Inflated responsibility

As defined by the OCCWG (2001, pp. 1002, 1003), inflated responsibility refers to ‘the belief that one has power that is pivotal to bring about or prevent subjectively crucial negative outcomes perceived as essential to prevent and that may have consequences in the real world and/or at a moral level’. This influential construct has received attention both as a singular entity and alongside other belief domains (e.g., Freeston, Rhéaume, & Ladouceur, Reference Freeston, Rhéaume and Ladouceur1996; OCCWG, 1997, 2005; Rachman, Reference Rachman1993; Rhéaume, Ladouceur, Freeston, & Letarte, Reference Rhéaume, Ladouceur, Freeston and Letarte1995; Salkovskis, Reference Salkovskis1985, Reference Salkovskis1989, Reference Salkovskis, Shafran, Rachman and Freeston1999; Salkovskis, Richards, & Forrester, Reference Salkovskis, Richards and Forrester1995; Van Oppen & Arntz, Reference Van Oppen and Arntz1994; Williams et al., Reference Williams, Salkovskis, Forrester and Allsopp2002). Salkovskis (Reference Salkovskis1985, Reference Salkovskis1989) originally proposed that intrusive thoughts are interpreted in a way that indicates personal responsibility for possible harm or negative consequences occurring to the self or to others. This may involve acts of omission (failure to prevent harm) or acts of commission (causing harm) (Wroe & Salkovskis, Reference Wroe and Salkovskis2000). A causal role is assigned to responsibility appraisals, whereby feeling personally responsible for preventing harm will directly result in compulsive behaviours aimed to decrease discomfort and the risk of a negative outcome (Salkovskis et al., Reference Salkovskis, Richards and Forrester1995). Following Salkovskis’ (Reference Salkovskis1985, Reference Salkovskis1989) proposal, the role of responsibility was highlighted by the OCCWG (1997) among the aforementioned five other belief domains, and later combined with threat estimation (OCCWG, 2005).

Within the substantial body of research focused on exploring the role of inflated responsibility appraisals in OCD, the large majority of studies are correlational and results appear mixed. First, several studies have found that responsibility beliefs are elevated in OCD patients compared to normal controls (Cartwright-Hatton & Wells, Reference Cartwright-Hatton and Wells1997; Freeston, Ladouceur, Gagnon, & Thibodeau, Reference Freeston, Ladouceur, Gagnon and Thibodeau1993; Kumar, Sharma, Kandavel, & Reddy, Reference Kumar, Sharma, Kandavel and Reddy2012; Niemeyer, Moritz, & Pietrowsky, Reference Niemeyer, Moritz and Pietrowsky2013; OCCWG, 2001, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Steketee, Frost, & Cohen, Reference Steketee, Frost and Cohen1998), especially those with checking compulsions (Foa, Amir, Bogert, Molnar, & Przeworski, Reference Foa, Amir, Bogert, Molnar and Przeworski2001; Foa, Sacks, Tolin, Prezworski, & Amir, Reference Foa, Sacks, Tolin, Prezworski and Amir2002). For example, Steketee et al. (Reference Steketee, Frost and Cohen1998) and Niemeyer et al. (Reference Niemeyer, Moritz and Pietrowsky2013) found that OCD participants scored higher than control participants on responsibility belief domains. This was further supported by a recent meta-analysis of 22 studies that found a large effect size favouring OCD patients compared to healthy controls (Pozza & Dèttore, Reference Pozza and Dèttore2014b). Responsibility beliefs have also been shown to correlate with obsessive compulsive (OC) symptoms (Niemeyer et al., Reference Niemeyer, Moritz and Pietrowsky2013; Steketee et al., Reference Steketee, Frost and Cohen1998) and significantly predict OC symptoms after controlling for metacognition (Niemeyer et al., Reference Niemeyer, Moritz and Pietrowsky2013). These associations are also evident in studies utilising non-clinical samples showing a link between responsibility and OC symptoms (Freeston et al., Reference Freeston, Ladouceur, Gagnon and Thibodeau1993; Gwilliam, Wells, & Cartwright‐Hatton, Reference Gwilliam, Wells and Cartwright‐Hatton2004; Myers, Fisher, & Wells, Reference Myers, Fisher and Wells2008; Parrish & Radomsky, Reference Parrish and Radomsky2011; Rhéaume, Freeston, Dugas, Letarte, & Ladouceur, Reference Rhéaume, Freeston, Dugas, Letarte and Ladouceur1995; Wilson & Chambless, Reference Wilson and Chambless1999). For example, Parrish and Radomsky (Reference Parrish and Radomsky2011) found that higher levels of responsibility were associated with higher anxiety and compulsive urges in a series of experimental vignettes. Similarly, Wilson and Chambless (Reference Wilson and Chambless1999) and Rhéaume et al. (Reference Rhéaume, Freeston, Dugas, Letarte and Ladouceur1995) found that inflated responsibility beliefs were significantly associated with OC symptoms in non-clinical adults and were a significant predictor of OC symptoms. However, such results are not found consistently; for example, recent research on student samples has found that responsibility is not a strong predictor of OC symptoms (Fitch & Cougle, Reference Fitch and Cougle2013; McKay et al., Reference McKay, Kim, Taylor, Abramowitz, Tolin, Coles and Olatunji2014) and does not predict OC symptoms when other belief domains are controlled for (Gwilliam et al., Reference Gwilliam, Wells and Cartwright‐Hatton2004; Myers et al., Reference Myers, Fisher and Wells2008; Myers & Wells, Reference Myers and Wells2005; Tolin, Woods, & Abramowitz, Reference Tolin, Woods and Abramowitz2003), which is surprising given its prominence in cognitive models of OCD (Salkovskis, Reference Salkovskis1985, Reference Salkovskis1989; Salkovskis et al., Reference Salkovskis, Shafran, Rachman and Freeston1999).

