Investigating User Engagement of Diabetes Applications and its Associations with Self-Management in Type 2 Diabetes

Abstract

Aim: The prevalence of Type 2 diabetes (T2D) is rising rapidly presenting a considerable burden to health economies worldwide. Diabetes smartphone apps represent a feasible, low-cost alternative to traditional face-to-face interventions, which can be time and resource intensive and often miss hard to reach populations. The aim of this study was to investigate user engagement of two popular, commercially available diabetes apps (Glucose Buddy and mySugr) over a two-week period and to examine whether user engagement was associated with self-management in T2D. Methods: 89 adults with T2D attending the Auckland diabetes centre were recruited and randomised into one of the two app groups. The primary outcome measure was adherence to self-care behaviours and secondary outcome measures included illness beliefs and user engagement. Mann-Whitney U tests and Wilcoxon-signed rank analyses were used to examine changes between and within groups. Spearman's correlations were performed to investigate relationships between outcome measures. A hierarchical linear regression was used to examine which factors predicted user engagement. Six participants were also interviewed to further explore their experience of using the app. Findings: There were no improvements in self-care behaviours or illness beliefs from baseline to follow-up. There was no clear evidence that mySugr was more engaging than Glucose Buddy, although there was a trend towards participants in the mySugr app group using the app on more days than participants in the Glucose Buddy app group. Blood glucose testing was significantly related to minutes of app use and several significant relationships between self-care behaviours and illness beliefs were also observed. Three illness beliefs (timeline beliefs, identity beliefs and type 2 diabetes understanding) were significant predictors of user engagement. The apps were considered acceptable, although difficulties with usability were identified. Conclusions: The quantitative findings and high attrition and non-adherence to the intervention suggest that the free versions of both the diabetes apps trialled in this study are not engaging to T2D patients. Results also indicate that more involvement of patients and healthcare professionals throughout the app development process is needed to ensure the apps provide value to users.