A controlled comparative investigation of large group therapy for generalised anxiety disorder - "stress control"
One hundred and nine generalised anxiety disorder (GAD) patients, referred by their General Practitioners to a clinical psychology primary care service, were assigned to either Cognitive, Behavioural, Cognitive-behavioural, Placebo or Waiting List conditions. `Stress Control' large group therapy combined didactic therapy with a workshop model and emphasised the aim of turning patients into their own `therapists' in order to enable them to deal with present and future problems. Patients were thus encouraged to view Stress Control as an `evening class' rather than `group therapy'. Measures of treatment process and outcome were obtained mainly from self-report instruments. Follow-up data were collected at six months post-treatment. At post-therapy, all active therapy conditions and, against expectation, the Placebo condition had shown significant time within treatment group change. The active therapy conditions, and to a lesser extent, the Placebo condition, were significantly different to the Waiting List condition, which, overall showed no evidence of improvement. At follow-up the active therapy condition generally enhanced therapy gains while the placebo condition maintained therapy gains. Process measures did not, with the exception of self-statement change, differentiate between the groups. Noted variable response in the main analyses was somewhat explained by various sub-group analyses. There appeared to be little benefit in dividing patients into those who experienced panic and those who did not. There was some evidence that `matching' patients to therapy, i.e. cognitive responders to cognitive therapy was of value at post-therapy although differences generally disappeared at follow-up. Synchronous change was associated with enhanced performance. Finally, attempts to predict response to Stress Control by a comparison of responders and non-responders were attempted and the results assessed in terms of clinical as opposed to statistical significance. The results of the present study are discussed with reference to other treatment outcome studies and an attempt to produce a model to account for the similar effects found across treatment conditions. The implications of these findings and some suggestions for future research for GAD and other diagnostic categories are discussed.