Preventing Mental Health Problems in Children After High Conflict Parental Separation/Divorce Study: An Optimization Randomized Controlled Trial Protocol

Abstract

Parental divorce is a childhood stressor that affects approximately 1.1 million children in the U.S. annually. The children at greatest risk for deleterious mental health consequences are those exposed to high interparental conflict (IPC) following the separation/divorce. Research shows that children’s emotional security and coping efficacy mediate the impact of IPC on their mental health. Interventions targeting their adaptive coping in response to IPC events may bolster their emotional security and coping efficacy. However, existing coping interventions have not been tested with children exposed to high post-separation/divorce IPC, nor has any study assessed the effects of individual intervention components on children’s coping with IPC and their mental health. This intensive longitudinal intervention study examines the mechanisms through which coping intervention components impact children’s responses to interactions in interparental relationships. A 2³ factorial experiment will assess whether, and to what extent, three candidate intervention components demonstrate main and interactive effects on children’s coping and mental health. Children aged 9–12 (target N = 144) will be randomly assigned to one of eight combinations of three components with two levels each: (1) reappraisal (present vs. absent), (2) distraction (present vs. absent), (3) relaxation (present vs. absent). The primary outcomes are child-report emotional security and coping efficacy at one-month post-intervention. Secondary outcomes include internalizing and externalizing problems at the three-month follow-up. Based on data from this optimization phase RCT, intervention components will be selected to comprise a multi-component intervention and assessed for effectiveness in a subsequent evaluation phase RCT.

Parental divorce is a prevalent stressor affecting approximately 1.1 million children in the U.S. annually (Kreider & Ellis, 2011). The children at greatest risk for mental health problems are those exposed to high interparental conflict (IPC) following the separation/divorce. (Amato, 2001) Post-separation/divorce IPC is associated with multiple issues in children, such as depression, anxiety, and conduct problems (van Dijk et al., 2020). Emotional security theory (EST; Davies, Harold, et al., 2002) posits that IPC threatens children’s psychological need to feel safe and secure in family relationships and when they do not, the stress of IPC elicits maladaptive emotional reactions and coping behaviors that over time, result in mental health problems (Davies, Harold, et al., 2002). Given the harmful effects of post-separation/divorce IPC, there is a pressing need for theory-driven, effective interventions that can be widely implemented to protect and promote children’s mental health.

Improving children’s ability to cope effectively when facing IPC and other stressors mitigates their impact (Compas et al., 2017) and existing child-focused interventions teach skills to help children cope with divorce-related stressors (Boring et al., 2015; Pedro-Carroll & Cowen, 1985; Stolberg & Garrison, 1985). However, IPC is an uncontrollable stressor that poses a special coping challenge, and strategies that generally predict better outcomes (e.g., problem-solving) may exacerbate distress in this context. Indeed, the broader coping and stress literature highlights the importance of “goodness of fit” when identifying effective coping strategies (Forsythe & Compas, 1987). For uncontrollable stressors, such as IPC, secondary control coping (focused on adaptation to the situation; e.g., reappraisal, distraction) is hypothesized to be more effective than primary control coping (focused on changing the situation or one’s reaction to it; e.g., problem-solving, emotion regulation) (Compas et al., 2017; Wadsworth & Compas, 2002). Primary control strategies, such as children attempting to change an uncontrollable IPC situation, may have different effects than other primary control strategies, such as children regulating intense emotional reactions to the IPC situation. To contextualize how strategies might fit this situation, the development of coping programs for children navigating high-IPC separation/divorce will substantially benefit from foundations rooted in the principles of the Emotional Security Theory (EST; Cummings & Davies, 2010) and the Cognitive Contextual Framework (CCF; Grych & Fincham, 1990).

The research showing the superiority of particular coping strategies is entirely correlational. To our knowledge, no experimental studies have examined the effects of any specific strategies for coping with IPC or other uncontrollable stressors. Thus, despite substantial evidence that child-focused coping interventions reduce the risk for problems in children’s mental health after separation/divorce (Stathakos & Roehrle, 2003) and that coping may moderate the deleterious effects of IPC (Nicolotti et al., 2003; Tu et al., 2016), coping interventions have also not been optimized. Consequently, they may include iatrogenic or inert components that undermine their effects. To address this problem, we adopted the multiphase optimization strategy (MOST; Collins, 2018), a strategic, phased, and data-driven approach to intervention development that balances effectiveness against affordability, scalability, and efficiency.

