2% of the worldwide population, this means around 150 million individuals, suffer from obsessive compulsive disorder (OCD; Sasson et al.,1997). A psychological disorder, characterised by obsessions, which are intrusive, repetitive and troubling thoughts, next to compulsions which manifest in repetitive ritualistic behaviours, mainly to neutralize obsessions (Williams & Grisham, 2013; Lack, 2012). When left untreated OCD had a chronic course (Abramowitz, Taylor & McKay, 2009) A construct that is highly related to OCD and hypothesized as a maintaining factor is Thought-action fusion (TAF; Storch & Lewin, 2016). TAF is an interpretation bias that leads to overvaluing your thoughts and the belief that thinking something, is equal to carrying out the action (Rassin, Diepstraten, Merkelbach & Muris, 2001; Siwiec, Davine, Kresser, Rohde & Lee, 2016). It consists of two subtypes, namely TAF moral, the belief that having a thought is comparable to carrying out the action and TAF likelihood, the believe that thinking of a concerning event increases the likelihood of that event happening (Siwiec et al., 2016). While most individuals have intrusive thoughts, people with OCD consider them as a threat for which they are responsible. Therefore, TAF leads to emotional distress, thought suppression and compulsions.
A relatively new way of modifying such a bias is through Cognitive Bias Modification for Interpretation (CBM-I). In the last couple of years there have been a few studies that aimed on modifying TAF with CBM-I (Clerking & Teachman, 2011; Williams, et al., 2013; Siwiec, et al., 2016). These studies used a sentence completion paradigm, in which they present participants with ambiguous information (e.g. eating with a friend the thought of poking my friend’s eye with my fork came into head) and steer them in the direction of a TAF incongruent interpretation (e.g. having this thought is meani_gl_ss) instead of a TAF congruent interpretation (e.g. unacc_pt_ble). Siwiec et al. (2016) examined TAF specifically and found that TAF could be reduced significantly by 31% of the participants in the experimental group (the TAF incongruent condition, TAFINC) with respect to 8% in the control group (the TAF congruent condition, TAFCON). Nevertheless, studies have always focused on students presenting OC-symptoms, therefore there is no evidence that these results generalize to clinical samples with OCD. The aim of this research is to replicate and extend the study of Siwiec et al. (2016), to test the added value of CBM-I in treating TAF in the OCD-population. Based on the research above it is hypothesized that after the CBM-I training (1) the TAFINC group shows significantly less TAF interpretation bias and (2) the TAFINC group shows significantly less distress.
Design and analyses
Participants
In extension of Siwiec et al. (2016), patients with an OCD-diagnosis should be recruited. This means individuals who meet the Diagnostic Statistic Manual (DSM) criteria for OCD (see appendix). Therefore, a non-probability form of sampling is used to recruit patients suffering from OCD. These individuals will be recruited through treatment centres and clinics.
Certain in- and exclusion criteria are set to make sure participants meet the criteria of OCD and form a representative sample. At first people are examined by trained assessors using the The Yale-Brown obsessive-compulsive scale (YBOCS; Goodman, et. al., 1989), a semi-structured interview that assesses the presence and severity of obsessions and compulsions. The YBOCS is a scale with good psychometric properties (Rapp, Bergman, Piacentini & McGuire, 2016). Adults (>18 years) with a score of eight or higher (mild to extreme OCD-symptoms meeting the DSM criteria) are included in the study. Interrater reliability will be taken into account. Since OCD is associated with substantial comorbidity (Ruscio, Stein, Chiu & Kessler, 2010), this will not be an exclusion criterion. With the exception of severe depression, suicidal ideation, substance abuse and a history of psychosis, because these symptoms can interfere with the CBM-I task (Williams & Grisham, 2013).
Measures
The Thought Action Fusion scale (TAFS; Shafran, Thordarson, & Rachman, 1996) is the first measure used. A scale developed to quantify TAF severity on a 5-point Likert scale from zero (disagree strongly) to four (agree strongly). It exists of 19 items and measures TAF moral (TAF-M) as well as TAF likelihood (TAF-L). Higher scores point out higher levels of TAF biases. The TAF-total and both its subscales are used as indicators of CBM-I treatment success. Rassin, Merckelbach, Muris en Schmidt (2001) found a good internal consistency and construct validity for the scale.
