A Trial of Two Universal Programs for Parents of Teenagers

BACKGROUND:

Despite the development of psychological treatments in recent decades, mental health disorders remain a main cause of disease, functional impairments, and deaths of adolescents across the world. UNICEF estimated that globally, over 13 percent of adolescents 10-19 live with a mental disorder. Moreover, suicide is the fifth most common cause of death for adolescents in the same age-span and the fourth most prevalent cause for adolescents 15-19 years old. Reports also show that youth mental illness have increased in recent years. Psychological treatments seems not enough to combat this negative development; preventive efforts are needed.

Family experiences can be a target of prevention given the strong impact on young peoples' current and future mental health. The quality of the parent-child relationship impacts health quality throughout life. Also, adverse experiences related to family and parenting dysfunction is one of the main predictors of mental ill-health. While adolescence is a period of increased independence, parents remain one of the most important people in the young person's life. A link between parenting factors during adolescence and anxiety and depression, which constitutes 40% of diagnosed adolescent mental disorders, is established. Parenting factors linked to depression and anxiety include low warmth, inter-parental conflict, over-involvement, and aversiveness.

Decades of research support the health benefits of parent training programs that help parents interact and communicate with their children, including face-to-face programs delivered in groups and internet-based programs. However, most studies have concerned children below 13 years of age. Given the impact of parenting factors during adolescence on mental health, it is important to investigate how parenting programs work for parents of adolescents.

Parenting programs on different prevention levels are needed to maximize the benefit on public health. However, most studies have concerned indicated/selected prevention or treatments. Outcomes have mostly concerned symptom reduction. To prevent health problems at a population scale, more studies are needed of how parenting programs benefit the general population (i.e., universal prevention). The general population includes those who do not presently experience clinical problems, but where future issues may be prevented through strengthening health-promoting factors (e.g., family relationships).

PURPOSE AND RESEARCH QUESTIONS:

The aim of this study is to assess the effectiveness of two universal preventive parenting programs for teenagers: ABC-teen and ParentWeb. A version of the ABC program for children 3-12 years old has been evaluated in a randomized controlled trial, but so far no study has investigated ABC-teen. Likewise, a longer version of the ParentWeb targeting a selected group of parents with elevated levels of conflict with their adolescents has been published, but no study on the universal version of ParentWeb. The effects of the programs will be evaluated both in comparison to each other and to a Wait-List control. Both programs aim to strengthen the parent-adolescent relationship and reduce negative communication.

1. What are the programs' effect on adolescent and parental mental health?
2. What are the programs' effect on parenting behaviors and parental self-efficacy?
3. What are the programs' effect on the relationship between parents and adolescents?
4. What are the programs' effect on school adjustment?
5. What are the programs' effect on the adolescents' use of tobacco, alcohol and drugs?
6. To what extent do the parents experience participation in the programs' as feasible and satisfying? Research question 1-6 will be reported in the main paper from the project, while the following research questions will be reported in following papers.
7. Are there predictors and moderators of program effects (e.g., demographic variables and variables related to implementation/adherence)?
8. What are the health economic outcomes of the programs?

METHOD:

Participants:

The sample consisted of 1247 families (1327 parents and 471 adolescents) from the general population, from 58 sites in 52 municipalities across Sweden. The participants were recruited through local advertisements in schools or online municipality channels. Parents who were interested were referred to a website with additional information and a consent form. All members in a family (parents and adolescents) were invited to contribute to the data collection in the study (questionnaires).

Design:

The study is a multi-site randomized controlled trial. Parents were randomized to ABC-teen, ParentWeb or a 4 months Wait-List control. After recruitment, participants responded to baseline measurements and were subsequently randomized to one of the three study conditions. The randomization was conducted centrally by the researchers, but with separate randomization list for each site (generated at randomize.org). The researchers then informed a local coordinator at the site about participants' allocation. Parents and adolescents who were included at baseline were invited to answer follow-up questionnaires after 4 and 12 months. The wait-list control group parents were invited to participate in either ABC-teen or ParentWeb (their choice) after 4 months. Thus, at the 12 month follow-up, analyses of program effects will be restricted to comparisons between the programs.

Sample size calculation:

As the effects of universal interventions are generally small, a relatively large study group was required for adequate statistical power, also considering that drop-out can be substantial in multi-site studies. For a statistical power of 80%, 200 participating families will be needed to detect small differences in effect size (Cohen's d = .20) with alpha = 0.05. Since the randomization was unbalanced (2-1-1 in ABC-teen, ParentWeb, and wait list), an additional 200 families had to be recruited to reach at least 200 in each condition (i.e., 400 in ABC-teen, 200 in ParentWeb, and 200 in the wait list). This means that a total of 800 families with complete data were needed for the study. To account for dropout, an additional 447 families were recruited. Actual n:s in each condition at each measurement point were as follows at the completion of data collection: ABC-teen - 669 (baseline), 520 (4 month), 458 (12 month); ParentWeb - 297 (baseline), 209 (4 month), 168 (12 month); Wait list - 281 (baseline), 210 (4 month).

Statistical analyses:

Linear Mixed Modeling (LMM) will be used for analyzing the effects of the programs (questions 1-5) in two steps: First, effects at 4 month will be analyzed comparing each of the program to the wait list control. Second, the two programs will be compared to each other at 4 and 12 month. For questions 6, t-tests will be conducted to investigate mean differences between the two program conditions. For question 7, LMM-analyses will be conducted with predictors and moderators entered into the equations, to test if they have (differential) effects on the slope of adolescent mental health and parenting variables from baseline to 12 month follow-up. For question 8, two different health economic analyses will be conducted - one analyzing the cost of health gains in terms of quality of life (by mapping scores from the SDQ to preference-based utility values), and another analyzing the cost of health gains in terms of reduced mental illness (recovered cases). All analyses will be conducted according to Intention-to-treat (primary results), as well as per protocol (sensitivity).