OBJECTIVES: Identify explanatory variables of hormone therapy (HT) adherence and persistence (A&P) in women with breast cancer (BC), and evaluating the effect of such variables in BC survival rates. METHODS: Retrospective longitudinal data from a cohort of 5861 women with BC, submitted to HT, was put together through linkage of the Brazilian National Cancer Institute datasets, including the control of medicines delivered at its Pharmacy. A logistic regression model was applied to study adherence. Cox proportional hazard models were used to estimate persistence and BC survival. RESULTS: The proportion of treatment adherent was 75.3%. At the end of the first and the fifth year of treatment, respectively, overall persistence to treatment was 79% and 31%, and survival was 94% and 71%. Similarly, better A&P to treatment, as well as BC survival, were associated with higher education, having a partner, lower cancer stages, being submitted to surgery, having less inpatient care, making outpatient visits to a Mastologist and a Clinical Oncologist, and the need of less exams. Older women were more likely to adhere and to persist to treatment, but those aged 70 years old or more presented higher hazard of death. Alcoholism and tobacco use was associated with lower A&P. Longer time between diagnosis and the beginning of HT and cancer family history were, respectively, a risk and a protective factor to treatment persistence and survival. Psychotherapy was protective for adherence and survival. Finally, treatment adherence was positively associated with BC survival, being combined tamoxifen and aromatase inhibitor explicative of lower adherence, while only aromatase inhibitor use was associated with higher hazard of death. CONCLUSIONS: In this cohort, ¼ of the patients did not adhere, only 31% completed the 5-year hormone treatment, and 71% were alive after five years. Socio-demographic, behavioral, clinical and health care aspects explained partially variations in these dependent variables.