Behaviour Research and Therapy 56 (2014) 30e38
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Family-based exposure and response prevention therapy for preschool-aged children with obsessive-compulsive disorder: A pilot randomized controlled trial Adam B. Lewin a, b, c, *, Jennifer M. Park b, Anna M. Jones b, Erika A. Crawford a, Alessandro S. De Nadai b, Jessie Menzel b, Elysse B. Arnold a, b, Tanya K. Murphy a, c, Eric A. Storch a, b, c a
Department of Pediatrics, University of South Florida, USA Department of Psychology, University of South Florida, USA c Department of Psychiatry & Behavioral Neurosciences, University of South Florida, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 September 2013 Received in revised form 10 February 2014 Accepted 11 February 2014 Available online 2 March 2014
Aims: To examine the feasibility, acceptability and preliminary efficacy of family-based exposure/ response prevention therapy (E/RP) versus treatment as usual (TAU) in a cohort of very young children with early onset obsessive-compulsive disorder (OCD). Methods: Thirty-one children ages 3e8 years (M ¼ 5.8 years) with a primary diagnosis of OCD were randomized to E/RP or TAU. The E/RP condition received 12 sessions of family-based E/RP twice weekly over 6 weeks. Families were assessed at baseline, post-treatment, 1-month and 3-month follow up. The Children’s Yale Brown Obsessive Compulsive Scale and Clinical Global Impression served as primary outcome measures. Results: A large group effect emerged in favor of the E/RP group (d ¼ 1.69). Sixty-five percent of the E/RP group was considered treatment responders as compared to 7% in the TAU group. Symptom remission was achieved in 35.2% of the E/RP group and 0% of the TAU group. There was no attrition and satisfaction was high; gains were maintained at 3 months. Conclusions: Even amongst children as young as 3 years, developmentally tailored E/RP is efficacious and well-tolerated in reducing OCD symptoms. Key adaptations for younger children include extensive parent involvement, targeting family accommodation, and frequent meetings while delivering a full course of E/RP. ClinicalTrials.gov Identifier: NCT01447966 http://clinicaltrials.gov/ct2/show/NCT01447966?term¼ocdþ andþstþpetersburg&rank¼1 Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: OCD Exposure-response prevention Very young children Accommodation Cognitive therapy Preschool
Pediatric obsessive compulsive disorder (OCD) is a chronic and disabling illness with a point-prevalence of 1e2% (Zohar, 1999). Age of symptom onset varies but generally occurs during key developmental periods such as pre-pubescence or adolescence/early adulthood, with up to 80% of adults reporting emergence of symptoms prior to the age of 18 years (AACAP, 2012). As pediatric OCD was previously believed to be rare amongst very young children, most
Abbreviations: E/RP, exposure and response prevention; TAU, treatment as usual; OCD, obsessive compulsive disorder. * Corresponding author. USF Dept. of Pediatrics, Rothman Center for Neuropsychiatry, 880 6th Street South Suite 460 Box 7523, Child Development & Rehabilitation Center, St. Petersburg, FL 33701, USA. E-mail address: [email protected] (A.B. Lewin). http://dx.doi.org/10.1016/j.brat.2014.02.001 0005-7967/Ó 2014 Elsevier Ltd. All rights reserved.
studies focus primarily on school-aged children above the age of 8 years. However, reports have noted OCD presentations amongst those as young as 2 years (Coskun & Zoroglu, 2009). While reports on early childhood OCD (i.e., onset of symptoms prior to age of 9 years e.g., Freeman et al., 2008) are scant, studies have noted that symptoms are associated with positive family history, increased distress, longer duration of illness, and higher rates of comorbid tics (Coskun & Zoroglu, 2009; Garcia et al., 2009; Nakatani et al., 2011). Those affected by early childhood OCD may be at particular risk as OCD symptoms may impede or hinder normal development during key milestone years (Flessner, Garcia, & Freeman, 2013; Valderhaug & Ivarsson, 2005). Consequently, early and aggressive intervention is paramount for younger children with OCD (Hirshfeld-Becker & Biederman, 2002; Lewin, Park, & Storch, 2013).
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
Early childhood OCD presents with a number of unique developmental factors, which should be addressed during the course of treatment. Children are generally enmeshed into the dynamics of the family and often do not have the ability to disengage with family related stressors such as marital dysfunction and parental mental health (Freeman et al., 2003). These factors may strongly influence the mental health of the child and affect treatment progress, as well as maintenance of gains (Kazdin, 1995; Kazdin & Weisz, 1998). Family accommodation, which refers to the involvement of family members in the child’s OCD symptoms, is highly prevalent in children and adolescents with OCD (Storch, Geffken, Merlo, Jacob, et al., 2007). Family members may modify their family routine to allow children to avoid anxiety-provoking situations, or engage in rituals such as providing reassurance, assisting in mealtime, bedtime, or toileting rituals (Peris et al., 2008; Storch, Geffken, Merlo, Jacob, et al., 2007). Because of the involvement of family members in OCD symptoms, as well as the influence of family dynamics upon the course of the treatment, it is imperative that family members are involved in the treatment process. There are two evidence-based treatments for pediatric OCD: serotonin reuptake inhibitors (SRIs) and exposure and response prevention (E/RP) (Lewin & Piacentini, 2009; POTS, 2004). Pooled effects suggest that E/RP has an advantage over SRI treatment alone (Abramowitz, Whiteside, & Deacon, 2005) leading to practice parameter recommendations that children receive E/RP alone in those with mild to moderate severity or with SRI therapy in more severe cases (AACAP, 2012). While efficacious, pharmacotherapy may have serious adverse events, somatic side effects (e.g., poor sleep, gastrointestinal complaints), and behavioral activation with increased risk at younger ages (Murphy, Segarra, Storch, & Goodman, 2008) even at small doses. For example, in an open trial of fluoxetine in six very young children with OCD, 50% developed significant medication-related side effects including two with behavioral disinhibition (Coskun & Zoroglu, 2009). Although a recent case series suggests fluoxetine produced clinical improvement in four treatment-refractory preschool aged youth with OCD (Ercan, Kandulu, & Akyol Ardic, 2012), medication is often undesirable to parents of young children, especially in cases of new onset/initial treatment. Conversely, E/RP for pediatric OCD is a well-tolerated and acceptable option for parents (Patel & Simpson, 2010). Welldesigned studies have had positive outcomes, particularly amongst children ages 7e17 years, with response rates ranging from 39% to 90% (Piacentini et al., 2011; POTS, 2004; Storch, Bussing, et al., 2013; Storch, Geffken, Merlo, Mann, et al., 2007). However, despite OCD being relatively common among very young children (Coskun, Zoroglu, & Ozturk, 2012; Flessner et al., 2013; Garcia et al., 2009) and practice parameters recommending E/RP as the first line intervention for pre-latency youth with OCD (AACAP, 2012; Gleason et al., 2007), strikingly few children below the age of 8 years have been included in the extant controlled research (Flessner et al., 2013). Indeed, Freeman et al. (2008) noted that within the Pediatric OCD Treatment Study I and Study II (POTS), which are considered the largest pediatric OCD treatment studies to date, only 10e16% were ages 7 or 8 years (POTS, 2004). Following the POTS trials, only one published trial to date has included youth under the age of 7 years (Freeman et al., 2008) although a larger trial for youth ages 5e8 years is underway (Freeman et al., 2012). In Freeman et al. (2008), a 12-session family-based E/RP course was found efficacious for treating youth ages 5e8 years with OCD. Completer analysis showed a large group difference (d ¼ 0.85) with greater obsessive-compulsive symptom reductions for youth receiving E/RP (n ¼ 16) versus a 12-session relaxation training control (n ¼ 15). Following the 12-session course, 69% of youth responded to E/RP compared to 20% of youth receiving relaxation
