[Epidemiologic methods of estimation of a secondary cancer associated with cancer treatment]

It is a prerequisite to define "secondary cancer associated with cancer treatment" before estimating it statistically. There are 4 major statistical methods for estimating secondary cancer associated with cancer treatment. Method A: Individually calculate a person-year and multiply it by the cancer incidence rate to obtain the expected number of secondary cancers. Then, compare this with the observed number. Method B: Estimate a person-year by multiplying the survival rate for the same type of cancer by the number of cancer patients. Then, calculate the expected number of secondary cancers as in Method A and compare the result with the observed number. Method C: Compare the proportions of site-specific cancer incidence in the study group with those of the control group. Method D: Compare the incidence rates of secondary cancer in two treatment groups in a randomized trial or in well matched groups. Method A is considered to be the best method, but it is necessary to use the best matched cancer incidence rate in estimating the expected number of secondary cancers. In evaluating the causal relationship it is necessary to take into account both biological significance and statistical significance.