J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

Contents lists available at ScienceDirect

Journal of Behavior Therapy and Experimental Psychiatry journal homepage: www.elsevier.com/locate/jbtep

Augmentation of Treatment As Usual with online Cognitive Bias Modification of Interpretation training in adolescents with Obsessive Compulsive Disorder: A pilot study Elske Salemink a, b, *, Lidewij Wolters b, c, Else de Haan b, d a Addiction, Development and Psychopathology Lab (Adapt Lab), Department of Developmental Psychology, Research Priority Areas ‘Yield’, and ‘Amsterdam Brain and Cognition’, University of Amsterdam, The Netherlands b Department of Obsessive Compulsive-, Anxiety- and Tic Disorders, Academic Center for Child and Adolescent Psychiatry, de Bascule, Amsterdam, The Netherlands c Academic Medical Center, Department of Child and Adolescent Psychiatry, Amsterdam, The Netherlands d Child Development and Education, University of Amsterdam, The Netherlands

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 July 2014 Received in revised form 6 February 2015 Accepted 8 February 2015 Available online 16 February 2015

Background and objectives: Cognitive Behavioral Therapy for children and adolescents with Obsessive Compulsive Disorder (OCD) is effective. However, since almost half of patients remain symptomatic after treatment, there remains room for improvement. Cognitive Bias Modification training of Interpretations (CBM-I) is a promising new intervention, as it targets misinterpretation of intrusions, which is seen as an important characteristic in OCD. To date, there have been no published studies of CBM-I in adolescents with OCD. The aim of the current pilot study was to examine the added value of online CBM-I training as an adjunctive treatment to the Treatment As Usual (TAU; that included CBT and pharmacotherapy) in adolescents with OCD. Methods: Patients receiving TAU were randomly assigned to either an additional CBM-I training (n ¼ 9), or to an additional placebo variant of this procedure (n ¼ 7). Results: Immediate, on-line interpretations changed in response to the CBM-I training, while no such effects were observed on slower retrospective off-line interpretations. Patients in the CBM-I training condition reported fewer obsessive compulsive symptoms after training, and clinicians rated them as having fewer obsessive symptoms (corresponding to medium-large effect sizes). No such changes were observed in the placebo group. Limitations: The small sample size precludes strong conclusions and replication is necessary to test the robustness of the findings. Conclusions: This small randomized controlled trial is suggestive, although not conclusive, regarding the promising additive value of OC-related CBM-I training as an adjunctive intervention to TAU in an adolescent clinical population. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Interpretive bias Cognitive Bias Modification Obsessive Compulsive Disorder Adolescents

1. Introduction Cognitive theories of Obsessive Compulsive Disorder (OCD) argue that the misinterpretation of normal intrusions is the core problem in the development and maintenance of OCD (Salkovskis, 1985). Patients with OCD are assumed to interpret such intrusions

* Corresponding author. Department of Developmental Psychology, University of Amsterdam, Weesperplein 4, 1018 XA, Amsterdam, The Netherlands. Tel.: þ31 205258663. E-mail address: [email protected] (E. Salemink). http://dx.doi.org/10.1016/j.jbtep.2015.02.003 0005-7916/© 2015 Elsevier Ltd. All rights reserved.

as potentially dangerous, bad, or as predicting harm, resulting in anxiety and distress. Consequently, compulsions (recurrent behaviors) are performed in an attempt to reduce the anxiety and distress. There is a wealth of empirical evidence suggesting that such misinterpretations of intrusions are related to OC symptoms in adults (Frost & Steketee, 2002) as well as in children and adolescents (Matthews, Reynolds, & Derisley, 2007; Reeves, Reynolds, Coker, & Wilson, 2010). According to this cognitive theory, more benign interpretations of normal intrusions may prevent anxiety and distress as well as compulsions. The first-line treatment for youth with OCD is cognitive behavioral therapy (CBT) (Geller & March, 2012), with exposure

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

with response prevention and the modification of misinterpretations as the most important strategies. Meta-analyses have shown that these treatment strategies are effective in nchez-Meca, reducing OC symptoms in youth (large effect sizes; Sa zar, Iniesta-Sepúlveda, & Rosa-Alca zar, 2014). However, Rosa-Alca as response rates vary between 40 and 65%, almost half of the patients continue to have significant levels of symptoms after standard treatment. There seems to be substantial room for improvement of OCD treatment in youth. A recent development is Cognitive Bias Modification of Interpretation (CBM-I) training; a computerized training paradigm developed to modify dysfunctional interpretations in anxiety. Participants receive short ambiguous scenarios with a missing word fragment in the last sentence. Resolution of the fragment results in a positive interpretation of the scenario. A recent meta-analysis focusing on adult samples revealed that CBM-I training was successful in increasing positive interpretations and decreasing negative mood states (Menne-Lothmann et al., 2014). In contrast to CBT, CBM-I changes interpretation bias directly, through practicing interpreting information more positively, rather than through verbal instructions and explicitly challenging dysfunctional thoughts (Baert, De Raedt, & Koster, 2011). Based on these differences, adding CBM-I training to CBT might enhance treatment effects and response rates. CBM-I studies in adolescents are accumulating. Promising effects of CBM-I in adolescents have been observed in unselected adolescent volunteers and sub-clinical samples (e.g., Lau, Belli, & Chopra, 2013; Lothmann, Holmes, Chan, & Lau, 2011; Salemink & Wiers, 2011; Vassilopoulos, Banerjee, & Prantzalou, 2009). In the only CBM-I study in a clinical sample of adolescents with anxiety disorders, one session of CBM-I training affected interpretations, but not anxiety (Fu, Du, Au, & Lau, 2013). There has been limited application of CBM-I to OCD; to the best of our knowledge, there are three studies in adults. Clerkin and Teachman (2011) presented students with elevated levels of OC symptoms with one session of CBM-I training designed to change interpretations of intrusive thoughts. Compared to a neutral training group, participants in the CBM-I group endorsed healthier OC-relevant interpretations and beliefs after the training. No OC symptoms were measured, but there was a non-significant trend for participants in the CBM-I condition to report less negative affect in response to a stressor. Recently, the effects of (single-session) CBM-I training were examined in a sample of students with high levels of OC symptoms and in a selected sample of adult community members with varying levels of OC symptoms (Beadel, Smyth, & Teachman, 2014; Williams & Grisham, 2013 respectively). Again, CBM-I training resulted in less OC-relevant interpretations and obsessive beliefs. However, more mixed (or even null) effects were observed on the more clinically relevant measures in response to a stressor. These latter findings might be related to the use of subclinical samples. They may also be related to the provision of just a single session of training as a recent meta-analysis indicated that providing repeated training sessions increased training effectiveness (Menne-Lothmann et al., 2014). In sum, negative interpretation of intrusive thoughts is assumed to play an important role in OCD and three studies in adults suggested that single-session CBM-I training can modify such interpretations. However, there is a lack of studies of CBM-I training in adolescents with OCD. The aim of the current pilot study was to explore the effectiveness of eight-sessions of online CBM-I training as an adjunctive treatment in a clinical sample of adolescents with OCD. In a small randomized placebo-controlled trial of adolescents who were already receiving Treatment As Usual (TAU) for OCD, we randomly allocated participants to receive i) additional CBM-I training to interpret intrusions as less threatening, or ii) placebo