An Optimized Intervention: The Multiphase Optimization Strategy (MOST) Approach

Given the high prevalence of divorce and robust data on the risk associated with exposure to high post-divorce IPC, developing effective coping interventions has important public health implications. However, effective interventions will not have a substantial public health impact if they are not also affordable, scalable, and efficient. One approach to developing an effective intervention that prioritizes these critical implementation determinants is through the lens of intervention optimization. Optimized interventions aim to achieve the best expected outcome under the specific constraints of the setting in which the intervention will be implemented. We have adopted the multiphase optimization strategy (MOST; Collins, 2018) framework to guide the systematic development of this intervention. MOST comprises three phases (preparation, optimization, evaluation) to develop optimized multicomponent interventions that are maximally effective, affordable, scalable, and efficient. In contrast to the traditional approach to intervention development, where interventions are tested as a package against a control condition in a parallel two-arm RCT immediately following pilot testing, researchers using MOST experimentally test individual components to ensure they demonstrate a meaningful effect in the desired direction before including them in the intervention package and evaluating its effectiveness. MOST prioritizes the implementation of effective interventions and addresses the “experimental therapeutics” approach to intervention development encouraged by the National Institute of Mental Health by systematically testing the underlying intervention conceptual model (Gordon, 2017).

Optimized Interventions for Family Courts

Using MOST has two advantages for developing interventions for children experiencing high-conflict divorce. First, this approach allows for more refined tests of individual component effects, which is critical given the mixed data on the effects of individual coping strategies for children coping with IPC. Second, it is an implementation-forward approach to intervention development that prioritizes achieving strategic balance among effectiveness, affordability, scalability, and efficiency, depending on the intervention’s intended delivery context. In this case, the intervention will reach the most children if implemented in family courts and with champions from the family law community. Although they constitute only 5–10% of separations (Neff & Cooper, 2004; O’Hara et al., 2019; Schepard, 2004), high-IPC cases have a substantial impact from a public health perspective. Based on nearly 5 million annual case filings reported by the National Center for State Courts, this translates to affecting up to 500,000 families a year (Court Statistics Project, 2020).

The family law community strives to safeguard the health and well-being of children impacted by parental legal disputes, including divorces and custody battles. The prevailing Best Interests of the Child doctrine has been adopted by all jurisdictions in the United States (Child Welfare Information Gateway, 2016). The Association of Family and Conciliation Courts (AFCC) and the National Center for State Courts (NCSC) seek to identify and advocate for best practices that promote favorable outcomes for families embroiled in family court processes (AFCC, 2023; NCSC, 2023). As part of this mission, many jurisdictions across the United States offer programs, some voluntary and some court-ordered, to families who are experiencing parental separation/divorce.

Most of these programs are directed toward the parent; Mayhew (2016) reported that family courts in 46 states mandate parent education programs. Although not well-documented, fewer jurisdictions appear to have programs for children. Many child-directed programs that are easily accessible do not have published reports of their effectiveness (e.g., The Center for Divorce Education, 2023; for an exception see Boring et al., 2015). Of those that have been studied, meta-analytic data indicate weak evidence for their effects on children’s mental health, likely due in part to unknown mechanisms of change (Herrero et al., 2023). Despite repeated documentation that children who experience high post-divorce IPC are at greatest risk for mental health problems, positive findings on interventions for children do not apply to children exposed to high IPC because there is only one published pilot study investigating the impacts of child programs for this high-risk subgroup (O′Hara et al., 2022).

Current Study

This study represents the optimization phase of MOST to develop a digital intervention for children coping with high-conflict parental separation/divorce. The decision to develop a digital intervention was based on pilot data with key stakeholders from the family law community indicating a preference for an easily accessible online program (Sullivan et al., 2023). To optimize the new digital intervention, we will identify which intervention components demonstrate a meaningful effect in the desired direction on putative mediators and clinical targets before combining them and testing an intervention package. The optimization randomized controlled trial (RCT) will use a factorial experiment to screen candidate intervention components for inclusion in the multicomponent intervention which will then be evaluated for effectiveness against an active control in a subsequent evaluation phase.

