What types of evidence are most needed to advance behavioral medicine?

Background This editorial presents a perspective on the types of evidence most needed to advance behavioral medicine given the current status of the field. Purpose The paper argues that the types of evidence most needed at present are evidence that is contextual, practical, and robust. Methods Each of the above issues is discussed with attention to characteristics of interventions; representativeness at the multiple levels of setting, clinical staff, and participants; and research design and measures. Arguments are made from philosophy of science, status of the literature, and future directions perspectives. Results The current dominant paradigm of reductionistic studies focused predominantly on internal validity using highly homogenous patients and academic settings is not and will not produce the desired translation to real-world practice and policy. Instead, broader "practical" clinical and behavioral trials are needed that address the influence of the context in which programs are conducted, that include outcomes important to decision makers and communities, and that focus on moderating, mediating, and economic issues. Conclusions To create programs that will be disseminable, a greater focus is needed on external validity and transparency of reporting. We need to realize that the world is complex and embrace and study this complexity to produce further progress. Such an approach can produce evidence that is both rigorous and relevant.