event was considered important to interpret the event defining the core outcome, and to better reflect the patient experience, at the expense of increasing the data collection burden, thereby reducing the likelihood of widespread uptake of the core outcome. When the core outcome is being defined, these competing principles will have to be balanced against each other. One option would be to incorporate more detail such as reasons for the event where it is directly relevant to a trial (e.g. where technique survival is a key outcome that is likely influenced by the intervention), and use the simpler measure by recording the event without further detail in all trials.

One of the drivers underpinning the interest in technique survival is seeking to understand differences in the longevity of PD as a therapy, although PDOPPS has now demonstrated that international differences in the duration of PD mostly reflect differences in transplantation, infection and death. ² It is also now clear that burnout occurs in people undertaking PD, ¹⁵ but neither of these results were available at the time of the workshop and might have influenced the discussions had they been. There is no standardisation in when PD is considered to have started, although the International Society for Peritoneal Dialysis (ISPD) have now convened a group to produce a standardised approach to the metric of time on PD. We also note that, despite the wide geographical representation in the group, this cannot be considered a perfect representation of all perspectives.

In conclusion, technique survival should be defined as transfer to HD. Further agreement on the detail required for a working definition was not possible, but the workshop did establish the principles necessary for this (Table 3). The ISPD working group for time on PD will apply these principles to produce a working definition, including appropriate international collaboration to ensure the definition is widely applicable, and relevant to all healthcare settings. Given that routine reporting of PD outcomes for meaningful quality assurance was a widely agreed principle in the workshop, subsequent endorsement by international societies and national registries will be necessary. To ensure that the outcome measure is reported within trials as intended, liaison with funding and research organisations may be necessary to compel researchers to include core outcome measures in future research, and journal editors to encourage routine reporting of the results (already required by some).

Table 3.

Summary of workshop recommendations for establishing a core outcome measure for technique survival.

Implications for establishing a core outcome measure for technique survival in patients receiving PD
The core outcome should be defined as a transfer to HD, with other events such as death or withdrawal of dialysis captured separately
The terminology used should reflect the event used to define it, avoiding ambiguous terms such as ‘technique’ or terms with positive or negative connotations
A minimum duration of time on HD to define a transfer should be stipulated in the definition, where the duration is considered long enough to imply that the transfer is permanent but not so long that it impairs the ability to record this in clinical studies
The definition of the outcome measure should consider collection of the reasons for transfers to HD, balancing the burden of data collection against the need to aid interpretability and external comparisons and capture important elements of the patient experience
The outcome measure must be relevant to both trials and registries, both of which require simplicity and a low burden of data capture

HD = haemodialysis; PD = peritoneal dialysis.