Creating the analyses

Study centric data management is available by default for new Cochrane reviews from April 25th 2023.  

• Authors with ongoing protocols should consider using study centric data management given the benefits listed on the intro page.
• Authors that have started the data extraction stage or who are updating a review may wish to use study centric data management, but it will require completed work to be repeated to ensure the correct setup. 

1. Types of analyses in RevMan

There are two ways to set up an analysis in RevMan: 

1. Using study data: All studies with results for the selected outcome and interventions will automatically be included in the analysis. Data are entered and stored at the included study level and can be re-used across different analyses.
2. Using custom input: Manually choose which studies are included in the analysis. Data are entered and stored specifically for this analysis and cannot be re-used across different analyses. See here for more details.

This page details the steps to create analyses using study data.

2. How are study centric data analyses generated?

Study centric data analyses are generated by RevMan based on the criteria you set up in each analysis. Depending on the data source, outcome, interventions, subgroup or filter for sensitivity, and the available data, RevMan will identify which results to include. To optimize the amount of data to include, RevMan will automatically transform result data with the following operations:

• Combine intervention arms when two or more intervention arms map to the interventions selected for the analysis (available for both contrast and arm level data)
• Transform arm level data to contrast level data when including both arm and contrast level data in an analysis
• Switch reference arm when a study reports the contrasts A-B and A-C were reported and you are interested in the contrast B-C.

Note that a result will only be included if there is enough data available. For some operations the covariance is required.

View a more complete list of the business rules applied to generate study centric analyses, and the calculations required to do so. 

Please note it is currently not possible to add footnotes to study centric data analyses.

3. Add an Analysis using study data

After setting up the review criteria (see step 1), it is recommended that authors set up their analyses at the protocol stage before entering any study results data. This enables you to check that you have included the Interventions, Intervention groupings, Outcomes and Covariates within the Review criteria correctly. It also means that once you import your results data, your analyses will be automatically populated (a key benefit of using study centric data management).  

First, set up an Analysis group as described here (create an Analysis group, name the Analysis group and create an analysis). 

Go back to the main Analyses page, expand the panel for the Analysis group that you just created by clicking on its name. Then click Add Analysis. You are prompted whether you would like to create an analysis based on study data or custom input. Choose Study Data. 

4. Define the name of the new analysis and data source

On the new Analysis page, first enter a name (in this case it makes sense to just enter the outcome name) and select the data source. The data source relates to the types of results data to include in your analyses, with the options: 

• Only arm-level data
• Only contrast-level data
• Contrast- and arm-level data (preferring arm-level data where both exist)
• Contrast- and arm-level data (preferring contrast-level data where both exist)  

More information about arm level and contrast level data is available and Section 5.3.6 of the Cochrane Handbook for Systematic Reviews of Interventions provides more guidance on the different types of data and in what circumstances either might be preferred in your analyses. 

5. Set the Synthesis PICO

On the new Analysis page, under Synthesis PICO, choose your: 

• Outcome
• Intervention grouping (for broader synthesis questions), experimental intervention (for narrower synthesis questions) and control intervention. 

The example below shows a broad synthesis question using intervention groupings. It includes 'Persistent symptoms after 7 days' as the Outcome, 'Antibiotics' (all antibiotics) as the Experimental Intervention and 'Placebo' as the Control Intervention. None of the following examples subgroup or filter for sensitivity analyses, so these options are left as default. 

The example below shows a narrower synthesis question using the more granular interventions based on a specific antibiotic versus placebo. It includes 'Persistent symptoms after 7 days' as the Outcome, 'Tetracycline' as the Experimental Intervention and 'Placebo' as the Control Intervention. 

The example below shows a narrower synthesis question using the more granular interventions based on antibiotic coverage. It includes 'Persistent symptoms after 7 days' as the Outcome, ' Broad-spectrum antibiotic' as the Experimental Intervention and ' Narrow spectrum antibiotics' as the Control Intervention.  

6. Check your Analyses

Once your Analyses are set up and you have entered your study results (step 2), you can now check your Analyses. 

On the main Analyses page, select the Analysis you want to check. You will notice the analysis table in now populated with the number of studies, participants and analyses data. Click Edit Analysis and select the Graph tab.  

For the example syntheses in point 5 above (setting up the Synthesis PICO) above, here are the example forest plots if you navigate to the Graph tab. 

The example below shows a broad synthesis question using intervention groupings. It includes 'Persistent symptoms after 7 days' as the Outcome, 'Antibiotics' (all antibiotics) as the Experimental Intervention and 'Placebo' as the Control Intervention.  

The example below shows a narrower synthesis question using the more granular interventions based on a specific antibiotic versus placebo. It includes 'Persistent symptoms after 7 days' as the Outcome, 'Tetracycline' as the Experimental Intervention and 'Placebo' as the Control Intervention. 

The example below shows a narrower synthesis question using the more granular interventions based on antibiotic coverage. It includes 'Persistent symptoms after 7 days' as the Outcome, ' Broad-spectrum antibiotic' as the Experimental Intervention and ' Narrow spectrum antibiotics' as the Control Intervention. 

4. How to use the covariance calculator

If your analysis should include results where RevMan needs to switch reference arms or combine arms for contrast-level data then you will need to enter a covariance for contrasts.

If you already have the covariance for the contrasts you can enter it directly. If you do not have the covariance the calculator can help you calculate it using three different methods depending on the information you have available.

Click on the calculator icon to open the covariance calculator.

Select tab depending on what information is available: ‘Data on another contrast’, ‘Data on the correlation between the contrasts’ or if there is no additional data, and you can assume the variance within the arms of the study are similar, you can choose to use the ‘Approximation’.

Enter the data required to calculate a covariance according to the three options available.

Data on another contrast

Let’s assume you have the standard error for the contrast Aureomycin vs Tetracycline (SE: 0.5). Select the tab ‘Data on another contrast’, select the contrast and enter the standard error. Note that any information on this contrast that is enough to calculate the variance will also do. 

Data on the correlation between contrasts

If we instead assume that you have data on the correlation between the contrast, you select the tab ‘Data on the correlation between the contrasts’, select the contrasts and enter the correlation. For this study select Contrast 1 Auremycin vs Placebo and Contrast 2 Tetracycline vs Placebo and the Correlation 0.59.

Approximation

If you do not have any additional data and you can assume that the variance within the arms of the study are similar, as a last resort you can use a method of approximation to calculate the Covariance.

For any of these options you will see that if the calculated covariance is not consistent with the standard errors given in the results table, you will be notified and you will not be able to update the Covariance. 

When you have calculated the Covariance, click on ‘Update covariance’ to save the covariance with the result.

Now that you have added the Covariance you can go back to the analysis and check that the study-result is calculated and included in the analysis for the study ‘Hoaglund 1950’.