What Happens When It’s Really Self-Administered? Parent Utilization of a Tablet-Based Parent Training Program

Introduction

mHealth parent training interventions are an efficient and evidence-based approach toward mitigating behavioral problems in young children. Critical to successful implementation of these interventions is an understanding of their real-world implementation. The purpose of this presentation is to explore, through an evaluation of digital tracking analytics, the ecological validity of the ezParent program during an effectiveness trial.

Methods

Digital utilization data in the effectiveness trial (study 2) will be compared to a previous efficacy trial (study 1). Parents in both studies were recruited from pediatric primary care clinics but intervention delivery and support differed. In study 1, parents (N=40) received a structured training of the ezParent program and demonstrated their ability to navigate the program. In study 2, parents (N=100) were given the ezParent and shown how to access a video tutorial, no other program support was provided. Ongoing program support differed across studies, in study 1, study staff conducted follow-up calls during the intervention period to support use and questions regarding the program. In study 2, follow up calls were conducted only as part of data collection. ezParent utilization was assessed via total modules completed of the 6-module program over the 3-month intervention period.

Results

Parents in study 2 and 1 were similar in age (32.7 vs 33.8), number of children (2.3 vs 2.8), and race/ethnicity (91% vs 90% African-American or Latinx). On average, parents in study 2 were older at age of first child (24.0 vs 21.6, p=0.03) and less likely to have a child with a physical or mental disability (9% vs 28%, p=0.01). Parents in study 2 completed an average of 1.17 fewer modules than study 2 parents (p<0.01). To control for potential confounders, number of modules completed was regressed on study and a number of potential covariates. This difference in utilization by study remained significant (β=-1.12, p=0.01), and the effects of covariates were all non-significant. Additional analyses will include alternative metrics for assessing intervention use.

Conclusions

These results warrant further investigation into the effects of simulating real-world circumstances for mHealth interventions. While additional factors will be explored and other implications investigated, these results suggest that even minimal procedural changes to improve ecological validity can significantly affect program utilization.