Incidence, prevalence, and evidence: Scientific problems in epidemiologic statistics for the occurrence of cancer

Abstract Major changes in policy for health, nutrition, and industrial safety have been proposed because of the "epidemic of apprehension" produced by statistics for the occurrence rates of cancer. The statistical information, however, contains gross violations of epidemiologic principles and scientific standards for credible evidence. The calculations often depend on an improper linkage of numerators and denominators; and the calculated rates, assembled from reports of overtly detected cases of cancer, represent neither incidence nor prevalence. Many of the secular or regional changes in rates are due to corresponding changes in the availability and dissemination of diagnostic technology, but the technologic changes have not been adequately evaluated or investigated. Improvements will require drastically altered approaches to the use of death certificates, tumor registries, and clues from necropsy data.