DIALysis or not: Outcomes in older kidney patients with GerIatriC Assessment (DIALOGICA): rationale and design

Background The incidence and prevalence of older patients with kidney failure who are dependent on dialysis is increasing. However, observational studies showed limited or no benefit of dialysis on mortality in subgroups of these patients when compared to conservative care. As the focus is shifting towards health-related quality of life (HRQoL), current evidence of effects of conservative care or dialysis on HRQoL in older patients is both limited and biased. Dialysis comes with both high treatment burden for patients and high costs for society; better identification of patients who might not benefit from dialysis could result in significant cost savings. The aim of this prospective study is to compare HRQoL, clinical outcomes, and costs between conservative care and dialysis in older patients. Methods The DIALysis or not: Outcomes in older kidney patients with GerIatriC Assessment (DIALOGICA) study is a prospective, observational cohort study that started in February 2020. It aims to include 1500 patients from 25 Dutch and Belgian centres. Patients aged ≥70 years with an eGFR of 10–15 mL/min/1.73m2 are enrolled in the first stage of the study. When dialysis is initiated or eGFR drops to 10 mL/min/1.73m2 or lower, the second stage of the study commences. In both stages nephrogeriatric assessments will be performed annually, consisting of questionnaires and tests to assess most common geriatric domains, i.e. functional, psychological, somatic, and social status. The primary outcome is HRQoL, measured with the Twelve-item Short-Form Health Survey. Secondary outcomes are clinical outcomes (mortality, hospitalisation, functional status, cognitive functioning, frailty), cost-effectiveness, and decisional regret. All outcomes are (repeated) measures during the first year of the second stage. The total follow-up will be a maximum of 4 years with a minimum of 1 year in the second stage. Discussion By generating more insight in the effects of conservative care and dialysis on HRQoL, clinical outcomes, and costs, findings of this study will help patients and physicians make a shared decision on the best individual treatment option for kidney failure. Trial registration The study was registered in the Netherlands Trial Register (NL-8352) on 5 February 2020. Supplementary Information The online version contains supplementary material available at 10.1186/s12882-021-02235-y.


Background
In the Netherlands, over 6200 patients (358 per million population) with kidney failure are currently dependent on dialysis with an annual incidence of over 1650 patients (95 per million population) [1]. The proportion of older patients has risen over the years and in 2019 45% of the incident dialysis patients were 70 years or older [2]. After dialysis initiation, the overall mortality in this age group is high: 18% in the first year, 72% after 5 years [2]. A Dutch retrospective single-centre study showed no survival benefit from dialysis initiation for patients aged 80 years or older and in those above 70 years with severe comorbidity, compared to conservative care (CC) [3]. These results were confirmed by a prospective multicentre study performed in the Netherlands [4]. Results such as these have shifted the focus from mortality towards health-related quality of life (HRQoL) as the most important outcome [5]. Dialysis is associated with a high treatment burden in older patients when compared to CC [6,7]. A recent meta-analysis reviewed 11 studies comparing HRQoL between older patients treated with CC or dialysis [7]. All studies had a small sample size, eight were single-centre, and eight were highly susceptible to selection bias and/or confounding. Thus, current evidence for the effects of CC or dialysis on HRQoL is both limited and biased. Despite these limitations, the authors conclude that CC has the potential to achieve similar HRQoL as dialysis.
Outcomes such as survival and HRQoL might differ greatly among older patients with kidney failure due to substantial heterogeneity with respect to functional and cognitive impairment and frailty. Establishing which of these conditions are associated with poor outcome may help to identify those patients at highest risk and thus guide informed treatment decisions [8]. To do this, some authors suggest tailoring the comprehensive geriatric assessments used in geriatric care to routine kidney failure care [9]. However, studies showing the benefit of these assessments on outcomes, such as survival and HRQoL, in these patients are lacking.
Besides the impact on survival and HRQoL, dialysis is also a highly expensive treatment: the costs in the Netherlands approximate € 92,000 per patient per year, making it among the most expensive treatment in internal medicine, both per individual patient as in total treatment costs [10]. If dialysis treatment appears to yield little or no benefit in selected older and/or frail patients, such as improved survival or HRQoL, CC could be considered more often as the preferred treatment option. This may result in significant cost savings for treatment of kidney failure.
Therefore, large and well-designed, prospective, multicentre studies with sufficient follow-up in older patients treated with CC or dialysis are needed. The DIALysis or not: Outcomes in older kidney patients with GerIatriC Assessment (DIALOGICA) study aims to compare HRQoL, clinical outcomes, and financial costs between CC and dialysis in patients of 70 years or older with kidney failure. Secondly, the study aims to associate clinical and geriatric patient characteristics in the decisionmaking trajectory with these outcomes. Our hypothesis is that CC is associated with comparable HRQoL, comparable clinical outcomes, and lower costs, compared to dialysis in subgroups of older patients with kidney failure.

