The Control Group Dilemma in Clinical Research

Abstract

Clinical research on psychosocial and behavioral medicine interventions has burgeoned in the past two decades, so much so that sole reliance on standard no-treatment control conditions may no longer be appropriate or feasible. We discuss the ethical, theoretical, scientific, and statistical considerations that shape current clinical outcomes research for psychosocial and behavioral medicine interventions. Secondary analysis of a psychosocial randomized trial (N = 127) illustrates some of these points. A new design for randomized clinical trials is described that does not require a no-treatment control group, and that reveals dose-response relationships between interventions and treatment outcomes.