Cognitive Biases and Obsessive-Compulsive Symptoms in Children: Examining the Role of Maternal Cognitive Bias and Child Age

Available online at www.sciencedirect.com

Behavior Therapy 43 (2012) 593 –605

www.elsevier.com/locate/bt

Cognitive Biases and Obsessive-Compulsive Symptoms in Children: Examining the Role of Maternal Cognitive Bias and Child Age Lara J. Farrell Allison M. Waters Melanie J. Zimmer-Gembeck Griffith University

Objective: Multiple cognitive biases associated with adult obsessive-compulsive disorder (OCD) were tested in a clinical sample of children (ages 7–11) and adolescents (12–17) and their mothers. This study examined (a) the associations between child cognitive biases and OCD severity, (b) maternal cognitive biases and child OCD severity, and (c) maternal cognitive bias and child cognitive bias. It was hypothesized that age would significantly moderate these relationships, with stronger associations with OCD severity for cognitive bias in adolescents (relative to children), and maternal cognitive bias in younger children (relative to adolescents). Method: Forty-six children and adolescents diagnosed with OCD and their mothers completed questionnaires assessing responsibility bias, thought–action fusion (TAF), thought suppression, and metacognitive beliefs. OCD symptoms were assessed using structured diagnostic interviews and semistructured symptom interviews. Results: As predicted, age significantly moderated associations between (a) child cognitive variables and OCD severity—specifically between child responsibility and child metacognition, which were associated with OCD severity for adolescents only; (b) maternal cognitive biases and child OCD severity—specifically for maternal responsibility and thought suppression, which were significantly and positively associated with child OCD severity but not adolescent OCD severity; and (c) maternal cognitive biases and child cognitive bias— such that significant associations were evident only in the younger child sample, and only between maternal TAF self Address correspondence to Lara Farrell, Ph.D., Griffith Health Institute, School of Applied Psychology, Griffith University, Gold Coast Campus, Brisbane, Queensland, 4222, Australia; e-mail: [email protected]. 0005-7894/43/593-605/$1.00/0 © 2011 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.

and metacognition, with child suppression and child TAF moral, respectively. Conclusion: Maternal cognitive biases are more consistently linked to greater OCD severity among younger children, whereas personal cognitive biases are associated with greater OCD symptoms in adolescents. Treatments for pediatric OCD are likely to be improved by age-specific considerations for the role of maternal and child cognitive biases associated with OCD.

Keywords: cognitive appraisals; meta-cognitive beliefs; OCD; thought–action fusion; thought suppression

R ESEARCH INTO OBSESSIVE - COMPULSIVE DISORDER (OCD) has continued to mount over the past decade (Boschen, 2008), with increasing evidence to support cognitive theoretical models (e.g., Taylor, Abramowitz, & McKay, 2006), and efficacy for cognitive-behavioral treatments (CBT) in both adults (e.g., Rosa-Alcázar, Sánchez-Meca, Gómez-Conesa, & Marín-Martínez, 2008) and children with OCD (e.g., Barrett, Farrell, Pina, Peris, & Piacentini, 2008). Despite the wealth of research into the cognitive underpinnings of OCD in adults (e.g., Frost & Steketee, 2002), there is comparatively limited research into the underlying mechanisms associated with the persistence and maintenance of this disorder during childhood. As such, examination of maladaptive beliefs, the development of such beliefs, and familial processes involved in childhood OCD warrants further investigation. Adult cognitive models of OCD focus on six core domains of cognition that have been identified by the Obsessive–Compulsive Cognitions Working Group (OCCWG, 1997, 2001) as centrally important to OCD. These cognitions include (a) inflated

594

farrell et al.

responsibility (Salkovskis, 1985, 1996), (b) overimportance of thoughts (e.g., thought–action fusion [TAF]; Rachman, 1993), (c) control of thoughts (e.g., thought suppression [Clark & de Silva, 1985] and meta-cognitive beliefs [Wells & Papageorgiou, 1998]), (d) overestimation of threat, (e) intolerance of uncertainly, and (f) perfectionism (OCCWG, 1997, 2001). Notably, all of these cognitive processes share a central tenet that an individual's beliefs and appraisals of OCD symptoms play an integral part in the development of the disorder. The extent to which these factors explain child and adolescent OCD remains unclear. Research with child samples is limited and has been restricted in focus, with much of the work examining inflated responsibility bias, metacognitions, and TAF (see Reynolds & Reeves, 2008, for a review). Moreover, of only 11 studies to date on cognitive models of OCD in children (Reynolds & Reeves, 2008), only four examined the applicability of cognitions in clinical samples (i.e., Barrett & Healy, 2003; Barrett & Healy-Farrell, 2003; Farrell & Barrett, 2006; Libby, Reynolds, Derisley, & Clark, 2004), and findings have been mixed. One study provided support for an inflated responsibility bias in 11- to 18-year-olds with OCD (i.e., Libby et al., 2004); however, another found responsibility biases were significantly lower for children versus adolescents and adults with OCD, whereas TAF, overestimation of threat, doubt, and cognitive control were similar across all age groups (Farrell & Barrett, 2006). Yet other work has shown that inflated responsibility was not associated with increased ratings of distress, avoidance or ritualizing in children with OCD, ages 7–17 years, during a behavioral avoidance task (Barrett & Healy-Farrell, 2003). Moreover, another study comparing 7- to 13-year-olds with OCD, nonclinical children, and children with other (nonOCD) anxiety disorders (Barrett & Healy, 2003) found that biases of responsibility, threat severity, TAF, and cognitive control were significantly higher in children with OCD and other anxiety disorders compared to nonclinic controls. Cognitive control was the only bias however, found to be specific to children with OCD. Thus, while there is support for cognitive models in understanding childhood OCD (Reynolds & Reeves, 2008), the inconsistent findings to date warrant further examination. More recently, the role of responsibility in nonclinical children (n = 81; ages 9–12 years) was assessed during a sorting task, whereby responsibility was manipulated as a function of whether an adult was assigned to check the children's sorting or not (Reeves, Reynolds, Coker, & Wilson, 2010). The manipulation was successful, with significantly higher ratings of perceived responsibility in the

