Asian Americans (AAs) are the only race for whom cancer is the leading cause of death in the US. Despite radiation therapy (RT) and surgery being curative treatments, prior work demonstrated AAs refuse RT/surgery at a higher rate than other races. Given AAs are often aggregated with Native Hawaiians and other Pacific Islanders (NHPIs), rates of RT/surgery refusal within this community are poorly understood. We aimed to 1) assess RT/surgery refusal on overall survival (OS) using propensity-score (PS) matched groups, 2) identify AA and NHPI populations that refuse RT/surgery and 3) determine predictors of refusing RT/surgery. A US hospital-based retrospective cohort study of the National Cancer Database was conducted. Patients included were ≥18 years old, AA by ethnogeographic region (South AA, East AA, and Southeast AA) or NHPI race, with a confirmed diagnosis of 1 of the 10 most common US cancers during 2004-2017 who were recommended for RT/surgery. Cox proportional hazard models with adjusted Hazard Ratios (aHR) assessed propensity-score matched groups (1:10) based on age, race, cancer stage, comorbidity index, rurality, facility type, facility location, and year of diagnosis. Adjusted odds ratios (aOR) were calculated and 95% confidence intervals (95% CI) for treatment refusal using logistic regression. Population heterogeneity for treatment refusal by race was assessed with likelihood ratio tests (p-heterogeneity). The cohort of 147,685 patients who met the inclusion criteria were predominantly East AA (43%), diagnosed with breast cancer (42%), and had a <2 comorbidity index (99%). Median age was 61 years. Median follow-up was 58 months. Overall, 2,888 (5%) patients refused RT, and 1,073 (1%) refused surgery. RT refusal by race was 5.7% (East AA), 7.9% (NHPI), 4.6% (South AA), and 4.6% (Southeast AA). Surgery refusal rates were 1.5% (East AA), 1.9% (NHPI), 1.1% (South AA) and 1.2% (Southeast AA). RT refusal significantly predicted poorer OS (aHR = 1.15, 95% CI = 1.06-1.25) whereas surgery refusal did not. Compared to East AA, NHPIs had a higher risk of RT refusal (aOR = 1.44, 95% CI = 1.27-1.62), whereas South AA (aOR = 0.85, 95% CI = 0.75-0.95) and Southeast AA (aOR = 0.82, 95% CI = 0.75-0.90) had a significantly lower risk (p-heterogeneity<.0001). Predictors of RT refusal were older patient age, high comorbidity index, oral cavity cancer, urban-rural residence, Midwest or West US region, and diagnosis 2011-2017. Among AA and NHPI patients with cancers in the US, RT refusal predicted poorer OS. NHPI had the highest risk of RT refusal. Given AA and NHPI are not monolithic groups, data disaggregation is necessary to understand racial/ethnic disparities for treatment refusal. Sociocultural and historical contexts of AA and NHPI populations on treatment refusal are necessary to improve cancer outcomes among these populations.
Read full abstract