Meta-analysis for orthodontists: Part I – How to choose effect measure and statistical model Spyridon N. Papageorgiou 1,2 1 Department of Orthodontics; 2 Department of Oral Technology, School of Dentistry, University of Bonn, Bonn, Germany Systematic reviews ideally provide a comprehensive and unbiased summary of existing evidence from clinical studies, whilst meta-analysis combines the results of these studies to produce an overall estimate. Collectively, this makes them invaluable for clinical decision-making. Although the number of published systematic reviews and meta-analyses in orthodontics has increased, questions are often raised about their methodological soundness. In this primer, the first steps of meta-analysis are discussed, namely the choice of an effect measure to express the results of included studies, and the choice of a statistical model for the meta-analysis. Clinical orthodontic examples are given to explain the various options available, the thought process behind the choice between them and their interpretation. Key words: Effect measure, orthodontics, meta-analysis, fixed-effect, random-effects Received 2 March 2014; accepted 5 June 2014 Introduction Systematic reviews summarise qualitatively the results of multiple studies, providing the highest grade of evidence to develop clinical guidelines. Meta-analysis goes one step further, to the quantitative synthesis of studies in order to provide an overall summary. Although the number of published systematic reviews and meta-analyses in orthodontics has increased, questions about the metho- dological soundness of some have been raised. In this paper, the first steps of meta-analysis will be discussed in the context of: (1) the choice of an effect measure to express the results of included studies and (2) selection of an appropriate model for subsequent statistical synthesis. For readers not familiar with meta-analysis, a short description of the forest plot and of between-study heterogeneity will also be given. Meta-analyses will be covered that compare two groups; either an intervention group versus a placebo/control group or an intervention group versus another intervention group. Meta-analyses of data from a single group (i.e. event rates, means of groups, etc.), meta-analyses of diagnostic accuracy or network meta-analyses will not be covered. Furthermore, emphasis will be given on basic principles, which can be easily applied in the freely-available program Review Manager (RevMan) (Review Manager, 2012) available from the Cochrane Collaboration. It is beyond the scope of this article to analytically describe the stages of a systematic review and when it is sensible to perform a meta-analysis or not. Interested readers should refer to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins and Green, 2011) and to the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta- Analyses) (Liberati et al., 2009), where further details can be found on how to respectively conduct and report systematic reviews and meta-analyses. Effect measures The first step in meta-analysis is to identify all relevant studies evaluating the same measure in order to directly compare and eventually synthesise them. There are a wide array of effect measures that can be used, but this paper will focus on five of the most common: (i) those for continuous outcomes (such as ANB angle): mean difference (MD) (also known as difference in the means) and standardised mean difference (SMD), (ii) those for binary (dichotomous) outcomes (such as bracket failure): odds ratio (OR), relative risk (RR) (also known as risk ratio) and risk difference (RD) (also known as absolute risk reduction). Continuous outcomes Effect measures. If all included studies measure the same continuous outcome and use the same measurement INVITATION TO SUBMIT Journal of Orthodontics, Vol. 41, 2014, 317–326 Address for correspondence: Spyridon N. Papageorgiou, DDS, Department of Orthodontics, School of Dentistry, University of Bonn, Welschnonnenstr. 17, 53111, Bonn, Germany. Email: snpapage@gmail.com # 2014 British Orthodontic Society DOI 10.1179/1465313314Y.0000000111