Cognitive appraisals in young peoplewith obsessive-compulsive disorder

Sarah Libby,1 Shirley Reynolds,2 Jo Derisley,3 and Sarah Clark2

1United Bristol Healthcare Trust, UK; 2University of East Anglia, UK; 3Norfolk Mental Health Care Trust, Bethel Childand Family Centre, UK

Background: A number of cognitive appraisals have been identified as important in the manifestationof obsessive-compulsive disorder (OCD) in adults. There have, however, been few attempts to explorethese cognitive appraisals in clinical groups of young people. Method: This study compared youngpeople aged between 11 and 18 years with OCD (N ¼ 28), young people with other types of anxietydisorders (N ¼ 28) and a non-clinical group (N ¼ 62) on three questionnaire measures of cognitiveappraisals. These were inflated responsibility (Responsibility Attitude Scale; Salkovskis et al., 2000),thought–action fusion – likelihood other (Thought–Action Fusion Scale; Shafran, Thordarson & Rach-man, 1996) and perfectionism (Multidimensional Perfectionism Scale; Frost, Marten, Luhart &Rosenblate, 1990). Results: The young people with OCD had significantly higher scores on inflatedresponsibility, thought–action fusion – (likelihood other), and one aspect of perfectionism, concern overmistakes, than the other groups. In addition, inflated responsibility independently predicted OCDsymptom severity. Conclusions: The results generally support a downward extension of the cognitiveappraisals held by adults with OCD to young people with the disorder. Some of the results, however,raise issues about potential developmental shifts in cognitive appraisals. The findings are discussed inrelation to implications for the cognitive model of OCD and cognitive behavioural therapy for youngpeople with OCD. Keywords: Cognitive models, inflated responsibility, obsessive-compulsive disorder,perfectionism, thought–action fusion. Abbreviations: ADIS-C: Anxiety Disorders Interview Schedulefor Children; ADIS-P: Anxiety Disorders Interview Schedule for Parents; E/RP: Exposure/ResponsePrevention; LOI-CV: Leyton Obsessional Inventory – Child Version; MPS: Multidimensional Perfec-tionism Scale; OCD: Obsessive-Compulsive Disorder; RAS: Responsibility Attitude Scale; TAF-LO:Thought–Action Fusion – (Likelihood Other).

Over recent years there have been proposals thatspecific cognitions are associated with Obsessive-Compulsive Disorder (OCD) in adults (e.g., Obses-sive-Compulsive Cognitions Working Group, 1997;Salkovskis, Forrester, & Richards, 1998; Steketee,Frost, Rheaume, & Wilhelm, 1998) and mountingevidence to support this (e.g., Frost & Steketee,1997; Salkovskis et al., 2000; Shafran et al., 1996).Evaluating whether or not young people with OCDhave similar cognitive appraisals could help informdebates on differences and similarities between OCDin childhood and adulthood. Although cognitiveabilities continue to develop through childhood andadolescence (Kail & Bisanz, 1991; Kuhn, Amsel, &O’Loughlin, 1988), studies with non-clinical samplessuggest that young people hold similar cognitiveappraisals about anxiety as adults (Chorpita,Albano, & Barlow, 1996; Hadwin, Frost, French, &Richards, 1997). Nevertheless, studies have rarelyincluded young people who meet diagnostic criteriafor anxiety disorders.

OCD was once thought to be rare in young people;however, it has now been established that OCD inchildren and adolescents is broadly similar in pre-valence and symptomatology when compared toadults with the disorder. Lifetime prevalence is es-timated at 3% for young people (Valleni-Basille et al.,1994) and 2.5% for adults (Karno, Golding, Sorenson,

& Burnham, 1988). Between half and three-quartersof adults with OCD identify the onset of their symp-toms before the age of 18 years (Pauls, Alsobrook,Goodman, Rasmussen, & Leckman, 1995; Rasmus-sen & Eisen, 1990). OCD is characterised by re-current, intrusive and distressing thoughts orimages (obsessions) that are typically ego-dystonic,going against the value system of the individual.Obsessional thoughts are usually confined to fears ofcontamination, fear of harm to self or others, sexualthemes, aggressive thoughts, religiosity, scru-pulosity and urge for symmetry or exactness (Swedo,Rapoport, Leonard, & Lenane, 1989). Compulsionsare completed in an attempt to alleviate the distressgenerated by the obsessional thoughts. They aremost commonly washing or checking (Riddle et al.,1990; Swedo et al., 1989), but can also present asrepeating, touching, counting, ordering and hoard-ing.

Behaviour therapy comprising Exposure/Re-sponse prevention (E/RP) is currently the dominanttherapeutic approach for young people with OCD(March, 1995). This approach uses graded exposureto the actual or imagined obsessional fear whilstresisting the urge to complete a ritual. E/RP isreported to be helpful for approximately 60% ofyoung people with OCD (De Haan, Hoogduin,Buitelaar, & Keijsers, 1998), but as many as a third

Journal of Child Psychology and Psychiatry 45:6 (2004), pp 1076–1084

� Association for Child Psychology and Psychiatry, 2004.Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

drop out of this treatment (Allsopp & Verduyn,1990). Furthermore, long-term follow-up studies ofyoung people with OCD indicate that it is one of themost intractable anxiety disorders, with approxim-ately 50% having persistent difficulties even withtherapeutic input (Bolton, Luckie, & Steinberg,1995; Flament et al., 1990).

It has been demonstrated that most adultsexperience intrusive thoughts and these are similarin content to those reported by individuals with OCD(Salkovskis & Harrison, 1984). It is hypothesisedthat for those who develop OCD it is the appraisal ofthese thoughts, rather than the thoughts per se, thatcauses distress. This has led to the proposal thatcertain cognitive appraisals are central to the devel-opment and maintenance of the disorder (Salkovskiset al., 1998; Steketee et al., 1998) and need to beaddressed within therapy.

