A Bayesian back-calculation method to estimate the risk of bovine spongiform encephalopathy (BSE) in Canada during the period 1996–2011

Abstract

ABSTRACTSeventeen typical cases of bovine spongiform encephalopathy (BSE) were detected in Canada the period of 2003-2011. The clinical incidence of BSE was censored by early slaughter, death, or exportation of infected cattle due to the long incubation period of BSE disease. The aim of this study was to estimate the infection incidence of BSE in birth cohorts during 1996–2004 and project infection frequency through to 2007. An estimate of the number of asymptomatic infected cattle slaughtered for human consumption is also provided. The number of incident, asymptomatic cases was assumed to follow a Poisson process. A Bayesian back-calculation approach was used to project the risk of contracting BSE in those birth cohorts. Model parameters and inputs were taken from scientific literature and governmental data sources. The projected number of infected cattle in birth cohorts spanning the period 1996–2007 was 492, with median 95% credible interval 258–830. If the requirement to remove specified risk material (SRM) from cattle prior to entering the food chain was not in place, the predicted number of slaughtered infected in the human food chain from 1996–2010 was 298, with a 95% credible interval 156–500. The magnitude of the BSE epidemic in Canada for 1996–2007 birth cohorts was estimated to be approximately 28-fold higher than the number of clinical cases detected through to October 2011. Although some of those cattle were slaughtered for human consumption, the requirement of SRM removal may have prevented most of the infectious material from entering the food chain.