OBJECTIVES: The purpose of this study is to estimate the impact of anxiety on absenteeism and employment in the United States using the 1997 Medical Expenditure Panel Survey. METHODS: Anxiety cases were defined by self-report by survey respondents who saw a physician during the previous year for a condition and recorded as a three digit ICD-9 code (Anxiety disorders=300). National population estimates were derived from patient level weights as determined by a nationally representative sample of the non-institutionalized civilian population of the United States in 1997. Descriptive analyses were used to examine individual demographic, socioeconomic, and co-morbidity characteristics of those respondents with anxiety. RESULTS: Of the 1997 US representative population of 34,551 survey respondents from the MEPS analysis, 848 persons, representing 684,312 persons (2.5% of the US population), reported having anxiety. Anxiety respondents were more likely to be female (69.2%), Caucasian (83.6%), and older than the overall population (mean age 44 years vs. 33 years). Among respondents aged 18-64 years, persons with anxiety were less likely to work than those without anxiety (males 59.1% vs. 70.8%; females 53.5 % vs. 56.9%). Those persons reporting anxiety were more likely to have missed days from work 39.1% vs. 28.2% for females and 35.9% vs. 27.5% for males. Persons with anxiety were more likely to change jobs through the year and a half interview reference period (57.8% vs.49.2% for females; 63.9% vs. 57.4% for males). CONCLUSIONS: Differences are seen at the workplace in patients with and without anxiety disorders. Patients with anxiety most often are female and have a greater frequency of not being employed, missing time from work, and changing jobs.