- Email: [email protected]
Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany
G Model PEC 5183 No. of Pages 9
Patient Education and Counseling xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou
Shared decision making
Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany Maxi Robinskia,* , Wilfried Maua , Andreas Wienkeb , Matthias Girndtc a
Institute for Rehabilitation Medicine, Medical Faculty of the Martin Luther University Halle-Wittenberg, Halle (Saale), Germany Institute of Medical Epidemiology, Biostatistics, and Informatics, Medical Faculty of the Martin Luther University Halle-Wittenberg, Halle (Saale), Germany c Department of Internal Medicine II, University Hospital of the Martin Luther University Halle-Wittenberg, Halle (Saale), Germany b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 11 June 2015 Received in revised form 25 September 2015 Accepted 17 October 2015
Objective: To compare differences in shared decision-making (SDM) and treatment satisfaction (TS) between haemodialysis (HD) and peritoneal dialysis (PD) patients. Methods: 6–24 months after initiation of dialysis, we surveyed 780 patients from throughout Germany (CORETH-project) regarding SDM, the reason for modality choice and TS. Data were compared between two age-, comorbidity-, education-, and employment status-matched groups (n = 482). Results: PD patients rated all aspects of SDM more positively than did HD patients (total score: MPD = 84.6, SD = 24.1 vs. MHD = 61.9, SD = 37.3; p 0.0001). The highest difference occurred for the item “announcement of a necessary decision” (delta = 1.3 points on a 6-point Likert-scale). PD patients indicated their desire for independence as a motivator for choosing PD (65%), whereas HD patients were subject to medical decisions (23%) or wanted to rely on medical support (20%). We found positive correlations between SDM and TS (0.16 r 0.48; p 0.0001). Conclusion: Our findings increase awareness of a participatory nephrological counseling-culture and imply that SDM can pave the way for quality of life and treatment success for dialysis patients. Practice implications: Practitioners can facilitate SDM by screening patient preferences at an early stage, being aware of biases in consultation, using easy terminology and encouraging passive patients to participate in the choice. ã 2015 Elsevier Ireland Ltd. All rights reserved.
Keywords: Dialysis modality choice Patient participation Treatment satisfaction Peritoneal dialysis Heuristic decision-making
1. Introduction 1.1. The choice of dialysis modality and shared decision-making Persons suffering from chronic kidney disease (CKD) often have the choice between two significantly different types of dialysis therapy [1]. On the one hand, hemodialysis (HD) is usually performed three times a week in an outpatient unit. Supervised by a medical team, the patient is connected to a blood purification system for four to five hours each session. Consequently, this is a passive treatment. The patient has to follow certain rules regarding diet and the correct medication. On the other hand, there is the option for peritoneal dialysis (PD), which can be performed at home by the patient himself. Through an implanted PD catheter, dialysis fluid is filled into the abdominal cavity, with the
* Corresponding author at: Medical Faculty of the Martin Luther University HalleWittenberg, C/o Institute for Rehabilitation Medicine, 06097 Halle (Saale), Germany. Fax: +49 3455574206. E-mail address: [email protected] (M. Robinski).
peritoneum functioning as a membrane. PD patients treat themselves several times a day by replacing the dialysis fluid. Furthermore, outpatient consultations are necessary only every four to eight weeks, and there are few dietary restrictions. At a minimum, every third CKD patient could opt for both modalities; however, this estimation varies among the nephrology expert community. Decisions in favor of PD only occur in 5% of all cases in Germany, although the two methods are considered to be equivalent in terms of mortality [2,3]. The literature is heterogeneous regarding the “optimal” assignment to one modality, even if some characteristics covariate with the choice [2]. Young and employed patients are, in general, assigned to PD rather than HD. Initiation of HD often occurs as an urgent lifesaving action or as a bridging treatment while waiting for transplantation. Moreover, patients who live closer to the dialysis unit or come from disadvantaged backgrounds are more likely to be treated with HD. The empirical trial to assign patients to one of the two modalities at random failed [4]. In addition, assignment is determined by characteristics of the consulting nephrologist, such as education or attitude toward PD [5].
http://dx.doi.org/10.1016/j.pec.2015.10.014 0738-3991/ ã 2015 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
2
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
Previous findings suggest that approximately one-third of CKD patients feel insufficiently involved in the treatment choice [6,7]. Especially in multi-morbid CKD patients aged 65 years and older, informed decision-making appears to be poorer [8,15]. More vulnerable CKD patients often appear to be excluded from extensive and fair consultation before their choice [16,17], even though they have the same right to obtain unbiased information and to take part in decision-making. However, there are empirical indications that up to 20% of HD patients would have chosen PD if they had received comprehensive consultation before selecting a treatment modality [9]. This number matters in light of the percentage of patients who receive PD treatment in Germany (only 5%; see above). A recent study shows that educating patients with decision aid tools led to a 50/50 distribution of PD and HD choice [34]. Furthermore, the use of decision aids corresponded to a high level of stability regarding the definitive modality, even in unplanned dialysis starters. Another patient survey in Germany that included various indication groups (N = 1500–1800) showed that more than half of the respondents would have preferred a participatory medical decision-making process [9]. This process is called “shared decision-making (SDM)” between doctor and patient. SDM is defined as a decision situation, in which 1) at least two participants are involved, who 2) both share information and 3) take steps to build a consensus about the preferred treatment, and where 4) an agreement is reached on the treatment with joint responsibility [10,11]. A systematic review summarises the barriers (Fig. 1) of SDM [18]. Situational facilitators are, for example, nurses who function as mediators between physicians and patients. SDM is
particularly eased when patients take responsibility for their treatment. In return, physicians create good conditions for SDM when they are able to recognize their patients’ needs (empathy). However, patients cope better with treatment failure if they can assign the responsibility to the physician [18]. A body of literature, including a Cochrane review [19–21], focuses on interventions to improve the adoption of SDM by healthcare professionals. Researchers recommend, for example, staff training and teaching communication skills for talking to patients, decision aids, coaching and prompt sheets for patients, the distribution of information material, and rigorous quality measures. On the one hand, the SDM concept is heterogeneously applied; there are no evidence-based guidelines and no data for CKD patients [12,13,34]. On the other hand, successful SDM can promote treatment satisfaction (TS), adherence and compliance, as well as knowledge about the disease. SDM can reduce symptoms, and hence, even indirect costs [9,14]. Even though SDM has been identified as a key for positive patient-centered outcomes, until now, it was not clear how medical counseling and SDM are perceived from a dialysis patient's perspective, what reasons influence the choice of dialysis modality and how CKD patients evaluate their participation in the process. Additionally, no evidence is available regarding differences in how PD and HD patients rate SDM. 1.2. Summary and research questions To address those gaps, our study aims to investigate CKD patients’ retrospective SDM ratings with respect to the two
Fig. 1. Barriers of SDM [20].
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
3
Fig. 2. Recruited dialysis units all over Germany and flow diagram of sampling. Copyright: card data ã2015 GeoBasis-DE/BKG [ã2009], Google. (Note: A large part of the more than 7300 available dialysis patients had to be excluded because patients were not in line with the inclusion criteria [especially the time criterion of 6–24 months]. A smaller number were not willing to provide consent).
common dialysis modalities (PD vs. HD). It should be noted that both groups of dialysis patients could differ strongly in bio-psychosocial characteristics. Because we cannot assign patients to a treatment option at random [4], we have to control these characteristics a posteriori. However, theoretically, all CKD patients have the right to obtain a comprehensive consultation and SDM, even if they are at a very advanced age or suffer from high morbidity and a lack of social support. Our research questions were as follows: 1. What differences in PD vs. HD patients occur when they retrospectively rate the SDM with their nephrologist at the particular point of dialysis modality choice? 2. How satisfied are PD vs. HD patients with their treatment option at present? 3. Is there an association between the SDM rating and the present TS in dialysis patients? 4. What argument is perceived by PD vs. HD patients as the dominant reason for their dialysis modality choice?
2. Methods 2.1. Design and sampling The study was carried out within the framework of the CORETH project as a two-armed, non-interventional, cross-sectional
multicentre survey, registered in the German Clinical Trials Register (#DRKS00006350) [22]. Patients were recruited from May 2014 to May 2015 from 55 dialysis units throughout Germany. Local nephrologists screened the patients (see Fig. 2); two trained study nurses obtained informed written consent and surveyed the patients using standardised questionnaires. End-stage renal disease patients (ICD: N18.5) were enrolled at 6–24 months (T1) after initiation of dialysis. At this time, the patients still remembered the circumstances of decision-making and had enough experience with their treatment to have a stable opinion of it. The timing also ensured the absence of any acute complications or adaptation problems during the early phase of dialysis. Moreover, the inclusion criteria (absence of acute psychiatric symptoms, ability to read and understand the questionnaire, ability to provide written consent, age 18 years or older) ensured that patients were able to self-rate the aspects of SDM and TS. 2.2. Instruments and outcome measures 2.2.1. SDM the shared decision-making questionnaire—German version (SDM-Q-G) consists of nine items that have been proven to have excellent internal consistency (Cronbach’s a = 0.94) [10] and has been standardised with 2351 respondents. The authors claim that it shows high face and factorial validity, whereby all items can be summarised to one SDM total score. The recommended method of
Fig. 3. Treatment satisfaction items.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
a The t-test was used for continuous variables and the Chi2-test was used for categorical variables. PD, peritoneal dialysis; HD, hemodialysis. The Charlson index was derived from the established 16-item checklist [30]. In the literature, it has proven high retest reliability and validity and is thus well established in the clinical context. However, the Charlson index fails to represent functional limitations from the patient’s perspective. Kuss [29] suggests the use of z-differences, since it is commonly agreed that statistical testing is insufficient for the assessment of the balance of covariates (p-value depends on sample size).
