Predicting outcome after treatment for generalised anxiety disorder

A previous paper, Butler and Anastasiades (Behaviour Research and Therapy 26, 531-534, 1988) presented evidence for three reliable predictors of response to Anxiety Management in patients with generalised anxiety disorder. It was argued there that these reflected severity of anxiety, demoralisation and depression. A second study (Butler, Fennell, Robson & Gelder, Journal of Consulting and Clinical Psychology, 59, 167-175, 1991) has compared two treatments for GAD: Behaviour Therapy and Cognitive Behaviour Therapy. Data from this study is used here to answer two questions: (i) do the same three variables predict outcome in the second study? The answer to this question is 'no'; and (ii) which variables contribute to prediction of outcome when these two treatments are compared? Information presented here suggests that this depends partly on the nature of the treatment given. Outcome after behaviour therapy is predicted by initial levels of anxiety only, but the gains made are also relatively modest. Outcome after cognitive behaviour therapy is predicted by the degree to which ambiguous (external) information is interpreted as threatening. Thus a cognitive variable contributes reliably to the prediction of outcome after a cognitive treatment, but does not predict in the same way to outcome after a behavioural treatment.