This paper describes a randomised controlled trial (RCT) investigation of the added value of systemic family therapy (SFT) over individually focused cognitive behavioural therapy (CBT) for families in which one or more members has suffered trauma and been referred to a community-based psychotherapy centre. The results illustrate how an apparently robust design can be confounded by high attrition rates, low average number of therapeutic sessions and poor protocol adherence. The paper highlights a number of general and specific lessons regarding the resources and processes involved that can act as a model for those planning to undertake studies of this type and scope. A key message is that the challenges of conducting RCTs in ‘real world’ settings should not be underestimated. The wider implications in relation to the place of RCTs within the creation of the evidence base for complex psycho-social interventions is discussed and the current movement towards a phased mixed-methods approach, including the appropriate use of RCTs, which some might argue is a return to the original vision of evidence-based practice (EBP), is affirmed.