Intervention Conceptual Model

In MOST, the intervention conceptual model outlines how each candidate component is expected to impact putative mediators and clinical targets (Collins, 2018). Our intervention conceptual model (Figure 1) posits that intervention-induced changes in children’s use of coping strategies in response to IPC events and, in turn, changes in emotional security and coping efficacy, will, over time, decrease or prevent mental health problems.

Figure 1.

Intervention Conceptual Model

Reappraisal, distraction, and relaxation coping strategies are related to fewer problems among children (Chorpita et al., 2005), making intervention components based on these strategies key candidates for inclusion in an optimized digital coping intervention for children exposed to post-separation/divorce IPC. These strategies are theorized to impact children’s emotional security and coping efficacy based on the principles of EST and CCF. For example, relaxation lessens physiological stress responses (Davies et al., 2016), distraction regulates exposure to conflict (Davies, Harold, et al., 2002), and reappraisal reduces negative, threatening, and self-relevant beliefs about IPC events (Grych, 1998; Sturge-Apple et al., 2008). For an extended discussion of how these frameworks have been integrated and applied to coping with post-separation/divorce conflict see (O’Hara et al., 2023). Notably, these are not the only promising candidates for intervention components for inclusion in an optimized digital coping intervention for children exposed to post-separation/divorce IPC; per the “continual improvement” principle of MOST, this optimization trial represents the first step in a programmatic approach to developing an optimized program for these high-risk children.

Decision-Making Philosophy

Our goal is to design a program for children from high-IPC separated/divorcing families that is comprised of all active ingredients that help children cope with IPC events. Thus, our primary focus in decision-making is to arrive at an efficient intervention. Based on preparation phase research with key stakeholders (Sullivan et al., 2023), a program with up to five digital modules that are available online for up to $50 per user is likely to be scalable. Given that we are only testing three components, which can all be delivered online for under $50 per person, there is no need to include cost as a constraint, so our optimization objective is the “all active components” criterion (Collins, 2018, p. 248). This optimization objective focuses on identifying an effective and efficient intervention by sorting candidate intervention components into a screened-in and screened-out set. All screened-in components are then bundled in an optimized intervention package to be evaluated in a large-scale two-arm RCT comparing the optimized intervention to a suitable control condition. Should the RCT show significant effects of the intervention to reduce internalizing and externalizing problems in children exposed to high-IPC parental separation/divorce, the next step will be to disseminate the intervention in family courts and other community settings to maximize its use and impact, and thus reduce the public health burden of divorce.

Decision-Making Procedures for the Screening Experiment

Aligned with MOST’s optimization phase objectives, we embrace a decision-priority perspective (i.e., prioritize practical decisions about which components will be screened in; prioritizes estimated effect sizes) rather than a conclusion-priority perspective (i.e., prioritize drawing scientifically valid conclusions about component efficacy; prioritizes statistical significance). In the optimization phase of MOST, the risk associated with a Type II error (i.e., discarding a potentially useful component) is considered more detrimental than the risk associated with a Type I error (mistakenly including a component with a smaller effect). Thus, the optimization RCT will use a priori effect sizes representing a meaningful change to interpret results whereas the subsequent evaluation RCT will use industry standards for statistical significance to interpret results.

Threshold of Meaningful Change.

We set thresholds for meaningful change on our four outcome measures that are statistically reliable, or not likely to be attributable to measurement error (Achenbach, 2018; Jacobson & Truax, 1992). Based on normative data (Piper et al., 2014), we found that reliable change in our measure of internalizing and externalizing problems corresponded to a minimum effect size of d = .17 and d = .24, respectively. Then, we used our conceptual model (i.e., intervention-induced changes in children’s problems are mediated through changes in emotional security and coping efficacy) to estimate reliable changes in emotional security and coping efficacy. Because the three links in a theoretical mediation model (a path, b path, c path) are mathematically interdependent, we could use prior data to estimate the b and c paths, and solve for the a path with the equation: a*b = c. We assumed complete mediation to be conservative (MacKinnon, 2008). First, we calculated the expected total effect (theoretical “c path;” the overall effect of the intervention on outcomes) based on statistically reliable (i.e., not likely to be attributable to chance) change on the measure of internalizing and externalizing problems (Piper et al., 2014). We then used available data on the correlations of emotional security and coping efficacy with children’s internalizing and externalizing problems in prior studies (Davies, Forman, et al., 2002; Forman & Davies, 2005; Sandler et al., 2000) to estimate the conceptual theory effect (i.e., theoretical “b path;” effect of mediators on clinical targets). Finally, we used the estimated total effect and the estimated conceptual theory effect to calculate an expected action theory effect (i.e., theoretical “a path;’ effect of the intervention on putative mediators). By doing so, we determined that reliable change in emotional security and coping efficacy corresponded to a minimum effect size of d = .18 and d = .33, respectively. See Appendix A for details on our procedures for calculating reliable change values and corresponding effect sizes. Based on all this information, we set the thresholds for meaningful change as an effect size of d ≥ .20 on internalizing or externalizing problems, d ≥ .33 on coping efficacy, or d ≥ .18 on emotional security.

