Realism in evidence based medicine: interpreting the randomised controlled trial

Abstract

Realism is emerging as a paradigm for research and explanation in the natural and social sciences. A realist framework is elaborated and applied to the four possible situations that may generate the observations of randomised, controlled trials. It is demonstrated that by using two realist concepts "mechanism" and "context" a number of misinterpretations of such trials from within the dominant empiricist paradigm may be rectified. Evidence based medicine should adopt realism to temper a misleading empiricism, this will involve relegating statistical arguments to their proper subsidiary place and adopting an adequate theory of causation.