<0.0001 2.00 2.6 (0.5) 0.558 4.9 (2.2) 5.6 (2.3) Mean Charlson index (SD)
4.8 (2.3)
6.0 (2.3)
0.52
5.0 (2.3) <0.0001
5.0 (2.3)
0.04
6.7 (2.1)
6.8 (2.0)
<0.0001 80.16 80.16 100.0 0.0 0.500 25.3 74.7 17.3 82.7 Employment (%) Employed Unemployed
28.3 71.7
12.1 87.9
12.00 12.00
<0.0001 25.1 74.9
24.9 75.1
0.20 0.20
4.7 95.3
1.4 98.6
<0.0001 10.69 8.46 33.33 0.0 40.0 60.6 0.083 24.9 58.7 16.4 Education (%) Lower Medium High
24.7 50.6 24.7
25.0 62.5 12.5
0.19 6.76 9.17
<0.0001 22.6 56.0 21.4
25.7 51.0 23.2
19.5 61.0 19.5
3.25 4.42 1.98
28.6 62.9 8.5
29.6 63.7 6.6
0.487 1.88 1.88 34.8 65.2 40.0 60.0 0.115 32.6 67.4 Gender (%) Female Male
28.3 71.7
34.6 65.4
3.75 3.75
0.079
31.1 68.9
27.8 72.2
34.4 65.6
3.13 3.13
35.0 65.0
1.62 70.2 (10.5) 52.7 (5.3) 0.509 63.2 (15.1) Mean age (SD)
58.6 (15.8)
65.4 (14.2)
0.45
<0.0001 59.3 (15.9)
58.8 (16.0)
59.8 (15.9)
0.06
69.6 (10.8)
zdifferencea PD HD (N = 10) (N = 287) Total (N = 297)
Non-matched
pvaluea zdifferencea HD (N = 241) PD (N = 241) Total (N = 482)
After Propensity Score-Matching
p-valuea zdifferencea HD (N = 529) PD (N = 251)
Before propensity score-matching
Total (N = 780) Characteristic
Table 1 Characteristics of the total sample and the subsamples before and after propensity score-matching.
<0.0001
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx p-valuea
4
scoring the SDM-Q total score is to sum all items and then transform them to a scale on 0–100 by multiplying the summed score by (20/9). Because the questionnaire was dedicated to dialysis modality choice in this survey, we marginally adapted the wording and substituted the word “treatment” in all items with “haemodialysis and peritoneal dialysis”. The items (see below) were rated on a six-point Likert-scale from 0 = “completely disagree” to 5 = “completely agree”. The SDM raw scores thus range from 0 to 5, with higher scores indicating a more positively perception of SDM. 2.2.2. TS TS was investigated via four items according to previous TS assessments in dialysis patients [23,24] (Fig. 3 ). The items were rated on a five-point-scale (from 0 = “not at all satisfied” to 4 = “completely satisfied”), with higher scores indicating greater TS. 2.2.3. Dominant reason for dialysis modality choice Patients were required to indicate the perceived primary reason for the former decision. A team of nephrological and psychological experts developed the items. They were investigated by means of nine binary categories (for example, PD: “I want to be more independent with PD.” vs. HD: “I want to rely on the medical staff's support.”). 2.3. Statistical analyses To model comparability of the PD and HD patients with regards to age, comorbidity (Charlson Comorbidity Index, CCI [30]), education, and employment status, we matched our data using a linear propensity score [25,33]. Each case was simulated as if the CKD patients may have been candidates for both dialysis modalities. The resulting groups were compared with respect to the outcomes. In doing so, our PD patients were compared only to those of our HD patients who showed a similar age, comorbidity, educational level, and employment status. The selection of propensity score-matching variables were based on interviews with nephrological experts experienced in educating and supporting dialysis candidates when being confronted with modality choice. Following expert statements, these four characteristics are usually considered during decision-making: younger, less comorbid and employed patients with a higher educational level are recommended to use PD [2]. Similarly, in previous studies, matching characteristics were selected depending on the research question. For example, in a PD vs. HD mortality comparison, researchers typically focus on socio-demographic factors (age, gender, race) and the comorbidity of patients [26,27]. The selection of matching characteristics is thus heterogeneous and conventionally “heuristic”. Based on the available literature and expert statements, our selected characteristics were reasonable regarding the present research question. The procedure involved a 1:1 matching with the logit-transformed propensity score and an optimal-matching-algorithm with a calliper of 0.2 SD from the linear predictor [28]. Kuss [29] suggests the use of z-differences because it is commonly agreed that statistical testing is insufficient for the assessment of the balance of covariates (p-value depends on sample size). A reduction of z-differences after matching can be interpreted as an approximation between matched groups. Nonmatched cases were subject to a separate description. Mean values (M), standard deviations (SD) and mean differences (delta) were calculated for the total and the subsamples. To evaluate the internal consistency of the SDM items, Cronbach’s a was calculated. The items from the SDM-Q-G were summed and transformed to a scale on 0–100 (total score). Differences were determined by means of t-tests for independent samples. Error
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
probability was set to a = 0.05 and was Bonferroni-adjusted. For TS items, the percentage distribution was inspected. Because TS was interval-scaled, with higher scores indicating higher TS, we assessed associations by correlating the SDM total score with the TS items (Pearson’s coefficient). The self-reported reasons for dialysis treatment were analysed by the inspection of the percentage distribution. Single responses in the category “Other” are reported separately. All analyses were carried out with SPSS 22.0 and R 2.15.0 for Windows.
5
did their matched counterparts. In contrast, the 10 non-matched PD patients were younger (M = 53, SD = 5) and had a lower comorbidity (M = 2.6, SD = 0.5). All of them were employed (100%) and had a medium (40%) or high (60%) educational level. In other words, propensity score-matching resulted in the exclusion of PD patients with very advantageous conditions and non-matched HD patients with significantly worse conditions compared to propensity score-matched HD patients. 3.2. Shared decision-making and treatment satisfaction
2.4. Ethical considerations The study was conducted in accordance with the Code of Ethics of the Declaration of Helsinki and approved by the leading Ethics Committee of the University of Halle-Wittenberg, Germany. The Ethics Committees at each study site also approved the study protocol. Data safety, in accordance with GCP-regulations, has been guaranteed by the “Coordination Centre for Clinical Studies Halle” (KKSH). 3. Results 3.1. Patient characteristics In all, 780 dialysis patients (87.5% of the eligible 891 patients, Fig. 2) participated in the study (PD = 32.2%, HD = 67.8%). Sample characteristics are shown in Table 1. Before propensity score-matching, we found large statistically significant differences in favour of PD patients (see z- and p-values) with respect to age, comorbidity, educational level, and employment status. These differences could be balanced after propensity score-matching so that PD and HD patients were 59 years old on average, had an average comorbidity of 5 points, and were distributed equally among the categories of educational level and employment status. The inspection of non-matched cases (N = 297, 38%) included mostly HD patients (n = 287) who were older (M = 70, SD = 10), had a higher comorbidity (M = 6.8, SD = 2.0), were less likely employed (1.4%) and more frequently had a low educational level (30%) than
Table 2 illustrates the item characteristics for the SDM-Q-G. The mean differences (delta) of the items ranged from deltamin = 0.80 for item #1 to deltamax = 1.28 for item #2. Item #1, on average, received the most positive support in both groups. Internal consistency was excellent (Cronbach’s a = 0.96), thus providing a nearly optimal reliability for the dialysis-specific reworded SDM assessment. The statistical comparison between propensity scorematched PD and HD patients yielded a significantly higher SDM total score in PD patients (t[479] = 7.86, p < 0.0001). Fig. 4 outlines the percent of each response for the TS assessment. In both groups, the majority of patients indicated “completely satisfied” with aspects of dialysis treatment. However, some differences emerged between the groups with PD patients being represented more in the “completely satisfied” category than their HD counterparts. The correlation analysis yielded statistically significantly positive correlations between the SDM total score and TS aspects, with the TS aspects showing moderate intercorrelations. The highest coefficient with respect to SDM occurred for information satisfaction. The lowest positive relationship was found between SDM and patients' satisfaction with non-medical support (Table 3). 3.3. Dominant reason for dialysis modality choice The vast majority of PD patients (65%) indicated “I want to be more independent with PD” as the dominant reason for their modality choice (Fig. 5). In the “Other” category (10%), patients
Table 2 Item characteristics for the SDM-Q-G [10]. #
Item wording
Total (N = 482) Mean (SD)
PD (N = 241) Mean (SD)
HD (N = 241) Mean (SD)
1
My doctor told me that there are different dialysis modalities (hemodialysis and peritoneal dialysis) for treating my medical condition My doctor made clear that a decision between hemodialysis and peritoneal dialysis needs to be made My doctor wanted to know exactly how I want to be involved in making the decision between hemodialysis and peritoneal dialysis My doctor precisely explained the advantages and disadvantages of hemodialysis and peritoneal dialysis My doctor helped me understand all of the information concerning hemodialysis and peritoneal dialysis My doctor asked me which dialysis treatment option (hemodialysis or peritoneal dialysis) I prefer My doctor and I thoroughly weighed the different dialysis treatment options (hemodialysis and peritoneal dialysis) My doctor and I selected a dialysis treatment option (hemodialysis or peritoneal dialysis) together My doctor and I reached an agreement on how to proceed
4.20 (1.62)a 3.75 (1.95) 3.60 (1.96) 3.75 (1.89) 3.82 (1.78) 3.59 (2.05) 3.41 (2.02) 3.37 (2.07) 3.43 (2.04)
4.60 (1.13) 4.38 (1.43) 4.20 (1.53) 4.34 (1.38) 4.37 (1.29) 4.14 (1.68) 3.95 (1.75) 3.93 (1.80) 3.98 (1.76)
3.80 (1.92) 3.10 (2.18) 2.98 (2.15) 3.15 (2.13) 3.28 (2.03) 3.03 (2.34) 2.87 (2.14) 2.80 (2.18) 2.86 (2.14)
SDM total score (transformed 0–100)
72.59 (33.24)
83.80 (24.42)
61.34 (36.97)
2 3 4 5 6 7 8 9
a
Missing
Delta MPD–MHD
1
0.80
6
1.28
10
1.22
4
1.19
3
1.09
5
1.11
5
1.08
2
1.13
6
1.12
1
22.46
0 = completely disagree; 5 = completely agree; PD, peritoneal dialysis; HD, hemodialysis.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
6
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
Fig. 4. Percentage distribution for the treatment satisfaction (TS) items (N = 482; item scaling from 0 = not at all satisfied to 4 = completely satisfied; PD = peritoneal dialysis, HD = haemodialysis; item “TS Non-Medical” had 2 missings).