31
training. Notably, the authors implemented modifications to the E/ RP protocol to address developmental, cognitive, and socioemotional differences specific to children age 8 years and younger. Specific modifications included: (1) Inclusion of parents as coaches to foster motivation, increase adherence, and reduce accommodation; (2) Cognitive therapy components were simplified as concrete, child-specific examples and to foster allegiance in combating ‘OCD’ as a common adversary; (3) Addressing other demands associated with treating anxiety in a very young sample including child oppositionality/resistance, managing out-ofsession E/RP homework, parental tolerance of their own distress during E/RP, and related logistical issues for parents. Despite this promising work, a clear need exists for replication and extension of these findings in youth age 8 years and under with OCD to help support E/RP as efficacious for this younger cohort. Consequently, the present study is an investigation of feasibility, tolerability and preliminary efficacy of a family-based E/RP protocol designed for the youngest cohort of youth with OCD to date. Rather than an exact replication of Freeman et al. (2008), we introduced a novel protocol with several modifications given (1) the younger cohort (32% of our sample was below the range included in Freeman et al. (2008)), (2) to reduce attrition (26% in Freeman et al. (2008)) and (3) bolster responsiveness to this promising approach without (4) compromising acceptability. Briefly, key modifications included: (a) a primary focus on E/RP versus the inclusion of other treatment elements (e.g., 4e6 sessions in Freeman et al. (2008) without primary E/RP focus); (b) eliminating most cognitive procedures; (c) parent/child inclusion in all sessions; (d) twice-weekly format to expediate response; (e) tenacious targeting of family accommodation starting at session 1. Contrasts with the promising Freeman et al. (2008, 2009) protocol were as follows. Treatment utilized in the present study was refined to emphasize the primacy of E/RP; rather than initiating E/ RP at session 5 or 6, our revised protocol implemented E/RP no later than the second visit (but often during session 1). Moreover, after the first session, E/RP was the primary focus of each of the remaining sessions with even less emphasis on cognitive strategies/child focused psychoeducation (given the very young age of the targeted cohort, i.e., inclusion of 3e5 year olds). Several other modifications were implemented given the very young age range. First, both parents and youth were included in all sessions together (unlike Freeman et al. (2008) which excluded youth until session 3). While there is support for children as young as 4 years benefiting from verbally-presented information regarding exposure therapy (Scheeringa et al., 2011), emphasis was placed on parental understanding of E/RP concepts (Choate-Summers et al., 2008). Although parents desire their child to “understand” therapeutic concepts (Labouliere, Arnold, Storch & Lewin, 2014), there is little benefit in spending additional session time on child focused psychoeducation with a 3e5 year old (Choate-Summers et al., 2008), especially at the expense of reducing the time spent in E/RP. Thus, having the child present for parental psychoeducation could appease parents while preserving time dedicated for E/RP. Consequently, instead of utilizing two sessions for education, our revised protocol was more streamlined in order to maximize the dosage of E/RP. Another advantage of keeping the parent and child together in all sessions was the ability to provide singular focus on E/RP rather than separate parent and child goals in each session (Freeman & Garcia, 2009). Additionally, our study protocol addresses accommodation beginning week 1, as a mechanism for conducting exposure. Family accommodation is a hallmark of early childhood OCD (Garcia et al., 2009; Flessner et al., 2013; Freeman et al., 2003) and consequently E/RP was adapted to aggressively target accommodation. Parent engagement in child rituals and/or parent facilitation of OCD (e.g., allowing extra time, purchasing special products, making
32
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
individual meals, etc.) was typically targeted prior to asking the child to engage in E/RP (e.g., engage with a feared object or change a ritualistic routine). Targeting accommodation as the first “exposure” avoided coercion and negotiation (the parent was choosing to change behavior) while still permitting habituation to distress and progression up a hierarchy. Further, this modification was particularly appropriate for younger children (e.g., 3e5 years) who had limited capacity to understand treatment related concepts and hierarchy development (Labouliere et al., 2014). Parents left session 1 with specific instructions to begin reducing accommodation (or were directed to construct a plan to initiate during the second visit later week 1). Additionally, we introduced a twice-weekly format to accelerate the treatment (especially E/RP). Given the aggressiveness of our approach (early introduction of E/RP without implementation of supportive therapeutic/cognitive strategies), the twice-weekly content allowed for frequent therapist contact and was feasible given the age of our cohort (i.e., less difficulty with missing school). Additional details are presented in the Methods. Accordingly, the present study aimed to examine feasibility, acceptability, and preliminary efficacy via a pilot randomized controlled trial of our adapted family based E/RP intervention compared to a Treatment as Usual (TAU) control condition. We hypothesized that family-based E/RP would be a highly acceptable and well-tolerated intervention with symptom reductions similar to the Freeman et al. (2008) trial, but with decreased attrition. Secondary aims included examination of changes in family accommodation, functional impairment, and maintenance of treatment gains at 1 and 3 month follow-up. Method Participants Participants were 31 children with a primary OCD diagnosis (ages of 3e8 years; M ¼ 5.8; SD ¼ 1.6) recruited through patient flow at a university-based specialty clinic. Thirty-six percent of youth were age 5 years or younger (breakdown by age was as follows: 2 [age 3], 8 [age 4], 1 [age 5] 7 [age 6] 9 [age 7] and 4 [age 8]. Inclusion required: (1) Clinician diagnosis of primary OCD with diagnoses verified by administration of the Anxiety Disorders Interview Schedule (ADIS) e Child and Parent Versions (Silverman & Albano, 1996) clinical severity rating 4; CYBOCS 16 (or CYBOCS Compulsions scale of > 8); between ages of 3e8 years inclusive; commitment of consistent parent to attend all sessions. Exclusion criteria included: (1) the presence of psychiatric illness that contraindicated study participation, e.g., autism; (2) IQ < 85; (3) nonEnglish speaking or absence of language; (4) history of E/RP; (5) a recent change in psychotropic medication1; or (6) concurrent psychotherapy or other behavioral interventions. Children randomized to TAU were allowed to make medication changes following randomization. All children randomized to E/RP remained stable on medications throughout the study (see Table 1). Fig. 1 presents descriptive data regarding participant flow through the study. There was no attrition from either condition. Procedure The IRB approved this study, and both parents and children provided written informed consent and assent respectively prior to
1 For youth on established pharmacotherapy, 8 weeks stability at current dose was required for SSRIs and 6 weeks for other medications. For newly initiated pharmacotherapy, 12 weeks stability for SSRIs and 8 weeks for other mediations was required.
Table 1 Sample demographics.
