“Fostering” Effective Foster Parent Training Programs : Parent-Child Interaction Therapy Adaptations for the Child Welfare Setting

Foster parents face considerable challenges in caring for children in the child welfare system, many of whom have significant behavioral difficulties [1]. Foster parents often lack the training and support needed to manage these externalizing behaviors, which contribute to parenting stress and are highly predictive of placement breakdowns [2, 3]. Although child welfare agencies provide foster parents with pre-service training experiences, they often lack the capacity and financial resources to implement gold standard, evidence-based interventions that address child behavior difficulties. Parent-Child Interaction Therapy (PCIT) has been well-established as an empirically supported treatment for disruptive behavior, yet standard delivery of PCIT to children in the foster care system is often impractical due to time, financial, childcare, and personnel constraints. Adaptations of PCIT for the foster care setting may remove some of these barriers to treatment. These adaptations have typically retained the parent-coaching principles inherent to PCIT but replaced the traditional 12to 20-week format with a shorter, less intensive treatment regimen in order to maintain feasibility within the child welfare context. Preliminary findings from studies using abbreviated formats of PCIT suggest effectiveness of such adaptations in reducing externalizing behavior in foster children and maintaining behavioral improvements several months after the end of the treatment.


INTRODUCTION
The role of foster parents in the child welfare system is multi-faceted; foster parents are responsible for meeting the physical, social-emotional, educational, and medical needs of the children in their care. These children often have histories of maltreatment or neglect that put them at risk for developing externalizing symptoms and, in many cases, psychological disorders [1 -4]. Data from a national survey of 2,813 foster children under the age of 6 indicated conduct problems to be the most common areas of difficulty, more so than cognitive, communication, social, or adaptive difficulties [1]. Previous studies have shown that disruptive behavior is associated with adverse outcomes for foster children, including greater placement instability [2,3] and lower rates of reunification with biological parents [5]. Many foster parents feel illequipped to handle severe conduct problems, which ultimately contribute to parenting stress and caregiver burden [6]. In fact, foster parents have identified the management of challenging behavior as one of their greatest training needs [7] and cited the inability to handle a foster child's behavior as a major consideration for ending a placement [8]. Thus, foster parents align with the previous findings demonstrated in Tennessee [11], wherein 43 to 50% of foster children displayed clinically significant behavior problems on CBCL narrow-band subscales (e.g., attention problems, aggressive behavior), broad-band scales (i.e., internalizing and externalizing), or Total Problems scale [12]. Heflinger,Simpkins, examined externalizing behaviors among foster children and analyzed data from 105 children (ages 2-18) in state custody through the Children's Problem Outcome Review Team. Heflinger et al. [14] utilized the CBCL to provide information about problems at a global level (i.e., Total Problems score), as well as two broad-bands of behavior: Internalizing Problems and Externalizing Problems [13]. At the global level, 33% of children in foster care reported having significant behavior problems; moreover, 31% of foster children demonstrated externalizing problems at either borderline or clinical levels [14]. Specifically, 20% were reported to have externalizing problems within the clinical range of severity, which is similar to children already receiving treatment in clinical settings [14].

Maltreatment
According to Clausen et al. [12] there are two reasons children in foster care are at higher risk for mental health problems: (1) child maltreatment and (2) adjustment difficulties inherent in foster care. Maltreatment negatively impacts children across multiple domains, such as increased disruptive behavior problems, social withdrawal, and difficulty developing secure attachments [15 -17]. Externalizing behavior problems often present as core symptoms in maltreated children; that is, externalizing behavior (e.g., oppositionality) is highly correlated with childhood physical abuse [18]. Maltreated children in foster care may benefit most from an intervention that aims to decrease externalizing behavior, provide foster caregivers with safe discipline strategies, and enhance the foster caregiver-child relationship.