A similar picture is evident when comparing OCD patients and patients with other anxiety disorders. While some studies report that responsibility beliefs are specific to individuals with OCD, others report that responsibility may be related to anxiety in general, if at all. For example, while Foa et al. (Reference Foa, Amir, Bogert, Molnar and Przeworski2001), the OCCWG (2003), and Steketee et al. (Reference Steketee, Frost and Cohen1998) found stronger responsibility beliefs in individuals with OCD compared to those with other anxiety disorders, McKay et al. (Reference McKay, Kim, Taylor, Abramowitz, Tolin, Coles and Olatunji2014) and Tolin et al. (Reference Tolin, Worhunsky and Maltby2006) found that OCD patients did not endorse responsibility more strongly than either anxious or non-clinical controls. Similarly, Cartwright-Hatton and Wells (Reference Cartwright-Hatton and Wells1997) found that OCD patients did not endorse responsibility beliefs more strongly than patients with generalised anxiety disorder (GAD), although both patient groups showed elevated responsibility beliefs compared to non-clinical controls. Results such as these have led authors to conclude that inflated responsibility may be common among individuals with anxiety disorders, although OCD may be higher on the continuum (Foa et al., Reference Foa, Amir, Bogert, Molnar and Przeworski2001). Further, some authors suggest that responsibility may only differ between OCD and anxious patients and predict OCD symptoms (e.g., checking) when combined with threat estimation (Rachman, Reference Rachman2002). Following from this, a recent meta-analysis of 58 studies conducted by Pozza and Dèttore (Reference Pozza and Dèttore2014a) found that responsibility was more strongly associated with OCD than depression symptoms, but not anxiety disorder symptoms. Further, when analyses were limited to responsibility measures designed by the OCCWG, the relationship between responsibility and OCD was not stronger than that of depression or anxiety disorders. The authors thus concluded that the specificity of responsibility to OCD was at present questionable, and causal inferences on the role of responsibility in OCD could not be made due to the cross-sectional nature of the studies.

In terms of experimental research, various studies have been conducted to examine the effect of inflated responsibility on OCD symptoms and processes (e.g., Arntz, Voncken, & Goosen, Reference Arntz, Voncken and Goosen2007; Ladouceur, Rhéaume, & Aublet, Reference Ladouceur, Rhéaume and Aublet1997; Ladouceur et al., Reference Ladouceur, Rhéaume, Freeston, Aublet, Jean, Lachance and de Pokomandy-Morin1995; Mancini, D'Olimpio, & Cieri, Reference Mancini, D'Olimpio and Cieri2004). For example, Lopatka and Rachman (Reference Lopatka and Rachman1995) and Shafran (Reference Shafran1997) manipulated responsibility with OCD patients and found that increases in perceived responsibility resulted in significant elevations in discomfort and the urge to engage in compulsive checking/neutralising. Similarly, Mancini et al. (Reference Mancini, D'Olimpio and Cieri2004) found increased doubt, error preoccupation, checking, hesitations, and trial latency for their high compared to low responsibility condition with an unselected sample. Not all studies, however, showed such results. For example, Jones and Menzies (Reference Jones and Menzies1997a) found that urges to neutralise anxiety, time spent neutralising, and ‘time in dirt’ were associated with danger (i.e., threat) but not responsibility appraisals in an OCD sample. To compare such studies and synthesise results, a recent systematic review was conducted (Mantz & Abbott, Reference Mantz and Abbott2017). Fourteen studies using adult samples were identified, with the earliest published in 1995 and the most recent published in 2014. Of the five studies with OCD patients, only two included a control group to provide a frame of reference for the effects found with OCD participants. The authors found that while responsibility manipulations had a consistently adverse effect on responsibility and threat appraisals, the effect on self and observer-rated variables (e.g., urge to check, neutralising behaviours) was inconsistent, with mixed results. Further, the effects of responsibility were generally not significantly greater for OCD participants. Thus, it was concluded that the specificity of responsibility to OCD was at present unclear and that support for a responsibility bias in OCD was inconclusive. The need to explore other appraisal types (e.g., uncertainty, perceived threat, control) was also emphasised, particularly given the strong overlap with findings showing threat appraisals as an important contributor to OCD symptomatology.

Overestimation of threat

As defined by the OCCWG (2001, p. 1004), overestimation of threat refers to ‘an exaggerated belief about the probability (likelihood) or severity (cost) of harm’. This belief domain has received attention from many theorists and clinicians who identify threat cognitions as common among individuals with OCD and important in the conceptualisation of the disorder (Beck, Emery, & Greenberg, Reference Beck, Emery and Greenberg1985; Carr, Reference Carr1974; Foa & Kozak, Reference Foa, Kozak, Tuma and Maser1985; Jones & Menzies, Reference Jones and Menzies1997a; Marks, Reference Marks1987; McFall & Wollersheim, Reference McFall and Wollersheim1979; OCCWG, 1997, 2001, 2003; Rachman & Hodgson, Reference Rachman and Hodgson1980; Reed, Reference Reed1985; Van Oppen & Arntz, Reference Van Oppen and Arntz1994; Woods, Frost, & Steketee, Reference Woods, Frost and Steketee2002). Different components of threat have been identified, including overestimations of personal vulnerability and overestimation of general threat, among others, with debate as to which is most important in threat estimation (Moritz & Jelinek, Reference Moritz and Jelinek2009). For example, a person with OCD may wash their hands excessively after touching a door handle due to an overestimation of the objective probability of contamination, or an overestimation of personal vulnerability to experience a negative event (Moritz & Jelinek, Reference Moritz and Jelinek2009). If the primary issue lies in an inaccurate overestimation of general threat (i.e., base rates), providing corrective information may well improve OCD symptoms (Jones & Menzies, Reference Jones and Menzies1997b). This is evidenced in DIRT, an effective treatment for OCD washing and checking symptoms that focuses on reducing threat expectancies through structured psychoeducation and cognitive therapy (Jones & Menzies, Reference Jones and Menzies1998; Krochmalik et al., Reference Krochmalik, Jones, Menzies and Kirkby2004; Vaccaro et al., Reference Vaccaro, Jones, Menzies and Wootton2014). Additionally, individuals with OCD may overestimate the consequences of negative events instead of or in addition to overestimating the probability of such events occurring (Moritz & Jelinek, Reference Moritz and Jelinek2009). For example, individuals with fear of contamination might overestimate the probability of contracting an illness, as well as the severity of the illness that occurs. Perceived control and ability to cope have also been identified as important factors pertaining to the construct of threat (Barlow, Reference Barlow2004; McFall & Wollersheim, Reference McFall and Wollersheim1979).