high-responsibility condition. Furthermore, this was associated with increased time taken to complete the task, increased checking, and more frequent hesitations, providing support for a role of responsibility in checking behaviors. However, unexpectedly there were no between-group differences on state anxiety, raising the question about the specificity of responsibility bias to anxiety and/or the development of clinical OCD. Coles and colleagues (2010) recently published the initial validation of the Obsessive-Belief Questionnaire–Child Version (OBQ-CV), which in the future will allow for standardized and reliable testing of obsessional beliefs in children across six core belief domains identified by the OCCWG (1997, 2001). The results provide support for a relationship between beliefs (OBQ-CV total scores) and self-reported OCD severity in two samples of children (8–18 years). Interestingly, however, total scores were not significantly correlated with OCD severity based on the Children's Yale–Brown Obsessive–Compulsive Scale (CY-BOCS; Scahill et al., 1997) total scores or compulsion scores, and there was only borderline significance in one sample and a nonsignificant trend in the other sample with the CY-BOCS obsession scores. While the child studies collectively provide preliminary support for a role of maladaptive beliefs, the results are somewhat inconsistent across samples, study methodologies, and age groups. Research specifically exploring the origins of obsessive beliefs may offer more prevailing information regarding the role of maladaptive beliefs in the development of OCD symptoms in childhood. Researchers have suggested that maladaptive beliefs may have their origins within a familial-based developmental context (Rector, Cassin, Richter, & Burroughs, 2009), whereby specific parenting behaviors, such as criticism, control, and strict codes of conduct have been postulated to be associated with the development of maladaptive perfectionism (e.g., Kawamura, Frost, & Harmatz, 2002) and inflated responsibility beliefs (e.g., Salkovskis, Shafran, Rachman, & Freeston, 1999). Rector and colleagues examined familial vulnerability for OCD by comparing maladaptive beliefs between adults with OCD, their nonaffected first-degree relatives, and nonaffected controls. Firstdegree relatives scored significantly higher than controls on inflated responsibility and overestimation of threat. Furthermore, relatives of adults with early onset OCD also scored significantly higher than controls on both inflated responsibility and overestimation of threat, suggesting a familial-based vulnerability for the development of certain maladaptive beliefs.

cognitive bias and ocd symptoms in children When relationships between cognitive biases in children (ages 9–17 years; n = 28) with OCD and their mothers have been investigated, a moderate and positive correlation was found between children's and mother's beliefs regarding inflated responsibility and overestimation of threat (Pietrefesa, Schofield, Whiteside, Sochting, & Coles, 2010). However, perfectionism and certainty beliefs were not significantly correlated, and unexpectedly, beliefs regarding the importance and control of thoughts were negatively correlated. These findings for responsibility biases and overestimation of threat are consistent with Rector et al. (2009), and suggest that familial loading may be particularly strong for specific maladaptive beliefs, namely, responsibility and threat beliefs. Coles and Schofield (2008) have developed and validated the Pathways to Inflated Responsibility Beliefs Scale (PIRBS; Coles & Schofield, 2008), which assesses five hypothesized familial-based developmental contexts that are proposed by Salkovskis and colleagues (1999) as important in the developmental of maladaptive beliefs of responsibility (e.g., heightened responsibility as a child, rigid and extreme codes of conduct). The use of this scale in future studies will greatly assist in developing our understanding of the origins of responsibility biases in adults with OCD. Moreover, research examining these issues in children is important to ascertain the developmental vulnerability for maladaptive biases associated with OCD. The major aim of the current study was to extend knowledge into the applicability of cognitive models to childhood OCD by examining the influence of family (mothers specifically) and age on children's cognitions and OCD symptoms. Specifically, the current study had three aims. The first aim was to determine whether child and adolescent OCD symptom severity (measured by the CY-BOCS; Scahill et al., 1997) was associated with cognitive biases of responsibility, TAF, thought suppression, and meta-cognitive beliefs, across a clinical sample of children and adolescents with OCD. These belief domains were chosen as they have received the most attention to date in the child literature (e.g., Farrell & Barrett, 2006; Reynolds & Reeves, 2008). The second study aim was to examine the relationship between mothers’ maladaptive beliefs and children's OCD severity. The third and final study aim was to examine whether there are associations between maternal cognitive beliefs and children's cognitive beliefs across multiple belief domains, replicating the study by Pietrefesa et al. (2010). Developmental research has highlighted critical differences in children's ability to appraise their social worlds and regulate their cognition, emotion,

595

and behavior across childhood versus adolescence, with these skills becoming more sophisticated, differentiated, organized, and self- regulated rather than other-regulated with increasing age (Calkins & Bell, 2010; Skinner & Zimmer-Gembeck, 2009; Zimmer-Gembeck & Skinner, 2011). Hence, it is not surprising that there has been disparity in results across different age samples in past research. To address this, we explored the above associations for children and adolescents separately. We examined age as two developmental groups (child vs. adolescent) rather than as a continuous measure, given previous evidence that differences can be found when comparing findings for children under age 12 to those between ages 12 and 17 (Farrell & Barrett, 2006). We hypothesised that the relationships between variables would be moderated by age, making three specific hypotheses. First, the associations between child cognitive biases and OCD severity were expected to be stronger among adolescents (ages 12–17 years) compared to children (ages 7– 11 years) based on past findings (i.e., Farrell & Barrett, 2006). Second, we expected that mothers’ cognitive biases would be associated with OCD severity, and this association would be stronger for children versus adolescents. This hypothesis was founded on developmental research suggesting that children's emotions and cognitions are more susceptible to socialization by parents, compared with adolescents’ emotions and cognitions (Aldwin, 2007; Zimmer-Gembeck & Skinner, 2011). Third, we hypothesised that maternal cognitive bias would be associated with child cognitive bias in both age groups, in line with a family vulnerability hypothesis. Moreover, given that maternal OCD symptoms might contribute to a family context that models and supports OCD-related cognitive biases and ritualized behavior, the associations of interest (i.e., one to three above) also were investigated within a subsample of dyads where mothers had either clinical or subclinical OCD across both age groups. It was hypothesised that there would be significant and large associations within this subsample between maternal cognitive biases and child OCD, and between maternal cognitive bias and child cognitive biases.

Method participants Participants were 46 children (7–11 years, n = 26) and adolescents (12–17 years, n = 20), with a mean age of 11.30 years (SD = 2.86), comprised of 30 males and 16 females, who were consecutively referred for free treatment of OCD to Griffith University from health professionals and parents

596

farrell et al.

following media announcements. Participants were selected into this study on the basis of a Diagnostic and Statistical Manual of Mental Disorders (DSMIV-TR; American Psychiatric Association, 2000) diagnosis of OCD, and suspected IQ within at least average range (based on parental screen). Exclusion criteria included psychosis, intellectual disability, mental retardation, or receiving concurrent psychotherapy. There were no referrals to the project during this time that met exclusion criteria. All children were offered free treatment following their assessment. Eight-three percent of the sample (n = 38) had a principal diagnosis of OCD, based on clinician severity rating (CSR) of the Anxiety Disorders Interview Schedule for Children–Parent Version (ADIS-P; Silverman & Albano, 1996), with 11% presenting with a secondary diagnosis of OCD (n = 5) and 4% presenting with a tertiary diagnosis of OCD (n = 2). This sample was typical of pediatric OCD, consisting of high comorbidity, with 85% presenting with a secondary diagnosis and 63% presenting with a tertiary diagnosis. Table 1 presents diagnosis information for the sample, including principal secondary and tertiary diagnoses. 1 In regard to severity of the sample, the mean CY-BOCS (Scahill et al., 1997) rating was 22.53 (SD = 7.66) indicating the sample was overall within the upper range of moderate severity. Severity of OCD did not differ across child (M = 21.24, SD = 8.16) versus adolescent (M = 23.96, SD = 6.93) age groups; however, the adolescent group reported significantly higher severity on the obsession (M = 13.16, SD = 3.11) subscale only, relative to children (M = 9.77, SD = 4.37), F(1,52) = 8.91, p b 0.005. Four mothers (9%) reported clinical OCD and a further 14 mothers (30%) reported subclinical OCD symptoms (without significant duration or interference; confirmed using the adult OCD diagnostic interview section from the Anxiety Disorders Interview Schedule for DSM-IV–Adult Version [ADIS-Adult]; Brown, Di Nardo, & Barlow, 1994). Within this subsample of children who had a mother with subclinical or clinical OCD (n = 18), the mean age of children was 11.11 years (SD = 2.4, age range 7–15 years).

