To quantify the economic burden associated with premenstrual dysphoric disorder (PMDD) by assessing health care service use and related expenditures, work loss, role limitation, and productivity. Women ages 21 to 45, randomly selected from membership of a northern California HMO (n = 1,194), provided prospective daily symptom ratings and survey data on health care use and productivity for two menstrual cycles. Summary measures of 12-month utilization and expenditures based on HMO encounter data also were constructed. Based on daily symptom ratings, we classified women as having minimal (n = 186), moderate (n = 801), and severe (n = 151) premenstrual symptoms, or PMDD (n = 56) and compared health care use and expenditures, predicted values of productivity and work loss, and marginal effects of symptom severity on outcome measures. Women with PMDD had higher degrees of luteal phase (premenstrual) productivity impairment than those with minimal symptoms. Compared with the minimal and moderate symptom groups, women with PMDD continued to report lower productivity (P <0.01) in the 5 to 10 days after onset of menses (follicular phase). We found little evidence that women spent more time in bed, reduced time at work, or decreased activities at home or school as a result of premenstrual symptoms. As symptom severity increased, the likelihood of health care service use increased only for an emergency department, obstetrician/gynecologist, or alternative medicine provider visit. There were no significant differences in health care expenditures across the symptom groups. The economic burden associated with PMDD manifests itself primarily in reported productivity decrements rather than health care utilization or costs associated with time away from work.