Evaluation of the power for meta-analyses

Meta-analysis is a method used to compound independent studies to explain their diversity and to ensure more reliable and accurate results. The most important issue, which is a sign of quality in meta-analyses, is power, and this power is assessed under two approaches: Fixed effect or random effect model with the existence of heterogeneity. Evaluation of power is influenced by effect size, sample size at smallest study and number of studies. It was aimed to determine the minimum sample size and number of studies in order to reach an acceptable power in this study. The mean and standard deviations of power values for different number of studies (2, 5, 7, 10, 15 and 18) and minimum sample sizes were determined with 1000 replication simulation studies taking into consideration different effect sizes (0.1-0.2, 0.20001-0.3, 0.30001-0.4, 0.40001-0.5 and 0.50001-0.7), different case-control rates (1:1, 1:2 and 1:3) under fixed and low, medium and high heterogeneous type random effect models. For the power evaluations obtained with the simulated study, the fixed effect, low, medium and high heterogeneous random effects models, the number of samples required was found to be higher for the models with more heterogeneity. Deterioration in case-control rates also increased sample size required for good power. By taking into consideration the factors that affect the power researchers can be guided to reach an acceptable quality in their studies.