113

training. Consistent with the cognitive model of OCD (Salkovskis, 1985), adolescents in the CBM-I condition were trained to interpret intrusive thoughts in a healthier, more benign way. We predicted that adolescents who completed the CBM-I training would 1) make fewer OC-relevant and more OCirrelevant interpretations of normal intrusions, and 2) have fewer OC symptoms compared to the placebo condition. We made a distinction between “on-line” interpretations that are made immediately at the time the ambiguous information is first encountered, and slower “off-line” interpretations in which judgments are made retrospectively (see also Hirsch & Clark, 2004). The time taken to solve OC-relevant and OC-irrelevant word fragments during the training was used as an index of immediate on-line interpretations and it was expected that, compared to the placebo control condition, adolescents who follow the CBM-I training would show faster responses to OC-irrelevant and slower responses to OC-relevant word fragments. To assess off-line interpretations, adolescents completed a recognition task before and after training (Salemink & van den Hout, 2010), and it was predicted that the CBM-I training would result in stronger endorsement of OC-irrelevant and less endorsement of OCrelevant interpretations. With respect to clinical outcome measures, we expected that adolescents who followed the CBM-I training would have fewer obsessive beliefs, as assessed with the Obsessive Beliefs Questionnaire-Child Version (OBQ-CV, Coles et al., 2010). A multi-informant approach was taken to assess CBM-I effects on OC symptoms; we expected a reduction in selfreported OC-symptoms assessed with the Revised Child Anxiety and Depression Scale e Child Version (RCADS, Chorpita, Moffitt, & Gray, 2005) and in clinician-rated OC-symptoms assessed with the Children's Yale-Brown Obsessive Compulsive Scale (CY-BOCS, Scahill et al., 1997) in the CBM-I group compared to the placebo group. Finally, we explored generalization to anxiety and depressive symptoms. 2. Method 2.1. Participants Participants were adolescents referred for either inpatient or outpatient treatment of OCD at the academic center for child and adolescent psychiatry, the Bascule, Amsterdam. Inclusion criteria for the study were a primary diagnosis of OCD according to DSMIV TR criteria (APA, 2000), a score of 8 or more on the CY-BOCS total score, and a score of 4 or more on the CY-BOCS obsession subscale (Scahill et al., 1997). Exclusion criteria were psychosis and insufficient Dutch literacy. Both adolescents and their parents received an information letter. The study was introduced as a test of a potentially new type of treatment: an online computerized training. The rationale of the study design, including the placebo control condition, was explained and participants were informed that they had equal chance of receiving ‘the real training’ or the placebo training. It was also explained that a lottery system determined allocation. All information and computerized instructions presented to the participants were identical in the two conditions. After written informed consent was obtained, 21 patients were randomly allocated (stratified on gender, age, and school level) to either the CBM-I training (n ¼ 12) or the placebo variant of this procedure (n ¼ 9). Four patients were excluded (one patient scored below the clinical cut-off and three patients did not complete the training and/or assessments); they did not differ significantly from the completers on any of the measures. Finally, one additional patient was excluded as they completed 60% of the word fragments incorrectly. The final sample consisted of 16 patients (10 females) with nine individuals in the CBM-I condition

114

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

Table 1 Socio-demographic and clinical characteristics. CBM-I Placebo control Statistic group (n ¼ 7) group (n ¼ 9) Age (in years) Gender (females/males) Education (lower/higher education) Duration of OC symptoms (in years) Setting (inpatient/outpatient clinic) Pharmacotherapy medication use (yes/no)

p- value

15.6 (2.4) 5/4 4/5

15.1 (2.0) 5/2 1/6

t ¼ 0.37 .72

6.2 (3.8) 5/4

4.5 (3.1) 3/4

t ¼ 0.93 .37

c2 ¼ 0.25 .61

3/6

5/2

c2 ¼ 2.29 .13

c2 ¼ 0.42 .52 c2 ¼ 1.67 .20

Standard deviations in parentheses. Lower education refers to VMBO (preparatory middle-level applied education) and MBO (middle-level applied education, vocational training) and higher education refers to HAVO/VWO (high school).

