Alprazolam and exposure for panic disorder with agoraphobia. Attribution of improvement to medication predicts subsequent relapse.

Patients with panic disorder plus agoraphobia had 8 weeks of drug treatment (alprazolam or placebo) plus psychological treatment (exposure or relaxation). At the end of treatment at week 8, 40 patients who had become much/very much improved rated how much their gains were attributable to medication or to their own efforts. During the tapering-off to week 16, and treatment-free follow-up to week 43, patients who at week 8 had attributed their gains to medication and felt less confident in coping without tablets had more severe withdrawal symptoms and greater loss of gains than did patients who at week 8 had attributed their gains to their own efforts during treatment. Baseline illness severity, greater age, higher expectations from drug treatment, and more side-effects of drugs during treatment all predicted more external attributions (i.e. to the effect of drugs) but did not independently predict relapse. Patients on alprazolam compared with placebo had more drug attributions. Though drug attributions predicted relapse in both alprazolam and placebo groups, predictions were stronger in the alprazolam group.