The majority of research investigating overestimation of threat has been conducted using self-report questionnaires. Several studies have found that beliefs about threat are elevated in OCD patients compared to normal controls (Exner, Zetsche, Lincoln, & Rief, Reference Exner, Zetsche, Lincoln and Rief2014; Freeston et al., Reference Freeston, Ladouceur, Gagnon and Thibodeau1993; Mavrogiorgou et al., Reference Mavrogiorgou, Bethge, Luksnat, Nalato, Juckel and Brüne2016; OCCWG, 2001, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Steketee et al., Reference Steketee, Frost and Cohen1998; Tolin et al., Reference Tolin, Worhunsky and Maltby2006). For example, Exner et al. (Reference Exner, Zetsche, Lincoln and Rief2014) found that participants with OCD performed significantly more poorly on a prediction task compared to healthy controls due to heightened threat overestimations. The tendency for OCD patients to underestimate their coping ability has also been explored, with poorer estimates of coping being predictive of more severe OCD (Woods et al., Reference Woods, Frost and Steketee2002). Threat estimation beliefs have also been shown to correlate with OC symptoms in OCD samples (OCCWG, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Steketee et al., Reference Steketee, Frost and Cohen1998). These associations are also evident in studies utilising non-clinical samples showing a link between threat estimation and OC symptoms (Freeston et al., Reference Freeston, Ladouceur, Gagnon and Thibodeau1993; Moulding et al., Reference Moulding, Anglim, Nedeljkovic, Doron, Kyrios and Ayalon2010; Myers et al., Reference Myers, Fisher and Wells2008; Parrish & Radomsky, Reference Parrish and Radomsky2011; Tolin et al., Reference Tolin, Abramowitz, Brigidi and Foa2003). For example, Tolin et al. (Reference Tolin, Abramowitz, Brigidi and Foa2003) found that threat estimation significantly predicted OC symptoms in a non-clinical sample, including checking, hoarding, neutralising, obsessions, and washing. Similarly, Parrish and Radomsky (Reference Parrish and Radomsky2011) found that increases in perceived threat were associated with increased anxiety and compulsive urges in a series of experimental vignettes.

When comparing OCD patients and patients with other anxiety disorders on threat estimation beliefs, some studies report significant differences (OCCWG, 2001; Steketee et al., Reference Steketee, Frost and Cohen1998; Tolin et al., Reference Tolin, Worhunsky and Maltby2006). For example, the original validation studies of the Obsessive Beliefs Questionnaire found that OCD patients scored higher on threat estimation compared to anxious control participants (OCCWG, 2001). Similarly, Tolin et al. (Reference Tolin, Worhunsky and Maltby2006) found that OCD patients held stronger beliefs about threat estimation compared to anxious controls, although this did not hold after controlling for depression and trait anxiety. Other results suggest that threat estimation may not be specific to OCD, but instead may be characteristic of anxiety disorders in general (OCCWG, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Tolin et al., Reference Tolin, Abramowitz, Brigidi and Foa2003). For example, Tolin et al. (Reference Tolin, Abramowitz, Brigidi and Foa2003) found that in addition to significantly predicting OC symptoms, threat estimation showed a significant relationship with non-OCD psychopathology (e.g., social anxiety, depression).

Within the relevant literature, some research supports the idea that adults with OCD perceive threatening information with enhanced perceptions of probability and severity for harm occurring, as well as enhanced personal vulnerability and ‘looming’ vulnerability (belief that a dreaded threat is rapidly approaching). For example, one study found that while participants with subclinical OCD perceived higher probability of contamination, more imminent contamination, and lack of control over contamination compared to control participants, looming vulnerability was the only variable that independently contributed to OC symptoms (Riskind, Abreu, Strauss, & Holt, Reference Riskind, Abreu, Strauss and Holt1997). Results vary in other studies; for example, in a study assessing differences in threat estimations between OCD patients and students, increased severity estimation and decreased coping ability were correlated with OC symptoms for both groups, whereas probability estimation was only associated with OC symptoms in the student sample (Woods et al., Reference Woods, Frost and Steketee2002). The need to determine which components of threat estimation are most problematic in OCD has been pointed out (e.g., personal versus general incidence probability). To this end, studies have found that individuals with OCD overestimate the likelihood of harm occurring to them and experience less relief than individuals without OCD when given factual information about the unlikely occurrence of dangerous events, leading some authors to conclude that OCD patients overestimate their personal vulnerability rather than having a knowledge deficit relating to overestimations of overall probability of negative events (Moritz & Jelinek, Reference Moritz and Jelinek2009; Moritz & Pohl, Reference Moritz and Pohl2009; Niemeyer et al., Reference Niemeyer, Moritz and Pietrowsky2013).

