The use of linear mixed models to estimate optimal vaccination timing for infectious bursal disease in broilers in paraguay

Abstract

The objectives of this study were (1) to fit linear mixed models for Maternally Derived Antibody (MDA) values for estimating optimal days of age for Infectious Bursal Disease (IBD) vaccination in broiler chicks and (2) to evaluate how optimal vaccination timing estimates varied, based on the field data collected in Paraguay. The MDA titres were measured by Enzyme-linked Immunosorbent Assay (ELISA) with sera collected from 20 chicks per flock (n = 14) at 1, 8 and 15 days of age. Both Restricted Maximum-likelihood (REML) and Markov Chain-Monte Carlo (MCMC) estimation methods were used to fit linear mixed models for the dependent variable log-transformed MDA. There is a slight dissimilarity of each estimate between the two models because of the difference in mathematical algorithms and handling method of missing data. The study flocks with the earliest and latest estimated optimal timing for vaccination had a mean days of age of 15.1 [95% Bayesian Credible Interval (BCI): 13.3-17.0] and 23.7 [95% BCI: 21.7-25.7], respectively. This variation could be partly explained by a limited understanding of the true biological variability representing the variety of factors affecting MDA in the study chicks. Although the results of the study can be used as a benchmark to establish IBD vaccination programmes in the study area, it is recommended for estimating optimal IBD vaccination timing to measure the MDA level by ELISA tests on a routine basis. © Asian Network for Scientific Information, 2009.