Table 3 Correlation matrix for the SDM total score and TS items.
SDM total score TS dialysis TS information TS medical TS non-medical
SDM total score
TS dialysis
TS information
TS medical
TS non-medical
1.00
0.19 1.00
0.46 0.50 1.00
0.19 0.47 0.51 1.00
0.16 0.38 0.40 0.59 1.00
Note: All correlations are significant with p < 0.0001. SDM, shared decision making; TS, treatment satisfaction.
stated an “enhanced quality of life”, the “opportunity to work or to study”, the “greater efficiency”, the “lower physical strain” or other “medical causes”, which were not further specified. The other preset reasons were represented with less than 10% in the subsample. None of the PD patients selected the category “I am not familiar with haemodialysis”. In contrast, almost one-quarter of the HD patients (23%) indicated that the choice had been the sole decision of their nephrologist (Fig. 6). Approximately 20% wanted to rely on the support in the dialysis unit by means of deciding on HD. As many as 10% specified that they are not familiar with PD and that this led to their HD treatment. In addition, 16% was sure that the PD catheter would have been embarrassing to them, and 17% chose the “Other” category. These included predominantly medical reasons such as a “scarred peritoneum”. Single HD patients stated that “PD treatment would have been too complicated” to them, that they
wanted to “spatially separate the disease and the private life” or that they wanted to “continue to go swimming”. 4. Discussion and conclusion 4.1. Discussion Our study shows that after propensity score-matching, the group of PD patients, compared to HD patients, indicate a more successful SDM with their nephrologist and higher treatment satisfaction. Additionally, our findings suggest a positive linear association between the retrospective SDM rating and the present treatment satisfaction. PD patients predominantly indicate their desire to be more independent with PD (compared to the option of HD) when asked about the dominant reason for their modality choice. In contrast, almost one-quarter of HD patients indicated that modality choice had been the decision of their nephrologist,
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
7
Fig. 5. Percentage distribution of self-reported dominant reasons for PD treatment (n = 241).
Fig. 6. Percentage distribution of self-reported dominant reasons for HD treatment (n = 241).
and one-fifth wanted to rely on the medical staff’s support. Our findings underpin existing research showing that a relatively large proportion of patients feel inadequately included in the decisionmaking process [6–8]. In addition, this study went beyond previous approaches by both “simulating” the case where CKD patients were eligible for both PD and HD based on the applied propensity score-matching criteria (age, comorbidity, educational level and employment status) and quantitatively comparing PD vs. HD patients with regard to an established SDM measure. The finding that the SDM total score was positively associated with treatment satisfaction in our CKD sample is consistent with existing literature [9,14]. This study is not without limitations. Recruitment by the dialysis centers might lead to selection bias because data collection was limited to voluntarily participating units. However, because dialysis is subject to strict quality control measures in Germany with mandatory communication to the authorities of parameters, such as treatment frequency, urea clearance, and haemoglobin levels, the quality of treatment will likely not be very diverse among the centers. Currently, more than 80,000 individuals in Germany are being treated with renal replacement therapy for
chronic kidney failure, with an annual increase in prevalence of approximately 2–4%. Within our study, we screened 7.312 individuals, with 780 eligible individuals who consented to participate. Hence, more than 79,000 potential CKD patients were not subjects in the study, and our findings should be generalized with caution. The representativeness on the patient level may be limited because we only included dialysis patients who were currently in the dialysis unit. Moreover, our findings should be interpreted with caution because the study is retrospective and not randomized. Although our study highlights significant aspects of the communication between patients and professionals, we cannot answer the question of whether PD patients had a different interaction with their nephrologist than did HD patients. All respondents were recruited 6–24 months after dialysis-initiation resulting in some participants having to remember SDM circumstances from a long time ago. Hence, we cannot draw causal conclusions and entirely rule out memory effects. However, because the moment of CKD diagnosis and its related decisionmaking requirement can both be seen as clinical milestones [31], we assume that patients can recall this autobiographical and emotional moment in detail [32].
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
8
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
The application of propensity score-matching in testing for differences in non-randomised studies has both advantages and disadvantages. On the one hand, the selection bias can be reduced and the comparability of matching variables among CKD patients can be increased. According to the literature, propensity scorematching demonstrates its superiority to the standard of regression adjustment [28]. We applied propensity score-matching to simulate comparability with respect to age, comorbidity, education and employment status between patients and the case that they could have opted for both dialysis modalities. Propensity score-matching can be understood as a form of adjusting for patient characteristics; however, we have to take into consideration that some of the matched HD patients may still not have been eligible for PD. Thus, their consultations may have been less likely to be focused on the decision between HD and PD. This may be especially applicable for the 10% of patients unfamiliar with PD. Additionally, it is reasonable to assume that some doctors may not offer a treatment that is contraindicated or not safe for the individual. However, it is debatable whether doctors could let patients know about treatment options and explained reasons why they ARE NOT eligible for the PD alternative. On the other hand, propensity score-matching has the disadvantage of creating a control group that may deviate significantly from the typical population of HD patients. Empirically, on average, HD-patients are older and show a greater comorbidity than do PD patients; however, these differences disappear in the final matched groups. Additionally, considering that on a theoretical level, every patient has the right to obtain good SDM, it may be irrelevant whether our participants could have opted for both modalities or only for one.
meaningful. Third, by applying a standardised questionnaire, the patient-focused SDM evaluation was unbiased and allowed for the identification of gaps in the individual decision-making process. Our study enhances awareness of the necessity of early screening and providing CKD patients with comprehensive information, combined with a successful and unbiased SDM when choosing a dialysis modality. Even if SDM is not completely applicable because of medical indications (e.g., if the patient is considered for HD only due to chronic peritonitis) or a mild cognitive impairment, the patient should always know about the different dialysis modalities and be involved in a minimum mutual dialog with the doctor regarding treatment options.
4.2. Practice implications
The members of the nursing staff from the different dialysis units that participated in this study are gratefully acknowledged for their cooperation. In particular, we thank the dialysis units in Aachen, Bad Bevensen, Berlin, Bernkastel-Kues, Bonn, Bottrop, Braunschweig, Chemnitz, Coburg, Cottbus, Dessau, Dillingen, Dresden, Duesseldorf I, Duesseldorf II, Erfurt, Essen, Genthin/ Tangermuende, Gera, Gifhorn, Guenzburg, Halle I, Halle II/ Merseburg/Querfurt, Hameln, Hannover I, Hannover II, Hannover III, Heidelberg, Heidenau, Hildesheim, Homburg, Jena, Karlsruhe, Cologne, Leipzig, Magdeburg, Marburg, Meiningen, Memmingen, Merzig, Minden, Muenster, Neusaess, Offenbach, Osnabrueck, Paderborn, Rodewisch, Siegen, Solingen, Stuttgart, Trier, Uelzen, Velbert, Wiesbaden, Zwickau. We also thank Juliane Lamprecht, Annemarie Schubert, Sabrina Frost, Merle Ottich and Denise Neumann for their support in study design, data collection and writing.
