Statistical issues in the design and analysis of animal health interventions for vector-borne disease

Methods for determining the efficacy of animal health interventions involving vector-borne diseases when the intervention attempts to control the vector itself are described. There are a number of possibilities. For example, a before/after study can be undertaken when data collected following the introduction of the intervention are compared with data collected earlier. An untreated area might also be selected, where vector control is not applied, in order to adjust for uncontrollable annual sources of variation. If several herds can be monitored then individual changes in herd performance can be regressed against corresponding individual changes in herd disease prevalence. Vector control studies, however, are expensive, and each of these methods has its drawbacks. An alternative solution might be to extrapolate from baseline information on the effects of disease on the performance of individual animals to model the potential effect of vector control. These different approaches are illustrated in a study of tsetse control in South West Ethiopia