Feasibility and Acceptability of a Self-Administered Mindfulness-Based Intervention for Stress Reduction in Adult Singaporeans: A Pilot Study

Background: Fully decentralized self-administered mindfulness (SAM) interventions show promise for stress reduction, but rigorous evaluations of their feasibility, acceptability, and effectiveness using both self-report and physiological measures remain limited. In Singapore, where mental health concerns rank as the top healthcare priority (46%), ahead of cancer (38%) and stress-related issues (35%), accessible and scalable interventions are urgently needed to address the significant economic burden of mental health conditions.

Objective: This pre-registered pilot study (NCT06765889) evaluated a decentralized, three-day self-administered mindfulness intervention versus sham control in Singaporean adults, examining effects on self-reported and physiologically measured stress responses. The inclusion of a sham control condition allows for disentangling mindfulness-specific effects from demand characteristics and expectation effects, addressing a critical methodological gap in digital mindfulness research.

Methods: Sixty participants were randomized to either a mindfulness intervention or structurally matched sham control. Stress was assessed via self-report (STAI-6) and physiologically through Heart Rate Variability (HRV). The study incorporated three methodological innovations: (1) A structurally equivalent sham control to match expectancy and credibility, (2) remote collection of HRV as an objective physiological biomarker, and (3) full decentralization via smartphone allowing unsupervised multi-platform delivery. Feasibility was evaluated through recruitment/retention rates, cross-platform delivery, protocol adherence, and data quality. Acceptability was assessed through quantitative ratings and qualitative feedback.

Results: The study demonstrated excellent feasibility with near perfect retention (98.3%) and moderate HRV data quality (69.8% valid signals). Acceptability ratings were strong (M = 4.17/5, SD = 0.53), with highest scores for comfort/engagement (M = 4.27/5, SD = 0.57), exceeding established usability benchmarks for digital health interventions. Qualitative feedback identified technical challenges (HRV instability, device overheating) and scheduling difficulties. While Bayesian analyses revealed no significant group differences in stress reduction (BF₁₀ = 0.03) or HRV improvement (BF₁₀ = 0.2), both groups showed significant stress reductions (BF₁ ₀= 3.01×10⁶), suggesting that observed benefits may stem from non-specific factors common to both interventions.

Conclusions: This study demonstrates (1) the feasibility of conducting fully decentralized mindfulness trials with multimodal assessment, (2) the value of mixed-methods acceptability evaluation, and (3) identifies key technical and control condition refinements necessary for future trials. By addressing methodological limitations in digital mindfulness research through improved control conditions and objective physiological measures, this work provides a foundation for more rigorous investigation of mindfulness-specific versus non-specific effects in remotely delivered stress reduction interventions.