and seven in the placebo condition. Age ranged from 12 to 19 years1 (M ¼ 15.4, SD ¼ 2.2) (Table 1). The study was approved by the Medical Ethical Committee of the Academic Medical Center Amsterdam (NL35351.018.11). 2.2. Materials

in the study by Clerkin and Teachman were adapted to reflect adolescent situations. We also developed new scenarios as our study utilized eight sessions of training instead of one. Scenarios were based on the domains described by the Obsessive Compulsive Cognitions Working Group (OCCWG, 2011): inflated responsibility, overestimation of threat, intolerance of uncertainty, overimportance of thoughts, beliefs about the importance of controlling thoughts, and perfectionism. Each scenario consisted of three lines that were ambiguous in terms of valence. The final sentence contained a missing word. After disappearance of the scenario, the omitted word was presented as a word fragment and disambiguated the scenario in a healthy, OC-irrelevant way. Participants were instructed to complete the word fragment as quickly and accurately as possible by typing the first missing letter. If the response time exceeded 10.000 ms, the task continued automatically. A comprehension question then appeared and patients pressed Y for yes and N for no. Feedback was presented (correct vs. incorrect answer) to reinforce the OCirrelevant, healthy interpretation. An example of a training scenario is as follows: You bought some roses for your mum and accidently got pricked by a thorn. You are bleeding and ask your mum to put a band-aid on it. It is _______ that this would make your mother sick. unli-ely

2.2.1. Treatment As Usual The TAU was cognitive behavioral therapy (CBT), based on the Dutch treatment manual ‘Bedwing je dwang’ (‘Control your OCD’; De Haan & Wolters, 2009), with or without pharmacotherapy. It was provided in both an inpatient and outpatient setting. In both settings, treatment involved psycho-education, an inventory and hierarchy of the obsessions and compulsions, exposure with response prevention, cognitive interventions, and relapse prevention. Exposure with response prevention is introduced early in treatment, followed by and combined with cognitive interventions. The treatment ended with relapse prevention. Parents were involved in the therapy and level of involvement depended on the child's developmental level, preferences of the child and parents, and clinical considerations. The treatment was given by trained clinicians experienced in treating OCD in children and adolescents. The therapists were specifically trained in the treatment protocol and attended group supervision every two weeks and optional individual supervision.

unlikely If you are bleeding, is there a great risk of making somebody else sick? Instead of training scenarios, participants in the placebo control condition received neutral (non-ambiguous) scenarios in order to control for time behind computer and practice effects (Salemink, Kindt, Rienties, & van den Hout, 2014). An example of a scenario from the placebo control condition is as follows: You are making a pizza. First, you made dough for the crust and now you are busy with the sauce. You are using _________ for that. toma-oes tomatoes Are you making spaghetti?

2.2.2. CBM-I Eight sessions of scenario-based CBM-I training were used (Clerkin & Teachman, 2011). In each session, six blocks of nine unique scenarios were presented on a computer, with an optional break given after each block. In total, 432 unique scenarios were presented. Each block contained seven training scenarios and two probe scenarios (see Test of Interpretation Bias I for a description of the latter). The OC-related training scenarios used

1 This is a broad age range, which might be related to compliance and/or engagement. First, we tested whether there was a significant difference in age between included (n ¼ 16, M ¼ 15.4, SD ¼ 2.2) and excluded (n ¼ 5, M ¼ 14.8, SD ¼ 2.2) patients, this was not the case, t(19) ¼ .52, p ¼ .61. To explore whether there is an association between age and training engagement, correlations were calculated between age and percentage of errors in the training. Age was not significantly correlated with percentage of errors completing word fragments, r(16) ¼ .13, p ¼ .64, nor with percentage of errors completing comprehension questions, r(16) ¼ .35, p ¼ .19.

2.2.3. Test of Interpretation Bias I: Reaction times Probe scenarios were interspersed with the training or placebo scenarios to assess immediate on-line interpretations. Probe scenarios were also three-lines in length and described an ambiguous situation. Each block contained an OC-irrelevant (i.e. healthy) and an OC-relevant probe, resulting in 12 probes per session. Both groups received the same probe scenarios. An example of an OC-irrelevant probe scenario is as following: You sit in the bus to school and an ill-looking man enters the bus. When passing by, he brushes against your shoulder and you think you might also become ill due to this. This is an _________ thought.

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

exagger-ted exaggerated An example of an OC-relevant probe scenario is as following: You walk to your train platform, where your train leaves in five minutes. You see a used injection needle lying on the floor. Seeing this needle is _________ danger-us dangerous The speed of correctly resolving these probe word fragments was used as an index for interpretations; a healthy non-OC interpretation style is expected to facilitate solving OC-irrelevant probes and impede solving OC-relevant probes.

2.2.4. Test of Interpretation Bias II: Recognition test The Recognition Test was used as a test of off-line interpretation bias and was completed before and after the entire training procedure. The recognition task consisted of two phases (Salemink & van den Hout, 2010). In phase one (encoding phase), six scenarios describing ambiguous OC-related situations were presented (different scenarios for the pre- and post-training test) that tapped into each of the six OCCWG defined belief domains (OCCWG, 2011). Each scenario had a title, accompanying picture, and consisted of three lines. A missing word in the final sentence appeared as a word fragment once participants had pressed the space bar. As in the training trials, participants were asked to complete the fragment, however here the valence of the scenario remained ambiguous. Afterwards, participants were asked to answer a question about an irrelevant aspect of the scenario to motivate reading the scenario carefully and then feedback was presented (correct answer vs. incorrect answer). An example of an ambiguous scenario from the first part of the recognition task is as follows: With a friend at McDonald's You are eating a McFlurry ice cream with a friend at McDonald's. Suddenly, you imagine pushing your ice cream in your friend's face. You don't want to do that, but can't stop thinking _____ it. ab-ut

115

Participants rated each interpretation for its similarity in meaning to the original scenario (presented in phase one). A 4point scale was used, ranging from 1 (no match at all) to 4 (excellent match). Endorsement of OC-relevant versus healthy OCirrelevant interpretations in this second phase was used as a second indicator of interpretation bias.