Child sex (male) Child age (mean, SD) Parent sex (female) Parent graduated from college Living Situation Married Both parents (same) Both parents (different residence) Lives with single parent (mom) Lives with mom/stepdad Child ethnic/race White Asian/Pacific islander Latino/Latina African American Comorbid diagnoses Any anxiety disorder Separation anxiety disorder Social phobia GAD Specific phobia ADHD Oppositional defiant disorder Psychiatric medication use SSRI Guanfacine Stimulant
E/RP n (%)
TAU n (%)
c2/t value
12 5.76 16 9
10 5.86 10 7
0.003 0.16
(71) (1.52) (94) (53)
(71) (1.70) (71) (50)
1.73 13 12 1 3 1
(76) (71) (6) (18) (6)
15 1 2 1
(88) (6) (12) (6)
12 4 8 7 5 5 5
(71) (24) (47) (41) (29) (29) (29)
12 12 1 1
(86) (86) (7) (7)
0 4.08
0
0 2 (14) 0 10 4 4 4 6 8 6
(71) (29) (29) (29) (35) (57) (35)
0.003 0.10 1.11 0.53 0.61 2.43 0.61 e
0 1 (6)
0
12 (86)
4 (29) 2 (14)
2.92 1.76
enrollment. At a baseline assessment, primary diagnosis of OCD was determined by clinical interview with a board certified child psychologist (and consensus review of all records by a second experienced licensed child psychologist). Diagnoses were verified by administration of the ADIS. Following baseline assessment, qualifying participants were randomized (using a 1:1 ratio) to E/RP or TAU; parents were informed of condition immediately by the study coordinator. Post-treatment assessments were administered within 1 week following the final E/RP session or at approximately 6e7 weeks following baseline assessment in the TAU condition. Follow-up assessments were conducted at 1-month and 3-months post-treatment (balanced for both groups to preserve blindness of ratings). Procedures consistent with other child randomized studies (Piacentini et al., 2011, 2010) were taken to preserve blindness (e.g., separate supervision of raters/therapists, reminding participants of the blind before each assessment, positioning independent evaluators separate from therapist, utilizing ancillary entrance/exit to avoid breaking the blind). However, notably, a recent study demonstrated that blinding is not essential for conducting accurate ratings in clinical trials of child OCD (Lewin, Peris, De Nadai, McCracken, & Piacentini, 2012). Treatment Family-based Exposure/Response Prevention (E/RP) The E/RP intervention consisted of 12 sessions, 60 minutes in length, administered twice weekly over the course of 6 weeks. The twice-weekly format was selected to provide more frequent contact and possibly reduce attrition (26.2% of the Freeman et al. (2008) sample did not complete the 12 session protocol implemented over 14 weeks suggesting protocol changes may be necessary). The E/RP content was based on a manualized, familybased treatment for young children with OCD (Freeman & Garcia, 2009) with modifications to emphasize E/RP earlier in the protocol, hold sessions twice-weekly, and aggressively target accommodation. This treatment targets OCD symptoms as well as accommodation by family members. All treatment sessions were
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
33
Fig. 1. Study consort diagram. Note: from Schulz KF, Altman DG, Moher D, for the CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ 2010; 340:c332. For more information, visit: www.consort-statement.org.
conducted with the child and at least one primary caregiver present at all times. A consistent parent was required for every session (although the other parent/guardian was encouraged to attend when feasible). Treatment components included psychoeducation for parents and children (allying against ‘OCD’; introducing developmentally appropriate metaphors/examples), parent tools (e.g., development of a rewards program, differential reinforcement, extinction, and modeling), E/RP, and relapse prevention planning (Freeman & Garcia, 2009; Freeman et al., 2008). Given the younger age in the present sample, cognitive elements were further reduced and emphasis was placed on parent-focused aspects of the protocol (e.g., reducing accommodation, training the parent in extinctionbased treatment and need for the child to habituate to anxious triggers, and implementing E/RP independently at home) (Kendall et al., 2005; Piacentini & Bergman, 2001). The revised treatment schedule was as follows: Session 1 e introduction, parent-focused psychoeducation on E/RP, allying against OCD, hierarchy development, and parent driven E/RP directed at family accommodation behaviors; Session 2 e continued hierarchy development, targeting accommodation (ignoring)/extinction, resisting OCD; introduce/ model E/RP with the child; Session 3 e more modeling exposures, practice in-session E/RP, developing reward program (as needed); Sessions 4e11 e in session E/RP, brief developmentally appropriate
cognitive interventions (e.g., labeling “It’s only my OCD,” resisting “I don’t have to listen to [OCD],” positive self-talk “I can win this fight against OCD.”), continued reduction of parental accommodation, increased parent-led E/RP, adjusting reward program as needed; Session 12 e continued E/RP, relapse prevention, and certificate/ brief celebration. Many of the procedures described in the existing literature (Choate-Summers et al., 2008; Flessner et al., 2013; Freeman et al., 2008, 2003) for treating youth aged 5e8 years were further modified for our even younger sample. Behavioral techniques were embedded within the E/RP practice to avoid losing session time dedicated to practicing exposures (without a need for adding additional sessions). In other words, parent training principles such as differential reinforcement of other behavior/alternative behavior (DRO/DRA), extinction, response cost, and behavioral reduction procedures were implemented within the context of E/RP practice (Fernandez et al., 2006; Labouliere et al., 2014) when necessary. Further, our protocol focused on reducing parental accommodation immediately in treatment. Similar to the Freeman et al. (2008) protocol, E/RP was parent driven. However, our protocol focused on reducing accommodation as a parent driven first-step of the E/RP hierarchy. Therapists were licensed clinical psychologists or advanced clinical psychology doctoral students (who had 3 years experience
34
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
in E/RP for OCD under the supervision of the principal investigator); therapists were assigned to participants at random. Therapist training included didactics with the principal investigator, advanced guided reading of the treatment manual, and completion of at least one supervised training case of a patient under age 8 using the manual with integrity checks. Treatment integrity was ensured through weekly supervision, reviewing content checklists that corresponded to the treatment manual, and evaluation of 15% of randomly selected audio-recorded therapy sessions using an 11item, 6-point scale. There was excellent adherence to session components (25.0/30.0), therapist competence (16.2/18.0), and therapist alliance (11.2/12.0). Treatment as usual The TAU condition was designed to reflect standard practice. All participants randomized to the TAU condition were instructed to continue receiving their prior interventions as recommended by their providers (e.g., psychotherapy, social skills training, pharmacological interventions, or parenting classes) or were permitted to seek new treatment/evaluation from providers not affiliated with our treatment program. The TAU control was chosen over relaxation/ education controls implemented in previous studies (e.g., Freeman et al., 2008; Piacentini et al., 2011) to allow families to pursue other potentially efficacious treatment rather than assignment to an intervention known to have low response rates (e.g., 20%; Freeman et al., 2008) and a relatively high attrition rate (27% among youth randomized to relaxation/education in Freeman et al. (2008) vs. 0% randomized to TAU in a recent child E/RP controlled trial for anxiety (Storch, Arnold, et al., 2013; Storch, Bussing, et al., 2013)). Further, the downward extension of age in our study limits the sophistication of relaxation techniques, decreasing the likelihood that relaxation would be an effective control. Consequently, we chose to allow families to select their own care rather than allocate a treatment with very limited support, given the lack of validated treatments for preschool age youth with OCD. Children and families in the TAU condition were assessed regularly using the Service Assessment for Children and Adolescents e Service Use Scale (SACA) (Horwitz, 2001). Ten of 14 TAU youth (71%) received services: 4 participated in individual (non-OCD specific) therapy with a psychologist that did not involve E/RP (average of 4.5 visits), 4 sought consultation with a psychiatrist of whom 2 were placed on an SSRI, and 2 youth received in-home parent management training (without E/RP; average 6 visits). Among the youth receiving pharmacotherapy, one also received therapy from a school counselor and the parent of another child sought a behavioral parenting consultation. Patients receiving TAU were made aware (at randomization) that they would be offered free E/RP at post-treatment regardless of response. Rater training All clinician-rated measures were administered by independent evaluators (IE) who were blinded to participants’ study conditions and were not involved in treatment. Interviewers had at least 2 years experience in the assessment of OCD. Training involved didactics on all study measures, and scoring to at least 90% agreement with a senior clinician on a video-recording, corating multiple live interviews, and conducting at least one evaluation under direct observation of the principal investigator. All assessments were audio-recorded and integrity checks were conducted at weekly supervision, as well as for 20% of primary outcome measures. Measures Anxiety Disorders Interview Schedule e Parent Version (ADIS) The ADIS (Silverman & Albano, 1996) is a clinician-administered, semi-structured interview that was developed based on DSM-IV