In practice, with regard to “good SDM” at the point of dialysis modality choice, nephrologists face specific challenges. These result from concerns when educating dialysis candidates, as well as particular fears in CKD patients (e.g., fear of death, fear of alarms of the dialysis machine, or fear of being limited through the PD catheter). Thus, doctors have to integrate patient history into consultation. Practitioners can avoid falling into the pitfalls of not applying SDM in this specific population when screening patients’ preferences and characteristics at an early stage, being aware of potential short cuts and biasing heuristics in consultation, using easy terminology when educating the patient and encouraging passive patients to ask questions and to participate in the decision. Moreover, to ease the evaluation of patient preferences, the development of valid screening tools and research at the moment of dialysis modality choice is indicated. To enhance the anticipation of being a dialysis patient in the future, the cooperation of peer patients as “models” can be beneficial [31] and is already applied in some dialysis units. According to previous approaches [19–21], potential interventions to improve SDM in the dialysis setting may refer to training dialysis staff and teaching communication skills in working with CKD patients, developing decision aids [34] and even evaluating the learning curve of patients with respect to specific modality knowledge. A first step could be to ask patients to share in developing items for the assessment of SDM and their main reason for modality choice. This is also important because at the diagnosis of end-stage renal failure, patients may experience a deep limiting or changing of their perception of information [31]. 4.3. Conclusion The above-mentioned limitations are balanced by several strengths of our study. First, this study received federal funding, and thus, neutrality of results is warranted. Second, we surveyed a large nationwide sample, and thus, the obtained results are
Conflict of interest None declared. Role of funding The CORETH-project is supported by a grant from the German Federal Ministry of Education and Research (#01GY1324). Authors’ contributions MR interpreted the data and wrote the manuscript. MG and WM designed the study and critically revised the paper. AW provided statistical advice. Acknowledgements
References [1] T. Sensky, Psychosomatic aspects of end-stage renal failure, Psychother. Psychosom. 59 (1993) 56–68. [2] S.V. Jassal, G. Krishna, N.P. Mallick, et al., Attitudes of British Isles nephrologists towards dialysis modality selection: a questionnaire study, Nephrol. Dial. Transplant. 17 (2002) 474–477. [3] E.F. Vonesh, J. Moran, Mortality in end-stage renal disease: a reassessment of differences between patients treated with haemodialysis and peritoneal dialysis, J. Am. Soc. Nephrol. 10 (1999) 354–365. [4] J.C. Korevaar, G.W. Feith, F.W. Dekker, et al., Effect of starting with haemodialysis compared with peritoneal dialysis in patients new on dialysis treatment: a randomized controlled trial, Kidney Int. 64 (2003) 2222–2228. [5] S. Klein, K. Lottmann, P. Gierling, H.H. Bleß, Status quo und Zukunft der Heimdialyse, Reihe: Forschung und Entwicklung im Gesundheitswesen, Bd. 5 [Status quo and future directions of home dialysis. Series: Research and development in health care, 5th ed.], Nomos, Baden-Baden, Germany, 2014. [6] P. Tuso, Choosing wisely and beyond: shared decision making and chronic kidney disease, Perm. J. 17 (2013) 75–78. [7] Bundesverband Niere [German National Kidney Association], CEAPIR– Patientenbefragung in Europa, Therapiewahl: Wie frei ist die Entscheidung? Ergebnisse, Diskussionen, Forderungen [CEAPIR patient survey in Europe. Therapy decision–making: How free is the choice? Results, Discussions, Requirements], Bundesverband Niere, Mainz, Germany, 2012, pp. 1–13.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx [8] M.K. Song, S.E. Ward, The extent of informed decision-making about starting dialysis: does patients’age matter? J. Nephrol. 27 (2014) 571–576. [9] B. Braun, G. Marstedt, Partizipative Entscheidungsfindung beim Arzt: Anspruch und Wirklichkeit [shared decision-making with the physician: theory and practice], Bertelsmann Gesundheitsmonitor 2 (2014) 1–11. [10] L. Kriston, I. Scholl, L. Hölzel, et al., The 9-item shared decision making questionnaire (SDM-Q-9): development and psychometric properties in a primary care sample, Patient Educ. Couns. 80 (2010) 94–99. [11] C. Charles, A. Gafni, T. Whelan, Decision-making in the physician–patient encounter: revisiting the shared treatment decision-making model, Soc. Sci. Med. 49 (1999) 651–661. [12] N. Moumjid, A. Gafni, A. Brémond, et al., Shared decision making in the medical encounter: are we all talking about the same thing, Med. Decis. Making 27 (2007) 539–546. [13] A. Lelie, Decision–making in nephrology: Shared decision making? Patient Educ. Couns. 39 (2000) 81–89. [14] L.E. Oshima, E.J. Emanuel, Shared decision making to improve care and reduce costs, N. Engl. J. Med. 368 (2013) 6–8. [15] E.M. Aasen, M. Kvangarsnes, K. Heggen, Perceptions of patient participation amongst elderly patients with end-stage renal disease in a dialysis unit, Scand. J. Caring Sci. 26 (2012) 61–69. [16] J. Sheu, P.L. Ephraim, N.R. Powe, et al., African American and non-African American patients’ and families’ decision making about renal replacement therapies, Qual. Health Res. 22 (2012) 997–1006. [17] G. Garcia-Garcia, V. Jha, CKD in disadvantaged populations, Kidney Int. 87 (2015) 251–253. [18] N. Joseph-Williams, G. Elwyn, A. Edwards, Knowledge is not power for patients: a systematic review and thematic synthesis of patient-reported barriers and facilitators to shared decision making, Patient Educ. Couns. 94 (2014) 291–309. [19] D.L. Frosch, S.G. May, K.A. Rendle, C. Tietbohl, G. Elwyn, Authoritarian physicians and patients’ fear of being labeled ‘difficult’ among key obstacles to shared decision making, Health Aff. 31 (2012) 1030–1038. [20] M. Körner, H. Ehrhardt, A.K. Steger, J. Bengel, Interprofessional SDM train-thetrainer program Fit for SDM: provider satisfaction and impact on participation, Patient Educ. Couns. 89 (2012) 122–128. [21] F. Légaré, S. Ratté, D. Stacey, J. Kryworuchko, K. Gravel, I.D. Graham, S. Turcotte, Interventions for improving the adoption of shared decision making by healthcare professionals, Cochrane Database Syst. Rev. 5 (2010) CD006732.
9
[22] M. Robinski, W. Mau, J. Lamprecht, et al., The choice of renal replacement therapy (coreth) project: study design and methods, Clin. Kidney J. 7 (2014) 575–581. [23] S.C. Palmer, G. de Berardis, J.C. Craig, et al., Patient satisfaction with in-centre haemodialysis care: an international survey, BMJ Open 4 (2014) e005020. [24] S.Z. Fadem, D.R. Walker, G. Abbott, et al., Satisfaction with renal replacement therapy and education: The American Association of Kidney Patients survey, Clin. J. Am. Soc. Nephrol 6 (2011) 605–612. [25] J.A. Rassen, A.A. Shelat, J.M. Franklin, et al., Matching by propensity score in cohort studies with three treatment groups, Epidemiology 24 (2013) 401–409. [26] C.Y. Chou, C.C. Liang, H.L. Kuo, C.T. Chang, J.H. Liu, H.H. Lin, et al., Comparing risk of new onset diabetes mellitus in chronic kidney disease patients receiving peritoneal dialysis and hemodialysis using propensity score matching, PLoS One 9 (2014) . [27] J.P. Lafrance, E. Rahme, S. Iqbal, N. Elftouh, M. Vallée, L.P. Laurin, et al., Association of dialysis modality with risk for infection-related hospitalization: a propensity score-matched cohort analysis, Clin. J. Am. Soc. Nephrol. 7 (2012) 1598–1605. [28] F.J. Thoemmes, E.S. Kim, A systematic review of propensity score methods in the social sciences, Multivar. Behav. Res. 46 (2011) 90–118. [29] O. Kuss, The z-difference can be used to measure covariate balance in matched propensity score analyses, J. Clin. Epidemiol. 66 (2013) 1302–1307. [30] M.E. Charlson, P. Pompei, K.L. Ales, C.R. MacKenzie, A new method of classifying prognostic comorbidity in longitudinal studies: development and validation, J. Chronic Dis. 40 (1987) 373–383. [31] F. Balck, U. Koch, H. Speidel, Psychonephrologie psychische Probleme bei Niereninsuffizienz [Psychonephrology: psychological problems concerning end-stage renal disease], Springer, Berlin, Germany, 1985. [32] E.A. Kensinger, Remembering the details: effects of emotion, Emot. Rev. 1 (2009) 99–113. [33] E.D. Weinhandl, R.N. Foley, D.T. Gilbertson, et al., Propensity-matched mortality comparison of incident haemodialysis and peritoneal dialysis patients, J. Am. Soc. Nephrol. 21 (2010) 499–506. [34] M. Prieto-Velasco, P. Quiros, C. Remon, Spanish Group for the Implementation of a Shared Decision Making Process for Choice with Patient Decision Aid Tools, The Concordance between Patients’ Renal Replacement Therapy Choice and Definitive Modality: Is It a Utopia? PLoS ONE 10 (10) (2015) e0138811, doi: http://dx.doi.org/10.1371/journal.pone.0138811.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
Patient Education and Counseling xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou
Shared decision making
Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany Maxi Robinskia,* , Wilfried Maua , Andreas Wienkeb , Matthias Girndtc a
Institute for Rehabilitation Medicine, Medical Faculty of the Martin Luther University Halle-Wittenberg, Halle (Saale), Germany Institute of Medical Epidemiology, Biostatistics, and Informatics, Medical Faculty of the Martin Luther University Halle-Wittenberg, Halle (Saale), Germany c Department of Internal Medicine II, University Hospital of the Martin Luther University Halle-Wittenberg, Halle (Saale), Germany b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 11 June 2015 Received in revised form 25 September 2015 Accepted 17 October 2015
Objective: To compare differences in shared decision-making (SDM) and treatment satisfaction (TS) between haemodialysis (HD) and peritoneal dialysis (PD) patients. Methods: 6–24 months after initiation of dialysis, we surveyed 780 patients from throughout Germany (CORETH-project) regarding SDM, the reason for modality choice and TS. Data were compared between two age-, comorbidity-, education-, and employment status-matched groups (n = 482). Results: PD patients rated all aspects of SDM more positively than did HD patients (total score: MPD = 84.6, SD = 24.1 vs. MHD = 61.9, SD = 37.3; p 0.0001). The highest difference occurred for the item “announcement of a necessary decision” (delta = 1.3 points on a 6-point Likert-scale). PD patients indicated their desire for independence as a motivator for choosing PD (65%), whereas HD patients were subject to medical decisions (23%) or wanted to rely on medical support (20%). We found positive correlations between SDM and TS (0.16 r 0.48; p 0.0001). Conclusion: Our findings increase awareness of a participatory nephrological counseling-culture and imply that SDM can pave the way for quality of life and treatment success for dialysis patients. Practice implications: Practitioners can facilitate SDM by screening patient preferences at an early stage, being aware of biases in consultation, using easy terminology and encouraging passive patients to participate in the choice. ã 2015 Elsevier Ireland Ltd. All rights reserved.