Study design
DIALOGICA is a multicentre, prospective, observational cohort study comparing CC with dialysis. The first patient was recruited in February 2020. The study aims to include 1500 patients from at least 25 academic and non-academic centres in the Netherlands and Belgium. DIALOGICA is one of the Leading the Change healthcare evaluation projects. Leading the Change finances and guides several healthcare evaluation projects in the Netherlands. The programme is financed by Dutch health insurers and aims to increase and implement acquired knowledge from these projects, thereby increasing the effectiveness of the Dutch healthcare system.
The study consists of two stages (Fig. 1). Patients remain in the first stage as long as their estimated glomerular filtration rate (eGFR), as calculated by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula or using mean urea and creatinine clearances in 24-h urine collections, stays above 10 mL/ min/1.73m 2 . The second stage commences at the start of dialysis or when the eGFR drops to or below 10 mL/ min/1.73m 2 . This creates two groups (patients treated with either dialysis or CC) which are used as the main determinant for the study.
DIALOGICA is aimed to run for a total of 4 years with a minimum follow-up of 1 year in the second stage. The study is conducted according to the principles of the Declaration of Helsinki and the ICH-GCP guidelines. Primary ethical approval was obtained from the medical research ethics committee of the The Hague region (METC Zuidwest Holland) on May 24th 2019 (reference number 19-071). Local medical research ethics committees in participating centres will also be asked for local ethical approval. The study is registered in the Netherlands Trial Register under NL-8352.

Study population
Inclusion criteria are age of 70 years and older and an eGFR between 10 and 15 mL/min/1.73m 2 at the moment of inclusion. Due to the nature of the assessments that will be performed in this study (described in detail later), illiterate patients and those unable to speak Dutch will be excluded. All patients have to provide written informed consent. When a patient appears to have severe cognitive impairment, a geriatrician will judge if a patient is mentally competent to provide the required consent.

Data registration
At baseline, all relevant demographical and clinical data will be collected. In the first stage of the study, a followup visit will be scheduled once-yearly until the start of the second stage. In the second stage, data collection will be performed at start, after 3 months, 6 months, 12 months, and thereafter annually until end of follow-up or end of the study. All data will be recorded in a database developed in collaboration with Nefrovisie. Nefrovisie is the Dutch Quality Institute for Nephrology which hosts the Renine-database, in which all Dutch patients on kidney replacement therapy are registered. Upon consent, data for DIALOGICA will be linked to the Renine-database and the databases of three Dutch studies with similar or complementary study aims and outcomes: the Dutch nOcturnal and hoME dialysis Study To Improve Clinical Outcomes (DOMESTICO) study will compare HRQoL, clinical outcomes and costs in home dialysis patients with in-centre haemodialysis patients; the Pathway for OLDer patients with End-stage Renal disease (POLDER) project will assess the feasibility of implementing a geriatric assessment in routine predialysis care for older patients with an eGFR < 20 mL/ min/1.73m 2 and the Optimizing Access Surgery In Senior haemodialysis patients (OASIS) study will compare three strategies of vascular access in older patients using a randomised design [11][12][13]. The DIALOGICA database will be archived for future research during a minimum of 15 years after the completion of this study.

