This study investigated whether facilitating a benign interpretive bias decreases negative thought intrusions in generalized anxiety disorder (GAD). Clients were randomly allocated to an interpretation modification condition in which they repeatedly accessed benign meanings of emotionally ambiguous homographs and scenarios, or to a control condition in which they accessed threat and benign meanings with equal frequency. Worry frequency was assessed using a breathing focus task that involved categorising the valence of thought intrusions before and after an instructed worry period. Interpretation bias was assessed during the modification tasks, and on a different measure of interpretation bias (sentence completion) following a period of worry. The experimental procedure modified interpretations made during training, and in the later sentence completion task. Furthermore, compared to the control group, the benign group showed fewer negative thought intrusions during breathing focus (as rated by both participants and an assessor). These findings show that it is possible to induce a more benign interpretive bias in GAD clients and that this reduces negative thought intrusions.
The key defining feature of Generalized Anxiety Disorder (GAD) is excessive and uncontrollable worry about multiple topics that is experienced as being difficult to stop. Supporting such reports, Ruscio and Borkovec (2004) found that, following instructions to worry about a current personally relevant topic, individuals with GAD had more negative thought intrusions during a subsequent task than did non-clinical high worriers. While cognitive-behaviour therapy (CBT) is effective in reducing the symptoms of GAD, around half of treated clients fail to achieve high end state functioning (Borkovec, Newman, Pincus, & Lytle, 2002). To improve effectiveness and efficiency of interventions it is important to gain a better understanding of the mechanisms that maintain the cardinal feature of GAD – excessive worry.
GAD is associated with the interpretation of ambiguous events in a threatening manner (Eysenck et al., 1987, Eysenck et al., 1991, Mathews et al., 1989 and Mogg et al., 1994), increasing perceived danger, and perhaps triggering worry (Mathews, 1990). In contrast, non-anxious individuals are characterised by an opposing bias favouring benign interpretations of ambiguous events (Hirsch and Mathews, 1997 and Hirsch and Mathews, 2000). However, correlations between a threatening interpretive bias and anxiety are clearly insufficient to demonstrate a causal relationship between them.
Recent research has demonstrated that a threatening interpretive bias can be induced in non-anxious individuals through repeated practice in accessing negative outcomes of emotionally ambiguous information, and causes heightened anxiety when confronted with a threatening event (e.g., Hirsch et al., 2007, Mathews and Mackintosh, 2000 and Wilson et al., 2006). Related methods have been used to induce a more benign interpretive bias in highly anxious individuals, leading to reductions in anxiety reactivity (e.g., Mathews et al., 2007 and Murphy et al., 2007).
Of most relevance to the current research, Hirsch, Hayes, and Mathews (2009) demonstrated that non-clinical high worriers given repeated practice accessing benign meanings of threat-related homographs and emotionally ambiguous scenarios showed less negative thought intrusions than a control group who accessed threatening and benign meanings with equal frequency. The question remains of whether it is possible to facilitate a more benign interpretive bias in clients with GAD and if so, whether this also leads to reductions in worry frequency. The present study investigated this question, using similar methods to Hirsch et al. (2009).
Worry frequency has been assessed in GAD clients by Ruscio and Borkovec (2004) using a task which required participants to focus on their breathing and report on thought intrusions both before and after an instructed worry period. This paradigm was adapted by Hirsch et al. (2009) to include assessor ratings of intrusions in addition to that of participants. The current study similarly included assessor ratings of the valence of thought intrusions reported by GAD clients. The persistence of interpretive bias changes was assessed using an adapted version of a sentence completion task developed by Huppert, Pasupuleti, Foa, and Mathews (2007).
