Contributions as of July 1, 2006

Towards Nephro-QUEST

QUEST is a multidimensional initiative conceived to serve the scope of increasing the qual­ity of renal care in Europe. We are striving to focus the attention of renal physicians on the importance of outcome research and epidemiological research in general. The new endeavour of the Registry is to organise two epidemiology courses/year in various European countries. This should be proof of the determination of the ERA-EDTA Registry to create valuable occasions for European nephrologists to be introduced into the fascinating field of epidemiology. For the whole initiative, a fundamental need is the transition of registries towards high quality clinical databases. This is made difficult by the fact that the use of paper collection forms is still diffuse.

Read more

Observational studies have dominated clinical nephrology for many years like they dominate the literature in general. Even if observational studies are extremely valuable, their main disadvan­tage is that they may, more than trials, suffer from confounding. The recognition of the rapidly growing number of people with chronic kidney disease (up to 16 % of the US population) leads to the need of evaluating treatment strategies for such disease. In order to over­come the drawbacks of observational studies, during recent years in­vestigators networked successfully (i.e. Collaborative Study Group, GISEN, EUVAS) and produced a number of early randomized con­trolled trials (RCTs).

Read more

An accurate prediction of survival is extremely useful when helping patients to understand the likely benefits of RRT, but in the mid nineteen nineties published prediction models showed low predictive accuracy. By starting this study we have attempted to over come limitations in previous studies by comparing the results of conventional and novel statistical modelling techniques. All participating renal centres in the ERA EDTA region were invited to join this registry project. The lack of a reliable historic “European” database for this effort was countered by using a mini questionnaire to identify a cohort of patients who died in the months of February 1997 or 1998 and another who started RRT during the same periods.

Read more

Confounding Bias

In many pa­pers that de­scribe the anal­ysis of patient outcomes the authors make use of statis­tical models which ‘adjust’ for confound­ing variables. In this note we will explain in short what con­founding is and how we may address it. Confounding can be described as a “mix­ing of effects”. This may occur if we study the relationship between an exposure E and a disease D (Figure 1) which might be confounded by a factor C.

Read more