Social risk factors and treatment preferences for patients with breast cancer.

Abstract

e23139 Background: Physicians often overestimate their understanding of patient preferences. While patient preferences are influenced by their values, social risk factors (i.e., social conditions associated with worse health) can constrain patients’ treatment choices. Disparities in breast cancer outcomes persist and may be influenced by social risk factors. By directly measuring both treatment preferences and social risk factors, we sought to determine the influence of social risk factors on stated preferences about a breast cancer treatment decision. Methods: From December 2024 to June 2025, we administered surveys to female patients with breast cancer. Patients were eligible if they were diagnosed with any type or stage of breast cancer within 5 years or on active treatment at the University of Chicago. Patients responded to a sample of 8 of 12 randomly selected hypothetical treatment decision scenarios with their preference for 6 versus 12 months of adjuvant treatment. Scenarios had varying efficacy, toxicity, and out-of-pocket costs. Preferences were elicited as a choice probability (percent chance they would choose 6 months) and a discrete choice (6 vs 12 months). Social risk factors of financial toxicity (COST item 12), prior discrimination in healthcare (BRFSS), limited cancer health literacy (CHLT-6), and limited reading ability (SILS-1) were dichotomized based on published cutoffs. Multivariable regression models clustered standard errors at the patient level and adjusted for sociodemographic factors, clinical factors, and scenario attributes. Results: Of the 248 participants, the mean age was 57 (SD 13), 34% were Black, and 38% were living without a partner. Patients had a median of 0 social risk factors (IQR 0-1). Twenty-one percent of patients had financial toxicity, 10% reported prior discrimination in healthcare, 10% had limited cancer health literacy, and 20% had limited reading ability. While neither the presence of any social risk factor nor the number of social risk factors were significantly associated with elicited choice probability or discrete choice, the specific social risk factor of limited cancer health literacy was associated with lower odds of choosing 6 months (aOR 0.28, p=0.005). Conclusions: While social risk factors generally were not associated with treatment preferences, women with limited cancer health literacy were more likely to choose a longer duration of adjuvant therapy when evaluating efficacy and toxicity tradeoffs. This may be due to those with limited cancer health literacy perceiving more treatment as better. When seeking patient treatment preferences, it is important to assess patient understanding of the information provided. Screening for cancer health literacy and then providing tailored information could help ensure accurate preference elicitation when engaging in shared decision-making.