Background: There is widespread interest in cancer statistics. One can expect that any national report on cancer statistics will quickly attract the attention of the mass media. It is important that statistical concepts be conveyed accurately, even if they are complicated.
Aim: In this presentation, we will explain some statistical concepts that are fundamental in statistical epidemiology. Some of these concepts are not as simple as one might think.
Methods: The basic statistical methods that are most useful in cancer are described in the text by Estève et al.1 We will focus on explaining four measures in cancer statistics, namely incidence, prevalence, risk, and survival rates. Understanding these measures is useful to health care professionals in sharing the latest information on cancer control.
Results: Users of these concepts should approach their application judiciously. There are several measures of incidence, namely crude incidence, incidence rate, age-standardised incidence rate, cumulative incidence rate, and each has its own use. Measures of prevalence, risk and survival require even more careful consideration2.
Conclusions: To interpret statistical reports in cancer correctly, it is essential to understand the definitions of the key concepts. In applications, one should use the relevant measures. These matters are important for researchers, policy makers, decision makers (including clinicians, patients and their carers who also make decisions), and those who summarise these reports for the media. Finally one should be mindful of the inherent difficulties associated with conveying statistical information, especially probabilities3.