Veterinary surgeons, nurses, and owners and farmers make clinical decisions every day that affect the health and welfare of the animals they care for. A large number of these decisions are focused on medical or surgical treatments. A recent study by Robinson and others (2017) found that for small animal consultations, therapeutic treatments were the most common (40.5 per cent, 1286 of 3192 problems). In an ideal situation the most efficacious treatment would be used if it is practical, ethical and economic to implement. To be able to do this clinical decision makers need a robust relevant evidence base. Evidence-based veterinary medicine has been defined as the ‘use of best relevant evidence in conjunction with clinical expertise to make the best possible decision about a veterinary patient’ (Centre for Evidence Based Veterinary Medicine [CEVM] 2017a). Therefore, to make evidence-based treatment decisions a relevant, ‘strong’ evidence base about the efficacy of all the possible treatments is required. Evidence is completely useless if it is not relevant to the population of animals that we treat in practice. Only a clinician can decide if the population in the paper they are reading is similar enough to the patients they see in practice and whether the findings can be applied to their clinical decisions. For many clinical questions about the treatments available in veterinary practice the relevant evidence base is very limited. Once it has been established that a paper concerning treatment is relevant to practice, the study design must be identified to see if it is suitable to provide evidence of efficacy. The ‘strongest’ (least subject to bias, see later) study design that can be used to compare the efficacy of two or more treatments is …