Keywords: Dialysis modality choice Patient participation Treatment satisfaction Peritoneal dialysis Heuristic decision-making
1. Introduction 1.1. The choice of dialysis modality and shared decision-making Persons suffering from chronic kidney disease (CKD) often have the choice between two significantly different types of dialysis therapy [1]. On the one hand, hemodialysis (HD) is usually performed three times a week in an outpatient unit. Supervised by a medical team, the patient is connected to a blood purification system for four to five hours each session. Consequently, this is a passive treatment. The patient has to follow certain rules regarding diet and the correct medication. On the other hand, there is the option for peritoneal dialysis (PD), which can be performed at home by the patient himself. Through an implanted PD catheter, dialysis fluid is filled into the abdominal cavity, with the
* Corresponding author at: Medical Faculty of the Martin Luther University HalleWittenberg, C/o Institute for Rehabilitation Medicine, 06097 Halle (Saale), Germany. Fax: +49 3455574206. E-mail address: [email protected] (M. Robinski).
peritoneum functioning as a membrane. PD patients treat themselves several times a day by replacing the dialysis fluid. Furthermore, outpatient consultations are necessary only every four to eight weeks, and there are few dietary restrictions. At a minimum, every third CKD patient could opt for both modalities; however, this estimation varies among the nephrology expert community. Decisions in favor of PD only occur in 5% of all cases in Germany, although the two methods are considered to be equivalent in terms of mortality [2,3]. The literature is heterogeneous regarding the “optimal” assignment to one modality, even if some characteristics covariate with the choice [2]. Young and employed patients are, in general, assigned to PD rather than HD. Initiation of HD often occurs as an urgent lifesaving action or as a bridging treatment while waiting for transplantation. Moreover, patients who live closer to the dialysis unit or come from disadvantaged backgrounds are more likely to be treated with HD. The empirical trial to assign patients to one of the two modalities at random failed [4]. In addition, assignment is determined by characteristics of the consulting nephrologist, such as education or attitude toward PD [5].
http://dx.doi.org/10.1016/j.pec.2015.10.014 0738-3991/ ã 2015 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
2
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
Previous findings suggest that approximately one-third of CKD patients feel insufficiently involved in the treatment choice [6,7]. Especially in multi-morbid CKD patients aged 65 years and older, informed decision-making appears to be poorer [8,15]. More vulnerable CKD patients often appear to be excluded from extensive and fair consultation before their choice [16,17], even though they have the same right to obtain unbiased information and to take part in decision-making. However, there are empirical indications that up to 20% of HD patients would have chosen PD if they had received comprehensive consultation before selecting a treatment modality [9]. This number matters in light of the percentage of patients who receive PD treatment in Germany (only 5%; see above). A recent study shows that educating patients with decision aid tools led to a 50/50 distribution of PD and HD choice [34]. Furthermore, the use of decision aids corresponded to a high level of stability regarding the definitive modality, even in unplanned dialysis starters. Another patient survey in Germany that included various indication groups (N = 1500–1800) showed that more than half of the respondents would have preferred a participatory medical decision-making process [9]. This process is called “shared decision-making (SDM)” between doctor and patient. SDM is defined as a decision situation, in which 1) at least two participants are involved, who 2) both share information and 3) take steps to build a consensus about the preferred treatment, and where 4) an agreement is reached on the treatment with joint responsibility [10,11]. A systematic review summarises the barriers (Fig. 1) of SDM [18]. Situational facilitators are, for example, nurses who function as mediators between physicians and patients. SDM is
particularly eased when patients take responsibility for their treatment. In return, physicians create good conditions for SDM when they are able to recognize their patients’ needs (empathy). However, patients cope better with treatment failure if they can assign the responsibility to the physician [18]. A body of literature, including a Cochrane review [19–21], focuses on interventions to improve the adoption of SDM by healthcare professionals. Researchers recommend, for example, staff training and teaching communication skills for talking to patients, decision aids, coaching and prompt sheets for patients, the distribution of information material, and rigorous quality measures. On the one hand, the SDM concept is heterogeneously applied; there are no evidence-based guidelines and no data for CKD patients [12,13,34]. On the other hand, successful SDM can promote treatment satisfaction (TS), adherence and compliance, as well as knowledge about the disease. SDM can reduce symptoms, and hence, even indirect costs [9,14]. Even though SDM has been identified as a key for positive patient-centered outcomes, until now, it was not clear how medical counseling and SDM are perceived from a dialysis patient's perspective, what reasons influence the choice of dialysis modality and how CKD patients evaluate their participation in the process. Additionally, no evidence is available regarding differences in how PD and HD patients rate SDM. 1.2. Summary and research questions To address those gaps, our study aims to investigate CKD patients’ retrospective SDM ratings with respect to the two
Fig. 1. Barriers of SDM [20].
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
3
Fig. 2. Recruited dialysis units all over Germany and flow diagram of sampling. Copyright: card data ã2015 GeoBasis-DE/BKG [ã2009], Google. (Note: A large part of the more than 7300 available dialysis patients had to be excluded because patients were not in line with the inclusion criteria [especially the time criterion of 6–24 months]. A smaller number were not willing to provide consent).
common dialysis modalities (PD vs. HD). It should be noted that both groups of dialysis patients could differ strongly in bio-psychosocial characteristics. Because we cannot assign patients to a treatment option at random [4], we have to control these characteristics a posteriori. However, theoretically, all CKD patients have the right to obtain a comprehensive consultation and SDM, even if they are at a very advanced age or suffer from high morbidity and a lack of social support. Our research questions were as follows: 1. What differences in PD vs. HD patients occur when they retrospectively rate the SDM with their nephrologist at the particular point of dialysis modality choice? 2. How satisfied are PD vs. HD patients with their treatment option at present? 3. Is there an association between the SDM rating and the present TS in dialysis patients? 4. What argument is perceived by PD vs. HD patients as the dominant reason for their dialysis modality choice?
2. Methods 2.1. Design and sampling The study was carried out within the framework of the CORETH project as a two-armed, non-interventional, cross-sectional
multicentre survey, registered in the German Clinical Trials Register (#DRKS00006350) [22]. Patients were recruited from May 2014 to May 2015 from 55 dialysis units throughout Germany. Local nephrologists screened the patients (see Fig. 2); two trained study nurses obtained informed written consent and surveyed the patients using standardised questionnaires. End-stage renal disease patients (ICD: N18.5) were enrolled at 6–24 months (T1) after initiation of dialysis. At this time, the patients still remembered the circumstances of decision-making and had enough experience with their treatment to have a stable opinion of it. The timing also ensured the absence of any acute complications or adaptation problems during the early phase of dialysis. Moreover, the inclusion criteria (absence of acute psychiatric symptoms, ability to read and understand the questionnaire, ability to provide written consent, age 18 years or older) ensured that patients were able to self-rate the aspects of SDM and TS. 2.2. Instruments and outcome measures 2.2.1. SDM the shared decision-making questionnaire—German version (SDM-Q-G) consists of nine items that have been proven to have excellent internal consistency (Cronbach’s a = 0.94) [10] and has been standardised with 2351 respondents. The authors claim that it shows high face and factorial validity, whereby all items can be summarised to one SDM total score. The recommended method of
Fig. 3. Treatment satisfaction items.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
a The t-test was used for continuous variables and the Chi2-test was used for categorical variables. PD, peritoneal dialysis; HD, hemodialysis. The Charlson index was derived from the established 16-item checklist [30]. In the literature, it has proven high retest reliability and validity and is thus well established in the clinical context. However, the Charlson index fails to represent functional limitations from the patient’s perspective. Kuss [29] suggests the use of z-differences, since it is commonly agreed that statistical testing is insufficient for the assessment of the balance of covariates (p-value depends on sample size).
