Evidence-based medicine in nephrology: identifying and critically appraising the literature

the management of ESRD will facilitate this form of research in the future. Background. Uncertainties about best management of end-stage renal disease ( ESRD) are reflected in wide Keywords: end-stage renal disease; randomized variations in practice. Systematic reviews aim to reduce controlled trials; systematic reviews uncertainty by strengthening the evidence base for clinical practice, allowing estimation of the benefits and risks of particular interventions, whilst minimizing the potential for bias. This paper describes the methods Introduction and conduct of six systematic reviews of aspects of the management of ESRD, and the yield in terms of There are marked variations in the management of trials found. patients with end-stage renal disease ( ESRD). For Methods. Our methodology was based on that recom- example, the proportion of patients treated by haemomended by the Cochrane Collaboration (an interna- dialysis or peritoneal dialysis varies widely between tional initiative set up to perform and disseminate centres in Europe, as does the use of specific componsystematic reviews of health care). It involved a system- ents of treatment such as type of haemodialysis mematic search of electronic databases and bibliographic brane (synthetic or cellulose), and type of peritoneal reference lists, together with handsearching of Kidney dialysis delivery/transfer set ( Y-set or non-Y-set). Such International for studies relevant to the management variations often reflect uncertainty about the true of ESRD, followed by a systematic assessment of benefits and risks of one treatment choice over another. study quality. Increasingly, both providers and consumers of health Results. Around 12 000 abstracts were assessed which care are demanding information on the most effective had been identified from electronic sources. Of these, and efficient therapies which can then be used to 2085 (18%) were deemed to be reports of possible develop standards and guidelines for high-quality clinrandomized or quasi-randomized controlled trials rel- ical practice. evant to the management of ESRD. Three hundred The randomized controlled trial (RCT ) is now and forty were relevant to the six specific reviews, and widely accepted as the ‘gold standard’ method for after assessment of the full manuscripts, 39 studies assessing the relative effectiveness of alternative treatwere finally included in our reviews. Reports of a ments [1,2]. This approach to evaluation ensures that further nine trials, which were identified from other selection and other common sources of bias are minimsources, were also included. The broad search adopted ized. The interpretation of the results of individual allowed the parallel development of a register of trials trials may, however, be difficult, as few are of sufficient of all aspects of the management of ESRD. size to rule out chance as an explanation for observed Conclusions. This study has demonstrated that the differences between treatments. In addition, entry crimethodology of systematic reviews, as promoted by teria for inclusion of patients into individual trials may the Cochrane Renal Group, is feasible but has signific- be too restrictive to allow confident extrapolation of ant resource implications. The development of a regis- the results to a particular patient. Considering all ter of randomized controlled trials (RCTs) related to relevant trials together should redress the limitations