Abstract
Despite substantial empirical support, correlates of retention and success in community-based parent–child interaction therapy (PCIT) implementation are not well defined. Widespread application of PCIT necessitates improved understanding of intervention components relating to family outcomes beyond highly controlled research trials. Using data collected as part of routine care, this study examined homework completion, time in intervention, and parent perceptions of pre-intervention behavioral issues as predictors of PCIT completion in a community-based sample. Subjects included 78 parents (49 women and 29 men) of 45 children (20 girls, 25 boys; mean age = 5.53 years) participating in PCIT in an outpatient behavioral health clinic in a small Midwestern U.S. city. Unlike previous controlled trials, homework completion did not predict child behavior growth or intervention completion. Reasons for early termination were examined thematically, and parent stress emerged as a possible avenue for future intervention in helping parents successfully complete PCIT. The issue of intervention dosage was also explored to see if families who prematurely terminated from PCIT still evidenced gains in child behavior.
Parent–child interaction therapy (PCIT; Eyberg, 1988) is a behavioral parent training intervention targeting externalizing behavior problems in young children between 2 and 7 years of age. Disruptive behaviors commonly include oppositionality, hyperactivity, non-compliance, and aggression (Tremblay, 2010). Children may have a variety of diagnoses relating to disruptive behavior to participate in PCIT, but common diagnoses include attention deficit hyperactivity disorder (ADHD), oppositional defiant disorder (ODD), and conduct disorder (Ward et al., 2016), and more recently autism spectrum disorder (ASD) when behavior problems co-occur (McNeil et al., 2018). PCIT combines the components of social learning theory (Bandura, 1977) and attachment theory (Ainsworth & Bowlby, 1991) to strengthen the quality of parent–child relationships and communication during child-led play and then teaches parents safe and effective behavior management skills. PCIT’s hallmark features include live clinician coaching feedback couched within an assessment- and data-driven process.
For the purposes of this study, “community-based” refers to PCIT that was offered as part of typical practice at outpatient community mental health centers or private practices that was not provided for the purpose of conducting research. In our study, PCIT was provided onsite in a private practice, but other community-based applications of PCIT include homes, through telehealth, and child welfare settings. Laboratory-based PCIT, however, refers to PCIT conducted for the purpose of research, usually within a university setting, under highly controlled conditions.
Substantial evidence supports the efficacy of PCIT (e.g., Cooley et al., 2014; Ward et al., 2016). Meta-analyses show large mean effects for decreases in child behavior problems (d = −1.06, Cooley et al., 2014; d = −1.65, Ward et al., 2016). Factors related to parents’ engagement in PCIT are critical to understand given known associations between parent background and outcome (Danko et al., 2016) and to inform better use of resources and to promote better child outcomes and improved parent satisfaction. It is critical to understand what potentially malleable factors may foster long-term engagement in this popular approach to managing young children’s behavioral concerns.
PCIT includes two stages, with the goal of first fostering positive parent–child relationships during the child-directed interaction (CDI) phase and then teaching parents effective and healthy behavioral management skills, such as precision instructions and appropriate reinforcement, and consequences during the parent-directed interaction (PDI) phase. At the outset of both phases, trained clinicians teach parents the requisite skills during a parent-only teaching session. In subsequent coaching sessions within each phase, clinicians provide immediate feedback as parents practice skills with their child in vivo through a one-way mirror and a “bug in the ear” (Eyberg, 1988). For instance, the parent may wear a wireless earphone connected to a phone call with the clinician who is observing and providing live feedback and recommendations from behind a one-way mirror. Skills taught in CDI include labeling behaviors, using specific praise, and reflecting the child’s speech during play. Parents also practice avoiding questions, critical statements, and commands during CDI. Skills taught in PDI include using direct and unambiguous commands, enthusiastic, specific praise for child compliance, giving warnings for time-out for non-compliance, and implementing a time-out procedure if non-compliance occurs after a warning. PCIT clinicians review graphs of progress monitoring data (using the tools described below) with parents weekly. Table 1 provides an overview of expectations for parents receiving PCIT.
Overview of Engagement Expectations for Families in PCIT.
