JMIR Publications

ABSTRACT

Background:

Internet-based cognitive behavioural therapy (iCBT) for children and adolescents is a ‘persuasive system’ in that it combines 3 major components to therapy—therapeutic content, technological features and interactions between the user and program—to reduce users’ anxiety symptoms. Several reviews report the effectiveness of iCBT; however, iCBT design and delivery components differ widely across programs raising important questions about how iCBT effects are produced and can be optimized. One approach to addressing these questions involves examining the persuasive design features of iCBT programs.

Objective:

We conducted a realist synthesis of iCBT literature using a persuasive system perspective to: (i) document the design and delivery components of iCBT, and (ii) explore how these components may explain the change in anxiety symptoms after completing iCBT.

Methods:

A multi-strategy search identified published and gray literature on iCBT for child and adolescent anxiety up until November 2017. Documents that met our pre-specified inclusion criteria were appraised for relevance and methodological rigor. Data extraction was guided by the persuasive systems design (PSD) model. The model describes 28 technological design features, organized into 4 categories, that help users meet their health goals: Primary task support, Dialogue support, System credibility support, and Social support. We generated hypotheses for how PSD features (Mechanisms) and program delivery features (Context of use) were linked to symptom changes (Outcomes) across iCBT programs using realist and meta-ethnographic techniques. These hypothesized Context-Mechanism-Outcome configurations were refined during analysis using evidence from the literature to improve their explanatory value.

Results:

Forty-five documents detailing 12 iCBT programs were included. Six iCBT programs were rated ‘high’ for relevance and most studies were of moderate-to-high methodological rigor. Eleven Context-Mechanism-Outcome configurations were generated. Configurations were primarily comprised of PSD features from the Primary task and Dialogue support categories. Several key PSD features (Self-monitoring, Simulation, Social role, Similarity, Social learning, Rehearsal) were consistently reported in programs shown to reduce anxiety; many features were employed simultaneously, suggesting synergy in their grouping. The proposed functions of PSD features in iCBT impacts were also included in the configurations. Adjunct support was an important aspect of Context that may have complemented certain PSD features in reducing users’ anxiety.

Conclusions:

The Context-Mechanism-Outcome configurations we developed suggest that, when delivered with adjunct support, PSD features may contribute to symptom reductions for children and adolescents with anxiety. These findings provide an improved understanding of the function, combination and impacts of iCBT program components thought to support desired program effects. Formal testing of the 11 configurations is required to confirm their impact on anxiety-based outcomes. From this we encourage a systematic and deliberate approach to iCBT design and evaluation to increase the pool of evidence-based interventions available to prevent and treat child and adolescents with anxiety.