[Methodology for survival assessment of cancer patients using population-based cancer registration data]

Evaluating and monitoring long-term survival of cancer patients and reporting the survival rate are routinely employed by cancer registries. Long-term survival rate is a necessary indicator in evaluating the effect of cancer therapy and cancer burden. Cohort method is a traditional approach for survival analysis, but it essentially reflects the survival expectations of patients diagnosed many years ago, therefore survival status of cancer patients was often disclosed with delay. Given the limitation of cohort method, period analysis and model-based period analysis are subsequently proposed and gradually applied in assessment of survival rates in recent years. Period analysis includes the patients of interest period, which reflects more up-to-date estimates of long-term survival of cancer patients. While model-based period analysis can use the existing data to calculate survival rates and to assess the trend, and predict survival rates in the future. Compared with cohort approach, period analysis and model-based period analysis are better in timeliness and precision in survival analysis. This article reviews the definition and theory, calculation and application of cohort method, period analysis and model-based period analysis, in order to provide a basis on up-to-date and precise assessment of survival rates of cancer patients.