The objective is to estimate parameters of the natural history of panic disorder, including its prodrome, incidence, recovery and recurrence.

In 1981 the Baltimore Epidemiologic Catchment Area Study interviewed 3481 individuals probabilistically selected from the household population. During 1993–1996, 1920 of these individuals (73% of survivors) were interviewed again. Baseline and follow-up interviews included the National Institute of Mental Health Diagnostic Interview Schedule. During the follow-up, a subsample was assessed by psychiatrists using the World Health Organization Schedules for Clinical Assessment in Neuropsychiatry (SCAN).

There were 35 new cases of panic disorder in 24 475 person years of exposure, yielding an annual incidence of 1.43 per 1000 per year. Data from the SCAN assessments suggest the incidence estimate is conservative. Incidence is greater in females and declines with age. About one-third of the new cases arise without agoraphobia, but about half have anxiety of some sort present for many years prior to meeting criteria for diagnosis. People with agoraphobia have less intense onsets but slower recoveries than those without agoraphobia.

Panic is heterogeneous in its pattern of onset and recovery. Some of the heterogeneity is associated with the presence of other anxiety over a long period of the life.