measures Diagnostic and Symptom Severity The ADIS-P (Silverman & Albano, 1996) was developed specifically to diagnose anxiety disorders 1

Given that the sample consisted of eight children who presented with a nonprimary OCD diagnosis, we ran the analyses both with and without these children. The results did not differ substantially with these children removed; hence, they are included as they provide more power to the study.

Table 1

Participant Diagnostic Presentation on Assessment at Pretreatment Diagnosis

OCD GAD SAD SoPH SpPH MDD DYS ADHD ODD PDD AGO PTSD TOTAL

Principal Diagnosis

Secondary Diagnosis

Third Diagnosis

N (%)

N (%)

N (%)

38 (83%) 1 (2%) 2 (4%) 0 1 (2%) 0 0 1 (2%) 0 3 (7%) 0 0 46 (100%)

5 (11%) 12 (22%) 1 (2%) 5 (11%) 5 (11%) 2 (4%) 1 (2%) 3 (7%) 1 (2%) 5 (11%) 1 (2%) 0 39 (85%)

2 (4%) 7 (15%) 2 (4%) 4 (9%) 4 (9%) 3 (7%) 1 (2%) 4 (9%) 0 1 (2%) 0 1 (2%) 29 (63%)

Note. OCD = obsessive–compulsive disorder, GAD = generalized anxiety disorder, SAD= separation anxiety disorder, SoPH= social phobia, SpPH= specific phobia, MDD = major depressive disorder, DYS = dysthymic disorder, ADHD = attention-deficit/hyperactivity disorder, ODD =oppositional defiant disorder, PDD = pervasive developmental disorder, AGO = agoraphobia, PTSD = post-traumatic stress disorder.

in children (Silverman & Eisen, 1992), and has good interrater and test–retest reliability. The ADIS-C/P has demonstrated good sensitivity to treatment effects in both childhood anxiety (Barrett, Dadds, & Rapee, 1996; Kendall, 1994) and childhood OCD research (Albano, Knox, & Barlow, 1996; Barrett, Healy-Farrell, & March, 2004; Waters, Barrett, & March, 2001). This interview was administered to the child's parent(s) in clinic to ascertain an OCD diagnosis. Diagnoses were confirmed based on a clinician severity rating of 4 (from 0 to 8) or above. Maternal OCD status was assessed via a screen for OCD diagnosis, using the OCD section of the ADISAdult (Brown et al., 1994). Interrater reliability has been conducted across 20% of the videotaped diagnostic interviews by an independent rater, with results indicating excellent reliability (primary diagnosis κ = 1.0, secondary diagnosis κ = 0.84, tertiary diagnosis κ = 0.83). The CY-BOCS (Scahill et al., 1997) is a widely used, clinician-rated, semistructured interview. The CY-BOCS rates severity of obsessions and compulsions across five scales: (a) time occupied by symptoms, (b) interference, (c) distress, (d) resistance, and (e) degree of control over symptoms, and also provides a total severity score. The CY-BOCS shows reasonable reliability and validity, with good to excellent interrater agreement and high internal consistency for total score (Scahill et. al., 1997). This interview was administered to children

cognitive bias and ocd symptoms in children (including parents for the younger sample, 7– 11 years) to assess overall OCD symptom severity. Cognitive Beliefs Questionnaire Package The Responsibility Attitudes Scale (RAS; Salkovskis et al., 2000) is a 26-item questionnaire designed to assess general beliefs about responsibility. Examples of items include “I often feel responsible for things that go wrong,” and “I am often close to causing harm.” Internal consistency for the RAS is high, with Cronbach's α of .92. Similarly, the RAS has excellent test–retest reliability of .94, and adequate concurrent validity with other measures of obsessionality (Salkovskis et al., 2000). Internal consistency within this sample was .88 for children, .98 for adolescents, and .90 for mothers. The revised TAF scale (Shafran, Thordarson, & Rachman, 1996) assesses the degree to which clients interpret intrusive thoughts as having personal significance, across both morality TAF and likelihood TAF. The scale consists of 19 items (rated 0 to 4). Twelve items evaluate morality (e.g., thinking blasphemous thoughts is almost as sinful as blasphemous actions), four items assess TAF for likelihood of negative events happening to others (e.g., thinking of a friend falling ill increases the chances that he or she will fall ill), and three items assess TAF for likelihood of negative events happening to oneself (e.g., if I think of myself being injured in a car accident, this will increase the chances that I will be injured). Studies have demonstrated that TAF is a highly reliable construct across student, adult, and obsessional samples and internal consistency has been shown to be excellent (Shafran et al., 1996). Internal consistency was .96 for children, .95 for adolescents, and .92 for mothers. The White Bear Suppression Inventory (WBSI; Wegner & Zanakos, 1994) assesses the degree to which people experience and attempt to suppress intrusive thoughts. The measure has 15 items, with response options from 1 (strongly disagree) to 5 (strongly agree). Examples of items include “There are things I would prefer not to think about” and “I have thoughts that I cannot stop.” The WBSI has demonstrated excellent internal stability, good test– retest reliability, and adequate convergent validity with measures of anxiety, depression, and OCD (Wegner & Zanakos, 1994). Internal consistency across items within this scale was .89 for children, .96 for adolescents and .95 for mothers. The Meta-Cognition Questionnaire for Adolescents (MCQ-A; Cartwright-Hatton et al., 2004) measures individual differences in a selection of metacognitive beliefs. This 30-item measure has five factors including cognitive confidence, positive beliefs about worry, cognitive self-consciousness,

597

negative beliefs about uncontrollability of thoughts and danger, and beliefs about need to control thoughts. The MCQ-A has demonstrated good internal consistency, convergent validity, and acceptable to good test–retest reliability. Furthermore, positive relationships between metacognitions and measures of emotional symptoms support the validity of this measure (see Wells & CartwrightHatton, 2004). This measure has been extended for use with children (ages 7–17 years) and demonstrated similar psychometrics (Barlow, Pincus, Ehrenreich, & Brody, 2009). For this paper, the acronym META will refer to the MCQ-A for ease of interpretation. Internal consistency was .87 for children, .92 for adolescents, and .92 for mothers. All adult cognitive questionnaires were adapted for use with children by way of minor changes to wording to ease interpretation by children and youth (see Farrell & Barrett, 2006).

