To evaluate the relative efficacy of an intervention, we typically come across a plethora of randomized control trials. These trials will most probably differ on their sample size, quality, risk of bias and results. We need to make a synopsis of all the available information, and this is achieved with the method of the systematic review, which identifies and evaluates all relevant studies satisfying a-priori defined eligibility criteria. Ideally, we would proceed with a quantitative synthesis of the trials’ results through the statistical method of meta-analysis. A systematic review with meta-analysis lies on top of the pyramid of evidence-based medicine and many international organizations (such as WHO) suggest its use for evaluating the efficacy of an intervention. Meta-analysis has a lot of benefits including an increase in precision and exploring heterogeneity of treatment effects. We can also proceed to subgroup or meta-regression analyses to further explore heterogeneity of treatment effects. We will present the most commonly used meta-analyses models with their benefits, limitations and criticisms.
Introduction to meta-analysis
See also
-
Scientific challenges and problems in football analytics
-
Estimating NBA players salary share according to their performance on court: A machine learning approach
-
Input-Biased Technical Progress and the Aggregate Elasticity of Substitution: Evidence from 14 EU Member States
-
Production Structure, International Trade and Carbon Footprint