Note. PCIT = Parent–child interaction therapy.
ECBI = Eyberg Child Behavior Inventory (Eyberg & Pincus, 1999). The ECBI is a parent-reported rating scale for child behavior concerns. b The total treatment course consists of the CDI phase and the PDI phase.
To address PCIT’s goals of improving the parent–child relationship and reducing child behavioral concerns, parent and child behaviors are assessed at baseline and are monitored weekly. Families “graduate” from PCIT after achieving mastery, demonstrated by displaying consistent use of effective parenting techniques, measured by the Dyadic Parent–Child Interaction Coding System (DPICS; Eyberg et al., 2013). The DPICS is a behavioral coding system measuring the quality of parent–child interactions, including the positive behaviors of providing labeled praise, reflections, and behavioral descriptions, and child compliance with parental directions. Parents must first demonstrate mastery of the positive parenting criterion skills in the CDI phase before moving onto the use of precision commands and contingencies in the PDI phase. A second critical element of PCIT progress is measured by parent-reported child behavioral issues through the Eyberg Child Behavior Inventory (ECBI; Eyberg & Pincus, 1999). To “graduate” from PCIT, parent-reported child behavior must improve to a specific cutoff score on the ECBI.
Although PCIT is effective in increasing quality parent–child interactions and reducing child behavior concerns (Cooley et al., 2014; Ward et al., 2016), progress may be hindered by a lack of parent engagement and commitment to the intervention process. Parent engagement is vital for successful intervention for many childhood disorders, especially for externalizing behavior problems (Nock & Ferriter, 2005). In behavioral parent training, parents’ adherence to the intervention is associated with motivation toward the intervention and session attendance (Nock & Kazdin, 2005). Intervention programs that address families’ logistical and psychological barriers to engagement improve initial participation, but long-term effects of these components are less compelling (Ingoldsby, 2010). PCIT includes a brief weekly check-in meant to target impediments to family engagement, but constructs measuring parent engagement in the intervention (e.g., homework completion and session attendance) are not well understood in relation to outcomes.
Issues with engagement are prevalent in behavioral parent training programs in general (Chacko et al., 2017) and in PCIT (Werba et al., 2006). Early dropout negatively affects gains measured at follow-up (Boggs et al., 2005). However, some research suggests moderate effects following just four sessions (Lieneman et al., 2019), thus gains may not be dependent on a full course of PCIT. Attrition rates vary for PCIT across laboratory- and community-based studies. For instance, families who received less than the full dose of PCIT ranged between 18% (n = 14) and 36% (n = 36) in laboratory-based studies by Stokes and colleagues (2016) and Fernandez and Eyberg’s (2009), respectively. In relatively less prevalent community-based studies of PCIT, families receiving less than a full dose of PCIT ranged between 35% (n = 14) and 67% (n = 8; Abrahamse et al., 2016; Lyon & Budd, 2010). In short, families’ dosage of PCIT varies with dropout, and mixed findings on the importance of time in intervention suggest a need to better understand the factors associated with dropout.
Early research on PCIT included some efforts to identify the predictors of attrition. Predictors in previous studies included lower socioeconomic status (SES), less frequent maternal praise pre-intervention, disagreement with the intervention philosophy (Fernandez & Eyberg, 2009), greater frequency of negative maternal talk as coded using the DPICS (Fernandez & Eyberg, 2009; Werba et al., 2006), and maternal distress at baseline (Werba et al., 2006). Some studies identified baseline child behavior problems as a predictor of intervention attrition or graduation but findings are mixed. Bagner (2013) found that greater severity of behavior problems was significantly associated with early dropout, while Werba and colleagues (2006) only found a small, non-significant effect size for baseline severity of child behavior problems on attrition. A few studies have reported reasons for attrition from PCIT, including family satisfaction with gains, severity of other family issues, and discontinuation for no specific reason (Abrahamse et al., 2016). In a laboratory-based study (Fernandez & Eyberg, 2009), commonly cited reasons for early dropout included disagreement with the PCIT approach, other parental stressors that interfered with engagement, scheduling conflicts, and logistical issues. Across previous PCIT studies, family characteristics, including low income and impaired functioning, emerge as risk factors for dropout.