There is a growing evidence base that cognitivetherapy is more effective than E/RP with adults (VanOppen et al., 1995), although when administered ina group format E/RP has been found to be superiorto cognitive therapy alone (McLean et al., 2001).There have been a few case studies indicating thatthese cognitive techniques can also be appliedeffectively with adolescents (Shafran & Somers,1998; Williams, Salkovskis, Forrester, & Allsopp,2002). These are promising findings but at presentthere is no direct evidence that the cognitions asso-ciated with OCD in adults are also widely held byyoung people with OCD.

A number of cognitive appraisals have been pro-posed as critical in OCD. Obsessive CompulsiveWorking Group (1997) suggest that beliefs pertainingto (1) inflated responsibility, (2) overimportance ofthoughts, (3) excessive concern about the importanceof controlling one’s thoughts, (4) overestimation ofthreat, (5) intolerance of uncertainty and (6) perfec-tionism are implicated. To date, the most empiricalevidence has been found for inflated responsibility,overimportance of thoughts and perfectionism andthese are evaluated in the current study.

Salkovskis (1989) suggested that individuals withOCD experience excessive responsibility for intrusivethoughts and external events and believe they couldbe accountable for the perceived consequence. Thisleads to distress and attempts to neutralise theintrusive thoughts by completing compulsive beha-viour. Evidence for inflated responsibility relating toOCD comes primarily from experimental studies(Ladouceur et al., 1995; Ladouceur, Rheaume, &Aublet, 1997) and questionnaires with non-clinicalsamples (Rheaume, Ladouceur, Freeston, & Letarte,1994; Wilson & Chambless, 1999). Salkovskis et al.(2000) is the only published study to compare clin-ical groups. They used the Responsibility AttitudeScale (RAS) to measure belief about responsibility.Adults with OCD had significantly higher scores onthis measure when compared to individuals withother anxiety problems and a non-clinical group.

Although Salkovskis, Shafran, Rachman, and Free-ston (1999) hypothesise that these beliefs emergefrom childhood experiences, this has not yet beenevaluated empirically.

Overestimating the importance of thoughts alsoappears to be critical in the presentation of OCD(Rachman, 1997). The specific error of equating anintrusive thought as increasing the likelihood of anevent occurring to significant others, for example ‘if Ithink of a relative falling ill this increases the riskthat he/she will fall ill’ is defined as Thought–ActionFusion – (Likelihood Other) (TAF-LO) (Shafran et al.,1996). Shafran et al. (1996) found that individualswith OCD had significantly higher scores on TAF-LOwhen compared to a student population and acommunity sample. Rassin, Merckelbach, Muris,and Schmidt (2001), however, found no differencebetween scores on TAF-LO in a group with OCD anda group with other anxiety disorders. More recently,Muris, Meesters, Rassin, Merckelbach, and Camp-bell (2001) developed a measure of thought–actionfusion for adolescents and also found that it wascorrelated with anxiety and depression symptoms aswell as OCD in a non-clinical sample. So, althoughTAF-LO is associated with obsessive-compulsivebehaviours, it may also be found in other disorders.

The other type of cognitive appraisal that has beenfound to be elevated in adults with OCD is perfec-tionism. Perfectionism is often defined as striving forhigh standards of performance that are accompaniedby overly critical evaluations (Antony, Purdon, Huta,& Swinson, 1998). It is underpinned by the beliefthat a perfect state exists and that one should alwaystry to attain this (Pacht, 1984). Research with adultshas predominantly used the Multidimensional Per-fectionism Scale (MPS, Frost et al., 1990). The MPSidentifies six factors, (a) perceiving high parentalexpectations, (b) perceiving parents as overly critical,(c) negative reaction to mistakes, (d) setting highpersonal standards, (e) doubting the quality of one’sactions and (f) emphasising the importance oforganisation. The doubts about action subscale isconfounded with OCD symptoms because theseitems are taken from the Maudsley ObsessionalCompulsive Inventory (Hodgson & Rachman, 1977),so this subscale does not reveal meaningful cognitiveappraisals (Shafran & Mansell, 2001) and will not beconsidered here.

Frost and Steketee (1997) found that a group ofindividuals with OCD had higher total perfectionismscores and concern over mistakes scores on the MPSthan a non-clinical group but were not distinguish-able from a group with panic disorder. Antony et al.(1998) compared a group of adults with OCD toindividuals with panic disorder, social phobia, spe-cific phobia and non-clinical controls. They alsofound that those with OCD did not differ from theother clinical groups on the other dimensions. Allgroups with anxiety disorders were significantlyhigher on concern over mistakes compared to the

Cognitive appraisals in OCD 1077

non-clinical controls. In non-clinical samples con-cern over mistakes is related to sub-clinical obses-sive-compulsive symptoms (Frost et al., 1990; Frost,Steketee, Cohn, & Griess, 1994). So, concern overmistakes appears to be the critical factor in distin-guishing non-anxious and anxious individuals.Parental criticism has also been found to be higher inadults with anxiety disorders compared to non-clinical groups (Antony et al., 1998; Frost & Steke-tee, 1997; Saboonchi, Lundh, & Ost, 1999),although it has not been found to be associated withOCD.

In summary, studies using self-report ques-tionnaires with adults indicate that cognitive ap-praisals associated with inflated responsibility areunique to individuals with OCD. Thought–actionfusion – likelihood other and concern over mistakesmay also be important in our understanding of OCDbut are also held by adults with other anxiety dis-orders. Theorists have proposed that these cognitiveappraisals emerge in childhood (Salkovskis et al.,1999), yet there have been few attempts to evaluatethis empirically. This study used the measures ofinflated responsibility (RAS, Salkovskis et al., 2000),thought–action fusion – (likelihood other) (TAF-LO,Shafran et al., 1996) and perfectionism (MPS, Frostet al., 1990) to evaluate cognitive appraisals in youngpeople with OCD, young people with other anxietyproblems and a non-clinical group of 11- to 18-year-olds. Although adapted measures of thought–actionfusion (Muris et al., 2001) and inflated responsibility(Williams, personal communication) are beingdeveloped specifically for adolescents, these were notavailable at the outset of this project. The study’sprincipal aim was to investigate whether the samepattern of cognitive appraisals found in studies withadults will be observed in this younger population. Asecondary aim of the study is to establish the rela-tionship between cognitive appraisals and the extentthese predict obsessive-compulsive symptoms.