2.2.5. Clinical measures The OBQ-CV (Coles et al., 2010; Wolters et al., 2011) is a selfreport questionnaire used for assessing obsessive beliefs. It contains 44 items rated on a 5-point scale (1 ¼ disagree very much to 5 ¼ agree very much). Higher scores indicate more obsessive beliefs. Cronbach's alpha of the Dutch version of the OBQ-CV was .95, and test-retest reliability was adequate (Wolters et al., 2011). The CY-BOCS (Scahill et al., 1997) is a clinician-rated semistructured interview used to assess severity of OC symptoms. The CY-BOCS severity scale is divided into an obsession and a compulsion subscale. Each subscale contains five items concerning frequency/time, interference, distress, resistance, and level of control which are rated on a 5-point scale ranging from 0 to 4 over the past week. In the current study, the clinicians were not blind to condition. Of note, research in this field of pediatric OCD suggests that therapists and independent evaluators produce similar ratings and that blindness is not essential for making accurate improvement ratings (Lewin, Peris, De Nadai, McCracken, & Piacentini, 2012). The total score, the sum of both subscales, ranges from 0 to 40 (higher scores reflect more symptom severity). The instrument has demonstrated good internal consistency, good testretest reliability and adequate divergent and convergent validity (Storch et al., 2004). The RCADS (Chorpita et al., 2005) is a self-report questionnaire to assess symptoms of anxiety and depression. The RCADS comprises 47 items and answers are given on a 4-point scale (0 ¼ never to 3 ¼ always). It contains six subscales: separation anxiety disorder, social phobia, obsessive-compulsive disorder, panic disorder, generalized anxiety disorder, and major depressive disorder. The OCD subscale has good internal consistency (Cronbach's a ¼ .82) and adequate divergent and convergent validity (Chorpita et al., 2005). The Children's Depression Inventory (CDI; Kovacs, 1992) is a 27item self-report questionnaire assessing depressive symptoms. Items are rated on a 3-point scale. Total scores range from 0 to 54 and higher scores reflect more depressive symptoms. Internal consistency (Cronbach's a) in a Dutch sample was .85 (Timbremont, Braet, & Roelofs, 2008).

about Were you eating a BigMac with your friend? The second phase of the Recognition test is the recognition phase. To cue recognition of a specific scenario from the encoding phase, participants first saw the identifying title with the accompanying picture. These were followed one at a time by two interpretations of that scenario: a healthy OC-irrelevant and an OCrelevant interpretation (see also Salemink & Wiers, 2011; Vassilopoulos et al., 2009). Two interpretations of the previously presented ambiguous test scenario are as following: a) You have a strange image of pushing your ice cream in your friend's face. When thinking about it, you have to laugh b) You have a strange image of pushing your ice cream in your friend's face. When thinking about it, you see it as a sign that you are a bad person

2.3. Procedure During the first session (at the Bascule), participants completed the self-report questionnaires (RCADS, OBQ-CV, and CDI), and the CY-BOCS interview was conducted. The computer program was then started, and participants were told to imagine themselves being in the situations described in the scenarios. They completed the pre-training Recognition test and three neutral practice trials of the scenarios. The CBM-I/placebo training was completed online at home or in any other location with internet access and lasted for approximately 30 min. Participants had to complete eight sessions in 11 days. Afterwards, participants completed the post-training Recognition test. The follow-up assessment (questionnaires, CY-BOCS interview) was approximately two weeks after initial assessment. Participants were debriefed and received a small financial reimbursement.

116

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

3. Results

Table 3 Mean recognition ratings for healthy OC-irrelevant and OC-relevant interpretations (SD in parentheses) pre- and post-training for the CBM-I and placebo control group.

3.1. Pre-training group comparisons

CBM-I group (n ¼ 9)

Comparison of the two groups at baseline revealed no significant differences in demographic characteristics, duration of OC symptoms, setting (inpatient vs. outpatient treatment), nor pharmacotherapy, p‘s > .13 (Table 1). In addition, the groups did not differ significantly on the pre-assessment obsessive beliefs and clinical outcome measures, p‘s > .28 (Table 2). Finally, groups did not differ significantly in their pre-training OC-irrelevant, t(14) ¼ .56, p ¼ .58, and OC-relevant, t(14) ¼ 1.48, p ¼ .16, interpretations (Table 3). 3.2. CBM-I effects on interpretation bias Reaction time data were excluded if the response to the probe word fragment (8.1%) or comprehension question (9.0%) was incorrect or if the latency was less than 200 ms (0.07%). Two patients missed one session and were excluded from these analyses. Independent samples t tests were conducted to examine whether the groups differed in their accuracy completing word fragments and comprehension questions in the probe scenarios. Results revealed that the groups did not differ significantly in the percentage of errors completing word fragments (MCBM-I group ¼ 8.3, SD ¼ 11.6, Mplacebo group ¼ 7.9, SD ¼ 4.7, t(14) ¼ 0.10, p ¼ .93), nor comprehension questions (MCBM-I group ¼ 9.8, SD ¼ 12.2, Mplacebo group ¼ 8.0, SD ¼ 6.4, t(14) ¼ 0.35, p ¼ .73). To test whether the CBM-I procedure was effective in influencing the time to complete OC-relevant and OC-irrelevant on-line interpretations, the reaction time data were subjected to a 2 (Group)  2 (Valence probe)  8 (Time) repeated measures mixed model ANOVA. This analysis revealed significant main effects of Valence, F(1, 12) ¼ 21.0, p ¼ .001, hp 2 ¼ 0.64, Time, F(7, 6) ¼ 6.7, p ¼ .02, hp 2 ¼ 0.89, and Group, F(1, 12) ¼ 4.8, p ¼ .05, hp 2 ¼ 0.29, indicating quicker responses to OC-irrelevant probes (M ¼ 1191, SD ¼ 59) vs. OC-relevant probes (M ¼ 1504, SD ¼ 105) and towards the end of training. Patients in the placebo condition responded more quickly (M ¼ 1177, SD ¼ 118) than patients in the training condition (M ¼ 1519, SD ¼ 102). More importantly, a significant Group  Valence interaction effect was observed, F(1, 12) ¼ 7.1, p ¼ .02, hp 2 ¼ 0.37 (Fig. 1). Adolescents who followed the CBM-I Table 2 Mean scores of the questionnaire outcome measures (SD in parentheses) pre- and post-training for the CBM-I and placebo control group. CBM-I group (n ¼ 9)

