Moderators and predictors of treatment outcome following adjunctive internet-delivered emotion regulation therapy relative to treatment as usual alone for adolescents with nonsuicidal self-injury disorder: Randomized controlled trial

Abstract

Background

Despite the wide-ranging negative consequences of nonsuicidal self-injury (NSSI), there are few evidence-based treatments for NSSI among adolescents and little is known about what treatments that work best for whom. The objective of this study was to investigate moderators (i.e., for whom a specific treatment works) and predictors (i.e., factors associated with treatment outcome independent of treatment type) of treatment outcome in a randomized clinical trial comparing internet-delivered emotion regulation individual therapy for adolescents (IERITA) plus treatment as usual (TAU) to TAU alone.

Methods

Adolescents (N = 166; mean [SD] age = 15.0 [1.2] years) with NSSI disorder were randomized to IERITA plus TAU (n = 84) or TAU-only (n = 82). Adolescent emotion regulation difficulties, suicidality, NSSI frequency, depressive symptoms, sleep difficulties, global functioning, and age, and parental invalidation, were measured pre-treatment and investigated as moderators and predictors of treatment outcome (i.e., NSSI frequency during treatment and for 4 weeks post-treatment). A zero-inflated negative binomial generalized linear mixed effects regression model was used to estimate the rate of NSSI change as a function of both treatment condition and moderator/predictor.

Results

No significant moderators of treatment outcome were found. Parental invalidation was a significant predictor of treatment outcome regardless of treatment condition, such that high levels of parental invalidation pre-treatment were associated with a less favorable NSSI frequency.

Conclusions

We did not find evidence of a differential treatment effect as a function of any of the examined client factors. Future research should investigate moderation in larger samples and with sufficient statistical power to detect moderation effects of smaller magnitude. Results suggest that parental invalidation may have a negative impact on treatment response and highlight the importance of further investigating parental invalidation in the context of NSSI treatments.