Method

Participants

Three groups of young people aged between 11 and18 years old were recruited for this study. Twenty-eight

had OCD, 28 had other anxiety disorders and 62 werein a non-clinical group. The non-clinical group wererecruited from a High School in the UK. The clinicalgroups were recruited by clinicians from a number ofChild and Adolescent Mental Health Services in the eastof England.

Parents of young people in the non-clinical groupwere asked to report if their child had previous orcurrent problems with anxiety and whether or not theyhad consulted a professional about these problems.Eight families reported problems with anxiety but nonehad sought help from a professional and remained inthe sample as they were classed as non-clinical. Oneyoung person was excluded from the non-clinicalgroup because of a score on the Leyton ObsessionalInventory – Child Version (LOI-CV, Berg, Whitaker,Davies, Flament, & Rapoport, 1988) that was inthe clinical range according to criteria laid down byFlament et al. (1988).

Diagnostic status for the clinical groups was estab-lished using two parallel versions of a structuredinterview for young people and parents. The AnxietyDisorders Interview Schedule for Children and Parents(ADIS-C and ADIS-P, Silverman & Albano, 1996)applies DSM-IV criteria (American Psychiatric Associa-tion, 1994) to diagnose anxiety disorder in childrenaged between 6 and 18 years. For the purposes of thisstudy, only the sections pertinent to the diagnoses forinclusion were used. If a young person met criteria formore than one diagnoses the primary diagnosis wasestablished by considering which produced the mostdistress from symptoms and the greatest interference inthe young person’s life. The diagnoses for the groupwith anxiety disorder were separation anxiety (N ¼ 4),generalised anxiety disorder (N ¼ 6), panic disorderwith agoraphobia (N ¼ 8) and social phobia (N ¼ 10). Allyoung people in the OCD group met criteria for thedisorder according to the ADIS. The level of comorbiditywith other anxiety disorders was high, with all theyoung people in the anxious group meeting criteria formore than one diagnosis and all but two cases (93%)with OCD meeting criteria for more than one diagnosis.The comorbidity across the anxiety disorders is dis-played in Table 1. Young people were excluded if thereferring clinician had diagnosed Asperger’s syndrome,psychosis or drug abuse. Individuals who had comor-bidity with other diagnoses (e.g., depression), accordingto clinician reports, remained in the sample.

Participants were not excluded if they were on medi-cation for their anxiety symptoms. In the OCD group 17(60.7%) young people were taking medication and in the

Table 1 Comorbid diagnoses of the participants in the clinical groups

Primary diagnosis

Comorbid diagnoses

Separationanxiety

Socialanxiety

Panicagoraphobia GAD

N N % N % N % N %

OCD 28 13 50.0% 14 50.0% 3 10.7% 20 71.4%Separation anxiety 4 – – 2 50.0% 2 50.0% 2 50.0%Social anxiety 10 2 20.0% – – 6 60.0% 7 70.0%Panic disorder & agoraphobia 8 2 25.0% 6 75.0% – – 6 75.0%Generalised anxiety disorder (GAD) 6 1 16.7% 3 50.0% 3 50.0% – –

1078 Sarah Libby et al.

anxious group 12 (42.8%) young people were takingmedication.

The mean age in years of the participants in the threegroups (OCD, M ¼ 14;08, SD ¼ 1;10; Anxious, M ¼14;09, SD ¼ 1;10; Non-clinical, M ¼ 14;08, SD ¼ 1;08)did not differ significantly (F(2, 115) ¼ .04, p ¼ .96).The gender ratio for the three groups (male:female, OCD12:16; Anxious 6:22; Non-clinical 31:31) was signific-antly different (v2(2) ¼ 6.51, Fisher’s exact test p ¼ .04).The anxious group had significantly more girls than theother two groups.

Measures

Leyton Obsessional Inventory – Child Version(Berg et al., 1988). The revised Leyton ObsessionalInventory – Child Version (LOI-CV) was used to measurethe severity of OCD symptoms. It is a 20-item self-reportmeasure with yes/no ratings; if a yes rating is given it isthen ranked for level of interference in daily life on a four-point scale. LOI-CV is not a diagnostic tool but a yesscore greater than 15 and an interference score greaterthan 25 have been used as indicative of obsessive-compulsive psychopathology in previous studies (e.g.,Flament et al., 1988). The LOI-CV has acceptable sen-sitivity but poor specificity (Flament et al., 1988). Thetotal scale has a Cronbach’s a of .82 (Berg et al., 1988).

Responsibility Attitude Scale (Salkovskis et al.,2000). The Responsibility Attitude Scale (RAS) reflectsgeneral beliefs that would predispose for inflatedresponsibility. It is the only measure of inflatedresponsibility that has been used with clinical groups ofadults. The RAS is a 26-item questionnaire whereindividuals rate a series of statements such as ‘I oftenfeel responsible for things which go wrong’ on a 7-pointscale.

The RAS has good test–retest reliability (k ¼ .94) andhigh internal consistency (a ¼ .92) (Salkovskis et al.,2000). Some wording was changed from the originalversion to make it suitable for a younger population.There were four changes: question 2 ‘see danger com-ing’ replaced ‘foresee danger’, question 14 ‘punished’replaced ‘condemned’, question 17 ‘unforgivable’ re-placed ‘inexcusable’ and question 18 ‘actions’ replaced‘intentions’.