Placement Stability
Behavior problems in foster children are associated with placement instability [19]. In a study examining the placement experiences of 184 foster children, Palmer [20] demonstrated that children's problem behavior was a stronger predictor of placement stability than the child's gender, parents' preparation of the child for separation, or social workers' training in separation. Newton,Litrownik,and Landsverk [21] conducted a prospective study examining the relationship between change in placement and problem behavior in foster children. Results from this study were consistent with Palmer's [20] earlier findings: externalizing behavior at entry into foster care was found to be the strongest predictor of placement changes [21]. This finding reinforces the notion that children entering the foster care system with high rates of externalizing behavior are less likely to have successful, stable placements.
Highlighting the need for foster children to find a permanent placement, Newton et al. [21] found that multiple placement disruptions can aggravate children's mental health. For foster children who were initially within normal limits on the CBCL, the number of placement breakdowns predicted increased internalizing and externalizing behaviors. While Palmer [20] asserted that foster care workers are already making strong efforts to help children modify their behavior, foster children may benefit from caregiver-child focused interventions aimed at reducing disruptive behavior and strengthening the caregiver-child bond.

FOSTER PARENT TRAINING PROGRAMS
Although foster parents receive mandatory pre-service training, the content of these programs varies widely by state and often does not include training in parenting skills that specifically target negative child behavior [22]. Two of the most commonly used pre-service training curricula are Model Approach to Partnerships in Parenting Group Preparation and Selection of Foster and/or Adoptive Families (MAPP/GPS) and Foster Parent Resources for Information, Development, and Education (PRIDE). Both models seek to orient potential foster parents to their roles and responsibilities as foster parents and members of the child welfare team; however, neither program focuses on teaching specialized skills to address challenging child behavior or traumatic stress [22]. Despite their widespread adoption by numerous child welfare agencies across the country, the MAPP and PRIDE programs lack empirical study and support [23]. There is emerging evidence that the PRIDE program can increase parent knowledge of the program curriculum [24] and maintain these gains 18 months after the end of the program [25], but there is no evidence that the program improves behavioral outcomes for foster children. Evaluations of the MAPP/GPS program found that the program did not sufficiently prepare foster parents for their roles [26] nor did it prepare them to address behavior challenges in foster children [27].
Several meta-analyses and literature reviews on parenting interventions in foster care have been conducted in recent years and have demonstrated positive parent and child outcomes [9, 28 -32]. Training programs used in the foster care setting are diverse in theoretical framework (e.g., behavioral, attachmentbased, cognitive-behavioral, psychodynamic) and service delivery methods (e.g., group training, wraparound services, phone consultations, home/clinic-based) [28]. Uretsky and Hoffman [31] reviewed 11 studies on group-based foster parent training that evaluated 1 of 4 models: Incredible Years (IY), Keeping Foster and Kin Parents Supported and Trained (KEEP), Middle School Success program, and Cognitive-Behavioral Parent Training (CBPT). Two of these interventions (IY and KEEP) were adaptations of existing and wellestablished parenting interventions. The meta-analysis showed significant decreases in child problem behavior across all studies, with treatment effects maintained for studies that included three or more time points [31]. Several of the studies reviewed also indicated decreases in parental depression and stress and increases in parenting skills. Solomon et al. [30] included 16 studies in their meta-analysis evaluating the efficacy of foster parent training in reducing child behavior problems. A number of different interventions were included in this meta-analysis: KEEP, adaptations of IY, Attachment and Behavioral Catch-Up (ABC), CBPT, and the Love and Logic Program, among others. Solomon and colleagues [30] found that foster parents in the treatment groups reported fewer child behavior problems (i.e., small, but significant mean effect size) and greater parenting skills and knowledge (moderate, significant mean effect size) than those in the control groups. This meta-analysis highlighted the importance of integrating empirically supported treatments for maladaptive child behavior into foster parent training. Schoemaker et al. [29] conducted a series of eight meta-analyses and found significant improvements in child behavior problems, as well as parental sensitivity, dysfunctional discipline, parenting knowledge, and parenting stress. However, no significant changes were observed in placement stability, children's cortisol levels, or attachment security following the intervention [29]. Certain elements of foster parent training programs seem to be essential to improving outcomes for both child and caregiver. A meta-analysis [32] highlighted the importance of enhancing the foster child-foster parent relationship, improving attachment quality for young children, and targeting behavior problems and symptoms of stress. In a review of foster parent training programs, Rork and McNeil [9] recommended the use of empirically supported treatments, especially those that are behaviorally based, to guide interventions for foster parents.