In terms of experimental studies, there is a lack of research examining the role of threat biases in OCD. Perhaps the most relevant research is that of Jones and Menzies (Reference Jones and Menzies1997a) who explored the relationship between threat cognitions (among others) and overt displays of OCD washing behaviours during a behavioural avoidance test. Results found that of the cognitive domains assessed, danger expectancies were the most likely mediator of OCD washing behaviours, showing a ‘strong, reliable and robust’ relationship with both covert (i.e., urge to wash, distress) and overt (i.e., avoidance, washing) OC symptoms. The authors concluded that substantial support had been found for threat-based models of OCD, and failed to support models emphasising responsibility and perfectionism. These results were further supported by a more recent experimental study with non-clinical participants, which found that danger expectancy concerning severity of disease almost entirely mediated the relationship between disgust sensitivity and washing time (Thorpe, Barnett, Friend, & Nottingham, Reference Thorpe, Barnett, Friend and Nottingham2011). Such studies suggest an important role of threat as a maintaining factor in OCD symptoms, with potentially greater impact than that of responsibility. The role of threat was also evident in a recent systematic review on responsibility manipulations, which found a significant effect of the ‘responsibility’ manipulations on threat ratings of probability and severity of harm in the majority of studies (Mantz & Abbott, Reference Mantz and Abbott2017). Further, a number of variables (e.g., checking, hesitations) showed a similar or pronounced effect for conditions targeting responsibility combined with threat (Ladouceur et al., Reference Ladouceur, Rhéaume and Aublet1997; Mancini et al., Reference Mancini, D'Olimpio and Cieri2004), while others only reached significance (e.g., doubt, modifications) when threat was included in the manipulation (Ladouceur et al., Reference Ladouceur, Rhéaume and Aublet1997). The authors thus concluded that threat appraisals are a likely contributor to OCD symptomatology and recommended that further experimental studies manipulating danger expectancies be undertaken to better understand the role of this belief domain on OCD phenomena.

Perfectionism and intolerance of uncertainty

Defined by the OCCWG (2001, p. 1004), perfectionism refers to ‘the belief that there is a perfect solution to every problem, that doing something perfectly is possible and necessary, and that even minor mistakes will have serious consequences’. The idea that perfectionism works as a cognitive mediator of OCD symptomatology has been given substantial consideration in the literature (Frost, Novara, & Rhéaume, Reference Frost, Novara, Rhéaume, Frost and Steketee2002; Guidano, Liotti, & Guidano, Reference Guidano, Liotti and Guidano1983; Mallinger, Reference Mallinger1984; McFall & Wollersheim, Reference McFall and Wollersheim1979; OCCWG, 1997; Pacht, Reference Pacht1984; Rasmussen & Eisen, Reference Rasmussen and Eisen1989). This notion is based on the idea that perfectionistic individuals struggle to tolerate mistakes or imperfection, and as such, set unrealistic performance standards whereby they attempt to uphold perfect conditions and evade feelings of failure (Frost & Steketee, Reference Frost and Steketee1997; Jones & Menzies, Reference Jones and Menzies1997a). For example, the OCD sufferer may need to fill out a form ‘perfectly’ or repeat a routine action until it feels ‘just right’ (Leckman, Walker, Goodman, Pauls, & Cohen, Reference Leckman, Walker, Goodman, Pauls and Cohen1994). The OCCWG (1997) suggest that perfectionism often works in conjunction with intolerance of uncertainty (IU), whereby perfectionistic behaviours aim to increase certainty about future outcomes. IU refers to ‘beliefs about the necessity for being certain, that one has poor capacity to cope with unpredictable change, and that it is difficult to function adequately in ambiguous situations’ (OCCWG, 2001, p. 1004). While previous research suggests that IU might be more specific to worry and generalised anxiety disorder (GAD) than other anxiety disorders (Dugas, Buhr, & Ladouceur, Reference Dugas, Buhr, Ladouceur, Heimberg, Turk and Mennin2004), researchers have also contended that a low tolerance for uncertainty or ambiguity may be a central component in OCD symptomatology (Carr, Reference Carr1974; Rasmussen & Eisen, Reference Rasmussen and Eisen1989), whereby uncertainty is experienced as upsetting and stressful, leading to compulsions that work to eliminate uncertainty and reduce distress (Holaway, Heimberg, & Coles, Reference Holaway, Heimberg and Coles2006; Sarawgi, Oglesby, & Cougle, Reference Sarawgi, Oglesby and Cougle2013; Tolin, Abramowitz, Brigidi, & Foa, Reference Tolin, Abramowitz, Brigidi and Foa2003).

As with the belief domains already reviewed, results pertaining to perfectionism appear mixed and experimental studies are scarce. Studies report that individuals with OCD have increased perfectionism scores in comparison to non-clinical controls (Frost & Steketee, Reference Frost and Steketee1997; OCCWG, 2001, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Tolin et al., Reference Tolin, Worhunsky and Maltby2006). Perfectionism beliefs have also been shown to correlate with OC symptoms (Ferrari, Reference Ferrari1995; Frost, Steketee, Cohn, & Griess, Reference Frost, Steketee, Cohn and Griess1994; Myers et al., Reference Myers, Fisher and Wells2008; OCCWG, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Tolin et al., Reference Tolin, Abramowitz, Brigidi and Foa2003), even after controlling for responsibility and responsibility/threat beliefs (Rhéaume et al., Reference Rhéaume, Freeston, Dugas, Letarte and Ladouceur1995; Wu & Cortesi, Reference Wu and Cortesi2009). This has led authors to conclude that perfectionism might play an important yet underestimated role in OCD and that cognitive models allocating a central role to responsibility are only partially supported. Results, however, are mixed as to whether perfectionism beliefs differ between OCD and other anxiety disorders, with some studies suggesting that perfectionism is specific to OCD (Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004), while others report no differences between participant groups (Frost & Steketee, Reference Frost and Steketee1997; OCCWG, 2003) or differences that fail to hold after controlling for depression and trait anxiety (Tolin et al., Reference Tolin, Worhunsky and Maltby2006). It has also been suggested that perfectionism may partially mediate the relationship between responsibility and OCD symptoms (Bouchard, Rhéaume, & Ladouceur, Reference Bouchard, Rhéaume and Ladouceur1999; Pleva & Wade, Reference Pleva and Wade2006). In terms of OCD symptom subtypes, perfectionism beliefs have been linked with symmetry, ordering, and arranging symptoms (Myers et al., Reference Myers, Fisher and Wells2008; Summerfeldt, Reference Summerfeldt2004; Tolin et al., Reference Tolin, Abramowitz, Brigidi and Foa2003). Other, studies have reported a link between maladaptive perfectionism domains (e.g., concern over mistakes, doubts about actions), not just right experiences (feelings of incompleteness, conceptually related to perfectionism) and OC symptoms, including checking, ordering, and doubting (Coles, Frost, Heimberg, & Rhéaume, Reference Coles, Frost, Heimberg and Rhéaume2003; Fitch & Cougle, Reference Fitch and Cougle2013; Moretz & McKay, Reference Moretz and McKay2009).