Thought–Action Fusion Scale (Shafran et al.,1996). The TAF has three subscales, Moral TAF, TAF-(likelihood other) and TAF-(likelihood self). Only theTAF-(likelihood other) subsection (TAF-LO) has beenfound to be associated with OCD in adults (Shafranet al., 1996) and was the only subsection used in thisstudy. This comprises four questions that evaluate theextent negative thoughts are believed to have conse-quences for other people (e.g., If I think of a friend/relative falling ill, this increases the risk that he/shewill fall ill). Items are rated on a five-point scale. Internalconsistency for TAF-LO is good, ranging from a’s of .93–.96 depending on the population sampled (Shafranet al., 1996).

Multidimensional Perfectionism Scale (Frost et al.,1990). The Multidimensional Perfectionism Scale

(MPS) has previously been used with adults with OCD(Antony et al., 1998; Frost & Steketee, 1997) and anon-clinical sample of 10- to 12-year-olds (Stumpf &Parker, 2000). The MPS is a 35-item questionnairewhich asks participants to rate statements on a five-point scale. It has six subscales that reflect concernover mistakes, personal standards, parental expecta-tions, parental criticism, doubts about actions andorganisation. All subscales have an internal con-sistency alpha value greater than .77 (Frost et al.,1990). The subscale Doubts about Action was notevaluated in this study because it is confounded withOCD symptoms (Shafran & Mansell, 2001). Wordingwas changed from the original so questions referring toparents were in the present tense (e.g., my parentswant me to be the best at everything) rather than thepast tense (e.g., my parents wanted me to be the bestat everything).

Procedure

Clinical group. Families in this group were given in-formation packs about the project by their clinicianand returned consent forms to the researchers if theywere interested in participating. The young person andparents were then interviewed by one of the authorseither in the clinic or in the family’s home. The youngperson was always interviewed first with the ADIS-Cfollowed by the parent using the ADIS-P. While theparent was being interviewed the young person com-pleted the four questionnaire measures. The ques-tionnaires were presented in a booklet and alwayscompleted in the same order, MPS, TAF-LO, RAS andLOI-CV.

Non-clinical group. Two hundred and fifty names wereselected randomly from the school register and sentinformation packs about the project. Consent formswere returned to the researchers and the question-naires were then posted to the young people to becompleted. Eighty-eight families (35.2%) consented toparticipate and were sent the questionnaires to com-plete; 63 returned the questionnaires; this is 71.6% ofthose sent the questionnaires and 25.2% of all thefamilies invited.

As most of these questionnaires had not been usedwith young people before, the participants were askedto place a star by any questions they failed to under-stand. Using a conservative measure including starsand items missed out, this occurred in 17 young people(14.4%). Forty-six questions in total were not answered,which is less than 1% of the total number of questionsposed.

Results

Internal consistency of questionnaires

As many of the questionnaires had not been usedwith this age group previously, internal consistencyfor each of the questionnaires and their subscaleswas measured using Cronbach a. These data arepresented in Table 2. All had acceptable internalconsistency ranging from .79 to .96.

Cognitive appraisals in OCD 1079

Group comparisons

Table 2 also presents the mean item scores for themeasures used in the study. Item means are used soparticipants who missed a question could be in-cluded in all the analyses. Using a one-way ANOVAand post-hoc Scheffe tests, the OCD group had sig-nificantly higher item means on the LOI-CV yes/noratings (F(2, 115) ¼ 27.93, p < .01) than the anxiousgroup (p < .01) and the non-clinical group (p < .01).The OCD group also had higher mean item scores onthe LOI-CV interference scale (F(2, 115) ¼ 48.45,p ¼ .01) than the anxious group (p < .01) and thenon-clinical group (p < .01).

The results of a one-way ANOVA and post-hocScheffe tests revealed a significant difference be-tween groups on the RAS (F(2, 115) ¼ 12.68,p < .01), with the young people in the OCD groupscoring significantly higher than the anxious group(p < .01) and the non-clinical group (p < .01). Scoreson the TAF-LO were not normally distributed andcould not be transformed; a Kruskal–Wallis H testwas significant (H(2,115) ¼ 24.75, p < .01). Mann–Whitney U tests identified significant differencesbetween the OCD and anxious group (z(56) ¼)2.57, p ¼ .01) and the OCD and non-clinical group(z(90) ¼ )1.82, p < .01). The anxious group alsohad significantly higher scores than the non-clinicalgroup (z(90) ¼ )2.32, p ¼ .02). On the MPS the onlysubscales that had significant differences were theconcern over mistakes subscale (F(2, 115) 4.71,p ¼ .01) and the parental expectations subscale(F(2, 115) ¼ 3.75, p ¼ .03). On the concern overmistakes subscale the OCD group scored signific-antly higher than the non-clinical group (p ¼ .02)but not the anxious group. On the parental expec-tations subscale the non-clinical group scored sig-

nificantly higher than the OCD group (p ¼ .05) butdid not differ significantly from the anxious group.The groups did not differ on the Parental Criticismsubscale as predicted from studies with adults.

Relationship between variables

Correlations between the measures are presented inTable 3. There are modest correlations between thevariables that were found to distinguish the OCDgroup. These variables also correlate with LOI-CVyes/no. Personal Standards (MPS-PS) and Organ-isation (MPS-O) also had low correlations with thismeasure of OCD.

Regression analysis

A standard multiple regression with simultaneousentry was conducted with RAS, TAF-LO and MPS-CM as predictors for obsessive-compulsive symp-toms (LOI-CV yes/no). The data from all participantswas used (N ¼ 118). As TAF-LO was not normallydistributed and could not be transformed, thevariable was dichotomised as recommended byTabachnick and Fidell (1996). The sample was di-vided into those who rated all questions on thismeasure as ‘strongly disagree’ (0) and those whorated one or more questions as ‘slightly disagree’ (1),‘neutral’ (2), ‘slightly agree’ (3) or ‘strongly agree’ (4).As there were many individuals who rated all ques-tions as 0 on this measure (43.2%), this seemed themost meaningful dichotomisation. The model had anR2 of .39 (F(3, 114) ¼ 23.85, p < .01); only RAS madea significant contribution to symptom severity,although Concern over Mistakes approaches sig-nificance (see Table 4).