diagnostic criteria for Axis I disorders. Given the young age of this sample, the ADIS was used to verify OCD diagnosis and screen for comorbidity based on symptom endorsement by the parent. Diagnostic presence and severity is established using a clinical severity rating (CSR) of 0e8 on a continuous rating scale with scores of 4 or higher indicating diagnostic presence and significant impairment/distress. The ADIS has been widely used and previous research has demonstrated validity and inter-rater agreement (Wood, Piacentini, Bergman, McCracken, & Barrios, 2002) including with youth as young as 3 years (Rapee, Kennedy, Ingram, Edwards, & Sweeney, 2005). The ADIS was administered at baseline, posttreatment, and follow-up assessments. Inter-rater reliability was calculated for 20% of the sample with full agreement on primary OCD diagnosis (Kappa ¼ 1.0). Peabody Picture Vocabulary Test (PPVT) The PPVT (Dunn & Dunn, 2007) is a standardized, normreferenced assessment of the child’s receptive vocabulary. This assessment was clinician-administered and was used during the baseline assessment only to verify the child’s threshold intellect. Standard scores 80 were used to indicate average intellectual ability. The PPVT has demonstrated excellent psychometric properties, construct validity, and cultural sensitivity (Dunn & Dunn, 2007). Children Yale-Brown Obsessive Compulsive Scale (CYBOCS) The CYBOCS (Scahill et al., 1997) is a semi-structured, clinicianadministered measure consisting of a checklist of common obsessions and compulsions and a 10-item severity scale, assessing symptoms over the course of the previous week. Each of the 10 severity-scale items is rated on a 5-point Likert scale which yields an Obsession and a Compulsion severity score (each based on 5 items assessing symptom frequency, intensity, interference, resistance, and control) as well as a total score. The CYBOCS was administered at all time-points, and scores served as a primary symptom severity outcome for the study. The CYBOCS is considered the gold standard for OCD severity in youth, (Lewin & Piacentini, 2010) and has been validated in very young children (Freeman, Flessner, & Garcia, 2011; Lewin et al., in press), with focus on obtaining information from the parent. The CYBOCS has demonstrated excellent inter-rater and testeretest reliability as well as construct validity (Freeman et al., 2011; Scahill et al., 1997; Storch et al., 2004). Initial severity scores are consistent with a moderate range of severity (approximately the 45th percentile) based on recent normative data of treatment seeking youth (Lewin et al., in press). Consistent with Freeman et al. (2008), remission was defined as a score of 12 or less. Excellent agreement was found for randomly selected recordings (ICC ¼ 0.84). Clinical Global Impression e Severity and e Improvement (CGI-S/-I) The CGI-S (National Institute of Mental, 1985) is a single item, 7point clinician rating of illness severity. The rating scale ranges from no illness to serious illness. The CGI-I (Guy, 1976) is a single item, 7point clinician rating of treatment response ranging from very much improved to very much worse. The CGI-S/I are widely used in clinical trials and have strong validity as primary outcome measures for pediatric OCD (Lewin et al., 2012). The CGI-S was administered at baseline, post-treatment, 1-month, and 3-month follow up assessments by an IE (CGI-I at all but the baseline). Consistent with prior research, youth rated by the IE as being much or very much improved on the CGI-I were considered treatment responders and youth rated as mild illness (or less) on the CGI-S were considered remitters (Lewin & Piacentini, 2010). Inter-rater agreement was strong for both the CGI-S/I respectively (ICC ¼ 0.89 and 0.94).
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
Pediatric Anxiety Rating Scale (PARS) The PARS (RUPP, 2002) is a clinician-rated, semi-structured interview (administered by the IE at all time-points). The five item PARS severity scale, which assesses frequency of anxiety symptoms, distress related to anxiety, avoidance associated with feared stimuli, interference at home, and interference outside the home (e.g., peer relationships, school functioning), was utilized. The psychometrics of the PARS are well established in populations of youth with anxiety (RUPP, 2002). National Institute of Mental Health Global OCD Scale (NIMH-GOCS) The NIMH-GOCS (Goodman & Price, 1992) is an anchored, clinician-administered scale that provides a rating of the overall severity of OCD symptoms as they relate to functional impairment. Conducted by the IE at all time points, NIMH-GOCS scores served as a secondary symptom severity outcome in the study. The scale is composed of a single item with a 15-point rating ranging from 1 (normal) to 15 (very severe). The NIMH-GOCS has been widely used in both adult and pediatric populations. Service Assessment for Children and Adolescents e service use scale (SACA) The SACA (Horwitz, 2001) is a standardized clinicianadministered interview and was used to assess mental health service utilization by the child and family (e.g., outpatient, schoolbased, family-based, inpatient). The SACA was administered at post-treatment (or week 7 in the TAU condition) to determine any services received by participants since baseline. Measure of Family Accommodation (FA) The FA (adapted from Calvocoressi et al., 1999) is a 13-item clinician-administered measure that assesses the degree to which family members have accommodated a child’s OCD symptoms over the past month. The measure also assesses the level of distress/ impairment that family members and the child experience as a result of accommodating or not accommodating the child over the past month. The IE administered the FA to the parent at all timepoints. Items are rated on a 5-point Likert scale from 0 (none/ never) to 4 (everyday/extreme). The FA has demonstrated good psychometric properties (Peris et al., 2008; Storch, Geffken, Merlo, Jacob, et al., 2007). Satisfaction with services Parent satisfaction was assessed at post-treatment with a selfreport measure (Hawley & Weisz, 2005). Parents were specifically queried about the acceptability of the E/RP focused intervention. Expectancy rating Parental expectations of treatment were obtained pretreatment via a single item self-report previously validated in a child OCD sample (Lewin, Peris, Bergman, McCracken, & Piacentini, 2011). Children’s Sheehan Disability Scale-parent The CSDS is a 5-item parent report of impairment in school, home, and social domains for the child, and home and work domains for the parent due to the child’s OCD symptoms (Whiteside, 2009). Items are rated on an 11-point Likert scale from 0 to 10. The CSDS has demonstrated good psychometric properties (Whiteside, 2009). Analytic plan Consistent with feasibility/pilot trial methodology, group differences on continuous outcomes were evaluated by ANCOVAs,
35
