Article Text

Download PDFPDF
Study size
  1. Angie Wade
  1. Department of Paediatric Epidemiology and Biostatistics, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK
  1. awade{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Research is only worth doing if it provides useful information. Medical research usually consists of studying groups of individuals with the aim of answering a predefined research question. Most commonly in the field of sexually transmitted infections (STI), the prevalence of a virus or abnormality is to be estimated or prevalences compared, either over time or between different groups of people. Alternatively, several therapies may be compared within a randomised controlled trial. One question that arises at the start of any study is “How many individuals should be included in this study?” There are several ways of answering this question. The number included may be based on practical issues—for instance, the length of time available to the researcher together with the time taken to recruit, treat, and test each individual and the expected patient accrual rate. These factors will vary from researcher to researcher and between different sources of patients—for example, accrual rates will differ between different hospitals. To use such variable quantities to determine the number needed to effectively answer a given research question is clearly flawed. Ethically it is wrong to either underrecruit or overrecruit. On the one hand we may be left with insufficient numbers to conclusively answer the question. On the other hand, if we overrecruit, then the best scenario is not only do we waste time, but also we subject more individuals than necessary to any inconvenience associated with being studied. In the worst scenario, we may be allowing individuals to receive inferior treatment after sufficient numbers have been recruited to ensure that the best treatment is known.

Many researchers associate sample size calculation purely with randomised controlled trials. Most of the studies presented in this journal do not fall into this category. However sample size estimation before study commencement is important for all types …

View Full Text