Nephrogeriatric assessment
To assess the most common geriatric domains (i.e. functional, psychological, somatic, and social status) and the study's outcome measures (i.e. HRQoL, clinical outcomes, and costs), a geriatric assessment tailored to the kidney failure population will be used. It is derived from the consensus-based test set used in a pilot study [11]. This nephrogeriatric assessment (NGA) consists of a combination of 12 questionnaires and seven tests and assessments (Table 1). Nine questionnaires are answered independently by the patient at home and three are conducted by an interviewer. All of the tests and assessments are performed in the participating centres. The NGA is performed at baseline and at a yearly interval during the first stage. Upon entering the second stage, a new NGA will be performed, repeated again at yearly intervals. The questionnaires on HRQoL, symptoms and costs are also repeated after 3 and 6 months into the second stage. To minimise inter-observer bias, the tests will be performed by trained geriatric or (pre) dialysis nurses only.

Functional status
Functionality will be assessed by using four instruments. Firstly, the Activity of Daily Living Scale by Katz et al. assesses the patient's ability to perform activities of daily living independently [14]. It ranks adequacy of performance  measures eight domains of more complex independent living skills, such as handling finances, medication, and driving [15]. Thirdly, hand grip strength will be measured using a Jamar handheld dynamometer. Lastly, patients' fall risk will be assessed by both asking for fall frequency in the last year and their fear of falling, using a numeric rating scale.

Psychological status
Psychological status is assessed by using a total of five instruments, including three cognitive tests. The Montreal Cognitive Assessment scores different types of cognitive abilities, such as orientation, short-term memory, executive function, and visuospatial abilities [17]. The Six-item Cognitive Impairment Test is a fast and simple tool to screen for dementia, consisting of only six questions [18]. Finally, the Letter Digit Subtraction Test is a speed-dependent task that measures complex neuropsychological processes including visual scanning, mental flexibilities, sustained attention, psychomotor speed, and information processing [19]. Depression is assessed in two steps: first two casefinding questions ('Whooley-questions') are used, extended with the fifteen-item Geriatric Depression Scale if one or both of the two initial questions are positive [21,22]. The Life Orientation Test -Revised is a tenitem measure of optimism versus pessimism [23,24,38].

Somatic status
Somatic status will be evaluated by using three instruments. Treating clinicians will use the Clinical Frailty Score, which grades patients from 1 ('very fit') to 9 ('terminally ill') [25]. Comorbidity will be assessed with the Charlson Comorbidity Index [26]. Finally, the surprise question, a screening tool in which caregivers answer the question "Would you be surprised if this patient died within the next twelve months?", will be applied [39]. The Mini Nutritional Assessment -Short Form is used to assess the nutritional status of patients [27]. It is a well validated screening tool for the geriatric population consisting of 18 questions, categorising patients as 'wellnourished', 'at risk of malnutrition' or 'malnourished'.

Social status
Relevant social data (e.g. marital and residential status, biographic data such as educational attainment, and support by home care professionals and/or caregivers) is registered.

HRQoL and symptoms
Patients' HRQoL will be measured by the Twelve-item Short Form Health Survey (SF-12) [28]. Together with the Dialysis Symptom Index (DSI), it was selected as the patient reported outcome measures in Dutch nephrological care by the Dutch Kidney Patients Association, the Dutch Federation for Nephrology, Nefrovisie, and Leiden University Medical Center [29,40]. The SF-12 and DSI were introduced for all Dutch patients on maintenance dialysis in 2018 [40]. The SF-12 consists of 12 questions regarding HRQoL and is the shorter version of the Short Form-36 (SF-36), the most widely used survey to assess HRQoL [28]. The SF-36 consists of eight domains: physical functioning, role-physical, bodily pain, general health, vitality, social function, role-emotional, and mental health. These domains are summarised in the Physical Component Summary Score and the Mental Component Summary Score. In the SF-12 these summary scores are calculated from the twelve most important items (explaining 90% variance) of the SF-36 questionnaire [41]. As the average difference in summary scores between SF-12 and SF-36 is quite small, for timeefficiency reasons, the SF-12 can be used reliably in cohort studies [42]. The DSI is a set of 30 questions evaluating the severity of symptoms in patients with kidney failure (Table S1) [29]. Patients report the level of burden of the symptoms on a five-point Likert scale, ranging from 'not at all bothersome' to 'very bothersome'. Since these symptoms, such as nausea, energy loss, and shortness of breath, are common in patients with kidney failure and not specifically related to dialysis treatment, this questionnaire is also applicable to patients treated with CC.