procedure Following referral, participants were screened over the telephone for eligibility into the study via a brief parent interview assessing for obsessive–compulsive symptomatology. If eligible, families attended an assessment at the university psychology clinic, conducted by the first author and a postgraduate clinically trained experimenter. On attending this interview, the research aims were explained to all participants and written informed consent was gained from parents. Initial assessment interviews involved ADIS-P interviews with parents and the CY-BOCS interview with children (including parents for younger children ages 7–11 years). Interviewers were previously trained in diagnostic interviews and CY-BOCS interviews. All procedures in this study had prior ethics clearance through the university human research ethics committee. During interviews with parents, parental diagnostic status of OCD was assessed via structured probe questions that assessed parental report of OCD symptoms, whether a past diagnosis had been made, and whether there were current subclinical or clinical OCD symptoms. In cases where parents reported some possible current OCD symptoms, a diagnostic interview for OCD was conducted following the ADIS-Adult interview OCD section (ADIS –Adult; Brown et al., 1994). Following diagnostic and symptom severity interviews, children were interviewed alone using the cognitive belief questionnaire package outlined above (included the RAS, TAF, WBSI, and META), which were read to children (ages 7–11 years) or given to adolescents to read and answer without assistance. Mothers were given the same adult versions of the cognitive belief questionnaires, and were asked to complete these while their children

598

farrell et al.

completed the forms with the examiner. After the initial interview and pretreatment assessment tasks, families were offered either individual or group cognitive-behavioral treatment free of charge.

Results overview of the analyses Because it was hypothesised that child age and maternal OCD status would each be important to consider when examining associations among child cognitive biases, maternal cognitive biases, and child OCD symptoms, correlations were examined among relevant subsamples addressing each of the three study hypotheses. These hypotheses were (a) child cognitive biases would be associated with OCD severity, which would be moderated by age; (b) maternal cognitive biases would be associated with OCD severity, which would be moderated by age; and (c) maternal cognitive biases would be associated with child and adolescent cognitive biases. Subsamples included data from (a) children (ages 7– 11 years), and (b) adolescents (ages 12–17 years), and also examined (c) child and mother dyads where mothers had either subclinical or clinical OCD within the child sample (n = 8) and the adolescent sample (n = 10). Moderated regression analyses were conducted where there were significant associations to examine whether there were significant age-group differences. To test these hypotheses multiple correlational analyses were conducted. While this may result in inflated Type 1 error, it has been argued that traditional Bonferroni methods with small clinical sample sizes further reduces statistical power, and therefore should not be routinely employed (see

Jaccard & Guilamo-Ramos, 2002). As such, this study employs Holm modified Bonferroni corrections (Holm, 1979) that are applied within each family of analyses corresponding to the study hypotheses (i.e., Hypothesis 1, six correlations; Hypothesis 2, six correlations; and Hypothesis 3, 36 correlations). The results largely did not differ when comparing before to after these corrections were made, except in regard to the third hypothesis, which requires applying α at .001 in the first instance. Given that there were two large effect sizes (r) found within the associations for Hypothesis 3 and they were significant at p b .01, we report these as their magnitude indicates meaningful and robust effects. Correlations that were significant at only p b .05 are, however, rejected based on the applied alpha corrections. Table 2 presents means and standard deviations for all measures for both the child and adolescent sample. Child Sample Hypothesis 1. When examining correlations between variables for the child subsample only, children's own cognitive biases were not associated with their OCD symptom severity. Hypothesis 2. Mothers’ cognitive biases were, however, associated with their children's OCD symptom severity. Maternal RAS was strongly associated with child OCD severity (r = .72, p b .008), as was maternal WBSI (r = .51, p b .01). Hypothesis 3. Regarding associations between mothers’ cognitive biases and their children's biases, there were a number of significant associations; however, with the alpha correction applied, two effects

Table 2

Means and Standard Deviations for Mothers and Children Across the Entire Sample, and for Children and Adolescents Measures

CY-BOCS total severity Mother TAF moral Child TAF moral Mother TAF other Child TAF other Mother TAF self Child TAF self Mother WBSI Child WBSI Mother RAS Child RAS Mother META Child META

Entire Sample

Children 7–11 years

Adolescents 12–17 years

(child n = 46; parent n = 42)

(child n = 24; parent n = 22)

(child n = 22; parent n = 20)

22.53; 7.66 18.29; 10.47 21.17; 12.52 3.50; 3.37 5.07; 4.26 3.19; 2.63 3.67; 3.46 44.38; 12.84 51.93; 13.49 110.32; 22.25 106.27; 31.65 56.00; 14.69 62.87; 15.67

21.24; 8.16 16.95; 8.78 22.63; 2.69 3.41; 3.23 5.25; 3.76 3.27; 2.51 3.58; 3.02 44.00; 13.88 52.21; 10.14 107.81; 25.42 106.38; 21.58 54.77; 15.49 59.92; 13.24

23.96; 6.93 19.75; 12.14 19.59; 2.52 3.60; 3.56 4.86; 4.83 3.10; 2.83 3.77; 3.96 45.85; 11.83 51.64; 16.64 112.95; 18.63 106.14; 40.85 57.35; 14.03 66.24; 17.87

Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, TAF = Thought–Action Fusion Scale, WBSI = White Bear Suppression Inventory, RAS = Responsibility Attitudes Scale, META = Metacognitive Beliefs Questionnaire; there were four less parents involved in this study due to missing/incomplete questionnaires.

599

cognitive bias and ocd symptoms in children were significant at p b .01. Maternal TAF was strongly correlated with child WBSI (r = .58, p b .01), and furthermore, maternal META was strongly correlated with child TAF moral (r = .55, p b .01).

Maternal OCD Subsample Hypothesis 1. When analyses were limited only to those families who had mothers reporting some level of OCD symptoms (clinical or subclinical OCD) within the child subsample, there were no associations between child cognitive biases and child OCD severity. Hypothesis 2. There were, however, very large positive correlations between maternal cognitive biases and child OCD severity, with maternal RAS strongly and positively correlated with child OCD severity (r = .93, p b .008), and maternal WBSI also strongly and positively correlated with child OCD severity (r = .90, p b .01). Hypothesis 3. There were no significant correlations between maternal cognitive biases and child cognitive biases. Table 3 presents the observed correlations and statistical significance among variables of interest.

