Background: There are indications that pancreatic cancer survival may differ according to sociodemographic factors, such as residential location. This may be due to differential access to curative resection. Understanding factors associated with the decision to offer a resection might enable strategies to increase the proportion of patients undergoing potentially curative surgery. Methods: Data were extracted from medical records and cancer registries for patients diagnosed with pancreatic cancer between July 2009 and June 2011, living in one of two Australian states. Among patients clinically staged with non-metastatic disease we examined factors associated with survival using Cox proportional hazards models. To investigate survival differences we examined determinants of: 1) attempted surgical resection overall; 2) whether patients with locally advanced disease were classified as having resectable disease; and 3) attempted resection among those considered resectable. Results: Data were collected for 786 eligible patients. Disease was considered locally advanced for 561 (71%) patients, 510 (65%) were classified as having potentially resectable disease and 365 (72%) of these had an attempted resection. Along with age, comorbidities and tumour stage, increasing remoteness of residence was associated with poorer survival. Remoteness of residence and review by a hepatobiliary surgeon were factors influencing the decision to offer surgery. Conclusions: This study indicated disparity in survival dependent on patients' residential location and access to a specialist hepatobiliary surgeon. Accurate clinical staging is a critical element in assessing surgical resectability and it is therefore crucial that all patients have access to specialised clinical services.