[Contribution of clinical epidemiology to evidence-based nephrology]

Improvement of modern medicine does not make day-to-day practice of nephrology easier. Decision-making remains difficult, often requiring numerical, probabilistic approach. Contribution of clinical epidemiology to evidence-based nephrology is not limited to randomised-controlled trial. The concept of natural history: spontaneous evolution without treatment, has been developed in epidemiology to describe diagnosis and prognosis of diseases. For example, in IgA nephropathy, it allowed to measure the prognostic significance of gross hematuria or to design classification of end-stage renal failure. Epidemiologic studies needs appropriate designs, representative samples and correct methods of analysis. Like other medical research, epidemiology has some important methodological limits that must be taking into account. Practising evidence-based nephrology improves confidence in management decisions.