Means and standard deviations were calculated for continuous variables, and frequency counts and percentages for categorical variables. Descriptive analyses were performed for the total survey population and stratified by Hb level, geographical region, and CKD stage. For the generic EQ-5D-3 L, Hb levels of < 8 g/dL, 8- < 10 g/dL, 10–12 g/dL, and > 12 g/dL were used; for the KDQOL-36 and WPAI, Hb levels of < 10 g/dL, 10–12 g/dL, and > 12 g/dL were used. CKD stages were 3a non-dialysis dependent (NDD), 3b NDD, 4 NDD, 5 NDD, and dialysis-dependent (DD).
The non-parametric Spearman’s rank correlation test was used to assess the correlation of Hb level with EQ-5D-3 L utility index and domains, EQ-5D-3 L VAS, SF-12 PCS, SF-12 MCS scores and the three subscales from the KDQOL-36. To adjust for potential confounding, linear regression analyses were performed on EQ-5D-3 L VAS, SF-12 PCS, SF-12 MCS and the three subscales from the KDQOL-36 as the outcome variables; independent variables included were Hb level (continuous), CKD stage, Hb and CKD stage interacted, sex, common comorbidities (diabetes, heart failure, stroke) and cardiovascular risk. Logistic regression analysis was performed with the same independent variables and EQ-5D-3 L utility index score (classified as ≥0.8 and < 0.8) as the outcome variable. Exploratory analyses compared KDQOL-36 and WPAI scores for patients with Hb < 8 g/dL with those for patients with higher Hb levels.
Patients who had completed a PSC, and for whom current CKD stage and Hb level were available were included in the analysis. Patients with no Hb level reported were included in the descriptive analysis of demographics and disease characteristics, but excluded from all other analyses. Missing data were not imputed but remained missing; therefore, the base of patients for analysis could vary between variables, and is reported for each analysis. All descriptive and exploratory analyses were conducted using IBM SPSS Data Collection Survey Reporter version 6 or later and all statistical testing was conducted in Stata v15.1 [37].
Do you have any questions about this protocol?
Post your question to gather feedback from the community. We will also invite the authors of this article to respond.