Adolescent Subsample Hypothesis 1. When examining intercorrelations among variables for the adolescent subsample, there were significant correlations between adolescent cognitive biases and OCD symptoms severity that were unlike the findings observed for the child sample. Adolescent META (r = .60, p b .008), WSBI (r = .57,

p b .01), and RAS (r = .55, p b .01) were each strongly and positively correlated with adolescent OCD severity. Hypothesis 2. In regard to maternal cognitive biases and adolescent OCD severity, there were moderate associations; however, these were not significant following corrections for Type 1 error. Of note, however, the direction of the associations were opposite to those observed in children, such that higher adolescent OCD severity was associated with lower ratings of maternal bias (i.e., META, WBSI, RAS, TAF self). Hypothesis 3. Interestingly, and again unlike that observed in the younger subsample, there were no significant correlations between maternal cognitive variables and adolescent cognitive variables.

Maternal OCD Subsample Hypothesis 1. When analyses were limited only to those families who had mothers reporting some level of OCD symptoms (clinical or subclinical OCD) within the adolescent subsample, there were large and positive associations between adolescent cognitive biases and OCD severity. Specifically, adolescent RAS was significantly correlated with OCD severity (r = .87, p b .008), as was adolescent WBSI (r = .77, p b .013). Hypothesis 2. Contrary to outcomes within the child subsample, there were, however, no significant association between maternal cognitive biases and adolescent OCD severity.

Table 3

Correlations Between Measures of Child Cognitive Bias, Maternal Cognitive Bias and Child OCD Severity for Children and Adolescents

OCD severity child OCD severity Youth Child RAS Youth RAS Child WBSI Youth WBSI Child META Youth META Child TAF self Youth TAF Self Child TAF other Youth TAF other Child TAF Moral Youth TAF Moral

OCD Severity (CY-BOCS)

Maternal RAS

Maternal WBSI

Maternal META

Maternal TAF Self

Maternal TAF Other

Maternal TAF Moral

1 1 –.16 .55* .11 .56*** –.13 .56*** –.14 .21 –.15 .15 .19 .09

.72*** –.43 –.23 .31 .15 –.15 .10 –.32 –.03 –.04 –.14 –.01 .29 .10

.51* –.49 –.24 .17 .25 –.10 .25 –.29 .05 0.14 –.02 .05 .44 –.02

.40 –.52* –.34 .18 .44 –.21 .39 –.44 .27 –.23 .16 –.26 .55*** –.19

.36 –.46 .19 .01 .58*** –.11 .05 .03 .09 .02 .06 –.05 –.02 .01

.25 –.21 .06 –.02 .31 –.05 .19 –.19 .02 .28 .04 .23 .05 .04

–.02 .02 –.09 .12 –.03 –.17 .26 –.10 .11 –.39 .26 –.42 –.11 .39

Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, TAF = Thought–Action Fusion Scale, WBSI = White Bear Suppression Inventory, RAS = Responsibility Attitudes Scale, META = Metacognitive Beliefs Questionnaire. * significance at 0.01, ** significance at 0.013, *** significance at 0.008.

600

farrell et al. 30

Hypothesis 3. There were no significant correlations between maternal cognitive biases and child cognitive biases.

Child Cognitive Bias and OCD Symptom Severity—Moderated by Age For child cognitive variables, age moderated the association between child RAS and OCD severity, β = −.56, t(44) = −1.96, p b .06; and between child META and OCD severity, β = .50, t(44) = 2.21, p b .05. The moderation regression equation was not significant for WBSI. As can be seen in Figures 1 and 2, the relationship of child RAS and child META with OCD severity was only significant for the adolescent subsample, such that higher responsibility and metacognition were each associated with more severe OCD in adolescents. Parent Cognitive Bias and OCD Symptom Severity—Moderated by Age Similar moderated regression analyses were estimated to examine maternal cognitive biases, OCD symptom severity, and child age group. Age moderated the association between maternal WBSI and OCD severity, β = −.61, t(41) = − 3.47, p b .001; maternal RAS and OCD severity, β = .66, t(40) = 4.26, p b .001; and maternal META and OCD severity, β = −.60, t(41) = −3.16, p b .01. These analyses revealed that higher maternal cognitive bias was

CY-BOCS

20 –.16 15 10 5 0 Low RAS child

High RAS child

Young (< 12 years)

Older (>11 years)

FIGURE 1 Relationship between child RAS scores and OCD severity as measured by the CY-BOCS, moderated by age. Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, RAS = Responsibility Attitudes Scale. * p b 0.06.

associated with more severe child OCD, but in contrast, higher maternal cognitive biases were associated with less severe adolescent OCD symptoms (see Figures 3, 4, and 5). Child Cognitive Bias and Maternal Cognitive Bias—Moderated by Age Given the contrary associations of child and parent cognitive variables with OCD symptoms for the two age groups, a series of moderated regression analyses were also estimated to examine whether the associations between maternal and child cognitive biases differed in the child versus the adolescent subsample. These analyses were only conducted with those variables that had significant simple correlations (i.e., child TAF moral with maternal META, and child WBSI with maternal TAF self).

.60*

30

–.13

25

CY-BOCS

moderated regression analyses Given the different patterns of correlations between cognitive variables and OCD severity when comparing findings for the child and adolescent subsamples, a series of moderated regression analyses were estimated to test whether associations between cognitive biases and OCD severity were significantly different between children and adolescents. With OCD severity as the dependent variable, regression analyses for each of the cognitive biases were conducted separately (where there were significant associations observed within either the child or adolescent sample), moderated by age group. Moderation effects were constructed by calculating product terms—age group (0 = child, 1 = adolescent) multiplied by cognitive bias score. The product terms (one for each cognitive bias) were produced after centering each measure of cognitive bias to reduce the likelihood of multicollinearity in the regression models, as recommended by Jaccard, Turrisi, and Wan (1990). In each model, OCD severity was the dependent variable, and the independent variables were a centered measure of child cognitive bias, age group (0 = child and 1 = adolescent), and the product term (Cognitive Bias × Age Group). To reduce length, only significant moderation effects are described.

.55* 25

20 15 10 5 0 Low Meta child Young (< 12 years)

High Meta child Older (>11 years)

FIGURE 2 Relationship between child META scores and OCD severity as measured by the CY-BOCS, moderated by age. Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, META = Metacognitive Beliefs Questionnaire. * p b 0.05.

601

cognitive bias and ocd symptoms in children 30

30

–0.52* –0.49*

25

20

CY-BOCS

CY-BOCS

25

0.51* 15

20

0.40*

15

10

10

5

5 0

0 Low WBSI parent

Young (< 12 years)

Low Meta parent

High WBSI parent

High Meta parent

Young (< 12 years)

Older (>11 years)

Older (>11 years)

FIGURE 3 Relationship between maternal WBSI scores and OCD severity as measured by the CY-BOCS, moderated by age. Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, WBSI = White Bear Suppression Inventory. * p b 0.001.

FIGURE 5 Relationship between maternal META scores and OCD severity as measured by the CY-BOCS, moderated by age. Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, META = Metacognitive Beliefs Questionnaire. * p b 0.01.

