Klin Onkol 2014; 27(2): 127-135. DOI: 10.14735/amko2014127.
Summary
Background: Calculating 5-year overall and relative survival is the standard method for population-based analyses in oncology. Survival rates based on population data do not, however, guarantee standardized benchmarks for comparison of different patient populations, which is especially true when compared populations differ considerably in age structure and representation of clinical stages. In this paper, we present and compare statistical methods for standardization of cancer survival rates. Patients and Methods: Using data of the Czech National Cancer Registry, we estimated 5-year overall and relative survival estimates for periods 2001 – 2005 and 2006 – 2010. To demonstrate the effect of standardization, we calculated crude and age- standardized survival rates as well as survival rates standardized for both age and clinical stage. Results: Our results show that the particular standardization method influences resulting 5-year overall and relative survival rates regarding both within and between time periods comparisons. In addition, our results document a recent improvement in 5-year relative survival between periods 2001 – 2005 and 2006 – 2010 for 19 of 20 evaluated diagnoses. All most prevalent cancers including prostate, lung, colorectal, breast, kidney, and uterine cancer and melanoma were observed among the diagnoses with statistically significantly improved patient survival. Conclusion: Unless the use of standardization to the age and stage of tumor is limited due to a small number of patients in individual age- and stage- specific subgroups, this method can be considered as a proper statistical methodology for the population assessment of Czech cancer patient survival rates.