PARENT-CHILD INTERACTION THERAPY
One parent training program that encompasses these key components (e.g., behaviorally based, targets externalizing behaviors, enhances the parent-child relationship) is Parent-Child Interaction Therapy (PCIT). PCIT is an evidence-based, manualized treatment for behavioral symptoms in children between the ages of two and seven years [33]. This behavioral parent-training program was first designed to treat disruptive behavior disorders through weekly, one-hour sessions involving the child and his or her caregivers [34]. Today, PCIT is indicated in treating symptoms of childhood anxiety [35], depression [36], Attention-Deficit/Hyperactivity Disorder (ADHD) [37], Autism Spectrum Disorder (ASD) [38], maltreatment [39], and other behavioral health concerns. PCIT is unique in that it involves live therapist coaching of the caregiver through an earpiece and two-way mirror, incorporates weekly behavioral assessment measures, and depends upon skill-based mastery criteria for progression through treatment.

PCIT Basics
PCIT is made up of two treatment phases [40]. In the first phase, Child-Directed Interaction (CDI), a therapist spends one session, teaching caregiver(s) foundational play therapy skills. The caregiver is taught to use "PRIDE" skills-praise, reflect, imitate, describe, enjoy-during one-on-one "special playtime" with the child during weekly therapy sessions and daily, 5minute practice sessions in the home. Caregivers also learn to avoid questions, commands, and negative talk, also known as the "don't" skills, during CDI. Using a microphone and earpiece, a PCIT therapist provides live coaching to caregivers from behind a two-way mirror as caregivers engage in play with their child. Coaching involves a parallel process wherein therapists employ the same behavioral principles (e.g., praising appropriate use of skills, avoiding negative talk) to improve caregivers' skills as caregivers are expected to use with their children. CDI coaching sessions continue until caregiver skills meet mastery criteria (i.e., 10 praises, 10 reflections, 10 descriptions, and 3 or fewer "don't skills" during a 5-minute coding period). The use of mastery criteria is intended to capitalize on the idea of overlearning skills; caregivers who demonstrate high ratios of positive to negative parenting skills are more likely to retain these habits outside of sessions [33]. CDI mastery typically occurs after about six to eight sessions of PCIT.
Following CDI mastery, the therapist teaches caregiver(s) a set of disciplinary strategies focused on compliance training. This second phase of treatment, Parent-Directed Interaction (PDI), aims to hone caregiver skills in giving effective commands (i.e., commands that are singular, specific, direct, and positively stated) and delivering consistent, predictable consequences (i.e., labeled praise for compliance and time-out from positive reinforcement for noncompliance). Again, caregivers are coached in PDI skills during therapy sessions and complete daily at-home practice of both CDI and PDI skills. Mastery of the PDI phase occurs when the caregiver gives 75% or greater effective commands and 75% or greater correct subsequent follow-through, during a 5-minute coding session. Caregivers often achieve PDI mastery in six to ten sessions.
Following mastery of CDI and PDI, PCIT therapists work ideographically with families to achieve a variety of remaining goals such as decreasing sibling conflict, improving public behavior, and generalizing behavioral gains to the classroom. Throughout PCIT, caregivers also complete a weekly selfreport measure of child behavior problems called the Eyberg Child Behavior Inventory (ECBI) [41]. In addition to CDI and PDI skills mastery, families must report child behavior problems within normal limits on the ECBI and confidence in managing their child's behavior in order to graduate from PCIT [33]. The typical course of PCIT lasts for about twelve to twenty sessions from baseline assessment to graduation.