As with the other belief domains, research on IU in adults is largely correlational. Studies report that individuals with OCD have increased IU scores in comparison to non-clinical controls (Freeston et al., Reference Freeston, Ladouceur, Gagnon and Thibodeau1993; Holaway et al., Reference Holaway, Heimberg and Coles2006; OCCWG, 2001, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Steketee et al., Reference Steketee, Frost and Cohen1998; Tolin et al., Reference Tolin, Worhunsky and Maltby2006). For example, Tolin et al. (Reference Tolin, Abramowitz, Brigidi and Foa2003) found that IU was greatest for OCD participants with compulsive checking compared to both non-checking OCD patients and non-anxious controls, who did not differ from each other. IU beliefs have also been shown to correlate with OC symptoms in clinical and non-clinical samples (Boelen & Reijntjes, Reference Boelen and Reijntjes2009; Calleo, Hart, Björgvinsson, & Stanley, Reference Calleo, Hart, Björgvinsson and Stanley2010; Dugas, Gosselin, & Ladouceur, Reference Dugas, Gosselin and Ladouceur2001; Fergus & Wu, Reference Fergus and Wu2010; Freeston et al., Reference Freeston, Ladouceur, Gagnon and Thibodeau1993; Holaway et al., Reference Holaway, Heimberg and Coles2006; OCCWG, 2003; Sarawgi et al., Reference Sarawgi, Oglesby and Cougle2013; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Sookman & Pinard, Reference Sookman, Pinard, Frost and Steketee2002; Steketee et al., Reference Steketee, Frost and Cohen1998). For example, Steketee et al. (Reference Steketee, Frost and Cohen1998) and Fergus and Wu (Reference Fergus and Wu2010) found that IU predicted OC symptoms beyond other belief domains (e.g., responsibility, threat estimation) after controlling for general distress. Further, Lind and Boschen (Reference Lind and Boschen2009) found that while both responsibility and IU belief domains were correlated with the incidence of compulsive checking in a mixed sample, the relationship between responsibility and checking was completely mediated by IU. More recently, Sarawgi et al. (Reference Sarawgi, Oglesby and Cougle2013) found that IU uniquely predicted OC symptoms, including ordering and arranging, checking, contamination fear, and obsession symptoms. This relationship held after controlling for depression and anxiety. Results are not always so consistent; for example, Tolin et al. (Reference Tolin, Abramowitz, Brigidi and Foa2003) found that tolerance for uncertainty failed to predict any OC subscales.

A substantial amount of research has investigated IU beliefs in both OCD and GAD samples. Some studies report significant differences (Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Steketee et al., Reference Steketee, Frost and Cohen1998), while others do not (Gentes & Ruscio, Reference Gentes and Ruscio2011; OCCWG, 2003). For example, Dugas et al. (Reference Dugas, Gosselin and Ladouceur2001) found that IU has a stronger role in GAD while Sica et al. (Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004) found that IU was one of two belief domains that best differentiated between OCD, GAD, and control samples, with stronger IU beliefs in OCD compared to the other groups. Other studies have found similar levels of IU in both GAD and OCD (Fergus & Wu, Reference Fergus and Wu2010; Holaway et al., Reference Holaway, Heimberg and Coles2006). In line with this, a meta-analysis assessing IU in GAD, OCD, and major depressive disorder (MDD) found that IU was significantly related to symptoms of each disorder and did not demonstrate narrow specificity (Gentes & Ruscio, Reference Gentes and Ruscio2011).

As with the other belief domains, experimental studies investigating the role of IU in producing compulsive behaviours are lacking. In one study, IU and perfectionism beliefs were the only significant predictor of urge to check and post-task anxiety, although the independent role of IU could not be assessed due to the use of a combined measure (Fitch & Cougle, Reference Fitch and Cougle2013). Further, a recent study incorporating OC-related in vivo tasks (e.g., stove checking, hand immersion in a contaminant) found that IU uniquely predicted greater anxiety and urge to check/wash, lower behavioural approach, washing duration and higher clutter-related distress, urge to straighten, and ordering and arranging duration (Sarawgi et al., Reference Sarawgi, Oglesby and Cougle2013). Experimental studies have also investigated repeated checking and its relationship to memory confidence and uncertainty in OCD. For example, it has been found that when healthy individuals engage in OCD-like repeated checking, confidence in memories related to the checking decreases (Ashbaugh & Radomsky, Reference Ashbaugh and Radomsky2007; Coles, Radomsky, & Horng, Reference Coles, Radomsky and Horng2006; van den Hout & Kindt, Reference van den Hout and Kindt2003) and uncertainty and doubt increases (van den Hout & Kindt, Reference van den Hout and Kindt2003). This counterproductive effect proves difficult for OCD sufferers who likely engage in compulsions, particularly checking, in an attempt to alleviate uncertainty related to obsessive beliefs (Tolin et al., Reference Tolin, Abramowitz, Brigidi, Amir, Street and Foa2001; van den Hout & Kindt, Reference van den Hout and Kindt2003).