with post-treatment scores predicted by treatment group while including pre-treatment scores as a covariate. Group differences for categorical outcomes were evaluated by Fisher’s exact tests. Effect sizes were converted into Cohen’s d based on formulas provided (Cooper, Hedges, & Valentine, 2009), where values of d of 0.2, 0.5, and 0.8 correspond to small, medium, and large effect sizes, respectively (Cohen, 1988). Decline of symptoms at 1-month and 3month follow up was evaluated via paired t-tests for continuous outcomes and Fisher’s exact tests for categorical outcomes, where only participants who returned for follow up were included in analyses. Multiple categorizations of treatment response and symptom remission were employed. Using the CGI, treatment response was defined by categorizations of “much better” or “very much better” on the CGI-I, and remission was defined by categorizing patients as having “mild symptoms” or less on the CGI-S, following research actuarial precedent (Lewin et al., 2011; Storch, Lewin, De Nadai, & Murphy, 2010). To provide comparison to other studies, we also considered remission outcome using CYBOCS scores of 12 at post-treatment reflecting symptom remission (mirroring criteria used by Freeman et al., 2008). In considering age group differences, the sample was split where youth below age 5 were considered the younger group (n ¼ 11), while all other participants were categorized as older youth (n ¼ 20). Missing data were imputed using Hotdeck imputation, which preserves validity of inference provided data were missing at random and the proportion of missing data are under 10% (Myers, 2011). Follow-up analysis included youth who were treatment responders. Comorbid treatment status was used as the adjustment variable for providing imputed values. Overall, missing data were very few (
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Family-based exposure and response prevention therapy for preschool-aged children with obsessive-compulsive disorder: A pilot randomized controlled trial Adam B. Lewin a, b, c, *, Jennifer M. Park b, Anna M. Jones b, Erika A. Crawford a, Alessandro S. De Nadai b, Jessie Menzel b, Elysse B. Arnold a, b, Tanya K. Murphy a, c, Eric A. Storch a, b, c a
Department of Pediatrics, University of South Florida, USA Department of Psychology, University of South Florida, USA c Department of Psychiatry & Behavioral Neurosciences, University of South Florida, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 September 2013 Received in revised form 10 February 2014 Accepted 11 February 2014 Available online 2 March 2014
Aims: To examine the feasibility, acceptability and preliminary efficacy of family-based exposure/ response prevention therapy (E/RP) versus treatment as usual (TAU) in a cohort of very young children with early onset obsessive-compulsive disorder (OCD). Methods: Thirty-one children ages 3e8 years (M ¼ 5.8 years) with a primary diagnosis of OCD were randomized to E/RP or TAU. The E/RP condition received 12 sessions of family-based E/RP twice weekly over 6 weeks. Families were assessed at baseline, post-treatment, 1-month and 3-month follow up. The Children’s Yale Brown Obsessive Compulsive Scale and Clinical Global Impression served as primary outcome measures. Results: A large group effect emerged in favor of the E/RP group (d ¼ 1.69). Sixty-five percent of the E/RP group was considered treatment responders as compared to 7% in the TAU group. Symptom remission was achieved in 35.2% of the E/RP group and 0% of the TAU group. There was no attrition and satisfaction was high; gains were maintained at 3 months. Conclusions: Even amongst children as young as 3 years, developmentally tailored E/RP is efficacious and well-tolerated in reducing OCD symptoms. Key adaptations for younger children include extensive parent involvement, targeting family accommodation, and frequent meetings while delivering a full course of E/RP. ClinicalTrials.gov Identifier: NCT01447966 http://clinicaltrials.gov/ct2/show/NCT01447966?term¼ocdþ andþstþpetersburg&rank¼1 Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: OCD Exposure-response prevention Very young children Accommodation Cognitive therapy Preschool
Pediatric obsessive compulsive disorder (OCD) is a chronic and disabling illness with a point-prevalence of 1e2% (Zohar, 1999). Age of symptom onset varies but generally occurs during key developmental periods such as pre-pubescence or adolescence/early adulthood, with up to 80% of adults reporting emergence of symptoms prior to the age of 18 years (AACAP, 2012). As pediatric OCD was previously believed to be rare amongst very young children, most
Abbreviations: E/RP, exposure and response prevention; TAU, treatment as usual; OCD, obsessive compulsive disorder. * Corresponding author. USF Dept. of Pediatrics, Rothman Center for Neuropsychiatry, 880 6th Street South Suite 460 Box 7523, Child Development & Rehabilitation Center, St. Petersburg, FL 33701, USA. E-mail address: [email protected] (A.B. Lewin). http://dx.doi.org/10.1016/j.brat.2014.02.001 0005-7967/Ó 2014 Elsevier Ltd. All rights reserved.
studies focus primarily on school-aged children above the age of 8 years. However, reports have noted OCD presentations amongst those as young as 2 years (Coskun & Zoroglu, 2009). While reports on early childhood OCD (i.e., onset of symptoms prior to age of 9 years e.g., Freeman et al., 2008) are scant, studies have noted that symptoms are associated with positive family history, increased distress, longer duration of illness, and higher rates of comorbid tics (Coskun & Zoroglu, 2009; Garcia et al., 2009; Nakatani et al., 2011). Those affected by early childhood OCD may be at particular risk as OCD symptoms may impede or hinder normal development during key milestone years (Flessner, Garcia, & Freeman, 2013; Valderhaug & Ivarsson, 2005). Consequently, early and aggressive intervention is paramount for younger children with OCD (Hirshfeld-Becker & Biederman, 2002; Lewin, Park, & Storch, 2013).
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
Early childhood OCD presents with a number of unique developmental factors, which should be addressed during the course of treatment. Children are generally enmeshed into the dynamics of the family and often do not have the ability to disengage with family related stressors such as marital dysfunction and parental mental health (Freeman et al., 2003). These factors may strongly influence the mental health of the child and affect treatment progress, as well as maintenance of gains (Kazdin, 1995; Kazdin & Weisz, 1998). Family accommodation, which refers to the involvement of family members in the child’s OCD symptoms, is highly prevalent in children and adolescents with OCD (Storch, Geffken, Merlo, Jacob, et al., 2007). Family members may modify their family routine to allow children to avoid anxiety-provoking situations, or engage in rituals such as providing reassurance, assisting in mealtime, bedtime, or toileting rituals (Peris et al., 2008; Storch, Geffken, Merlo, Jacob, et al., 2007). Because of the involvement of family members in OCD symptoms, as well as the influence of family dynamics upon the course of the treatment, it is imperative that family members are involved in the treatment process. There are two evidence-based treatments for pediatric OCD: serotonin reuptake inhibitors (SRIs) and exposure and response prevention (E/RP) (Lewin & Piacentini, 2009; POTS, 2004). Pooled effects suggest that E/RP has an advantage over SRI treatment alone (Abramowitz, Whiteside, & Deacon, 2005) leading to practice parameter recommendations that children receive E/RP alone in those with mild to moderate severity or with SRI therapy in more severe cases (AACAP, 2012). While efficacious, pharmacotherapy may have serious adverse events, somatic side effects (e.g., poor sleep, gastrointestinal complaints), and behavioral activation with increased risk at younger ages (Murphy, Segarra, Storch, & Goodman, 2008) even at small doses. For example, in an open trial of fluoxetine in six very young children with OCD, 50% developed significant medication-related side effects including two with behavioral disinhibition (Coskun & Zoroglu, 2009). Although a recent case series suggests fluoxetine produced clinical improvement in four treatment-refractory preschool aged youth with OCD (Ercan, Kandulu, & Akyol Ardic, 2012), medication is often undesirable to parents of young children, especially in cases of new onset/initial treatment. Conversely, E/RP for pediatric OCD is a well-tolerated and acceptable option for parents (Patel & Simpson, 2010). Welldesigned studies have had positive outcomes, particularly amongst children ages 7e17 years, with response rates ranging from 39% to 90% (Piacentini et al., 2011; POTS, 2004; Storch, Bussing, et al., 2013; Storch, Geffken, Merlo, Mann, et al., 2007). However, despite OCD being relatively common among very young children (Coskun, Zoroglu, & Ozturk, 2012; Flessner et al., 2013; Garcia et al., 2009) and practice parameters recommending E/RP as the first line intervention for pre-latency youth with OCD (AACAP, 2012; Gleason et al., 2007), strikingly few children below the age of 8 years have been included in the extant controlled research (Flessner et al., 2013). Indeed, Freeman et al. (2008) noted that within the Pediatric OCD Treatment Study I and Study II (POTS), which are considered the largest pediatric OCD treatment studies to date, only 10e16% were ages 7 or 8 years (POTS, 2004). Following the POTS trials, only one published trial to date has included youth under the age of 7 years (Freeman et al., 2008) although a larger trial for youth ages 5e8 years is underway (Freeman et al., 2012). In Freeman et al. (2008), a 12-session family-based E/RP course was found efficacious for treating youth ages 5e8 years with OCD. Completer analysis showed a large group difference (d ¼ 0.85) with greater obsessive-compulsive symptom reductions for youth receiving E/RP (n ¼ 16) versus a 12-session relaxation training control (n ¼ 15). Following the 12-session course, 69% of youth responded to E/RP compared to 20% of youth receiving relaxation
