Correcting the completeness bias of observed prevalence

Aims and background: The cancer prevalence in given areas can be estimated on the basis of data supplied by cancer registries. As the obtained estimat e of prevalence depends on the length of the cancer registry's observation period, it is generally lower than the total prevalence in the considered a rea. In the present work we propose a method to calculate a correction fact or of this bias in order to obtain an approximation to the total prevalence . Methods & study design: The method is based on the relationship between rel ative survival and incidence by age for a specific cancer site. Results and conclusions: We provide values of the correction factor, the co mpleteness index R, relative to the most important cancer sites, for specif ic ages and periods of observation of the cancer registries in Italy, In ad dition, we provide indications for extended use of the index when substanti al variations from the basic pattern of relative survival are observed in p ractical situations. Furthermore, we give helpful suggestions to obtain app roximate values of the correction factor to be used for ages and periods of observation that are intermediate between the ones presented in this paper .