Evaluation of hierarchical model of trait and disorder- specic intolerance of uncertainty in anxiety disorder symptoms

Background: Research suggests that Intolerance of Uncertainty is a transdiagnostic risk and maintaining factor in anxiety disorders. The aim of this study was to evaluate the hierarchical model of trait Intolerance of Uncertainty and disorder-specic Intolerance of Uncertainty in symptoms of anxiety disorders in Iranian university students. Methods: Participants (N = 561) completed a battery of questionnaires which assess intolerance of uncertainty, disorder-specic intolerance of uncertainty, metacognition beliefs, fear of negative evaluation, obsessive cognitions, agoraphobic cognitions, obsessive compulsive symptoms, generalized anxiety disorder symptoms, social phobia symptoms, and panic disorder symptoms. Results: Results showed hierarchical model of trait Intolerance of Uncertainty and disorder-specic Intolerance of Uncertainty t with data. Trait intolerance of uncertainty and disorder-specic intolerance of uncertainty had signicant direct and indirect association with various cognitive vulnerabilities and symptoms of anxiety disorders. Conclusion: The ndings showed trait intolerance of uncertainty and disorder-specic intolerance of uncertainty have signicant contribution to various anxiety disorders. These ndings have important implications in conceptualization, prevention and treatment of anxiety disorders. Questionnaire (MCQ-30) (31): The MCQ-30 was used to measure metacognitive beliefs of participants about their worry and thoughts. This scale consists of ve subscales; positive beliefs about worry, negative metacognitions about the uncontrollability and danger of worry, cognitive condence, need to control thoughts, and cognitive self-consciousness (31). Responses were rated on a four-point Likert scale from do not agree (1) to agree very much (4). The alpha coecients for Iranian sample for total scale was .91 and for the subscales were ranged from .71 to .87 (32). In the current study, we used negative metacognitions about the uncontrollability and danger of worry subscales. Internal consistency of the subscale was .87. and validity (45). In the current study, internal of the scale was .93. study evaluated validity of a hierarchical model of trait intolerance of uncertainty and disorder-specic intolerance of uncertainty (18) in Iranian university students, after controlling cognitive vulnerabilities. Results model Iranian students. This nding is in


Introduction
A number of cognitive behavioral models have been proposed to explain the etiology of anxiety disorders (1). New conceptualizations emphasize of shared risk factors in development of anxiety disorders (2). According to the triple vulnerability model of Barlow (2), there are three sets of factors, including general biological vulnerability, general psychological vulnerability, and speci c psychological vulnerability interacting with each other to form and develop some speci c emotional disorders. Research has also suggested that some of the vulnerability factors are related to various disorders and are therefore transdiagnostic (3).
In recent studies, increasing attention has been paid to intolerance of uncertainty (IU) as a general psychological risk factor to explain development and maintenance of anxiety disorders (4,5). IU is considered as negative thoughts about uncertainty and involves negative reactions to uncertain situations and events" (6). IU is driven from unknown and perception of uncontrollability of emotions and events (7). Research has recognized IU as a two dimensional concept: prospective anxiety and inhibitory anxiety (8). Initially, it was assumed that IU had a speci c relationship with generalized anxiety disorder (GAD) and worry (9). However, recent research supported the transdiagnostic role of IU in social anxiety (10), post-traumatic stress disorder (PTSD) (11), obsessive-compulsive disorder (OCD) (12), and panic disorder (PD) (13). Furthermore, IU may lead to anxiety and safety behaviors through overestimating the possibility of threat (14). Additionally, high-level perception of IU and interpreting it as threat may lead to avoidant behaviors (15). Furthermore, some dysfunctional cognitions (e.g. worry, obsessional doubt, cognitive distortions) and behaviors (e.g. compulsions) may be an effort to decrease uncertainty and gain control (16). Therefore, IU is considered as an important transdiagnostic variable in anxiety disorders (17).
Although previous research emphasized on the importance of trait IU (7), recent evidence suggests that trait IU (i.e. general experience of uncertainty) and disorder-speci c IU (DSIU) (i.e. speci c focus on uncertainty in anxiety disorders) may be different (16). Existing evidence proposed that although individuals can tolerate uncertainty in some situations, however, they may cannot face with the uncertainty related to some speci c personally distressing situations (14). In other words, uncertainty about the likelihood of making harm in obsessive-compulsive disorder may differ from uncertainty about social evaluative cues in social anxiety disorder (14). Therefore, distinguishing between role of trait IU and disorder-speci c IU in anxiety disorders is important.
In an attempt to respond to this neccesity, Shihata et al. (18) proposed a hierarchical model of trait IU and disorder-speci c IU in anxiety disorder symptoms. The proposed model examined relationship between trait IU as a higher-order factor and disorder-speci c IU as intermediate factor, and disorder symptomology (symptoms of generalized anxiety disorder, social anxiety disorder, obsessive-compulsive disorder, and panic disorder). Shihata et al. (18) considered an established cognitive vulnerability as a mediator in each disorder symptoms: negative metacognition in generalized anxiety disorder (19); fear of negative evaluation in social anxiety disorder (20); in ated responsibility in obsessive-compulsive disorder (21); and agoraphobic cognition in panic disorder (22). They found that the proposed model t with the data of the students. The results showed that disorder speci c IU along with trait IU explained a signi cant variance of anxiety disorders symptoms.
However, as far as we know, there is no published study which evaluate structural validity of the Shihata et al. (18) model. Given that evaluating the existence of cultural differences in dealing with depression and anxiety is generally encouraged (23), we tried to evaluate this model in Iranian population. Studies showed that anxiety disorders are the most prevalent group of psychiatric disorders in Iran (15.6%) (24). In addition, in light of disagreements on the speci c protocols for each disorder, and cost of protocols (25), development and evaluation of integrated explanatory models is a clinical and research necessity. Therefore, further investigating the transdiagnostic role of IU in anxiety disorders can help to develop and modify therapeutic protocols of these disorders in Iranian culture.
Thus, aim of this study was to evaluate structural validity of hierarchical model of trait IU and disorder-speci c IU in symptoms of generalized anxiety disorder, social anxiety disorder, obsessive-compulsive disorder, and panic disorder in Iranian students. Our hypotheses were: 1) trait IU and disorder-speci c IU would signi cantly predict negative metacognition and symptoms of generalized anxiety disorder, 2) trait IU and disorder-speci c IU would signi cantly predict fear of negative evaluation and symptoms of social anxiety disorder, 3) trait IU and disorder-speci c IU would signi cantly predict in ated responsibility and symptoms of obsessive-compulsive disorder, 4) trait IU and disorder-speci c IU would signi cantly predict agoraphobic cognition and symptoms of panic disorder, 5) each of the disorder-speci c IU and cognitive vulnerabilities mediate relationship of trait IU and disorder-speci c symptoms.

