LBA6002 Background: Generalized competing event (GCE) models have been used to stratify patients with cancer according to their relative hazard for cancer death versus death from other causes. We evaluated outcomes for head and neck cancer (HNC) patients treated at Kaiser Permanente Northern California (KPNC) based on demographic data and comorbidities using a GCE model. Methods: We identified 884 HNC patients diagnosed 2000-2015 from the KPNC cancer registry, age 18-85 and stage II-IVB by AJCC 7th edition. Using the GCE proportional relative hazards model, controlling for age, sex, tumor site, and Charlson comorbidity index (CCI), we identified associations between these factors and the relative hazard for HNC-specific mortality (ω+ ratio, ‘gcerisk’ package in R). Death, disenrollment, and end of study (12/31/2016) were used as censoring events. Logistic regression models estimated the odds of receiving intensive treatment (platinum based regimen), adjusting for the same covariates plus stage, smoking, and alcohol abuse history. Results: With a median follow-up of 2.9 years, 271 patients died of cancer, and 93 of non-cancer causes. Compared to male, females were less likely to receive intensive chemotherapy (35% vs. 46%, p = 0.006) and radiation (60% vs. 70%, p = 0.008). On GCE analysis, female patients had an increased relative hazard ratio (RHR) for death from HNC vs. other causes (adjusted RHR 1.92; 95% CI 1.07-3.43), indicating they may be relatively undertreated. Conclusions: Female patients in our cohort may be undertreated in clinical practice, potentially missing the opportunity to aggressively treat their HNC. This study supports the use of a GCE methodology to objectively identify patients more likely to benefit from treatment intensification. These findings may help guide future research in health disparities.[Table: see text]
The impact of health insurance status on the survival of patients with head and neck cancer
