Modelling the trend of bovine spongiform encephalopathy prevalence in France: Use of restricted cubic spline regression in age-period-cohort models to estimate the efficiency of control measures.

An age-period-cohort (APC) analysis was used to assess the trend in prevalence of bovine spongiform encephalopathy (BSE) in France over time in relation to the control measures adopted since onset of the epidemic. Restricted cubic regression splines were used to model the functional forms of the non-linear effects of age at screening, birth cohort and date of diagnosis of the tested animals. The data of the 2001-2007 period of surveillance was analysed using 1-year categorisation. A categorical analysis was performed as control to check the accuracy of the sets of knots in the spline models, which were selected according to the Akaike Information Criterion (AIC). Knot selection was based on a priori knowledge of the disease and the dates of implementation of the five main BSE control measures. It was assumed that disease prevalence was a function of exposure to BSE and that changes in the exposure of cattle to BSE were mainly due to the control measures. The effects of the five main control measures were discussed in relation to the trend in BSE risk for the successive birth cohorts. The six selected models confirmed that all measures participated in disease control. However, characterization of the respective effect of individual measures was not straightforward due to the very low disease prevalence, incompletely tested cohorts and probably cumulative and overlapping effects of successive measures. The ban of importation of meat and bone meal (MBM) from the UK and the ban of use of MBM in bovines were insufficient to control the epidemic. The decline in the BSE epidemic more likely originated from implementation of the ban of MBM use in all ruminants in 1994, whose effect was probably reinforced by the evolution in perception of the BSE risk following evidence of BSE transmission to humans. Finally, the respective effects of the last two measures (prohibition of the use of specific risk material in 1996 and total MBM ban in 2000) could not be characterized as they were implemented when the disease trend was already in strong decline.