Estimation of the Decision Inconsistency Index and the Across-Studies Inconsistency Index


How to use this tool:

1. This app allows the estimation of the Decision Inconsistency index (DI) and of the Across-Studies Inconsistency index (ASI) for any given meta-analysis.
2. Start by uploading a .csv file containing at least (i) one column with the effect sizes for each primary study, and (ii) one column with the respective variances. Each row should correspond to a different primary study. If you are doing meta-analysis of risk ratio, odds ratio or hazard ratio, please provide your data in the log-transformed form.
3. In the left panel, please indicate if you want to perform Bayesian or frequentist meta-analysis.
4. Then, please indicate the name of the columns containing the effect sizes and the variances, as well as the applied effect size measures.
5a. If you are doing meta-analysis of risk ratio, odds ratio or hazard ratio, please indicate the baseline risk, and either (i) directly the thresholds you want to apply (as absolute values) or (ii) the utility of the corresponding health state.
5b. If you are doing meta-analysis of mean differences or standardised mean differences, please indicate the thresholds you want to directly apply (as absolute values).




      

    

    

      

      
Want to try an example?

      

        Please download an example dataset here.
         This corresponds to the same example described in the reference paper. The effect size measure corresponds to the variable yi while the variance corresponds to the variable vi
      

      

      

      
Authors: Bernardo Sousa-Pinto, Ignacio Neumann, Rafael José Vieira, Antonio Bognanni, Manuel Marques-Cruz, Sara Gil-Mata, Simone Mordue, Clareece Nevill, Gianluca Baio, Paul Whaley, Guido Schwarzer, James Steele, Gavin Stewart, Holger J Schunemann, Luís Filipe Azevedo

      
Any question should be sent to Bernardo Sousa-Pinto (bernardo@med.up.pt), Faculty of Medicine, University of Porto, Porto, Portugal