An application of hidden Markov models to the French variant Creutzfeldt-Jakob disease epidemic

In 1996, the discovery of variant Creutzfeldt-Jakob disease in the UK raised serious concerns about a large-scale epidemic. These concerns have been heightened by the recent discovery of people in Britain who were infected through blood transfusion. The outbreak of variant Creutzfeldt-Jakob disease in France emerged more recently with 23 cases observed to date. We use a hidden Markov model to predict the scale of the epidemic in France. As accurate data on the most important epidemiological parameters are scarce, we incorporate estimates from previous studies. Parameter estimation is performed by using Markov chain Monte Carlo methods from which credible intervals for our predictions are obtained. The sensitivity of these predictions to important assumptions regarding population exposure is assessed. Copyright (c) 2010 Royal Statistical Society.