PURPOSE Despite the benefits of adjuvant endocrine therapy (AET) for reducing recurrence and mortality risks after hormone-sensitive breast cancer, AET adherence is sub-optimal for a high proportion of women. However, little is known about long-term patterns of AET adherence over the minimally recommended 5 years. Our objectives were to: (1) identify 5-year AET adherence trajectory groups; (2) describe trajectory groups according to adherence measures traditionally used (i.e., Proportion of Days Covered); and (3) explore factors associated with trajectories. METHODS We conducted a 5-year cohort study using data from a French national study that included AET dispensing data. Women diagnosed with first non-metastatic breast cancer and having at least 1 AET dispensing in the 12 months after diagnosis were included. Group-based trajectory modeling was used to identify adherence trajectory groups by clustering similar patterns of monthly AET dispensing. Multinomial logistic regressions were used to identify factors associated with trajectories. RESULTS Among 674 women, five AET adherence trajectory groups were identified: (1) quick decline and stop (5.2% of women); (2) moderate decline and stop (6.4%); (3) slow decline (17.2%); (4) high adherence (30.0%); and (5) maintenance of very high adherence (41.2%). Mean 5-year Proportion of Days Covered varied from 10 to 97% according to trajectories. Women who did not receive chemotherapy or a personalized care plan were more likely to belong to trajectories where AET adherence declined and stopped. CONCLUSION Our results provide information on the diversity of longitudinal AET adherence patterns, the timing of decline and discontinuation and associated factors that could inform healthcare professionals.

中文翻译：

---

非转移性乳腺癌女性开始辅助治疗后五年内的坚持轨迹：一项使用行政数据库进行的队列研究。

目的尽管辅助内分泌疗法（AET）可以降低激素敏感性乳腺癌患者的复发和死亡风险，但对于大部分女性而言，AET依从性并不理想。但是，对最少推荐的5年内AET依从性的长期模式知之甚少。我们的目标是：（1）确定5年的AET遵守轨迹组；（2）根据传统使用的依从性度量方法（即所涵盖天数的比例）来描述轨迹组；（3）探索与轨迹相关的因素。方法我们使用来自法国国家研究的数据（包括AET配药数据）进行了为期5年的队列研究。纳入诊断为首次非转移性乳腺癌且在诊断后的12个月内至少分配1次AET的女性。通过对每月AET分配的相似模式进行聚类，使用基于组的轨迹建模来识别依从轨迹组。多项式逻辑回归用于确定与轨迹相关的因素。结果在674名妇女中，确定了五个AET依从性轨迹组：（1）快速下降和停止（5.2％的妇女）；（2）适度下降和停止（6.4％）；（3）缓慢下降（17.2％）；（4）依从性高（30.0％）; （5）保持很高的依从性（41.2％）。根据轨迹，平均5年覆盖天数比例从10％到97％不等。未接受化学疗法或个性化护理计划的女性更有可能属于AET依从性下降和停止的轨迹。结论我们的结果提供了有关纵向AET依从性模式多样性的信息，