Participants
The participants were 614 university students from universities of Tehran, Iran (university of Tehran, Shahid Beheshti University, and Alzahra University). 32 students missed more than 10% of the items and their data was excluded from analyses. 21 students were excluded from analyses based on outliers from the box diagram. Thus, the data of 561 subjects (265 males, 296 females) were analyzed. The mean age of the males was 23.39 (SD = 7.38) and females was 21.72 (SD = 5.72). Other demographic information is presented in Table 1. Inclusion criteria were being over 18 years old.  type scale from disagree very much (-3) to neutral (0) to agree very much (+ 3). This scale has good internal consistency and criterion validity in clinical and non-clinical samples (38). In Iranian population, it showed good test-retest and internal consistency reliability and discriminant validity (39). In the current study we used OBQ-RT subscale Internal consistency of the subscale was .88.
Agoraphobic Cognitions Questionnaire (ACQ) (40): This 14-item scale assesses thoughts concerning negative consequences of experiencing anxiety and was designed to assess aspects of fear (panic attacks) in agoraphobics. The ACQ comprises two 7-item subscales: physical concerns and social/behavioral concerns. Responses were rated on a 5-point Likert scale from thought never occurs (1) to thought always occurs (5). Researches have shown good discriminant and construct validity and reliability of the scale (40). Psychometric research showed good sensitivity and validity in Iranian population (41). In the current study, internal conssitency of the scale was .86.
Generalized Anxiety Disorder-7 (GAD-7) (42): The 7-item GAD-7 was designed to identify probable cases of generalized anxiety disorder. Individuals were asked "how often, during the last 2 weeks, they were bothered by such symptoms." Responses were rated on a 4-point Likert scale from not at all (0) to nearly every day (4). Psychometric research in Iranian population showed good reliability and validity (43

Procedure
Subjects were selected from the city of Tehran's student population using convenience sampling. After coordinating with the universities' educational departments and the instructors, the second researcher (FM) referred to the classes and explained purpose and procedure of the study. participants who agreed to sign a written informed consent were asked to complete the questionnaires. Ethical review board of University of Social Welfare and Rehabilitation Sciences approved the study procedure (Code of Ethics: 1397.144). All methods were carried out in accordance of the institutional guidelines and conforming to the ethical standards of the declaration of Helsinki. There were no rewards/incentives for completing the survey.

