Estimating and generalizing with clustered sampling in general practice

Abstract

General practice research focusing on patients may involve risk factors, morbidity, medication use or patient satisfaction. When collecting information about patients it is often easier, cheaper and more appropriate to enlist the support of a number of general practitioners (GPs) who provide access to a number of patients. Such studies utilise a 'cluster sampling' (CS) design, as clusters or groups of patients around a GP are used for the investigation. In analysing data from these studies it is necessary to consider the impact that the study design will have on the variance structure of data collected.