ABSTRACT
124The sensitivity analysis is the study of influence and robustness of different statistical methodologies on the results of meta-analysis. In case of influence meta-analysis, the influence of each study can be estimated by deleting each in turn from the analysis and noting the degree to which the size and significance of the treatment effect changes. The subgroup analysis is the study of variation by different categories of patients on the results of meta-analysis. When the heterogeneity statistic is significant, it is not feasible to assume that the given treatment effect is same across different groups of patients. The subgroup analysis is carried out when there is indirect evidence suggesting considerable difference between categories. The subgroup analysis should best be used as hypothesis-generating tools. The usefulness of subgroup analysis may be limited due to small number of studies included in meta-analysis. The cumulative meta-analysis is a repeated performance of meta-analysis whenever a new relevant study is available for inclusion. This allows a retrospective identification of the patient in time whenever a treatment effect first reach conventional level of statistical significance. As in primary studies with regression, or multiple regressions, in meta-regression analysis, we assess the relationship between one or more covariates and a dependent variable. Meta-regression has become a commonly used tool for investigating whether clinical heterogeneity may explain the statistical heterogeneity. The meta-regression analysis is used to suggest reasons for statistical heterogeneity. The multiple meta-regressions can encompass two or more study characteristics simultaneously as independent variables. The main objectives of meta-cluster analysis is to find out which studies in a meta-analysis are similar and which studies dissimilar with respect to relevant variables related to the particular summary statistic of the analysis.