Overimportance of thoughts and need to control thoughts

Overimportance of thoughts (OT) refers to ‘the belief that the mere presence of a thought indicates that it is important’ (OCCWG, 2001, p. 1003). Individuals with OCD often believe that merely having a thought implies that it is important or reflects their true desires (OCCWG, 1997; Rheaume, Ladouceur, & Freeston, Reference Rhéaume, Ladouceur and Freeston1997; Thordarson & Shafran, Reference Thordarson, Shafran, Frost and Steketee2002). This cognitive process has been linked with both thought-action fusion and magical thinking, of which the beliefs may reflect either (OCCWG, 2001). Thought-action fusion, for example, refers to the belief that thoughts are able to affect the external world (e.g., I am more likely to have a car accident if I think about it) or that thinking about doing something is morally equivalent to actually doing it (e.g., committing adultery) (Amir, Freshman, Ramsey, Neary, & Brigidi, Reference Amir, Freshman, Ramsey, Neary and Brigidi2001; Shafran et al., Reference Shafran, Thordarson and Rachman1996). Given this appraisal of intrusive thoughts being important and potentially harmful, the presence of such thoughts is viewed negatively and likely leads to a need to resist or control one's thoughts (Clark & Purdon, Reference Clark and Purdon1993; OCCWG, 1997). The need to control thoughts (NCT) refers to ‘the belief that it is important, possible, and desirable to exert complete control over intrusive thoughts, images, and impulses’ (OCCWG, 2001, p. 1003). Theorists have asserted that individuals with OCD believe it is necessary and possible to control one's thoughts and will often make attempts to control their thoughts through means such as thought suppression (Abramowitz, Whiteside, Kalsy, & Tolin, Reference Abramowitz, Whiteside, Kalsy and Tolin2003; Clark & Purdon, Reference Clark and Purdon1993; Moulding & Kyrios, Reference Moulding and Kyrios2006; Purdon & Clark, Reference Purdon, Clark, Frost and Steketee2002). However, individuals with OCD typically struggle to control their thoughts (Tolin, Abramowitz, Przeworski, & Foa, Reference Tolin, Abramowitz, Przeworski and Foa2002; Tolin, Hamlin, & Foa, Reference Tolin, Hamlin and Foa2002), thus resulting in enhanced anxiety and negative judgements (Tolin, Abramowitz, Hamlin, Foa, & Synodi, Reference Tolin, Abramowitz, Hamlin, Foa and Synodi2002).

As with the belief domains already reviewed, results pertaining to OT and NCT lack consistency, and experimental studies are limited. Studies have found that OT and/or NCT are elevated in individuals with OCD compared to non-clinical samples (Kumar et al., Reference Kumar, Sharma, Kandavel and Reddy2012; Mavrogiorgou et al., Reference Mavrogiorgou, Bethge, Luksnat, Nalato, Juckel and Brüne2016; OCCWG, 2001, 2003; Pazvantoğlu et al., Reference Pazvantoğlu, Algül, Ateş, Sarısoy, Ebrinc, Başoğlu and Cetin2013; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Tolin et al., Reference Tolin, Worhunsky and Maltby2006) and correlate with measures of OC symptoms (Cucchi et al., Reference Cucchi, Bottelli, Cavadini, Ricci, Conca, Ronchi and Smeraldi2012; Emmelkamp & Aardema, Reference Emmelkamp and Aardema1999; Gwilliam et al., Reference Gwilliam, Wells and Cartwright‐Hatton2004; Myers & Wells, Reference Myers and Wells2005; OCCWG, 2003; Tolin et al., Reference Tolin, Abramowitz, Brigidi and Foa2003). For example, Tolin et al. (Reference Tolin, Abramowitz, Brigidi and Foa2003) found that OT significantly predicted OC neutralising symptoms, and NCT significantly predicted OC obsessing symptoms in a non-clinical sample. In more recent studies, Cucchi et al. (Reference Cucchi, Bottelli, Cavadini, Ricci, Conca, Ronchi and Smeraldi2012) and Pazvantoglu et al. (Reference Pazvantoğlu, Algül, Ateş, Sarısoy, Ebrinc, Başoğlu and Cetin2013) found that beliefs about the uncontrollability and danger of thoughts and NCT were positively correlated with degree of doubt/indecisiveness in OCD patients. However, such results are not found consistently; for example, Sica et al. (Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004) found that OT and NCT were not significantly correlated with OC symptoms in an OCD sample.

When compared to other clinical groups, studies suggest that OT and/or NCT are more strongly endorsed by OCD patients compared to individuals with other anxiety disorders and/or depression (Morillo et al., Reference Morillo, Belloch and García-Soriano2007; OCCWG, 2003; Sica et al., Reference Sica, Coradeschi, Sanavio, Dorz, Manchisi and Novara2004; Tolin et al., Reference Tolin, Worhunsky and Maltby2006). For example, Morillo et al. (Reference Morillo, Belloch and García-Soriano2007) found that worries about thoughts coming true and the importance of controlling thoughts were specific to OCD patients when compared with depressed patients, anxious control patients, and non-clinical participants. Other results suggest that OT and NCT may be OCD-relevant but not OCD-specific. For example, Cucchi et al. (Reference Cucchi, Bottelli, Cavadini, Ricci, Conca, Ronchi and Smeraldi2012) found that NCT beliefs did not differ between OCD patients and those with panic disorder, although both scored significantly higher than healthy control participants. More recently, McKay et al. (Reference McKay, Kim, Taylor, Abramowitz, Tolin, Coles and Olatunji2014) conducted a large study comparing individuals with OCD, other anxiety disorders, and undergraduate students and found that while responsibility beliefs were prominent in all three groups, NCT was the only subscale unique to OCD, although it was unexpectedly lower than the other groups. Similarly, Tolin et al. (Reference Tolin, Worhunsky and Maltby2006) found that of the six belief domains, only NCT was able to differentiate between OCD and other anxiety disordered patients. It has thus been suggested that beliefs about controlling one's thoughts may have the clearest relationship with OCD compared to other anxiety disorders (McKay et al., Reference McKay, Kim, Taylor, Abramowitz, Tolin, Coles and Olatunji2014).