These thresholds for meaningful change were appropriate given the larger literature on interventions for children. A recent meta-analysis showed that single-session interventions have a small-medium average effect size for youth mental health problems (g = .32, 95% confidence interval = .17 - .46.) (Schleider & Weisz, 2017). We also considered the possibility that individual components may demonstrate smaller main effects (e.g., d = .20) but interact synergistically to achieve effect sizes comparable to those observed in prior evaluations of child coping interventions (d = .31 - .55) (Boring et al., 2015; Compas et al., 2010; Sandler et al., 2003). Given all these considerations, setting our thresholds to range from .18-.33 seemed reasonable.

Direction of Change.

We will calculate individual-level outcomes for each participant using an optimal-fitting model adapted from Jacobson and Traux’s (1991) Reliable Change Index (RCI) to assess whether components have a potentially iatrogenic effect. The new optimal-fitting model for estimating the RCI uses moderated nonlinear factor analysis scoring with measurement error correction to produce reliable estimates of the proportion of participants showing statistically reliable improvement or deterioration (Morgan-Lopez et al., 2022). Components with ≥ 10% of participants showing reliable deterioration will be screened out, regardless of their group-level main effect size.

Decision-making Criterion.

Components demonstrating no iatrogenic effects that meet the threshold on at least one outcome measure – internalizing problems, externalizing problems, coping efficacy, emotional security – will be screened in unless an iatrogenic interaction effect involving that component results in a net detriment on one or more outcome measures.

Hypotheses

We expect only synergistic interactions (i.e., the combined effect is more favorable than either main effect) among candidate intervention components on all outcomes. The Reappraisal and Distraction components are anticipated to moderately enhance emotional security and coping efficacy and reduce internalizing and externalizing problems. Two conflicting hypotheses exist for the Relaxation component: based on correlational studies, it may have insignificant or minor effects due to its lesser effectiveness in correlational studies (Compas et al., 2017); however, considering its common presence in successful child coping interventions (Chorpita & Daleiden, 2009), it could have a moderate impact on increasing emotional security and coping efficacy, and reducing problems.

Significance

This study will identify components for inclusion in a digital coping intervention to reduce mental health problems among children exposed to high-IPC parental separation/divorce. Components that meet a priori thresholds for meaningful change in the desired direction will be combined to form an intervention package that will be evaluated against an active control in a subsequent evaluation phase randomized controlled effectiveness trial. If the effectiveness trial shows significant intervention effects to reduce problems in children exposed to high-IPC parental separation/divorce, the next step will be to disseminate the intervention in family courts and other community settings to promote its widespread use and maximize public health impact.

Method

Design

This study is a randomized and counterbalanced 2³ factorial experimental to test the main and interactive effects of three candidate intervention components (i.e., Reappraisal, Distraction, Relaxation) (ClinicalTrials.gov Identifier: NCT05822687). Each candidate intervention component has two levels – present or absent. Participants will be randomized to eight experimental conditions representing all possible combinations of candidate intervention components (see Table 1). We developed an allocation table using the RANDARRAY function in Excel. We used REDCap to implement the randomization model and monitor random assignment. We will counterbalance the order in which components are presented to control for order effects (see Table 2).

Table 1.

Experimental Conditions in 2^x2^x2 Factorial Design

Experimental Condition	Introduction Component	Reappraisal Component	Distraction Component	Relaxation Component	Sample Size (n)
1	YES	NO	NO	NO	18
2	YES	NO	NO	YES	18
3	YES	NO	YES	NO	18
4	YES	NO	YES	YES	18
5	YES	YES	NO	NO	18
6	YES	YES	NO	YES	18
7	YES	YES	YES	NO	18
8	YES	YES	YES	YES	18

Note. For intervention component columns, YES = component is included, and NO = component is excluded. In experimental conditions that include more than one component (i.e., 4, 6, 7, 8), the order of components will be counterbalanced.