<0.0001 2.00 2.6 (0.5) 0.558 4.9 (2.2) 5.6 (2.3) Mean Charlson index (SD)
4.8 (2.3)
6.0 (2.3)
0.52
5.0 (2.3) <0.0001
5.0 (2.3)
0.04
6.7 (2.1)
6.8 (2.0)
<0.0001 80.16 80.16 100.0 0.0 0.500 25.3 74.7 17.3 82.7 Employment (%) Employed Unemployed
28.3 71.7
12.1 87.9
12.00 12.00
<0.0001 25.1 74.9
24.9 75.1
0.20 0.20
4.7 95.3
1.4 98.6
<0.0001 10.69 8.46 33.33 0.0 40.0 60.6 0.083 24.9 58.7 16.4 Education (%) Lower Medium High
24.7 50.6 24.7
25.0 62.5 12.5
0.19 6.76 9.17
<0.0001 22.6 56.0 21.4
25.7 51.0 23.2
19.5 61.0 19.5
3.25 4.42 1.98
28.6 62.9 8.5
29.6 63.7 6.6
0.487 1.88 1.88 34.8 65.2 40.0 60.0 0.115 32.6 67.4 Gender (%) Female Male
28.3 71.7
34.6 65.4
3.75 3.75
0.079
31.1 68.9
27.8 72.2
34.4 65.6
3.13 3.13
35.0 65.0
1.62 70.2 (10.5) 52.7 (5.3) 0.509 63.2 (15.1) Mean age (SD)
58.6 (15.8)
65.4 (14.2)
0.45
<0.0001 59.3 (15.9)
58.8 (16.0)
59.8 (15.9)
0.06
69.6 (10.8)
zdifferencea PD HD (N = 10) (N = 287) Total (N = 297)
Non-matched
pvaluea zdifferencea HD (N = 241) PD (N = 241) Total (N = 482)
After Propensity Score-Matching
p-valuea zdifferencea HD (N = 529) PD (N = 251)
Before propensity score-matching
Total (N = 780) Characteristic
Table 1 Characteristics of the total sample and the subsamples before and after propensity score-matching.
<0.0001
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx p-valuea
4
scoring the SDM-Q total score is to sum all items and then transform them to a scale on 0–100 by multiplying the summed score by (20/9). Because the questionnaire was dedicated to dialysis modality choice in this survey, we marginally adapted the wording and substituted the word “treatment” in all items with “haemodialysis and peritoneal dialysis”. The items (see below) were rated on a six-point Likert-scale from 0 = “completely disagree” to 5 = “completely agree”. The SDM raw scores thus range from 0 to 5, with higher scores indicating a more positively perception of SDM. 2.2.2. TS TS was investigated via four items according to previous TS assessments in dialysis patients [23,24] (Fig. 3 ). The items were rated on a five-point-scale (from 0 = “not at all satisfied” to 4 = “completely satisfied”), with higher scores indicating greater TS. 2.2.3. Dominant reason for dialysis modality choice Patients were required to indicate the perceived primary reason for the former decision. A team of nephrological and psychological experts developed the items. They were investigated by means of nine binary categories (for example, PD: “I want to be more independent with PD.” vs. HD: “I want to rely on the medical staff's support.”). 2.3. Statistical analyses To model comparability of the PD and HD patients with regards to age, comorbidity (Charlson Comorbidity Index, CCI [30]), education, and employment status, we matched our data using a linear propensity score [25,33]. Each case was simulated as if the CKD patients may have been candidates for both dialysis modalities. The resulting groups were compared with respect to the outcomes. In doing so, our PD patients were compared only to those of our HD patients who showed a similar age, comorbidity, educational level, and employment status. The selection of propensity score-matching variables were based on interviews with nephrological experts experienced in educating and supporting dialysis candidates when being confronted with modality choice. Following expert statements, these four characteristics are usually considered during decision-making: younger, less comorbid and employed patients with a higher educational level are recommended to use PD [2]. Similarly, in previous studies, matching characteristics were selected depending on the research question. For example, in a PD vs. HD mortality comparison, researchers typically focus on socio-demographic factors (age, gender, race) and the comorbidity of patients [26,27]. The selection of matching characteristics is thus heterogeneous and conventionally “heuristic”. Based on the available literature and expert statements, our selected characteristics were reasonable regarding the present research question. The procedure involved a 1:1 matching with the logit-transformed propensity score and an optimal-matching-algorithm with a calliper of 0.2 SD from the linear predictor [28]. Kuss [29] suggests the use of z-differences because it is commonly agreed that statistical testing is insufficient for the assessment of the balance of covariates (p-value depends on sample size). A reduction of z-differences after matching can be interpreted as an approximation between matched groups. Nonmatched cases were subject to a separate description. Mean values (M), standard deviations (SD) and mean differences (delta) were calculated for the total and the subsamples. To evaluate the internal consistency of the SDM items, Cronbach’s a was calculated. The items from the SDM-Q-G were summed and transformed to a scale on 0–100 (total score). Differences were determined by means of t-tests for independent samples. Error
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
probability was set to a = 0.05 and was Bonferroni-adjusted. For TS items, the percentage distribution was inspected. Because TS was interval-scaled, with higher scores indicating higher TS, we assessed associations by correlating the SDM total score with the TS items (Pearson’s coefficient). The self-reported reasons for dialysis treatment were analysed by the inspection of the percentage distribution. Single responses in the category “Other” are reported separately. All analyses were carried out with SPSS 22.0 and R 2.15.0 for Windows.
5
did their matched counterparts. In contrast, the 10 non-matched PD patients were younger (M = 53, SD = 5) and had a lower comorbidity (M = 2.6, SD = 0.5). All of them were employed (100%) and had a medium (40%) or high (60%) educational level. In other words, propensity score-matching resulted in the exclusion of PD patients with very advantageous conditions and non-matched HD patients with significantly worse conditions compared to propensity score-matched HD patients. 3.2. Shared decision-making and treatment satisfaction
2.4. Ethical considerations The study was conducted in accordance with the Code of Ethics of the Declaration of Helsinki and approved by the leading Ethics Committee of the University of Halle-Wittenberg, Germany. The Ethics Committees at each study site also approved the study protocol. Data safety, in accordance with GCP-regulations, has been guaranteed by the “Coordination Centre for Clinical Studies Halle” (KKSH). 3. Results 3.1. Patient characteristics In all, 780 dialysis patients (87.5% of the eligible 891 patients, Fig. 2) participated in the study (PD = 32.2%, HD = 67.8%). Sample characteristics are shown in Table 1. Before propensity score-matching, we found large statistically significant differences in favour of PD patients (see z- and p-values) with respect to age, comorbidity, educational level, and employment status. These differences could be balanced after propensity score-matching so that PD and HD patients were 59 years old on average, had an average comorbidity of 5 points, and were distributed equally among the categories of educational level and employment status. The inspection of non-matched cases (N = 297, 38%) included mostly HD patients (n = 287) who were older (M = 70, SD = 10), had a higher comorbidity (M = 6.8, SD = 2.0), were less likely employed (1.4%) and more frequently had a low educational level (30%) than
Table 2 illustrates the item characteristics for the SDM-Q-G. The mean differences (delta) of the items ranged from deltamin = 0.80 for item #1 to deltamax = 1.28 for item #2. Item #1, on average, received the most positive support in both groups. Internal consistency was excellent (Cronbach’s a = 0.96), thus providing a nearly optimal reliability for the dialysis-specific reworded SDM assessment. The statistical comparison between propensity scorematched PD and HD patients yielded a significantly higher SDM total score in PD patients (t[479] = 7.86, p < 0.0001). Fig. 4 outlines the percent of each response for the TS assessment. In both groups, the majority of patients indicated “completely satisfied” with aspects of dialysis treatment. However, some differences emerged between the groups with PD patients being represented more in the “completely satisfied” category than their HD counterparts. The correlation analysis yielded statistically significantly positive correlations between the SDM total score and TS aspects, with the TS aspects showing moderate intercorrelations. The highest coefficient with respect to SDM occurred for information satisfaction. The lowest positive relationship was found between SDM and patients' satisfaction with non-medical support (Table 3). 3.3. Dominant reason for dialysis modality choice The vast majority of PD patients (65%) indicated “I want to be more independent with PD” as the dominant reason for their modality choice (Fig. 5). In the “Other” category (10%), patients
Table 2 Item characteristics for the SDM-Q-G [10]. #
Item wording
Total (N = 482) Mean (SD)
PD (N = 241) Mean (SD)
HD (N = 241) Mean (SD)
1
My doctor told me that there are different dialysis modalities (hemodialysis and peritoneal dialysis) for treating my medical condition My doctor made clear that a decision between hemodialysis and peritoneal dialysis needs to be made My doctor wanted to know exactly how I want to be involved in making the decision between hemodialysis and peritoneal dialysis My doctor precisely explained the advantages and disadvantages of hemodialysis and peritoneal dialysis My doctor helped me understand all of the information concerning hemodialysis and peritoneal dialysis My doctor asked me which dialysis treatment option (hemodialysis or peritoneal dialysis) I prefer My doctor and I thoroughly weighed the different dialysis treatment options (hemodialysis and peritoneal dialysis) My doctor and I selected a dialysis treatment option (hemodialysis or peritoneal dialysis) together My doctor and I reached an agreement on how to proceed
4.20 (1.62)a 3.75 (1.95) 3.60 (1.96) 3.75 (1.89) 3.82 (1.78) 3.59 (2.05) 3.41 (2.02) 3.37 (2.07) 3.43 (2.04)
4.60 (1.13) 4.38 (1.43) 4.20 (1.53) 4.34 (1.38) 4.37 (1.29) 4.14 (1.68) 3.95 (1.75) 3.93 (1.80) 3.98 (1.76)
3.80 (1.92) 3.10 (2.18) 2.98 (2.15) 3.15 (2.13) 3.28 (2.03) 3.03 (2.34) 2.87 (2.14) 2.80 (2.18) 2.86 (2.14)
SDM total score (transformed 0–100)
72.59 (33.24)
83.80 (24.42)
61.34 (36.97)
2 3 4 5 6 7 8 9
a
Missing
Delta MPD–MHD
1
0.80
6
1.28
10
1.22
4
1.19
3
1.09
5
1.11
5
1.08
2
1.13
6
1.12
1
22.46
0 = completely disagree; 5 = completely agree; PD, peritoneal dialysis; HD, hemodialysis.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
6
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
Fig. 4. Percentage distribution for the treatment satisfaction (TS) items (N = 482; item scaling from 0 = not at all satisfied to 4 = completely satisfied; PD = peritoneal dialysis, HD = haemodialysis; item “TS Non-Medical” had 2 missings).
