JMIR Publications

ABSTRACT

Background:

Theories that propose that brief mobile fully-automated mindfulness ecological interventions (MEMIs) could efficaciously enhance emotion regulation (ER) and self-compassion domains and that such changes might be mechanisms of change have not been rigorously tested.

Objective:

The current assessor-blinded, parallel-group, randomized controlled trial (RCT) thus aimed to test these theories in social anxiety disorder (SAD).

Methods:

Participants with SAD were randomized to 14-day fully-automated MEMI (n = 96) or a self-monitoring app (SM; n = 95). They completed web-based self-reports of ER and self-compassion at pre-randomization and 7-day mid-intervention and six clinical outcome measures at pre-randomization, 14-day post-intervention, and 1-month follow-up (1MFU). Multilevel modeling determined the efficacy of MEMI on ER and self-compassion domains from pre-mid-intervention. Bootstrapped parallel multilevel mediation analysis examined the mediating role of pre-mid ER and self-compassion domains on the efficacy of MEMI on six pre-1MFU clinical outcomes.

Results:

Participants demonstrated strong compliance, with 78% engaging in at least 80% of MEMI and SM prompts. MEMI was more efficacious than SM in improving ER goals (between-group Cohen's d = -0.24, 95% confidence intervals [CI] [-0.39, -0.10]), clarity (d = 0.16 [0.02, 0.31]), SC isolation (d = 0.19 [0.04, 0.33]), over-identification (d = 0.16 [0.01, 0.30]), and self-kindness (d = 0.19 [0.04, 0.33]) domains from pre-mid-intervention. The within-group effect sizes from pre-mid-intervention on these domains were larger in MEMI (ER goals: d = -0.73 [-0.88, -0.58]; clarity: d = -0.40 [-0.54, -0.25]; SC isolation: d = 0.45 [0.30, 0.60]; over-identification: d = 0.63 [0.48, 0.78]; self-kindness: d = 0.36 [0.21, 0.50]) than SM (ER goals: d = -0.29 [-0.44, -0.15]; clarity: d = -0.21 [-0.35, -0.06]; SC isolation: d = 0.19 [0.04, 0.33]; over-identification: d = 0.48 [0.33, 0.62]; self-kindness: d = 0.10 [-0.05, 0.24]). SM, but not MEMI, enhanced ER awareness (between-group d = 0.11 [-0.03, 0.26]; within-group: d = -0.29 [-0.44, -0.14] vs. d = -0.13 [-0.27, 0.02]) and reduced SC common humanity (between-group d = 0.26 [0.11, 0.40]; within-group: d = -0.23 [-0.38, -0.08] vs. d = 0.13 [-0.02, 0.27]). No ER and self-compassion domains were mediators of the effect of MEMI on SAD symptoms (P = .110 to .976), generalized anxiety symptoms (P = .120 to .906), depression severity (P = .157 to .895), repetitive negative thinking (P = .191 to .896), and trait mindfulness (P = .148 to .893).

Conclusions:

Brief, fully automated, mobile MEMIs efficaciously enhanced specific self-compassion domains (reducing isolation and over-identification and promoting self-kindness) and improved emotion regulation domains associated with goal pursuit and clarity of thoughts and actions from pre-mid-intervention. Higher-intensity MEMIs may be required to pinpoint the specific change mechanisms in ER and self-compassion domains for SAD. Clinical Trial: NCT04846777 on ClinicalTrials.gov and Open Science Framework (OSF) (doi: 10.17605/OSF.IO/M3KXZ)