Table 2.

Counterbalancing Details

Experimental Condition	Counterbalanced Condition	Introduction (I)	Reappraisal (A)	Distraction (B)	Relaxation (C)	Order of Administration	Sample Size (n)
1	1	YES	NO	NO	NO	I	18
2	2	YES	YES	NO	NO	I | A	18
3	3	YES	NO	YES	NO	I | B	18
4	4	YES	NO	NO	YES	I | C	18
5	5	YES	YES	YES	NO	I | A | B	9
5	6	YES	YES	YES	NO	I | B | A	9
6	7	YES	YES	NO	YES	I | A | C	9
6	8	YES	YES	NO	YES	I | C | A	9
7	9	YES	NO	YES	YES	I | B | C	9
7	10	YES	NO	YES	YES	I | C | B	9
8	11	YES	YES	YES	YES	I | ABC	3
8	12	YES	YES	YES	YES	I | ACB	3
8	13	YES	YES	YES	YES	I | BAC	3
8	14	YES	YES	YES	YES	I | BCA	3
8	15	YES	YES	YES	YES	I | CAB	3
8	16	YES	YES	YES	YES	I | CBA	3

Note. For intervention component columns, YES = component is included, and NO = component is excluded. The Intervention Component Combination column indicates which of the eight combinations of candidate intervention components will be included. The Study Condition column indicates the sixteen conditions which represent both the combination of candidate intervention components and the counterbalanced order of administration of the candidate intervention components. Shading indicates conditions in which participants will receive the same intervention components, but just in different orders of administration. I = Introduction, A = Reappraisal, B = Distraction, C = Relaxation.

The trial is approved by the Arizona State University Institutional Review Board (STUDY00016078). Informed consent will be obtained from all participants. Recruitment started in January 2023.

Participants and Sample Size

Participants will be 144 children ages 9–12 whose parents are separated or divorced. The 9–12 age bracket was selected based on meta-analytic findings, which indicate this group experiences the most significant benefits from cognitive-behavioral programs compared to younger or older children (Stathakos & Roehrle, 2003). Additionally, data highlight pivotal developmental changes in middle childhood, particularly in their capacity to utilize the cognitive coping strategies that are represented in the candidate intervention components being tested in this study (Skinner & Zimmer-Gembeck, 2007; Zimmer-Gembeck & Skinner, 2011).

A sample size of 144 will yield power to detect medium-sized effects (d = .47, power = .80; alpha = .05). A medium effect size is reasonable given the published effect sizes for coping interventions in prior studies of children exposed to family stressors (Boring et al., 2015). In a full factorial 2^x2^x2 design, there are approximately 72 participants, or half the sample, in each of the two levels (present vs. absent) of each factor representing the three candidate intervention components (reappraisal, relaxation, distraction). Statistical power for each estimated effect in a full factorial design is based on the factor level, thus the main and interactive effects of each candidate intervention component will be estimated using the full sample (i.e., half the sample is in the included group and half the sample is in the excluded group for each component).

Procedures

Children will be recruited through their parents from a variety of settings including court- or community-based parent education courses, referrals by family law professionals, schools, university and professional listservs, and social media. Study information will be conveyed to parents through a brief video, a short description in a newsletter, an email, and/or a digital or printed flyer. Interested parents will access the web-based screening survey on the study website.

A web-based screening survey will assess preliminary inclusion criteria: (1) a child between ages 9–12 whose parents are separated or divorced, (2) a child and one parent fluent in English, (3) a parent with sufficient contact (i.e., at least four overnights and/or 12 daytime visits per month) with the child who is willing to complete study assessments and has the legal right to permit for the child to participate, and (4) child not in active therapy. One child will be randomly selected in families with multiple age-eligible children. Preliminarily eligible parents will be directed to a scheduling website to select a virtual intake session appointment.

At the start of the intake session, study staff will explain and answer questions about study procedures and obtains informed consent, parental permission, and child assent. Then parents and children will complete the pretest to determine study eligibility. The first inclusion criterion is the child’s report of high exposure to IPC. In this study, high IPC is defined as a mean Z score ≥ 40^th percentile on measures of IPC (score standardization based on data from 559 youth from a previous trial (Sandler et al., 2020). The second inclusion criterion is elevated internalizing or externalizing problems (T score ≥ 60, according to either child or parent report). Children who score ≥ 99th percentile on either internalizing and externalizing problems or who endorse suicidality will be ineligible for the study and referred for a higher level of care.