Costs
For cost-effectiveness analyses, two questionnaires are used to investigate healthcare resources utilisation and patient costs. The Five-level EQ-5D (EQ-5D-5L) is a short and widely used questionnaire in both clinical and health-economic research as its scores can easily be translated into quality adjusted life years (QALYs), the primary outcome measure for most cost-effectiveness research [30,31]. It consists of five questions on domains of HRQoL, such as mobility, pain, and daily functioning. In addition, patients' and healthcare costs and costs with regard to productivity losses will be assessed with a subset of questions from the Institute for Medical Technology Assessment (iMTA) Productivity Cost Questionnaire (iPCQ) and the iMTA Medical Cost Questionnaire (iMCQ) [32,33].

Decisional regret
Decisional regret regarding the choice for CC or dialysis, will be determined with the validated Decisional Regret Scale [36]. Multiple studies assessed the regret of patients to start dialysis with results ranging from 7.4% in a Dutch survey up to 61% in a Canadian study [43][44][45][46].
Currently, there is a lack of data regarding the decisional regret comparing CC with dialysis. An additional questionnaire (Treatment Choice & Decisional Conflict Scale) regarding the treatment choice made will be added to gain more insight in the decisionmaking process regarding the choice between CC and dialysis [34,35].

Outcome variables
The primary outcome parameter is HRQoL (Table 2). DIALOGICA will assess three secondary outcome parameters: clinical outcomes, cost-effectiveness, and decisional regret. The clinical outcomes that will be assessed are mortality, hospitalisation, functional status, cognitive functioning, and frailty. All outcomes are assessed with (repeated) measures during the first year of the second stage.

HRQoL
HRQoL is assessed by comparing repeated measurements of both the calculated Physical Component Summary Score and the Mental Component Summary Score of the SF-12 at the start of the second stage and after 3, 6 and 12 months after the start of the second stage.

Clinical outcomes
Mortality will be analysed as all-cause mortality within the first year of the second stage. The cause of death will also be categorised into seven categories, based on the United Kingdom Renal Registry, using ERA-EDTA codes (Table S2) [16,47]. Hospitalisation will be analysed as all-cause hospitalisation within the first year of the second stage and each individual episode will be categorised into seven categories, using ICD-10 codes (Table S3) [48]. Functional status, cognitive functioning, and frailty will be repeated measures (at the start of the second stage and after 12 months) using their respective instruments, described in the previous paragraphs.

Cost-effectiveness
QALYs will be calculated using the EQ-5D-5L [30]. Total healthcare costs will be calculated using a subset of questions from the iPCQ and iMCQ [32,33]. Costeffectiveness will be assessed as total costs per QALY using incremental cost-effectiveness ratio's (ICERs).

Decisional regret
Decisional regret will be measured using the Decisional Regret Scale, measured at 12 months after the start of the second stage only [36].

Statistical analysis
All statistical analyses will be performed using statistical software, such as SPSS and R. Univariable and multivariable regression analysis to correct for possible confounders will be used to compare groups. Longitudinal data will be analysed with linear and logistic mixed models and presented as estimated coefficients and odds ratios with 95% confidence intervals. Cumulative incidence of hospitalisation and mortality will be reported in Kaplan Meier curves. A Cox proportional hazards model will be used to compare the rate of mortality and hospitalisation between groups. Overall costs and ICERs will be compared across the two groups using linear regression. Discounting, a mathematical procedure for adjusting future costs to their 'present day value', will be applied, as requested for all health economic evaluation exceeding a time frame of 1 year. To deal with missing data, multiple imputation by fully conditional specification will be applied.  [16,47] Hospitalisation (all-cause) ICD-10 codes [48] Functional status Katz-ADL [14], Lawton-iADL [15] Cognitive functioning MoCA [17], 6-CIT [18], LDST [20] Frailty Clinical Frailty Scale [25] Cost-effectiveness ICERs, calculated using EQ-5D-5 L [30] and iPCQ/iMCQ [32,33] Decisional regret Decisional Regret Scale [36] All outcomes are repeated measures during the first year of the second stage (

