Survival data for pancreatic cancer are usually based on actuarial calculations and actual long-term survival rates are rarely reported. Here we use population-level data from the Surveillance, Epidemiology, and End Results program for patients with microscopically confirmed pancreatic ductal adenocarcinoma diagnosed from 1975 to 2011. A total of 84,275 patients with at least 5 years of follow-up were evaluated (follow-up cutoff date: December 31, 2016). Actual 5-year survival for pancreatic cancer increased from 0.9% in 1975 to 4.2% in 2011 in patients of all stages (p < 0.001), while in surgically resected patients, it rose from 1.5% to 17.4% (p < 0.001). In non-resected patients, the actual 5-year survival remained unchanged over the same time period (0.8% vs 0.9%; p = 0.121). Multivariable analysis of surgically resected patients diagnosed in the recent time era (2004–2011) showed that age, gender, grade, tumour size, TNM-stage and chemotherapy were significant independent predictors of actual 5-year survival, while age, grade and TNM-stage were significant independent predictors in non-resected patients. However, unfavourable clinicopathological factors did not preclude long-term survival. Collectively, our findings indicate that actual 5-year survival for pancreatic cancer is still below 5% despite improvement of survival for the subset of patients undergoing surgical resection.