Satisfaction of patients on chronic haemodialysis and peritoneal dialysis

Background: In contrast to quality of life, patient satisfaction on chronic haemodialysis (HD) and peritoneal dialysis (PD) has only rarely been studied. Patients and methods: All chronic HD and PD patients of the 19 centres located in western Switzerland were asked to complete a specific questionnaire, assessing dialysis centre characteristics, treatment modalities, and information received before and during dialysis treatment. Comparison between satisfaction with PD and HD was carried out on the patients in the nine centres offering both treatment modalities. Results: Of the 558 questionnaires distributed to chronic HD patients, 455 were returned (response rate 82%). Fifty of 64 PD patients (78%) returned the questionnaire. The two groups were similar in age, gender, and duration of dialysis treatment. Completion rates were >90% for a majority of questions, with the lowest rate for information on sexuality (49% in HD and 54% in PD respectively). The lowest scores were recorded for information received about complications and costs of dialysis, and impact of end-stage kidney disease on sexuality. Satisfaction was lower in anonymous questionnaires. Satisfaction of PD patients was significantly better in 50% of the questions, particularly session tolerance (p <0.001), information about dialysis sessions (p = 0.007), and complications (p = 0.006). Conclusions: PD patients were on average more satisfied with their treatment than HD patients. Satisfaction could be improved with more information about potential adverse treatment effects.

The evaluation of patients' satisfaction when dealing with chronic illnesses is an increasingly important domain.Firstly, patient satisfaction is associated with adherence to treatment regimens.Furthermore, patient satisfaction is also increasingly considered as an important outcome in its own right.For instance, it is one of nine key dimensions of quality addressed in the European Foundation for Quality Management's excellence model [1].
In contrast to quality of life, only few previous studies have addressed patient satisfaction with renal replacement therapy.In a peritoneal dialysis (PD) facility, determinants of patient satisfaction were shown to include the amount of information provided by the staff, the personal atmosphere of the facility, the efficiency of the delivery of dialysis supplies, and the availability of the primary nurse [2].Among patients on haemodialysis (HD), patient satisfaction has been linked with adherence to dialysis treatment regimens [3], and with a favourable health outcome [4].Only one questionnaire is specific for this patient population [5], but it is of considerable length, and exists only in English.A shorter form was recently used for comparing patient ratings of PD versus HD, seven weeks after treatment initiation, and showed better results for PD [6].However, these last two instruments include different dimensions such as quality of life or patient preferences, and questions about satisfaction are only part of them.We therefore adapted a generic questionnaire for assessment of patient satisfaction applicable to both chronic HD and PD.The purpose of this study was to investigate the satisfaction with care experienced by patients on HD and PD, and to compare the dimensions shared by the two treatment modalities in centres offering both types of treatment.

Introduction Study population
Eligible subjects for this cross-sectional survey were all end-stage renal failure (ESRF) patients treated in all the dialysis centres (19 offering HD and 9 HD and PD) of Western Switzerland (population of about 1.8 million inhabitants), as part of a larger examination of quality of care in chronic dialysis [7].Two centres treating only one patient each were not included in the present analysis.Each patient was asked to complete a questionnaire assessing dialysis centre characteristics, treatment modalities (nurses' and physicians' involvement, security and tolerance of treatment), and the information received before the initiation of dialysis treatment or later during dialysis care (drug treatment, side-effects, dietary restrictions, physical activity, sexuality, cost of care, contact in case of an emergency, and global satisfaction).
Questionnaires were distributed by the nursing staff, and the patients were asked to respond to the questions at home, and to return the questionnaire to the dialysis centre.Collection boxes were provided for the anonymous return of questionnaires.The patients had the possibility to fill in their name at the beginning of the questionnaire if they wanted to do so.After collection of the questionnaires at the foreseen date, no recall procedure was carried out.The protocol was approved by the research ethics committee of Lausanne and Geneva University Hospitals.

