Using effectiveness studies for prescribing research, part 1.

The process of evaluating pharmaceuticals has become highly conceptualized in contrast to the lack of formal rules for assessing effects of interventions on practice. We argue that clinical audit is a key factor prior to instigating an intervention and that randomized controlled evaluations are preferable. We discuss the need for small-scale experiments prior to full trials to validate the underlying concept of an intervention with the recognition that different approaches may be necessary. This includes open rather than blind assessments and greater emphasis on qualitative issues during development of interventions followed by quantitative appraisal of their impact.