Whether healthcare interventions are conducted in an in-person setting or via an app, developing a deeper understanding of what works for whom is imperative. What makes some interventions successful and others sub-optimal? The science of behavioral medicine was founded in the 1940s. Since then, many theories and frameworks have been developed and have come to prominence in the 1980s. More recently, behavior change techniques, like goals setting and self-monitoring, have been linked to health outcomes.
In some cases, when we don’t have a robust evidence-base to work from, getting a group of deep experts together to map a future state is a prudent step. A recent publication in the Annals of Behavioral Medicine by Lauren Connell and colleagues from the University of London set out to map behavior change techniques against mechanisms of action in interventions to further clarify what influences behavior. An example of a mechanism of action includes knowledge, intentions, needs, attitudes, and beliefs.
The expert consensus brought one hundred and five behavior change experts together to rate the top 61 most used behavior change techniques against 26 mechanisms of action. The goal was to develop a shared understanding in the field of behavior change on where the strongest overlap for influence lives between technique and mechanism. In the review process, the goal was to reach 80% agreement between technique and mechanism, the panel reviewed 1,586 potential links between behavior change techniques and mechanisms of action (61 x 26). The panel then developed heat maps depicting the strength of the degree of overlap between the two. I have included the heat mapping for the techniques and mechanisms that panelists felt linked had the strongest relationship. The greyer the bar, the greater the relationship in terms of the degree of connection.
Source: https://www.ncbi.nlm.nih.gov/pubmed/30452535. A heat map indicating the proportion of experts rating a behavior change technique (BCT) was “definitely” linked to a mechanism of action (MoA). Values range from 0 to 1, with values closer to 1 shaded in the darkest grey. A 1 indicates 100% of experts agreed on a BCT that was definitely linked to an MoA. M.A.D.P. = Memory, Attention, and Decision Processes; P.S.V. = Perceived Susceptibility and Vulnerability; S.P.R.I = Social/Professional Role and Identity; B. Con. = Beliefs about Consequences; G.A.B. = General Attitudes and Beliefs; A.T.B = Attitude towards the Behavior; B.R. = Behavioral Regulation; B.Cap. = Beliefs about Capabilities.
The value of this process is manifold:
1 It allows for more interventions to build off the evidence by linking techniques and mechanisms,
2 It allows for continued hypothesis testing in areas where consensus was trending toward 80% but didn’t quite meet that bar, and
3 The usual path is examining the evidence base via theories and frameworks to develop interventions to test. This work takes a lot of the guesswork out of what might work in a real-world setting and brings us closer to building and testing interventions outside the halls of academia.
In essence, it allows us to start at a deeper level of understanding and sophistication of the science base. One challenge in the digital health space is a lack of depth and breadth in product development. This work allows for better translation into a more robust user experience. Of course, any expert consensus is only as good as its experts and the methodologies used. Given the caliber of people involved, this work is a significant contribution to the field and so crucial to those of us who work in real-world settings. I hope that more digital health companies will take this science on board as part of their product roadmap development journey.
Thanks for reading – Trina
(Opinions are my own)
References
Links between behavior change techniques and mechanisms of action: An expert consensus study.
