Host animal efficacy studies are a natural situation for the consideration of the use of mixed models. Random effects associated with the design structure, which reflect the hierarchy associated with units of housing, animal sources, etc., are generally incorporated in the analysis of the study data from these studies. Recently, the use of generalized linear mixed models (GLMMs) has become more common as these models have the ability to model the design structure in a manner that reflects restrictions in the study randomization for non-normal data. In this presentation, we will identify key considerations in designing studies in Veterinary Biologics, with an emphasis on non-normal outcomes. These considerations include the statistic used to estimate efficacy, the multiple objectives associated with some studies, and a look at the appropriateness of the use of GLMMs for some study design and treatment structures.