Article type
Year
Abstract
Comprehensive Meta Analysis is a computer program for meta-analysis that was developed under a grant from the National Institutes of Health. The program, which runs under Windows 95© or Microsoft NT© incorporates a wide array of features including the following.
Database. The program allows the user to enter data for any number of citations. For each citation, the user may enter data enter data for one or more outcomes (e.g. for disease recurrence and also for mortality) and for one or more subgroups. Optionally, the user may import the full text of the abstract.
Outcomes. The program will accept a variety of effect size indices, including indices based on a Mean difference. Correlation, Rate difference, Relative risk, or Odds Ratio. For each type of index the program will accept data in a variety of formats, and will perform all conversions automatically.
Sub-groups. The program allows the user to define any number of variables, enter data for these variables, and include these variables in the analyses. For example, the researcher might elect to code patient type (acute or chronic) for each study. The program could then generate graphs, source tables and diagnostics that show how effect size varies as a function of patient type.
Analyses. The program allows the user to define (and save) any number of meta-analysis. Each analysis may be based on any subset of the database and is linked dynamically so that any changes to the data are reflected automatically in the analyses.
Graphs. The program features a wide selection of graphs including graphs that show the contribution of each study to the combined effect; graphs that show how effect size varies as a function of categorical or continuous variables, and graphs of cumulative meta-analyses. The program includes a sophisticated graphics engine for producing publication-quality prints and slides.
Computational options. The program includes a wide variety of computational options for both fixed effects and random effects models. Additionally, the program allows the user to view all details of the computations.
Database. The program allows the user to enter data for any number of citations. For each citation, the user may enter data enter data for one or more outcomes (e.g. for disease recurrence and also for mortality) and for one or more subgroups. Optionally, the user may import the full text of the abstract.
Outcomes. The program will accept a variety of effect size indices, including indices based on a Mean difference. Correlation, Rate difference, Relative risk, or Odds Ratio. For each type of index the program will accept data in a variety of formats, and will perform all conversions automatically.
Sub-groups. The program allows the user to define any number of variables, enter data for these variables, and include these variables in the analyses. For example, the researcher might elect to code patient type (acute or chronic) for each study. The program could then generate graphs, source tables and diagnostics that show how effect size varies as a function of patient type.
Analyses. The program allows the user to define (and save) any number of meta-analysis. Each analysis may be based on any subset of the database and is linked dynamically so that any changes to the data are reflected automatically in the analyses.
Graphs. The program features a wide selection of graphs including graphs that show the contribution of each study to the combined effect; graphs that show how effect size varies as a function of categorical or continuous variables, and graphs of cumulative meta-analyses. The program includes a sophisticated graphics engine for producing publication-quality prints and slides.
Computational options. The program includes a wide variety of computational options for both fixed effects and random effects models. Additionally, the program allows the user to view all details of the computations.