A new paradigm for behavior change

Abstract

Drs. Bartholomew and Mullen present an articulate and weighty discourse describing strategies to advance behavior change research, especially health disparity research. For too long we have limited ourselves to conducting surveys, participating in “intervention” testing, and summarizing our findings for colleagues. Rarely have we reported breakthroughs that could inform the next level of research, often because the determinants of the behavior change were not measured or manipulated. As Drs. Bartholomew and Mullen point out, we have often conflated the theoretical method with the mechanism underlying the behavior, and in the end reported neither. Thus, we are left with three tasks if we are to accept their message and move forward. We must provide guidance in selecting theoretical models on which to design our research, dissect/refine the determinants of the problem and causal model, and implement a systematic approach that allows manipulating, testing, and reporting results in a much finer detail than before. One strategy advocated by the authors is the use of logic models, which force careful consideration of the health problem, determinants, and the putative causality from both individual and environmental perspectives. The authors are careful to point out the recursive nature of this process, and therein lay the challenge. How can we more effectively build on each others' work when separation often exists between the theoretical framing, the variables tested, and the analysis plans pursued so that we are reduced to a cursory description of a multifaceted approach? One way is to convene forums in which we have ongoing dialogue about the pathways through which behavior change occurs. It is no longer sufficient merely to publish thoughtfully, but we must consider the next generation of studies that can be built upon our work. For instance, the reader may have noted that the matrices and models presented by Bartholomew and Mullen do not address the growing segment of society who can only access episodic or problem-driven care (1). Thus, working within a routine care system will only partially address the problem of late-stage diagnosis and poor survival from oral and pharyngeal cancer. Wade and colleagues provide evidence that the elderly, low socio-economic and the at-risk are not receiving “routine dental” care and are frequently seen for episodic care only. Thus, our logic model must be built on a broad understanding of the health problem including those of invisible populations. We need to test interventions that focus on cancer screenings for routine, as well as emergency, dental care patients. Moreover, interventions are needed that elucidate the pathways (2) through which behavior is changed (3) that will lead to reducing late-stage diagnosis among individuals outside the routine care system. Otherwise, we will only increase the disparity between those within and outside the routine care system. Finally, if we are to succeed in our forums, we will need to bring together colleagues who work with behavioral theories in laboratories, are skilled at community engagement, and are behavior-intervention experts. Big science and big efforts addressing big problems require new paradigms. The author has received an NIDCR grant.