A Systematic Review and Meta-Analytic Gaussian Network Aggregation of Anxious Symptoms

Aims: Numerous studies have modelled symptoms of Generalised Anxiety Disorder (GAD) as interacting nodes within complex systems, and sought to identify the most influential symptoms and mechanisms of anxious symptoms and disorders. However, meta-analytic evidence is lacking. This review and meta-analysis pooled data to investigate which features are most influential within network models of anxiety, with a particular focus on worry as a mechanism.Method: 78 cross-sectional network studies (N = 303, 151) measuring anxious symptoms using the Generalised Anxiety Disorder Questionnaire (GAD-7) were included. The review considered sample characteristics as well as how networks were conducted and reported, with regard to established reporting standards. Correlation matrices were extracted from studies and used to estimate a Meta-Analytic Gaussian Network Aggregation (MAGNA). Separate networks were estimated and compared across clinical and non-clinical samples.Results: Uncontrollable worry was the most central node in the MAGNA, followed by breadth of worries. The model resulted in a one-factor solution. Findings were stable across samples varying ingender, nationality, age and clinical status. Studies showed some risk of bias in terms of representativeness of samples and inconsistent reporting of some aspects of network analyses. Conclusions: Results robustly corroborate the role of worrying—and particularly the controllablilty of worry—as a central mechanism maintaining anxious symptomatology. This further highlights the therapeutic utility in targeting worry in clinical work where anxiety is either the primary presentation or co-occurring.