Inflated responsibility

Research assessing the role of responsibility in OCD among children and adolescents is sparse and based largely on observational studies, preventing a causative relationship from being established. For example, Magnúsdóttir and Smári (Reference Magnúsdóttir and Smári2004) investigated the correlation between responsibility attitudes and OC symptoms in a sample of non-clinical children and found that responsibility attitudes added significantly to the prediction of obsessive symptoms. Similarly, Matthews, Reynolds, and Derisley (Reference Matthews, Reynolds and Derisley2007) found that inflated responsibility beliefs were a significant independent predictor of OC symptoms in non-clinical adolescents, although when comparing cognitive appraisals in children with OCD, anxious children, and non-clinical controls, Barrett and Healy (Reference Barrett and Healy2003) found that children with OCD reported significantly higher ratings of responsibility than non-clinical, but not anxious children. In contrast, Libby, Reynolds, Derisley, and Clark (Reference Libby, Reynolds, Derisley and Clark2004) found that young people with OCD had significantly higher scores on inflated responsibility and thought-action fusion compared to young people with an anxiety disorder or no diagnosis. However, inflated responsibility was the only significant predictor of OC symptoms. Research also suggests that responsibility appraisals may be have an impact developmentally, given findings that children with OCD report significantly less inflated responsibility appraisals than adolescents and adults with OCD who do not differ from each other (Farrell & Barrett, Reference Farrell and Barrett2006; Pişgin & Özen, Reference Pişgin and Özen2010). Similarly, Farrell, Waters, and Zimmer-Gembeck (Reference Farrell, Waters and Zimmer-Gembeck2012) found that responsibility and metacognition were associated with OCD severity for adolescents but not for children. Interestingly, Lawrence and Williams (Reference Lawrence and Williams2011) found that adolescents with a history of OCD reported increased responsibility for significant negative incidents prior to OCD onset compared to adolescents with no history of OCD. Authors thus concluded that an inflated sense of responsibility, combined with particular events, may contribute to the development of OCD. To assess the role of responsibility in OCD further, a recent meta-analysis was conducted and found inconclusive support for the specificity of responsibility to OCD (Pozza & Detore, Reference Pozza and Dèttore2014a). Authors also reported that responsibility was more strongly related to OCD symptoms in adult than in child or adolescent samples, and recommended further research with experimental designs to better ascertain the role of responsibility beliefs in OCD symptom changes (Pozza & Dèttore, Reference Pozza and Dèttore2014a).

In terms of experimental research with an in vivo manipulation designed to manipulate levels of responsibility, only two studies have been conducted to investigate the role of inflated responsibility on OCD symptoms with a paediatric sample (i.e., Barrett & Healy-Farrell, Reference Barrett and Healy-Farrell2003; Reeves, Reynolds, Coker, & Wilson, Reference Reeves, Reynolds, Coker and Wilson2010). Results were mixed; for example, both studies found that ratings of perceived responsibility were significantly greater in the high compared to moderate and low responsibility conditions, thus supporting an impact of the experimental manipulation on responsibility. However, only one study found a significant increase in OCD-related behaviours (e.g., neutralising, hesitations, overall time taken) for the high compared to the low responsibility condition (Reeves et al., Reference Reeves, Reynolds, Coker and Wilson2010), while the other found no effect of inflated responsibility on avoidance behaviours or ritualising (Barrett & Healy-Farrell, Reference Barrett and Healy-Farrell2003). Based on the latter study, the authors concluded that inflated responsibility may not be an important cognitive process in childhood OCD (Barrett & Healy-Farrell, Reference Barrett and Healy-Farrell2003). Similarly, a more recent study on young people with OCD using an experimental cognitive appraisal task with vignettes to manipulate perceived responsibility found that responsibility was not related to urge to check. Authors concluded that responsibility biases may not be central to childhood OCD (Farrell, Waters, Boschen, & Milliner, Reference Farrell, Waters, Boschen and Milliner2011). Thus, the role of responsibility in child and adolescent OCD requires further investigation.

Overestimation of threat

A number of studies have assessed threat cognitions among anxious children, with studies showing that clinically anxious children interpret ambiguous scenarios to be more threatening than non-clinical children (e.g., Barrett, Rapee, Dadds, & Ryan, Reference Barrett, Rapee, Dadds and Ryan1996; Bögels & Zigterman, Reference Bögels and Zigterman2000; Creswell, Schniering, & Rapee, Reference Creswell, Schniering and Rapee2005; Muris, Rapee, Meesters, Schouten, & Geers, Reference Muris, Rapee, Meesters, Schouten and Geers2003; Suarez & Bell-Dolan, Reference Suarez and Bell-Dolan2001; Waters, Craske, Bergman, & Treanor, Reference Waters, Craske, Bergman and Treanor2008). In terms of young people with OCD, there is relatively less research examining threat cognitions in comparison to other belief domains. This was evident in a systematic review of cognitive processes in paediatric OCD, which gave little consideration to cognitions other than responsibility, thought-action fusion, and metacognitions (Reynolds & Reeves, Reference Reynolds and Reeves2008). Of these, some studies have reported a significant difference between OCD patients and normal controls. For example, when compared with other groups, Barrett and Healy (Reference Barrett and Healy2003) found that children with OCD reported higher severity, but not probability, of threat in comparison to non-clinical controls, although there was not a clear differentiation between OCD and anxious children. Similarly, Farrell, Hourigan, Waters, and Harrington (Reference Farrell, Hourigan, Waters and Harrington2015) found that children with OCD perceived ambiguous scenarios as being more difficult to manage compared to non-clinical children, although these differences did not hold after self-reported anxiety symptoms were controlled, suggesting that interpretation biases may be more closely related to anxiety symptoms than OCD. Further, cognitive responses concerning threat-based thoughts did not show an inflated threat interpretation bias, inconsistent with paediatric anxiety research showing enhanced threat-based thinking (e.g., Creswell et al., Reference Creswell, Schniering and Rapee2005; Waters et al., Reference Waters, Craske, Bergman and Treanor2008). The authors suggested that this may be related to the developmental stage of the included sample, or that threat cognitions and coping appraisals may be specific to OCD-related situations. In line with this, some studies suggest that threat estimations may be affected developmentally, with less prominence in children compared to adults (Pişgin & Özen, Reference Pişgin and Özen2010). For example, one study found that children with OCD showed lower probability of threat (but similar perceived severity of harm) compared to adolescents and adults with OCD (Farrell & Barrett, Reference Farrell and Barrett2006). Another study found that negative cognitions concerning physical threat, social threat, and personal failure were related to severity of obsessions, but not compulsions, in OCD adolescents aged 13–18 (Verhaak & De Haan, Reference Verhaak and De Haan2007). However, no significant correlations were found for a group of OCD children aged 8–12 years, supporting a developmental difference in the belief domains (Verhaak & De Haan, Reference Verhaak and De Haan2007). Authors have thus recommended that cognitive models of OCD consider developmental differences in the cognitive processing of threat, and further examine the occurrence and nature of these biases in children, adolescents and adults (Farrell & Barrett, Reference Farrell and Barrett2006; Farrell et al., Reference Farrell, Hourigan, Waters and Harrington2015).

