OBJECTIVE To examine the impact of physician counseling on perceived risks, benefits, and likelihood of use of anxiolytic pharmacotherapy during pregnancy among women with a history of anxiety. METHOD We surveyed 96 nonpregnant women, aged 21-45 years, with panic disorder and/or generalized anxiety disorder (DSM-IV criteria) recruited by their family physicians to participate in an anxiety treatment trial from 7 primary care practices in Pittsburgh, Pennsylvania. Trained research assistants telephoned study participants to assess sociodemographics, psychiatric history, comorbidities, and anxiety severity. Respondents were asked to assess risks, benefits, and likelihood of taking a prescribed anxiolytic during pregnancy using 3 Likert scales at baseline. Respondents were then asked to indicate whether their perceptions would change with (1) a US Food and Drug Administration (FDA) warning reporting a 5% chance of birth defects with use and (2) physician counseling that the medication was safe during pregnancy despite the warning. Data were collected from January 1, 2005, through December 30, 2007. RESULTS In this study, 46% (44/96) of respondents had generalized anxiety disorder, 14% (14/96) had panic disorder, and 40% (38/96) had both generalized anxiety disorder and panic disorder. The mean baseline Hamilton Anxiety Rating Scale score was 25.6 (SD = 8.4). Respondents were less likely to perceive risk if counseled by their primary care physicians that medication use was safe despite FDA warning. They also saw more benefit in use and reported being more likely to take anxiolytic medications during pregnancy if counseled that doing so was safe. Age, ethnicity, and severity of anxiety did not modify the effect of physician counseling. However, college educated women were less likely to be reassured by primary care physician counseling (P = .05) that anxiolytic use during pregnancy was safe. CONCLUSIONS Women with anxiety disorders are often hesitant to use anxiolytic medications during pregnancy. Physician counseling may change some women's perceptions of risk and decisions regarding use during pregnancy. TRIAL REGISTRATION Clinicaltrials.gov Identifier: NCT00158327.