Intervention Components

The intervention is comprised of four components – one constant component and three candidate intervention components. All participants will complete the constant component, which provides basic education about divorce and brain science-based explanations of the connections among feelings, thoughts, and actions. The content is based on empirical research and cognitive behavioral therapy principles. It draws on recent literature advocating for the intentional design of youth brief interventions to maximize intervention acceptance and engagement (Schleider et al., 2019). Each candidate intervention component will teach one coping strategy (i.e., reappraisal, distraction, relaxation). All components are tailored to coping with post-divorce IPC but represent standard coping strategies used in evidence-based coping interventions for children, including home practice assignments to encourage the use of coping strategies during their daily lives (Chorpita & Daleiden, 2009). Prior research has established the feasibility of incorporating home practice in digital interventions. Studies of an online intervention for children experiencing divorce showed that children reported high home practice assignment completion (89% in Boring et al., 2015; 74% in O′Hara et al., 2022).

Home Practice Review

After completing each candidate intervention component, children will receive a text or email link to complete a home practice review survey, which will assess whether the child completed their home practice assignment and their experience doing the home practice assignment. They will be guided through a problem-solving activity if they report they did not complete their home practice assignment, or if their home practice experience was anything other than “great” or “good.”

Assessment

Assessments will occur at pretest, posttest (one month later), and three-month follow-up. At each assessment, the child will report their exposure to IPC, global coping style, emotional security, coping efficacy, and internalizing and externalizing symptoms. Parents will report on children’s exposure to IPC and internalizing and externalizing problems.

Completion monitoring and engagement

The completion of intervention components will be monitored by study staff. If participants do not complete their assigned intervention component as scheduled, they will receive reminder phone calls and texts. We will use several engagement strategies to retain participants and encourage participation. We have designed the intervention components to be highly interactive and fun, with input from representative children like those who will participate in this study. Parents and children will earn e-gift cards and other prizes for participating in study activities.

Measures

We will collect demographic (i.e., parent and child race, ethnicity, age, and gender; parents’ current and former marital status; time since separation; family income; parent education level) and contact information for the participants. We will also ask for permission to contact two informants who would know the parent’s whereabouts if their contact information changed.

Covariates

Interparental Conflict.

Children will report their exposure to IPC via two measures: (1) Children’s Perception of Interparental Conflict Scale Properties Subscale Short Version (CPIC; 17 items, Grych et al., 1992); (2) Caught in the Middle scale (CIM; 7 items, Buchanan et al., 1991). The CPIC items are rated on a 3-point scale, 1 [true] to 3 [false], and assess frequency (e.g., “My parents hardly ever argue.”), intensity (e.g., “My parents get really mad when they argue.”), resolution (e.g., “When my parents disagree, they usually come up with a solution.”), child-centered (e.g., “My parents usually argue because of things I do.”), and triangulation (e.g., “I feel caught in the middle when my parents argue.”) properties of IPC events. The CIM items are rated on a 4-point scale, 1 [never] to 4 [very often] and assess the extent to which the child perceives being involved in IPC events (e.g., “how often does [parent] ask you to give messages or information to [parent].” CPIC and CIM have well-established psychometric properties and are validated for children in this age range (Buchanan et al., 1991; Grych et al., 1992). Parents will also report on their child’s exposure to IPC using a parent version of the CPIC.

Global Coping Style.

Children will report on their typical coping responses to IPC events via the Responses to Stress Questionnaire – Parental Conflict (RSQ-PIC; 57 items, 1 [not at all] to 4 [a lot]; Connor-Smith et al., 2000). Items are rated on a 4-point scale, 1 [not at all] to 4 [a lot], and assess different types of coping such as avoidance (“When dealing with my parents arguing, I try not to feel anything.”), problem-solving (“…I try to think of ways to change or fix the situation.”), reappraisal (“I tell myself that everything will be alright.”), and emotion regulation (“I do something to calm myself down.”) The RSQ-PC has demonstrated high reliability across samples (α = .81 – .87; Connor-Smith et al., 2000; ), including a sample of children exposed to family conflict (α =.88; Wadsworth & Compas, 2002), and correlates with children’s mental health problems (Compas et al., 2017).

