Dear Editor:

In a recent letter titled “A discussion on quality”1 the author states that “in order to demonstrate the virtues of the quality indicators, some of the articles use very weak baseline data”.2,3 We feel that this hypothesis could easily be refuted with objective data. We will compare variables from the clinical results of the observational study titled Dialysis Outcomes and Practice Pattern Study (DOPPS),4 which included 575 patients from 20 different centres in Spain, with the baseline results of our study (313 patients from four centres)2:mean haemoglobin 10.8 vs. 11.7 ± 1.4g/dl, phosphorus 5.5 vs 5.3 ± 1.6mg/dl, Kt/Vsp 1.31 vs. 1.37 ± 0.29, ferritin 288 vs. 370 ± 290mg/ml and percentage of autologous arteriovenous fistulas 81 vs. 79.9 (DOPPS vs. our own study)2 (the standard deviation for the DOPPS study is not mentioned because it does not appear in the publication). After seeing the results from both studies, we can state that variables from the clinical results of the DOPPS study could be considered worse than, or at best similar to, those presented by the patients in our study. The conclusion that we reach is not different when we analyse the European population (excluding Spain), which is also represented in the DOPPS study. The comparison with the study carried out by Plantinga et al. is more complex due to the form in which the results are expressed, but in general, although these results are worse than the Spanish and European results, they are similar to those from the rest of the population of the United States. Comparisons of variables from clinical results in centres should be carried out with representative samples from the general population, and not with samples representing select centres. The author does not mention what studies the cited studies are compared with. As Fink et al. describe, the variability of results from centre to centre is welldemonstrated (they call this phenomenon the “centre effect”).5 We heartily agree with the other statements expressed in the letter. Meanwhile, we confirm the limitations of our study (which were not mentioned by the writer of the letter) which were listed in the original publication.