Does online parent training measure up to group parent training in real-world settings?

Background: Disruptive behavior problems involve recurrent patterns of defiance, aggression, and hostility that interfere with normal functioning. Disruptive behaviors are commonly observed in children and adolescents and are linked to a heightened risk of academic and vocational underachievement, substance use, criminal activities, depression, and anxiety in adulthood. Parent training programs are well-established and recommended by guidelines as effective treatments. However, due to limited accessibility, there is a need for alternative approaches to deliver parent training to reach more families beyond traditional methods.

Aim: The thesis’ overall aim was to evaluate treatment for children in primary care with disruptive behavior problems. Study I evaluated if a parenting program (Comet) when delivered online (iComet) would be noninferior in reducing disruptive behavior problems in children to Comet when delivered in its standard face-to-face group format (gComet). Study II assessed predictors and moderators of effects, engagement in and completion of treatment. Study III was a health economic evaluation examining costeffectiveness and cost-utility of the treatments.

Methods: The three studies were based on data from a randomized noninferiority trial involving 161 children with disruptive behavior problems and their parents. Participants were patients in primary care in Stockholm who consented to participate in the trial. They were randomized to receive either gComet (n=86) or iComet (n=75). Assessments took place at baseline and after 3, 6, and 12 months. In Study I, the primary outcome was disruptive behavior problems measured by the Eyberg Child Behavior Inventory (ECBI). Secondary outcomes encompassed the behaviors and well-being of both children and parents, along with treatment satisfaction. Noninferiority analysis was conducted by examining one-sided 95% confidence intervals for the mean difference between gComet and iComet using multilevel modeling. In Study II, linear mixed effects models analyzed predictors and moderators of change in disruptive behavior and treatment engagement and completion from baseline to 3- and 12-month follow-ups. In Study III, the economic evaluation included a cost-effectiveness analysis and cost-utility analysis. Outcomes included recovered and reliably improved cases of disruptive behavior, quality-adjusted life-years (QALYs), costs from a healthcare perspective as well as a wider societal perspective. Statistical analysis involved logistic regression, generalized linear models, and incremental cost-effectiveness ratios (ICERs).

Results: Study I found iComet to be noninferior to gComet at all follow-ups, with small mean differences in reduction of disruptive behavior (d =-0.02 to 0.13) with the upper limit of the one-sided 95% CI below the noninferiority margin of d = 0.43 at 3-, 6-, and 12-month follow-ups (upper limits of 95% CIs between d= .2 and .38). The statistically significant differences in secondary outcomes were clinician-assessed ADHD symptoms, parenting behavior at 3-month follow-up and satisfaction with treatment, all favoring gComet. There were no statistical differences at 12-month follow-up. In Study II, most variables did not predict nor moderate outcomes. Initial problem severity of disruptive behavior and ADHD-symptoms predicted larger decreases in disruptive behavior. Comorbid emotional problems and coercive family dynamics both predicted and moderated effects. Parents’ education level also moderated effect. The three moderators were associated with higher effects in gComet. The only predictor of treatment completion and engagement was matching treatment preference, parents who were allocated to their preferred format completed treatment to a greater extent. Results in Study III showed that healthcare costs were lower for iComet (-$1002., 95% CI -1484, -585), and that gComet resulted in non-significantly higher rates of recovered and reliably improved cases (23 % vs 12 %, p = .129 and 34 % vs 30 %, p = .593). iComet yielded marginally fewer QALYs than gComet for children (-.013, p = .014) and borderline so for the child-parent dyads (-.016, p=0.05). There was no difference in QALYs for parents (-.002, p = .73). The cost-effectiveness results indicate that iComet leads to cost savings while being slightly less effective.

Conclusions: iComet was noninferior to gComet in reducing disruptive behavior. Most variables did not predict or moderate treatment effect. Coercive family dynamics, comorbid emotional problems, and parent education level did moderate the treatment effect with gComet leading to stronger effects. Treatment preference predicted treatment completion. iComet demonstrated cost savings with comparable clinical outcomes, except for slightly higher QALY gain for children in gComet. The combined results lead to the conclusion that internet-delivered parent training can be a viable alternative to group parent training in clinical care.

List of scientific papers

I. Engelbrektsson, J., Salomonsson, S., Högström, J., Sorjonen, K., Sundell, K., & Forster, M. (2023). Parent Training via Internet or in Group for Disruptive Behaviors: A Randomized Clinical Noninferiority Trial. Journal of the American Academy of Child and Adolescent Psychiatry. 62(9), 987–997.
https://doi.org/10.1016/j.jaac.2023.01.019

II. Engelbrektsson, J., Salomonsson, S., Högström, J., Sorjonen, K., Sundell, K., & Forster, M. (2023). Is internet-based parent training for everyone? Predictors and moderators of outcomes in group vs. internet-based parent training for children with disruptive behavior problems. Behaviour Research and Therapy. 171, 104426–104426.
https://doi.org/10.1016/j.brat.2023.104426

III. Engelbrektsson, J., van Leuven, L. Salomonsson, S., Högström, J., Sundell, K., Forster, M., Sampaio, F. The cost-effectiveness of online versus groupbased delivery of a parenting program: evidence from the Comet trial. [Manuscript]