Category: Adult Anxiety - GAD

PS10- #A29 - Contribution of the Proposed Worry Behaviors Criterion to the DSM-5 Definition of GAD

Saturday, Nov 18
11:00 AM – 12:00 PM
Location: Indigo Ballroom CDGH, Level 2, Indigo Level

Keywords: GAD (Generalized Anxiety Disorder) | Worry | DSM-5

Generalized anxiety disorder (GAD) is characterized by difficult to control and excessive worry about a number of life matters, accompanied by at least 3 of 6 associated symptoms (e.g., irritability and muscle tension). Research suggests that GAD is associated with safety behaviors such as avoidance of potentially negative events, (over)preparing for possible negative outcomes, procrastination, and reassurance seeking. Indeed, a worry behaviors criterion entailing the presence of at least 1 of these 4 worry behaviors was proposed for inclusion in DSM-5 GAD (Andrews et al., 2010).  Despite avoidance behaviors being a diagnostic feature of the other major anxiety disorder categories, this criterion was ultimately not included in the DSM-5 definition of GAD, in part because of insufficient research about these behaviors in GAD. This study provided the first large-scale evaluation of these worry behaviors and their contribution to the diagnosis of GAD.


The sample was 702 adult outpatients seeking treatment at an anxiety and related disorders clinic (307 met criteria for GAD). Patients were assessed using the Anxiety and Related Disorders Interview Schedule for DSM-5 (ADIS-5). In the entire sample, ADIS-5 clinicians rated the excessiveness of 8 worry areas, the severity of the 6 associated symptoms, and the severity of the 4 aforementioned worry behaviors. Per the DSM-5 proposal, patients who endorsed at least one worry behavior at a clinical level (“sometimes” or higher) were deemed to meet the worry behaviors criterion.


Most patients with GAD (282 of 307, 91.9%) met the proposed worry behaviors criterion; rates of endorsement for each behavior were: avoidance (49.8%), over-preparation (46.3%), procrastination (67.8%), and reassurance seeking (57.0%). Bivariate analyses indicated that the presence of each behavior increased the odds of having GAD by at least 5 times and meeting the worry behaviors criterion was associated with a 12.55 times greater odds of GAD. Hierarchical logistic regressions attested to the incremental validity of the worry behaviors criterion by showing that the criterion, as well as each behavior individually, contributed significantly to the prediction of GAD caseness above and beyond the core defining features of GAD (excessive worry, associated symptoms). The addition of this criterion resulted in a modest improvement in classification accuracy (2%), reflecting the fact that 7 cases who were classified as false negatives based on existent DSM-5 GAD features were correctly classified as having GAD after the worry behaviors criterion was added to the analysis. A more stringent version of the worry behaviors criterion (requiring 2 or more behaviors) did not improve classification accuracy, and indicated that another 62 (20.2%) patients currently meeting DSM-5 GAD would not be assigned this diagnosis if this criterion was required.  In addition to presenting further results (e.g., interrater reliability, characterization of GAD cases who do not meet this criterion), the findings will be discussed with regard to their implications and future research on the potential role of worry behaviors in the formal diagnostic definition of GAD.

Esther S. Tung

PhD candidate
Boston University
Boston, Massachusetts

Timothy A. Brown

Boston University