Withdrawal time estimation of veterinary drugs: extending the range of statistical methods.

In order to use a drug in a food producing animal, evidence has to be provided that after a certain withdrawal time, drug residues in tissues, such as muscle meat, fat, liver, kidney etc., are below a given maximum residue limit (MRL), for a majority of animals. Several statistical methods, both regression based and nonparametric based methods, have been proposed, each relying on different sets of assumptions, which may or may not hold for the specific data situation. The purpose of this paper is to enrich the range of methods, i.e. to provide approaches for situations where current methods are inappropriate. Bayesian methods, using Markov chain Monte Carlo, are proposed to derive inference on the parameters of interest.