OBJECTIVE: To examine the medical services and treatment for anxiety disorders reported by patients who had either panic disorder with agoraphobia or else social phobia. DESIGN: Archival research of consecutive records of psychiatric interviews conducted between January 1990 and December 1991. The records were examined by a trained research assistant who had had no contact with the patients. PATIENTS: One hundred patients who had panic disorder with agoraphobia and twenty-eight patients who had social phobia. SETTING: An anxiety disorders clinic in a university-affiliated psychiatric institute. OUTCOME MEASURES: Variables related to the use of medical services included history of hospitalization, emergency department visits and referrals to specialists. Variables related to treatment included types of medication received, whether behaviour therapy was received and types of health care professionals seen. RESULTS: Almost 30% of the patients with panic disorder and more than 20% of those with social phobia had a history of a major depressive episode at some time in their lives; 30% and 25% respectively had a current nonpsychiatric medical diagnosis. In the past year nearly one-third of both patient groups had seen three or more different health care professionals and almost one-fifth of those with panic disorder had gone to a general hospital emergency department. Of the patients with panic disorder 9% had previously been assessed by a cardiologist and 17% by a neurologist. At least two-thirds of each group had received benzodiazepines, often for use as needed. Although most of the patients in both groups had been seen by mental health professionals such as psychiatrists, few had received optimal treatment. Of those with panic disorder, only 15% had received the tricyclic antidepressant imipramine, 13% alprazolam and 11% cognitive-behavioural therapy. Only 4% of the patients with social phobia had received cognitive-behavioural therapy. CONCLUSIONS: Both groups of patients, and particularly those with panic disorder, are frequent users of medical services. Although most have had contact with mental health professionals, few have received appropriate treatment. Benzodiazepines appear to be overprescribed, whereas forms of treatment that have been shown to reduce the use of medical services, such as cognitive-behavioural therapy, are infrequently given.