Insights From Longitudinal Evaluation of Moodie Mental Health App

Abstract

Recently, mental health applications have gained increasing attention. This is because research shows that users find mental health apps as a good alternative for self-management of mental conditions, especially in the last two years when access to physicians was limited because of the pandemic. Despite the existence of several mobile apps for mental health, most available apps are not supported by empirical and scientific evidence, or they are designed based on experimental rather than real data. This work represents our first step toward designing evidence-based AI-driven mental health applications. In this paper, we show our initial results and discussion of in-the-wild users' interactions and engagement with a mental health app (called Moodie) for a period of two years. Specifically, we investigate the interactions and engagements of 434 users (who used the app for two years) with the mood tracking feature where the app allows users to enter their moods and the corresponding factors associated with the mood. Chi-square analysis showed a significant correlation between moods and factors (or user's activities). The results also show that Home, Work, Relaxation, and Family-related activities are the most common factors that affect moods either positively or negatively and that these factors have a different impact on moods during different times of the year.