Conversely, some studies have examined factors that bolster engagement and success in parent-mediated interventions. Among parents participating in behavioral parent training programs in general, belief in the credibility and potential success of the program predicts likelihood of engagement and successful completion (Nock et al., 2007). In PCIT-specific studies, the role of the clinician is highlighted. For example, Barnett and colleagues (2017) found that a more reflective and responsive rather than directive clinician coaching style that reinforces parents’ efforts is associated with higher completion rates. In addition, Harwood and Eyberg (2004) found that the quality of the therapeutic alliance between parent and clinician as measured by clinician verbalization style was key for intervention success.
Beyond attrition or successful intervention completion, homework completion is a seldom-studied dimension of parental engagement. Homework in PCIT is the parents’ time to practice skills learned during sessions and to spend “special time” with their child. Completion of weekly homework is a proxy for parents’ intervention adherence, as parents are asked to report weekly on their homework completion. Homework procedures are detailed in the “Method” section. Because PCIT is a manualized and highly standardized intervention, clinicians explain goals and expectations for sessions and between sessions using the scripts in the PCIT Therapy Protocol (Eyberg & Funderburk, 2011). However, it is often difficult for clinicians to ensure that parents complete homework assignments, and studies of behavioral parent training in general report homework completion around 50% (Chacko et al., 2009).
In PCIT, homework completion rates tend to range from approximately 40% to 77% (Berkovits et al., 2010; Lyon & Budd, 2010). Common reasons for homework non-completion include lack of time, forgetting, and difficulty implementing skills (Chacko et al., 2013). Some PCIT studies indicate that homework completion is a moderate predictor of faster time to completion (Danko et al., 2016; Stokes et al., 2016), higher likelihood of graduation from intervention, greater satisfaction with the intervention (Danko et al., 2016), reduced parenting stress, and increased use of positive parenting strategies (Ros et al., 2016). Conversely, Stokes and colleagues (2016) found that homework completion was not related to differences in disruptive behavior after controlling for baseline levels of child behavior. It is necessary to clarify the relation between homework completion and child outcomes in community-based samples where motivation to complete homework may wane over the course of the intervention without the typical monetary incentives available in many laboratory-based studies.
Yet another factor that may be particularly relevant to community-based PCIT is time in intervention. Meta-analytic studies of PCIT report an average of 12 to 14 sessions for completion (Cooley et al., 2014; Thomas et al., 2017), and each session lasts for approximately 1 hour. However, some community-based studies show longer times to completion, such as the mean 22 sessions in Abrahamse and colleagues’ (2016) Netherlands-based study. It is important to note that the duration of intervention is dependent on the number of sessions needed for parents to master strategies in each phase. In community-based samples, there is a need for interventions in which the effects are achieved within a reasonable time frame, given potential strains on families’ time and finances (e.g., insurance co-payments; out-of-pocket expenses). Notably, some community samples in PCIT studies have included families with overall lower SES (e.g., Abrahamse et al., 2016; Lyon & Budd, 2010) compared to the laboratory-based samples (e.g., Fernandez & Eyberg, 2009).
Current Study
The growing popularity and widespread use of PCIT in community practices necessitates improved understanding of the influence of malleable factors related to family outcomes. The contribution of these components to success in PCIT has seldom been explored, particularly in community-based studies. The current study examined the relations between parental perceptions of child behavior, PCIT homework completion, time to completion, and child behavior outcomes in a community-based sample that received PCIT through a private outpatient behavioral health clinic in a small city in the Midwestern United States. Reasons for which families terminated prematurely were also examined qualitatively. We designed this preliminary study in conjunction with clinicians to inform future practices in enhancing parent engagement and to develop future research.