Table 2 Internal consistency and comparison of the groups on item means for each of the questionnaire measures

Cronbach a

OCD(N ¼ 28) Anxious(N ¼ 28)

Non-clinical(N ¼ 62)

M SD M SD M SD

Leyton yes/no .88 .75a .22 .47b .24 .39b .39 F(2,115) ¼ 27.93**Leyton interference .93 1.33a .79 .54b .35 .27b .31 F(2,115) ¼ 48.45**Responsibility Attitude Scale .95 4.75a 1.25 3.88b .98 3.57b .94 F(2,115) ¼ 12.68**Thought Action Fusion – Likelihood Other1 .96 1.96a 1.29 1.01b 1.04 .56c 1.04 H(2, 115) ¼ 24.75**Multidimensional Perfectionism Scale(Concern over Mistakes)

.91 2.59a .96 2.41 .92 2.04b .66 F(2,115) ¼ 4.71**

Multidimensional Perfectionism Scale(Personal Standards)

.83 3.07 .79 2.75 .79 2.99 .73 F(2, 115) ¼ 1.16

Multidimensional Perfectionism Scale(Parental Expectations)

.83 2.13a .75 2.28 1.01 2.62b .84 F(2, 115) ¼ 3.75*

Multidimensional Perfectionism Scale(Parental Criticism)

.79 1.91 .80 2.07 .74 2.00 .81 F(2, 115) ¼ .23

Multidimensional Perfectionism Scale(Organisation)

.90 3.48 .88 3.02 .83 3.42 .83 F(2, 115) ¼ 2.24

Different superscripts indicate significant differences between the groups.*Significant at p < .05.**Significant at p < .01.All analyses were based on one-way ANOVAs and post-hoc Scheffe tests except1 where Kruskal–Wallis and Mann–Whitney U testswere used because the data was not normally distributed.

1080 Sarah Libby et al.

Discussion

The results of this study demonstrate that youngpeople with OCD have higher levels of inflatedresponsibility than young people with other anxietydisorders and a non-clinical group. In addition,inflated responsibility was the only significant pre-dictor of OCD. These results map on to the findingswith adults and the mean scores obtained in thisstudy for the three groups are very close to thosereported for adults (Salkovskis et al., 2000). Thefindings support Salkovskis et al.’s (1999) proposi-tion that inflated responsibility is associated withOCD in young people. It remains to be established ifcognitions associated with inflated responsibilityprecede the development of OCD, as predicted bySalkovksis et al. (1999), or emerge following theonset and only have a maintaining role.

Some recent studies have found thought–actionfusion to be associated with other anxiety symptomsas well as OCD (Muris et al., 2001; Rassin et al.,2001). This study, however, found that TAF-LO wassignificantly higher in the OCD group compared withthe anxious group and the non-clinical group, sug-gesting a specific association. The anxious group,

however, had scores that were significantly higherthan the non-clinical group and this may indicatethat TAF-LO has a minor role in anxiety disorders inyoung people. It is of interest to note that the meantotal score obtained by the OCD group in this study(M ¼ 7.82) was higher than that observed in studieswith adults who have OCD (e.g., Rassin et al., 2001,M ¼ 4.43; Shafran et al., 1996, M ¼ 4.77). Scores forthe anxious group were broadly similar to the scoresobtained in studies with adults (Rassin et al., 2001;Shafran et al., 1996). The non-clinical group’s totalscore (M ¼ 2.25) was similar to scores reported byShafran et al. (1996) for students (M ¼ 2.59) buthigher than scores found for a general non-clinicalpopulation (M ¼ 1.03). The findings suggest thatTAF-LO may be more important in the presentationof OCD in young people and may be more prevalentin young people generally. This difference may reflectdevelopmental influences. For example, it may beassociated with the emergence of formal operationalthinking at the age of 11 years which is character-ised by abstract reasoning, hypothetical deductionand metacognition, including the ability to reflect onone’s own thinking (Inhelder & Piaget, 1958). Elkind(1967) has argued that the emergence of formal op-erations results in young people believing that theyare the focus of others’ attention and that they areunique and omnipotent. This cognitive backdropmay make young people more likely to reportthought–action fusion. Thought–action fusion hasalso been associated with a greater sense of uncer-tainty (Keinan, 1994). Uncertainty is inevitable foradolescents and young adults and may help explainsome of these data.

On the MPS, concern over mistakes was the onlysubsection that was significantly higher for the OCDgroup compared to the non-clinical group. The

Table 3 Correlations between measures for all three groups combined

RAS TAF-LOa MPS-CM MPS-PS MPS-PC MPS-PE MPS-O

LOI-CV .58** .51** .49** .31** .14 .04 .31**yes/noRAS – .57** .65** .46** .25* .15 .15TAF-LOa – – .47** .28* .22* .08 ).06MPS-CM – – – .60** .39** .35* .03MPS-PS – – – –.01 .20* .36**MPS-PC – – – – – .75** ).24*MPS-PE – – – – – – ).14a

*Significant at 0.05 level.**Significant at 0.01 level.aCorrelations with TAF-LO were calculated using Spearman’s rho Correlation Coefficient. All other correlations were calculatedusing Pearson’s Product Moment Correlation Coefficient.Abbreviations:LOI-CV yes/no (Leyton Obsessional Inventory – Child Version yes/no rating).RAS (Responsibility Attitude Scale).TAF-LO (Thought–Action Fusion – Likelihood Other).MPS (Multidimensional Perfectionism Scale).CM (Concern Over Mistakes).PS (Personal Standards).PC (Parental Criticism).PE (Parental Expectations).

