Step 4: Subgroup analysis by covariate

Study-centric data is only available for the reviews that enter the pilot. If you are interested in learning more about the pilot and how to join, please contact revman@cochrane.org.

Let's assume that you have created an analysis (see Step 3) and you are now interested in doing a subgroup analysis based on whether or not there were dropouts in the included studies. With study-centric data you can enter the study characteristic 'Dropouts' as a covariate. Covariates are characteristics of studies that might influence the size of the intervention effect. They can be used to perform subgroup analysis. A covariate is defined in terms of categories. Categories of a covariate can be used to divide studies into subgroups. In our example the review-level covariate will be study dropout and the categories will be 'No dropouts' and 'Dropouts present'. The restriction is however that the studies only report one result per covariate in order to subgroup by it. We assume that all our studies are like that for this example. Currently this is the only type of subgrouping you can do with study centric data structures.

1. Set up the concepts (PICO)

The first step is to define the covariate and the categories hat you want for the subgrouping. Go back to the PICO page and click the Covariates tab.

Click Add Covariate button and type 'Dropouts' as the Name and define two Categories: 'No dropouts' and 'Dropouts present'.

2. Enter study results

Assuming that all of our (already defined) studies fall into one of the two categories, you can now go back to Studies in the navigation panel and choose Edit Study for the first one 'Herne 1980'. Select the Covariates tab and set the 'Dropouts' category to 'No dropouts'. Navigate to the next study using the Next button at the top and enter the covariates for other studies (Hoaglund 1950 - No dropouts; Kaiser 1996 - Dropouts present; Lexomboon 1971 - No dropouts; McKerrow 1961 - No dropouts; Taylor 1977 - Dropouts present).

3. Create the subgroup analysis

You're now ready to see the subgroup analysis results. Go back to our analysis group 'Antibiotics versus placebo' and expand the one you created previously called 'Persistent symptoms after 7 days'. Click on 'Add analysis' and select 'Study Data'.

Set up a subgroup analysis by selecting the PICO elements, including the subgroup factor (covariate) - in this example, dropouts. You can indicate in the name that this is a subgroup analysis.

When you click on the 'Graphs' tab, you will get the forest plot with the subgroup analysis by whether or not there were dropouts in the included studies.