Research questions
The study sought to examine whether homework completion rates, pre-intervention child behavior scores, and time in intervention predicted successful completion of PCIT intervention and whether parents who perceived their child’s behavior to be more or less problematic at baseline differed in their homework completion, time in PCIT, and behavior tolerance and child behavior outcomes post-PCIT. In addition, it sought to determine among parents of children who graduated from PCIT if shorter time in intervention predicted more positive child behavior outcomes. Final goals were to understand why some families terminated early from PCIT prior to meeting graduation criteria and whether there were significant reductions in child behavioral concerns reported for early termination families who completed at least four sessions.
Method
Participants
Participants were 78 parents of 45 children who participated in PCIT at a private outpatient behavioral health clinic located in a small university city in the Midwestern United States between September 2016 and June 2019. This city’s population is approximately 120,000 (United States Census Bureau, 2020). Services were provided through private medical insurance and private pay. See Table 2 for parent and child demographic and child diagnostic information. Diagnoses were retrieved from each child’s electronic medical record and confirmed by clinicians overseeing each case. Primary diagnoses are presented in Table 2 and included ASD, ADHD, other disruptive behavior disorders, and anxiety and adjustment disorders. It should be noted that 10 children were diagnosed with more than one disorder and that this clinic is known for their work with children with ASD. Children’s ages ranged from 2.47 to 7.87 years (M = 5.53 years; SD = 1.24). Overall, 62% (n = 28) of families graduated from PCIT, meaning they successfully completed all DPICS and ECBI mastery criteria, resulting in an attrition rate of 38%. Race, ethnicity, and SES information were not available for the sample. The median household income for the city is US$65,745, with 22% of families living below the poverty line. The population is 68.6% White, 15.9% Asian, 6.89% Black/African American, and 4.38% Hispanic/Latino (United States Census Bureau, 2020).
Demographic Characteristics of Participants.
Note. ASD = autism spectrum disorder; ADHD = attention deficit hyperactivity disorder. Behavioral disorders included oppositional defiant disorder (i.e., ODD), unspecified disruptive impulse control and conduct disorder, and a genetic disorder with associated externalizing symptoms; anxiety disorders included generalized anxiety disorder (i.e., GAD), unspecified anxiety disorder, and selective mutism. In 10 cases, children were diagnosed with more than one disorder.
Measures
Time in intervention
Time in intervention reflected the number of intervention sessions attended by families across both CDI and PDI phases.
Homework completion
Clinicians asked parents to use a homework tracking sheet dividing the number of days of the week that they completed homework by 7 days in the week. This calculation yielded a percentage for each week for each parent. Percentages were averaged for each parent for CDI, PDI, and overall intervention. Completion was calculated beginning with the third session as there was no homework assigned prior to this session. “Homework” is the term used in PCIT for practicing the skill learned that week between sessions. During CDI, homework consists of spending 5 minutes in child-directed play (e.g., “special time”), and during PDI, it consists of spending 5 minutes in child-directed play practicing CDI skills followed by a clean-up session in which the parent practices PDI skills, such as using effective commands. Parents are also instructed to use PDI commands and follow-through throughout the day as needed and to keep track of completing “homework” on a form provided by clinicians (Eyberg & Funderburk, 2011).
Graduation from PCIT
PCIT clinicians determined successful graduation from the intervention based on predetermined PCIT mastery criteria, as measured by meeting pre-set mastery criteria (i.e., certain frequencies of clinician-observed parenting behaviors relevant to CDI and PDI skills) based on the DPICS and a child ECBI intensity score ≤ 114. For our study, graduation/completion was dichotomously coded “yes” or “no.”
Child behavior issues and parental behavior tolerance
Child behavior issues and parental tolerance of these behaviors were measured at the beginning of each PCIT session using the ECBI, a parent-report measure (Eyberg & Pincus, 1999). The ECBI contains 36 items assessing common child behavior problems among children with disruptive behavior disorders. Parents rated the frequency of each behavior on a scale from 1 (Never) to 7 (Always). Scores from each item were summed to produce a total ECBI intensity score. Scores of 131 or greater on the intensity scale are considered clinically significant, indicating the child’s behavior likely poses a significant challenge to the family (Eyberg & Pincus, 1999). The child’s ECBI intensity score must be 114 or lower before a family can meet PCIT graduation criteria.