Empirical Support for PCIT
For more than four decades, researchers have investigated the impact of PCIT through efficacy and effectiveness trials in a variety of contexts. See Lieneman,Brabson,Highlander,Wallace,and McNeil [42] for a recent comprehensive literature review. Effect sizes for pre-to post-treatment decreases in child behavior problems as reported on the ECBI are routinely measured in the "very large" to "huge" range in samples from community-based treatment settings (e.g., d = 2.30 [43], d = 1.65 [44]). See Cohen [45] and Sawilowsky [46] for more information on effect size classifications. Families report maintenance of treatment gains at one-, two-, three-, and sixyear follow-up assessments subsequent to PCIT [47 -49]. Drawing from a meta-analysis of common effective components of behavioral parent training programs, PCIT's large effects can be attributed in part to its core features of: (1) increasing positive interactions, (2) in-session practice, (3) emotional communication, and (4) time-out of from positive reinforcement [50,51]. The positive effects of PCIT have been measured through caregiver-based outcomes as well. A 2007 meta-analysis of PCIT research demonstrated improvements across studies in positive and negative parenting practices following treatment [52]. Improvements in caregiver emotion regulation [53], as well as decreases in parenting stress [54 -56] and risk for future maltreatment [39,57] are also evident.

PCIT IN THE CHILD WELFARE CONTEXT
Several studies have examined the effectiveness of PCIT for foster child-foster parent dyads [75,76]. A case study on a foster mother and her four-year-old foster son demonstrated significant reductions in parenting stress and child behavior problems that were no longer in the clinical range on the ECBI and CBCL [75]. Observational data from the Dyadic Parent-Child Interaction Coding System (DPICS) [77] showed an increase in the foster mother's use of praise and behavioral descriptions and a decrease in questions and commands. Fricker-Elhai, Ruggiero, and Smith [78] conducted a clinical case study of PCIT with two biological siblings with significant maltreatment histories and their foster parents. PCIT was effective in bringing the ECBI Problem and Intensity scores below clinical range for both children and to normal limits on the Teacher Rating Form (TRF) [13] completed by the children's teachers, except for an elevated Aggressive Behavior score for one sibling. In a study comparing nonabusive biological parent-child dyads to foster parent-foster child dyads, Timmer and colleagues [76] found that PCIT yielded strong treatment effects on child problem behavior and caregiver distress, but no major differences were found between the two parent groups.

Feasibility
Despite the overwhelming evidence that PCIT is effective in improving behavioral outcomes for children with challenging behavior or histories of maltreatment, the reality is that PCIT may not be feasible in some child welfare contexts due to a number of logistical and financial restraints. PCIT may be cost-prohibitive for some child welfare agencies, with startup costs estimated at $14,000 and treatment costs estimated at $1,000 per client [79]. PCIT also requires delivery by trained mental health providers with masters' degrees, which for many agencies may require training of staff and restructuring of personnel [80]. High attrition rates pose another problem for child welfare agencies looking to use PCIT as a model for preservice training. Standard PCIT has an estimated 33% dropout rate [81], and non-kin foster parents have higher attrition rates than kinship foster parents [82]. Mersky and colleagues [83] cited the intensity and duration of treatment as a major barrier to the implementation of PCIT in the foster care setting. Foster parents typically have multiple children in their care and may be less inclined to participate in an intensive 12-to 20-week program focused on a single child. The recognition of these barriers to the treatment has served as a catalyst for the development of PCIT adaptations for the child welfare system.

PCIT with in-home Coaching
One such adaptation maintained all the components of standard PCIT but added one hour per week of in-home coaching [84]. The rationale for the additional in-home coaching practice was to promote the generalizability of skills and allow the caregiver to practice skills with therapist support in a real-world setting, complete with siblings, other adults, and environmental distractions. Although the study did not focus on foster parents, non-kin foster parents were included in the sample, along with biological parents and kin caregivers. Participants were randomly assigned to either receive additional in-home PCIT coaching or in-home social support services. Both PCIT with and without in-home coaching yielded similar improvements in child and caregiver behavior; caregivers who received additional in-home coaching sessions did not fare better in terms of skill acquisition, quality of performance by the last session, or speed in completing treatment than those in the social support group. However, those in the in-home coaching group used significantly more positive verbalizations, had lower parental stress, and experienced child behavior as less problematic (as measured by the ECBI Problem Scale).

