Why Behavioral Theories Need an Update: An Interview with Susan Michie

Lorien C. Abroms, ScD, Theories and Techniques of Behavior Change Interventions SIG junior co-chair

The Society of Behavioral Medicine’s (SBM’s) Theories and Techniques of Behavior Change Interventions Special Interest Group (TTBCI SIG) recently interviewed SBM Fellow Susan Michie, DPhil, CPsych, FBPS, about the past, present, and future of behavioral theories.

TTBCI SIG: Many of our behavioral theories—such as social cognitive theory—were developed more than 40 years ago. What do you see as the limitations of widely used behavioral theories?

Michie: “Many of our theories were developed with incomplete information—for example, with cross-sectional samples and focused on predicting differences between groups of people rather than on explaining behavior and behavior change within individuals. They were also developed without objective measurement of behavior, relying almost exclusively on self-report. In some cases, these theories were developed on the basis of proxies for behavior such as ‘behavioral intention’ rather than behavior itself. We now have much better objective measurement of people’s everyday environments, through smartphone apps, wearables, and environmental and physiological sensors.

“Additionally, most past theories were partial in that they did not include the range of aspects of capability, opportunity, and motivation that are needed to explain the range of behaviors and behavior change. They were also poorly specified both in terms of the constructs and the relations between them, which makes testing their propositions and systematically accumulating evidence about theory very difficult. This became evident in a systematic literature review we conducted which identified 83 theories of behavior change. This was reported in an article in Health Psychology in 2015 and more fully in the book, The ABCs of Behavior Change. Only three of these theories set out to be comprehensive, and there were many overlapping constructs—we identified more than 1,700 across the theories. For these reasons, we are in need of better behavioral theories that are ecologically valid, and we now have the technology to begin to make rapid advances.”

TTBCI SIG: What will new theories offer us?

Michie: “We should start by understanding our current theories’ strengths and limitations, then build on their strengths, especially those that are dynamic and account for change, such as Operant Learning Theory and Control Theory. One of my current projects has specified all 83 theories in terms of specific definitions of constructs and relationships between them—we have identified 14 types of relationships—and we have checked these with the theory authors. These have been specified in databases that allow them to be programmed computationally. We are working with computer scientists to write a program to identify ‘iconic’ theories representing the strongest aspects of the 83 theories. This should make it easier for people to work with behavioral theory. Currently many people do not know which to choose or what criteria to use to select a theory to work with.

“We also have exciting new possibilities of advancing theory with better specification and measurement of behavior in relation to people’s thoughts and feelings, their physiological response, and their environments. However, the analysis of the huge quantities of ecologically valid streams of data provided by sensors, wearables, and smartphone apps is considerably challenging.”

TTBCI SIG: What theory-based projects are you working on?

Michie: “We are just starting a project that is a very exciting collaboration between behavioral, informational, and computer scientists, called the Human Behavior Change Project. This aims to revolutionize methods for synthesizing evidence about behavior change in real time, generating new insights about behavior change. It will create an up-to-date knowledge base to help design effective interventions tailored to particular populations, settings, and behaviors. It will do this by building on the behavior change technique taxonomy. We will create an ‘ontology’ to organize knowledge from published research reports, and will use natural language processing to automatize the process of extracting information from the reports into the ontology. We will use machine learning to identify patterns and make inferences beyond what we already know. Finally, we will create a user interface so that researchers, policymakers, and practitioners can get answers to variants of the question, ‘What works for whom in what settings for what populations to change what behaviors, and how?’ This is a hugely ambitious project to pursue over the next 4 years; its success will very much depend on the engagement of the international scientific community.”

TTBCI SIG: What top three readings would you recommend to someone interested in behavior change theory?

Michie: “Alexander Rothman’s ‘Is there Nothing More Practical than a Good Theory? Why Innovations and Advances in Health Behavior Change will Arise if Interventions are Used to Test and Refine Theory.’ Also The ABCs of Behavior Change and a forthcoming American Journal of Preventive Medicine article from Heckler, Michie, et. al on the methodology of designing with theory.

TTBCI SIG: What are some training opportunities for those interested in advance the field of behavior change theory?

Michie: “I am director of University College London’s Centre for Behaviour Change, which brings together academic expertise across disciplines and translates that to policymakers, practitioners, researchers, and all those interested in behavior change. Training is part of what we offer. Our week-long summer schools are always oversubscribed, so sign up early if you want to come next year. We are also starting a multidisciplinary master’s course in behavior change, which we think is the first of its kind in the world that will bring together perspectives across disciplines. If others exist, we are very happy to hear about them!”