When child TAF moral was the dependent variable, age moderated the association with maternal META, β = −.48, t(41) = −2.43, p b .05. This finding revealed that for younger children there was a significant positive association between child TAF moral and maternal META, such that a higher child cognitive rating was associated with a higher maternal rating. In contrast, this association was negative in the adolescent subsample indicating that a higher maternal rating was associated with a lower adolescent cognitive rating (see Figure 6). When child WBSI was the dependent variable, age moderated the association with maternal TAF self; however, these effects were marginal, β = −.46, t(39) = −2.02, p b .06. These results demonstrated that higher maternal TAF self was associated with higher child WBSI for the child subgroup; however, the

opposite association was found for the adolescent subgroup (see Figures 6 and 7).

Discussion Cognitive factors are implicated as central mechanisms in theoretical models of OCD in adults; however, the nature of these processes in pediatric OCD remains largely unknown. The results of the current investigation extend current research and highlight the importance of considering both age and maternal OCD status when investigating the role of maladaptive beliefs and the influence of parents’ beliefs on children's OCD symptoms. Overall, the results of this study suggest that age significantly moderates associations between child cognitive biases, maternal cognitive biases, and child OCD severity in three important ways. First, and consistent with our hypotheses, age moderated associations between child cognitive variables and

30 –0.43*

200

20 15

0.55*

150 0.72*

Child TAF

CY-BOCS

25

10 5 0 Low RAS parent Young (< 12 years)

High RAS Parent Older (>11 years)

100 –0.20 50

0 Low Meta parent -50 Young (< 12 years)

FIGURE 4 Relationship between maternal RAS self scores and OCD severity as measured by the CY-BOCS, moderated by age. Note. CY-BOCS = Children's Yale–Brown Obsessive–Compulsive Scale, RAS = Responsibility Attitudes Scale. * p b 001.

High Meta parent

Older (>11 years)

FIGURE 6 Relationship between child TAF moral and maternal META, moderated by age. Note. TAF = Thought–Action Fusion Scale, META = Metacognitive Beliefs Questionnaire. * p b 0.05.

602

farrell et al. 65 60

Child WBSI

–0.11 55 50

0.58*

45 40 Low TAF self - parent

Young (< 12 years)

High TAF self - parent

Older (>11 years)

FIGURE 7 Relationship between child WBSI and maternal TAF self, moderated by age. Note. WBSI = White Bear Suppression Inventory, META = Metacognitive Beliefs Questionnaire. * p b 0.06.

OCD severity—specifically between child responsibility and OCD severity, and child metacognition and OCD severity—demonstrating that associations between children's maladaptive beliefs and OCD severity were only significant for the adolescent subsample, with no significant associations present in the younger child subsample. Of note, our age groups were not significantly different on overall OCD severity, indicating that the observed age differences are not simply associated with increased OCD severity. The adolescent subsample was, however, significantly higher on obsessional ratings of OCD, suggesting that OCD in adolescence is more likely to have a stronger obsessional component, which may explain the development of maladaptive beliefs during this time. In this way, it may be that cognitive biases are not in fact important in terms of etiological perspectives of childhood OCD but rather develop as a result of obsessive–compulsive symptoms that become more cognitive and obsessionally driven in adolescence. For example, Verhaak and de Hann (2007) propose that cognitive biases may be best understood as an “attribution of meaning that takes place after the perceived need to perform a compulsive behaviour, an intellectual maneuver that develops in adolescence (p. 360).” We also considered that age of onset of OCD and duration of OCD may play a role in the development of biases, such that earlier onset OCD or longer OCD duration may at least partly account for stronger associations. Inspection of the data, 2

2

These data are not presented in the current paper due to the already lengthy presentation of variables and data. Results are available from the first author upon request.

however, indicated that there were no significant differences between age subsamples in terms of age of onset, or duration of illness, indicating a developmental stage difference versus being a function of illness course. These findings have important implications for treatment, such that cognitive approaches are more likely to be relevant for treating adolescents with OCD, however, may be less relevant for the treatment of OCD in younger children. Instead, younger children may benefit more from a predominantly family based, behavioral approach to managing OCD. Second, and again consistent with our hypotheses, age significantly moderated the associations between maternal cognitive biases and child OCD severity. Maternal responsibility and thought suppression were significantly and positively associated with child OCD severity; however, there were no significant associations between maternal beliefs and adolescent OCD severity. Of interest, however, and unexpectedly, we found that maternal beliefs were negatively associated with OCD severity in adolescents, with lower maternal ratings of cognitive bias associated with increased OCD severity in youth. Finally, contrary to our hypothesis, age also moderated the associations between child cognitive bias and maternal cognitive bias, specifically child TAF moral and thought suppression with maternal metacognition and TAF self, respectively. For the child subsample, there was a significant positive association between child TAF moral and maternal metacognition, and between child suppression and maternal TAF self, such that higher maternal ratings were associated with higher child cognitive ratings. In contrast, these associations were negative in direction (although nonsignificant) for the adolescent subsample. Furthermore, inconsistent with Pietrefesa et al. (2010) and Rector et al. (2009), this study did not find a significant relationship between mother and child responsibility biases. The adolescent associations that were observed to be in the opposite direction to what was expected are somewhat consistent with findings from Pietrefesa and colleagues (2010) who also found negative associations among children with OCD and their parents on ratings of importance of thoughts and control of thoughts in their sample ages 9–17 years. One hypothesis that is shared by Pietrefesa and colleagues is that mothers may be masking or hiding their own experiences of these maladaptive beliefs. It is intriguing in our study that there was a clear interaction effect for age, suggesting that mothers of adolescents might be more likely to deny these symptoms in themselves, as they become more

cognitive bias and ocd symptoms in children evident in their child during adolescence. An alternate hypothesis for the moderating effects of age on child and maternal cognitive biases is that there may be a reciprocal influence of cognition that exerts its effects on both child and mothers’ threat appraisals. For example, results suggest that maternal cognitive biases may play a pivotal role in the development and/or maintenance of OCD in younger samples, which might then lead to the development of more obsessional symptoms and child maladaptive biases in adolescence. It may be that increased obsessional concerns and maladaptive biases in adolescents have an ameliorating affect on mothers’ own biases. Longitudinal studies are needed to explore the interacting effects of maternal maladaptive biases and children's OCD symptoms. When examining data for the subsample of dyads where a mother reported some degree of OCD symptoms, the results were almost identical to the larger sample, except that significant associations were more robust among this subgroup. These findings demonstrate that associations among child and maternal cognitive biases and child OCD severity differ as a function of children's age, and these patterns of results are even stronger when mothers have OCD symptoms. In regard to clinical implications, these results lend support for the importance of involving families in the treatment of pediatric OCD, especially with a preadolescent sample (e.g., Farrell & Barrett, 2007; Piacentini & Langley, 2004). Based on the current study, findings suggest that treatment for younger children may involve incorporating cognitive approaches to the parent component of therapy, rather than as an addition to children's treatment. This approach deviates from traditional parental involvement in treatment, which usually focuses on parental management skills of OCD and addresses family accommodation of symptoms, but does not address parents’ own cognitive processing of thoughts/worry. Cognitive approaches to parent training might involve teaching parents to identify rationale and flexible cognitive appraisals and cognitive processes, which may then have an influence on the development of children's own cognitive biases and OCD severity. For adolescents, the inclusion of cognitive components to their therapy seems warranted, and should focus on at least addressing maladaptive responsibility beliefs and metacognitive beliefs regarding intrusive thoughts through cognitiveenhanced exposure therapy (e.g., Fama & Whilhelm, 2005), whereby adolescents are assisted to increase awareness of maladaptive beliefs and develop strategies for being more flexible and rational in their appraisals of threat.

