Predicting Parent Participation in Two Parent-Training Delivery Models: A Person-Centered Approach

Introduction: Poor parent participation in parent-training programs represents a significant barrier to program implementation and effectiveness. Several parent characteristics may affect participation in these programs, including stress, mental health, and motivation to engage in treatment (Kazdin, Holland, Crowley, & Breton, 1997). There is also evidence that delivery method (e.g., in-home vs. center) impacts parent participation (Bloomquist, August, Lee, Piehler, & Jensen, 2011). However, little research has examined how parent characteristics interact with delivery method in promoting parent participation. The current study aimed to 1) utilize a person-centered approach to identify common profiles across several parent characteristics linked to participation and 2) examine how membership in those profile subgroups may moderate the impact of randomization to two delivery models of parent training (in-home vs. center based) on participation.

Methods: Participants included 246 parents of school-aged children identified by teachers as displaying elevated levels of behavioral problems. The ethnically diverse sample was randomly assigned to participate in one of two delivery models of the Early Risers Conduct Problems Prevention Program: home-based (Outreach) or community-center based (Center). Participation was tracked throughout the 18-month intervention through minutes of contact. Hypothesized predictors of participation measured at baseline included parent motivation, extraversion, interpersonal support, depression, anxiety, discipline practices, and parenting confidence and efficacy.

Results: A latent profile analysis using 8 hypothesized predictors of participation as class indicators revealed a best-fitting 3-class solution. These classes were categorized as high distress (9%; high mental health symptoms, low support, high motivation, and low parenting confidence), low parenting confidence (23%; low parenting confidence and moderate mental health symptoms), and low distress (68%; low motivation, high parenting confidence). Controlling for demographic variables, class membership significantly moderated the impact of intervention delivery model on participation. The high distress class demonstrated higher participation in Center when compared to Outreach, while the low distress class demonstrated relatively equal participation in both delivery models.

Conclusions: Person-centered methods such as latent profile analysis have the potential to yield valuable insights into participant response to prevention programming. The current findings highlight specific parent profiles associated with variable participation in two parent-training delivery models. These results may lead to tailored prevention approaches associated with improved engagement and outcomes.