Placebo control group (n ¼ 7)

Pre-training Post-training Pre-training OC-irrelevant 2.3 (0.5) interpretations OC-relevant 2.6 (0.6) interpretations

Post-training

2.5 (0.6)

2.5 (0.9)

3.0 (0.5)

2.2 (0.7)

2.1 (0.8)

2.0 (0.5)

training were significantly slower in completing OC-relevant probes compared to individuals in the placebo condition, t(12) ¼ 2.5, p ¼ .03, Cohen's d ¼ .64. In other words, it was more difficult (took longer) for adolescents in the CBM-I training condition to interpret ambiguity in an OC-related manner. There was no group difference in time to solve OC-irrelevant probes, t(12) ¼ 1.4, p ¼ .20, d ¼ 0.44. Note that the three-way Group  Valence  Time interaction effect was not significant, F(7, 6) ¼ 2.5, p ¼ .15, hp 2 ¼ 0.74. A 2 (Group)  2 (Valence)  2 (Time) repeated measures mixed model ANOVA was conducted on mean recognition ratings to assess CBM-I effects on more slower off-line interpretations and revealed a significant Valence  Time interaction effect, F(1, 14) ¼ 12.4, p ¼ .003, hp 2 ¼ 0.47. OC-irrelevant interpretations increased significantly over time, t(15) ¼ 3.0, p ¼ .01, d ¼ 0.63, while there was a trend for a decrease in OC-relevant interpretations, t(15) ¼ 1.9, p ¼ .08, d ¼ 0.37. Note, these effects are independent of training group; the three-way interaction effect of Group  Valence  Time interaction effect was not significant, F(1, 14) < 0.001, p ¼ .99, hp 2 < 0.01 (see Table 3). 3.3. CBM-I effects on obsessive beliefs To examine the effects of CBM-I on self-reported obsessive beliefs, a 2 (Group)  2 (Time) repeated measures mixed model ANOVA was carried out on OBQ-CV scores (see Table 2) and revealed a significant decrease in obsessive beliefs, Time: F(1, 14) ¼ 5.9, p ¼ .029, hp 2 ¼ 0.30, independent of group assignment (Group  Time interaction effect F(1, 14) ¼ 1.1, p ¼ .32, hp 2 ¼ 0.07). 3.4. CBM-I effects on OC symptoms To examine the impact of CBM-I on OC symptoms, 2 (Group)  2 (Time) repeated measures mixed model ANOVAs were carried out

Placebo control group (n ¼ 7)

Pre-training Post-training Pre-training Post-training Obsessive beliefs 133.0 (18.9) 115.4 (31.6) 127.9 (37.5) 120.7 (38.4) (OBQ-CV) Obsessions 12.8 (4.4) 10.0 (4.4) 10.6 (3.2) 10.0 (2.5) (CY-BOCS subscale) Compulsions 11.1 (3.9) 9.6 (4.8) 9.9 (2.5) 8.0 (3.1) (CY-BOCS subscale) OC-symptoms 23.9 (7.6) 19.6 (9.1) 20.4 (4.3) 18.0 (4.4) (CY-BOCS Total score) OC-symptoms 10.5 (2.8) 7.8 (4.1) 9.7 (3.4) 9.7 (3.0) (RCADS OCD subscale) Anxiety symptoms 38.2 (8.7) 29.7 (10.4) 37.4 (17.5) 35.9 (16.8) (RCADS anxiety score) Depressive symptoms 14.6 (7.5) 13.0 (9.0) 14.6 (6.6) 16.3 (8.5) (CDI) The OBQ-CV stands for Obsessive Beliefs Questionnaire-Child Version, CY-BOCS for Child Yale-Brown Obsessive Compulsive Scale, RCADS for Revised Child Anxiety and Depression Scale Child Version, and CDI for Children's Depression Inventory.

Fig. 1. Mean reaction times and standard errors (in ms) to solve healthy OC-irrelevant and OC-relevant probes during training depicted for both the CBM-I training and the placebo control group.

