User engagement and utilisation of an app for anxiety management: analysis and evaluation of usage patterns (Preprint)

Abstract

Background: SAM is a mobile app providing self-help for anxiety management. Launched in 2013, the app has achieved over a 1 million downloads and favourable ratings on the platform app stores. Key features of the app are anxiety monitoring, self-help techniques and social support via an online forum ("the social cloud"). This paper presents unique insights into e-mental health app usage patterns and explores user behaviour and usage of self-help techniques. Objective: To investigate behavioural engagement and to establish discernible usage patterns of the app linked to the features of: anxiety monitoring, rating of self-help techniques and social participation. Methods: We use data mining techniques on aggregate data from registered users of the app's cloud services Results: Engagement in general conforms to common online participation patterns, with an inverted pyramid or "funnel" of engagement of increasing intensity. We further identify four distinct groups of behavioural engagement, differentiated by levels of activity in anxiety monitoring and social feature usage. Anxiety levels among all monitoring users show a marked reduction in the first few days of usage, with some "bounce back" effect thereafter. A small group of users with demonstrable long-term anxiety reduction (using a robust measure) typically monitored for 12-110 days with 10-30 discrete updates and showed low levels of social participation. Conclusions:The data supports our expectation of different usage patterns given flexible user journeys and varying commitment in an unstructured mobile usage setting. We nevertheless show an aggregate trend of reduction in self-reported anxiety across all minimally engaged users. We find several commonalities between these patterns and traditional therapy engagement. Keywords: IntroductionAnxiety is one of the most common mental health problems: in 2013 there were 8.2 million cases reported in the UK. The cost of anxiety -treatment, healthcare and indirect costs such as loss of employment and productivity -was estimated at €11.6 billion [1]. In July of that year we launched SAM, a free app to provide self-help for anxiety management. While it has been successful in the terms of commercial app culture -downloads, ranking, ratings and reviews -we wished to examine how it is being used as a precursor to assessing its therapeutic impact in the mHealth domain. This paper reports on the analysis of user data and its therapeutic implications from the first three years of SAM's availability to a global population of users. Developing SAMThe development of SAM was driven by a desire to produce a generic, flexible tool for anxiety self-help that provided ease of access and embodied high standards of usability. Thus our therapeutic design assumed that:1. Anxiety does not need to meet diagnosable levels of severity to justify intervention by mHealth devices. 2. People will engage with digital mental health devices in accordance with their unique needs and preferences. 3. Aspects of face-to-face psychotherapy ca...