JMIR Publications

ABSTRACT

Background:

Most adolescents own smartphones and many of them are almost constantly online. Around one-tenth of adolescents are diagnosed with an emotional disorder and many more face significant emotional problems during adolescence. However, a small portion of adolescents uses health applications to improve mental health, highlighting a gap between high ownership, widespread usage and lack of interest in using mHealth applications for mental health improvement among adolescents with significant emotional problems. Attrition rates in mHealth interventions targeting populations of anxious and/or depressed adolescents are high, and attrition is commonly portrayed in two-point fashion. Research on mental mHealth interventions among adolescents have frequently lacked detailed time-related attrition data alongside accurate definitions and analysis of attrition reasons through usage data.

Objective:

The objective was to obtain daily attrition rates among adolescents with clinical anxiety and/or depression symptoms in an mHealth intervention while simultaneously obtaining knowledge of attrition reasons through evaluation of application usage data and frequency, time and amount of health behaviour in three categories; nutrition, physical activity, and mental health.

Methods:

A randomized controlled trial with 121 adolescents suffering from anxiety or depression, undertaken to measure outcomes over a 42-day period across control, treatment-as-usual, and intervention groups. Primary outcomes were attrition, time and frequency of in-app health tasks and anxiety and depression symptoms. Secondary outcome measures were self-efficacy levels and significant sleep problems. Statistical tests included t-tests, χ2, Pearson’s correlations, ANOVAs, Kaplan-Meier survival analysis, and Cox proportional hazard regression models to analyze attrition and health task impacts. Multiple regression models assessed predictors for changes in anxiety and depression symptoms.

Results:

Significant differences in attrition were evident between intervention group (37%) and TAU group (93.70%) x2(1)= 27.11, P.01. Days of usage were 6.31 (95% CI 2.90-9.73) in TAU group and 29.96 (95% CI 22.58-33.34) in intervention group. A significant decrease (13.469, SD 15.486) in completion of health exercises from 1st to 2nd week was evident in TAU group (t31=4.920, P.01) while weekly usage rose by 7.954% from the first week to the second (P=.590). A significant increase in health exercises was obtained in the last week of the trial compared to the 5th week (M=35.261, SD=72.013) in the intervention group (t45=2.344, P=.05), not the TAU group. Days of in-app usage significantly predicted differences in pre- and post-measures on RCADS total score (β=-.252, P.01), anxiety symptoms (β = -.334, P.05) as well as depressive symptoms (β = -.307, P.05).

Conclusions:

This mHealth RCT demonstrated differences in attrition and usage between groups of adolescents with clinical anxiety and depressive symptoms. Indications towards sensitivity periods in completion of health tasks were obtained that warrant further research. Focus on time-specific attrition alongside daily times and type of in-app health behaviour seem a promising avenue for research on mHealth interventions for adolescent populations with anxiety and depression with the aim to ultimately lower attrition rates that continue to be too high. Clinical Trial: ClinicalTrials.gov NCT05912439; https://clinicaltrials.gov/study/NCT05912439