Table 3 Correlation matrix for the SDM total score and TS items.
SDM total score TS dialysis TS information TS medical TS non-medical
SDM total score
TS dialysis
TS information
TS medical
TS non-medical
1.00
0.19 1.00
0.46 0.50 1.00
0.19 0.47 0.51 1.00
0.16 0.38 0.40 0.59 1.00
Note: All correlations are significant with p < 0.0001. SDM, shared decision making; TS, treatment satisfaction.
stated an “enhanced quality of life”, the “opportunity to work or to study”, the “greater efficiency”, the “lower physical strain” or other “medical causes”, which were not further specified. The other preset reasons were represented with less than 10% in the subsample. None of the PD patients selected the category “I am not familiar with haemodialysis”. In contrast, almost one-quarter of the HD patients (23%) indicated that the choice had been the sole decision of their nephrologist (Fig. 6). Approximately 20% wanted to rely on the support in the dialysis unit by means of deciding on HD. As many as 10% specified that they are not familiar with PD and that this led to their HD treatment. In addition, 16% was sure that the PD catheter would have been embarrassing to them, and 17% chose the “Other” category. These included predominantly medical reasons such as a “scarred peritoneum”. Single HD patients stated that “PD treatment would have been too complicated” to them, that they
wanted to “spatially separate the disease and the private life” or that they wanted to “continue to go swimming”. 4. Discussion and conclusion 4.1. Discussion Our study shows that after propensity score-matching, the group of PD patients, compared to HD patients, indicate a more successful SDM with their nephrologist and higher treatment satisfaction. Additionally, our findings suggest a positive linear association between the retrospective SDM rating and the present treatment satisfaction. PD patients predominantly indicate their desire to be more independent with PD (compared to the option of HD) when asked about the dominant reason for their modality choice. In contrast, almost one-quarter of HD patients indicated that modality choice had been the decision of their nephrologist,
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
7
Fig. 5. Percentage distribution of self-reported dominant reasons for PD treatment (n = 241).
Fig. 6. Percentage distribution of self-reported dominant reasons for HD treatment (n = 241).
and one-fifth wanted to rely on the medical staff’s support. Our findings underpin existing research showing that a relatively large proportion of patients feel inadequately included in the decisionmaking process [6–8]. In addition, this study went beyond previous approaches by both “simulating” the case where CKD patients were eligible for both PD and HD based on the applied propensity score-matching criteria (age, comorbidity, educational level and employment status) and quantitatively comparing PD vs. HD patients with regard to an established SDM measure. The finding that the SDM total score was positively associated with treatment satisfaction in our CKD sample is consistent with existing literature [9,14]. This study is not without limitations. Recruitment by the dialysis centers might lead to selection bias because data collection was limited to voluntarily participating units. However, because dialysis is subject to strict quality control measures in Germany with mandatory communication to the authorities of parameters, such as treatment frequency, urea clearance, and haemoglobin levels, the quality of treatment will likely not be very diverse among the centers. Currently, more than 80,000 individuals in Germany are being treated with renal replacement therapy for
chronic kidney failure, with an annual increase in prevalence of approximately 2–4%. Within our study, we screened 7.312 individuals, with 780 eligible individuals who consented to participate. Hence, more than 79,000 potential CKD patients were not subjects in the study, and our findings should be generalized with caution. The representativeness on the patient level may be limited because we only included dialysis patients who were currently in the dialysis unit. Moreover, our findings should be interpreted with caution because the study is retrospective and not randomized. Although our study highlights significant aspects of the communication between patients and professionals, we cannot answer the question of whether PD patients had a different interaction with their nephrologist than did HD patients. All respondents were recruited 6–24 months after dialysis-initiation resulting in some participants having to remember SDM circumstances from a long time ago. Hence, we cannot draw causal conclusions and entirely rule out memory effects. However, because the moment of CKD diagnosis and its related decisionmaking requirement can both be seen as clinical milestones [31], we assume that patients can recall this autobiographical and emotional moment in detail [32].
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
8
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx
The application of propensity score-matching in testing for differences in non-randomised studies has both advantages and disadvantages. On the one hand, the selection bias can be reduced and the comparability of matching variables among CKD patients can be increased. According to the literature, propensity scorematching demonstrates its superiority to the standard of regression adjustment [28]. We applied propensity score-matching to simulate comparability with respect to age, comorbidity, education and employment status between patients and the case that they could have opted for both dialysis modalities. Propensity score-matching can be understood as a form of adjusting for patient characteristics; however, we have to take into consideration that some of the matched HD patients may still not have been eligible for PD. Thus, their consultations may have been less likely to be focused on the decision between HD and PD. This may be especially applicable for the 10% of patients unfamiliar with PD. Additionally, it is reasonable to assume that some doctors may not offer a treatment that is contraindicated or not safe for the individual. However, it is debatable whether doctors could let patients know about treatment options and explained reasons why they ARE NOT eligible for the PD alternative. On the other hand, propensity score-matching has the disadvantage of creating a control group that may deviate significantly from the typical population of HD patients. Empirically, on average, HD-patients are older and show a greater comorbidity than do PD patients; however, these differences disappear in the final matched groups. Additionally, considering that on a theoretical level, every patient has the right to obtain good SDM, it may be irrelevant whether our participants could have opted for both modalities or only for one.
meaningful. Third, by applying a standardised questionnaire, the patient-focused SDM evaluation was unbiased and allowed for the identification of gaps in the individual decision-making process. Our study enhances awareness of the necessity of early screening and providing CKD patients with comprehensive information, combined with a successful and unbiased SDM when choosing a dialysis modality. Even if SDM is not completely applicable because of medical indications (e.g., if the patient is considered for HD only due to chronic peritonitis) or a mild cognitive impairment, the patient should always know about the different dialysis modalities and be involved in a minimum mutual dialog with the doctor regarding treatment options.
4.2. Practice implications
The members of the nursing staff from the different dialysis units that participated in this study are gratefully acknowledged for their cooperation. In particular, we thank the dialysis units in Aachen, Bad Bevensen, Berlin, Bernkastel-Kues, Bonn, Bottrop, Braunschweig, Chemnitz, Coburg, Cottbus, Dessau, Dillingen, Dresden, Duesseldorf I, Duesseldorf II, Erfurt, Essen, Genthin/ Tangermuende, Gera, Gifhorn, Guenzburg, Halle I, Halle II/ Merseburg/Querfurt, Hameln, Hannover I, Hannover II, Hannover III, Heidelberg, Heidenau, Hildesheim, Homburg, Jena, Karlsruhe, Cologne, Leipzig, Magdeburg, Marburg, Meiningen, Memmingen, Merzig, Minden, Muenster, Neusaess, Offenbach, Osnabrueck, Paderborn, Rodewisch, Siegen, Solingen, Stuttgart, Trier, Uelzen, Velbert, Wiesbaden, Zwickau. We also thank Juliane Lamprecht, Annemarie Schubert, Sabrina Frost, Merle Ottich and Denise Neumann for their support in study design, data collection and writing.
