OBJECTIVES: Lack of non-compliance is often underestimated in breast cancer treatment. The aim of our study was to analyze the persistence with tamoxifen (TAM) and aromatase inhibitors (AI) in women with breast cancer (BC) and to identify reasons of treatment discontinuation and determinants of non-persistency. METHODS: We used data of the Oncology Analyzer database, which includes individual information on patient history related to the treatment of patients across all cancer types. This enables a complete overview of cancer patient care from diagnosis onward, facilitating research in areas such as treatment changesn, dosing and regimen compliance, market sizing and off-label use. We identified 7063 breast cancer patients with a start of TAM or AI therapy from 1990 until 2011 and with a treatment duration of at least 365 days. RESULTS: After one year of follow up, 11.2% of TAM, and 18.4% of AI treated patients discontinued their treatment. In these patients with , the reasons for stopping were: progressive disease including local and distant progression (68.1%), side effects (15.5%), patient’s choice (8.6%) and other reasons (7.8%). The multivariate hazard ratios of the cox regression models showed that patients younger than 50 were most likely to discontinue initial therapy when compared with the reference group of women over 70 years of age (HR: 2.30, p=0.01). In contrast, patients treated in gynaecologist or oncologist practice had significantly longer persistence than patients who obtained their prescriptions in general hospital or academic cancer facility (HR: 0.47, p=0.02). Additionally, patients with therapy initiation in gynecological practices had significantly longer persistence than in oncological practice (HR: 0.68, p=0.04). Additionally, metastases were associated with strongly increased risk of treatment discontinuation (HR: 3.81, p<0.01). CONCLUSIONS: The proportion of breast cancer patients with therapy discontinuation within first year after therapy start is high and needs to be significantly increased to improved outcome in clinical practice.