Different from the study of Siwiec et al. (2016) this study will use an interpretation recognition task (IREC-T Matthews & Mackintosh, 2000) to have a more objective measure of TAF. This task checks whether there is indeed a change in interpretation instead of an induced response bias (Beard, 2011). The task exists of similar scenario’s as the training session which remain ambiguous, participants should then rate four sentences on their correspondence with this scenario. One of the sentences represents a TAFINC interpretation, one a TAFCON interpretation and the other two will be foil sentences, which are not possible interpretations of the scenario. Each of those four interpretations can be rated on a scale of one (not at all) to four (fully). The recognition task, is a regular used method after CBM-I (Mackintosh et. al., 2006; Klein, de Voogd, Wiers & Salemink, 2017; Yiend, Mackintosh, & Matthews, 2005) and has been found to be reliable (Mackintosh et al., 2006) and valid (Salemink & van den Hout, 2010).
To measure the amount of distress participants experience before and after the CBM-I training, the Revised obsessional intrusions inventory-Distress (ROII-Distress; Purdon & Clark, 1994) is used. This inventory measures how distressing various intrusive thoughts are based on 52 different possible mental intrusions. This is done with a 7-point Likert scale from zero (not distressing) tot six (extremely distressing). At last the Mini International Neuropsychiatric Interview (MINI; Sheehan et al., 1998) is used to terminate whether participants have comorbid disorders. Research supports the MINI’s validity and reliability (Sheehan et. al, 1998). When analysing comorbidity can be taken into account. The MINI will be administered by two trained assessors.
Procedure
The study starts with a screening using the YBOCS and the MINI, after this the TAFS and ROII are administered as a pre-test. Following all participants are randomly assigned to either the TAFINC or the control condition. Because this study examines participants with an actual OCD diagnosis there is chosen to deviate from a TAFCON condition, because of the amount of distress this can bring on patients. The CBM-I treatment in the control condition will therefore exist of both TAFINC as TAFCON items and will form a neutral CBM-I training.
In the CBM-I training the experimental group, TAFINC, starts by seeing a TAF inducing thought on the computer (e.g. you are driving and suddenly you think of hitting the driver in front of you). The thought ends with a sentence the participant should complete in a TAF incongruent matter (e.g. this thought is n_rm_l, because a lot of people think this). In each sentence, there is a key word missing two letters which participants should fill in correctly. The control condition will have these TAFINC examples as well but these will alternate with TAFCON sentences (e.g. you are driving and suddenly you think of hitting the driver in front of you, this thought is d_ng_rous, because it increases the chance of it happening). If people make a mistake an error will be shown and they should correct their answer. Half of the scenarios will be TAF-M items and half of them TAF-L. After people have given the right answer they are shown a compliment (e.g. good work, go on). The CBM-I session will last around 20 minutes. After the CBM-I training the TAFS and ROII are administered again, together with the recognition task.
Data analysis
The pre- and post-test of both conditions are analysed by using the Statistical Package for the Social Sciences (SPSS, 2016). To start the chi-square (χ2) and independent sample t-test are used to test whether there are significant differences between the TAFINC condition and the control condition. This will be done for both demographics as outcome measures, this means the TAFS, ROII, the behavioural tasks and the YBOCS.
To test the hypotheses, post-test scores on the TAFS, recognition task and the ROII are analysed using ANCOVA. This will control for possible pre-training differences. To analyse if there is a correlation between the severity of symptoms measured with the YBOCS and the effect on TAFS a two-sided Pearson’s correlation will be used. This Pearson’s correlation will also be used for the amount of distress people experience on the ROII and the effect on TAFS. These analyses taken into account, the sample size could be calculated using the program G*power (Erdfelder, Faul & Buchner, 1996). A meta-analysis of Sharpe and Jones (2017), showed that CBM-I effect sizes (ES) lie in between 0.52-0.81 therefore an ES of 0.52 is used to calculate the sample size. The ES in combination with a power of 0.80 and a significance level of 0.05 compute a sample size of at least 47 participants in each group.