Context: Many of the interventions that nurses develop and implement are in themselves complex and have to operate in situations of irreducible complexity and uncertainty.
Main Argument: This article argues that the primary means of generating knowledge for the evidence-based deployment of complex interventions should be the pragmatic randomised controlled trial. Randomised controlled trials represent the only research design to adequately deal with that which we know and (far more importantly) that which we do not.
Literary Method: Using the example of practice development as an exemplar for complexity, and drawing on the objections often voiced as a response to calls to make use of randomised controlled trials in nursing and nursing research, the article presents a developmental framework and some methodological solutions to problems often encountered.
Conclusion: Randomised controlled trials, whilst undoubtedly methodologically and strategically challenging, offer the most robust basis for developing primary research knowledge on the effects of complex interventions in nursing and their active components.