Designs that are used, or should be used, for dissemination and implementation research
C. Hendricks Brown, Ph.D.
Northwestern University
ABSTRACT:
The fields of medicine and public health have made great progress in determining whether an intervention is efficacious or effective by conducting carefully crafted randomized clinical trials. In contrast to these designs to evaluate an intervention’s efficacy or effectiveness, the designs for dissemination and implementation (D&I) research are not yet well established, a factor that has no doubt has impeded developing our knowledge of effective D&I. By its very nature D&I research is intimately connected to understanding how programs, practices, or policies work in different contexts, so there is more attention in D&I research on external validity, as contrasted to the heavy emphasis on internal validity that many of the randomized efficacy and some effectiveness trials address.
Results: This presentation is a product of a workgroup meeting of 10 scientists and NIH staff, convened by NIH to facilitate D&I research. This committee addressed differences in terminology and provided a summary of the designs that have been used in D&I research, including both randomized and non-randomized studies. We identified 27 designs and found it useful to categorize these designs into several broad categories. One category of designs involves what can be termed the “traditional translational pipeline” of interventions that move step by step from efficacy, to effectiveness, to implementation research. A second major class of designs involves “hybrid designs,” which combine elements of effectiveness and implementation research in one single design. Thirdly, we describe designs that are focused on quality improvement as the primary goal, in contrast to producing generalizable knowledge. Several of these latter designs borrow from diverse areas of engineering.
Discussion: We provide illustrations of these alternative designs and discuss cross-cutting issues, including community engagement and ethics in conducting implementation research.