The ECBI is intervention-sensitive (Schuhmann et al., 1998) and has been established as psychometrically adequate. The ECBI intensity scale and problem scale both have high internal consistency (α > .90; Colvin et al., 1999). In addition, inter-rater reliability coefficients are between .79 and .86, and test–retest reliability coefficients are between .86 and .88 (Eyberg & Pincus, 1999). The ECBI intensity scale shows convergent validity with measures of child behavior problems, including the Strengths and Difficulties Questionnaire (SDQ; r = .068; Axberg et al., 2008; Butler, 2011) and the Child Behavior Checklist (CBCL; r = .75; Boggs et al., 1990).
We computed child behavior change scores by subtracting the post-intervention ECBI intensity score from the baseline ECBI intensity score. Change score calculation yielded negative numbers if child behavior improved. Parents indicated “yes” or “no” for each behavior, indicating whether or not they found it troublesome. The sum of “yes” answers yielded the ECBI problem score, which was used as a measure of parent tolerance of problem behaviors. Scores of 15 or higher on the ECBI problem scale may indicate significant parental distress due to how troublesome they find their child’s behavior problems (Eyberg & Pincus, 1999). For the current study, parent behavioral tolerance was dichotomously scored as “low” (<15) and “high” (≥15) based on pre-intervention scores.
Family termination
Clinicians recorded termination reasons for families who did not graduate from PCIT based on information reported by families and/or issues observed by clinicians during PCIT.
Procedure
We analyzed data previously collected during PCIT at the clinic. The project was reviewed and approved by the University’s Institutional Review Board. Data were de-identified before being shared with the research team. All analyses were conducted using SPSS (Version 25).
Data Analysis
Research question 1: Predictors of PCIT completion
We used logistic regression to assess the influence of average homework completion rate, pre-intervention child behavior issues (ECBI intensity), and total number of intervention sessions on the likelihood that parents would successfully complete PCIT, controlling for child gender. Logistic regression is used to examine the predictive effects of multiple independent variables on a dichotomous outcome variable. Results are presented as odds ratios, or the likelihood of the outcome occurring due to the influence of each independent variable (Stoltzfus, 2011).
Research question 2: Influence of pre-intervention behavioral tolerance on parental engagement in PCIT
We performed a one-way, between-groups multivariate analysis of variance (MANOVA) to investigate differences in homework completion, time in intervention, behavior tolerance post-intervention, and child behavior outcomes across high (n = 45) and low (n = 20) pre-intervention behavior tolerance. Because preliminary assumption testing was violated for homogeneity of error variance for homework completion, the alpha level for significance was modified to .01.
Research question 3: Influence of time in intervention on child behavior outcomes for PCIT completers
We conducted a linear regression to determine, among children whose parents completed PCIT (n = 43), whether the number of intervention sessions predicted child behavior outcomes (ECBI intensity change score).
Research question 4: Reasons for premature termination from PCIT
To answer this question, we conducted a thematic analysis of reasons cited by clinicians for families’ early termination (n = 18) using Braun and Clarke’s (2006) thematic analysis technique. Braun and Clarke’s technique involves five steps: (a) familiarizing oneself with the data, (b) generating initial codes by systematically noting interesting features across all data, (c) searching for themes by assembling codes into groups of potential themes and collating information relevant to the theme, (d) reviewing the themes to see if they make sense among the codes and within the data set as a whole, (e) refining, defining, and naming the themes, and (f) reporting on the analysis. After working through this process, we conducted inter-observer agreement for the thematic analysis by asking an independent researcher (a doctoral school psychology student) to match each family’s early termination reason to the themes identified by the first author. We calculated inter-observer agreement by dividing the number of themes on which the two researchers matched by the total number of themes. Inter-observer agreement was considered acceptable at 88%.
Research question 5: Child behavior outcomes for early termination families
A dependent sample t-test was used to evaluate the significance of pre–post change in child behavior (ECBI intensity score) for children of parents who terminated early but completed at least four sessions (n = 22).
Results
Descriptive Statistics
Descriptive statistics for the variables of interest can be found in Table 3.
Descriptive Statistics.