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

on the clinician rated CY-BOCS obsession and compulsion scales, and on the RCADS OCD scale. For the CY-BOCS obsession scale, analyses revealed a significant main effect of Time, F(1, 14) ¼ 8.4, p ¼ .01, hp 2 ¼ 0.38, and a trend for the predicted Group  Time interaction effect, F(1, 14) ¼ 3.7, p ¼ .08, hp 2 ¼ 0.21. We inspected this interaction effect further given our specific hypotheses. In line with our prediction, there was a significant decrease in obsessions in the CBM-I group, t(8) ¼ 3.8, p ¼ .005, d ¼ 0.64, while there was no significant change in the placebo group, t(6) ¼ 0.6, p ¼ .56, d ¼ 0.21. For the compulsion scale, a significant decrease was observed, Time, F(1, 14) ¼ 9.3, p ¼ .009, hp 2 ¼ 0.40, independent of group assignment (Group  Time effect, F(1, 14) ¼ 0.1, p ¼ .80, hp 2 ¼ 0.005). Analyses of the RCADS OCD subscale revealed a trend for the predicted Group  Time interaction effect, F(1, 14) ¼ 3.7,p ¼ .08, hp 2 ¼ 0.21. Again, we inspected this interaction effect. OC symptoms decreased significantly in the CBM-I group, t(8) ¼ 3.0, p ¼ .018, d ¼ 0.78, but did not in the placebo group, t(6) ¼ 0.0, p ¼ 1.0, d ¼ 0. 3.5. Generalization to anxiety and depressive symptoms We also explored CBM-I effects on anxiety and depressive symptoms (Table 2). The two-way repeated measures mixed model ANOVA with RCADS anxiety (minus OCD) revealed a significant effect of Time, F(1, 14) ¼ 9.5, p ¼ .008, hp 2 ¼ 0.40, and a trend for a Group  Time interaction effect, F(1, 14) ¼ 4.5, p ¼ .052, hp 2 ¼ 0.24. While there was no significant change in the placebo group, t(6) ¼ 1.0, p ¼ .37, d ¼ 0.09, there was a significant reduction in the CBM-I group, t(8) ¼ 3.3, p ¼ .01, d ¼ 0.89. For depressive symptoms, we found no significant effects, p's > .11. 4. Discussion The aim of the current study was to test whether eight-sessions of OC-specific Cognitive Bias Modification training of Interpretations (CBM-I) could improve TAU for adolescents with OCD. OCD patients receiving TAU (CBT with or without pharmacotherapy) were randomly allocated to online CBM-I training or to an online placebo-control training condition. Results indicated that CBM-I was successful in reducing the speed of making OC-related immediate on-line interpretations; that is, adolescents who received the CBM-I training needed more time to complete the OC-relevant probes compared to adolescents who received the placebo condition. There was no significant difference in the time to complete OC-irrelevant probes, suggesting that the training resulted in reducing the likelihood or speed of interpreting information in an OC-related manner specifically. Note that while previous CBM-I studies on OCD did not observe any effects on probes (Clerkin & Teachman, 2011) or did not assess it (Beadel et al., 2014; Williams & Grisham, 2013), this CBM-I effect of slower responding to threatening probes has been observed previously (Salemink et al., 2014; Salemink, van den Hout, & Kindt, 2009). Little is known regarding the underlying mechanism of CBM-I and what might explain the differential effects on the two types of probes (i.e., OC-relevant vs. OC-irrelevant). One possibility is that due to the repeated pairing of ambiguous OC-relevant situations with positive interpretations, more positive interpretations come to mind when encountering ambiguous information, without affecting the speed of interpreting ambiguity positively (i.e., no faster responding to OC-irrelevant probes). The suggested increased number of positive interpretations might be associated with a delay in negative interpretations coming to mind, which is then reflected in the slower resolution of OC-relevant probes. Another possible explanation for the discrepant CBM-I finding on OC-relevant and irrelevant probes might be that CBM-I training is

117

associated with increased control over threat-related information, in this case OC-relevant interpretations. Participants in the CBM-I condition might have learned to inhibit OC-related interpretations, as supported by their slower responses to the OCrelevant probes. This is consistent with recent CBM-attention training findings, where selective inhibition of threat stimuli mediated the effects of training on anxiety (Browning, Holmes, Murphy, Goodwin, & Harmer, 2010). Gaining control over threatrelated stimuli might be (one of) the mechanisms of CBM training. Admittedly, these are post-hoc and tentative explanations, though the latter fits well with a cognitive model of anxiety that described biased information processing as the result of an interplay between a threat and a regulatory control system (Mathews & Mackintosh, 1998) and with recent empirical findings illustrating the beneficial role of strong regulatory control processes in adolescent threat-related interpretive bias (Salemink & Wiers, 2012). With respect to slower retrospective off-line interpretations; an increase in healthy OC-irrelevant interpretations was observed, independent of group assignment. Thus, all adolescents interpreted ambiguity presented in the recognition task in a more healthy way. As all adolescents were receiving TAU for their OCD (CBT with or without pharmacotherapy), this likely reflects the effect of the TAU on these off-line interpretations. One possibility is that TAU was so effective in modifying these interpretations (i.e., a ceiling effect) that there was no room to detect any additional benefits of CBM-I training. CBM-I's extra therapeutic benefit may be in changing the immediate on-line interpretations. A tentative conclusion might be that TAU and CBM-I are complementary interventions that influence different aspects of interpretation bias. In CBT, cognitive restructuring techniques are used to guide patients to reevaluate the meaning and interpretation of intrusions. Compared to CBM-I, this seems a more explicit, re-interpretation strategy that potentially also affects retrospective off-line interpretations the most. In contrast, CBM-I training does not require explicit reevaluation, but patients repeatedly practice interpreting ambiguity in a healthy, OC-irrelevant way. They might learn to associate ambiguity with healthy interpretations directly, thus having the greatest impact on immediate, on-line interpretations. Regarding clinical outcomes, consistent with expectations, patients reported significantly fewer OC symptoms after CBM-I training, while patients in the placebo condition did not. Similarly, clinicians rated patients in the CBM-I condition as having significantly fewer obsessive symptoms, but not patients in the placebo group. Furthermore, CBM-I effects generalized to selfreported anxiety symptoms. These findings are consistent with a cognitive model of OCD (Salkovskis, 1985) that describes a causal link between misinterpretation of intrusions and anxiety. Note that no effects were observed on obsessive beliefs, compulsive behaviors, or depressive symptoms. Previous studies examining OCrelated CBM-I training in adults with elevated levels of OC symptoms failed to observe consistent effects on either more clinical outcome measures such as distress in response to an OC-stressor or compulsive behavior (Beadel et al., 2014; Clerkin & Teachman, 2011; Williams & Grisham, 2013). This might not be entirely surprising as these previous studies used just a single session of training. It seems likely that multiple sessions of training are necessary to provide time and opportunity for the trained interpretations to affect the processing of ambiguity in daily life and consequently to have effect on OC symptoms (see also metaanalysis Menne-Lothmann et al., 2014). Although the durability of the current findings remains to be tested, and the findings need to be replicated, the effects in our study on OC symptoms conceptually replicate previous findings of the added value of CBM-attention training in adolescents with anxiety disorders (Riemann,

