Treatment of locally advanced pancreatic cancer in the real world: Population-based practices and effectiveness

Purpose : To evaluate the use and effectiveness of cancer-directed therapy in elderly patients with locally advanced pancreatic cancer (LAPC). Methods: We used the linked Surveillance, Epidemiology, and End Results Medicare database to perform a retrospective cohort study in 1,696 patients diagnosed with LAPC between 1991 and 1996. We calculated cancer-directed treatment use rates, then used logistic regression to identify patient and health system factors that were associated with receipt of treatment. Effectiveness of treatment was estimated using Cox proportional hazards models and propensity score methods. Results: In our cohort, 44% of patients received some form of cancer-directed therapy (24% radiation with concurrent chemotherapy, 13% radiation alone, and 7% chemotherapy alone). Older age, lower socioeconomic status, presence of comorbid illness, no care in a teaching hospital, and residence in the western United States were associated with a lower likelihood of receiving treatment (P less than or equal to 0.05). Among those treated, younger age and certain geographic locations were the only predictors of receiving combined-modality therapy. The adjusted hazard ratio for death associated with any treatment in the Cox model was 0.53 (P < .0001). Effectiveness estimates obtained using propensity score methods were similar. Conclusion: This analysis supports the effectiveness of cancer-directed treatment in elderly patients with LAPC, but use is low. Receipt of treatment is strongly correlated with non-disease-related factors, especially sociodemographic characteristics, indicating possible disparities in access to care. (C) 2003 by American Society of Clinical Oncology.