Questionnaire
Development of the questionnaire was carried out in 2 steps.First, patients and nurses in one dialysis centre offering HD and PD identified dimensions of satisfaction through a focus group, similar to the process reported by Bass et al. [8].The identified dimensions were grouped into three chapters: access, dialysis care, and information received.Secondly, the relevant questions covering these dimensions were directly selected from the validated French version of the Sequs ® data bank of questions for assessing patient satisfaction with outpatient care [9].They included both "positive aspects of care", such as the benefits linked with treatment, as well as "negative aspects of care", such as side effects, costs and handicap.The only modification carried out was replacing the term "treatment" by "dialysis" in order to be more specific, and avoid confusion between dialysis sessions and other treatments, such as medications, often associated in these conditions.Response choices consisted of a five point Likert scale with the following response options: fully agree / rather agree / rather disagree / totally disagree, and a last option (not applicable or does not know).Patients also had the option of adding individual personal comments to each dimension and at the end of the questionnaire.The final form of the questionnaire, without the response scales, is displayed in the Appendix.
Adaptation of the questionnaire to PD was made by omiting all questions about access to the HD centre and HD sessions, except for a question about the outpatient visit facility.

Data analysis
Responses were coded with a value of 1 for "fully agree", 2 for "partially agree", 3 for "partially disagree", 4 for "totally disagree", and missing for "not applicable" or "does not know".Hence, the greater the agreement with a statement, the closer to 1 was the observed value.Computation of mean values and standard deviations were carried out after exclusion of "not applicable" or "does not know" responses, to keep only interpretable answers.Categorical variables were compared with the chi-square test and continuous variables with the Mann-Whitney U or Student's t-tests.Analyses were carried out on the whole set of responses for computing question completion rates.We computed Cronbach's alpha coefficients to assess the internal consistency of domain-specific scales [10].Multiple correlations were carried out with Spearman's correlation coefficients in order to detect potentially redundant questions.The results for each specific centre were compared to the overall mean.Results were then analyzed by dialysis treatment modality.
Comparison between HD and PD results was carried out on the 335 HD patients and 50 PD patients from the nine centres offering both treatment modalities.We used a generalized linear mixed model for ordinal response treating centres as a random effect, and age, gender, duration of dialysis treatment, as well as type of treatment as fixed effects [11].Satisfaction scores were recoded as 0 for "fully agree", 1 for "partially agree", and 2 for "partially disagree" or "totally disagree".Using the "gllamm" procedure of STATA [12], models were constructed using c-variables values centered for the mean for age, and duration of treatment, and binary values for gender (female = 1, male = 0), and treatment modality (PD = 1, HD = 0).The reference patient was a 64-year old man, on HD for 8 years.For the 24 models analyzed, negative regression coefficients and z values indicate better satisfaction for PD patients as compared with HD patients.Statistical significance was defined as a p value <0.05.