Data analysis
We used SPSS 25.0 to calculate descriptive statistics and correlation matrix between measured variables. Correlation matrix were examined by calculating Pearson correlation coe cient. The correlation matrix between the observed variables showed no multiple linearity between them (49).
The measurement models were evaluated by con rmatory factor analysis (CFA) and hypothesized model was assessed by SEM with maximum likelihood estimation conducted in Mplus 7.4. For evaluation of measurement and structural model, t indices, factor loadings, and modi cation indices were considered. Model t indices included the x 2 /df ratio, Tucker-Lewis index (TLI), comparative t index (CFI), root mean square error of approximation (RMSEA), and standard root mean square residual (SRMR). For evaluating the strength of structural pathways, standardized estimates were utilized. Bootstrapping with at least 1000 repeated samples was used for estimating the strength of the total and speci c indirect effects.

Correlation between variables
The bivariate correlations showed moderate to large signi cant associations between trait IU, all disorder-speci c IU subscales, cognitive vulnerabilities, and disorder symptoms. Descriptive statistics indices and correlations of research variables are reported in Table 2.

Measurement models
In order to test the dimensionality of each subscale, independent CFA was performed to examine the measurement model of each variable employed in structural model. Modi cation indices and error covariance were performed for models with poor t. The factor loadings of measurement models were signi cant and ranged from .63 to .98. All variables showed a unidimensional construct.

Structural model
The goodness of t indices con rmed that the Shihata et al.'s model (2017) on uncertainty provided an acceptable t with the data (χ2 /df = 2.75, CFI = .92, TLI = .91, SRMR = .06, RMSEA = .05. The structural model with standardized parameter is displayed in Fig. 1.

Disorder
Disorder-speci c IU Cognitive Vulnerability Symptoms Trait intolerance of uncertainty showed strong association with IU-GAD and negative metacognition. This nding was consistent with previous research representing the association between intolerance of uncertainty and other vulnerabilities (4). Contrary to our hypothesis, the direct effect between trait intolerance of uncertainty and generalized anxiety disorder symptoms was not signi cant, when IU-GAD and negative metacognition were considered in the path. This is not consistent with research indicating direct relationship between trait intolerance of uncertainty and generalized anxiety disorder symptoms (16). However, it should be considered that previous research studied only trait intolerance of uncertainty, not disorder-speci c intolerance of uncertainty. Another potential explanation could be that the measure of disorder-speci c IU-GAD assesses uncertainty widely (e.g. uncertainty about everything) and thus, this is similar to the trait intolerance of uncertainty (18). It seems that negative metacognition play role as a mediator between trait intolerance of uncertainty and generalized anxiety disorder symptoms. In other words, when negative metacognition was eliminated from the model, trait intolerance of uncertainty showed signi cant association with generalized anxiety disorder symptoms. According to these ndings, it might be suggested that trait intolerance of uncertainty leads to symptoms of generalized anxiety disorder via metacognitive beliefs.
Our results suggest that there is a positive and signi cant association between trait intolerance of uncertainty and disorderspeci c IU-SAD, fear of negative evaluation, and social anxiety disorder symptoms. This is consistent with previous studies suggesting intolerance of uncertainty is an important factor in development and maintenance of social anxiety disorder symptoms (5,18). The ndings of the model on social anxiety disorder showed that the direct path of trait intolerance of uncertainty to social anxiety disorder symptoms was signi cant, but when the disorder speci c intolerance of uncertainty was included in the path as mediator variable, the statistical power of the path increased. This nding is in line with the research of Shihata et al. (18). Similarly, Boelen & Reijntjes (50) showed that trait intolerance of uncertainty predict social anxiety disorder symptoms after negative evaluation was controlled. The ndings of this study may indicate that social anxiety symptoms are associated with IU-SAD rather than trait intolerance of uncertainty. Therefore, individuals with social anxiety disorder symptoms may bene t from treatments which focused on intolerance of uncertainty in social situations.
The results showed that there is a signi cant relationship between trait intolerance of uncertainty and IU-OCD, in ated responsibility, and obsessive-compulsive disorder symptoms. The trait intolerance of uncertainty was signi cantly correlated with symptoms of obsessive-compulsive disorder, which was consistent with previous studies showing that intolerance of uncertainty was an important factor in obsessive-compulsive disorder symptoms (18).This nding probably indicates that people with symptoms of obsessive-compulsive disorder are unable to adapt to unpredictable changes and need to assure that unpleasant changes will not occur. In line with this hypothesis, Buhr and Dugas (6) showed that doubts and checking behaviors in people with obsessive-compulsive disorder symptoms are most associated with intolerance of uncertainty.
However, contrary to the research hypothesis and previous research ndings (18), the relationship between IU-OCD and obsessive-compulsive disorder symptoms was not signi cant. That is, IU-OCD did not lead to the obsessive-compulsive disorder symptoms. This nding may indicate that obsessive-compulsive disorder symptoms are more related to trait intolerance of uncertainty rather than IU-OCD. In addition, contrary to our hypothesis, the indirect pathway between trait intolerance of uncertainty and symptoms of obsessive-compulsive disorder was not signi cant through in ated responsibility. This nding was in line with the research of Shihata et al. (18). Also, Myers et al. (51) represented that in ated responsibility did not predict obsessive compulsive disorder symptoms. This nding was inconsistent with previous research showing the relationship between in ated responsibility and obsessive-compulsive disorder symptoms (18). The nding may suggest that people with symptoms of obsessive-compulsive who are not able to tolerate general uncertain situations, may represent obsessive-compulsive disorder symptoms, even without in ated responsibility. Therefore, intolerance of uncertainty may be more important in obsessive-compulsive disorder symptoms.
The ndings of this study showed that trait intolerance of uncertainty lead to symptoms of panic disorder through two pathways. The rst pathway was mediated by IU-PD, which was the strongest pathway for predicting symptoms of panic disorder and consistent with prior research (18). It can be said that uncertainty about the likelihood of a panic attack in the future eventuates attention to physical symptoms (such as a gradual increase in heart rate) and worries, and one may develop maladaptive avoidance behaviors that reduce uncertainty (52). symptoms. This nding was in contradiction with evidence showing that trait intolerance of uncertainty has a signi cant direct relationship with panic disorder symptoms (16, 52). It is worth mentioning that these studies only examined trait intolerance of uncertainty, and IU-PD had not been taken into account. However, the current nding was in line with some previous research (17,30), which exhibited that IU-PD compared to trait intolerance of uncertainty has a greater effect on panic disorder symptoms. In general, it can be said that a cognitive core such as agoraphobic cognitions as well as IU-PD about the potential catastrophic consequences of somatic and physical symptoms may be the maintaining factors of panic disorder symptoms.