In practice, with regard to “good SDM” at the point of dialysis modality choice, nephrologists face specific challenges. These result from concerns when educating dialysis candidates, as well as particular fears in CKD patients (e.g., fear of death, fear of alarms of the dialysis machine, or fear of being limited through the PD catheter). Thus, doctors have to integrate patient history into consultation. Practitioners can avoid falling into the pitfalls of not applying SDM in this specific population when screening patients’ preferences and characteristics at an early stage, being aware of potential short cuts and biasing heuristics in consultation, using easy terminology when educating the patient and encouraging passive patients to ask questions and to participate in the decision. Moreover, to ease the evaluation of patient preferences, the development of valid screening tools and research at the moment of dialysis modality choice is indicated. To enhance the anticipation of being a dialysis patient in the future, the cooperation of peer patients as “models” can be beneficial [31] and is already applied in some dialysis units. According to previous approaches [19–21], potential interventions to improve SDM in the dialysis setting may refer to training dialysis staff and teaching communication skills in working with CKD patients, developing decision aids [34] and even evaluating the learning curve of patients with respect to specific modality knowledge. A first step could be to ask patients to share in developing items for the assessment of SDM and their main reason for modality choice. This is also important because at the diagnosis of end-stage renal failure, patients may experience a deep limiting or changing of their perception of information [31]. 4.3. Conclusion The above-mentioned limitations are balanced by several strengths of our study. First, this study received federal funding, and thus, neutrality of results is warranted. Second, we surveyed a large nationwide sample, and thus, the obtained results are
Conflict of interest None declared. Role of funding The CORETH-project is supported by a grant from the German Federal Ministry of Education and Research (#01GY1324). Authors’ contributions MR interpreted the data and wrote the manuscript. MG and WM designed the study and critically revised the paper. AW provided statistical advice. Acknowledgements
References [1] T. Sensky, Psychosomatic aspects of end-stage renal failure, Psychother. Psychosom. 59 (1993) 56–68. [2] S.V. Jassal, G. Krishna, N.P. Mallick, et al., Attitudes of British Isles nephrologists towards dialysis modality selection: a questionnaire study, Nephrol. Dial. Transplant. 17 (2002) 474–477. [3] E.F. Vonesh, J. Moran, Mortality in end-stage renal disease: a reassessment of differences between patients treated with haemodialysis and peritoneal dialysis, J. Am. Soc. Nephrol. 10 (1999) 354–365. [4] J.C. Korevaar, G.W. Feith, F.W. Dekker, et al., Effect of starting with haemodialysis compared with peritoneal dialysis in patients new on dialysis treatment: a randomized controlled trial, Kidney Int. 64 (2003) 2222–2228. [5] S. Klein, K. Lottmann, P. Gierling, H.H. Bleß, Status quo und Zukunft der Heimdialyse, Reihe: Forschung und Entwicklung im Gesundheitswesen, Bd. 5 [Status quo and future directions of home dialysis. Series: Research and development in health care, 5th ed.], Nomos, Baden-Baden, Germany, 2014. [6] P. Tuso, Choosing wisely and beyond: shared decision making and chronic kidney disease, Perm. J. 17 (2013) 75–78. [7] Bundesverband Niere [German National Kidney Association], CEAPIR– Patientenbefragung in Europa, Therapiewahl: Wie frei ist die Entscheidung? Ergebnisse, Diskussionen, Forderungen [CEAPIR patient survey in Europe. Therapy decision–making: How free is the choice? Results, Discussions, Requirements], Bundesverband Niere, Mainz, Germany, 2012, pp. 1–13.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014
G Model PEC 5183 No. of Pages 9
M. Robinski et al. / Patient Education and Counseling xxx (2015) xxx–xxx [8] M.K. Song, S.E. Ward, The extent of informed decision-making about starting dialysis: does patients’age matter? J. Nephrol. 27 (2014) 571–576. [9] B. Braun, G. Marstedt, Partizipative Entscheidungsfindung beim Arzt: Anspruch und Wirklichkeit [shared decision-making with the physician: theory and practice], Bertelsmann Gesundheitsmonitor 2 (2014) 1–11. [10] L. Kriston, I. Scholl, L. Hölzel, et al., The 9-item shared decision making questionnaire (SDM-Q-9): development and psychometric properties in a primary care sample, Patient Educ. Couns. 80 (2010) 94–99. [11] C. Charles, A. Gafni, T. Whelan, Decision-making in the physician–patient encounter: revisiting the shared treatment decision-making model, Soc. Sci. Med. 49 (1999) 651–661. [12] N. Moumjid, A. Gafni, A. Brémond, et al., Shared decision making in the medical encounter: are we all talking about the same thing, Med. Decis. Making 27 (2007) 539–546. [13] A. Lelie, Decision–making in nephrology: Shared decision making? Patient Educ. Couns. 39 (2000) 81–89. [14] L.E. Oshima, E.J. Emanuel, Shared decision making to improve care and reduce costs, N. Engl. J. Med. 368 (2013) 6–8. [15] E.M. Aasen, M. Kvangarsnes, K. Heggen, Perceptions of patient participation amongst elderly patients with end-stage renal disease in a dialysis unit, Scand. J. Caring Sci. 26 (2012) 61–69. [16] J. Sheu, P.L. Ephraim, N.R. Powe, et al., African American and non-African American patients’ and families’ decision making about renal replacement therapies, Qual. Health Res. 22 (2012) 997–1006. [17] G. Garcia-Garcia, V. Jha, CKD in disadvantaged populations, Kidney Int. 87 (2015) 251–253. [18] N. Joseph-Williams, G. Elwyn, A. Edwards, Knowledge is not power for patients: a systematic review and thematic synthesis of patient-reported barriers and facilitators to shared decision making, Patient Educ. Couns. 94 (2014) 291–309. [19] D.L. Frosch, S.G. May, K.A. Rendle, C. Tietbohl, G. Elwyn, Authoritarian physicians and patients’ fear of being labeled ‘difficult’ among key obstacles to shared decision making, Health Aff. 31 (2012) 1030–1038. [20] M. Körner, H. Ehrhardt, A.K. Steger, J. Bengel, Interprofessional SDM train-thetrainer program Fit for SDM: provider satisfaction and impact on participation, Patient Educ. Couns. 89 (2012) 122–128. [21] F. Légaré, S. Ratté, D. Stacey, J. Kryworuchko, K. Gravel, I.D. Graham, S. Turcotte, Interventions for improving the adoption of shared decision making by healthcare professionals, Cochrane Database Syst. Rev. 5 (2010) CD006732.
9
[22] M. Robinski, W. Mau, J. Lamprecht, et al., The choice of renal replacement therapy (coreth) project: study design and methods, Clin. Kidney J. 7 (2014) 575–581. [23] S.C. Palmer, G. de Berardis, J.C. Craig, et al., Patient satisfaction with in-centre haemodialysis care: an international survey, BMJ Open 4 (2014) e005020. [24] S.Z. Fadem, D.R. Walker, G. Abbott, et al., Satisfaction with renal replacement therapy and education: The American Association of Kidney Patients survey, Clin. J. Am. Soc. Nephrol 6 (2011) 605–612. [25] J.A. Rassen, A.A. Shelat, J.M. Franklin, et al., Matching by propensity score in cohort studies with three treatment groups, Epidemiology 24 (2013) 401–409. [26] C.Y. Chou, C.C. Liang, H.L. Kuo, C.T. Chang, J.H. Liu, H.H. Lin, et al., Comparing risk of new onset diabetes mellitus in chronic kidney disease patients receiving peritoneal dialysis and hemodialysis using propensity score matching, PLoS One 9 (2014) . [27] J.P. Lafrance, E. Rahme, S. Iqbal, N. Elftouh, M. Vallée, L.P. Laurin, et al., Association of dialysis modality with risk for infection-related hospitalization: a propensity score-matched cohort analysis, Clin. J. Am. Soc. Nephrol. 7 (2012) 1598–1605. [28] F.J. Thoemmes, E.S. Kim, A systematic review of propensity score methods in the social sciences, Multivar. Behav. Res. 46 (2011) 90–118. [29] O. Kuss, The z-difference can be used to measure covariate balance in matched propensity score analyses, J. Clin. Epidemiol. 66 (2013) 1302–1307. [30] M.E. Charlson, P. Pompei, K.L. Ales, C.R. MacKenzie, A new method of classifying prognostic comorbidity in longitudinal studies: development and validation, J. Chronic Dis. 40 (1987) 373–383. [31] F. Balck, U. Koch, H. Speidel, Psychonephrologie psychische Probleme bei Niereninsuffizienz [Psychonephrology: psychological problems concerning end-stage renal disease], Springer, Berlin, Germany, 1985. [32] E.A. Kensinger, Remembering the details: effects of emotion, Emot. Rev. 1 (2009) 99–113. [33] E.D. Weinhandl, R.N. Foley, D.T. Gilbertson, et al., Propensity-matched mortality comparison of incident haemodialysis and peritoneal dialysis patients, J. Am. Soc. Nephrol. 21 (2010) 499–506. [34] M. Prieto-Velasco, P. Quiros, C. Remon, Spanish Group for the Implementation of a Shared Decision Making Process for Choice with Patient Decision Aid Tools, The Concordance between Patients’ Renal Replacement Therapy Choice and Definitive Modality: Is It a Utopia? PLoS ONE 10 (10) (2015) e0138811, doi: http://dx.doi.org/10.1371/journal.pone.0138811.
Please cite this article in press as: M. Robinski, et al., Shared decision-making in chronic kidney disease: A retrospection of recently initiated dialysis patients in Germany, Patient Educ Couns (2015), http://dx.doi.org/10.1016/j.pec.2015.10.014