A hierarchical uncertainty framework of anxiety: Preregistered meta-analytic structural model of intolerance of uncertainty and worry across anxiety disorders.

Intolerance of uncertainty (IU) has been widely conceptualized as a transdiagnostic vulnerability factor for worry and anxiety; however, the extent to which the associations among IU, worry, and anxiety outcomes vary across clinical phenotypes and broader contextual and methodological conditions has not yet been fully integrated within a unified framework. This pre-registered meta-analysis tested a structural model of IU → worry → anxiety while also examining the moderating role of key clinical, developmental, and measurement-related factors, synthesizing 738 effect sizes from 115 studies/dissertations comprising 138 independent samples (N = 28,693). Beyond confirming robust associations between IU-worry (r = .59), the analysis reveals a systematic hierarchical pattern in IU-anxiety relationships across disorders. IU showed strongest associations with GAD (r = .54) and non-specific anxiety (r = .48), followed by PTSD (r = .46), social anxiety disorder (r = .45) and OCD (r = .41), and comparatively weaker links with panic disorder (r = .27). This gradient provides the first quantitative evidence that the transdiagnostic relevance of IU varies in a theoretically coherent hierarchy across anxiety outcomes. Worry demonstrated even stronger associations, particularly with GAD (r = .67), and remained significantly linked to anxiety after accounting for IU. Meta-analytic structural equation modeling confirmed worry as a partial mediator of the IU-anxiety relationship, with the largest indirect effects for GAD (r = .32), followed by non-specific anxiety (r = .22), social anxiety disorder (r = .17), panic disorder (r = .15), PTSD (r = .14), and the smallest effect observed for OCD (r = .13). Moderator analyses were significant and revealed stronger effects for IU-GAD, IU-non-specific anxiety, and worry-non-specific anxiety when assessed with overlapping self-report measures (p < .05), indicating greater construct proximity. In contrast, IU-OCD showed weaker effects under overlapping measurement conditions. IU-anxiety associations were weaker in the presence of depressive symptoms and varied across anxiety phenotypes. Overall, findings characterize IU as a distal factor and worry as a proximal cognitive process, refining transdiagnostic models through a hierarchical uncertainty-related architecture and complementary intervention targets.