Objective Our objective was to analyze the time trend variation of 30-day mortality after lung cancer surgery, and to quantify the impact of surgeon and hospital volumes over a 5-year period in France. Methods We used Epithor, the French national thoracic database and benchmark tool, which catalogues more than 180,000 procedures of 89 private and public hospitals in France. From January 2005 to December 2010, 19,556 patients who underwent major lung resection (lobectomy, bilobectomy, pneumonectomy) were included in our study. Multilevel logistic models were designed to investigate the relationship between 30-day mortality and surgeon (model 1) or hospital (model 2) volumes. The 3 levels considered were the patient, the surgeon, and the hospital. Results From 2005 to 2007, the 30-day mortality of patients who underwent major lung resection averaged 10%, and then decreased until it reached 3.8% in 2010 ( P  < .0001). A significant decrease in 30-day mortality was observed over time ( P  = .0046). During the study period, the mean annual number of procedures per surgeon was 46.1 (standard deviation SD = 23.6) and per hospital was 97.9 (SD = 50.8). Model 1 showed that surgeon volume had a significant impact on 30-day mortality ( P  = .03), whereas model 2 failed to show that hospital volume influenced 30-day mortality ( P  = .75). Conclusions Since 2007, when France's first National Cancer Plan became effective, 30-day mortality of primary lung cancer surgery has decreased and currently measures 3.8%. Low mortality was correlated with higher surgeon volume but was not influenced by hospital volume, which cannot be considered a proxy measure for determining the safety of lung cancer surgery.