Note. CDI = child-directed interaction; PDI = parent-directed interaction; ECBI = Eyberg Child Behavior Inventory (Eyberg & Pincus, 1999). Overall, average weekly homework completion occurred across both CDI and PDI treatment phases or however long participants stayed in treatment. Child behavior growth scores were calculated using post and pretreatment ECBI intensity scores.
Research Question 1: Predictors of PCIT Completion
The full logistic regression model containing all predictors was statistically significant, χ2(3, N = 73) = 28.01, p < .001, indicating that the model was able to distinguish between respondents who did and did not complete PCIT. The model as a whole explained 31.9% (Cox and Snell R square; Cox & Snell, 1989) of the variance in graduation status and correctly classified 86.3% of cases. Only one variable, total number of sessions, contributed significantly to the model (OR = 1.29; p < .01). That is, for every additional intervention session completed, parents were 1.29% more likely to be PCIT graduates.
Research Question 2: Influence of Pre-intervention Behavioral Tolerance on Parental Engagement in PCIT
There was no statistically significant difference between high- and low-tolerance parents on the combined dependent variables F(4, 60) = 0.83, p = .51; Wilks’ Lambda = .95; partial eta squared = .05. None of the individual dependent variables were statistically significant.
Research Question 3: Influence of Time in Intervention on Child Behavior Outcomes for PCIT Completers
Results of the linear regression were not significant, R2 = .05, F(1, 41) = 2.06, p = .16. Number of intervention sessions did not predict child behavior outcomes.
Research Question 4: Reasons for Premature Termination from PCIT
Multiple reasons were sometimes cited for early termination (n = 18), thus some families were represented across multiple themes. The most common reasons for early dropout were need for different/more intensive services (e.g., applied behavior analysis, anger management; n = 7; 39%) and parent stress beyond the parent–child relationship (n = 5; 28%). Other reasons included disagreement with the intervention philosophy (n = 3; 17%), other interventions or therapies considered more optimal or of higher priority due to child’s comprehension of “if-then” statements (n = 3; 17%), life situations beyond parents’ control (e.g., moving away, insurance issues; n = 2; 11%), and parent beliefs that the child no longer needed intervention due to sufficient behavioral improvement (n = 2; 11%).
Research Question 5: Child Behavioral Outcomes for Early Termination Families
Among the 22 parents of 14 children who terminated from PCIT before meeting graduation criteria with available data who completed at least four sessions, child behavioral concerns were significantly lower for the final session (M = 118.32, SD = 18.86) compared to the initial session (M = 147.23, SD = 30.01) (t21 = 5.18, p = .000, d = 1.15). These parents completed an average of 11.55 sessions (range = 4-17). Five parents of three children were not included in this analysis because they did not have full ECBI data available. One parent of one child completed less than four sessions and was also not included in this analysis.
Discussion
Although PCIT is well established as an efficacious intervention (Cooley et al., 2014; Ward et al., 2016), much of the literature includes highly controlled, laboratory-based samples or families referred for child welfare concerns. There is a need for more research studying PCIT implementation in diverse, community-based settings. This study contributes to the literature by providing insight into the correlates of intervention outcomes in a community-based PCIT sample and informs future efforts to retain parents in PCIT conducted in similar settings. Moreover, this study includes children with and without ASD which mirrors the child population at many clinics serving young children with behavior disorders (Maenner et al., 2020).
Findings suggest that number of intervention sessions attended was related to PCIT completion. Among all parents included in this sample, every additional intervention session increased the likelihood of PCIT graduation by 1.29%. This finding is somewhat intuitive, as a certain number of sessions are necessary to provide time to meet mastery criteria for graduation. It is likely, however, that the positive influence of time in intervention reaches an asymptote and no longer supports the potential for graduation, given our other finding that time in intervention was not related to child behavioral improvements among parents who were PCIT graduates. It is possible that early intervention attendance has a greater impact on likelihood of graduation, as parents and children form connections to the clinician and work to change unhelpful habits. Other research focused on the importance of the family–clinician relationship in PCIT supports this supposition, particularly for parents who initially present with low motivation for change (Chaffin et al., 2009).
