It is debated whether treating cancer patients in high-volume surgical centres can lead to improvement in outcomes, such as shorter length of hospital stay, decreased frequency and severity of post-operative complications, decreased re-admission, and decreased mortality. The dataset for this analysis was based on cancer registration and hospital discharge data and comprised information on 15,738 non-small-cell lung cancer patients resident and diagnosed in England in 2006-2010 and treated by surgical resection. The number of lung cancer resections was computed for each hospital in each calendar year, and patients were assigned to a hospital volume quintile on the basis of the volume of their hospital. Hospitals with large lung cancer surgical resection volumes were less restrictive in their selection of patients for surgical management and provided a higher resection rate to their geographical population. Higher volume hospitals had shorter length of stay and the odds of re-admission were 15% lower in the highest hospital volume quintile compared with the lowest quintile. Mortality risks were 1% after 30 d and 3% after 90 d. Patients from hospitals in the highest volume quintile had about half the odds of death within 30 d than patients from the lowest quintile. Variations in outcomes were generally small, but in the same direction, with consistently better outcomes in the larger hospitals. This gives support to the ongoing trend towards centralisation of clinical services, but service re-organisation needs to take account of not only the size of hospitals but also referral routes and patient access.