Table 4 Multiple regression on RAS, TAF-LO and MPS-CMpredicting LOI-CV Yes / no scores

Independent variables b t p Partial r

Responsibility Attitude Scale .49 4.76 <.01 .41Thought–Action Fusion –Likelihood Other

.13 1.36 .18 .13

Multidimensional PerfectionismScale (Concern Over Mistakes)

.06 .76 .45 .07

(R2 of .39 (F(3, 114) ¼ 23.85, p < .01).

Cognitive appraisals in OCD 1081

scores obtained by the anxious group fell betweenthe two other groups, but did not differ significantlyfrom either. Studies with adults have found concernover mistakes to be high in individuals with OCD andin those with other anxiety problems (Antony et al.,1998; Frost & Steketee, 1997). The difference in thisstudy appears to relate to the anxious group attain-ing lower scores than studies with adults with panicdisorder (Antony et al., 1998; Frost & Steketee,1997) or social phobia (Antony et al. 1998; Saboon-chi et al., 1999). The OCD group and the non-clinicalgroup had scores that were comparable to thoseobserved in adult studies. This is an interestingfinding but is not easily explained. There may bespecific developmental shifts associated with thepresentation of anxiety or the findings may be rela-ted to some aspect of the study’s design; furtherinvestigation is required. Concern over mistakescould link to inflated responsibility and thought–action fusion because all are associated with a fear ofbeing judged by others or having intolerable anxietyif they make an error.

Parental criticism was not significantly higher inthe clinical groups as predicted from studies withadults and parental expectation was significantlylower in the OCD group compared with the non-clinical group. The cognitive model emphasises theimportance of belief systems about parents beingimportant in the development of disorders (Beck,1976) but there is no evidence for this in this study.The difference in the findings here compared to theadult literature could be accounted by a number offactors. Adults with anxiety disorders may only per-ceive their parents as critical retrospectively butthese beliefs may not have existed or been accessibleat the conscious level when they were children.Alternatively, there may have been a selection biasagainst young people who had these beliefs abouttheir parents, with these families choosing not totake part in the study.

The findings broadly replicate studies with adultswho have OCD, suggesting that there is continuity inthese cognitive appraisals from adolescence toadulthood. There are, however, some interestingdifferences in the extent to which some of these be-liefs are held. Further studies are required that di-rectly compare individuals at different points acrossthe lifespan or collect longitudinal data. These datacould enrich cognitive models by providing a devel-opmental perspective on risk factors for specificcognitions.

The study demonstrates that these questionnairesare valid when used with young people and may behelpful in assessment and evaluating change intherapy. The findings support the incorporation ofcognitive techniques in the treatment of youngpeople with OCD. These may include the use ofresponsibility pie charts (van Oppen & Arntz, 1994),thought–action defusion (Wells, 1997) and normal-isation of intrusive thoughts (Freeston et al., 1996)

as a precursor or adjunct to E/RP. Using genericcognitive techniques to address concern over mis-takes is also likely to be important with some indi-viduals. Within E/RP it is also important that thetherapist remains aware of whether or not the childis passing their sense of responsibility to the ther-apist or a parent in order to complete tasks. If this isoccurring it will prevent behavioural tasks resultingin any cognitive change and is likely to compromiseimprovements in symptoms. Consideration of theother belief domains highlighted by the ObsessiveCognitions Working Group (1997) would be ofinterest in future studies. Further case studies withyoung people that expand on the work completed byWilliams et al. (2002) and Shafran and Somers(1998) is an important next step. Randomised con-trol trials are also critical to evaluate if the cognitivecomponent adds to the efficacy of treatment as hasbeen found in adults (McLean et al., 2001; VanOppen et al., 1995).

The current study is limited by a number of fac-tors. Like many questionnaire studies, only aproportion of families asked to participate agreed, sothe participants may not be representative of thesampled population. It is impossible to know if thechildren who decided to participate had differentprofiles from those who did not. The OCD group wastreated as homogenous and there was no attempt toevaluate tic disorders that may be related to a dis-tinctive subgroup in children (Eichstedt & Arnold,2001). The anxious group was also treated ashomogenous, although adult studies have founddifferences in cognitive appraisals associated withspecific disorders. The study did not control foranxiety and depression and there is some evidencefrom studies with adults that some of the measuresevaluated here (e.g., TAF-LO, MPS) would be asso-ciated with these symptoms and, therefore, futurestudies may consider evaluating them.

In conclusion, this study has demonstrated thatthe use of questionnaires designed to assess cognit-ive appraisals in adults with OCD can be used withyoung people. The findings indicate that youngpeople with OCD and adults with the disorder holdsimilar cognitive appraisals. There are, however,possible developmental shifts in the extent to whichthese beliefs are held and this warrants furtherinvestigation. Consideration of cognitive appraisalsin therapy and future research with young peoplewith OCD could lead to fruitful advances in boththeory and practice.

Acknowledgements

The authors would like to thank all the families whoparticipated in this study. We are also grateful to allthe people who helped in the recruitment of familiesand the many people who offered comments on thispiece of research.

1082 Sarah Libby et al.

Correspondence to

Sarah Libby, Child & Adolescent Mental HealthService, United Bristol Healthcare Trust, BristolRoyal Children’s Hospital, Paul O’Gorman Building,Upper Maudlin Street, Bristol, BS2 8BJ, UK;Tel: 01179 294530; Email: sarah.libby@ubht.swest.nhs.uk

ReferencesAllsopp, M., & Verduyn, C. (1990). Adolescents withobsessive-compulsive disorder: A case note review ofconsecutive patients referred to a provincial regionaladolescent psychiatry unit. Journal of Adolescence,13, 157–169.

