[Methodological approach to human iatrogenic cancers]

Epidemiological methods in iatrogenic cancers do not differ from those used in cancers from other etiology. Two approaches have to be associated in an order related to the novelty of the treatment already studied: --on one hand in vitro or in vivo experimentation (either tests of mutagenicity or carcinogenicity); --on the other, epidemiological surveys, what ever their types (prospective sample, exposed/non exposed, case-control ...). These two approaches must converge to provide sufficient evidence about the causal relationship between treatment and cancer.