Two-day PCIT Workshop
McNeil, Herschell, Gurwitch, and Clemens-Mower [85] conducted an exploratory single-group study on a condensed version of PCIT with 30 foster child-foster parent dyads. Instead of the individualized, multi-week format, foster parents in this study attended a two-day group training. On the first day of training, foster parents attended a didactic session without their foster children in order to learn CDI skills and engage in discussions and role-plays to enhance skill acquisition. On the second day of training, foster parents brought their foster child with them and were trained in PDI skills. They were then given an opportunity to practice discipline skills with their children while a therapist coached from behind a one-way mirror. Other foster parents watched these parent-child interactions from the observation room, and childcare was provided any time a parent-child dyad was not being coached. The workshop ended with the development of a plan for each foster parent to roll out PCIT into the home setting.
In this study, the key components of PCIT were maintained (e.g., CDI and PDI skills, live caregiver coaching). However, a number of elements from the standard format were excluded in order to adapt PCIT to the child welfare context: at-home practice, coding of CDI and PDI skills, progress based on mastery, and the use of a back-up room in the time-out procedure (i.e., foster parents had the option of using a back-up room or another consequence such as restriction of privilege). This modified format of PCIT presented a number of advantages for the foster parent population. The two-day workshop significantly reduced the time commitment for caregivers and the cost of training for child welfare agencies, thereby increasing the feasibility of implementation. The group-based format lent itself to the discussion, observational learning, and social support among the foster parent participants.
After the 2-day workshop, 27 of the 30 participants completed a 1-month follow-up, and 8 participants completed a 5-month follow-up. Although the average pre-treatment ECBI Problem and Intensity Scale scores were measured in the clinical range, these scores fell in the average range one month after treatment. Interview data collected at the 1-month followup indicated that 80% of foster parents were using skills acquired during the workshop and 68% were satisfied with the time-out procedure. Scores on a treatment satisfaction survey were high immediately following the training but saw a small, yet statistically significant drop in satisfaction ratings at follow-up. Parents who completed the 5-month follow-up continued to report ECBI Intensity scores below clinical cutoffs and a decline in Problem scores, suggesting that these parents were better able to manage their children's behavior. The findings from this single-group study were promising in that they offered preliminary support for the use of a two-day workshop format for foster parents and their children. The methodological weaknesses of this study (e.g., lack of a control group, small sample size) prompted further evaluation of this abbreviated format.

Project Connect
Project Connect is a PCIT adaptation that was developed as an extension of the McNeil et al. [85] study in collaboration with two community partner agencies [86]. The group format was retained, but biweekly phone consultations and homework were added to support the transfer of skills to the home environment. Observational data using the DPICS-III [87] was also added to track changes in parenting behavior from baseline to eight weeks after baseline. The authors examined two variations of the model: (1) brief PCIT, a two-day workshop followed by six phone consultations conducted over eight weeks and (2) extended PCIT, a three-day workshop followed by four phone consultations conducted over six weeks [80,86]. Trainings were delivered by PCIT-trained therapists and took place on Saturdays for approximately eight hours with a group of six to eight families. Phone consultations were 15 to 20 minutes in length and typically occurred in the evenings. Additionally, childcare was provided for all Saturday sessions to remove barriers to participation, namely for foster parents with multiple children in their care.
The schedule of the first Saturday workshop began with didactic instruction on CDI skills, followed by role-playing to practice these skills. The therapists then coached the foster parents as they practiced CDI skills on one another. The bulk of the training was dedicated to live coaching; there were 3, 45minute live coaching sessions, and all parents received at least 2 coaching sessions. Foster parents not actively engaged in live coaching observed another parent's live coaching session from an observation room. The session ended with a discussion of questions and homework, and phone consultations were scheduled with each family. The second workshop followed the same format as the first but with PDI skills instead of CDI skills. For the three-day workshop, foster parents received an additional Saturday session at the eight-week mark. Phone consultations allowed providers to follow up with families on homework, answer any questions, and ensure foster parents were implementing PCIT skills in the home.
Families were randomly assigned to the two-day workshop condition, three-day workshop condition, or a waitlist control group. Rates of attrition were similar across the 3 groups (i.e., approximately 25%), a relatively low rate of attrition compared to other PCIT studies [56,81]. The fact that the attrition rate did not differ between the brief and extended conditions indicated that the booster session did not negatively impact foster parent engagement [86]. Mersky and colleagues [80,83] reported more detailed findings from this investigation. Foster parents in the two treatment conditions reported significant decreases in parenting stress on the Parenting Stress Index-Short Form (PSI-SF) [88] and improvements in parenting behavior [83]. Foster parents improved on measures of labeled praise, positive parenting behavior, and negative parenting behavior, but no differences between groups were found on the use of negative talk. There were no significant outcome differences between the brief PCIT and extended PCIT versions, which supports the use of the more time-and costeffective two-day model over the three-day model [83]. Both interventions yielded significant changes in child behavioral outcomes [80]. Scores on the ECBI Intensity and Problem Scales decreased for all 3 groups, and all groups had ECBI Problem scores below the clinical cutoff at 14 weeks after the baseline assessment. However, only the two treatment groups saw scores drop below the clinical cutoff on the ECBI Intensity Scale. Similarly, scores on the CBCL Externalizing and Internalizing Scales significantly decreased for all three groups, but only the two intervention groups had scores that fell below the clinical threshold post-treatment. Upon further analysis, Mersky et al. [80] found that those in the extended condition experienced further improvement in internalizing and externalizing behavior after the 8-week mark, but this was not the case for those in the brief condition. Although these results must be bolstered by further study, they suggest that a booster session may help with the maintenance of gains seen in foster parent and foster child behavior.