603

This study provides an extension to the current literature by investigating the important role of age and maternal cognitive biases, in examining the role of maladaptive beliefs associated with OCD in children and adolescents. Strengths of this study include a relatively large clinical sample with a wide age range allowing for examination of age-group differences, comprehensive assessment of OCDrelated cognitive bias, diagnostic assessment of child OCD, consideration of mother OCD symptoms, and diagnostic screen to verify mothers’ OCD status. This study is not without limitations, including (a) evaluation of only a selection of cognitive biases proposed by the OCCWG (1997, 2001) to be important in cognitive accounts of OCD; (b) the absence of father report; (c) the correlational nature of the data analysis making it impossible to draw conclusions regarding causal roles of cognition; (d) limited power in the current study as a function of the sample size relative to the number of analyses conducted; and (e) reliance on essentially adult measures of self-report of maladaptive beliefs, somewhat limiting the validity of the findings. This study relies on adult measures of cognitive beliefs (apart from the MCQ-A), adapted for children and used in past research (e.g., Farrell & Barrett, 2006), given that the OBQ-CV (or similar child measures) was not available at the time of this study. Moreover, this study relied on child report only during the CY-BOCS interview and parent report only during the diagnostic interview. Given that there is frequently discrepancies between child and parent report on OCD severity (e.g., Canavera, Wilkins, Pincus, & Ehrenreich-May, 2009), it would be a more reliable approach to use interview data from both children and parents in future studies. Future research utilizing longitudinal designs, studies with nonclinical control groups, and incorporating idiographic experimental tasks, as well as child-specific validated measures of beliefs (e.g., OBQ-CV; Coles et al., 2010) to assess cognitive appraisals, would enhance our understanding of the development of OCD-related beliefs in pediatric OCD. In conclusion, this study provides preliminary support for the role of maternal cognitive biases associated with child OCD symptoms and highlights the changing nature of cognitive biases with advancing age in children and adolescents with OCD. These findings are partially consistent with cognitive models of childhood OCD and suggest that psychological approaches to treatment may be improved by targeting maternal cognitive biases in children with OCD, whereas adolescents may benefit from an increased focus on cognitive variables in treatment.

604

farrell et al.

References Albano, A., Knox, L., & Barlow, D. (1996). Parental involvement in the treatment of childhood obsessive–compulsive disorder: A multiple-baseline examination incorporating parents. Behavior Therapy, 27, 93–115. Aldwin, C. M. (2007). Stress, coping, and development: An integrative perspective. New York: Guilford Press. American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: American Psychiatric Press. Barlow, T. L., Pincus, D. B., Ehrenreich, J. T., & Brody, L. R. (2009). The meta-cognitions questionnaire for children: Development and validation in a clinical sample of children and adolescents with anxiety disorders. Journal of Anxiety Disorders, 23(6), 727–736. Barrett, P., Dadds, M., & Rapee, R. (1996). Family treatment of childhood anxiety: A controlled trial. Journal of Consulting and Clinical Psychology, 64, 333–342. Barrett, P., & Healy-Farrell, L. (2003). Perceived responsibility in juvenile obsessive–compulsive disorder: An experimental manipulation. Journal of Clinical Child and Adolescent Psychology, 32, 430–441. Barrett, P. M., Farrell, L., Pina, A. A., Peris, T. S., & Piacentini, J. (2008). Evidence-based psychosocial treatments for child and adolescent obsessive–compulsive disorder. Journal of Clinical Child and Adolescent Psychology, 37(1), 131–155. Barrett, P. M., & Healy, L. J. (2003). An examination of the cognitive processes involved in childhood obsessive– compulsive disorder. Behaviour Research and Therapy, 41, 285–299. Barrett, P. M., Healy-Farrell, L. J., & March, J. S. (2004). Cognitive-behavioral family treatment of childhood obsessive–compulsive disorder: A controlled trial. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 46–62. Boschen, M. J. (2008). Publication trends in individual anxiety disorders: 1980–2015. Journal of Anxiety Disorders, 22, 570–575. Brown, T. A., Di Nardo, P. A., & Barlow, D. H. (1994). The Anxiety Disorder Interview Schedule for DSM-IV; Adult Version: Client Interview Schedule. San Antonio, TX: Psychological Corporation. Calkins, S. D., & Bell, M. A. (2010). Child development at the intersection of emotion and cognition. Washington, DC: American Psychological Association. Canavera, K., Wilkins, K. C., Pincus, D. B., & Ehrenreich-May, J. T. (2009). Parent–child agreement in the assessment of obsessive–compulsive disorder. Journal of Clinical Child and Adolescent Psychology, 38(6), 909–915. Cartwright-Hatton, S., Mather, A., Illingworth, V., Brocki, J., Harrington, R., & Wells, A. (2004). Development and preliminary validation of the Meta-Cognitions Questionnaire– Adolescent Version. Journal of Anxiety Disorders, 18, 411–422. Clark, D. A., & de Silva, P. (1985). The nature of depressive and anxious thoughts: Distinct or uniform phenomena? Behaviour Research and Therapy, 23, 279–282. Coles, M. E., & Schofield, C. A. (2008). Assessing the development of inflated responsibility beliefs: The Pathways to Inflated Responsibility Beliefs Scale. Behavior Therapy, 39, 322–335. Coles, M. E., Wolters, L. J., Sochting, I., de Haan, E., Pietrefesa, A., & Whiteside, S. P. (2010). Development and validation of the Obsessive Belief Questionnaire–Child Version (OBQCV). Depression and Anxiety, 27, 982–991.

