Some Problems in the Estimation of Familial Risks of Disease

Publisher Website
Google Scholar
Abstract
Some differences between investigations of familial distribution of disease and the distribution of recessive characteristics in individuals are discussed. While the biological model for inheritance of recessive characteristics supports the use of a binomial distribution in the estimation of their frequencies, from index-case data, its application in studies of disease is questionable. Different approaches to the calculation of familial risk of disease, from index-case and control data, and their underlying assumptions are discussed and illustrated with stomach-cancer data reported by Videbaek and Mosbech. The model of equal probability of ascertaining affected individuals is also considered for familial studies of disease. Tabulations of the case-order distribution of index cases are advocated as a means of investigating whether ascertained index cases reflect a distribution of affected individuals or of families with affected individuals.