Our finding underscores the need for work that identifies the minimum dose of PCIT needed for positive outcomes. A recent large study (N = 2,787; Lieneman et al., 2019) suggests that while the most robust outcomes regarding improvement in child behavior problems occurred when families graduated from a full course of PCIT, there were still medium to large effect sizes for families who completed at least four sessions. In contrast, families who terminated after attending fewer than four sessions only showed a small effect size. Thus, the first four sessions of PCIT, including CDI teaching and coaching sessions, may be sufficient to improve parent and child outcomes. This suggests a possible reconceptualization of early PCIT dropout from “failure” to a less robust, while still significant, intervention course. We also found a significant reduction in child behavior with a large effect size among parents who terminated early from the intervention but completed at least four sessions. This provides additional support for the notions that impactful behavior change may occur early in PCIT and that a full course of PCIT may not be necessary for all families.
We also found that baseline child behavioral issues did not affect the likelihood of PCIT graduation. Previous findings are mixed. For example, Werba and colleagues (2006) did not find that baseline level of child behavior was a significant predictor of graduation, while Bagner (2013) did. Although it is not clear why our results align with Werba and researchers (2006) yet differ from Bagner (2013), our sample was unique from these other studies in that we included children with ASD.
Unlike previous controlled trials (Danko et al., 2016; Ros et al., 2016), homework completion (i.e., between-session skill practice) also did not predict successful intervention completion in this community-based sample. However, our finding that homework completion did not relate to a decrease in child symptomology was similar to Stokes and colleagues’ (2016) finding. This suggests that homework completion may not be as crucial in PCIT as some have indicated, at least in certain populations. Stokes and colleagues hypothesized that potential reasons for their finding were including only families who had successfully completed PCIT and the inclusion of some families with less possible room for improvement due to subclinical pre-intervention ECBI scores. Additional research is needed to determine if there is a threshold for the minimum amount of homework completion needed to successfully complete PCIT.
Other findings from our study suggest that parents with low and high pre-intervention tolerance for their children’s behavioral issues did not differ on several factors central to PCIT, including homework completion and post-intervention behavioral tolerance and outcomes. These findings were somewhat surprising, given the previous literature suggesting a relationship between parent stress and child externalizing behavior problems (e.g., Jones et al., 2017). Future research and practice should include a validated measure of parent stress to clarify the associations between parent stress and child behavior following PCIT participation and the relationship between parent stress and tolerance for their children’s behaviors. Other possible reasons for the null relation between parental behavioral tolerance, homework, and child outcomes include the makeup of the sample. It is possible that this particular community-based sample had resources that helped them to manage other potential life stressors, such that they could focus primarily on practicing PCIT skills and addressing child behavioral concerns. It is also important to note that findings related to homework completion may have been influenced by the variability in homework completion rates between participants.
When we examined reasons for early dropout from PCIT, those most commonly cited were a need for different/more intensive services, parent stress beyond the parent–child relationship, disagreement with the intervention philosophy, and PCIT being deemed inappropriate for the child due to co-occurring conditions. The common reasons for dropout were generally similar to those found in previous literature (e.g., Abrahamse et al., 2016; Fernandez & Eyberg, 2009). Themes also align with factors deemed relevant in intervention outcomes for children with ASD, who comprised nearly half of our sample. Relevant factors include negative effects of parent stress on child outcomes (Shalev et al., 2020), the importance of assuring family buy-in to intervention (Wainer et al., 2017), and complexities in designing effective interventions for children with ASD who experience cumulative impairments from co-occurring disorders (Antshel et al., 2011). Importantly, families who terminated early from PCIT after four or more sessions reported meaningful and significant reductions in behavioral concerns. Thus, it is possible that at least some of these families experienced satisfactory improvements in their children’s behavior which led to the decision to prioritize other needs and/or intervention modalities.
