A Design Framework for Microintervention Software Technology in Digital Health: Critical Interpretive Synthesis (Preprint)

<sec> <title>BACKGROUND</title> Although many digital health interventions have been studied in human-computer interaction, long-term engagement remains a challenge. Microinterventions that leverage small but meaningful steps toward health outcomes have been proposed as a method for providing flexible and engaging, yet short and focused, interventions. One way to mitigate the potentially limiting scope of each microintervention is through narratives combining multiple different microinterventions over time to facilitate greater effects and engagement. As an emerging field of research, significant focus has thus far been placed on the creation and testing of individual microinterventions with little consideration for the consolidated experience, that is, the narrative over time. A more concise terminology and framework could make it easier to design, evaluate, and compare systems for microinterventions, facilitate collaboration in the field, and foster communication across required disciplines. </sec> <sec> <title>OBJECTIVE</title> This paper aims to provide a framework for the creation of microintervention systems and the narratives served by these through their microinterventions. </sec> <sec> <title>METHODS</title> To synthesize a framework, we carried out a critical interpretive synthesis involving reciprocal translational analysis, refutational synthesis, and lines-of-argument synthesis, based on a systematic review of microintervention literature following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines across the Association for Computing Machinery Digital Library, PubMed, Scopus, and Web of Science. To assess the resulting framework, we applied it retroactively back onto the literature. </sec> <sec> <title>RESULTS</title> Based on a review including 35 studies, of which 80% (n=28) were published after 2020, we present and rationalize the Design for Microintervention Software Technology (D-MIST) framework, which can be used to better classify, design, and evaluate microinterventions. D-MIST defines a microintervention system on three levels: (1) the narrative, (2) the microintervention, and (3) the event. The narrative is the consolidated experience of engaging with multiple microinterventions over time, supporting a narrative goal, for example, a clinical goal. A microintervention is a highly focused intervention consisting of self-contained events aiming to facilitate small but meaningful steps toward the narrative goal. The event attempts a positive momentary change through a resource, for example, an exercise. Events may, for example, be designed for single use, present variations to facilitate engagement, or present order-dependent sequential resources. We additionally discuss supporting constructs, such as decision rules, conceptual models, and the system’s interaction model, that is, how the system interacts with the user through its design and other actors to create a consolidated narrative. </sec> <sec> <title>CONCLUSIONS</title> The resulting D-MIST framework, based on current state-of-the-art literature, describes the components of microintervention systems and how these contribute to a consolidated experience. Based hereon, we suggest opportunities for personalization of microintervention systems, discuss design guidelines, and the roles of actors in creating health narratives aiming to obtain better distal health outcomes. </sec>