Associations between HIV infection status, psychosocial factors, and adjuvant endocrine therapy adherence among South African women with early-stage breast cancer.

Abstract

522 Background: South African (SA) women living with comorbid estrogen receptor positive (ER+) breast cancer (BC) and HIV have higher mortality than other SA women with ER+ BC. We aimed to evaluate for associations between HIV status, psychosocial factors, and adjuvant endocrine therapy (AET) adherence as a potential contributor to this disparity. Methods: From Chris Hani Baragwanath Hospital, Johannesburg, we enrolled adult women with stage I-III ER+ BC in remission and currently prescribed tamoxifen or an aromatase inhibitor to a prospective cohort study. We performed AET pill counts at enrollment, 12 weeks, and 24 weeks, and calculated adherence ratios from the difference in pills between visits divided by the number of days between visits. Women completed questionnaires on sources of social support, beliefs about medication, health literacy, and self-efficacy at enrollment and questionnaires about mental health (depression, anxiety, and acute stress) symptom burden and AET-related side effects at the 12- and 24-week visits. We also collected household wealth data and clinical data on BC and HIV. AET adherence ratios were compared between women with and without comorbid HIV using Wilcoxon rank sum testing. We also used structural equation modeling techniques to refine and evaluate an a priori model of AET adherence with latent variables for mental health, socioeconomic (SES) status, healthcare savvy, and AET side effect burden and subgroup analyses of the resulting model in women with and without HIV. Results: Between April 26, 2022, and July 11, 2023, we enrolled 239 women, 63 (26.4%) of whom had co-morbid HIV. Adherence data was available from 106 women at 12 and 24 weeks, from 22 women at 12 weeks only, and from 111 women at 24 weeks only. Comparing women with and without HIV, median AET adherence ratio was 0.88 vs 0.89, respectively (p = 0.65). Our final model of AET adherence achieved good fit (comparative fit index = 0.96, RMSEA 90% CI: 0.03-0.06). In the full cohort, SES quintile showed a trend towards association with AET adherence (β 0.02, SE 0.01, p=0.09); mental health, healthcare savvy, and side effect burden latent variables were not significantly associated with adherence. In the subgroup of women living with HIV, SES quintile (β 0.04, SE 0.02, p=0.08) and mental health (β -0.02, SE 0.01, p=0.10) showed trends toward association with adherence; these relationships did not persist for women without HIV. Conclusions: HIV status is not predictive of AET adherence among SA women with ER+ BC, but in women with both diagnoses, decreasing SES status and increasing mental health symptoms trended towards an association with decreased adherence. Interventions to improve the mental health of SA breast cancer patients living with HIV may promote AET adherence.