Future research should examine whether parent stress, as measured by a standardized and validated measure, relates to successful PCIT completion and behavioral outcomes in a similar community sample since parent stress was the most malleable of the commonly cited reasons for dropout. Indeed, research has indicated that highly stressed parents of children with behavior problems may benefit not only from behavioral parent training but also from adult stress reduction interventions, such as mindfulness training (Jones et al., 2017). While behavioral parent training, such as PCIT, has been shown to help decrease parent stress (e.g., Cooley et al., 2014; Thomas et al., 2017), this is usually secondary to—and perhaps a result of—the main focus on decreasing child behavior problems. Finally, because the most common reason for early dropout in this sample was need for different or more intensive services, clinicians providing PCIT may consider use of thorough screening processes that consider a wider range of intervention options (e.g., applied behavior analysis (ABA) for some children with autism) to determine the best course of action.
Limitations
Limitations of this study should be acknowledged. First, data regarding racial/ethnic and SES of participants were not available as these data were not collected in standard practice at the clinic. This information would be helpful in examining whether cultural or demographic factors influenced any of the outcome variables. For example, Fernandez and Eyberg’s (2009) findings demonstrated that lower SES may be related to attrition from PCIT. Another limitation was our small sample size, which reduces statistical power and generalizability. It is also possible that clinicians’ impressions regarding reasons for client attrition were biased based on their work with clients, thus reasons for attrition should be interpreted with this context in mind. Future studies and practice should incorporate a standardized exit survey for parents. In addition, specific background information about clinicians, such as years of experience and training were unavailable and are all factors that may have influenced their responses.
This sample’s large percentage of children diagnosed with ASD may be a limitation to the study. While this is a strength given the dearth of research including children with ASD in community-based research on PCIT, parent stress for caregivers of children with ASD is often higher than in other diagnostic populations (Dillenburger et al., 2010). The unique challenges of a high number of participants parenting children with ASD could have influenced results. Important new research supports the efficacy of PCIT with children on the autism spectrum (McNeil et al., 2018). Future research should continue to examine the differences in outcome and engagement between parents of children with and without ASD in a community sample.
Implications for Practice
Findings from this study inform practice and efforts to retain families in PCIT services. Our findings suggest that baseline behavioral concerns and parents’ homework completion during PCIT may be less indicative of eventual graduation from PCIT, while session attendance, at least to some degree, is related to graduation. Taken with other research indicating positive effects for families who attended at least four sessions of PCIT (Lieneman et al., 2019), findings suggest that clinicians should emphasize the importance of consistent session attendance to families to facilitate positive intervention effects. To guard against early dropout, clinicians should be sure to conduct a comprehensive pre-intervention screening process and consider several possible therapeutic options to ensure that PCIT is the best intervention for the child and family based on presenting problems, comorbidities, and the root causes of behavioral concerns. Furthermore, parent stress should be addressed to guard against early dropout. Moreover, PCIT clinicians should consider adding a well validated parent stress scale, such as the Parenting Stress Index, Fourth Edition (PSI-4; Abidin, 2012), as a pre- and post-intervention measure to more formally evaluate PCIT’s effectiveness relating to stress. Clinicians may also consider adding to their post-intervention battery an acceptability or social validity measure to better understand what parents thought of their experience with PCIT including what went well and what might have been done differently. Such information could be valuable to understand the client experience in cases of both full intervention completion and early termination.
Clinicians may consider incorporating additional techniques, such as mindfulness, to aid parents in managing stressors both within and outside of the parent–child relationship to further facilitate continued attendance and positive PCIT effects. Finally, our results underscore the potential for PCIT to address child behavioral problems across a range of severity levels. Importantly, PCIT may be a suitable intervention for a family even if parents present with different perspectives about the severity of the child’s behavioral issues. Together, our findings emphasize the applicability of PCIT across children with varying levels of behavioral needs, inform efforts to retain families in the intervention, and stress the importance of session attendance to facilitate positive outcomes.
Footnotes
Authors’ Note
E.W.N. does not have any interests or activities that may have influenced the research; K.M.R. is an independently contracted psychologist for Sunfield Center for Autism, ADHD, and Behavioral Health; R.P. is an independently contracted social worker for Sunfield Center for Autism, ADHD, and Behavioral Health; and S.N. is the Owner and Clinical Director of Sunfield Center for Autism, ADHD, and Behavioral Health.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