Fama, J., & Whilhelm, S. (2005). Formal cognitive therapy: A new treatment for OCD. Concepts and controversies in Obsessive-Compulsive Disorder, Series in Anxiety and Related Disorders, Behavioral Science, 3, 263–281. Farrell, L., & Barrett, P. (2006). Obsessive–compulsive disorder across developmental trajectory: Cognitive processing of threat in children, adolescents and adults. British Journal of Psychology, 97, 95–114. Farrell, L., & Barrett, P. (2007). The function of the family in childhood OCD: Family interactions and accommodation. In E. Storch, G. Geffken, & T. Murphy (Eds.), A comprehensive handbook of child and adolescent obsessive–compulsive disorder (pp. 313–332). Hillsdale, NJ: Erlbaum. Frost, R. O., & Steketee, G. (Eds.). (2002). Cognitive approaches to obsessions and compulsions: Theory, assessment, and treatment Amsterdam, Netherlands: Pergamon/Elsevier Science. Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6(2), 65–70. Jaccard, J., & Guilamo-Ramos, V. (2002). Analysis of variance in clinical child and adolescent psychology: Issues and recommendations. Journal of Clinical Child and Adolescent Psychology, 31, 130–146. Jaccard, J., Turrisi, R., & Wan, C. K. (1990). Interaction effects in multiple regression. Newbury Park, CA: Sage. Kawamura, K. Y., Frost, R. O., & Harmatz, M. G. (2002). The relationship of perceived parenting styles to perfectionism. Personality and Individual Differences, 32, 317–327. Kendall, P. C. (1994). Treating anxiety disorders in youth: Results of a randomized clinical trial. Journal of Consulting and Clinical Psychology, 62, 100–110. Libby, S., Reynolds, S., Derisley, J., & Clark, S. (2004). Cognitive appraisals in young people with obsessive– compulsive disorder. Journal of Child Psychology and Psychiatry, 45, 1076–1084. Obsessive–Compulsive Cognition's Working Group (1997). Cognitive assessment of obsessive–compulsive disorder. Behaviour Research and Therapy, 35(7), 667–681. Obsessive–Compulsive Cognition's Working Group (2001). Development and initial validation of the Obsessive Beliefs Questionnaire and the Interpretation of Intrusions Inventory. Behaviour Research and Therapy, 39, 987–1006. Piacentini, J., & Langley, A. K. (2004). Cognitive-behavioral therapy for children who have obsessive–compulsive disorder. Journal of Clinical Psychology, 60(11), 1181–1194. Pietrefesa, A. S., Schofield, C. A., Whiteside, S. P., Sochting, I., & Coles, M. E. (2010). Obsessive beliefs in youth with OCD and their mothers. Journal of Cognitive Psychotherapy: An International Quarterly, 24(3), 187–197. Rachman, S. (1993). Obsessions, responsibility, and guilt. Behaviour Research and Therapy, 31(2), 149–154. Rector, N. A., Cassin, S. E., Richter, M. A., & Burroughs, E. (2009). Obsessive beliefs in first degree relatives of patients with OCD: A test of the cognitive vulnerability model. Journal of Anxiety Disorders, 23, 145–149. Reeves, J., Reynolds, S., Coker, S., & Wilson, C. (2010). An experimental manipulation of responsibility in children: A test of the inflated responsibility model of obsessive– compulsive disorder. Journal of Behaviour Therapy and Experimental Psychiatry, 41(3), 228–233. Reynolds, S., & Reeves, J. (2008). Do cognitive models of obsessive–compulsive disorder apply to children and adolescents? Behavioral and Cognitive Psychotherapy, 36, 463–471. Rosa-Alcázar, A. I., Sánchez-Meca, J., Gómez-Conesa, A., & Marín-Martínez, F. (2008). Psychological treatments for obsessive–compulsive disorder: A meta-analysis. Clinical Psychology Review, 28(8), 1310–1325.

cognitive bias and ocd symptoms in children Salkovskis, P., Shafran, R., Rachman, S., & Freeston, M. H. (1999). Multiple pathways to inflated responsibility beliefs in obsessional problems: Possible origins and implications for therapy and research. Behaviour Research and Therapy, 37(11), 1055–1072. Salkovskis, P. M. (1985). Obsessional compulsive problems: A cognitive-behavioural analysis. Behaviour Research and Therapy, 23(5), 571–583. Salkovskis, P. M. (1996). Cognitive-behavioural approaches to the understanding of obsessional problems. In R. M. Rapee (Ed.), Current controversies in the anxiety disorders New York: Guilford Press. Salkovskis, P. M., Wroe, A. L., Gledhill, A., Morrison, N., Forrester, E., Richards, C., . . ., & Thorpe, S. (2000). Responsibility attitudes and interpretations are characteristic of obsessive compulsive disorder. Behaviour Research and Therapy, 38(4), 347–372. Scahill, L., Riddle, M. A., McSwiggan-Hardin, M. T., Ort, S. I., King, R. A., Goodman, W. K., . . ., & Leckman, J. F. (1997). The Children's Yale–Brown Obsessive–Compulsive Scale: Preliminary report of reliability and validity. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 844–853. Shafran, R., Thordarson, M. A., & Rachman, S. (1996). Thought–action fusion in obsessive–compulsive disorder. Journal of Anxiety Disorders, 10, 379–391. Silverman, W. K., & Albano, A. M. (1996). Anxiety Disorders Interview Schedule for DSM-IV: Parent Version. San Antonio, TX: Graywing. Silverman, W. K., & Eisen, A. R. (1992). Age differences in the reliability of parent and child reports of child anxious symptomatology using a structured interview. Journal of American Academy of Child and Adolescent Psychiatry, 31 (1), 117–124.

605

Skinner, E. A., & Zimmer-Gembeck, M. J. (Eds.). (2009). Coping and the development of regulation. R. W. Larson & L. A. Jensen (Eds.), New Directions in child and adolescent development San Francisco: Jossey-Bass. Taylor, S., Abramowitz, J. S., & McKay, D. (2006). Cognitivebehavioral models of obsessive–compulsive disorder. In M. M. Antony, C. Purdon, & L. Summerfeldt (Eds.), Psychological treatment of OCD: Fundamentals and beyond Washington, DC: American Psychological Association Press. Waters, T., Barrett, P., & March, J. (2001). Cognitivebehavioral family treatment of childhood obsessive–compulsive disorder: An open clinical trial. American Journal of Psychotherapy, 55, 372–387. Wegner, D. M., & Zanakos, S. (1994). Chronic thought suppression. Journal of Personality, 62(4), 615–640. Verhaak, L. M., & de Hann, E. (2007). Cognitions in children with OCD. A pilot study for age specific relations with severity. European Child and Adolescent Psychiatry, 16, 353–361. Wells, A., & Cartwright-Hatton, S. (2004). A short form of the meta-cognitions questionnaire: Properties of the MCQ-30. Behaviour Research and Therapy, 42, 385–396. Wells, A., & Papageorgiou, C. (1998). Relationships between worry, obsessive–compulsive symptoms and metacognitive beliefs. Behaviour Research and Therapy, 36, 899–913. Zimmer-Gembeck, M. J., & Skinner, E. A. (2011). The development of coping across childhood and adolescence: An integrative review and critique of research. International Journal of Behavioral Development, 35, 1–17.

R E C E I V E D : April 13, 2011 A C C E P T E D : October 30, 2011 Available online 12 November 2011