Mapping User Trajectories to Examine Behavior and Outcomes in Digital Health Intervention Data

Digital health interventions pose a tremendous opportunity to increase the possibilities of treatment delivery, and with the innovations we continue to see in technologies, they are becoming increasingly pervasive. However, there are also challenges. We present FlowVis, an application which facilitates visual exploration of participants' experiences over the course of a digital health intervention. This application is based on the concept of participant trajectories, defined as paths that an individual might traverse during an intervention, characterized by key characteristics such as outcomes, usage, and qualitative data. FlowVis supports identification of subgroups of participants by similarities in their trajectories over time. We present a case study of the use of FlowVis in three parts. First, we perform exploratory analyses of the utility of different clustering methods for exploring digital health intervention data. Second, we present a flow-based visualization examining changes in participants' outcomes over the course of the intervention. Third, we compare subgroups of participants by their app usage behaviors. Though we illustrate usage of the application to explore data from one digital health intervention, the application can be applied to analyze other intervention datasets and potentially other types of performance data. This paper contributes to extant research in the visualization of digital health intervention data by proposing a visual method to examine similarities in evolutionary patterns of participants' experiences, behaviors, and outcomes over time.