31
training. Notably, the authors implemented modifications to the E/ RP protocol to address developmental, cognitive, and socioemotional differences specific to children age 8 years and younger. Specific modifications included: (1) Inclusion of parents as coaches to foster motivation, increase adherence, and reduce accommodation; (2) Cognitive therapy components were simplified as concrete, child-specific examples and to foster allegiance in combating ‘OCD’ as a common adversary; (3) Addressing other demands associated with treating anxiety in a very young sample including child oppositionality/resistance, managing out-ofsession E/RP homework, parental tolerance of their own distress during E/RP, and related logistical issues for parents. Despite this promising work, a clear need exists for replication and extension of these findings in youth age 8 years and under with OCD to help support E/RP as efficacious for this younger cohort. Consequently, the present study is an investigation of feasibility, tolerability and preliminary efficacy of a family-based E/RP protocol designed for the youngest cohort of youth with OCD to date. Rather than an exact replication of Freeman et al. (2008), we introduced a novel protocol with several modifications given (1) the younger cohort (32% of our sample was below the range included in Freeman et al. (2008)), (2) to reduce attrition (26% in Freeman et al. (2008)) and (3) bolster responsiveness to this promising approach without (4) compromising acceptability. Briefly, key modifications included: (a) a primary focus on E/RP versus the inclusion of other treatment elements (e.g., 4e6 sessions in Freeman et al. (2008) without primary E/RP focus); (b) eliminating most cognitive procedures; (c) parent/child inclusion in all sessions; (d) twice-weekly format to expediate response; (e) tenacious targeting of family accommodation starting at session 1. Contrasts with the promising Freeman et al. (2008, 2009) protocol were as follows. Treatment utilized in the present study was refined to emphasize the primacy of E/RP; rather than initiating E/ RP at session 5 or 6, our revised protocol implemented E/RP no later than the second visit (but often during session 1). Moreover, after the first session, E/RP was the primary focus of each of the remaining sessions with even less emphasis on cognitive strategies/child focused psychoeducation (given the very young age of the targeted cohort, i.e., inclusion of 3e5 year olds). Several other modifications were implemented given the very young age range. First, both parents and youth were included in all sessions together (unlike Freeman et al. (2008) which excluded youth until session 3). While there is support for children as young as 4 years benefiting from verbally-presented information regarding exposure therapy (Scheeringa et al., 2011), emphasis was placed on parental understanding of E/RP concepts (Choate-Summers et al., 2008). Although parents desire their child to “understand” therapeutic concepts (Labouliere, Arnold, Storch & Lewin, 2014), there is little benefit in spending additional session time on child focused psychoeducation with a 3e5 year old (Choate-Summers et al., 2008), especially at the expense of reducing the time spent in E/RP. Thus, having the child present for parental psychoeducation could appease parents while preserving time dedicated for E/RP. Consequently, instead of utilizing two sessions for education, our revised protocol was more streamlined in order to maximize the dosage of E/RP. Another advantage of keeping the parent and child together in all sessions was the ability to provide singular focus on E/RP rather than separate parent and child goals in each session (Freeman & Garcia, 2009). Additionally, our study protocol addresses accommodation beginning week 1, as a mechanism for conducting exposure. Family accommodation is a hallmark of early childhood OCD (Garcia et al., 2009; Flessner et al., 2013; Freeman et al., 2003) and consequently E/RP was adapted to aggressively target accommodation. Parent engagement in child rituals and/or parent facilitation of OCD (e.g., allowing extra time, purchasing special products, making
32
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
individual meals, etc.) was typically targeted prior to asking the child to engage in E/RP (e.g., engage with a feared object or change a ritualistic routine). Targeting accommodation as the first “exposure” avoided coercion and negotiation (the parent was choosing to change behavior) while still permitting habituation to distress and progression up a hierarchy. Further, this modification was particularly appropriate for younger children (e.g., 3e5 years) who had limited capacity to understand treatment related concepts and hierarchy development (Labouliere et al., 2014). Parents left session 1 with specific instructions to begin reducing accommodation (or were directed to construct a plan to initiate during the second visit later week 1). Additionally, we introduced a twice-weekly format to accelerate the treatment (especially E/RP). Given the aggressiveness of our approach (early introduction of E/RP without implementation of supportive therapeutic/cognitive strategies), the twice-weekly content allowed for frequent therapist contact and was feasible given the age of our cohort (i.e., less difficulty with missing school). Additional details are presented in the Methods. Accordingly, the present study aimed to examine feasibility, acceptability, and preliminary efficacy via a pilot randomized controlled trial of our adapted family based E/RP intervention compared to a Treatment as Usual (TAU) control condition. We hypothesized that family-based E/RP would be a highly acceptable and well-tolerated intervention with symptom reductions similar to the Freeman et al. (2008) trial, but with decreased attrition. Secondary aims included examination of changes in family accommodation, functional impairment, and maintenance of treatment gains at 1 and 3 month follow-up. Method Participants Participants were 31 children with a primary OCD diagnosis (ages of 3e8 years; M ¼ 5.8; SD ¼ 1.6) recruited through patient flow at a university-based specialty clinic. Thirty-six percent of youth were age 5 years or younger (breakdown by age was as follows: 2 [age 3], 8 [age 4], 1 [age 5] 7 [age 6] 9 [age 7] and 4 [age 8]. Inclusion required: (1) Clinician diagnosis of primary OCD with diagnoses verified by administration of the Anxiety Disorders Interview Schedule (ADIS) e Child and Parent Versions (Silverman & Albano, 1996) clinical severity rating 4; CYBOCS 16 (or CYBOCS Compulsions scale of > 8); between ages of 3e8 years inclusive; commitment of consistent parent to attend all sessions. Exclusion criteria included: (1) the presence of psychiatric illness that contraindicated study participation, e.g., autism; (2) IQ < 85; (3) nonEnglish speaking or absence of language; (4) history of E/RP; (5) a recent change in psychotropic medication1; or (6) concurrent psychotherapy or other behavioral interventions. Children randomized to TAU were allowed to make medication changes following randomization. All children randomized to E/RP remained stable on medications throughout the study (see Table 1). Fig. 1 presents descriptive data regarding participant flow through the study. There was no attrition from either condition. Procedure The IRB approved this study, and both parents and children provided written informed consent and assent respectively prior to
1 For youth on established pharmacotherapy, 8 weeks stability at current dose was required for SSRIs and 6 weeks for other medications. For newly initiated pharmacotherapy, 12 weeks stability for SSRIs and 8 weeks for other mediations was required.
Table 1 Sample demographics.
