Delay discounting as a predictor of adjuvant endocrine therapy adherence among breast cancer survivors.

Abstract

e24172 Background: Despite the demonstrated mortality benefit, 30-70% of hormone-responsive breast cancer (BC) survivors struggle to adhere to adjuvant endocrine therapy (ET). Little is known about patient-specific behavioral traits and their role in ET non-adherence. To gain a survival benefit, patients must adhere to ET for years. In contrast, the benefits of discontinuing ET (i.e., avoiding toxicity) are immediate. Thus, adherence requires one’s behavior to be guided by future outcomes, rather than immediate gratification. We therefore hypothesized that ET adherence may be related to the behavioral economic process of delay discounting (DD; i.e. devaluation of delayed outcomes), which provides a measure of how individuals value the future. Methods: BC survivors (age 18-80) prescribed adjuvant ET within the last 5 years were recruited. ET adherence was measured by pill count and self-report. DD was measured using two methods: monetary DD was measured using a validated task in which participants made repeated, hypothetical choices between a larger amount ($100 or $1000) available after a time delay or half of this amount ($50 or $500) available now. Across trials, the delay to the larger amount is titrated based on previous choices until reaching a delay at which the participant becomes indifferent to waiting for the larger reward. Cancer DD was measured using an investigational task adapted from the adjusting-delay task described above. This task presented repeated, hypothetical choices between preventing BC recurrence after varying time delays and avoiding ET side effects immediately. Results: Eighty nine BC survivors participated. Average pill count adherence was near 50%, but self-reported adherence was near 100%. Monetary DD was significantly associated with pill counts; adherence improved with increasing log indifference delays (standardized b 0.328, p = 0.005). Self-report adherence (categorized as discontinued/never started, suboptimal, and perfect) was significantly associated with Cancer DD and income (OR 3.764, p < 0.001 and OR 1.01, p = 0.04, respectively) but not monetary DD (OR 0.978, p = 0.951). No association was found between ET non-adherence and medication type, side effect severity, menopausal status, age, education or minority status. Conclusions: DD may promote ET non-adherence among BC survivors. Behavioral therapies which reduce DD may provide a targeted strategy to improve adherence. More research is needed to develop tools that reliably assess ET adherence and that tailor DD assessments to the cancer survivor population.