Implementing Digital Peer Support Training for Adolescent Mental Health: Study Protocol For A Type III Hybrid Randomized Controlled Trial

Background: The digital peer support training for adolescent mental health project aims to implement a multidisciplinary program, which translates theoretical and empirical work from implementation science, communication science, and developmental psychology, to train adolescents in effective digital peer support for mental well-being. Nearly 50% of all mental health issues has an early onset in adolescence, and mental health conditions have long lasting and far-reaching effects for adolescents. Although there is an emerging body of work that underscores digital platform as an important interpersonal context for peer support in managing well-being in adolescence, there is no implementation of an evidence-based intervention to train adolescents in effective online peer support.

Methods: Using a hybrid type III trial with a cluster randomized design, a total of 100 Singaporean students from 4 classes in a high school will be randomly allocated to the intervention or wait-list control arm. The intervention arm will undergo bite size modules on the four active ingredients of youth mental well-being-specifically, Mattering, selfhood, compassion and mindfulness, which will be delivered through training workshops, simulation activities, and homework assignments. Students in the wait-list control arm will receive the the training program after the conclusion of the trial with the intervention arm. The pre-post with control evaluation design will be utilized to evaluate program outcomes. Fidelity will be the primary outcome assessed to demonstrate the effectiveness of the digital peer support training program. It will be measured by the extent to which adolescents' responses to real-cases of peer disclosure indicate Mattering, selfhood, compassion and mindfulness. Reach, acceptability, cost-effectiveness, and adolescent self-reported psychological well-being will be assessed as secondary outcomes. Cost-effectiveness analysis will inform the development of a scalability and sustainability plan.

Data Collection, Management and Analyses: We conducted power analysis using to ascertain the adequate sample size needed for the randomized controlled trial, specifically in comparing the intervention and control groups on different implementation and clinical outcomes, so that the study has sufficient power to detect valid effects. The study sample size of n = 50 for intervention and n = 50 for control is needed because of multiple dependent variables measuring implementation outcomes included in the study--specifically, fidelity, reach, acceptability, and cost-effectiveness, and the testing of the clinical outcome on psychological well-being.

Missing data In dealing with missing data, we will determine if missing data is missing completely at random. If missing data is not systematic, as indicated by the non-significant result from Little's Missing Completely at Random Test, we will handle missing data using full information maximum likelihood imputation.

Qualitative analysis We will conduct qualitative analyses with NVivo 10 on the secondary outcome on acceptability of the program. This outcome data will be gathered from both the adolescents and their teachers using focus group discussions post-intervention to elucidate the facilitators and barriers at the individual and organization levels that contribute to and hinder adolescents in providing effective digital peer support, respectively. Participants' responses will be subjected to topic modelling using Latent Dirichlet Allocation (LDA) analyses with R to filter the huge amount of data available and drill down relevant themes and topics. LDA is an advanced statistical method that increases precision and accuracy in coding.