Child sex (male) Child age (mean, SD) Parent sex (female) Parent graduated from college Living Situation Married Both parents (same) Both parents (different residence) Lives with single parent (mom) Lives with mom/stepdad Child ethnic/race White Asian/Pacific islander Latino/Latina African American Comorbid diagnoses Any anxiety disorder Separation anxiety disorder Social phobia GAD Specific phobia ADHD Oppositional defiant disorder Psychiatric medication use SSRI Guanfacine Stimulant
E/RP n (%)
TAU n (%)
c2/t value
12 5.76 16 9
10 5.86 10 7
0.003 0.16
(71) (1.52) (94) (53)
(71) (1.70) (71) (50)
1.73 13 12 1 3 1
(76) (71) (6) (18) (6)
15 1 2 1
(88) (6) (12) (6)
12 4 8 7 5 5 5
(71) (24) (47) (41) (29) (29) (29)
12 12 1 1
(86) (86) (7) (7)
0 4.08
0
0 2 (14) 0 10 4 4 4 6 8 6
(71) (29) (29) (29) (35) (57) (35)
0.003 0.10 1.11 0.53 0.61 2.43 0.61 e
0 1 (6)
0
12 (86)
4 (29) 2 (14)
2.92 1.76
enrollment. At a baseline assessment, primary diagnosis of OCD was determined by clinical interview with a board certified child psychologist (and consensus review of all records by a second experienced licensed child psychologist). Diagnoses were verified by administration of the ADIS. Following baseline assessment, qualifying participants were randomized (using a 1:1 ratio) to E/RP or TAU; parents were informed of condition immediately by the study coordinator. Post-treatment assessments were administered within 1 week following the final E/RP session or at approximately 6e7 weeks following baseline assessment in the TAU condition. Follow-up assessments were conducted at 1-month and 3-months post-treatment (balanced for both groups to preserve blindness of ratings). Procedures consistent with other child randomized studies (Piacentini et al., 2011, 2010) were taken to preserve blindness (e.g., separate supervision of raters/therapists, reminding participants of the blind before each assessment, positioning independent evaluators separate from therapist, utilizing ancillary entrance/exit to avoid breaking the blind). However, notably, a recent study demonstrated that blinding is not essential for conducting accurate ratings in clinical trials of child OCD (Lewin, Peris, De Nadai, McCracken, & Piacentini, 2012). Treatment Family-based Exposure/Response Prevention (E/RP) The E/RP intervention consisted of 12 sessions, 60 minutes in length, administered twice weekly over the course of 6 weeks. The twice-weekly format was selected to provide more frequent contact and possibly reduce attrition (26.2% of the Freeman et al. (2008) sample did not complete the 12 session protocol implemented over 14 weeks suggesting protocol changes may be necessary). The E/RP content was based on a manualized, familybased treatment for young children with OCD (Freeman & Garcia, 2009) with modifications to emphasize E/RP earlier in the protocol, hold sessions twice-weekly, and aggressively target accommodation. This treatment targets OCD symptoms as well as accommodation by family members. All treatment sessions were
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
33
Fig. 1. Study consort diagram. Note: from Schulz KF, Altman DG, Moher D, for the CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ 2010; 340:c332. For more information, visit: www.consort-statement.org.
conducted with the child and at least one primary caregiver present at all times. A consistent parent was required for every session (although the other parent/guardian was encouraged to attend when feasible). Treatment components included psychoeducation for parents and children (allying against ‘OCD’; introducing developmentally appropriate metaphors/examples), parent tools (e.g., development of a rewards program, differential reinforcement, extinction, and modeling), E/RP, and relapse prevention planning (Freeman & Garcia, 2009; Freeman et al., 2008). Given the younger age in the present sample, cognitive elements were further reduced and emphasis was placed on parent-focused aspects of the protocol (e.g., reducing accommodation, training the parent in extinctionbased treatment and need for the child to habituate to anxious triggers, and implementing E/RP independently at home) (Kendall et al., 2005; Piacentini & Bergman, 2001). The revised treatment schedule was as follows: Session 1 e introduction, parent-focused psychoeducation on E/RP, allying against OCD, hierarchy development, and parent driven E/RP directed at family accommodation behaviors; Session 2 e continued hierarchy development, targeting accommodation (ignoring)/extinction, resisting OCD; introduce/ model E/RP with the child; Session 3 e more modeling exposures, practice in-session E/RP, developing reward program (as needed); Sessions 4e11 e in session E/RP, brief developmentally appropriate
cognitive interventions (e.g., labeling “It’s only my OCD,” resisting “I don’t have to listen to [OCD],” positive self-talk “I can win this fight against OCD.”), continued reduction of parental accommodation, increased parent-led E/RP, adjusting reward program as needed; Session 12 e continued E/RP, relapse prevention, and certificate/ brief celebration. Many of the procedures described in the existing literature (Choate-Summers et al., 2008; Flessner et al., 2013; Freeman et al., 2008, 2003) for treating youth aged 5e8 years were further modified for our even younger sample. Behavioral techniques were embedded within the E/RP practice to avoid losing session time dedicated to practicing exposures (without a need for adding additional sessions). In other words, parent training principles such as differential reinforcement of other behavior/alternative behavior (DRO/DRA), extinction, response cost, and behavioral reduction procedures were implemented within the context of E/RP practice (Fernandez et al., 2006; Labouliere et al., 2014) when necessary. Further, our protocol focused on reducing parental accommodation immediately in treatment. Similar to the Freeman et al. (2008) protocol, E/RP was parent driven. However, our protocol focused on reducing accommodation as a parent driven first-step of the E/RP hierarchy. Therapists were licensed clinical psychologists or advanced clinical psychology doctoral students (who had 3 years experience
34
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
in E/RP for OCD under the supervision of the principal investigator); therapists were assigned to participants at random. Therapist training included didactics with the principal investigator, advanced guided reading of the treatment manual, and completion of at least one supervised training case of a patient under age 8 using the manual with integrity checks. Treatment integrity was ensured through weekly supervision, reviewing content checklists that corresponded to the treatment manual, and evaluation of 15% of randomly selected audio-recorded therapy sessions using an 11item, 6-point scale. There was excellent adherence to session components (25.0/30.0), therapist competence (16.2/18.0), and therapist alliance (11.2/12.0). Treatment as usual The TAU condition was designed to reflect standard practice. All participants randomized to the TAU condition were instructed to continue receiving their prior interventions as recommended by their providers (e.g., psychotherapy, social skills training, pharmacological interventions, or parenting classes) or were permitted to seek new treatment/evaluation from providers not affiliated with our treatment program. The TAU control was chosen over relaxation/ education controls implemented in previous studies (e.g., Freeman et al., 2008; Piacentini et al., 2011) to allow families to pursue other potentially efficacious treatment rather than assignment to an intervention known to have low response rates (e.g., 20%; Freeman et al., 2008) and a relatively high attrition rate (27% among youth randomized to relaxation/education in Freeman et al. (2008) vs. 0% randomized to TAU in a recent child E/RP controlled trial for anxiety (Storch, Arnold, et al., 2013; Storch, Bussing, et al., 2013)). Further, the downward extension of age in our study limits the sophistication of relaxation techniques, decreasing the likelihood that relaxation would be an effective control. Consequently, we chose to allow families to select their own care rather than allocate a treatment with very limited support, given the lack of validated treatments for preschool age youth with OCD. Children and families in the TAU condition were assessed regularly using the Service Assessment for Children and Adolescents e Service Use Scale (SACA) (Horwitz, 2001). Ten of 14 TAU youth (71%) received services: 4 participated in individual (non-OCD specific) therapy with a psychologist that did not involve E/RP (average of 4.5 visits), 4 sought consultation with a psychiatrist of whom 2 were placed on an SSRI, and 2 youth received in-home parent management training (without E/RP; average 6 visits). Among the youth receiving pharmacotherapy, one also received therapy from a school counselor and the parent of another child sought a behavioral parenting consultation. Patients receiving TAU were made aware (at randomization) that they would be offered free E/RP at post-treatment regardless of response. Rater training All clinician-rated measures were administered by independent evaluators (IE) who were blinded to participants’ study conditions and were not involved in treatment. Interviewers had at least 2 years experience in the assessment of OCD. Training involved didactics on all study measures, and scoring to at least 90% agreement with a senior clinician on a video-recording, corating multiple live interviews, and conducting at least one evaluation under direct observation of the principal investigator. All assessments were audio-recorded and integrity checks were conducted at weekly supervision, as well as for 20% of primary outcome measures. Measures Anxiety Disorders Interview Schedule e Parent Version (ADIS) The ADIS (Silverman & Albano, 1996) is a clinician-administered, semi-structured interview that was developed based on DSM-IV