Results
Centres treated a median of 24 HD patients (range 6-68), and were equipped with a median of 10 dialysis beds (range 3-26, total 190).In the nine units performing PD, a median of 6 PD patients (1-20, total 64) were treated per centre.
Of the 558 questionnaires distributed to chronic HD patients, 455 were returned (response rate 82%), of which only 127 (28%) were submitted anonymously.This response rate ranged from 55.6% to 100% among centres, with a median 92.9%.For PD patients, of the 64 questionnaires distributed, 50 were returned (response rate 78%), of which 17 (34%) were anonymous.The response rate ranged from 50% to 100% for the nine centres under study, with a median of 90%.Almost all patients (98%) were Caucasian.HD and PD patients were not statistically different with respect to age (64 ± 15 vs. 60 ± 17 years), gender (male gender 63% vs. 55%), and duration of dialysis treatment (4.0 ± 4.5 vs. 3.2 ± 4.9 years).These participants' characteristics were also similar to those of the whole dialysis population treated in the study's dialysis centres, and no differences were observed among the different centres.
Altogether, the questionnaire was well accepted, as completion rates per question were above 90% for 17 out of 39 questions in HD and 19 out of 26 questions in PD, and above 80% for an additional 6 and 5 questions, respectively.The lowest question completion rates were recorded for the question about receiving information on sexual activity (49% in HD, and 54% in PD, respectively).Responses were distributed in all four grades of response categories.The only exception was nursing care, for which no patient used the most negative response option.Cronbach's alpha, ranged between 0.77 (dialysis centre characteristics) and 0.89 (patient information) for HD patients and between 0.72 (patient care) and 0.94 (patient information) for PD patients.
The correlation coefficients between the different questions were always smaller than 0.50, except for questions regarding day and time of dialysis treatment (r = 0.67), modality of transportation and transportation time (r = 0.59), medical care and frequency of medical visits (r = 0.55), dialysis room organization, and armchair or bed installation (r = 0.51), Finally, three correlations were of moderate importance (i.e. between 0.30 and 0.50): between information about dialysis possibilities and dialysis complications (r = 0.48); between information about drug treatment and complications of drug treatment (r = 0.43); and between global satisfaction and centre recommendations to family (r = 0.38).Altogether, there were no statistically significant differences in the distribution of results among the 19 different HD centres or the 9 PD centres.
Findings about dialysis centre access and other facility attributes are displayed in table 1. Question completion rates varied greatly depending on the specific question, as all questions did not apply to all patients.Some patients chose not to answer questions that were not applicable to them, while others responded by checking "not applicable" or "does not know".In general, the questions for which there were the greatest amount of missing data were also those for which there was the highest frequency of responses of "not applicable or does not know", translated into "no opinion" in the tables.
The distribution of responses about dialysis care for HD and PD is displayed in table 2, and about satisfaction with the receipt of health information in table 3. "Positive" information was rated higher than "negative" information, such as information about dialysis complications, drug treatment complications, restrictions of physical activity or particularly sexuality, and cost of treatment.Globally, satisfaction of PD patients was rated higher than satisfaction of HD patients, and anonymous responses were on average lower rated than signed ones: our HD patients who chose to remain anonymous were up to 13% less satisfied than those who signed their questionnaire.This difference also existed in PD patients, with patients who chose to remain anonymous being up to 24% less satisfied.Comparison of satisfaction between HD and PD patients of the 9 centres offering both treatment modalities are displayed in table 4.Among the covariables, age played a significant role in 4 models (frequency of medical visit [p = 0.025], ease of access to medical advice [p = 0.034], confidentiality of discussion about health [p = 0.022], and information about costs [p = 0.008]), younger patients being more satisfied than older patients.Similarly, women were statistically significantly more satisfied than men as to the tolerance of dialysis sessions (p = 0.034), information about dialysis sessions (p = 0.041), and were more willing to recommend the centre to relatives (p = 0.035).Pa-tients on dialysis for a shorter time were statistically more satisfied about information on food restriction (p = 0.044), physical activity (p = 0.039), and costs (p = 0.022).Statistically significant differences between HD and PD patients were observed in 12 of the 24 models tested.All were in favour of PD (negative coefficients), and of greater magnitude than the differences observed on the other co-variables.The most important one was about tolerance of dialysis session (p <0.001), followed by information about complications, dialysis session, and confidentiality of discussion about health.Detailed results are displayed in table 4.

