Development of an Ontology of Engagement with Behaviour Change Interventions

Background Participants’ engagement with behaviour change interventions is crucial for their effectiveness. However, engagement is conceptualised and measured inconsistently across research domains, limiting the ability to compare and synthesise evidence about engagement and identify strategies to enhance engagement. This study aimed to develop an ontology—a classification framework—to precisely specify and define aspects of engagement with behaviour change interventions. Methods The Intervention Engagement Ontology was developed in seven steps: (1) specifying the ontology’s scope, (2) reviewing intervention reports to identify key classes (categories) of engagement, (3) refining the ontology through literature annotations, (4) a stakeholder review on the ontology’s clarity and comprehensiveness, (5) testing inter-rater reliability in applying the ontology for annotations, (6) specifying relationships between classes, and (7) making the ontology machine-readable. Results Participant engagement with interventions was defined as “An individual human activity of an intervention participant within one or more parts of the intervention.” Through Steps 1–4, an initial ontology with 48 classes was developed, including 37 engagement-specific and 11 structurally supporting classes (e.g., emotional process). Inter-rater reliability for applying these engagement classes was ‘acceptable’ for researchers familiar (α = 0.71) and unfamiliar (α = 0.78) with the ontology. After further refinements (Steps 6-7), the published ontology included 54 classes - 44 engagement-specific and 10 supporting classes. The engagement classes were structured around three key engagement types: (1) behavioural, (2) emotional, and (3) cognitive. Behavioural engagement aspects, such as frequency and duration, were also represented in the ontology. Discussion The Intervention Engagement Ontology provides a structured framework for specifying and defining participant engagement with behaviour change interventions, facilitating clearer communication, comparison and evidence synthesis across research studies and domains. Future work will refine the ontology based on further feedback and empirical validation, enhancing its applicability.