Article Text

  1. G M Fitzmaurice1,2,3
  1. 1Department of Biostatistics, Harvard School of Public Health, USA,
  2. 2Department of Psychiatry, Harvard Medical School, USA,
  3. 3Laboratory for Psychiatric Biostatistics, McLean Hospital, Belmont, Massachusetts, USA


Objective: A multicentre study is a single study involving several centres (or investigators) where the combined data contributed by all centres is intended to be analysed as a whole. In general, multicentre studies are conducted at several centres either to increase the number of participants that would otherwise be available at any single centre, or to assess the effectiveness of the treatment or intervention across a broader range of settings. By performing the study in multiple centres, and often in multiple countries, a multicentre study design allows for the sampling of a broader patient population and the completion of the study within a reasonable timeframe. Typically, differences between centres (eg, different study investigators, differences in the centre patient populations, differences in medical care practices among centres) may result in extra variability in the treatment effect estimates. This extra variation must be properly accounted for in the analysis. In this presentation, we consider some alternative statistical methods for the analysis of data collected from multicentre studies. We highlight important distinctions between alternative methods and make some recommendations concerning appropriate methods of analysis.

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.