Primary Outcomes – Putative Mediators

Emotional Security.

Children will report on their emotional security via two measures: (1) Security in the Interparental Subsystem Scale – Short Form (SIS-SF;17 items; Holt et al., 2020); (2) Security in the Family System Scale . SIS assesses perceived security in the interparental subsystem. Items are rated on a 4-point scale, 1 [not at all true of me] to 4 [very true of me] and assess three domains including emotional reactivity (“When my parents argue, I feel scared.”), regulation of affect exposure (“When my parents argue, I try to be really quiet”), and internal representations (“When my parents argue, I feel like it’s my fault.”). SIFS holistically assesses perceived security in the family unit. Items are rated on a 4-point scale, 1 [completely disagree] to 4 [completely agree], and assess three domains including preoccupied (“I feel like something could go very wrong in my family at any time”), secure (“I feel I can count on my family to give me help and advice when I need it”), dismissive (“I don’t care what goes on in my family”). These scales demonstrated adequate reliability and correlated with children’s reports of emotional security and parent and child reports of mental health problems (Davies, Forman, et al., 2002).

Coping Efficacy.

Children will report on their coping efficacy via the Coping Efficacy Scale (CES; Sandler et al., 2000). The CES is rated on a 4 point scale, 1 [not at all] to 4 [very] and measures satisfaction with how they have handled problems (i.e., in general, across situations) in the past (“How good do you think that you have been in handling your problems?”) and how they anticipate handling problems in the future (“How good do you think you’ll be at making things better when problems coming up in the future?). The CES demonstrates a good fit as a seven-item, one-dimensional model (CFI = .95) with high reliability (α = .74; Sandler et al., 2000) and is negatively related to children’s internalizing and externalizing problems (O’Hara et al., 2019; Sandler et al., 2000). We will explore children’s coping efficacy with specific IPC events, but we will use the general coping efficacy scale to assess intervention effects since it has been shown to predict mental health outcomes.

Secondary Outcome – Clinical Target

Mental Health Problems.

Mental health problems will be assessed using the Brief Problem Monitor (BPM; parent and child report; Achenbach et al., 2011) internalizing and externalizing subscales (19 items; “too fearful or anxious” and “argues a lot,” respectively). The BPM is rated on a 3-point scale, 1 [not true] to 3 [very true]. It has demonstrated reliability (e.g., α = .78 [internalizing subscale], .86 [externalizing subscale]; Piper et al., 2014) and validity (e.g., r = .86 - .95 with the Child Behavior Checklist [Achenbach et al., 2001] and significantly higher scores among children with mental health diagnoses) (Piper et al., 2014).

Data Analysis Plan

Data will be analyzed using a linear mixed model with effect coding and will include all possible terms from the 2³ factorial experiment. The dependent variables will be operationalized as the child-report posttest and follow-up scores on the CES (coping efficacy; putative mediator), SIS/SIFS (emotional security; putative mediator), and child- and parent-report on the BPM (internalizing and externalizing problems; clinical target). We will include exposure to IPC, global coping style, gender, age, and baseline measures of the outcome as covariates.

The main and interactive effects of each intervention component on primary and secondary outcomes will be estimated as the difference between the mean score of conditions in which the candidate component is included and the mean score of conditions in which the candidate component is excluded, collapsing across the levels of the other components. Thus, each effect estimate will be based on data from the entire sample (Collins, 2018). Participants are “recycled” during data analysis and contribute to each estimated effect, a key benefit (i.e., efficiency) of factorial experiments as compared to two-arm RCTs (Collins, 2018). The aggregates of cells comprise the intervention and control conditions, which vary depending on which effect is estimated (Collins, 2018). For example, to estimate the main effect of Reappraisal, the mean score of experimental conditions 5, 6, 7, and 8 (see Table 1, column 1 specifies the experimental condition; included components have a “YES” in the Reappraisal column) will be compared to the mean score of experimental conditions 1, 2, 3, 4 (“NO” in the Reappraisal column).