Sample size calculation
For the primary outcome, we aim to have 80% power to detect a difference of 3.0 points in the SF-12 summary scores 12 months after the start of the second stage. To reject the null hypothesis of equal means with a standard deviation for both groups of 9.0 with a significance level (alpha) of 0.05 using a two-sided two-sample equalvariance Z-test, a required total sample size of 443 patients was calculated (89 patients on CC, 354 on dialysis, 1:4 ratio based on unpublished data from three Dutch centres). When taking into account a dropout rate of 10%, 99 patients on CC and 394 on dialysis are needed, a total of 493 patients. Based on a study from the United States, we expected one third of patients to progress from stage one to two within 3 years, so a cohort of 1500 patients will be needed [49].

Discussion
With an ageing population with kidney failure, more patients will face the decisional moment whether to start dialysis or not. As the survival benefit of dialysis treatment in the geriatric population is debatable, focus has shifted towards HRQoL as primary outcome measure. The impact of dialysis on HRQoL of older patients is probably large, but current data comparing it to CC are both limited and biased, creating an urgent need for high-quality studies [7]. DIALOGICA will be the first large study to prospectively assess the differences between patients choosing CC or dialysis. With a total follow-up of up to 4 years, outcome parameters can be assessed in more detail. Biological age, which also takes factors such as functional status, cognitive functioning, and somatic functioning into account, might influence these outcomes more than calendar age. This study will help to determine individual patient characteristics to identify patients less likely to benefit from dialysis initiation. This will support patients with kidney failure and their nephrologists in making a well-informed and shared decision when discussing renal replacement therapy. To do so, DIALOGICA combines a practical test set of well-validated questionnaires and assessments that can be performed in less than 1 hour per patient. With better identification of patients who might not benefit from dialysis, a significant reduction in costs for treatment of kidney failure can be achieved if these patients choose for CC, since CC has far lower annual treatment costs compared to dialysis [6]. DIALOGICA has an observational study design instead of a randomised controlled design. We deemed the patient's choice between CC and dialysis too fundamentally different to let it be determined by fate. Randomised studies for treatment modalities have been shown to be challenging in patients with kidney failure in general as they have large and different implications for patients' daily life. For example, a previous Dutch study that tried to randomise patients between haemodialysis and peritoneal dialysis failed and the choice whether or not to start dialysis arguably has even more impact [50]. To our knowledge, only one study comparing CC and dialysis is currently recruiting patients using a randomised design [51]. The implication of the observational design of DIALOGICA is that it is more sensitive to confounding by indication, since observed and unobserved variables can influence outcome of treatment. However, selective inclusion into randomised controlled studies can also lead to poor external validity of their results and an observational study may provide a better reflection of daily clinical practice. To account for the influence of the observed variables on treatment outcome and thus to limit confounding by indication, extensive correction for confounding will be applied, using a large set of patients' characteristics.
Treating physicians and patients are not blinded for the results of the NGA, because this could hamper appropriate care for newly diagnosed geriatric impairments. Moreover, as NGAs are becoming standard of care in most Dutch centres, it is ethically undesirable to withhold the results of the NGA from treating physicians and participating patients. Since the results of the NGA might influence the treatment decision both for patients and for treating physicians this can introduce selection bias between patients choosing for CC and dialysis, and could be considered a limitation. Where needed, correction for baseline NGA discrepancies will be applied.

Conclusion
Data on relevant outcomes are needed to answer the question whether CC is a serious alternative to dialysis in (a subgroup of) older patients with kidney failure. In the upcoming years, DIALOGICA will investigate the effect of CC on HRQoL, clinical outcomes, and costeffectiveness in comparison to dialysis in this population, generating more insight to aid doctors and patients in the shared decision making process.
Additional file 1: Table S1. The Dialysis Symptom Index.
Additional file 2: Table S2. Categories for mortality, using ERA-EDTA codes, based on the United Kingdom Renal Registry.