BACKGROUND: We performed a population-based study of patients from the deep South of the United States (with >25% black residents) to evaluate the survival rate of patients with pancreatic cancer. Our aims were to analyze prognostic factors influencing pancreatic cancer survival using the population-based Alabama Statewide Cancer Registry and to determine whether race/ethnicity is an independent determinant of outcomes in patients with pancreatic cancer. METHODS: Eligible participants included all persons diagnosed with pancreatic cancer from 1996 to 2000 and reported to the Alabama Statewide Cancer Registry. Survival time was calculated from time of diagnosis to death for pancreatic cancer deaths or to date of last contact or death from other causes for censored participants. Risk factors associated with survival were assessed with the Kaplan-Meier survival method and the log-rank test. Demographic, tumor, and treatment variables were assessed using the Cox proportional hazards model. RESULTS: Of 2230 patients, the median age at diagnosis was 71 years and the male to female ratio was approximately 1:1. Seventy-three percent of patients were white, and 27% of patients were black. The distribution by stage was 12.5% localized disease, 29.6% regional, 35.3% distant, and 22.6% unstaged. The median survival time for all patients was .39+/-.01 years. Patients who underwent surgical treatment were less likely to die of pancreatic cancer (hazard ratio, .48; 95% confidence interval, .41-.56). Similarly, patients who underwent either chemotherapy or radiation therapy had improved survival rates (hazard ratio, .62; 95% confidence interval, .53-.73). Across all stages, black patients were significantly less likely to receive chemotherapy compared with white patients (26.7% vs 32.3%, P=.02), and were less likely to receive surgical intervention (14.02% vs 17.0%, P=.09). When examining patients who were offered their therapy of choice but refused, we found across all stages that a greater proportion of black patients refused therapies versus whites: 5.6% versus 2.9% (P=.02) for chemotherapy, 3.8% versus 1.6% (P=.04) for radiation, and 9.0% versus 3.3% (P=.001 for surgery). The Cox proportional hazard model showed no effect of race on overall survival time while controlling for stage at presentation, type of therapy received, age at diagnosis, and site of primary tumor. CONCLUSIONS: Survival in patients with pancreatic cancer remains dismal. Tumor characteristics and treatment factors are related directly to survival time in patients with pancreatic cancer. Black patients were less likely to receive therapy but also were more likely to refuse the indicated therapy. Factors leading to racial disparity in the treatment of pancreatic cancer warrant further investigation.