When considering experimental studies examining overestimation of threat in children and adolescents, research is scarce and inconclusive. Two experimental studies measuring probability and severity of harm during a ‘responsibility’ manipulation were identified in a recent systematic review (Mantz & Abbott, Reference Mantz and Abbott2017). One study reported no significant difference for probability or severity between responsibility conditions in an OCD sample (Barrett & Healy-Farrell, Reference Barrett and Healy-Farrell2003), while the other study failed to report results pertaining to threat appraisals (Reeves et al., Reference Reeves, Reynolds, Coker and Wilson2010). Further experimental studies are thus warranted to better ascertain the role of threat cognitions in the onset and maintenance of paediatric OCD. This will ultimately prove useful in determining therapeutic targets requiring emphasis in treatment.

Perfectionism and intolerance of uncertainty (IU)

Although limited in number, some research support is evident for the role of these belief domains in child and adolescent OCD. One study found that young people with OCD had significantly higher ratings of perfectionism beliefs (i.e., concern over mistakes) compared to both anxious and non-clinical controls (Libby et al., Reference Libby, Reynolds, Derisley and Clark2004). A more recent study, however, found that perfectionism and certainty beliefs in an unselected sample of adolescents were similar to levels found in previous studies of young people with an OCD diagnosis (Coles, Ravid, Franklin, Storch, & Khanna, Reference Coles, Ravid, Franklin, Storch and Khanna2014). Perfectionism beliefs have also been shown to correlate with OC symptoms (Coles et al., Reference Coles, Ravid, Franklin, Storch and Khanna2014; Huan, Rice, & Storch, Reference Huan, Rice and Storch2008; Soreni et al., Reference Soreni, Streiner, McCabe, Bullard, Swinson, Greco and Szatmari2014). For instance, Soreni et al. (Reference Soreni, Streiner, McCabe, Bullard, Swinson, Greco and Szatmari2014) found that perfectionism beliefs were associated with severity of OCD and depressive symptoms in children and adolescents with OCD. In terms of development, it has been suggested that perfectionism and IU beliefs are not limited to adulthood but also observed during childhood and adolescence (Pişgin & Özen, Reference Pişgin and Özen2010).

Overimportance of thoughts and need to control thoughts

Studies have found that young people with OCD have significantly higher ratings of thought-action fusion beliefs and less cognitive control compared to anxious and/or non-clinical controls (Barrett & Healy, Reference Barrett and Healy2003; Kadak, Balsak, Besiroglu, & Çelik, Reference Kadak, Balsak, Besiroglu and Çelik2014; Libby et al., Reference Libby, Reynolds, Derisley and Clark2004). Beliefs related to OT (i.e., thought-action fusion, magical thinking) have also been shown to correlate with OC symptoms in non-clinical samples (Bolton, Dearsley, Madronal‐Luque, & Baron‐Cohen, Reference Bolton, Dearsley, Madronal‐Luque and Baron‐Cohen2002; Evans, Milanak, Medeiros, & Ross, Reference Evans, Milanak, Medeiros and Ross2002; Matthews et al., Reference Matthews, Reynolds and Derisley2007; Muris, Meesters, Rassin, Merckelbach, & Campbell, Reference Muris, Meesters, Rassin, Merckelbach and Campbell2001). For example, Evans et al. (Reference Evans, Milanak, Medeiros and Ross2002) found that magical thinking beliefs were significantly correlated with rituals and compulsive behaviours in non-clinical children. However, Farrell et al. (Reference Farrell, Waters and Zimmer-Gembeck2012) and Verhaak and De Haan (Reference Verhaak and De Haan2007) found no significant relationship between magical thinking/thought-action fusion beliefs and OCD severity in a sample of children and adolescents with OCD, although thought suppression and metacognition (including NCT beliefs, among others) were strongly correlated with OCD severity in adolescents (Farrell et al., Reference Farrell, Waters and Zimmer-Gembeck2012), indicating the need for further research on these constructs. In terms of development, it has been suggested that OCD-type cognitions related to overimportance of thoughts (Pişgin & Özen, Reference Pişgin and Özen2010), thought-action fusion, and controllability of thoughts are comparable across age groups (Farrell & Barrett, Reference Farrell and Barrett2006), although concerns about controlling one's thoughts may be more pronounced in adulthood (Pişgin & Özen, Reference Pişgin and Özen2010).