Metaanalysis: methodology, biostatistic rules, application in onco-hematology

Metaanalysis are proposed in order to increase statistical power and draw better conclusion about the efficacy of a treatment. After inclusion of trials corresponding to an objective, a statistical procedure is performed permitting to obtain an estimate of the common effect with the confidence interval after model validation. The main problem is posed by the selection of trials in order to avoid positive bias linked to the greater number of positive published studies. Application of metaanalysis in oncohematology should be proposed in case of weakness size of population in the individual trials.