Characterizing Emotional State Transitions During Prolonged Use of a Mindfulness and Meditation App: An Observational Study (Preprint)
BACKGROUND The increasing demand for mental health care, shortages in mental healthcare providers, and unequal access to health care generally has created a need for innovative approaches to mental health care. Digital device apps – including ‘digital therapeutics’ – that provide recommendations and feedback for dealing with stress, depression, and other mental health issues, can be used to adjust mood and show promise for helping meet this demand. In addition, the recommendations delivered through such apps can also be tailored to an individual’s needs (i.e., personalized) and thereby potentially provide greater benefits than traditional ‘one-size-fits-all’ recommendations. OBJECTIVE We sought to characterize individual transitions from one emotional state to another during the prolonged use of a digital app designed to provide a user with guided meditations based on their initial, potentially negative, emotional state. Understanding the factors that mediate such transitions can lead to improved recommendations for specific mindfulness and meditation interventions or activities (MMAs) provided in a mental health app. METHODS We analyzed data collected during the use of the Stop, Breathe and Think (SBT) mindfulness app. The SBT app prompts users to input their emotional state prior to, and immediately after, engaging with MMAs recommended by the app. Data were collected on more than 650,000 SBT users engaging in nearly 5 million MMAs. We limited the scope of our analysis to users with 10 or more MMA sessions that included at least 6 basal emotional state evaluations. Using clustering techniques, we grouped emotions recorded by individual users and then applied longitudinal mixed effect models to assess the effects that individual recommended MMAs had on transitions from one group of emotions to another. RESULTS We found that basal emotional states have a strong influence on transitions from one emotional state to another after MMA engagements. We also found that different MMAs impact these transitions, and many were effective in eliciting a healthy transition but only under certain conditions. In addition, we also observed gender and age effects on these transitions. CONCLUSIONS We find that the initial emotional state of an SBT app user has an impact on which SBT MMAs will have a favorable effect on their transition from one emotional state to another. Our results have implications for the design and use of guided mental health recommendations for digital device apps. CLINICALTRIAL