Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group


First Advisor

Dianne L. Chambless


Aromatase inhibitors (AIs) are a daily endocrine therapy for post-menopausal breast cancer survivors and often cause difficult side effects. Women who do not adhere to AI treatment are at higher risk of cancer recurrence and mortality, yet poor adherence is a prevalent problem. This dissertation explores potentially modifiable, psychosocial predictors of AI adherence. In Chapter 1, to determine how to measure adherence in subsequent studies, we examined the psychometric properties of two self-report adherence measures. Using estrogen assays to assess convergent validity, we found that neither measure performed well; however, in exploratory analyses, a “yes/no” item about use of an AI in the past month was found to be associated with estrogen metabolite levels. In Chapter 2, we examined whether survivors’ health beliefs about AI treatment predicted their adherence behaviors. Higher perceived barriers to AI treatment, but not perceived susceptibility to cancer recurrence or perceived benefits of AI treatment, predicted non-adherence. In Chapter 3, we examined the role of arthralgia-associated aging perceptions. After controlling for the severity of arthralgia (a common side effect of AIs), we found that the risk of non-adherence was higher among women who perceived that they had aged rapidly while on the treatment. We used a mixed-methods approach in Chapter 4 to examine the role of partner support. Our qualitative findings suggest that integration of partners into follow-up care and partner support around body changes and sexual dysfunction associated with AIs may contribute to women’s ability to persist with their treatment. Our quantitative results demonstrated that satisfaction with partner support around AI treatment was determined by affection and emotional support, whereas informational and tangible support had no bearings on their satisfaction levels. Support preferences did not differ between women with lower and higher levels of pain from AI side effects; however, women with higher levels of pain tended to receive more informational support. Together, these findings can inform intervention content for women who are at risk of not adhering to AI treatment.