American Psychiatric Association. (1994). Diagnosticand statistical manual of mental disorders (4th edn).Washington DC: American Psychiatric Association.

Antony, M.M., Purdon, C.L., Huta, V., & Swinson, R.P.(1998). Dimensions of perfectionism across theanxiety disorders. Behaviour Research and Therapy,36, 1143–1154.

Beck, A.T. (1976). Cognitive therapy and the emotionaldisorders. New York: International Universities Press.

Berg, C.Z., Whitaker, A., Davies, M., Flament, M.F., &Rapoport, J.L. (1988). The survey form of the LeytonObsessional Inventory-Child version: Norms from anepidemiological study. Journal of the American Acad-emy of Child and Adolescent Psychiatry, 27, 759–763.

Bolton, D., Luckie, M., & Steinberg, D. (1995). Long-term course of obsessive-compulsive disordertreated in adolescence. Journal of the AmericanAcademy of Child and Adolescent Psychiatry, 34,1441–1450.

Chorpita, B.F., Albano, A.M., & Barlow, D.H. (1996).Cognitive processing in children: Relation to anxietyand family influences. Journal of Clinical ChildPsychology, 25, 170–176.

De Haan, E., Hoogduin, K.A.L., Buitelaar, J.K., &Keijsers, G.P.J. (1998). Behaviour therapy versusclomipramine for the treatment of obsessive-compul-sive disorder in children and adolescents. Journal ofthe American Academy of Child and AdolescentPsychiatry, 37, 1022–1029.

Eichstedt, J.A., & Arnold, S.L. (2001). Childhood-onsetobsessive-compulsive disorder. A tic-related subtypeof OCD? Clinical Psychology Review, 21, 137–158.

Elkind, D. (1967). Egocentrism in adolescence. ChildDevelopment, 38, 1025–1034.

Flament, M.F., Koby, E., Rapoport, J.L., Berg, C.J.,Zahn, T., Cox, C., Denckla, M., & Lenane, M. (1990).Childhood obsessive-compulsive disorder: A prospec-tive study. Journal of Child Psychology and Psychi-atry, 31, 363–380.

Flament, M.F., Whitaker, A., Rapoport, J.L., Davies, M.,Berg, C.Z., Kalikow, K., Sceery, W., & Shaffer, D.(1988). Obsessive compulsive disorder in adoles-cence: An epidemiological study. Journal of theAmerican Academy of Child and Adolescent Psychi-atry, 27, 764–771.

Freeston, M.H., Rheaume, J., & Ladouceur, R. (1996).Correcting faulty appraisals of obsessional thought.Behaviour Research and Therapy, 34, 433–446.

Frost, R.O., Marten, P., Lahart, C., & Rosenblate, E.(1990). The dimensions of perfectionism. CognitiveTherapy and Research, 14, 449–468.

Frost, R.O., & Steketee, G. (1997). Perfectionism inobsessive-compulsive disorder patients. BehaviourResearch and Therapy, 35, 291–296.

Frost, R.O., Steketee, G., Cohn, L., & Griess, K. (1994).Personality traits in subclinical and non-obsessive-compulsive volunteers and their parents. BehaviourResearch and Therapy, 32, 47–56.

Hadwin, J., Frost, S., French, C.C., & Richards, A.(1997). Cognitive processing and trait anxiety intypically developing children: Evidence for an inter-pretation bias. Journal of Abnormal Psychology, 106,486–490.

Hodgson, R.J., & Rachman, S. (1977). Obsessional-compulsive complaints. Behaviour Research andTherapy, 15, 389–395.

Inhelder, B., & Piaget, J. (1958). The growth of logicalthinking from childhood to adolescents. New York:Basic Books.

Kail, R., & Bisanz, J. (1991). The information process-ing perspective on cognitive development in childhoodand adolescence. In J. McShane (Ed.), Cognitivedevelopment: An information processing approach.London: Basil Blackwell.

Karno, M., Golding, J.M., Sorenson, S.B., & Burnham,M.A. (1988). The epidemiology of obsessive-compul-sive disorder in 5 US communities. Archives ofGeneral Psychiatry, 45, 1094–1099.

Keinan, G. (1994). Effects of stress and tolerance ofambiguity on magical thinking. Journal of Personalityand Social Psychology, 67, 48–55.

Kuhn, D., Amsel, E., & O’Loughlin, M. (1988). Thedevelopment of scientific thinking skills. AcademicPress: San Diego.

Ladouceur, R., Rheaume, J., & Aublet, F. (1997).Excessive responsibility in obsessional concerns:A fine grained experimental analysis. BehaviourResearch and Therapy, 35, 423–427.

Ladouceur, R., Rheaume, J., Freeston, M.H., Aublet, F.,Jean, K., LaChance, S., Langlois, F., & De Pokoman-dy-Morin, K. (1995). Experimental manipulation ofresponsibility: An analogue test for models of ob-sessive compulsive disorder. Behaviour Research andTherapy, 35, 937–946.

March, J.S. (1995). Cognitive-behavioral psychother-apy for children and adolescents with OCD: A reviewand recommendations for treatment. Journal of theAmerican Academy of Child and Adolescent Psychi-atry, 34, 7–18.

McLean, P.D., Whittal, M.L., Thordarson, D.S., Taylor,S., Sochting, I., Koch, W.J., Paterson, R., & Anderson,K.W. (2001). Cognitive versus behavior therapy in thegroup treatment of Obsessive-Compulsive Disorder.Journal of Consulting and Clinical Psychology, 69,205–214.

Muris, P., Meesters, C., Rassin, E., Merckelbach, H., &Campbell, J. (2001). Thought–action fusion andanxiety disorder symptoms in normal adolescents.Behaviour Research and Therapy, 39, 843–852.

Obsessive Compulsive Cognitions Working Group.(1997). Cognitive assessment of obsessive-compuls-ive disorder. Behaviour Research and Therapy, 35,667–681.

