| Abstract : | Meta-analysis is a statistical procedure for combining information from alreadyexisting studies. In the last thirty years, meta-analysis met a rapid increase in themedical world due to greater emphasis of evidence-based medicine and the need formore reliable findings. Most meta-analyses within the field of medical research havebeen conducted on clinical trials and especially on randomized controlled trials. De-spite the development of meta-analytic techniques, the meta-analysis of randomizedcontrolled trials with the occurrence of zero events is still a statistical challenge.The current thesis aims to introduce the general concept of meta-analysis andfocuses on meta-analysis of randomized controlled trials with the occurrence of zeroevents. After we review the basic ideas of meta-analysis, we will examine the classicmeta-analytic methods in the presence of zero events. Then, we will proceed tothe presentation of alternative approaches. A real dataset with zero events, whichinvestigates the risk for major adverse cardiovascular events in patients treated withanti-IL-12/23 biological agents for chronic plaque psoriasis will be used for analysis.The objective is to apply the methods under investigation to the real dataset andevaluate their behaviour. Afterwards, we will compare the performance of somemeta-analytic methods in a simulation study. Finally, the conclusions of this thesiswill be discussed.
|
|---|
Copyright © 2013 Library AUEB. All rights reserved.