118

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119

Kuckertz, Rozenman, Weersing, & Amir, 2013). The findings also point to the clinical potential of CBM-I training as an add-on to more traditional treatments. Furthermore, since the training can be completed at home, does not involve stressful exercises, nor involvement of therapists, it is potentially an easy to implement, and cost-efficient novel intervention. Some study limitations should be acknowledged. First, as the sample size is small, replication with more participants is necessary to examine the robustness of the observed effects. Second, our Group  Time interaction effects for clinician rated obsessive symptoms, self-reported OC-, and anxiety symptoms were only a trend. This seems likely due to our small sample size, as the sizes of these interaction effects were large (hp 2 effect size estimates  0.21) and the changes in the CBM-I group were of medium-to-large size (Cohen's d ¼ 0.64e0.89). Third, the clinical outcome measures were obtained two weeks after completing the training, thus providing some information regarding the stability of change. To test the longevity of the training effects, future studies should include longer follow-up assessments. Such studies could also include a diary assessment to assess the occurrence (e.g., Lang, Moulds, & Holmes, 2009; Woud, Holmes, Postma, Dalgleish, & Mackintosh, 2012) and interpretation of intrusions in daily life. Fourth, in the current study, clinicians were not blind to condition. While it has been shown that nonblinded clinicians provide ratings that are consistent with blind independent evaluators' ratings of improvement of youth with OCD (Lewin et al., 2012), we cannot exclude the possibility that it might have affected clinicians' ratings in the current study. A final point concerns our training stimuli. Consistent with other OC-related CBM-I studies (Beadel et al., 2014; Clerkin & Teachman, 2011; Williams & Grisham, 2013), we developed one set of scenarios for all patients, containing beliefs covering all domains described by the OCCWG (2011). However, as an individual patient with OCD is unlikely to endorse all of these beliefs, it might be better in the future to match the content of the scenarios to the specific beliefs of a participant (tailored training). Future research is necessary to test whether such a tailored training increases the beneficial effects of CBM-I. Notwithstanding the above limitations, this small, randomized controlled trial provided the first evidence that eight sessions of computerized CBM-I training, without therapist assistance, augments treatment as usual (CBT with or without pharmacotherapy) in adolescents with OCD. CBM-I training reduced the speed of making OC-related interpretations and resulted in significant reductions in clinician-rated obsessive symptoms and self-reported OC-symptoms. No such changes were observed in the placebo condition. Findings tentatively support the added clinical potential of CBM-I to TAU treatment, and provide the basis for more largescale assessment of this novel treatment tool. Acknowledgments Elise Clerkin and Bethany Teachman are acknowledged for providing their CBM-I training scenarios. We are grateful to Carina Veldkamp, Ellen Holt, and Vivian op de Beek for help in developing the scenarios and collecting the data used in this study. We thank Kristen Anderson and Belinda Platt for helpful comments on an earlier draft of the manuscript. Elske Salemink is supported by a grant from the Dutch Scientific Foundation (NWO VENI 451-10-029). References American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association. Baert, S., De Raedt, R., & Koster, E. H. W. (2011). Modification of information processing biases in emotional disorders: clinically relevant developments in

