We collected data from January 2018 to November 2018 within a network of 20 dialysis clinics belonging to one private network (FMC-dialysis services Slovakia) in Slovakia (covering about 20% of the total Slovak dialysis population). This dialysis therapy is fully reimbursed by a compulsory health insurance system. Inclusion criteria were age over 18 years, a diagnosis of CKD-5 and undergoing haemodialysis treatment for at least 90 days. Exclusion criteria were an inability to fill in the questionnaire (due to dementia or mental retardation, inability to read the Slovak language) and acute severe intercurrent illness, obtained from medical records. All patients who met the eligibility criteria were asked to participate in the study.

Data were obtained by questionnaires filled in by patients during their routine visits to the dialysis clinic. Patients agreed to participate in the study by signing an informed consent prior to the study. They then filled in the questionnaires using tablets, with data recorded to an online platform with full confidentiality assured by means of a personal identification code. A research assistant was available for technical support.

We included 567 patients on maintenance haemodialysis (70.1% of those approached), 25 of whom were excluded due to not filling in the questionnaire related to HL (n = 9) or missing too many items in it (n = 16), leading to a final sample of 542 patients.

The study was approved by the Ethics Committee of the Faculty of Medicine of P.J. Safarik University (15N/2017) and the Ethics Committee of FMC-dialysis services (November 23, 2017). All data were collected in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Health-related quality of life was measured by the Kidney Disease Quality of Life – Short Form, version 1.3 (KDQoL-SF™, [10]). This is a self-report measure developed for CKD patients and those on dialysis and is widely used in research [24, 25]. It consists of 43 kidney disease targeted items within eight scales and three additional quality of life scales. These scales are: Symptom Problem Scale (SPS), Effects of Kidney Disease (EKD), Burden of Kidney Disease (BKD), Work Status Scale (WSS), Cognitive Function Scale (CFS), Quality of Social Interaction Scale (QSIS), Sexual Function Scale (SXFS), Sleep Scale (SS), Social Support Scale (SSS), Dialysis Staff Encouragement Scale (DSES) and Patient Satisfaction Item (PSI). The second part of the questionnaire is the 36-item health survey (SF-36) consisting of eight scales: Physical Functioning Scale (PFS), Role – Physical Scale (RPS), Pain Scale (PS), General Health Scale (GHS), Emotional Well-being Scale (EWS), Role – Emotional Scale (RES), Social Function Scale (SFS) and Energy Fatigue Scale (EFS). The scales score ranges from 0 to 100. A higher score reflects a better quality of life. The Cronbach’s Alpha in our sample ranges from 0.43 to 0.94. We did not include the Symptom Problem Scale for peritoneal dialysis, as our sample consists only of haemodialyzed patients. See Appendix A for more information on the measurement tool.

Health Literacy (HL) was measured using the Slovak version of the Health Literacy Questionnaire (HLQ, [26]), a multidimensional tool [27] consisting of nine domains of HL related to accessing, understanding and using information to make decisions about health (Cronbach’s Alpha in our sample ranges from 0.77 to 0.90). A higher mean score in a particular domain indicates better HL in that domain [27]. We categorized this measure using hierarchical cluster analysis [28] on standardized z-scores of all HL domain, leading to clusters of cases with similar HL characteristics. This method minimizes the within-cluster variance in a stepwise manner leading to clusters that are as different as possible. Three clusters were used for further analyses (low HL group, moderate HL group, high HL group), representing different levels of HL consistently across all domains in a particular cluster. The mean HLQ score for the nine domains of the three HL groups of patients are described in Table 1.

Sociodemographic data were measured using the questionnaire and included gender, education (lower education: elementary education and apprenticeship vs. higher education: high school and university), marital status (with partner vs. without a partner) and living conditions (living alone vs. with family member/s). We compared patients in productive age (≤50) with those in late productive and post-productive age (>50) as this might be associated with their social participation.

First, we assessed the sociodemographic characteristics of the sample and the three HL groups. Second, we assessed associations between the HL groups and HRQoL (continuous level) using one-way ANOVA and the Kruskal-Wallis test. The statistical significance of differences between the HL groups was tested using the post hoc Bonferroni tests in the case of ANOVA and Dunn’s tests with Bonferroni correction in the case of Kruskal-Wallis tests. Statistical analyses were performed using SPSS v. 23.0 for Windows [29].

Of the 542 patients (mean age 63.6 years, standard deviation = 14.12), most were men (61%) and most older than 50 years (82%); almost half of the patients had a lower education (49%) and were without a partner (42%), and 18% of patients lived alone (Table 2).

We found HRQoL to differ between the three HL groups in 15 out of the 19 HRQoL scales (Table 3). Post hoc comparisons (Figure 1) revealed that low HL patients had a worse HRQoL than moderate and high HL patients. Moreover, patients in the moderate HL group had a worse HRQoL than the high HL group for seven HRQoL scales. The greatest differences between the three HL groups regarded the scales EKD, CFS, QSIS, SSS, DSES, PSI, PFS, PS, EWS and SFS.

The major strengths of our study regard the representativeness of our sample, which covered ESRD patients undergoing haemodialysis in 20 dialysis clinics in Slovakia, and the relatively high response rate (70%). The use of a disease-specific HRQoL-related questionnaire (KDQoL – SF) as well as the generic core for QoL (SF-36) enabled us to bring detailed information on HRQoL profiles in a homogenous group of dialyzed patients.

Our study has some limitations as well. As the study had a cross-sectional design, we are unable to make causal inferences. Furthermore, the data are self-reported, which can result in some social desirability and thus in some information bias. Using self-report questionnaires may have led to some selection bias, with people with very low (health) literacy excluded, and thus to some underestimation of the real differences. Finally, some scales (WSS, QSIS and GHS) of this Slovak version showed a lower internal consistency than was found in the validation studies of the English original [10]. This may have added some measurement error, and thus an underestimation of the associations. This also requires further study on e.g., the impact of cultural factors and the results of our research should be interpreted with caution regarding the scales mentioned, as we used the best method to translate, i.e., forward-backward.

Our findings that patients with low HL show worse HRQoL than patients with higher HL suggest that it is important to support patients with limited HL to maintain their HRQoL. Such support could relate to their capacities to understand, appraise and adequately use relevant health information to secure proper adherence to treatment and good cooperation and communication with health care providers. Another way of helping may be to offer psychological support for better coping with the disease. In addition, the responsiveness of the health care system to the needs of low HL patients may be improved [34, 35]. This may include effective patient-centred care taking into account specific barriers and limitations connected with low HL [31].

In future research it will be important to study the mechanisms responsible for the association between health literacy and HRQoL. Insight into the role of potential mediators, such as lifestyle, medication adherence, perceived control or the quality of the communication by health care professionals, will support the improvement of interventions aimed at maintaining HRQoL in dialyzed patients.

We found that dialyzed patients differed in the HRQoL profile according to the level of their HL. Recognizing HL needs and limitations in dialyzed patients and tailoring care and health related communication towards those with low HL may help improve their HRQoL.