Cognitive appraisals in OCD 1083

Pacht, A.R. (1984). Reflections on perfection. AmericanPsychologist, 39, 386–390.

Pauls, D.L., Alsobrook, J.P., Goodman, W., Rasmussen,S., & Leckman, J. (1995). A family study of obsessive-compulsive disorder. American Journal of Psychiatry,152, 76–84.

Rachman, S. (1997). A cognitive theory of obsessions.Behavior Research and Therapy, 35, 793–802.

Rasmussen, S.A., & Eisen, J.L. (1990). Epidemiology ofobsessive compulsive disorder. Journal of ClinicalPsychiatry, 55(Suppl.), 10–14.

Rassin, E., Merckelbach, H., Muris, P., & Schmidt, H.(2001). The thought–action fusion scale: Furtherevidence for its reliability and validity. BehaviourResearch and Therapy, 39, 537–544.

Rheaume, J., Ladouceur, R., Freeston, M.H., & Letarte,H. (1994). Inflated responsibility in obsessive-com-pulsive disorder: Psychometric studies of a semi-idiographic measure. Journal of Psychopathology andBehavioural Assessment, 16, 265–276.

Riddle, M., Scahill, L., King, R., Hardin, M.T., Towbin,K.E., Ort, S.I., Leckman, J.F., & Cohen, D.J. (1990).Obsessive compulsive disorder in children and ado-lescents: Phenomenology and family history. Journalof the American Academy of Child and AdolescentPsychiatry, 29, 766–772.

Saboonchi, F., Lundh, L.-G., & Ost, L.-G. (1999).Perfectionism and self-consciousness in social phobiaand panic disorder with agoraphobia. BehaviourResearch and Therapy, 37, 799–808.

Salkovskis, P.M. (1989). Cognitive-behavioural factorsand the persistence of intrusive thoughts in obses-sional problems. Behaviour Research and Therapy,27, 677–682.

Salkovskis, P., Forrester, E., & Richards, C. (1998).Cognitive-behavioural approach to understandingobsessional thinking. British Journal of Psychiatry,173, 53–63.

Salkovskis, P.M., & Harrison, J. (1984). Abnormaland normal obsessions – a replication. BehaviourResearch and Therapy, 22, 549–552.

Salkovskis, P., Shafran, R., Rachman, S., & Freeston,M.H. (1999). Multiple pathways to inflated respons-ibility beliefs in obsessional problems: Possibleorigins and implications for therapy and research.Behaviour Research and Therapy, 37, 1055–1072.

Salkovskis, P.M., Wroe, A.L., Gledhill, A., Morrison, N.,Forrester, E., Richards, C., Reynolds, M., & Thorpe, S.(2000). Responsibility, attitudes and interpretationsare characteristic of obsessive compulsive disorder.Behaviour Research and Therapy, 38, 347–372.

Shafran, R., & Mansell, W. (2001). Perfectionism andpsychopathology: A review of research and treatment.Clinical Psychology Review, 21, 879–906.

Shafran, R., & Somers, J. (1998). Treating adolescentobsessive-compulsive disorder: Applications of thecognitive theory. Behaviour Research and Therapy,36, 93–97.

Shafran, R., Thordarson, D.S., & Rachman, S. (1996).Thought–action fusion in obsessive compulsive dis-order. Journal of Anxiety Disorders, 10, 379–391.

Silverman, W.K., & Albano, A.M. (1996). Anxiety Disor-ders Interview Schedule for DSM-IV: Child Version.San Antonio, TX: Psychological Corporation.

Steketee, G.S., Frost, R.A., Rheaume, J., & Wilhelm, S.(1998). Cognitive theory and treatment of obsessive-compulsive disorder. In M. Jenike & L. Baer (Eds.),Obsessive compulsive disorders: Practical manage-ment (3rd edn, 368–399). St.Louis: Mosby.

Stumpf, H., & Parker, W.D. (2000). A hierarchicalstructural analysis of perfectionism and its relationto other personality characteristics. Personality andIndividual Differences, 28, 837–852.

Swedo, S.E., Rapoport, J.L., Leonard, H., & Lenane, M.(1989). Obsessive-compulsive disorder in childrenand adolescents. Archives of General Psychiatry, 46,335–341.

Tabachnick, B.G., & Fidell, L.S. (1996). Using multi-variate statistics (3rd edn). New York: HarperCollins.

Valleni-Basile, L.A., Garrison, C.Z., Jackson, K.L.,Waller, J.L., McKeown, R.E., Addy, C.L., & Cuffe,S.P. (1994). Frequency of obsessive-compulsive dis-order in a community sample of young adolescents.Journal of the American Academy of Child andAdolescent Psychiatry, 33, 782–791.

Van Oppen, P., & Arntz, A. (1994). Cognitive therapy forObsessive Compulsive Disorder. Behaviour Researchand Therapy, 32, 79–87.

Van Oppen, P., De Haan, E., Van Balkom, A.J.L.M.,Spinhoven, P., Hoodgduin, K., & Van Dyck, R.,(1995). Cognitive therapy and exposure in vivoin the treatment of obsessive-compulsive disorder.Behaviour Research and Therapy, 33, 379–390.

Wells, A. (1997). Cognitive therapy of anxiety disorders:A practical manual and conceptual guide. New York:John Wiley & Son.

Williams, T.I., Salkovskis, P.M., Forrester, E.A., &Allsopp, M.A. (2002). Changes in symptoms of OCDand appraisal of responsibility during cognitivebehavioural treatment: A pilot study. Behaviouraland Cognitive Psychotherapy, 30, 69–78.

Wilson, K.A., & Chambless, D.L. (1999). Inflatedperceptions of responsibility and obsessive-compuls-ive symptoms. Behaviour Research and Therapy, 37,325–335.

Manuscript accepted 1 October 2003

1084 Sarah Libby et al.