Brief PCIT
An exploratory study based on the abbreviated PCIT models for foster parents utilized a 5-to 7-session, clinicbased, individualized format [89]. Brief PCIT maintained more of the standard format of PCIT than the group-based formats, and thus provided foster parents with more opportunities for individualized coaching. The first session was a typical intake/observation session, followed by 2 sessions dedicated to CDI teaching and coaching. An optional CDI coach session was added if needed. Foster parents then received 2 sessions of PDI teaching and coaching with the option of a third PDI session if needed. Preliminary findings from a sample of six participants demonstrated a decrease in scores on both ECBI Intensity and Problem Scales for all but one participant. Two of the participants had clinically significantly pre-treatment scores on both ECBI scales that were no longer in the clinical range post-treatment. A qualitative study on Brief PCIT in the child welfare context identified additional barriers by asking foster parents and providers about their experiences with the intervention [90]. In this particular study, PCIT providers added elements to Brief PCIT in order to address child traumatic stress (e.g., psychoeducation for parents on trauma and traumatic stress and physiological/cognitive relaxation strategies). Overall, foster parents cited minimal challenges to participating in treatment but stated that the time commitment was difficult to manage. Foster parents deemed the therapeutic alliance with their PCIT provider to be a tremendous source of support that promoted success and retention. They also found the integration of trauma-informed treatment to be helpful in understanding their foster children's behavior and emotions. The practitioners, on the other hand, cited two main barriers to treatment: (1) foster parents' perceived stigma associated with participating in treatment and (2) foster parents' tendency to overreport child behavior difficulties. To problem-solve around these issues, practitioners formed small groups or "learning collaboratives" comprised of approximately seven therapists to discuss challenges, brainstorm possible solutions, and promote fidelity to the model. The work of Blair et al. [89,90] has introduced yet another abbreviated adaptation of PCIT with its own modifications for both foster parents and their providers.

CONCLUSION
Preliminary studies on PCIT adaptations for foster parents have demonstrated improvements in child behavior problems, parental stress, and positive parenting behavior. There is empirical support for an abbreviated version of PCIT for the child welfare context and the addition of in-home coaching or follow-up phone consultations. The abbreviated models of PCIT reduce many of the barriers to treatment and make this adaptation a more feasible option for both foster parents and child welfare agencies. However, further study is needed to disaggregate the effects of these adaptations and determine which elements are most effective in reducing problem behavior, parental stress, and negative parenting behavior. Although there is some evidence that a booster session may be helpful in solidifying parental knowledge and use of skills, more research needs to be conducted on dosage effects with more follow-up assessments to determine if effects wear off over time. There is also evidence that some behavior (e.g., negative talk0 [83] may be harder to change with only a twoor three-day workshop and may require more intensive intervention to see significant improvement. Studying individual components of the adaptation (e.g., phone consultations, observational learning, group discussions, booster sessions) will be helpful in building a more robust parenting intervention that child welfare agencies can readily and widely adopt.

CONSENT FOR PUBLICATION
Not applicable.

FUNDING
None.