Aims: To determine the frequency, timing and patterns of endocrine therapy switching in Australian practice for postmenopausal women with primary breast cancer. Methods: We identified postmenopausal women in a population-based cohort commencing endocrine therapy for invasive primary breast cancer between December 2005 and December 2008 (n = 645). Individual-level administrative health records and self-report data were used to determine women's demographic and clinical characteristics, including preexisting and newly-treated comorbidities, and switches in endocrine therapy. Time to therapy switching was calculated. Chi-square tests compared the characteristics of women who did and did not switch, and those switching within 2 years or after 2 years of commencing therapy. Results: Twenty-eight percent of women switched from their initial endocrine therapy, most commonly from tamoxifen to anastrozole, or the converse. A small number of anastrozole-to-exemestane and letrozole-to-exemestane switches were observed (n = 19). Most women (>80%) who switched therapies did not have newly-treated comorbidities. Few women (<5%) switched before completing 2 years of therapy, but these women were significantly more likely to have preexisting antidepressant use than women switching later (43% vs 23%, P = 0.048) and remained on the subsequent therapy for less time (6 months vs 2.7 years, P < 0.001). Conclusions: Approximately one-quarter of postmenopausal women with primary breast cancer switched endocrine therapies. The findings suggest that the majority of switching in Australian practice was planned; occurring after 2–3 years of, not precipitated by comorbidity, and in a sequence supported by trial evidence. Early switching, however, was associated with preexisting depression and appeared to be a marker of poor persistence.