Abstract A014: Endocrine therapy adherence patterns in the Carolina Breast Cancer Study using latent class analysis

Abstract Introduction: Endocrine therapy (ET) is prescribed to hormone-receptor positive breast cancer patients. Lower rates of ET adherence are linked to differences in breast cancer recurrence and mortality. However, adherence patterns are complex, including different underlying barriers. We hypothesized that subtypes of non-adherent patients are identifiable and vary in frequency by clinical and social characteristics. Methods: Using Carolina Breast Cancer Study (CBCS3), we identified women with Stage I-III HR+ tumors who initiated endocrine therapy and completed an endocrine therapy (ET) use questionnaire containing both quantitative and qualitative questions at 25 months follow-up (n = 1042). Using latent class analysis, we grouped patients according to patterns of adherence, including 15 questions such as number of days participants missed pills over a 2-week period and qualitative measures of affordability, patient-provider shared decision making, and medication concerns. We estimated associations between latent classes of adherence and clinical and social characteristics using multinomial regression with post-hoc estimated marginal means. Results: Based on distributions of self-reported ET adherence, we identified 3 latent classes, which we conceptualized as “high adherence”, “adherence variability”, and “substantial barriers”. Women classified as having “substantial barriers” have less patient-provider shared decision making and face three to tenfold greater difficulties in treatment regimens, affordability, side effect severity, and medication tradeoffs. Comparing latent classes to a previously developed binary variable of self-reported adherence vs not (SRA) in CBCS3, “high adherence” had 98.5% SRA, “adherence variability” had 80.9% SRA, and “substantial barriers” had 18.1% SRA. Compared to non-Black women (n = 624), Black women (n = 418) were less frequently in the “high adherence” group (57% vs 72%; RFD = -14.5;-22.8,-6.1) and more frequently in the “substantial barriers” group (13% vs 8%; RFD = 5.5; 0.04, 11.1). Similar associations were observed after adjusting for age and stage. Age ≥50 was also associated with “high adherence” for both Black (RFD = 24.6; 11.5,37.8) and non-Black women (RFD =13.8; 3.7,23.9). Chemo use and stage were not associated with class membership, however, overall, poverty was associated with “substantial barriers” (RFDadj = 10.1; 1.6,18.5) and inversely associated with “high adherence” (RFDadj = -19.4; -30,-8.8). Black women below the 2010 poverty level were significantly less likely to have “high adherence” (RFDadj = -16.5; -29.3,-3.6), compared to non-Black women below the 2010 poverty level (RFDadj = -8.3; -28.7,12.1). Conclusion: Multidimensional measures identify women experiencing barriers and less likely to experience adherence benefits. Younger women, Black women, and women living in poverty are at higher risk of non-adherence. Upcoming analyses will evaluate whether these adherence groups have distinct recurrence patterns. Citation Format: Srijon Mukhopadhyay, Sarah C. Van Alsten, Terry Hyslop, Melissa A. Troester. Endocrine therapy adherence patterns in the Carolina Breast Cancer Study using latent class analysis [abstract]. In: Proceedings of the 18th AACR Conference on the Science of Cancer Health Disparities; 2025 Sep 18-21; Baltimore, MD. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2025;34(9 Suppl):Abstract nr A014.