Geographic location and other determinants of patient decision-making to decline chemotherapy in advanced breast cancer.

Abstract

111 Background: Much work has been done to explore how social determinants of health are associated with disparities in the use of chemotherapy for breast cancer. We used a large national registry to explore whether patient decision-making to accept or decline chemotherapy varies geographically across the United States. Methods: This analysis used the National Cancer Database to study decision-making in patients diagnosed with advanced breast cancer (AJCC clinical stage III-IV) between 2004-2017. We identified patients recommended chemotherapy by their physicians, but who declined treatment. We performed multivariate logistic regression modeling to explore variables associated with the decision to decline chemotherapy. Results: A total of n = 167,647 patients with stage III-IV breast cancer were included. After controlling for age, race, ethnicity, Charlson/Deyo comorbidity scoring, receptor status, histology, grade, patient education, insurance status, and facility type, geographic region was found to be significantly associated with the decision to either accept or decline chemotherapy (p < 0.001). Patients in the New England area (i.e. CT, MA, ME, NH, RI and VT) exhibited the highest rate of refusal of chemotherapy (9.3%). In contrast, patients in East South Central states (i.e. AL, KY, MS, TN) were the least likely to decline chemotherapy at 5.5% rate of refusal (OR 0.55, 95% CI: 0.47-0.66, p < 0.001). Rates of refusal were lower for patients with TNBC at 4.1% versus 7.7% in those with HR+/HER2 tumors. Patients aged > 70 had increased odds of declining chemotherapy (OR 9.30, 95% CI: 8.17-10.59, p < 0.001) compared to patients < 50. Race and level of education were not associated with patient decision-making in this cohort, however, Hispanics had lower odds (OR 0.70, 95% CI: 0.60-0.82, p < 0.001) of refusing chemotherapy as compared to Non-Hispanics. Patients with higher comorbidity were also more likely to decline chemotherapy (p < 0.05). As compared to the uninsured, Medicare-insured (OR 0.69, 95% CI: 0.60-0.80, p < 0.001), Medicaid-insured (OR 0.65, 95% CI: 0.56-0.76, p < 0.001) and privately-insured patients (OR 0.39, 95% CI: 0.33-0.44, p < 0.001) were less likely to decline chemotherapy. Lastly, patients treated at academic institutions (OR 0.90, 95% CI: 0.81-0.99, p = 0.029) were less likely to decline chemotherapy compared to those in community cancer programs. Conclusions: This study showed that after adjusting for known social determinants of health, there was a significant difference in rates of refusal of recommended chemotherapy by geographic region, with highest refusal rates in New England states. Of interest, Hispanic ethnicity was associated with lower likelihood to decline chemotherapy, while uninsured status was associated with increased refusal. Researching this geographic differential further will help improve efficacy of care delivery in patients with cancer.