Statistics Seminars: 'Bayesian Analysis of Misclassified Binary Data from Case-Control Studies with a Validation Sub-Study'
11 January 2002 14:00 in CM221
"An aim of many epidemiological studies is to assess the association between a binary exposure variable and the presence or absence of a disease. We suppose the exposure variable is subject to misclassification, perhaps due to forgetfulness or misreporting, but true exposure has been determined for a subset of the study group using an expensive "gold-standard" measure. A two-stage Bayesian method will be presented for analysing the study. The first stage analyses data from the validation sub-study group, whose exposure has been determined using both an inexact and the gold-standard measures. The second stage analyses data from the main study group for whom only the inexact measure has been used. The first stage is analytically tractable and MCMC methods are used for the second stage. The posterior distribution from the first stage becomes the prior distribution for the second stage, thus transferring all relevant information between the stages. Both unmatched case-control studies and matched case-control studies will be considered and the task of choosing an appropriate vague prior distribution will also be addressed."
Contact firstname.lastname@example.org for more information