diagnostic criteria for Axis I disorders. Given the young age of this sample, the ADIS was used to verify OCD diagnosis and screen for comorbidity based on symptom endorsement by the parent. Diagnostic presence and severity is established using a clinical severity rating (CSR) of 0e8 on a continuous rating scale with scores of 4 or higher indicating diagnostic presence and significant impairment/distress. The ADIS has been widely used and previous research has demonstrated validity and inter-rater agreement (Wood, Piacentini, Bergman, McCracken, & Barrios, 2002) including with youth as young as 3 years (Rapee, Kennedy, Ingram, Edwards, & Sweeney, 2005). The ADIS was administered at baseline, posttreatment, and follow-up assessments. Inter-rater reliability was calculated for 20% of the sample with full agreement on primary OCD diagnosis (Kappa ¼ 1.0). Peabody Picture Vocabulary Test (PPVT) The PPVT (Dunn & Dunn, 2007) is a standardized, normreferenced assessment of the child’s receptive vocabulary. This assessment was clinician-administered and was used during the baseline assessment only to verify the child’s threshold intellect. Standard scores 80 were used to indicate average intellectual ability. The PPVT has demonstrated excellent psychometric properties, construct validity, and cultural sensitivity (Dunn & Dunn, 2007). Children Yale-Brown Obsessive Compulsive Scale (CYBOCS) The CYBOCS (Scahill et al., 1997) is a semi-structured, clinicianadministered measure consisting of a checklist of common obsessions and compulsions and a 10-item severity scale, assessing symptoms over the course of the previous week. Each of the 10 severity-scale items is rated on a 5-point Likert scale which yields an Obsession and a Compulsion severity score (each based on 5 items assessing symptom frequency, intensity, interference, resistance, and control) as well as a total score. The CYBOCS was administered at all time-points, and scores served as a primary symptom severity outcome for the study. The CYBOCS is considered the gold standard for OCD severity in youth, (Lewin & Piacentini, 2010) and has been validated in very young children (Freeman, Flessner, & Garcia, 2011; Lewin et al., in press), with focus on obtaining information from the parent. The CYBOCS has demonstrated excellent inter-rater and testeretest reliability as well as construct validity (Freeman et al., 2011; Scahill et al., 1997; Storch et al., 2004). Initial severity scores are consistent with a moderate range of severity (approximately the 45th percentile) based on recent normative data of treatment seeking youth (Lewin et al., in press). Consistent with Freeman et al. (2008), remission was defined as a score of 12 or less. Excellent agreement was found for randomly selected recordings (ICC ¼ 0.84). Clinical Global Impression e Severity and e Improvement (CGI-S/-I) The CGI-S (National Institute of Mental, 1985) is a single item, 7point clinician rating of illness severity. The rating scale ranges from no illness to serious illness. The CGI-I (Guy, 1976) is a single item, 7point clinician rating of treatment response ranging from very much improved to very much worse. The CGI-S/I are widely used in clinical trials and have strong validity as primary outcome measures for pediatric OCD (Lewin et al., 2012). The CGI-S was administered at baseline, post-treatment, 1-month, and 3-month follow up assessments by an IE (CGI-I at all but the baseline). Consistent with prior research, youth rated by the IE as being much or very much improved on the CGI-I were considered treatment responders and youth rated as mild illness (or less) on the CGI-S were considered remitters (Lewin & Piacentini, 2010). Inter-rater agreement was strong for both the CGI-S/I respectively (ICC ¼ 0.89 and 0.94).
A.B. Lewin et al. / Behaviour Research and Therapy 56 (2014) 30e38
Pediatric Anxiety Rating Scale (PARS) The PARS (RUPP, 2002) is a clinician-rated, semi-structured interview (administered by the IE at all time-points). The five item PARS severity scale, which assesses frequency of anxiety symptoms, distress related to anxiety, avoidance associated with feared stimuli, interference at home, and interference outside the home (e.g., peer relationships, school functioning), was utilized. The psychometrics of the PARS are well established in populations of youth with anxiety (RUPP, 2002). National Institute of Mental Health Global OCD Scale (NIMH-GOCS) The NIMH-GOCS (Goodman & Price, 1992) is an anchored, clinician-administered scale that provides a rating of the overall severity of OCD symptoms as they relate to functional impairment. Conducted by the IE at all time points, NIMH-GOCS scores served as a secondary symptom severity outcome in the study. The scale is composed of a single item with a 15-point rating ranging from 1 (normal) to 15 (very severe). The NIMH-GOCS has been widely used in both adult and pediatric populations. Service Assessment for Children and Adolescents e service use scale (SACA) The SACA (Horwitz, 2001) is a standardized clinicianadministered interview and was used to assess mental health service utilization by the child and family (e.g., outpatient, schoolbased, family-based, inpatient). The SACA was administered at post-treatment (or week 7 in the TAU condition) to determine any services received by participants since baseline. Measure of Family Accommodation (FA) The FA (adapted from Calvocoressi et al., 1999) is a 13-item clinician-administered measure that assesses the degree to which family members have accommodated a child’s OCD symptoms over the past month. The measure also assesses the level of distress/ impairment that family members and the child experience as a result of accommodating or not accommodating the child over the past month. The IE administered the FA to the parent at all timepoints. Items are rated on a 5-point Likert scale from 0 (none/ never) to 4 (everyday/extreme). The FA has demonstrated good psychometric properties (Peris et al., 2008; Storch, Geffken, Merlo, Jacob, et al., 2007). Satisfaction with services Parent satisfaction was assessed at post-treatment with a selfreport measure (Hawley & Weisz, 2005). Parents were specifically queried about the acceptability of the E/RP focused intervention. Expectancy rating Parental expectations of treatment were obtained pretreatment via a single item self-report previously validated in a child OCD sample (Lewin, Peris, Bergman, McCracken, & Piacentini, 2011). Children’s Sheehan Disability Scale-parent The CSDS is a 5-item parent report of impairment in school, home, and social domains for the child, and home and work domains for the parent due to the child’s OCD symptoms (Whiteside, 2009). Items are rated on an 11-point Likert scale from 0 to 10. The CSDS has demonstrated good psychometric properties (Whiteside, 2009). Analytic plan Consistent with feasibility/pilot trial methodology, group differences on continuous outcomes were evaluated by ANCOVAs,
35
with post-treatment scores predicted by treatment group while including pre-treatment scores as a covariate. Group differences for categorical outcomes were evaluated by Fisher’s exact tests. Effect sizes were converted into Cohen’s d based on formulas provided (Cooper, Hedges, & Valentine, 2009), where values of d of 0.2, 0.5, and 0.8 correspond to small, medium, and large effect sizes, respectively (Cohen, 1988). Decline of symptoms at 1-month and 3month follow up was evaluated via paired t-tests for continuous outcomes and Fisher’s exact tests for categorical outcomes, where only participants who returned for follow up were included in analyses. Multiple categorizations of treatment response and symptom remission were employed. Using the CGI, treatment response was defined by categorizations of “much better” or “very much better” on the CGI-I, and remission was defined by categorizing patients as having “mild symptoms” or less on the CGI-S, following research actuarial precedent (Lewin et al., 2011; Storch, Lewin, De Nadai, & Murphy, 2010). To provide comparison to other studies, we also considered remission outcome using CYBOCS scores of 12 at post-treatment reflecting symptom remission (mirroring criteria used by Freeman et al., 2008). In considering age group differences, the sample was split where youth below age 5 were considered the younger group (n ¼ 11), while all other participants were categorized as older youth (n ¼ 20). Missing data were imputed using Hotdeck imputation, which preserves validity of inference provided data were missing at random and the proportion of missing data are under 10% (Myers, 2011). Follow-up analysis included youth who were treatment responders. Comorbid treatment status was used as the adjustment variable for providing imputed values. Overall, missing data were very few (