Regression model
Regression

Discussion
This study demonstrated that on average, patient satisfaction with care on chronic dialysis, as recorded with the validated Sequs ® data bank of questions, was better on PD than on HD, but not homogenous across the different dimensions included in the study questionnaire, and was associated with patient characteristics.
Patient satisfaction with the "negative" aspects of dialysis care, such as information about complications, costs, or daily restrictions, was lower than satisfaction with "positive" aspects, such as nursing or medical care, or information about dialysis sessions, or emergency contact.However, information should be provided along all these dimensions, if physicians and nurses caring for these patients want to provide better care.The distribution of responses across the different response categories was highly variable according to the specific question.The response distribution to the question about tolerance to dialysis sessions, which showed the second worst score was striking.The question on information about sexual function was also remarkable for its low question completion rate, despite many patients stating that they wanted, but did not receive information on sexuality.Globally however, patients were satisfied with their dialysis treatment and the information provided to them.
This study also showed that tolerance of dialysis treatment was a major concern for these patients, and was 30% better rated in PD than in HD.Altogether, PD patients were more satisfied, and seemed to be more involved in their treatment than HD patients.These results are in agreement with those recently published with the CHOICE questionnaire, comparing patient ratings seven weeks after dialysis initiation, which showed that PD patients were more likely to give excellent ratings of dialysis care and specific aspects of care than HD patients [6].In our study, this observation proved to be long term, as mean treatment duration was 4.0 and 3.2 years in HD and PD patients, respectively.
In addition, our results confirmed that providing information to patients is a difficult task that must be tailored to patients' need, and take into account the fact that patients do not always feel comfortable voicing their needs and concerns, as anonymous responses were on average lower rated than signed ones.
Centre characteristics are also likely to play a role in patient satisfaction.It is well known that practice variations exist in all fields in medicine and can have an important impact on patient outcome [13].Similarly, nephrologists have various opinions and practice about the value and the usefulness of exercise counselling [14].Therefore, if patient satisfaction is to be addressed optimally, information should always be available, preferably in written form, so that it can be provided whenever the patient requests it.This can prove difficult in countries with patients of different ethnicities and levels of education.Our results clearly showed room for improvement in this domain.
Few data have been reported about patient satisfaction with dialysis care.Emphasis has been given to psychosocial factors [4,15,16], and their association with compliance and outcome.Depression has been identified as the most common psychiatric disorder in dialysis patients [4,15].In end-stage renal disease, no association between satisfaction with care and level of depression has been identified [3], but satisfaction with nephrologists care was clearly correlated with physician attendance at dialysis sessions, as well as with the perception of social support, and indicators of quality of dialysis, such as measurement of protein catabolic rate, and serum albumin concentration [3].All these results indicate that optimal care of these patients must be interdisciplinary.
Our study has a number of limitations.Firstly, it was designed to be an anonymous prevalence study and used only satisfaction data, for which no gold standard exists.As a consequence, no systematic analysis could be performed that included case-mix variables, process of care variables, such as time after dialysis initiation, or outcome vari-ables, such as adequacy of dialysis, which all were found to have an impact both on physical and emotional dimensions of quality of life as assessed by SF-36 questionnaire [17], or on satisfaction with care as mentioned above [3].Secondly, it included some dialysis centres with very few patients, which may have biased answers for fear of loss of anonymity.However, this bias seemed to be rather limited, and was clearly noticeable only in one centre treating six patients only.Finally, as our study was a prevalence study, it cannot be excluded that modality selection influenced the results as more independent patients might be over-represented in PD.
As compared with other instruments such as the CHOICE study questionnaire with its 21 dimensions and 83 questions [5], our instrument assessed only 3 dimensions covering patient satisfaction with dialysis centre characteristics, dialysis sessions course, and information.As a consequence, the results obtained from the two types of instruments cannot be compared.However, as our instrument's development involved patients and dialysis nurses for the selection of questions to be included, it should closely reflect their preoccupations, which sometimes can be specific to our health care system.Adaptation to another health care system's characteristics might be required, but this limitation also applies to other existing instruments such as the KDQol [5].
In conclusion, because it showed heterogeneous responses between patient characteristics and the different dimensions of satisfaction, this questionnaire could be a promising surveillance tool for dialysis patients.It was well accepted, easy to use, and responsive in our HD and PD population.Its results allow each individual dialysis centre to set up corrective and preventive measures aimed at improving patient satisfaction with care in very specific domains.The very high response rate recorded in this study was also encouraging.It showed that patients were willing to participate in this kind of survey.tAs a consequence, it could be extended to other dialysis centres.
Med Wochenschr (1871-2000) Swiss Med Wkly (continues Schweiz Med Wochenschr from 2001) Editores Medicorum Helveticorum • SMW's impact factor has been steadily rising, to the current 1.537 • Open access to the publication via the Internet, therefore wide audience and impact • Rapid listing in Medline • LinkOut-button from PubMed with link to the full text website http://www.smw.ch(direct link from each SMW record in PubMed) • No-nonsense submission -you submit a single copy of your manuscript by e-mail attachment • Peer review based on a broad spectrum of international academic referees • Assistance of our professional statistician for every article with statistical analyses • Fast peer review, by e-mail exchange with the referees • Prompt decisions based on weekly conferences of the Editorial Board • Prompt notification on the status of your manuscript by e-mail • Professional English copy editing • No page charges and attractive colour offprints at no extra cost International Advisory Committee Prof. K. E. Juhani Airaksinen, Turku, Finland Prof. Anthony Bayes de Luna, Barcelona, Spain Prof. Hubert E. Blum, Freiburg, Germany Prof. Walter E. Haefeli, Heidelberg, Germany Prof. Nino Kuenzli, Los Angeles, USA Prof. René Lutter, Amsterdam, The Netherlands Prof. Claude Martin, Marseille, France Prof. Josef Patsch, Innsbruck, Austria Prof. Luigi Tavazzi, Pavia, Italy We evaluate manuscripts of broad clinical interest from all specialities, including experimental medicine and clinical investigation.We look forward to receiving your paper!Guidelines for authors: http://www.smw.ch/set_authors.htmlAll manuscripts should be sent in electronic form, to: Official journal of the Swiss Society of Infectious disease the Swiss Society of Internal Medicine the Swiss Respiratory Society