Objective: To estimate the rate of early discontinuation of oral adjuvant endocrine therapy by women with early-stage breast cancer in a commercially insured population. Methods: The study sample consists of all women from a commercially insured population who underwent mastectomy for breast cancer then filled a first prescription for an oral adjuvant hormonal agent (OAHA, defined as tamoxifen or an aromatase inhibitor) within one year after surgery, and had continuous eligibility for pharmacy benefits from six months prior to surgery. Patients were excluded if they had claims coded for distant metastasis or chemotherapy agents specific for advanced cancer. Days covered by OAHA are deduced from dispensed dates and days supplied. Time to nonpersistence (defined as 180 days without OAHA coverage) is estimated using a Kaplan-Meier analysis and the relation assessed between time to nonpersistence and age and history of cytotoxic adjuvant chemotherapy or radiation therapy preceding endocrine therapy. Results: A total fo 3654 women (age mean 59.8, SD 12.4) were identified who satisfied study criteria, underwent mastectomy between July1998 and December 2006, and had pharmacy benefits eligibility extending through 2007 or at least 180 days after deduced exhaustion of last OAHA supply. A total of 33.2% had claims consistent with cytotoxic adjuvant chemotherapy and 65% had claims for radiation therapy. Including as right-censored patients still receiving therapy at study end (n=1516) and those lost to follow-up (n=969), the cumulative nonpersistence rate is estimated as 24% at three years. Nonpersistence rates were higher for the youngest and oldest patients, and lower for patients with a preceding history of cytotoxic chemotherapy or radiation therapy. Conclusion: It is important to increase understanding of the determinants of persistence with cancer therapies administered orally for long time periods.