International comparisons of survival from lung cancer: pitfalls and warnings.

Population-based survival data can provide valuable comparative data on outcome but should be interpreted with caution. Differences in data collection and analysis, patient and tumor characteristics and treatment options can have an impact on reported results. Ideally, data from the whole population, including clinical-only diagnoses, should be reported and the methods of case identification described. The relative survival rates should preferably be given. Data on patient characteristics such as age, sex, ethnicity and socioeconomic deprivation should be described, together with tumor details such as pathology and clinical stage. Whenever possible, details on the use of treatments should be reported.