Group Transdiagnostic Cognitive Behavioural Therapy for Intolerance of Uncertainty Versus Applied Relaxation

Rationale. Intolerance of uncertainty (IU) is a recognised transdiagnostic risk factor for a broad range of mental health problems, including generalised anxiety, social anxiety, panic, depression, and obsessive-compulsive symptoms. Cognitive Behavioural Therapy for Intolerance of Uncertainty (CBT-IU) has demonstrated efficacy in previous randomised trials. However, it was primarily developed and tested in the context of generalised anxiety disorder and further inquiry into its effectiveness across diagnoses is required. To that end, a treatment that targets a clinical sample characterised by high IU rather than by a specific diagnosis would be more readily deployed than one tied to a single disorder. A second gap concerns how therapy changes uncertainty processing, and whether such changes drive clinical improvement. Existing interventions and trials rely heavily on self-report scales such as the Intolerance of Uncertainty Scale (IUS), which are informative about dispositional reactivity to uncertainty but are largely agnostic to how individuals represent different forms of uncertainty or why uncertainty is experienced as aversive. Recent work in computational psychiatry reframes IU not as a unitary disposition but as a set of individual differences in how people estimate distinct forms of uncertainty, the probability of outcomes, the volatility of the environment. Recent work also shows that experiences of IU may depend on the contextual factors that govern a situation. This suggests that distinct profiles of uncertainty processing may characterise different presentations of anxiety, and that identifying which of these profiles shift in response to IU-focused intervention is essential for refining both the construct of IU and the treatments that target it. Accordingly, the trial combines established IU questionnaires with a computational task battery designed to estimate distinct forms of uncertainty across contexts. The primary hypothesis of this study is that both therapies will reduce anxiety, with CBT-IU producing greater reduction in self-reported IU than applied relaxation. Applied Relaxation will serve as an active comparator matched on group format, duration, and therapist contact but which does not directly target uncertainty. To gain additional insights into the mechanisms underlying changes in IU, the computational battery will be used to construct an individualised "uncertainty fingerprint". Due to the novelty of this approach, these analyses will be exploratory, where the investigators will examine whether components of the fingerprint change differentially between arms, and whether such change relates to symptom reduction at 3-month follow-up. Design. This is a single-blind, parallel-group mechanistic randomised controlled trial conducted at the MRC Cognition and Brain Sciences Unit (MRC-CBU), University of Cambridge, UK. The trial compares two active psychological treatments on anxiety and self-reported IU, with an embedded mechanistic component examining change in uncertainty processing. Adults scoring above threshold for IU and anxiety symptoms will be randomised to one of two 12-session group-based interventions running over 14 weeks: (1) Group Transdiagnostic CBT-IU, drawing on the Dugas and Robichaud model, and Making Friends with Uncertainty by Mofrad and colleagues; or (2) Applied Relaxation. Recruitment is transdiagnostic and no specific diagnostic category is required, consistent with the scope of the treatment manual and with IU being a transdiagnostic construct. Both interventions comprise ten weekly plus two fortnightly 2-hour sessions, delivered in groups of 6-9 participants and co-facilitated by a clinical psychologist and an assistant. Sample size was determined via a Bayesian hierarchical mediation simulation fitted with JAGS, using 1,000 replications per candidate design, to reflect the planned analysis and data structure: participants nested within therapy groups of 6-9, with the mediated effect a × b (with the IUS being the mediator) estimated alongside between-group variance under weakly informative priors. Effect size estimations were drawn from the Miller and McGuire (2023) meta-analysis. The trial is powered for a scenario with a = 0.89 (meta-analytic mean for the effect of CBT-IU on intolerance of uncertainty), b = 0.60 (the mediator-outcome path), and a between-group SD of 0.2, deliberately larger than the empirical SDs typically reported in the literature to provide a conservative scenario. Two criteria informed the target: power, defined as the proportion of replications in which the 95% highest-density interval (HDI) for the mediation effect excluded zero, and the width of that HDI as an index of estimation precision. Power exceeds 80% at 14 groups and nears 100% from 16 groups onward; at 16 groups the 95% HDI width is \~0.77 and stabilises thereafter. Sixteen groups would therefore satisfy both criteria, but a target of 18 therapy groups (9 per arm; approximately N = 135 participants) was selected to account for attrition. Outcomes. The co-primary outcomes are the IUS-12 and the Generalized Anxiety Disorder-7 (GAD-7). Exploratory measures comprise a battery of computational tasks indexing learning and decision-making under uncertainty across cognitive contexts, from which the per-participant uncertainty fingerprint is derived. Secondary outcomes include the Penn State Worry Questionnaire (PSWQ) and the Patient Health Questionnaire (PHQ-9), the Work and Social Adjustment Scale (WSAS), and the Affective Response to Daily Uncertainty Scale (ARDUS), alongside measures of putative maintaining processes including worry behaviours, rumination, emotion regulation difficulties, emotional reactivity, fear of negative evaluation, and beliefs about losing control. Unless otherwise specified, all measures are administered at baseline, post-treatment (week 14), and 3-month follow-up. Treatment credibility and expectancy, working alliance, and group climate are collected as process measures, and the CORE-10 is administered weekly for monitoring. In addition, a lifetime history of potentially traumatic experiences (TEC) will be assessed at baseline as a candidate moderator of treatment response. Randomisation and blinding. Stratified minimisation based on baseline GAD-7 and IUS-12 severity scores and demographics, executed by the trial statistician who is blind to study objectives. Outcome assessors at baseline, post-treatment, and follow-up are blind to treatment allocation.