A: Certainly there is reason to try a different approach, especially if you have clinical and/or empirical evidence from research that would justify an alternative or novel approach. A single negative study is not enough to dismiss an intervention. In a recent case, a randomized controlled trial failed to show positive results for Behavioral Family Systems Therapy (BFST) as a method of improving adherence to medical treatment in children with diabetes. However, in an efficacy study, Wysocki and colleagues found a significant positive effect on adherence and medical outcome (glycemic control) (Wysocki, et al., 2006).

A: In their decision to design it as an efficacy study, the research team considered the possible reasons for failure of the previous study. Was the hypothesis based on incomplete data? Was the model not a good fit for some families? Was it not a good fit for this particular illness? e.g., did the model fail to address specific diabetes adherence-related behaviors? They then decided to revise the model to better fit their population. Because there had been little evidence that interventions significantly effect adherence or glycemic control, and because they modified the model to better fit the population, the research team decided to conduct an efficacy trial rather than an effectiveness trial. (An efficacy trial is set up under 'ideal' - controlled - conditions; whereas an effectiveness trial looks at how well the treatment works in the real world, such as in a primary care setting). They designed a randomized controlled trial with 3 groups (standard care, educational support, and the revised BFST). Results of the study showed that the revised BFST significantly reduced family conflict, and improved both adherence and glycemic control.

The decision of whether or not to conduct an efficacy or effectiveness study can be influenced by a number of factors. If you are planning to make modifications to the intervention that has not shown an effect under controlled conditions, it may be best to start with an efficacy study. An efficacy trial will tell you if the treatment works in a highly controlled laboratory-like setting, controlling for as many variables as you can. For example, you might select only families with both parents living in the house, or parents with a college-level education. You might provide incentives, such as paying them, or offering flexible scheduling of appointments. Another reason for beginning in the lab is monetary. It generally takes more time, money, and research staff to conduct a study in a community setting. Once an intervention has been shown to be effective under highly controlled conditions, then you're ready to apply it in real world settings such as schools or primary care.

A: While an intervention may be effective in real world situations, there is still the question of clinical significance. The investigator should ask, how many children need to be treated before one child shows positive results? This is really the question of efficiency. If several hundred children are given the treatment, yet only a few improve, it is not a efficient method. Yet often this question is left unanswered. Establishing clinical significance is a 'final frontier' of research on interventions in the field of pediatric psychology, and if they have any chance of successful dissemination, behavioral interventions will need to be implemented as resourcefully as possible.