We will explore differential intervention effects by racial/ethnic group and gender. Given that these interaction effects are likely to be smaller than the experiment is adequately powered to detect, we do not expect to find statistically significant effects unless the interaction effect size is large. We will report effect sizes and confidence intervals. Missing data will be handled using multiple imputation (Enders, 2010). We will use auxiliary variables to improve power and reduce bias. Although we do not expect significant differences, we will assess order effects for counterbalanced conditions. We will estimate unbiased intervention effects using treatment non-adherence models (Sagarin et al., 2014).

Optimized Intervention Decision-Making Plan

Our goal in the optimization phase is to decide which components will be included in the intervention package that will be formally evaluated for effectiveness against an active control using a two-arm RCT in the subsequent evaluation phase. We will not rely on formal mediation analyses in this phase because they will not directly affect our optimization phase decision-making. Given the well-documented associations between emotional security and coping efficacy and children’s problems, we have elected to consider intervention effects on theoretical mediators and clinical targets equally in our decision to include or exclude a candidate intervention component. We will use estimated effect sizes instead of the statistical significance of candidate intervention components to conclude whether each component has demonstrated a meaningful effect in the desired direction to minimize Type II errors in the optimization phase. We will follow the decision-making process outlined in Collins’ book (Collins, 2018) to identify an optimized intervention. We are aware of emerging research on this decision-making process (Collins, 2022) and will describe alternative approaches if appropriate. We will make provisional inclusion/exclusion decisions based on the main effect of each candidate component. If a candidate intervention component demonstrates a large enough main effect on any outcome of interest (i.e., (−) d ≥ .20 on internalizing or externalizing problems, (+) d ≥ .33 on coping efficacy, or (+) d ≥ .18 on emotional security), it will be provisionally screened in. We will reevaluate these provisional decisions in light of interaction effects. For example, we will reverse the screened-in decision to exclude any component if, because of an antagonistic interaction (i.e., where the combined effect is less favorable than the main effects), its inclusion would result in a net detriment on any outcome of interest.

The optimized intervention will be comprised of all components that are not inert or iatrogenic or involved in an antagonistic interaction. If we do not identify any intervention components that demonstrate a meaningful effect in the desired direction on putative mediators or clinical targets, we will return to the preparation phase of MOST to refine our conceptual model and consider other possible candidate intervention components before moving to the evaluation phase of MOST. Following the “continual improvement” principle of MOST, we will continue to investigate the unique effects of other promising candidate intervention components in an ongoing effort to further optimize this program for youth exposed to high-conflict parental separation/divorce. For example, given the importance of a robust repertoire of coping strategies and skills to strategically employ particular strategies depending on the situation context (Bonanno & Burton, 2013; Cheng et al., 2014; Forsythe & Compas, 1987), it will be important to develop a toolbox of strategies that help youth cope adaptively in the face of post-separation/divorce IPC.

Limitations

Children aged 9–12 were chosen based on both meta-analytic findings and developmental factors. This study specifically focuses on this age group, and results might not apply to children outside of it. In the future, it’s essential to tailor the intervention components for younger children and adolescents and evaluate their effectiveness. The current study will include intervention components offered only in English and with children living in the United States. Emphasizing the intervention’s effectiveness in samples of children from different cultural and ethnic backgrounds, and geographical regions, and who speak different languages, will be a key consideration in subsequent studies with larger sample sizes. As with all clinical trials, we anticipate encountering practical issues during implementation of the study and have planned for these challenges to the extent possible. For example, to counteract low recruitment, we have plans to continue developing community partnerships with family courts across the United States. We will document all practical challenges and protocol deviations so that we can include a comprehensive discussion on these issues when we report the study’s findings.

Exploratory Analyses

This protocol paper represents the first aim of a clinical trial. We included only the first aim in the protocol paper because the other two aims (i.e., intervention mediation effects and acceptability/feasibility) are exploratory. The full clinical trial protocol, including the hypotheses and analysis plan for all aims, are available on the Open Science Framework and clinicaltrials.gov.

Highlights.

• Interparental conflict amplifies the risk of adverse consequences associated with children’s parental separation/divorce experience. There is a pressing need for effective interventions to reduce such risk.

• Effective, affordable, scalable, and efficient interventions have the greatest public health impact. Optimized interventions include only active components for best-expected outcomes under real-world implementation constraints.

• Adaptive coping mitigates the effects of stressors on children’s outcomes but assessing “goodness-of-fit” is needed to identify effective coping strategies for particular contexts.

• Effective interventions that preserve children’s emotional security and coping efficacy hold promise for protecting their long-term mental health.