Conclusion
The ndings suggest that individuals with anxiety disorders tend to have threatening interpretations of uncertain information, leading to higher levels of worry. In general, the results revealed that in social anxiety disorder symptoms and panic disorder symptoms, the disorder-speci c intolerance of uncertainty explained more variance of the symptoms.
However, in obsessive-compulsive disorder symptoms and generalized anxiety disorder symptoms, the trait intolerance of uncertainty explained more variance of the disorder symptoms, indicating that general states of uncertainty may be involved in the development and maintenance of symptoms of the obsessive-compulsive disorder and generalized anxiety disorder.
Therefore, the ndings of this study were in synchrony with a large number of studies identifying intolerance of uncertainty as a metacognition and general vulnerability to anxiety (4).
Interpretation and generalization of the ndings of the present study should be done according to its limitations. First, the present study was conducted using a cross-sectional approach, which prevents the inference of causal relationships between variables. Therefore, future studies can examine the relationships between the variables studied in the present study through prospective longitudinal, and experimental studies. Second, in the present study, self-report tools were used to measure variables. This may put the ndings at risk of bias. For this reason, prospective studies are needed to investigate more precisely the role of Intolerance of Uncertainty in anxiety disorders using multiple assessment tools.

Consent for publication
Identi able demographic information has been removed from this manuscript to ensure anonymity. Thus, the consent to publish is not applicable.
Availability of the data University of Social Welfare and Rehabilitation Sciences has approved and supported that only researchers of the manuscript will have access to the dataset, so the data used in this study is not available for public view. Still, requests can be written o cially to the Farhad Taremian, Email: fa.taremian@uswr.ac.ir., Tel: (98) 9121451697.

Competing Interest
The authors have no actual or potential con icts of interest including any nancial, personal or other relationships with other people or organizations within three years of beginning the work submitted that could inappropriately in uence their work.

Funding
University of Social Welfare and Rehabilitation Sciences nancially supported this research. However, the university had no role in designing, gathering and analyzing the data, and preparing the manuscript.
Authors' Contribution RM, FT designed and supervised the research. FM and HP conducted the study. Also, RM analyzed the data and wrote the manuscript. All authors have read and approved the manuscript.