experimental psychopathology. International Journal of Cognitive Therapy, 4, 208e222. Beadel, J. R., Smyth, F. L., & Teachman, B. A. (2014). Change processes during Cognitive Bias Modification for obsessive compulsive beliefs. Cognitive Therapy and Research, 38, 103e119. Browning, M., Holmes, E. A., Murphy, S. E., Goodwin, G. M., & Harmer, C. J. (2010). Lateral prefrontal cortex mediates the cognitive modification of attentional bias. Biological Psychiatry, 67, 919e925. Chorpita, B. F., Moffitt, C. E., & Gray, J. (2005). Psychometric properties of the Revised Child Anxiety and Depression Scale in a clinical sample. Behaviour Research and Therapy, 43, 309e322. Clerkin, E. M., & Teachman, B. A. (2011). Training interpretation biases among individuals with symptoms of obsessive compulsive disorder. Journal of Behavior Therapy and Experimental Psychiatry, 42, 337e343. Coles, M. E., Wolters, L. H., Sochting, I., De Haan, E., Pietrefesa, A. S., & Whiteside, S. P. (2010). Development and initial validation of the Obsessive Belief Questionnaire-Child Version (OBQ-CV). Depression and Anxiety, 27, 982e991. De Haan, E., & Wolters, L. (2009). Behandeling van de dwangstoornis bij kinderen en adolescenten. Met het cognitief-gedragstherapeutisch protocol Bedwing je Dwang [Treatment of OCD in children and adolescents. With the cognitive behavioural treatment protocol Control your OCD]. Houten: Bohn Stafleu van Loghum. Frost, R. O., & Steketee, G. (2002). Cognitive approaches to obsessions and compulsions: Theory, assessment, and treatment. Access Online via Elsevier. Fu, X., Du, Y., Au, S., & Lau, J. Y. F. (2013). Reducing negative interpretations in adolescents with anxiety disorders: a preliminary study investigating the effects of a single session of Cognitive Bias Modification training. Developmental Cognitive Neuroscience, 4, 29e37. Geller, D. A., & March, J. (2012). Practice parameter for the assessment and treatment of children and adolescents with obsessive-compulsive disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 51, 98e113. Hirsch, C. R., & Clark, D. M. (2004). Information-processing bias in social phobia. Clinical Psychology Review, 24, 799e825. Kovacs, M. (1992). Children's depression inventory manual. North Tonawanda, NY: Multi Health Systems. Lang, T. J., Moulds, M. L., & Holmes, E. A. (2009). Reducing depressive intrusions via a computerized Cognitive Bias Modification of appraisals task: developing a cognitive vaccine. Behaviour Research and Therapy, 47, 139e145. Lau, J. Y. F., Belli, S. R., & Chopra, R. B. (2013). Cognitive Bias Modification training in adolescents reduces anxiety to a psychological challenge. Clinical Child Psychology and Psychiatry, 18, 322e333. Lewin, A. B., Peris, T. S., De Nadai, A. S., McCracken, J. T., & Piacentini, J. (2012). Agreement between therapists, parents, patients, and independent evaluators on clinical improvement in pediatric obsessive-compulsive disorder. Journal of Consulting and Clinical Psychology, 80, 1103e1107. Lothmann, C., Holmes, E. A., Chan, S. W. Y., & Lau, J. Y. F. (2011). Cognitive Bias Modification training in adolescents: effects on interpretation biases and mood. The Journal of Child Psychology and Psychiatry, 52, 24e32. Mathews, A., & Mackintosh, B. (1998). A cognitive model of selective processing in anxiety. Cognitive Therapy and Research, 22, 539e560. Matthews, L., Reynolds, S., & Derisley, J. (2007). Examining cognitive models of obsessive symptoms in adolescence: a preliminary investigation. Journal of Clinical Child and Adolescent Psychology, 33, 743e749. €hn, P., Kasanova, Z., Haller, S. P., Menne-Lothmann, C., Viechtbauer, W., Ho Drukker, M., et al. (2014). How to boost positive interpretations? A metaanalysis of the effectiveness of Cognitive Bias Modification for Interpretation. PloS One, 9, e100925. Obsessive Compulsive Cognitions Working Group (OCCWG). (2011). Development and initial validation of the obsessive beliefs questionnaire and the interpretation of intrusions inventory. Behaviour Research and Therapy, 39, 987e1006. 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, 228e233. Riemann, B. C., Kuckertz, J. M., Rozenman, M., Weersing, V. R., & Amir, N. (2013). Augmentation of youth cognitive behavioral and pharmacological intervenetions with attention modification: a preliminary investigation. Depression and Anxiety, 30, 822e828. Salemink, E., Kindt, M., Rienties, H., & van den Hout, M. (2014). Internet-delivered Cognitive Bias Modification of Interpretations in patients with anxiety disorders: a randomised controlled trial. Journal of Behavior Therapy and Experimental Psychiatry, 45, 186e195. Salemink, E., & van den Hout, M. (2010). Validation of the “recognition task” used in the training of interpretation biases. Journal of Behavior Therapy and Experimental Psychiatry, 41, 140e144. Salemink, E., van den Hout, M., & Kindt, M. (2009). Effects of positive interpretive bias modification in highly anxious individuals. Journal of Anxiety Disorders, 23, 676e683. Salemink, E., & Wiers, R. W. (2011). Modifying threat-related interpretive bias in adolescents. Journal of Abnormal Child Psychology, 39, 967e976. Salemink, E., & Wiers, R. W. (2012). Adolescent threat-related interpretive bias and its modification: the moderating role of regulatory control. Behaviour Research and Therapy, 50, 40e46. Salkovskis, P. M. (1985). Obsessional-compulsive problems: a cognitive-behavioural analysis. Behaviour Research and Therapy, 23, 571e583.

E. Salemink et al. / J. Behav. Ther. & Exp. Psychiat. 49 (2015) 112e119 zar, A. I., Iniesta-Sepúlveda, M., & Rosa-Alca zar, A. S anchez-Meca, J., Rosa-Alca (2014). Differential efficacy of cognitive-behavioral therapy and pharmacological treatments for pediatric obsessive-compulsive disorder: a meta-analysis. Journal of Anxiety Disorders, 28, 31e44. Scahill, L., Riddle, M. A., McSwiggin-Hardin, M., Ort, S. I., King, R. A., Goodman, W. K., et al. (1997). Children's Yale-Brown Obsessive Compulsive Scale: reliability and validity. Journal of the American Academy of Child & Adolescent Psychiatry, 36, 844e852. Storch, E. A., Murphy, T. K., Geffken, G. R., Soto, O., Sajid, M., Allen, P., et al. (2004). Psychometric evaluation of the Children's Yale-Brown obsessive-compulsive scale. Psychiatry Research, 129, 91e98. Timbremont, B., Braet, C., & Roelofs, J. (2008). Children's depression inventory. Nederlandstalige versie. Handleiding [Children's Depression Inventory. Dutch version. Manual]. Amsterdam: Pearson Assessment and Information.

119

Vassilopoulos, S. P., Banerjee, R., & Prantzalou, C. (2009). Experimental modification of interpretation bias in socially anxious children: changes in interpretation, anticipated interpersonal anxiety, and social anxiety symptoms. Behaviour Research and Therapy, 47, 1085e1089. Williams, A. D., & Grisham, J. R. (2013). Cognitive Bias Modification (CBM) of obsessive compulsive beliefs. BMC Psychiatry, 13, 256. Wolters, L. H., Hogendoorn, S. M., Koolstra, T., Vervoort, L., Boer, F., Prins, P. J. M., et al. (2011). Psychometric properties of a Dutch version of the Obsessive Beliefs Questionnaire-Child Version (OBQ-CV). Journal of Anxiety Disorders, 25, 714e721. Woud, M. L., Holmes, E. A., Postma, P., Dalgleish, T., & Mackintosh, B. (2012). Ameliorating intrusive memories of distressing experiences using computerized reappraisal training. Emotion, 12, 778e784.

Augmentation of Treatment As Usual with online Cognitive Bias Modification of Interpretation training in adolescents with Obsessive Compulsive Disorder: A pilot study.

Cognitive Behavioral Therapy for children and adolescents with Obsessive Compulsive Disorder (OCD) is effective. However, since almost half of patient...
343KB Sizes 0 Downloads 7 Views