A test for the radomness of the occurrence of a disease trait in familial or other similar ordered sequences of epidemiological data.

摘要

Through the construction of a probability distribution, a test is proposed for the randomness of the occurrence of an epidemiological event in a series of structured ordered sequences. Such sequences may be of persons admitted to hospital in some specified order, individuals seen in a large set of extended screening program, or individuals in a large set of extended genealogies. Modifications to account for the limited length of ascertained sequences are also given. The method will be particularly important in the first-stage evaluation of genealogical data to indicate the existence and degree of familial aggregation of diseases such as cancer and to justify more complex methods of analysis designed to evaluate the specific elements of causation that may be operating. The statistical power of this approach is also derived, considering the anticipated sample sizes in its application, to detect some forms of familial aggregation in a large genealogical data base from Laredo, Texas.