Quantitative Synthesis IInteractive Quiz Prepared for: The Agency for Healthcare Research and Quality (AHRQ) Training Modules for Systematic Reviews Methods Guide www.ahrq.gov
Binary Data For a meta-analysis of binary data, which of the following metrics is likely to be more statistically heterogeneous? • The odds ratio • The risk difference
Reasons for a Meta-analysis Which of the choices listed below is not a reason to carry out a meta-analysis? • Explore heterogeneity of results across studies • Reduce uncertainty in the estimate of treatment effect • Correct errors in previous publications
Fixed Effects ModelVersus Random Effects Model What is the key difference between the fixed effect model and the random effects model? • The spelling • The random effects model will always give a wider confidence interval • The fixed effect model considers only within-study variations
Assessing Study Conclusions This is part of a table from a meta-analysis of genetic factors that concludes: “polymorphisms of both GSTT1 and GSTP1 genes seem [to be] associated with elevated breast cancer risk in a race-specific manner.” This conclusion is based on a significant association among “non-Chinese,” versus a nonsignificant association among “Chinese.” Do you agree with the conclusion? • Yes, I agree • No, I do not agree CI = confidence interval; OR = odds ratio Sergentanis TN, Economopoulos KP. Breast Cancer Res Treat 2010;12:195-202.
CI = confidence interval F = fixed effects model OR = odds ratio Assessing Study Methods As described in the methods section of a paper, the authors opted to use random or fixed effects models according to whether a heterogeneity test was significant. Do you agree or disagree with their choice? • Yes • No Data on GSTP1 Sergentanis TN, Economopoulos KP. Breast Cancer Res Treat 2010;12:195-202.
Summary • It is important to critically appraise the methodological quality and reporting of information of studies included in meta-analyses. • In most meta-analyses, treatment effects across studies are more likely to be heterogeneous in the risk difference metric when compared with the risk ratio or the odds ratio. • The fixed effect model considers only the within-study variation, whereas the random effects model incorporates both within-study variation and between-study variations.
Authors • This quiz was prepared by Joseph Lau, M.D., and Thomas Trikalinos, M.D., Ph.D., members of the Tufts Medical Center Evidence-based Practice Center. • The information in this module is based on Chapter 9 in Version 1.0 of the Methods Guide for Comparative Effectiveness Reviews (available at: http://www.effectivehealthcare.ahrq.gov/repFiles/2007_10DraftMethodsGuide.pdf).